Merge pull request #1630 from minrk/mergekernel

Merge divergent Kernel implementations

IPython/zmq/serialize.py

@@ -0,0 +1,179 b''
+"""serialization utilities for apply messages
+
+Authors:
+
+* Min RK
+"""
+#-----------------------------------------------------------------------------
+#  Copyright (C) 2010-2011  The IPython Development Team
+#
+#  Distributed under the terms of the BSD License.  The full license is in
+#  the file COPYING, distributed as part of this software.
+#-----------------------------------------------------------------------------
+
+#-----------------------------------------------------------------------------
+# Imports
+#-----------------------------------------------------------------------------
+
+# Standard library imports
+import logging
+import os
+import re
+import socket
+import sys
+
+try:
+    import cPickle
+    pickle = cPickle
+except:
+    cPickle = None
+    import pickle
+
+
+# IPython imports
+from IPython.utils import py3compat
+from IPython.utils.pickleutil import can, uncan, canSequence, uncanSequence
+from IPython.utils.newserialized import serialize, unserialize
+
+if py3compat.PY3:
+    buffer = memoryview
+
+#-----------------------------------------------------------------------------
+# Serialization Functions
+#-----------------------------------------------------------------------------
+
+def serialize_object(obj, threshold=64e-6):
+    """Serialize an object into a list of sendable buffers.
+    
+    Parameters
+    ----------
+    
+    obj : object
+        The object to be serialized
+    threshold : float
+        The threshold for not double-pickling the content.
+        
+    
+    Returns
+    -------
+    ('pmd', [bufs]) :
+        where pmd is the pickled metadata wrapper,
+        bufs is a list of data buffers
+    """
+    databuffers = []
+    if isinstance(obj, (list, tuple)):
+        clist = canSequence(obj)
+        slist = map(serialize, clist)
+        for s in slist:
+            if s.typeDescriptor in ('buffer', 'ndarray') or s.getDataSize() > threshold:
+                databuffers.append(s.getData())
+                s.data = None
+        return pickle.dumps(slist,-1), databuffers
+    elif isinstance(obj, dict):
+        sobj = {}
+        for k in sorted(obj.iterkeys()):
+            s = serialize(can(obj[k]))
+            if s.typeDescriptor in ('buffer', 'ndarray') or s.getDataSize() > threshold:
+                databuffers.append(s.getData())
+                s.data = None
+            sobj[k] = s
+        return pickle.dumps(sobj,-1),databuffers
+    else:
+        s = serialize(can(obj))
+        if s.typeDescriptor in ('buffer', 'ndarray') or s.getDataSize() > threshold:
+            databuffers.append(s.getData())
+            s.data = None
+        return pickle.dumps(s,-1),databuffers
+            
+        
+def unserialize_object(bufs):
+    """reconstruct an object serialized by serialize_object from data buffers."""
+    bufs = list(bufs)
+    sobj = pickle.loads(bufs.pop(0))
+    if isinstance(sobj, (list, tuple)):
+        for s in sobj:
+            if s.data is None:
+                s.data = bufs.pop(0)
+        return uncanSequence(map(unserialize, sobj)), bufs
+    elif isinstance(sobj, dict):
+        newobj = {}
+        for k in sorted(sobj.iterkeys()):
+            s = sobj[k]
+            if s.data is None:
+                s.data = bufs.pop(0)
+            newobj[k] = uncan(unserialize(s))
+        return newobj, bufs
+    else:
+        if sobj.data is None:
+            sobj.data = bufs.pop(0)
+        return uncan(unserialize(sobj)), bufs
+
+def pack_apply_message(f, args, kwargs, threshold=64e-6):
+    """pack up a function, args, and kwargs to be sent over the wire
+    as a series of buffers. Any object whose data is larger than `threshold`
+    will not have their data copied (currently only numpy arrays support zero-copy)"""
+    msg = [pickle.dumps(can(f),-1)]
+    databuffers = [] # for large objects
+    sargs, bufs = serialize_object(args,threshold)
+    msg.append(sargs)
+    databuffers.extend(bufs)
+    skwargs, bufs = serialize_object(kwargs,threshold)
+    msg.append(skwargs)
+    databuffers.extend(bufs)
+    msg.extend(databuffers)
+    return msg
+
+def unpack_apply_message(bufs, g=None, copy=True):
+    """unpack f,args,kwargs from buffers packed by pack_apply_message()
+    Returns: original f,args,kwargs"""
+    bufs = list(bufs) # allow us to pop
+    assert len(bufs) >= 3, "not enough buffers!"
+    if not copy:
+        for i in range(3):
+            bufs[i] = bufs[i].bytes
+    cf = pickle.loads(bufs.pop(0))
+    sargs = list(pickle.loads(bufs.pop(0)))
+    skwargs = dict(pickle.loads(bufs.pop(0)))
+    # print sargs, skwargs
+    f = uncan(cf, g)
+    for sa in sargs:
+        if sa.data is None:
+            m = bufs.pop(0)
+            if sa.getTypeDescriptor() in ('buffer', 'ndarray'):
+                # always use a buffer, until memoryviews get sorted out
+                sa.data = buffer(m)
+                # disable memoryview support
+                # if copy:
+                #     sa.data = buffer(m)
+                # else:
+                #     sa.data = m.buffer
+            else:
+                if copy:
+                    sa.data = m
+                else:
+                    sa.data = m.bytes
+    
+    args = uncanSequence(map(unserialize, sargs), g)
+    kwargs = {}
+    for k in sorted(skwargs.iterkeys()):
+        sa = skwargs[k]
+        if sa.data is None:
+            m = bufs.pop(0)
+            if sa.getTypeDescriptor() in ('buffer', 'ndarray'):
+                # always use a buffer, until memoryviews get sorted out
+                sa.data = buffer(m)
+                # disable memoryview support
+                # if copy:
+                #     sa.data = buffer(m)
+                # else:
+                #     sa.data = m.buffer
+            else:
+                if copy:
+                    sa.data = m
+                else:
+                    sa.data = m.bytes
+
+        kwargs[k] = uncan(unserialize(sa), g)
+    
+    return f,args,kwargs
+

IPython/config/application.py

@@ -1,508 +1,511 b''
 # encoding: utf-8
 """
 A base class for a configurable application.
 
 Authors:
 
 * Brian Granger
 * Min RK
 """
 
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2008-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 import logging
 import os
 import re
 import sys
 from copy import deepcopy
 from collections import defaultdict
 
 from IPython.external.decorator import decorator
 
 from IPython.config.configurable import SingletonConfigurable
 from IPython.config.loader import (
     KVArgParseConfigLoader, PyFileConfigLoader, Config, ArgumentError, ConfigFileNotFound,
 )
 
 from IPython.utils.traitlets import (
     Unicode, List, Enum, Dict, Instance, TraitError
 )
 from IPython.utils.importstring import import_item
 from IPython.utils.text import indent, wrap_paragraphs, dedent
 
 #-----------------------------------------------------------------------------
 # function for re-wrapping a helpstring
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Descriptions for the various sections
 #-----------------------------------------------------------------------------
 
 # merge flags&aliases into options
 option_description = """
 Arguments that take values are actually convenience aliases to full
 Configurables, whose aliases are listed on the help line. For more information
 on full configurables, see '--help-all'.
 """.strip() # trim newlines of front and back
 
 keyvalue_description = """
 Parameters are set from command-line arguments of the form:
 `--Class.trait=value`.
 This line is evaluated in Python, so simple expressions are allowed, e.g.::
 `--C.a='range(3)'` For setting C.a=[0,1,2].
 """.strip() # trim newlines of front and back
 
 subcommand_description = """
 Subcommands are launched as `{app} cmd [args]`. For information on using
 subcommand 'cmd', do: `{app} cmd -h`.
 """.strip().format(app=os.path.basename(sys.argv[0]))
 # get running program name
 
 #-----------------------------------------------------------------------------
 # Application class
 #-----------------------------------------------------------------------------
 
 @decorator
 def catch_config_error(method, app, *args, **kwargs):
     """Method decorator for catching invalid config (Trait/ArgumentErrors) during init.
 
     On a TraitError (generally caused by bad config), this will print the trait's
     message, and exit the app.
     
     For use on init methods, to prevent invoking excepthook on invalid input.
     """
     try:
         return method(app, *args, **kwargs)
     except (TraitError, ArgumentError) as e:
         app.print_description()
         app.print_help()
         app.print_examples()
         app.log.fatal("Bad config encountered during initialization:")
         app.log.fatal(str(e))
         app.log.debug("Config at the time: %s", app.config)
         app.exit(1)
 
 
 class ApplicationError(Exception):
     pass
 
 
 class Application(SingletonConfigurable):
     """A singleton application with full configuration support."""
 
     # The name of the application, will usually match the name of the command
     # line application
     name = Unicode(u'application')
 
     # The description of the application that is printed at the beginning
     # of the help.
     description = Unicode(u'This is an application.')
     # default section descriptions
     option_description = Unicode(option_description)
     keyvalue_description = Unicode(keyvalue_description)
     subcommand_description = Unicode(subcommand_description)
 
     # The usage and example string that goes at the end of the help string.
     examples = Unicode()
 
     # A sequence of Configurable subclasses whose config=True attributes will
     # be exposed at the command line.
     classes = List([])
 
     # The version string of this application.
     version = Unicode(u'0.0')
 
     # The log level for the application
     log_level = Enum((0,10,20,30,40,50,'DEBUG','INFO','WARN','ERROR','CRITICAL'),
                     default_value=logging.WARN,
                     config=True,
                     help="Set the log level by value or name.")
     def _log_level_changed(self, name, old, new):
         """Adjust the log level when log_level is set."""
         if isinstance(new, basestring):
             new = getattr(logging, new)
             self.log_level = new
         self.log.setLevel(new)
+    
+    log_format = Unicode("[%(name)s] %(message)s", config=True,
+        help="The Logging format template",
+    )
+    log = Instance(logging.Logger)
+    def _log_default(self):
+        """Start logging for this application.
+
+        The default is to log to stdout using a StreaHandler. The log level
+        starts at loggin.WARN, but this can be adjusted by setting the
+        ``log_level`` attribute.
+        """
+        log = logging.getLogger(self.__class__.__name__)
+        log.setLevel(self.log_level)
+        if sys.executable.endswith('pythonw.exe'):
+            # this should really go to a file, but file-logging is only
+            # hooked up in parallel applications
+            _log_handler = logging.StreamHandler(open(os.devnull, 'w'))
+        else:
+            _log_handler = logging.StreamHandler()
+        _log_formatter = logging.Formatter(self.log_format)
+        _log_handler.setFormatter(_log_formatter)
+        log.addHandler(_log_handler)
+        return log
 
     # the alias map for configurables
     aliases = Dict({'log-level' : 'Application.log_level'})
 
     # flags for loading Configurables or store_const style flags
     # flags are loaded from this dict by '--key' flags
     # this must be a dict of two-tuples, the first element being the Config/dict
     # and the second being the help string for the flag
     flags = Dict()
     def _flags_changed(self, name, old, new):
         """ensure flags dict is valid"""
         for key,value in new.iteritems():
             assert len(value) == 2, "Bad flag: %r:%s"%(key,value)
             assert isinstance(value[0], (dict, Config)), "Bad flag: %r:%s"%(key,value)
             assert isinstance(value[1], basestring), "Bad flag: %r:%s"%(key,value)
 
 
     # subcommands for launching other applications
     # if this is not empty, this will be a parent Application
     # this must be a dict of two-tuples,
     # the first element being the application class/import string
     # and the second being the help string for the subcommand
     subcommands = Dict()
     # parse_command_line will initialize a subapp, if requested
     subapp = Instance('IPython.config.application.Application', allow_none=True)
 
     # extra command-line arguments that don't set config values
     extra_args = List(Unicode)
 
 
     def __init__(self, **kwargs):
         SingletonConfigurable.__init__(self, **kwargs)
         # Ensure my class is in self.classes, so my attributes appear in command line
         # options and config files.
         if self.__class__ not in self.classes:
             self.classes.insert(0, self.__class__)
 
-        self.init_logging()
-
     def _config_changed(self, name, old, new):
         SingletonConfigurable._config_changed(self, name, old, new)
         self.log.debug('Config changed:')
         self.log.debug(repr(new))
 
-    def init_logging(self):
-        """Start logging for this application.
-
-        The default is to log to stdout using a StreaHandler. The log level
-        starts at loggin.WARN, but this can be adjusted by setting the
-        ``log_level`` attribute.
-        """
-        self.log = logging.getLogger(self.__class__.__name__)
-        self.log.setLevel(self.log_level)
-        if sys.executable.endswith('pythonw.exe'):
-            # this should really go to a file, but file-logging is only
-            # hooked up in parallel applications
-            self._log_handler = logging.StreamHandler(open(os.devnull, 'w'))
-        else:
-            self._log_handler = logging.StreamHandler()
-        self._log_formatter = logging.Formatter("[%(name)s] %(message)s")
-        self._log_handler.setFormatter(self._log_formatter)
-        self.log.addHandler(self._log_handler)
-
     @catch_config_error
     def initialize(self, argv=None):
         """Do the basic steps to configure me.
 
         Override in subclasses.
         """
         self.parse_command_line(argv)
 
 
     def start(self):
         """Start the app mainloop.
 
         Override in subclasses.
         """
         if self.subapp is not None:
             return self.subapp.start()
 
     def print_alias_help(self):
         """Print the alias part of the help."""
         if not self.aliases:
             return
 
         lines = []
         classdict = {}
         for cls in self.classes:
             # include all parents (up to, but excluding Configurable) in available names
             for c in cls.mro()[:-3]:
                 classdict[c.__name__] = c
 
         for alias, longname in self.aliases.iteritems():
             classname, traitname = longname.split('.',1)
             cls = classdict[classname]
 
             trait = cls.class_traits(config=True)[traitname]
             help = cls.class_get_trait_help(trait).splitlines()
             # reformat first line
             help[0] = help[0].replace(longname, alias) + ' (%s)'%longname
             if len(alias) == 1:
                 help[0] = help[0].replace('--%s='%alias, '-%s '%alias)
             lines.extend(help)
         # lines.append('')
         print os.linesep.join(lines)
 
     def print_flag_help(self):
         """Print the flag part of the help."""
         if not self.flags:
             return
 
         lines = []
         for m, (cfg,help) in self.flags.iteritems():
             prefix = '--' if len(m) > 1 else '-'
             lines.append(prefix+m)
             lines.append(indent(dedent(help.strip())))
         # lines.append('')
         print os.linesep.join(lines)
 
     def print_options(self):
         if not self.flags and not self.aliases:
             return
         lines = ['Options']
         lines.append('-'*len(lines[0]))
         lines.append('')
         for p in wrap_paragraphs(self.option_description):
             lines.append(p)
             lines.append('')
         print os.linesep.join(lines)
         self.print_flag_help()
         self.print_alias_help()
         print
 
     def print_subcommands(self):
         """Print the subcommand part of the help."""
         if not self.subcommands:
             return
 
         lines = ["Subcommands"]
         lines.append('-'*len(lines[0]))
         lines.append('')
         for p in wrap_paragraphs(self.subcommand_description):
             lines.append(p)
             lines.append('')
         for subc, (cls, help) in self.subcommands.iteritems():
             lines.append(subc)
             if help:
                 lines.append(indent(dedent(help.strip())))
         lines.append('')
         print os.linesep.join(lines)
 
     def print_help(self, classes=False):
         """Print the help for each Configurable class in self.classes.
 
         If classes=False (the default), only flags and aliases are printed.
         """
         self.print_subcommands()
         self.print_options()
 
         if classes:
             if self.classes:
                 print "Class parameters"
                 print "----------------"
                 print
                 for p in wrap_paragraphs(self.keyvalue_description):
                     print p
                     print
 
             for cls in self.classes:
                 cls.class_print_help()
                 print
         else:
             print "To see all available configurables, use `--help-all`"
             print
 
     def print_description(self):
         """Print the application description."""
         for p in wrap_paragraphs(self.description):
             print p
             print
 
     def print_examples(self):
         """Print usage and examples.
 
         This usage string goes at the end of the command line help string
         and should contain examples of the application's usage.
         """
         if self.examples:
             print "Examples"
             print "--------"
             print
             print indent(dedent(self.examples.strip()))
             print
 
     def print_version(self):
         """Print the version string."""
         print self.version
 
     def update_config(self, config):
         """Fire the traits events when the config is updated."""
         # Save a copy of the current config.
         newconfig = deepcopy(self.config)
         # Merge the new config into the current one.
         newconfig._merge(config)
         # Save the combined config as self.config, which triggers the traits
         # events.
         self.config = newconfig
 
     @catch_config_error
     def initialize_subcommand(self, subc, argv=None):
         """Initialize a subcommand with argv."""
         subapp,help = self.subcommands.get(subc)
 
         if isinstance(subapp, basestring):
             subapp = import_item(subapp)
 
         # clear existing instances
         self.__class__.clear_instance()
         # instantiate
         self.subapp = subapp.instance()
         # and initialize subapp
         self.subapp.initialize(argv)
     
     def flatten_flags(self):
         """flatten flags and aliases, so cl-args override as expected.
         
         This prevents issues such as an alias pointing to InteractiveShell,
         but a config file setting the same trait in TerminalInteraciveShell
         getting inappropriate priority over the command-line arg.
 
         Only aliases with exactly one descendent in the class list
         will be promoted.
         
         """
         # build a tree of classes in our list that inherit from a particular
         # it will be a dict by parent classname of classes in our list
         # that are descendents
         mro_tree = defaultdict(list)
         for cls in self.classes:
             clsname = cls.__name__
             for parent in cls.mro()[1:-3]:
                 # exclude cls itself and Configurable,HasTraits,object
                 mro_tree[parent.__name__].append(clsname)
         # flatten aliases, which have the form:
         # { 'alias' : 'Class.trait' }
         aliases = {}
         for alias, cls_trait in self.aliases.iteritems():
             cls,trait = cls_trait.split('.',1)
             children = mro_tree[cls]
             if len(children) == 1:
                 # exactly one descendent, promote alias
                 cls = children[0]
             aliases[alias] = '.'.join([cls,trait])
         
         # flatten flags, which are of the form:
         # { 'key' : ({'Cls' : {'trait' : value}}, 'help')}
         flags = {}
         for key, (flagdict, help) in self.flags.iteritems():
             newflag = {}
             for cls, subdict in flagdict.iteritems():
                 children = mro_tree[cls]
                 # exactly one descendent, promote flag section
                 if len(children) == 1:
                     cls = children[0]
                 newflag[cls] = subdict
             flags[key] = (newflag, help)
         return flags, aliases
 
     @catch_config_error
     def parse_command_line(self, argv=None):
         """Parse the command line arguments."""
         argv = sys.argv[1:] if argv is None else argv
         
         if argv and argv[0] == 'help':
             # turn `ipython help notebook` into `ipython notebook -h`
             argv = argv[1:] + ['-h']
 
         if self.subcommands and len(argv) > 0:
             # we have subcommands, and one may have been specified
             subc, subargv = argv[0], argv[1:]
             if re.match(r'^\w(\-?\w)*$', subc) and subc in self.subcommands:
                 # it's a subcommand, and *not* a flag or class parameter
                 return self.initialize_subcommand(subc, subargv)
 
         if '-h' in argv or '--help' in argv or '--help-all' in argv:
             self.print_description()
             self.print_help('--help-all' in argv)
             self.print_examples()
             self.exit(0)
 
         if '--version' in argv or '-V' in argv:
             self.print_version()
             self.exit(0)
         
         # flatten flags&aliases, so cl-args get appropriate priority:
         flags,aliases = self.flatten_flags()
 
         loader = KVArgParseConfigLoader(argv=argv, aliases=aliases,
                                         flags=flags)
         config = loader.load_config()
         self.update_config(config)
         # store unparsed args in extra_args
         self.extra_args = loader.extra_args
 
     @catch_config_error
     def load_config_file(self, filename, path=None):
         """Load a .py based config file by filename and path."""
         loader = PyFileConfigLoader(filename, path=path)
         try:
             config = loader.load_config()
         except ConfigFileNotFound:
             # problem finding the file, raise
             raise
         except Exception:
             # try to get the full filename, but it will be empty in the
             # unlikely event that the error raised before filefind finished
             filename = loader.full_filename or filename
             # problem while running the file
             self.log.error("Exception while loading config file %s",
                             filename, exc_info=True)
         else:
             self.log.debug("Loaded config file: %s", loader.full_filename)
             self.update_config(config)
 
     def generate_config_file(self):
         """generate default config file from Configurables"""
         lines = ["# Configuration file for %s."%self.name]
         lines.append('')
         lines.append('c = get_config()')
         lines.append('')
         for cls in self.classes:
             lines.append(cls.class_config_section())
         return '\n'.join(lines)
 
     def exit(self, exit_status=0):
         self.log.debug("Exiting application: %s" % self.name)
         sys.exit(exit_status)
 
 #-----------------------------------------------------------------------------
 # utility functions, for convenience
 #-----------------------------------------------------------------------------
 
 def boolean_flag(name, configurable, set_help='', unset_help=''):
     """Helper for building basic --trait, --no-trait flags.
 
     Parameters
     ----------
 
     name : str
         The name of the flag.
     configurable : str
         The 'Class.trait' string of the trait to be set/unset with the flag
     set_help : unicode
         help string for --name flag
     unset_help : unicode
         help string for --no-name flag
 
     Returns
     -------
 
     cfg : dict
         A dict with two keys: 'name', and 'no-name', for setting and unsetting
         the trait, respectively.
     """
     # default helpstrings
     set_help = set_help or "set %s=True"%configurable
     unset_help = unset_help or "set %s=False"%configurable
 
     cls,trait = configurable.split('.')
 
     setter = {cls : {trait : True}}
     unsetter = {cls : {trait : False}}
     return {name : (setter, set_help), 'no-'+name : (unsetter, unset_help)}
 

IPython/core/history.py

@@ -1,976 +1,977 b''
 """ History related magics and functionality """
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2010-2011 The IPython Development Team.
 #
 #  Distributed under the terms of the BSD License.
 #
 #  The full license is in the file COPYING.txt, distributed with this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 from __future__ import print_function
 
 # Stdlib imports
 import atexit
 import datetime
 from io import open as io_open
 import os
 import re
 try:
     import sqlite3
 except ImportError:
     sqlite3 = None
 import threading
 
 # Our own packages
 from IPython.core.error import StdinNotImplementedError
 from IPython.config.configurable import Configurable
 from IPython.external.decorator import decorator
 from IPython.testing.skipdoctest import skip_doctest
 from IPython.utils import io
 from IPython.utils.path import locate_profile
 from IPython.utils.traitlets import Bool, Dict, Instance, Integer, List, Unicode
 from IPython.utils.warn import warn
 
 #-----------------------------------------------------------------------------
 # Classes and functions
 #-----------------------------------------------------------------------------
 
 class DummyDB(object):
     """Dummy DB that will act as a black hole for history.
     
     Only used in the absence of sqlite"""
     def execute(*args, **kwargs):
         return []
     
     def commit(self, *args, **kwargs):
         pass
     
     def __enter__(self, *args, **kwargs):
         pass
     
     def __exit__(self, *args, **kwargs):
         pass
     
 @decorator
 def needs_sqlite(f,*a,**kw):
     """return an empty list in the absence of sqlite"""
     if sqlite3 is None:
         return []
     else:
         return f(*a,**kw)
 
 class HistoryAccessor(Configurable):
     """Access the history database without adding to it.
     
     This is intended for use by standalone history tools. IPython shells use
     HistoryManager, below, which is a subclass of this."""
 
     # String holding the path to the history file
     hist_file = Unicode(config=True,
         help="""Path to file to use for SQLite history database.
         
         By default, IPython will put the history database in the IPython profile
         directory.  If you would rather share one history among profiles,
         you ca set this value in each, so that they are consistent.
         
         Due to an issue with fcntl, SQLite is known to misbehave on some NFS mounts.
         If you see IPython hanging, try setting this to something on a local disk,
         e.g::
         
             ipython --HistoryManager.hist_file=/tmp/ipython_hist.sqlite
         
         """)
 
 
     # The SQLite database
     if sqlite3:
         db = Instance(sqlite3.Connection)
     else:
         db = Instance(DummyDB)
     
     def __init__(self, profile='default', hist_file=u'', config=None, **traits):
         """Create a new history accessor.
         
         Parameters
         ----------
         profile : str
           The name of the profile from which to open history.
         hist_file : str
           Path to an SQLite history database stored by IPython. If specified,
           hist_file overrides profile.
         config :
           Config object. hist_file can also be set through this.
         """
         # We need a pointer back to the shell for various tasks.
         super(HistoryAccessor, self).__init__(config=config, **traits)
         # defer setting hist_file from kwarg until after init,
         # otherwise the default kwarg value would clobber any value
         # set by config
         if hist_file:
             self.hist_file = hist_file
         
         if self.hist_file == u'':
             # No one has set the hist_file, yet.
             self.hist_file = self._get_hist_file_name(profile)
 
         if sqlite3 is None:
             warn("IPython History requires SQLite, your history will not be saved\n")
             self.db = DummyDB()
             return
         
         try:
             self.init_db()
         except sqlite3.DatabaseError:
             if os.path.isfile(self.hist_file):
                 # Try to move the file out of the way
                 base,ext = os.path.splitext(self.hist_file)
                 newpath = base + '-corrupt' + ext
                 os.rename(self.hist_file, newpath)
                 print("ERROR! History file wasn't a valid SQLite database.",
                 "It was moved to %s" % newpath, "and a new file created.")
                 self.init_db()
             else:
                 # The hist_file is probably :memory: or something else.
                 raise
     
     def _get_hist_file_name(self, profile='default'):
         """Find the history file for the given profile name.
         
         This is overridden by the HistoryManager subclass, to use the shell's
         active profile.
         
         Parameters
         ----------
         profile : str
           The name of a profile which has a history file.
         """
         return os.path.join(locate_profile(profile), 'history.sqlite')
     
     def init_db(self):
         """Connect to the database, and create tables if necessary."""
         # use detect_types so that timestamps return datetime objects
         self.db = sqlite3.connect(self.hist_file, detect_types=sqlite3.PARSE_DECLTYPES|sqlite3.PARSE_COLNAMES)
         self.db.execute("""CREATE TABLE IF NOT EXISTS sessions (session integer
                         primary key autoincrement, start timestamp,
                         end timestamp, num_cmds integer, remark text)""")
         self.db.execute("""CREATE TABLE IF NOT EXISTS history
                 (session integer, line integer, source text, source_raw text,
                 PRIMARY KEY (session, line))""")
         # Output history is optional, but ensure the table's there so it can be
         # enabled later.
         self.db.execute("""CREATE TABLE IF NOT EXISTS output_history
                         (session integer, line integer, output text,
                         PRIMARY KEY (session, line))""")
         self.db.commit()
 
     def writeout_cache(self):
         """Overridden by HistoryManager to dump the cache before certain
         database lookups."""
         pass
 
     ## -------------------------------
     ## Methods for retrieving history:
     ## -------------------------------
     def _run_sql(self, sql, params, raw=True, output=False):
         """Prepares and runs an SQL query for the history database.
 
         Parameters
         ----------
         sql : str
           Any filtering expressions to go after SELECT ... FROM ...
         params : tuple
           Parameters passed to the SQL query (to replace "?")
         raw, output : bool
           See :meth:`get_range`
 
         Returns
         -------
         Tuples as :meth:`get_range`
         """
         toget = 'source_raw' if raw else 'source'
         sqlfrom = "history"
         if output:
             sqlfrom = "history LEFT JOIN output_history USING (session, line)"
             toget = "history.%s, output_history.output" % toget
         cur = self.db.execute("SELECT session, line, %s FROM %s " %\
                                 (toget, sqlfrom) + sql, params)
         if output:    # Regroup into 3-tuples, and parse JSON
             return ((ses, lin, (inp, out)) for ses, lin, inp, out in cur)
         return cur
 
     @needs_sqlite
     def get_session_info(self, session=0):
         """get info about a session
 
         Parameters
         ----------
 
         session : int
             Session number to retrieve. The current session is 0, and negative
             numbers count back from current session, so -1 is previous session.
 
         Returns
         -------
 
         (session_id [int], start [datetime], end [datetime], num_cmds [int], remark [unicode])
 
         Sessions that are running or did not exit cleanly will have `end=None`
         and `num_cmds=None`.
 
         """
 
         if session <= 0:
             session += self.session_number
 
         query = "SELECT * from sessions where session == ?"
         return self.db.execute(query, (session,)).fetchone()
 
     def get_tail(self, n=10, raw=True, output=False, include_latest=False):
         """Get the last n lines from the history database.
 
         Parameters
         ----------
         n : int
           The number of lines to get
         raw, output : bool
           See :meth:`get_range`
         include_latest : bool
           If False (default), n+1 lines are fetched, and the latest one
           is discarded. This is intended to be used where the function
           is called by a user command, which it should not return.
 
         Returns
         -------
         Tuples as :meth:`get_range`
         """
         self.writeout_cache()
         if not include_latest:
             n += 1
         cur = self._run_sql("ORDER BY session DESC, line DESC LIMIT ?",
                                 (n,), raw=raw, output=output)
         if not include_latest:
             return reversed(list(cur)[1:])
         return reversed(list(cur))
 
     def search(self, pattern="*", raw=True, search_raw=True,
                                                         output=False):
         """Search the database using unix glob-style matching (wildcards
         * and ?).
 
         Parameters
         ----------
         pattern : str
           The wildcarded pattern to match when searching
         search_raw : bool
           If True, search the raw input, otherwise, the parsed input
         raw, output : bool
           See :meth:`get_range`
 
         Returns
         -------
         Tuples as :meth:`get_range`
         """
         tosearch = "source_raw" if search_raw else "source"
         if output:
             tosearch = "history." + tosearch
         self.writeout_cache()
         return self._run_sql("WHERE %s GLOB ?" % tosearch, (pattern,),
                                     raw=raw, output=output)
     
     def get_range(self, session, start=1, stop=None, raw=True,output=False):
         """Retrieve input by session.
 
         Parameters
         ----------
         session : int
             Session number to retrieve.
         start : int
             First line to retrieve.
         stop : int
             End of line range (excluded from output itself). If None, retrieve
             to the end of the session.
         raw : bool
             If True, return untranslated input
         output : bool
             If True, attempt to include output. This will be 'real' Python
             objects for the current session, or text reprs from previous
             sessions if db_log_output was enabled at the time. Where no output
             is found, None is used.
 
         Returns
         -------
         An iterator over the desired lines. Each line is a 3-tuple, either
         (session, line, input) if output is False, or
         (session, line, (input, output)) if output is True.
         """
         if stop:
             lineclause = "line >= ? AND line < ?"
             params = (session, start, stop)
         else:
             lineclause = "line>=?"
             params = (session, start)
 
         return self._run_sql("WHERE session==? AND %s""" % lineclause,
                                     params, raw=raw, output=output)
 
     def get_range_by_str(self, rangestr, raw=True, output=False):
         """Get lines of history from a string of ranges, as used by magic
         commands %hist, %save, %macro, etc.
 
         Parameters
         ----------
         rangestr : str
           A string specifying ranges, e.g. "5 ~2/1-4". See
           :func:`magic_history` for full details.
         raw, output : bool
           As :meth:`get_range`
 
         Returns
         -------
         Tuples as :meth:`get_range`
         """
         for sess, s, e in extract_hist_ranges(rangestr):
             for line in self.get_range(sess, s, e, raw=raw, output=output):
                 yield line
 
 
 class HistoryManager(HistoryAccessor):
     """A class to organize all history-related functionality in one place.
     """
     # Public interface
 
     # An instance of the IPython shell we are attached to
     shell = Instance('IPython.core.interactiveshell.InteractiveShellABC')
     # Lists to hold processed and raw history. These start with a blank entry
     # so that we can index them starting from 1
     input_hist_parsed = List([""])
     input_hist_raw = List([""])
     # A list of directories visited during session
     dir_hist = List()
     def _dir_hist_default(self):
         try:
             return [os.getcwdu()]
         except OSError:
             return []
 
     # A dict of output history, keyed with ints from the shell's
     # execution count.
     output_hist = Dict()
     # The text/plain repr of outputs.
     output_hist_reprs = Dict()
 
     # The number of the current session in the history database
     session_number = Integer()
     # Should we log output to the database? (default no)
     db_log_output = Bool(False, config=True)
     # Write to database every x commands (higher values save disk access & power)
     #  Values of 1 or less effectively disable caching. 
     db_cache_size = Integer(0, config=True)
     # The input and output caches
     db_input_cache = List()
     db_output_cache = List()
     
     # History saving in separate thread
     save_thread = Instance('IPython.core.history.HistorySavingThread')
     try:               # Event is a function returning an instance of _Event...
         save_flag = Instance(threading._Event)
     except AttributeError:         # ...until Python 3.3, when it's a class.
         save_flag = Instance(threading.Event)
     
     # Private interface
     # Variables used to store the three last inputs from the user.  On each new
     # history update, we populate the user's namespace with these, shifted as
     # necessary.
     _i00 = Unicode(u'')
     _i = Unicode(u'')
     _ii = Unicode(u'')
     _iii = Unicode(u'')
 
     # A regex matching all forms of the exit command, so that we don't store
     # them in the history (it's annoying to rewind the first entry and land on
     # an exit call).
     _exit_re = re.compile(r"(exit|quit)(\s*\(.*\))?$")
 
     def __init__(self, shell=None, config=None, **traits):
         """Create a new history manager associated with a shell instance.
         """
         # We need a pointer back to the shell for various tasks.
         super(HistoryManager, self).__init__(shell=shell, config=config,
             **traits)
         self.save_flag = threading.Event()
         self.db_input_cache_lock = threading.Lock()
         self.db_output_cache_lock = threading.Lock()
-        self.save_thread = HistorySavingThread(self)
-        self.save_thread.start()
+        if self.hist_file != ':memory:':
+            self.save_thread = HistorySavingThread(self)
+            self.save_thread.start()
 
         self.new_session()
 
     def _get_hist_file_name(self, profile=None):
         """Get default history file name based on the Shell's profile.
         
         The profile parameter is ignored, but must exist for compatibility with
         the parent class."""
         profile_dir = self.shell.profile_dir.location
         return os.path.join(profile_dir, 'history.sqlite')
     
     @needs_sqlite
     def new_session(self, conn=None):
         """Get a new session number."""
         if conn is None:
             conn = self.db
         
         with conn:
             cur = conn.execute("""INSERT INTO sessions VALUES (NULL, ?, NULL,
                             NULL, "") """, (datetime.datetime.now(),))
             self.session_number = cur.lastrowid
             
     def end_session(self):
         """Close the database session, filling in the end time and line count."""
         self.writeout_cache()
         with self.db:
             self.db.execute("""UPDATE sessions SET end=?, num_cmds=? WHERE
                             session==?""", (datetime.datetime.now(),
                             len(self.input_hist_parsed)-1, self.session_number))
         self.session_number = 0
                             
     def name_session(self, name):
         """Give the current session a name in the history database."""
         with self.db:
             self.db.execute("UPDATE sessions SET remark=? WHERE session==?",
                             (name, self.session_number))
                             
     def reset(self, new_session=True):
         """Clear the session history, releasing all object references, and
         optionally open a new session."""
         self.output_hist.clear()
         # The directory history can't be completely empty
         self.dir_hist[:] = [os.getcwdu()]
         
         if new_session:
             if self.session_number:
                 self.end_session()
             self.input_hist_parsed[:] = [""]
             self.input_hist_raw[:] = [""]
             self.new_session()
     
     # ------------------------------
     # Methods for retrieving history
     # ------------------------------
     def _get_range_session(self, start=1, stop=None, raw=True, output=False):
         """Get input and output history from the current session. Called by
         get_range, and takes similar parameters."""
         input_hist = self.input_hist_raw if raw else self.input_hist_parsed
             
         n = len(input_hist)
         if start < 0:
             start += n
         if not stop or (stop > n):
             stop = n
         elif stop < 0:
             stop += n
         
         for i in range(start, stop):
             if output:
                 line = (input_hist[i], self.output_hist_reprs.get(i))
             else:
                 line = input_hist[i]
             yield (0, i, line)
     
     def get_range(self, session=0, start=1, stop=None, raw=True,output=False):
         """Retrieve input by session.
         
         Parameters
         ----------
         session : int
             Session number to retrieve. The current session is 0, and negative
             numbers count back from current session, so -1 is previous session.
         start : int
             First line to retrieve.
         stop : int
             End of line range (excluded from output itself). If None, retrieve
             to the end of the session.
         raw : bool
             If True, return untranslated input
         output : bool
             If True, attempt to include output. This will be 'real' Python
             objects for the current session, or text reprs from previous
             sessions if db_log_output was enabled at the time. Where no output
             is found, None is used.
             
         Returns
         -------
         An iterator over the desired lines. Each line is a 3-tuple, either
         (session, line, input) if output is False, or
         (session, line, (input, output)) if output is True.
         """
         if session <= 0:
             session += self.session_number
         if session==self.session_number:          # Current session
             return self._get_range_session(start, stop, raw, output)
         return super(HistoryManager, self).get_range(session, start, stop, raw, output)
 
     ## ----------------------------
     ## Methods for storing history:
     ## ----------------------------
     def store_inputs(self, line_num, source, source_raw=None):
         """Store source and raw input in history and create input cache
         variables _i*.
 
         Parameters
         ----------
         line_num : int
           The prompt number of this input.
 
         source : str
           Python input.
 
         source_raw : str, optional
           If given, this is the raw input without any IPython transformations
           applied to it.  If not given, ``source`` is used.
         """
         if source_raw is None:
             source_raw = source
         source = source.rstrip('\n')
         source_raw = source_raw.rstrip('\n')
 
         # do not store exit/quit commands
         if self._exit_re.match(source_raw.strip()):
             return
 
         self.input_hist_parsed.append(source)
         self.input_hist_raw.append(source_raw)
 
         with self.db_input_cache_lock:
             self.db_input_cache.append((line_num, source, source_raw))
             # Trigger to flush cache and write to DB.
             if len(self.db_input_cache) >= self.db_cache_size:
                 self.save_flag.set()
 
         # update the auto _i variables
         self._iii = self._ii
         self._ii = self._i
         self._i = self._i00
         self._i00 = source_raw
 
         # hackish access to user namespace to create _i1,_i2... dynamically
         new_i = '_i%s' % line_num
         to_main = {'_i': self._i,
                    '_ii': self._ii,
                    '_iii': self._iii,
                    new_i : self._i00 }
 
         self.shell.push(to_main, interactive=False)
 
     def store_output(self, line_num):
         """If database output logging is enabled, this saves all the
         outputs from the indicated prompt number to the database. It's
         called by run_cell after code has been executed.
 
         Parameters
         ----------
         line_num : int
           The line number from which to save outputs
         """
         if (not self.db_log_output) or (line_num not in self.output_hist_reprs):
             return
         output = self.output_hist_reprs[line_num]
 
         with self.db_output_cache_lock:
             self.db_output_cache.append((line_num, output))
         if self.db_cache_size <= 1:
             self.save_flag.set()
 
     def _writeout_input_cache(self, conn):
         with conn:
             for line in self.db_input_cache:
                 conn.execute("INSERT INTO history VALUES (?, ?, ?, ?)",
                                 (self.session_number,)+line)
 
     def _writeout_output_cache(self, conn):
         with conn:
             for line in self.db_output_cache:
                 conn.execute("INSERT INTO output_history VALUES (?, ?, ?)",
                                 (self.session_number,)+line)
 
     @needs_sqlite
     def writeout_cache(self, conn=None):
         """Write any entries in the cache to the database."""
         if conn is None:
             conn = self.db
 
         with self.db_input_cache_lock:
             try:
                 self._writeout_input_cache(conn)
             except sqlite3.IntegrityError:
                 self.new_session(conn)
                 print("ERROR! Session/line number was not unique in",
                       "database. History logging moved to new session",
                                                 self.session_number)
                 try: # Try writing to the new session. If this fails, don't recurse
                     self._writeout_input_cache(conn)
                 except sqlite3.IntegrityError:
                     pass
             finally:
                 self.db_input_cache = []
 
         with self.db_output_cache_lock:
             try:
                 self._writeout_output_cache(conn)
             except sqlite3.IntegrityError:
                 print("!! Session/line number for output was not unique",
                       "in database. Output will not be stored.")
             finally:
                 self.db_output_cache = []
 
 
 class HistorySavingThread(threading.Thread):
     """This thread takes care of writing history to the database, so that
     the UI isn't held up while that happens.
 
     It waits for the HistoryManager's save_flag to be set, then writes out
     the history cache. The main thread is responsible for setting the flag when
     the cache size reaches a defined threshold."""
     daemon = True
     stop_now = False
     def __init__(self, history_manager):
         super(HistorySavingThread, self).__init__()
         self.history_manager = history_manager
         atexit.register(self.stop)
 
     @needs_sqlite
     def run(self):
         # We need a separate db connection per thread:
         try:
             self.db = sqlite3.connect(self.history_manager.hist_file)
             while True:
                 self.history_manager.save_flag.wait()
                 if self.stop_now:
                     return
                 self.history_manager.save_flag.clear()
                 self.history_manager.writeout_cache(self.db)
         except Exception as e:
             print(("The history saving thread hit an unexpected error (%s)."
                    "History will not be written to the database.") % repr(e))
 
     def stop(self):
         """This can be called from the main thread to safely stop this thread.
 
         Note that it does not attempt to write out remaining history before
         exiting. That should be done by calling the HistoryManager's
         end_session method."""
         self.stop_now = True
         self.history_manager.save_flag.set()
         self.join()
 
 
 # To match, e.g. ~5/8-~2/3
 range_re = re.compile(r"""
 ((?P<startsess>~?\d+)/)?
 (?P<start>\d+)                    # Only the start line num is compulsory
 ((?P<sep>[\-:])
  ((?P<endsess>~?\d+)/)?
  (?P<end>\d+))?
 $""", re.VERBOSE)
 
 def extract_hist_ranges(ranges_str):
     """Turn a string of history ranges into 3-tuples of (session, start, stop).
 
     Examples
     --------
     list(extract_input_ranges("~8/5-~7/4 2"))
     [(-8, 5, None), (-7, 1, 4), (0, 2, 3)]
     """
     for range_str in ranges_str.split():
         rmatch = range_re.match(range_str)
         if not rmatch:
             continue
         start = int(rmatch.group("start"))
         end = rmatch.group("end")
         end = int(end) if end else start+1   # If no end specified, get (a, a+1)
         if rmatch.group("sep") == "-":       # 1-3 == 1:4 --> [1, 2, 3]
             end += 1
         startsess = rmatch.group("startsess") or "0"
         endsess = rmatch.group("endsess") or startsess
         startsess = int(startsess.replace("~","-"))
         endsess = int(endsess.replace("~","-"))
         assert endsess >= startsess
 
         if endsess == startsess:
             yield (startsess, start, end)
             continue
         # Multiple sessions in one range:
         yield (startsess, start, None)
         for sess in range(startsess+1, endsess):
             yield (sess, 1, None)
         yield (endsess, 1, end)
 
 def _format_lineno(session, line):
     """Helper function to format line numbers properly."""
     if session == 0:
         return str(line)
     return "%s#%s" % (session, line)
 
 @skip_doctest
 def magic_history(self, parameter_s = ''):
     """Print input history (_i<n> variables), with most recent last.
 
     %history [-o -p -t -n] [-f filename] [range | -g pattern | -l number]
 
     By default, input history is printed without line numbers so it can be
     directly pasted into an editor. Use -n to show them.
 
     By default, all input history from the current session is displayed.
     Ranges of history can be indicated using the syntax: 
     4      : Line 4, current session
     4-6    : Lines 4-6, current session
     243/1-5: Lines 1-5, session 243
     ~2/7   : Line 7, session 2 before current
     ~8/1-~6/5 : From the first line of 8 sessions ago, to the fifth line
                 of 6 sessions ago.
     Multiple ranges can be entered, separated by spaces
 
     The same syntax is used by %macro, %save, %edit, %rerun
 
     Options:
 
       -n: print line numbers for each input.
       This feature is only available if numbered prompts are in use.
 
       -o: also print outputs for each input.
 
       -p: print classic '>>>' python prompts before each input.  This is useful
        for making documentation, and in conjunction with -o, for producing
        doctest-ready output.
 
       -r: (default) print the 'raw' history, i.e. the actual commands you typed.
 
       -t: print the 'translated' history, as IPython understands it.  IPython
       filters your input and converts it all into valid Python source before
       executing it (things like magics or aliases are turned into function
       calls, for example). With this option, you'll see the native history
       instead of the user-entered version: '%cd /' will be seen as
       'get_ipython().magic("%cd /")' instead of '%cd /'.
 
       -g: treat the arg as a pattern to grep for in (full) history.
       This includes the saved history (almost all commands ever written).
       Use '%hist -g' to show full saved history (may be very long).
 
       -l: get the last n lines from all sessions. Specify n as a single arg, or
       the default is the last 10 lines.
 
       -f FILENAME: instead of printing the output to the screen, redirect it to
        the given file.  The file is always overwritten, though *when it can*,
        IPython asks for confirmation first. In particular, running the command
        "history -f FILENAME" from the IPython Notebook interface will replace
        FILENAME even if it already exists *without* confirmation.
 
     Examples
     --------
     ::
 
       In [6]: %hist -n 4-6
       4:a = 12
       5:print a**2
       6:%hist -n 4-6
 
     """
 
     if not self.shell.displayhook.do_full_cache:
         print('This feature is only available if numbered prompts are in use.')
         return
     opts,args = self.parse_options(parameter_s,'noprtglf:',mode='string')
 
     # For brevity
     history_manager = self.shell.history_manager
 
     def _format_lineno(session, line):
         """Helper function to format line numbers properly."""
         if session in (0, history_manager.session_number):
             return str(line)
         return "%s/%s" % (session, line)
 
     # Check if output to specific file was requested.
     try:
         outfname = opts['f']
     except KeyError:
         outfile = io.stdout  # default
         # We don't want to close stdout at the end!
         close_at_end = False
     else:
         if os.path.exists(outfname):
             try:
                 ans = io.ask_yes_no("File %r exists. Overwrite?" % outfname)
             except StdinNotImplementedError:
                 ans = True
             if not ans:
                 print('Aborting.')
                 return
             print("Overwriting file.")
         outfile = io_open(outfname, 'w', encoding='utf-8')
         close_at_end = True
 
     print_nums = 'n' in opts
     get_output = 'o' in opts
     pyprompts = 'p' in opts
     # Raw history is the default
     raw = not('t' in opts)
 
     default_length = 40
     pattern = None
 
     if 'g' in opts:         # Glob search
         pattern = "*" + args + "*" if args else "*"
         hist = history_manager.search(pattern, raw=raw, output=get_output)
         print_nums = True
     elif 'l' in opts:       # Get 'tail'
         try:
             n = int(args)
         except ValueError, IndexError:
             n = 10
         hist = history_manager.get_tail(n, raw=raw, output=get_output)
     else:
         if args:            # Get history by ranges
             hist = history_manager.get_range_by_str(args, raw, get_output)
         else:               # Just get history for the current session
             hist = history_manager.get_range(raw=raw, output=get_output)
 
     # We could be displaying the entire history, so let's not try to pull it
     # into a list in memory. Anything that needs more space will just misalign.
     width = 4
 
     for session, lineno, inline in hist:
         # Print user history with tabs expanded to 4 spaces.  The GUI clients
         # use hard tabs for easier usability in auto-indented code, but we want
         # to produce PEP-8 compliant history for safe pasting into an editor.
         if get_output:
             inline, output = inline
         inline = inline.expandtabs(4).rstrip()
 
         multiline = "\n" in inline
         line_sep = '\n' if multiline else ' '
         if print_nums:
             print(u'%s:%s' % (_format_lineno(session, lineno).rjust(width),
                     line_sep),  file=outfile, end=u'')
         if pyprompts:
             print(u">>> ", end=u"", file=outfile)
             if multiline:
                 inline = "\n... ".join(inline.splitlines()) + "\n..."
         print(inline, file=outfile)
         if get_output and output:
             print(output, file=outfile)
 
     if close_at_end:
         outfile.close()
 
 
 def magic_rep(self, arg):
     r"""Repeat a command, or get command to input line for editing.
     
     %recall and %rep are equivalent.
 
     - %recall (no arguments):
 
     Place a string version of last computation result (stored in the special '_'
     variable) to the next input prompt. Allows you to create elaborate command
     lines without using copy-paste::
 
          In[1]: l = ["hei", "vaan"]
          In[2]: "".join(l)
         Out[2]: heivaan
          In[3]: %rep
          In[4]: heivaan_ <== cursor blinking
 
     %recall 45
 
     Place history line 45 on the next input prompt. Use %hist to find
     out the number.
 
     %recall 1-4
 
     Combine the specified lines into one cell, and place it on the next
     input prompt. See %history for the slice syntax.
 
     %recall foo+bar
 
     If foo+bar can be evaluated in the user namespace, the result is
     placed at the next input prompt. Otherwise, the history is searched
     for lines which contain that substring, and the most recent one is
     placed at the next input prompt.
     """
     if not arg:                 # Last output
         self.set_next_input(str(self.shell.user_ns["_"]))
         return
                                 # Get history range
     histlines = self.history_manager.get_range_by_str(arg)
     cmd = "\n".join(x[2] for x in histlines)
     if cmd:
         self.set_next_input(cmd.rstrip())
         return
 
     try:                        # Variable in user namespace
         cmd = str(eval(arg, self.shell.user_ns))
     except Exception:           # Search for term in history
         histlines = self.history_manager.search("*"+arg+"*")
         for h in reversed([x[2] for x in histlines]):
             if 'rep' in h:
                 continue
             self.set_next_input(h.rstrip())
             return
     else:
         self.set_next_input(cmd.rstrip())
     print("Couldn't evaluate or find in history:", arg)
 
 def magic_rerun(self, parameter_s=''):
     """Re-run previous input
 
     By default, you can specify ranges of input history to be repeated
     (as with %history). With no arguments, it will repeat the last line.
 
     Options:
 
       -l <n> : Repeat the last n lines of input, not including the
       current command.
 
       -g foo : Repeat the most recent line which contains foo
     """
     opts, args = self.parse_options(parameter_s, 'l:g:', mode='string')
     if "l" in opts:         # Last n lines
         n = int(opts['l'])
         hist = self.history_manager.get_tail(n)
     elif "g" in opts:       # Search
         p = "*"+opts['g']+"*"
         hist = list(self.history_manager.search(p))
         for l in reversed(hist):
             if "rerun" not in l[2]:
                 hist = [l]     # The last match which isn't a %rerun
                 break
         else:
             hist = []          # No matches except %rerun
     elif args:              # Specify history ranges
         hist = self.history_manager.get_range_by_str(args)
     else:                   # Last line
         hist = self.history_manager.get_tail(1)
     hist = [x[2] for x in hist]
     if not hist:
         print("No lines in history match specification")
         return
     histlines = "\n".join(hist)
     print("=== Executing: ===")
     print(histlines)
     print("=== Output: ===")
     self.run_cell("\n".join(hist), store_history=False)
 
 
 def init_ipython(ip):
     ip.define_magic("rep", magic_rep)
     ip.define_magic("recall", magic_rep)
     ip.define_magic("rerun", magic_rerun)
     ip.define_magic("hist",magic_history)    # Alternative name
     ip.define_magic("history",magic_history)
 
     # XXX - ipy_completers are in quarantine, need to be updated to new apis
     #import ipy_completers
     #ipy_completers.quick_completer('%hist' ,'-g -t -r -n')

IPython/core/interactiveshell.py

@@ -1,2853 +1,2865 b''
 # -*- coding: utf-8 -*-
 """Main IPython class."""
 
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2001 Janko Hauser <[e-mail]>
 #  Copyright (C) 2001-2007 Fernando Perez. <[e-mail]>
 #  Copyright (C) 2008-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 from __future__ import with_statement
 from __future__ import absolute_import
 
 import __builtin__ as builtin_mod
 import __future__
 import abc
 import ast
 import atexit
 import os
 import re
 import runpy
 import sys
 import tempfile
 import types
 import urllib
 from io import open as io_open
 
 from IPython.config.configurable import SingletonConfigurable
 from IPython.core import debugger, oinspect
 from IPython.core import page
 from IPython.core import prefilter
 from IPython.core import shadowns
 from IPython.core import ultratb
 from IPython.core.alias import AliasManager, AliasError
 from IPython.core.autocall import ExitAutocall
 from IPython.core.builtin_trap import BuiltinTrap
 from IPython.core.compilerop import CachingCompiler
 from IPython.core.display_trap import DisplayTrap
 from IPython.core.displayhook import DisplayHook
 from IPython.core.displaypub import DisplayPublisher
 from IPython.core.error import UsageError
 from IPython.core.extensions import ExtensionManager
 from IPython.core.fakemodule import FakeModule, init_fakemod_dict
 from IPython.core.formatters import DisplayFormatter
 from IPython.core.history import HistoryManager
 from IPython.core.inputsplitter import IPythonInputSplitter
 from IPython.core.logger import Logger
 from IPython.core.macro import Macro
 from IPython.core.magic import Magic
 from IPython.core.payload import PayloadManager
 from IPython.core.plugin import PluginManager
 from IPython.core.prefilter import PrefilterManager, ESC_MAGIC
 from IPython.core.profiledir import ProfileDir
 from IPython.core.pylabtools import pylab_activate
 from IPython.core.prompts import PromptManager
 from IPython.utils import PyColorize
 from IPython.utils import io
 from IPython.utils import py3compat
 from IPython.utils import openpy
 from IPython.utils.doctestreload import doctest_reload
 from IPython.utils.io import ask_yes_no
 from IPython.utils.ipstruct import Struct
 from IPython.utils.path import get_home_dir, get_ipython_dir, get_py_filename, unquote_filename
 from IPython.utils.pickleshare import PickleShareDB
 from IPython.utils.process import system, getoutput
 from IPython.utils.strdispatch import StrDispatch
 from IPython.utils.syspathcontext import prepended_to_syspath
 from IPython.utils.text import (format_screen, LSString, SList,
                                 DollarFormatter)
 from IPython.utils.traitlets import (Integer, CBool, CaselessStrEnum, Enum,
                                      List, Unicode, Instance, Type)
 from IPython.utils.warn import warn, error
 import IPython.core.hooks
 
 #-----------------------------------------------------------------------------
 # Globals
 #-----------------------------------------------------------------------------
 
 # compiled regexps for autoindent management
 dedent_re = re.compile(r'^\s+raise|^\s+return|^\s+pass')
 
 #-----------------------------------------------------------------------------
 # Utilities
 #-----------------------------------------------------------------------------
 
 def softspace(file, newvalue):
     """Copied from code.py, to remove the dependency"""
 
     oldvalue = 0
     try:
         oldvalue = file.softspace
     except AttributeError:
         pass
     try:
         file.softspace = newvalue
     except (AttributeError, TypeError):
         # "attribute-less object" or "read-only attributes"
         pass
     return oldvalue
 
 
 def no_op(*a, **kw): pass
 
 class NoOpContext(object):
     def __enter__(self): pass
     def __exit__(self, type, value, traceback): pass
 no_op_context = NoOpContext()
 
 class SpaceInInput(Exception): pass
 
 class Bunch: pass
 
 
 def get_default_colors():
     if sys.platform=='darwin':
         return "LightBG"
     elif os.name=='nt':
         return 'Linux'
     else:
         return 'Linux'
 
 
 class SeparateUnicode(Unicode):
     """A Unicode subclass to validate separate_in, separate_out, etc.
 
     This is a Unicode based trait that converts '0'->'' and '\\n'->'\n'.
     """
 
     def validate(self, obj, value):
         if value == '0': value = ''
         value = value.replace('\\n','\n')
         return super(SeparateUnicode, self).validate(obj, value)
 
 
 class ReadlineNoRecord(object):
     """Context manager to execute some code, then reload readline history
     so that interactive input to the code doesn't appear when pressing up."""
     def __init__(self, shell):
         self.shell = shell
         self._nested_level = 0
 
     def __enter__(self):
         if self._nested_level == 0:
             try:
                 self.orig_length = self.current_length()
                 self.readline_tail = self.get_readline_tail()
             except (AttributeError, IndexError):   # Can fail with pyreadline
                 self.orig_length, self.readline_tail = 999999, []
         self._nested_level += 1
 
     def __exit__(self, type, value, traceback):
         self._nested_level -= 1
         if self._nested_level == 0:
             # Try clipping the end if it's got longer
             try:
                 e = self.current_length() - self.orig_length
                 if e > 0:
                     for _ in range(e):
                         self.shell.readline.remove_history_item(self.orig_length)
 
                 # If it still doesn't match, just reload readline history.
                 if self.current_length() != self.orig_length \
                     or self.get_readline_tail() != self.readline_tail:
                     self.shell.refill_readline_hist()
             except (AttributeError, IndexError):
                 pass
         # Returning False will cause exceptions to propagate
         return False
 
     def current_length(self):
         return self.shell.readline.get_current_history_length()
 
     def get_readline_tail(self, n=10):
         """Get the last n items in readline history."""
         end = self.shell.readline.get_current_history_length() + 1
         start = max(end-n, 1)
         ghi = self.shell.readline.get_history_item
         return [ghi(x) for x in range(start, end)]
 
 #-----------------------------------------------------------------------------
 # Main IPython class
 #-----------------------------------------------------------------------------
 
 class InteractiveShell(SingletonConfigurable, Magic):
     """An enhanced, interactive shell for Python."""
 
     _instance = None
 
     autocall = Enum((0,1,2), default_value=0, config=True, help=
         """
         Make IPython automatically call any callable object even if you didn't
         type explicit parentheses. For example, 'str 43' becomes 'str(43)'
         automatically. The value can be '0' to disable the feature, '1' for
         'smart' autocall, where it is not applied if there are no more
         arguments on the line, and '2' for 'full' autocall, where all callable
         objects are automatically called (even if no arguments are present).
         """
     )
     # TODO: remove all autoindent logic and put into frontends.
     # We can't do this yet because even runlines uses the autoindent.
     autoindent = CBool(True, config=True, help=
         """
         Autoindent IPython code entered interactively.
         """
     )
     automagic = CBool(True, config=True, help=
         """
         Enable magic commands to be called without the leading %.
         """
     )
     cache_size = Integer(1000, config=True, help=
         """
         Set the size of the output cache.  The default is 1000, you can
         change it permanently in your config file.  Setting it to 0 completely
         disables the caching system, and the minimum value accepted is 20 (if
         you provide a value less than 20, it is reset to 0 and a warning is
         issued).  This limit is defined because otherwise you'll spend more
         time re-flushing a too small cache than working
         """
     )
     color_info = CBool(True, config=True, help=
         """
         Use colors for displaying information about objects. Because this
         information is passed through a pager (like 'less'), and some pagers
         get confused with color codes, this capability can be turned off.
         """
     )
     colors = CaselessStrEnum(('NoColor','LightBG','Linux'),
                              default_value=get_default_colors(), config=True,
         help="Set the color scheme (NoColor, Linux, or LightBG)."
     )
     colors_force = CBool(False, help=
         """
         Force use of ANSI color codes, regardless of OS and readline
         availability.
         """
         # FIXME: This is essentially a hack to allow ZMQShell to show colors
         # without readline on Win32. When the ZMQ formatting system is
         # refactored, this should be removed.
     )
     debug = CBool(False, config=True)
     deep_reload = CBool(False, config=True, help=
         """
         Enable deep (recursive) reloading by default. IPython can use the
         deep_reload module which reloads changes in modules recursively (it
         replaces the reload() function, so you don't need to change anything to
         use it). deep_reload() forces a full reload of modules whose code may
         have changed, which the default reload() function does not.  When
         deep_reload is off, IPython will use the normal reload(), but
         deep_reload will still be available as dreload().
         """
     )
     disable_failing_post_execute = CBool(False, config=True,
         help="Don't call post-execute functions that have failed in the past."""
     )
     display_formatter = Instance(DisplayFormatter)
     displayhook_class = Type(DisplayHook)
     display_pub_class = Type(DisplayPublisher)
 
     exit_now = CBool(False)
     exiter = Instance(ExitAutocall)
     def _exiter_default(self):
         return ExitAutocall(self)
     # Monotonically increasing execution counter
     execution_count = Integer(1)
     filename = Unicode("<ipython console>")
     ipython_dir= Unicode('', config=True) # Set to get_ipython_dir() in __init__
 
     # Input splitter, to split entire cells of input into either individual
     # interactive statements or whole blocks.
     input_splitter = Instance('IPython.core.inputsplitter.IPythonInputSplitter',
                               (), {})
     logstart = CBool(False, config=True, help=
         """
         Start logging to the default log file.
         """
     )
     logfile = Unicode('', config=True, help=
         """
         The name of the logfile to use.
         """
     )
     logappend = Unicode('', config=True, help=
         """
         Start logging to the given file in append mode.
         """
     )
     object_info_string_level = Enum((0,1,2), default_value=0,
                                     config=True)
     pdb = CBool(False, config=True, help=
         """
         Automatically call the pdb debugger after every exception.
         """
     )
     multiline_history = CBool(sys.platform != 'win32', config=True,
         help="Save multi-line entries as one entry in readline history"
     )
 
     # deprecated prompt traits:
     
     prompt_in1 = Unicode('In [\\#]: ', config=True,
         help="Deprecated, use PromptManager.in_template")
     prompt_in2 = Unicode('   .\\D.: ', config=True,
         help="Deprecated, use PromptManager.in2_template")
     prompt_out = Unicode('Out[\\#]: ', config=True,
         help="Deprecated, use PromptManager.out_template")
     prompts_pad_left = CBool(True, config=True,
         help="Deprecated, use PromptManager.justify")
     
     def _prompt_trait_changed(self, name, old, new):
         table = {
             'prompt_in1' : 'in_template',
             'prompt_in2' : 'in2_template',
             'prompt_out' : 'out_template',
             'prompts_pad_left' : 'justify',
         }
         warn("InteractiveShell.{name} is deprecated, use PromptManager.{newname}\n".format(
                 name=name, newname=table[name])
         )
         # protect against weird cases where self.config may not exist:
         if self.config is not None:
             # propagate to corresponding PromptManager trait
             setattr(self.config.PromptManager, table[name], new)
     
     _prompt_in1_changed = _prompt_trait_changed
     _prompt_in2_changed = _prompt_trait_changed
     _prompt_out_changed = _prompt_trait_changed
     _prompt_pad_left_changed = _prompt_trait_changed
     
     show_rewritten_input = CBool(True, config=True,
         help="Show rewritten input, e.g. for autocall."
     )
     
     quiet = CBool(False, config=True)
 
     history_length = Integer(10000, config=True)
 
     # The readline stuff will eventually be moved to the terminal subclass
     # but for now, we can't do that as readline is welded in everywhere.
     readline_use = CBool(True, config=True)
     readline_remove_delims = Unicode('-/~', config=True)
     # don't use \M- bindings by default, because they
     # conflict with 8-bit encodings. See gh-58,gh-88
     readline_parse_and_bind = List([
             'tab: complete',
             '"\C-l": clear-screen',
             'set show-all-if-ambiguous on',
             '"\C-o": tab-insert',
             '"\C-r": reverse-search-history',
             '"\C-s": forward-search-history',
             '"\C-p": history-search-backward',
             '"\C-n": history-search-forward',
             '"\e[A": history-search-backward',
             '"\e[B": history-search-forward',
             '"\C-k": kill-line',
             '"\C-u": unix-line-discard',
         ], allow_none=False, config=True)
 
     # TODO: this part of prompt management should be moved to the frontends.
     # Use custom TraitTypes that convert '0'->'' and '\\n'->'\n'
     separate_in = SeparateUnicode('\n', config=True)
     separate_out = SeparateUnicode('', config=True)
     separate_out2 = SeparateUnicode('', config=True)
     wildcards_case_sensitive = CBool(True, config=True)
     xmode = CaselessStrEnum(('Context','Plain', 'Verbose'),
                             default_value='Context', config=True)
 
     # Subcomponents of InteractiveShell
     alias_manager = Instance('IPython.core.alias.AliasManager')
     prefilter_manager = Instance('IPython.core.prefilter.PrefilterManager')
     builtin_trap = Instance('IPython.core.builtin_trap.BuiltinTrap')
     display_trap = Instance('IPython.core.display_trap.DisplayTrap')
     extension_manager = Instance('IPython.core.extensions.ExtensionManager')
     plugin_manager = Instance('IPython.core.plugin.PluginManager')
     payload_manager = Instance('IPython.core.payload.PayloadManager')
     history_manager = Instance('IPython.core.history.HistoryManager')
 
     profile_dir = Instance('IPython.core.application.ProfileDir')
     @property
     def profile(self):
         if self.profile_dir is not None:
             name = os.path.basename(self.profile_dir.location)
             return name.replace('profile_','')
 
 
     # Private interface
     _post_execute = Instance(dict)
 
     def __init__(self, config=None, ipython_dir=None, profile_dir=None,
                  user_module=None, user_ns=None,
                  custom_exceptions=((), None)):
 
         # This is where traits with a config_key argument are updated
         # from the values on config.
         super(InteractiveShell, self).__init__(config=config)
         self.configurables = [self]
 
         # These are relatively independent and stateless
         self.init_ipython_dir(ipython_dir)
         self.init_profile_dir(profile_dir)
         self.init_instance_attrs()
         self.init_environment()
         
         # Check if we're in a virtualenv, and set up sys.path.
         self.init_virtualenv()
 
         # Create namespaces (user_ns, user_global_ns, etc.)
         self.init_create_namespaces(user_module, user_ns)
         # This has to be done after init_create_namespaces because it uses
         # something in self.user_ns, but before init_sys_modules, which
         # is the first thing to modify sys.
         # TODO: When we override sys.stdout and sys.stderr before this class
         # is created, we are saving the overridden ones here. Not sure if this
         # is what we want to do.
         self.save_sys_module_state()
         self.init_sys_modules()
 
         # While we're trying to have each part of the code directly access what
         # it needs without keeping redundant references to objects, we have too
         # much legacy code that expects ip.db to exist.
         self.db = PickleShareDB(os.path.join(self.profile_dir.location, 'db'))
 
         self.init_history()
         self.init_encoding()
         self.init_prefilter()
 
         Magic.__init__(self, self)
 
         self.init_syntax_highlighting()
         self.init_hooks()
         self.init_pushd_popd_magic()
         # self.init_traceback_handlers use to be here, but we moved it below
         # because it and init_io have to come after init_readline.
         self.init_user_ns()
         self.init_logger()
         self.init_alias()
         self.init_builtins()
 
         # pre_config_initialization
 
         # The next section should contain everything that was in ipmaker.
         self.init_logstart()
 
         # The following was in post_config_initialization
         self.init_inspector()
         # init_readline() must come before init_io(), because init_io uses
         # readline related things.
         self.init_readline()
         # We save this here in case user code replaces raw_input, but it needs
         # to be after init_readline(), because PyPy's readline works by replacing
         # raw_input.
         if py3compat.PY3:
             self.raw_input_original = input
         else:
             self.raw_input_original = raw_input
         # init_completer must come after init_readline, because it needs to
         # know whether readline is present or not system-wide to configure the
         # completers, since the completion machinery can now operate
         # independently of readline (e.g. over the network)
         self.init_completer()
         # TODO: init_io() needs to happen before init_traceback handlers
         # because the traceback handlers hardcode the stdout/stderr streams.
         # This logic in in debugger.Pdb and should eventually be changed.
         self.init_io()
         self.init_traceback_handlers(custom_exceptions)
         self.init_prompts()
         self.init_display_formatter()
         self.init_display_pub()
         self.init_displayhook()
         self.init_reload_doctest()
         self.init_magics()
         self.init_pdb()
         self.init_extension_manager()
         self.init_plugin_manager()
         self.init_payload()
         self.hooks.late_startup_hook()
         atexit.register(self.atexit_operations)
 
     def get_ipython(self):
         """Return the currently running IPython instance."""
         return self
 
     #-------------------------------------------------------------------------
     # Trait changed handlers
     #-------------------------------------------------------------------------
 
     def _ipython_dir_changed(self, name, new):
         if not os.path.isdir(new):
             os.makedirs(new, mode = 0777)
 
     def set_autoindent(self,value=None):
         """Set the autoindent flag, checking for readline support.
 
         If called with no arguments, it acts as a toggle."""
 
         if value != 0 and not self.has_readline:
             if os.name == 'posix':
                 warn("The auto-indent feature requires the readline library")
             self.autoindent = 0
             return
         if value is None:
             self.autoindent = not self.autoindent
         else:
             self.autoindent = value
 
     #-------------------------------------------------------------------------
     # init_* methods called by __init__
     #-------------------------------------------------------------------------
 
     def init_ipython_dir(self, ipython_dir):
         if ipython_dir is not None:
             self.ipython_dir = ipython_dir
             return
 
         self.ipython_dir = get_ipython_dir()
 
     def init_profile_dir(self, profile_dir):
         if profile_dir is not None:
             self.profile_dir = profile_dir
             return
         self.profile_dir =\
             ProfileDir.create_profile_dir_by_name(self.ipython_dir, 'default')
 
     def init_instance_attrs(self):
         self.more = False
 
         # command compiler
         self.compile = CachingCompiler()
 
         # Make an empty namespace, which extension writers can rely on both
         # existing and NEVER being used by ipython itself.  This gives them a
         # convenient location for storing additional information and state
         # their extensions may require, without fear of collisions with other
         # ipython names that may develop later.
         self.meta = Struct()
 
         # Temporary files used for various purposes.  Deleted at exit.
         self.tempfiles = []
 
         # Keep track of readline usage (later set by init_readline)
         self.has_readline = False
 
         # keep track of where we started running (mainly for crash post-mortem)
         # This is not being used anywhere currently.
         self.starting_dir = os.getcwdu()
 
         # Indentation management
         self.indent_current_nsp = 0
 
         # Dict to track post-execution functions that have been registered
         self._post_execute = {}
 
     def init_environment(self):
         """Any changes we need to make to the user's environment."""
         pass
 
     def init_encoding(self):
         # Get system encoding at startup time.  Certain terminals (like Emacs
         # under Win32 have it set to None, and we need to have a known valid
         # encoding to use in the raw_input() method
         try:
             self.stdin_encoding = sys.stdin.encoding or 'ascii'
         except AttributeError:
             self.stdin_encoding = 'ascii'
 
     def init_syntax_highlighting(self):
         # Python source parser/formatter for syntax highlighting
         pyformat = PyColorize.Parser().format
         self.pycolorize = lambda src: pyformat(src,'str',self.colors)
 
     def init_pushd_popd_magic(self):
         # for pushd/popd management
         self.home_dir = get_home_dir()
 
         self.dir_stack = []
 
     def init_logger(self):
         self.logger = Logger(self.home_dir, logfname='ipython_log.py',
                              logmode='rotate')
 
     def init_logstart(self):
         """Initialize logging in case it was requested at the command line.
         """
         if self.logappend:
             self.magic_logstart(self.logappend + ' append')
         elif self.logfile:
             self.magic_logstart(self.logfile)
         elif self.logstart:
             self.magic_logstart()
 
     def init_builtins(self):
         # A single, static flag that we set to True.  Its presence indicates
         # that an IPython shell has been created, and we make no attempts at
         # removing on exit or representing the existence of more than one
         # IPython at a time.
         builtin_mod.__dict__['__IPYTHON__'] = True
 
         # In 0.11 we introduced '__IPYTHON__active' as an integer we'd try to
         # manage on enter/exit, but with all our shells it's virtually
         # impossible to get all the cases right.  We're leaving the name in for
         # those who adapted their codes to check for this flag, but will
         # eventually remove it after a few more releases.
         builtin_mod.__dict__['__IPYTHON__active'] = \
                                           'Deprecated, check for __IPYTHON__'
 
         self.builtin_trap = BuiltinTrap(shell=self)
 
     def init_inspector(self):
         # Object inspector
         self.inspector = oinspect.Inspector(oinspect.InspectColors,
                                             PyColorize.ANSICodeColors,
                                             'NoColor',
                                             self.object_info_string_level)
 
     def init_io(self):
         # This will just use sys.stdout and sys.stderr. If you want to
         # override sys.stdout and sys.stderr themselves, you need to do that
         # *before* instantiating this class, because io holds onto
         # references to the underlying streams.
         if sys.platform == 'win32' and self.has_readline:
             io.stdout = io.stderr = io.IOStream(self.readline._outputfile)
         else:
             io.stdout = io.IOStream(sys.stdout)
             io.stderr = io.IOStream(sys.stderr)
 
     def init_prompts(self):
         self.prompt_manager = PromptManager(shell=self, config=self.config)
         self.configurables.append(self.prompt_manager)
         # Set system prompts, so that scripts can decide if they are running
         # interactively.
         sys.ps1 = 'In : '
         sys.ps2 = '...: '
         sys.ps3 = 'Out: '
 
     def init_display_formatter(self):
         self.display_formatter = DisplayFormatter(config=self.config)
         self.configurables.append(self.display_formatter)
 
     def init_display_pub(self):
         self.display_pub = self.display_pub_class(config=self.config)
         self.configurables.append(self.display_pub)
 
     def init_displayhook(self):
         # Initialize displayhook, set in/out prompts and printing system
         self.displayhook = self.displayhook_class(
             config=self.config,
             shell=self,
             cache_size=self.cache_size,
         )
         self.configurables.append(self.displayhook)
         # This is a context manager that installs/revmoes the displayhook at
         # the appropriate time.
         self.display_trap = DisplayTrap(hook=self.displayhook)
 
     def init_reload_doctest(self):
         # Do a proper resetting of doctest, including the necessary displayhook
         # monkeypatching
         try:
             doctest_reload()
         except ImportError:
             warn("doctest module does not exist.")
     
     def init_virtualenv(self):
         """Add a virtualenv to sys.path so the user can import modules from it.
         This isn't perfect: it doesn't use the Python interpreter with which the
         virtualenv was built, and it ignores the --no-site-packages option. A
         warning will appear suggesting the user installs IPython in the
         virtualenv, but for many cases, it probably works well enough.
         
         Adapted from code snippets online.
         
         http://blog.ufsoft.org/2009/1/29/ipython-and-virtualenv
         """
         if 'VIRTUAL_ENV' not in os.environ:
             # Not in a virtualenv
             return
         
         if sys.executable.startswith(os.environ['VIRTUAL_ENV']):
             # Running properly in the virtualenv, don't need to do anything
             return
         
         warn("Attempting to work in a virtualenv. If you encounter problems, please "
              "install IPython inside the virtualenv.\n")
         if sys.platform == "win32":
             virtual_env = os.path.join(os.environ['VIRTUAL_ENV'], 'Lib', 'site-packages') 
         else:
             virtual_env = os.path.join(os.environ['VIRTUAL_ENV'], 'lib',
                        'python%d.%d' % sys.version_info[:2], 'site-packages')
         
         import site
         sys.path.insert(0, virtual_env)
         site.addsitedir(virtual_env)
 
     #-------------------------------------------------------------------------
     # Things related to injections into the sys module
     #-------------------------------------------------------------------------
 
     def save_sys_module_state(self):
         """Save the state of hooks in the sys module.
 
         This has to be called after self.user_module is created.
         """
         self._orig_sys_module_state = {}
         self._orig_sys_module_state['stdin'] = sys.stdin
         self._orig_sys_module_state['stdout'] = sys.stdout
         self._orig_sys_module_state['stderr'] = sys.stderr
         self._orig_sys_module_state['excepthook'] = sys.excepthook
         self._orig_sys_modules_main_name = self.user_module.__name__
+        self._orig_sys_modules_main_mod = sys.modules.get(self.user_module.__name__)
 
     def restore_sys_module_state(self):
         """Restore the state of the sys module."""
         try:
             for k, v in self._orig_sys_module_state.iteritems():
                 setattr(sys, k, v)
         except AttributeError:
             pass
         # Reset what what done in self.init_sys_modules
-        sys.modules[self.user_module.__name__] = self._orig_sys_modules_main_name
+        if self._orig_sys_modules_main_mod is not None:
+            sys.modules[self._orig_sys_modules_main_name] = self._orig_sys_modules_main_mod
 
     #-------------------------------------------------------------------------
     # Things related to hooks
     #-------------------------------------------------------------------------
 
     def init_hooks(self):
         # hooks holds pointers used for user-side customizations
         self.hooks = Struct()
 
         self.strdispatchers = {}
 
         # Set all default hooks, defined in the IPython.hooks module.
         hooks = IPython.core.hooks
         for hook_name in hooks.__all__:
             # default hooks have priority 100, i.e. low; user hooks should have
             # 0-100 priority
             self.set_hook(hook_name,getattr(hooks,hook_name), 100)
 
     def set_hook(self,name,hook, priority = 50, str_key = None, re_key = None):
         """set_hook(name,hook) -> sets an internal IPython hook.
 
         IPython exposes some of its internal API as user-modifiable hooks.  By
         adding your function to one of these hooks, you can modify IPython's
         behavior to call at runtime your own routines."""
 
         # At some point in the future, this should validate the hook before it
         # accepts it.  Probably at least check that the hook takes the number
         # of args it's supposed to.
 
         f = types.MethodType(hook,self)
 
         # check if the hook is for strdispatcher first
         if str_key is not None:
             sdp = self.strdispatchers.get(name, StrDispatch())
             sdp.add_s(str_key, f, priority )
             self.strdispatchers[name] = sdp
             return
         if re_key is not None:
             sdp = self.strdispatchers.get(name, StrDispatch())
             sdp.add_re(re.compile(re_key), f, priority )
             self.strdispatchers[name] = sdp
             return
 
         dp = getattr(self.hooks, name, None)
         if name not in IPython.core.hooks.__all__:
             print "Warning! Hook '%s' is not one of %s" % \
                   (name, IPython.core.hooks.__all__ )
         if not dp:
             dp = IPython.core.hooks.CommandChainDispatcher()
 
         try:
             dp.add(f,priority)
         except AttributeError:
             # it was not commandchain, plain old func - replace
             dp = f
 
         setattr(self.hooks,name, dp)
 
     def register_post_execute(self, func):
         """Register a function for calling after code execution.
         """
         if not callable(func):
             raise ValueError('argument %s must be callable' % func)
         self._post_execute[func] = True
     
     #-------------------------------------------------------------------------
     # Things related to the "main" module
     #-------------------------------------------------------------------------
 
     def new_main_mod(self,ns=None):
         """Return a new 'main' module object for user code execution.
         """
         main_mod = self._user_main_module
         init_fakemod_dict(main_mod,ns)
         return main_mod
 
     def cache_main_mod(self,ns,fname):
         """Cache a main module's namespace.
 
         When scripts are executed via %run, we must keep a reference to the
         namespace of their __main__ module (a FakeModule instance) around so
         that Python doesn't clear it, rendering objects defined therein
         useless.
 
         This method keeps said reference in a private dict, keyed by the
         absolute path of the module object (which corresponds to the script
         path).  This way, for multiple executions of the same script we only
         keep one copy of the namespace (the last one), thus preventing memory
         leaks from old references while allowing the objects from the last
         execution to be accessible.
 
         Note: we can not allow the actual FakeModule instances to be deleted,
         because of how Python tears down modules (it hard-sets all their
         references to None without regard for reference counts).  This method
         must therefore make a *copy* of the given namespace, to allow the
         original module's __dict__ to be cleared and reused.
 
 
         Parameters
         ----------
           ns : a namespace (a dict, typically)
 
           fname : str
             Filename associated with the namespace.
 
         Examples
         --------
 
         In [10]: import IPython
 
         In [11]: _ip.cache_main_mod(IPython.__dict__,IPython.__file__)
 
         In [12]: IPython.__file__ in _ip._main_ns_cache
         Out[12]: True
         """
         self._main_ns_cache[os.path.abspath(fname)] = ns.copy()
 
     def clear_main_mod_cache(self):
         """Clear the cache of main modules.
 
         Mainly for use by utilities like %reset.
 
         Examples
         --------
 
         In [15]: import IPython
 
         In [16]: _ip.cache_main_mod(IPython.__dict__,IPython.__file__)
 
         In [17]: len(_ip._main_ns_cache) > 0
         Out[17]: True
 
         In [18]: _ip.clear_main_mod_cache()
 
         In [19]: len(_ip._main_ns_cache) == 0
         Out[19]: True
         """
         self._main_ns_cache.clear()
 
     #-------------------------------------------------------------------------
     # Things related to debugging
     #-------------------------------------------------------------------------
 
     def init_pdb(self):
         # Set calling of pdb on exceptions
         # self.call_pdb is a property
         self.call_pdb = self.pdb
 
     def _get_call_pdb(self):
         return self._call_pdb
 
     def _set_call_pdb(self,val):
 
         if val not in (0,1,False,True):
             raise ValueError,'new call_pdb value must be boolean'
 
         # store value in instance
         self._call_pdb = val
 
         # notify the actual exception handlers
         self.InteractiveTB.call_pdb = val
 
     call_pdb = property(_get_call_pdb,_set_call_pdb,None,
                         'Control auto-activation of pdb at exceptions')
 
     def debugger(self,force=False):
         """Call the pydb/pdb debugger.
 
         Keywords:
 
           - force(False): by default, this routine checks the instance call_pdb
           flag and does not actually invoke the debugger if the flag is false.
           The 'force' option forces the debugger to activate even if the flag
           is false.
         """
 
         if not (force or self.call_pdb):
             return
 
         if not hasattr(sys,'last_traceback'):
             error('No traceback has been produced, nothing to debug.')
             return
 
         # use pydb if available
         if debugger.has_pydb:
             from pydb import pm
         else:
             # fallback to our internal debugger
             pm = lambda : self.InteractiveTB.debugger(force=True)
 
         with self.readline_no_record:
             pm()
 
     #-------------------------------------------------------------------------
     # Things related to IPython's various namespaces
     #-------------------------------------------------------------------------
     default_user_namespaces = True
 
     def init_create_namespaces(self, user_module=None, user_ns=None):
         # Create the namespace where the user will operate.  user_ns is
         # normally the only one used, and it is passed to the exec calls as
         # the locals argument.  But we do carry a user_global_ns namespace
         # given as the exec 'globals' argument,  This is useful in embedding
         # situations where the ipython shell opens in a context where the
         # distinction between locals and globals is meaningful.  For
         # non-embedded contexts, it is just the same object as the user_ns dict.
 
         # FIXME. For some strange reason, __builtins__ is showing up at user
         # level as a dict instead of a module. This is a manual fix, but I
         # should really track down where the problem is coming from. Alex
         # Schmolck reported this problem first.
 
         # A useful post by Alex Martelli on this topic:
         # Re: inconsistent value from __builtins__
         # Von: Alex Martelli <[e-mail]>
         # Datum: Freitag 01 Oktober 2004 04:45:34 nachmittags/abends
         # Gruppen: comp.lang.python
 
         # Michael Hohn <[e-mail]> wrote:
         # > >>> print type(builtin_check.get_global_binding('__builtins__'))
         # > <type 'dict'>
         # > >>> print type(__builtins__)
         # > <type 'module'>
         # > Is this difference in return value intentional?
 
         # Well, it's documented that '__builtins__' can be either a dictionary
         # or a module, and it's been that way for a long time. Whether it's
         # intentional (or sensible), I don't know. In any case, the idea is
         # that if you need to access the built-in namespace directly, you
         # should start with "import __builtin__" (note, no 's') which will
         # definitely give you a module. Yeah, it's somewhat confusing:-(.
 
         # These routines return a properly built module and dict as needed by
         # the rest of the code, and can also be used by extension writers to
         # generate properly initialized namespaces.
         if (user_ns is not None) or (user_module is not None):
             self.default_user_namespaces = False
         self.user_module, self.user_ns = self.prepare_user_module(user_module, user_ns)
 
         # A record of hidden variables we have added to the user namespace, so
         # we can list later only variables defined in actual interactive use.
         self.user_ns_hidden = set()
 
         # Now that FakeModule produces a real module, we've run into a nasty
         # problem: after script execution (via %run), the module where the user
         # code ran is deleted.  Now that this object is a true module (needed
         # so docetst and other tools work correctly), the Python module
         # teardown mechanism runs over it, and sets to None every variable
         # present in that module.  Top-level references to objects from the
         # script survive, because the user_ns is updated with them.  However,
         # calling functions defined in the script that use other things from
         # the script will fail, because the function's closure had references
         # to the original objects, which are now all None.  So we must protect
         # these modules from deletion by keeping a cache.
         #
         # To avoid keeping stale modules around (we only need the one from the
         # last run), we use a dict keyed with the full path to the script, so
         # only the last version of the module is held in the cache.  Note,
         # however, that we must cache the module *namespace contents* (their
         # __dict__).  Because if we try to cache the actual modules, old ones
         # (uncached) could be destroyed while still holding references (such as
         # those held by GUI objects that tend to be long-lived)>
         #
         # The %reset command will flush this cache.  See the cache_main_mod()
         # and clear_main_mod_cache() methods for details on use.
 
         # This is the cache used for 'main' namespaces
         self._main_ns_cache = {}
         # And this is the single instance of FakeModule whose __dict__ we keep
         # copying and clearing for reuse on each %run
         self._user_main_module = FakeModule()
 
         # A table holding all the namespaces IPython deals with, so that
         # introspection facilities can search easily.
         self.ns_table = {'user_global':self.user_module.__dict__,
                          'user_local':self.user_ns,
                          'builtin':builtin_mod.__dict__
                          }
     
     @property
     def user_global_ns(self):
         return self.user_module.__dict__
 
     def prepare_user_module(self, user_module=None, user_ns=None):
         """Prepare the module and namespace in which user code will be run.
         
         When IPython is started normally, both parameters are None: a new module
         is created automatically, and its __dict__ used as the namespace.
         
         If only user_module is provided, its __dict__ is used as the namespace.
         If only user_ns is provided, a dummy module is created, and user_ns
         becomes the global namespace. If both are provided (as they may be
         when embedding), user_ns is the local namespace, and user_module
         provides the global namespace.
 
         Parameters
         ----------
         user_module : module, optional
             The current user module in which IPython is being run. If None,
             a clean module will be created.
         user_ns : dict, optional
             A namespace in which to run interactive commands.
 
         Returns
         -------
         A tuple of user_module and user_ns, each properly initialised.
         """
         if user_module is None and user_ns is not None:
             user_ns.setdefault("__name__", "__main__")
             class DummyMod(object):
                 "A dummy module used for IPython's interactive namespace."
                 pass
             user_module = DummyMod()
             user_module.__dict__ = user_ns
             
         if user_module is None:
             user_module = types.ModuleType("__main__",
                 doc="Automatically created module for IPython interactive environment")
         
         # We must ensure that __builtin__ (without the final 's') is always
         # available and pointing to the __builtin__ *module*.  For more details:
         # http://mail.python.org/pipermail/python-dev/2001-April/014068.html
         user_module.__dict__.setdefault('__builtin__', builtin_mod)
         user_module.__dict__.setdefault('__builtins__', builtin_mod)
         
         if user_ns is None:
             user_ns = user_module.__dict__
 
         return user_module, user_ns
 
     def init_sys_modules(self):
         # We need to insert into sys.modules something that looks like a
         # module but which accesses the IPython namespace, for shelve and
         # pickle to work interactively. Normally they rely on getting
         # everything out of __main__, but for embedding purposes each IPython
         # instance has its own private namespace, so we can't go shoving
         # everything into __main__.
 
         # note, however, that we should only do this for non-embedded
         # ipythons, which really mimic the __main__.__dict__ with their own
         # namespace.  Embedded instances, on the other hand, should not do
         # this because they need to manage the user local/global namespaces
         # only, but they live within a 'normal' __main__ (meaning, they
         # shouldn't overtake the execution environment of the script they're
         # embedded in).
 
         # This is overridden in the InteractiveShellEmbed subclass to a no-op.
         main_name = self.user_module.__name__
         sys.modules[main_name] = self.user_module
 
     def init_user_ns(self):
         """Initialize all user-visible namespaces to their minimum defaults.
 
         Certain history lists are also initialized here, as they effectively
         act as user namespaces.
 
         Notes
         -----
         All data structures here are only filled in, they are NOT reset by this
         method.  If they were not empty before, data will simply be added to
         therm.
         """
         # This function works in two parts: first we put a few things in
         # user_ns, and we sync that contents into user_ns_hidden so that these
         # initial variables aren't shown by %who.  After the sync, we add the
         # rest of what we *do* want the user to see with %who even on a new
         # session (probably nothing, so theye really only see their own stuff)
 
         # The user dict must *always* have a __builtin__ reference to the
         # Python standard __builtin__ namespace,  which must be imported.
         # This is so that certain operations in prompt evaluation can be
         # reliably executed with builtins.  Note that we can NOT use
         # __builtins__ (note the 's'),  because that can either be a dict or a
         # module, and can even mutate at runtime, depending on the context
         # (Python makes no guarantees on it).  In contrast, __builtin__ is
         # always a module object, though it must be explicitly imported.
 
         # For more details:
         # http://mail.python.org/pipermail/python-dev/2001-April/014068.html
         ns = dict()
         
         # Put 'help' in the user namespace
         try:
             from site import _Helper
             ns['help'] = _Helper()
         except ImportError:
             warn('help() not available - check site.py')
 
         # make global variables for user access to the histories
         ns['_ih'] = self.history_manager.input_hist_parsed
         ns['_oh'] = self.history_manager.output_hist
         ns['_dh'] = self.history_manager.dir_hist
 
         ns['_sh'] = shadowns
 
         # user aliases to input and output histories.  These shouldn't show up
         # in %who, as they can have very large reprs.
         ns['In']  = self.history_manager.input_hist_parsed
         ns['Out'] = self.history_manager.output_hist
 
         # Store myself as the public api!!!
         ns['get_ipython'] = self.get_ipython
         
         ns['exit'] = self.exiter
         ns['quit'] = self.exiter
 
         # Sync what we've added so far to user_ns_hidden so these aren't seen
         # by %who
         self.user_ns_hidden.update(ns)
 
         # Anything put into ns now would show up in %who.  Think twice before
         # putting anything here, as we really want %who to show the user their
         # stuff, not our variables.
 
         # Finally, update the real user's namespace
         self.user_ns.update(ns)
     
     @property
     def all_ns_refs(self):
         """Get a list of references to all the namespace dictionaries in which
         IPython might store a user-created object.
         
         Note that this does not include the displayhook, which also caches
         objects from the output."""
         return [self.user_ns, self.user_global_ns,
                 self._user_main_module.__dict__] + self._main_ns_cache.values()
 
     def reset(self, new_session=True):
         """Clear all internal namespaces, and attempt to release references to
         user objects.
 
         If new_session is True, a new history session will be opened.
         """
         # Clear histories
         self.history_manager.reset(new_session)
         # Reset counter used to index all histories
         if new_session:
             self.execution_count = 1
 
         # Flush cached output items
         if self.displayhook.do_full_cache:
             self.displayhook.flush()
 
         # The main execution namespaces must be cleared very carefully,
         # skipping the deletion of the builtin-related keys, because doing so
         # would cause errors in many object's __del__ methods.
         if self.user_ns is not self.user_global_ns:
             self.user_ns.clear()
         ns = self.user_global_ns
         drop_keys = set(ns.keys())
         drop_keys.discard('__builtin__')
         drop_keys.discard('__builtins__')
         drop_keys.discard('__name__')
         for k in drop_keys:
             del ns[k]
         
         self.user_ns_hidden.clear()
         
         # Restore the user namespaces to minimal usability
         self.init_user_ns()
 
         # Restore the default and user aliases
         self.alias_manager.clear_aliases()
         self.alias_manager.init_aliases()
 
         # Flush the private list of module references kept for script
         # execution protection
         self.clear_main_mod_cache()
 
         # Clear out the namespace from the last %run
         self.new_main_mod()
 
     def del_var(self, varname, by_name=False):
         """Delete a variable from the various namespaces, so that, as
         far as possible, we're not keeping any hidden references to it.
 
         Parameters
         ----------
         varname : str
             The name of the variable to delete.
         by_name : bool
             If True, delete variables with the given name in each
             namespace. If False (default), find the variable in the user
             namespace, and delete references to it.
         """
         if varname in ('__builtin__', '__builtins__'):
             raise ValueError("Refusing to delete %s" % varname)
 
         ns_refs = self.all_ns_refs
         
         if by_name:                    # Delete by name
             for ns in ns_refs:
                 try:
                     del ns[varname]
                 except KeyError:
                     pass
         else:                         # Delete by object
             try:
                 obj = self.user_ns[varname]
             except KeyError:
                 raise NameError("name '%s' is not defined" % varname)
             # Also check in output history
             ns_refs.append(self.history_manager.output_hist)
             for ns in ns_refs:
                 to_delete = [n for n, o in ns.iteritems() if o is obj]
                 for name in to_delete:
                     del ns[name]
 
             # displayhook keeps extra references, but not in a dictionary
             for name in ('_', '__', '___'):
                 if getattr(self.displayhook, name) is obj:
                     setattr(self.displayhook, name, None)
 
     def reset_selective(self, regex=None):
         """Clear selective variables from internal namespaces based on a
         specified regular expression.
 
         Parameters
         ----------
         regex : string or compiled pattern, optional
             A regular expression pattern that will be used in searching
             variable names in the users namespaces.
         """
         if regex is not None:
             try:
                 m = re.compile(regex)
             except TypeError:
                 raise TypeError('regex must be a string or compiled pattern')
             # Search for keys in each namespace that match the given regex
             # If a match is found, delete the key/value pair.
             for ns in self.all_ns_refs:
                 for var in ns:
                     if m.search(var):
                         del ns[var]
 
     def push(self, variables, interactive=True):
         """Inject a group of variables into the IPython user namespace.
 
         Parameters
         ----------
         variables : dict, str or list/tuple of str
             The variables to inject into the user's namespace.  If a dict, a
             simple update is done.  If a str, the string is assumed to have
             variable names separated by spaces.  A list/tuple of str can also
             be used to give the variable names.  If just the variable names are
             give (list/tuple/str) then the variable values looked up in the
             callers frame.
         interactive : bool
             If True (default), the variables will be listed with the ``who``
             magic.
         """
         vdict = None
 
         # We need a dict of name/value pairs to do namespace updates.
         if isinstance(variables, dict):
             vdict = variables
         elif isinstance(variables, (basestring, list, tuple)):
             if isinstance(variables, basestring):
                 vlist = variables.split()
             else:
                 vlist = variables
             vdict = {}
             cf = sys._getframe(1)
             for name in vlist:
                 try:
                     vdict[name] = eval(name, cf.f_globals, cf.f_locals)
                 except:
                     print ('Could not get variable %s from %s' %
                            (name,cf.f_code.co_name))
         else:
             raise ValueError('variables must be a dict/str/list/tuple')
 
         # Propagate variables to user namespace
         self.user_ns.update(vdict)
 
         # And configure interactive visibility
         user_ns_hidden = self.user_ns_hidden
         if interactive:
             user_ns_hidden.difference_update(vdict)
         else:
             user_ns_hidden.update(vdict)
     
     def drop_by_id(self, variables):
         """Remove a dict of variables from the user namespace, if they are the
         same as the values in the dictionary.
         
         This is intended for use by extensions: variables that they've added can
         be taken back out if they are unloaded, without removing any that the
         user has overwritten.
         
         Parameters
         ----------
         variables : dict
           A dictionary mapping object names (as strings) to the objects.
         """
         for name, obj in variables.iteritems():
             if name in self.user_ns and self.user_ns[name] is obj:
                 del self.user_ns[name]
                 self.user_ns_hidden.discard(name)
 
     #-------------------------------------------------------------------------
     # Things related to object introspection
     #-------------------------------------------------------------------------
 
     def _ofind(self, oname, namespaces=None):
         """Find an object in the available namespaces.
 
         self._ofind(oname) -> dict with keys: found,obj,ospace,ismagic
 
         Has special code to detect magic functions.
         """
         oname = oname.strip()
         #print '1- oname: <%r>' % oname  # dbg
         if not py3compat.isidentifier(oname.lstrip(ESC_MAGIC), dotted=True):
             return dict(found=False)
 
         alias_ns = None
         if namespaces is None:
             # Namespaces to search in:
             # Put them in a list. The order is important so that we
             # find things in the same order that Python finds them.
             namespaces = [ ('Interactive', self.user_ns),
                            ('Interactive (global)', self.user_global_ns),
                            ('Python builtin', builtin_mod.__dict__),
                            ('Alias', self.alias_manager.alias_table),
                            ]
             alias_ns = self.alias_manager.alias_table
 
         # initialize results to 'null'
         found = False; obj = None;  ospace = None;  ds = None;
         ismagic = False; isalias = False; parent = None
 
         # We need to special-case 'print', which as of python2.6 registers as a
         # function but should only be treated as one if print_function was
         # loaded with a future import.  In this case, just bail.
         if (oname == 'print' and not py3compat.PY3 and not \
             (self.compile.compiler_flags & __future__.CO_FUTURE_PRINT_FUNCTION)):
             return {'found':found, 'obj':obj, 'namespace':ospace,
                     'ismagic':ismagic, 'isalias':isalias, 'parent':parent}
 
         # Look for the given name by splitting it in parts.  If the head is
         # found, then we look for all the remaining parts as members, and only
         # declare success if we can find them all.
         oname_parts = oname.split('.')
         oname_head, oname_rest = oname_parts[0],oname_parts[1:]
         for nsname,ns in namespaces:
             try:
                 obj = ns[oname_head]
             except KeyError:
                 continue
             else:
                 #print 'oname_rest:', oname_rest  # dbg
                 for part in oname_rest:
                     try:
                         parent = obj
                         obj = getattr(obj,part)
                     except:
                         # Blanket except b/c some badly implemented objects
                         # allow __getattr__ to raise exceptions other than
                         # AttributeError, which then crashes IPython.
                         break
                 else:
                     # If we finish the for loop (no break), we got all members
                     found = True
                     ospace = nsname
                     if ns == alias_ns:
                         isalias = True
                     break  # namespace loop
 
         # Try to see if it's magic
         if not found:
             if oname.startswith(ESC_MAGIC):
                 oname = oname[1:]
             obj = getattr(self,'magic_'+oname,None)
             if obj is not None:
                 found = True
                 ospace = 'IPython internal'
                 ismagic = True
 
         # Last try: special-case some literals like '', [], {}, etc:
         if not found and oname_head in ["''",'""','[]','{}','()']:
             obj = eval(oname_head)
             found = True
             ospace = 'Interactive'
 
         return {'found':found, 'obj':obj, 'namespace':ospace,
                 'ismagic':ismagic, 'isalias':isalias, 'parent':parent}
 
     def _ofind_property(self, oname, info):
         """Second part of object finding, to look for property details."""
         if info.found:
             # Get the docstring of the class property if it exists.
             path = oname.split('.')
             root = '.'.join(path[:-1])
             if info.parent is not None:
                 try:
                     target = getattr(info.parent, '__class__')
                     # The object belongs to a class instance.
                     try:
                         target = getattr(target, path[-1])
                         # The class defines the object.
                         if isinstance(target, property):
                             oname = root + '.__class__.' + path[-1]
                             info = Struct(self._ofind(oname))
                     except AttributeError: pass
                 except AttributeError: pass
 
         # We return either the new info or the unmodified input if the object
         # hadn't been found
         return info
 
     def _object_find(self, oname, namespaces=None):
         """Find an object and return a struct with info about it."""
         inf = Struct(self._ofind(oname, namespaces))
         return Struct(self._ofind_property(oname, inf))
 
     def _inspect(self, meth, oname, namespaces=None, **kw):
         """Generic interface to the inspector system.
 
         This function is meant to be called by pdef, pdoc & friends."""
         info = self._object_find(oname)
         if info.found:
             pmethod = getattr(self.inspector, meth)
             formatter = format_screen if info.ismagic else None
             if meth == 'pdoc':
                 pmethod(info.obj, oname, formatter)
             elif meth == 'pinfo':
                 pmethod(info.obj, oname, formatter, info, **kw)
             else:
                 pmethod(info.obj, oname)
         else:
             print 'Object `%s` not found.' % oname
             return 'not found'  # so callers can take other action
 
     def object_inspect(self, oname, detail_level=0):
         with self.builtin_trap:
             info = self._object_find(oname)
             if info.found:
                 return self.inspector.info(info.obj, oname, info=info,
                             detail_level=detail_level
                 )
             else:
                 return oinspect.object_info(name=oname, found=False)
 
     #-------------------------------------------------------------------------
     # Things related to history management
     #-------------------------------------------------------------------------
 
     def init_history(self):
         """Sets up the command history, and starts regular autosaves."""
         self.history_manager = HistoryManager(shell=self, config=self.config)
         self.configurables.append(self.history_manager)
 
     #-------------------------------------------------------------------------
     # Things related to exception handling and tracebacks (not debugging)
     #-------------------------------------------------------------------------
 
     def init_traceback_handlers(self, custom_exceptions):
         # Syntax error handler.
         self.SyntaxTB = ultratb.SyntaxTB(color_scheme='NoColor')
 
         # The interactive one is initialized with an offset, meaning we always
         # want to remove the topmost item in the traceback, which is our own
         # internal code. Valid modes: ['Plain','Context','Verbose']
         self.InteractiveTB = ultratb.AutoFormattedTB(mode = 'Plain',
                                                      color_scheme='NoColor',
                                                      tb_offset = 1,
                                    check_cache=self.compile.check_cache)
 
         # The instance will store a pointer to the system-wide exception hook,
         # so that runtime code (such as magics) can access it.  This is because
         # during the read-eval loop, it may get temporarily overwritten.
         self.sys_excepthook = sys.excepthook
 
         # and add any custom exception handlers the user may have specified
         self.set_custom_exc(*custom_exceptions)
 
         # Set the exception mode
         self.InteractiveTB.set_mode(mode=self.xmode)
 
     def set_custom_exc(self, exc_tuple, handler):
         """set_custom_exc(exc_tuple,handler)
 
         Set a custom exception handler, which will be called if any of the
         exceptions in exc_tuple occur in the mainloop (specifically, in the
         run_code() method).
 
         Parameters
         ----------
 
         exc_tuple : tuple of exception classes
             A *tuple* of exception classes, for which to call the defined
             handler.  It is very important that you use a tuple, and NOT A
             LIST here, because of the way Python's except statement works.  If
             you only want to trap a single exception, use a singleton tuple::
 
                 exc_tuple == (MyCustomException,)
 
         handler : callable
             handler must have the following signature::
 
                 def my_handler(self, etype, value, tb, tb_offset=None):
                     ...
                     return structured_traceback
 
             Your handler must return a structured traceback (a list of strings),
             or None.
 
             This will be made into an instance method (via types.MethodType)
             of IPython itself, and it will be called if any of the exceptions
             listed in the exc_tuple are caught. If the handler is None, an
             internal basic one is used, which just prints basic info.
 
             To protect IPython from crashes, if your handler ever raises an
             exception or returns an invalid result, it will be immediately
             disabled.
 
         WARNING: by putting in your own exception handler into IPython's main
         execution loop, you run a very good chance of nasty crashes.  This
         facility should only be used if you really know what you are doing."""
 
         assert type(exc_tuple)==type(()) , \
                "The custom exceptions must be given AS A TUPLE."
 
         def dummy_handler(self,etype,value,tb,tb_offset=None):
             print '*** Simple custom exception handler ***'
             print 'Exception type :',etype
             print 'Exception value:',value
             print 'Traceback      :',tb
             #print 'Source code    :','\n'.join(self.buffer)
         
         def validate_stb(stb):
             """validate structured traceback return type
             
             return type of CustomTB *should* be a list of strings, but allow
             single strings or None, which are harmless.
             
             This function will *always* return a list of strings,
             and will raise a TypeError if stb is inappropriate.
             """
             msg = "CustomTB must return list of strings, not %r" % stb
             if stb is None:
                 return []
             elif isinstance(stb, basestring):
                 return [stb]
             elif not isinstance(stb, list):
                 raise TypeError(msg)
             # it's a list
             for line in stb:
                 # check every element
                 if not isinstance(line, basestring):
                     raise TypeError(msg)
             return stb
 
         if handler is None:
             wrapped = dummy_handler
         else:
             def wrapped(self,etype,value,tb,tb_offset=None):
                 """wrap CustomTB handler, to protect IPython from user code
                 
                 This makes it harder (but not impossible) for custom exception
                 handlers to crash IPython.
                 """
                 try:
                     stb = handler(self,etype,value,tb,tb_offset=tb_offset)
                     return validate_stb(stb)
                 except:
                     # clear custom handler immediately
                     self.set_custom_exc((), None)
                     print >> io.stderr, "Custom TB Handler failed, unregistering"
                     # show the exception in handler first
                     stb = self.InteractiveTB.structured_traceback(*sys.exc_info())
                     print >> io.stdout, self.InteractiveTB.stb2text(stb)
                     print >> io.stdout, "The original exception:"
                     stb = self.InteractiveTB.structured_traceback(
                                             (etype,value,tb), tb_offset=tb_offset
                     )
                 return stb
 
         self.CustomTB = types.MethodType(wrapped,self)
         self.custom_exceptions = exc_tuple
 
     def excepthook(self, etype, value, tb):
       """One more defense for GUI apps that call sys.excepthook.
 
       GUI frameworks like wxPython trap exceptions and call
       sys.excepthook themselves.  I guess this is a feature that
       enables them to keep running after exceptions that would
       otherwise kill their mainloop. This is a bother for IPython
       which excepts to catch all of the program exceptions with a try:
       except: statement.
 
       Normally, IPython sets sys.excepthook to a CrashHandler instance, so if
       any app directly invokes sys.excepthook, it will look to the user like
       IPython crashed.  In order to work around this, we can disable the
       CrashHandler and replace it with this excepthook instead, which prints a
       regular traceback using our InteractiveTB.  In this fashion, apps which
       call sys.excepthook will generate a regular-looking exception from
       IPython, and the CrashHandler will only be triggered by real IPython
       crashes.
 
       This hook should be used sparingly, only in places which are not likely
       to be true IPython errors.
       """
       self.showtraceback((etype,value,tb),tb_offset=0)
 
     def _get_exc_info(self, exc_tuple=None):
         """get exc_info from a given tuple, sys.exc_info() or sys.last_type etc.
         
         Ensures sys.last_type,value,traceback hold the exc_info we found,
         from whichever source.
         
         raises ValueError if none of these contain any information
         """
         if exc_tuple is None:
             etype, value, tb = sys.exc_info()
         else:
             etype, value, tb = exc_tuple
 
         if etype is None:
             if hasattr(sys, 'last_type'):
                 etype, value, tb = sys.last_type, sys.last_value, \
                                    sys.last_traceback
         
         if etype is None:
             raise ValueError("No exception to find")
         
         # Now store the exception info in sys.last_type etc.
         # WARNING: these variables are somewhat deprecated and not
         # necessarily safe to use in a threaded environment, but tools
         # like pdb depend on their existence, so let's set them.  If we
         # find problems in the field, we'll need to revisit their use.
         sys.last_type = etype
         sys.last_value = value
         sys.last_traceback = tb
         
         return etype, value, tb
     
 
     def showtraceback(self,exc_tuple = None,filename=None,tb_offset=None,
                       exception_only=False):
         """Display the exception that just occurred.
 
         If nothing is known about the exception, this is the method which
         should be used throughout the code for presenting user tracebacks,
         rather than directly invoking the InteractiveTB object.
 
         A specific showsyntaxerror() also exists, but this method can take
         care of calling it if needed, so unless you are explicitly catching a
         SyntaxError exception, don't try to analyze the stack manually and
         simply call this method."""
 
         try:
             try:
                 etype, value, tb = self._get_exc_info(exc_tuple)
             except ValueError:
                 self.write_err('No traceback available to show.\n')
                 return
 
             if etype is SyntaxError:
                 # Though this won't be called by syntax errors in the input
                 # line, there may be SyntaxError cases with imported code.
                 self.showsyntaxerror(filename)
             elif etype is UsageError:
                 self.write_err("UsageError: %s" % value)
             else:
                 if etype in self.custom_exceptions:
                     stb = self.CustomTB(etype, value, tb, tb_offset)
                 else:
                     if exception_only:
                         stb = ['An exception has occurred, use %tb to see '
                                'the full traceback.\n']
                         stb.extend(self.InteractiveTB.get_exception_only(etype,
                                                                          value))
                     else:
                         stb = self.InteractiveTB.structured_traceback(etype,
                                                 value, tb, tb_offset=tb_offset)
 
                         self._showtraceback(etype, value, stb)
                         if self.call_pdb:
                             # drop into debugger
                             self.debugger(force=True)
                         return
 
                 # Actually show the traceback
                 self._showtraceback(etype, value, stb)
 
         except KeyboardInterrupt:
             self.write_err("\nKeyboardInterrupt\n")
 
     def _showtraceback(self, etype, evalue, stb):
         """Actually show a traceback.
 
         Subclasses may override this method to put the traceback on a different
         place, like a side channel.
         """
         print >> io.stdout, self.InteractiveTB.stb2text(stb)
 
     def showsyntaxerror(self, filename=None):
         """Display the syntax error that just occurred.
 
         This doesn't display a stack trace because there isn't one.
 
         If a filename is given, it is stuffed in the exception instead
         of what was there before (because Python's parser always uses
         "<string>" when reading from a string).
         """
         etype, value, last_traceback = self._get_exc_info()
 
         if filename and etype is SyntaxError:
             try:
                 value.filename = filename
             except:
                 # Not the format we expect; leave it alone
                 pass
         
         stb = self.SyntaxTB.structured_traceback(etype, value, [])
         self._showtraceback(etype, value, stb)
 
     # This is overridden in TerminalInteractiveShell to show a message about
     # the %paste magic.
     def showindentationerror(self):
         """Called by run_cell when there's an IndentationError in code entered
         at the prompt.
 
         This is overridden in TerminalInteractiveShell to show a message about
         the %paste magic."""
         self.showsyntaxerror()
 
     #-------------------------------------------------------------------------
     # Things related to readline
     #-------------------------------------------------------------------------
 
     def init_readline(self):
         """Command history completion/saving/reloading."""
 
         if self.readline_use:
             import IPython.utils.rlineimpl as readline
 
         self.rl_next_input = None
         self.rl_do_indent = False
 
         if not self.readline_use or not readline.have_readline:
             self.has_readline = False
             self.readline = None
             # Set a number of methods that depend on readline to be no-op
             self.readline_no_record = no_op_context
             self.set_readline_completer = no_op
             self.set_custom_completer = no_op
             self.set_completer_frame = no_op
             if self.readline_use:
                 warn('Readline services not available or not loaded.')
         else:
             self.has_readline = True
             self.readline = readline
             sys.modules['readline'] = readline
 
             # Platform-specific configuration
             if os.name == 'nt':
                 # FIXME - check with Frederick to see if we can harmonize
                 # naming conventions with pyreadline to avoid this
                 # platform-dependent check
                 self.readline_startup_hook = readline.set_pre_input_hook
             else:
                 self.readline_startup_hook = readline.set_startup_hook
 
             # Load user's initrc file (readline config)
             # Or if libedit is used, load editrc.
             inputrc_name = os.environ.get('INPUTRC')
             if inputrc_name is None:
                 inputrc_name = '.inputrc'
                 if readline.uses_libedit:
                     inputrc_name = '.editrc'
                 inputrc_name = os.path.join(self.home_dir, inputrc_name)
             if os.path.isfile(inputrc_name):
                 try:
                     readline.read_init_file(inputrc_name)
                 except:
                     warn('Problems reading readline initialization file <%s>'
                          % inputrc_name)
 
             # Configure readline according to user's prefs
             # This is only done if GNU readline is being used.  If libedit
             # is being used (as on Leopard) the readline config is
             # not run as the syntax for libedit is different.
             if not readline.uses_libedit:
                 for rlcommand in self.readline_parse_and_bind:
                     #print "loading rl:",rlcommand  # dbg
                     readline.parse_and_bind(rlcommand)
 
             # Remove some chars from the delimiters list.  If we encounter
             # unicode chars, discard them.
             delims = readline.get_completer_delims()
             if not py3compat.PY3:
                 delims = delims.encode("ascii", "ignore")
             for d in self.readline_remove_delims:
                 delims = delims.replace(d, "")
             delims = delims.replace(ESC_MAGIC, '')
             readline.set_completer_delims(delims)
             # otherwise we end up with a monster history after a while:
             readline.set_history_length(self.history_length)
 
             self.refill_readline_hist()
             self.readline_no_record = ReadlineNoRecord(self)
 
         # Configure auto-indent for all platforms
         self.set_autoindent(self.autoindent)
 
     def refill_readline_hist(self):
         # Load the last 1000 lines from history
         self.readline.clear_history()
         stdin_encoding = sys.stdin.encoding or "utf-8"
         last_cell = u""
         for _, _, cell in self.history_manager.get_tail(1000,
                                                         include_latest=True):
             # Ignore blank lines and consecutive duplicates
             cell = cell.rstrip()
             if cell and (cell != last_cell):
                 if self.multiline_history:
                       self.readline.add_history(py3compat.unicode_to_str(cell,
                                                                 stdin_encoding))
                 else:
                     for line in cell.splitlines():
                         self.readline.add_history(py3compat.unicode_to_str(line,
                                                                 stdin_encoding))
                 last_cell = cell
 
     def set_next_input(self, s):
         """ Sets the 'default' input string for the next command line.
 
         Requires readline.
 
         Example:
 
         [D:\ipython]|1> _ip.set_next_input("Hello Word")
         [D:\ipython]|2> Hello Word_  # cursor is here
         """
         self.rl_next_input = py3compat.cast_bytes_py2(s)
 
     # Maybe move this to the terminal subclass?
     def pre_readline(self):
         """readline hook to be used at the start of each line.
 
         Currently it handles auto-indent only."""
 
         if self.rl_do_indent:
             self.readline.insert_text(self._indent_current_str())
         if self.rl_next_input is not None:
             self.readline.insert_text(self.rl_next_input)
             self.rl_next_input = None
 
     def _indent_current_str(self):
         """return the current level of indentation as a string"""
         return self.input_splitter.indent_spaces * ' '
 
     #-------------------------------------------------------------------------
     # Things related to text completion
     #-------------------------------------------------------------------------
 
     def init_completer(self):
         """Initialize the completion machinery.
 
         This creates completion machinery that can be used by client code,
         either interactively in-process (typically triggered by the readline
         library), programatically (such as in test suites) or out-of-prcess
         (typically over the network by remote frontends).
         """
         from IPython.core.completer import IPCompleter
         from IPython.core.completerlib import (module_completer,
                 magic_run_completer, cd_completer, reset_completer)
 
         self.Completer = IPCompleter(shell=self,
                                      namespace=self.user_ns,
                                      global_namespace=self.user_global_ns,
                                      alias_table=self.alias_manager.alias_table,
                                      use_readline=self.has_readline,
                                      config=self.config,
                                      )
         self.configurables.append(self.Completer)
 
         # Add custom completers to the basic ones built into IPCompleter
         sdisp = self.strdispatchers.get('complete_command', StrDispatch())
         self.strdispatchers['complete_command'] = sdisp
         self.Completer.custom_completers = sdisp
 
         self.set_hook('complete_command', module_completer, str_key = 'import')
         self.set_hook('complete_command', module_completer, str_key = 'from')
         self.set_hook('complete_command', magic_run_completer, str_key = '%run')
         self.set_hook('complete_command', cd_completer, str_key = '%cd')
         self.set_hook('complete_command', reset_completer, str_key = '%reset')
 
         # Only configure readline if we truly are using readline.  IPython can
         # do tab-completion over the network, in GUIs, etc, where readline
         # itself may be absent
         if self.has_readline:
             self.set_readline_completer()
 
     def complete(self, text, line=None, cursor_pos=None):
         """Return the completed text and a list of completions.
 
         Parameters
         ----------
 
            text : string
              A string of text to be completed on.  It can be given as empty and
              instead a line/position pair are given.  In this case, the
              completer itself will split the line like readline does.
 
            line : string, optional
              The complete line that text is part of.
 
            cursor_pos : int, optional
              The position of the cursor on the input line.
 
         Returns
         -------
           text : string
             The actual text that was completed.
 
           matches : list
             A sorted list with all possible completions.
 
         The optional arguments allow the completion to take more context into
         account, and are part of the low-level completion API.
 
         This is a wrapper around the completion mechanism, similar to what
         readline does at the command line when the TAB key is hit.  By
         exposing it as a method, it can be used by other non-readline
         environments (such as GUIs) for text completion.
 
         Simple usage example:
 
         In [1]: x = 'hello'
 
         In [2]: _ip.complete('x.l')
         Out[2]: ('x.l', ['x.ljust', 'x.lower', 'x.lstrip'])
         """
 
         # Inject names into __builtin__ so we can complete on the added names.
         with self.builtin_trap:
             return self.Completer.complete(text, line, cursor_pos)
 
     def set_custom_completer(self, completer, pos=0):
         """Adds a new custom completer function.
 
         The position argument (defaults to 0) is the index in the completers
         list where you want the completer to be inserted."""
 
         newcomp = types.MethodType(completer,self.Completer)
         self.Completer.matchers.insert(pos,newcomp)
 
     def set_readline_completer(self):
         """Reset readline's completer to be our own."""
         self.readline.set_completer(self.Completer.rlcomplete)
 
     def set_completer_frame(self, frame=None):
         """Set the frame of the completer."""
         if frame:
             self.Completer.namespace = frame.f_locals
             self.Completer.global_namespace = frame.f_globals
         else:
             self.Completer.namespace = self.user_ns
             self.Completer.global_namespace = self.user_global_ns
 
     #-------------------------------------------------------------------------
     # Things related to magics
     #-------------------------------------------------------------------------
 
     def init_magics(self):
         # FIXME: Move the color initialization to the DisplayHook, which
         # should be split into a prompt manager and displayhook. We probably
         # even need a centralize colors management object.
         self.magic_colors(self.colors)
         # History was moved to a separate module
         from IPython.core import history
         history.init_ipython(self)
 
     def magic(self, arg_s, next_input=None):
         """Call a magic function by name.
 
         Input: a string containing the name of the magic function to call and
         any additional arguments to be passed to the magic.
 
         magic('name -opt foo bar') is equivalent to typing at the ipython
         prompt:
 
         In[1]: %name -opt foo bar
 
         To call a magic without arguments, simply use magic('name').
 
         This provides a proper Python function to call IPython's magics in any
         valid Python code you can type at the interpreter, including loops and
         compound statements.
         """
         # Allow setting the next input - this is used if the user does `a=abs?`.
         # We do this first so that magic functions can override it.
         if next_input:
             self.set_next_input(next_input)
 
         magic_name, _, magic_args = arg_s.partition(' ')
         magic_name = magic_name.lstrip(prefilter.ESC_MAGIC)
 
         fn = getattr(self,'magic_'+magic_name,None)
         if fn is None:
             error("Magic function `%s` not found." % magic_name)
         else:
             magic_args = self.var_expand(magic_args,1)
             # Grab local namespace if we need it:
             if getattr(fn, "needs_local_scope", False):
                 self._magic_locals = sys._getframe(1).f_locals
             with self.builtin_trap:
                 result = fn(magic_args)
             # Ensure we're not keeping object references around:
             self._magic_locals = {}
             return result
 
     def define_magic(self, magicname, func):
         """Expose own function as magic function for ipython
         
         Example::
 
           def foo_impl(self,parameter_s=''):
               'My very own magic!. (Use docstrings, IPython reads them).'
               print 'Magic function. Passed parameter is between < >:'
               print '<%s>' % parameter_s
               print 'The self object is:', self
 
           ip.define_magic('foo',foo_impl)
         """
         im = types.MethodType(func,self)
         old = getattr(self, "magic_" + magicname, None)
         setattr(self, "magic_" + magicname, im)
         return old
 
     #-------------------------------------------------------------------------
     # Things related to macros
     #-------------------------------------------------------------------------
 
     def define_macro(self, name, themacro):
         """Define a new macro
 
         Parameters
         ----------
         name : str
             The name of the macro.
         themacro : str or Macro
             The action to do upon invoking the macro.  If a string, a new
             Macro object is created by passing the string to it.
         """
 
         from IPython.core import macro
 
         if isinstance(themacro, basestring):
             themacro = macro.Macro(themacro)
         if not isinstance(themacro, macro.Macro):
             raise ValueError('A macro must be a string or a Macro instance.')
         self.user_ns[name] = themacro
 
     #-------------------------------------------------------------------------
     # Things related to the running of system commands
     #-------------------------------------------------------------------------
 
     def system_piped(self, cmd):
         """Call the given cmd in a subprocess, piping stdout/err
 
         Parameters
         ----------
         cmd : str
           Command to execute (can not end in '&', as background processes are
           not supported.  Should not be a command that expects input
           other than simple text.
         """
         if cmd.rstrip().endswith('&'):
             # this is *far* from a rigorous test
             # We do not support backgrounding processes because we either use
             # pexpect or pipes to read from.  Users can always just call
             # os.system() or use ip.system=ip.system_raw
             # if they really want a background process.
             raise OSError("Background processes not supported.")
 
         # we explicitly do NOT return the subprocess status code, because
         # a non-None value would trigger :func:`sys.displayhook` calls.
         # Instead, we store the exit_code in user_ns.
         self.user_ns['_exit_code'] = system(self.var_expand(cmd, depth=2))
 
     def system_raw(self, cmd):
         """Call the given cmd in a subprocess using os.system
 
         Parameters
         ----------
         cmd : str
           Command to execute.
         """
         cmd = self.var_expand(cmd, depth=2)
         # protect os.system from UNC paths on Windows, which it can't handle:
         if sys.platform == 'win32':
             from IPython.utils._process_win32 import AvoidUNCPath
             with AvoidUNCPath() as path:
                 if path is not None:
                     cmd = '"pushd %s &&"%s' % (path, cmd)
                 cmd = py3compat.unicode_to_str(cmd)
                 ec = os.system(cmd)
         else:
             cmd = py3compat.unicode_to_str(cmd)
             ec = os.system(cmd)
         
         # We explicitly do NOT return the subprocess status code, because
         # a non-None value would trigger :func:`sys.displayhook` calls.
         # Instead, we store the exit_code in user_ns.
         self.user_ns['_exit_code'] = ec
 
     # use piped system by default, because it is better behaved
     system = system_piped
 
     def getoutput(self, cmd, split=True):
         """Get output (possibly including stderr) from a subprocess.
 
         Parameters
         ----------
         cmd : str
           Command to execute (can not end in '&', as background processes are
           not supported.
         split : bool, optional
 
           If True, split the output into an IPython SList.  Otherwise, an
           IPython LSString is returned.  These are objects similar to normal
           lists and strings, with a few convenience attributes for easier
           manipulation of line-based output.  You can use '?' on them for
           details.
           """
         if cmd.rstrip().endswith('&'):
             # this is *far* from a rigorous test
             raise OSError("Background processes not supported.")
         out = getoutput(self.var_expand(cmd, depth=2))
         if split:
             out = SList(out.splitlines())
         else:
             out = LSString(out)
         return out
 
     #-------------------------------------------------------------------------
     # Things related to aliases
     #-------------------------------------------------------------------------
 
     def init_alias(self):
         self.alias_manager = AliasManager(shell=self, config=self.config)
         self.configurables.append(self.alias_manager)
         self.ns_table['alias'] = self.alias_manager.alias_table,
 
     #-------------------------------------------------------------------------
     # Things related to extensions and plugins
     #-------------------------------------------------------------------------
 
     def init_extension_manager(self):
         self.extension_manager = ExtensionManager(shell=self, config=self.config)
         self.configurables.append(self.extension_manager)
 
     def init_plugin_manager(self):
         self.plugin_manager = PluginManager(config=self.config)
         self.configurables.append(self.plugin_manager)
         
 
     #-------------------------------------------------------------------------
     # Things related to payloads
     #-------------------------------------------------------------------------
 
     def init_payload(self):
         self.payload_manager = PayloadManager(config=self.config)
         self.configurables.append(self.payload_manager)
 
     #-------------------------------------------------------------------------
     # Things related to the prefilter
     #-------------------------------------------------------------------------
 
     def init_prefilter(self):
         self.prefilter_manager = PrefilterManager(shell=self, config=self.config)
         self.configurables.append(self.prefilter_manager)
         # Ultimately this will be refactored in the new interpreter code, but
         # for now, we should expose the main prefilter method (there's legacy
         # code out there that may rely on this).
         self.prefilter = self.prefilter_manager.prefilter_lines
 
     def auto_rewrite_input(self, cmd):
         """Print to the screen the rewritten form of the user's command.
 
         This shows visual feedback by rewriting input lines that cause
         automatic calling to kick in, like::
 
           /f x
 
         into::
 
           ------> f(x)
 
         after the user's input prompt.  This helps the user understand that the
         input line was transformed automatically by IPython.
         """
         if not self.show_rewritten_input:
             return
         
         rw = self.prompt_manager.render('rewrite') + cmd
 
         try:
             # plain ascii works better w/ pyreadline, on some machines, so
             # we use it and only print uncolored rewrite if we have unicode
             rw = str(rw)
             print >> io.stdout, rw
         except UnicodeEncodeError:
             print "------> " + cmd
 
     #-------------------------------------------------------------------------
     # Things related to extracting values/expressions from kernel and user_ns
     #-------------------------------------------------------------------------
 
     def _simple_error(self):
         etype, value = sys.exc_info()[:2]
         return u'[ERROR] {e.__name__}: {v}'.format(e=etype, v=value)
 
     def user_variables(self, names):
         """Get a list of variable names from the user's namespace.
 
         Parameters
         ----------
         names : list of strings
           A list of names of variables to be read from the user namespace.
 
         Returns
         -------
         A dict, keyed by the input names and with the repr() of each value.
         """
         out = {}
         user_ns = self.user_ns
         for varname in names:
             try:
                 value = repr(user_ns[varname])
             except:
                 value = self._simple_error()
             out[varname] = value
         return out
 
     def user_expressions(self, expressions):
         """Evaluate a dict of expressions in the user's namespace.
 
         Parameters
         ----------
         expressions : dict
           A dict with string keys and string values.  The expression values
           should be valid Python expressions, each of which will be evaluated
           in the user namespace.
 
         Returns
         -------
         A dict, keyed like the input expressions dict, with the repr() of each
         value.
         """
         out = {}
         user_ns = self.user_ns
         global_ns = self.user_global_ns
         for key, expr in expressions.iteritems():
             try:
                 value = repr(eval(expr, global_ns, user_ns))
             except:
                 value = self._simple_error()
             out[key] = value
         return out
 
     #-------------------------------------------------------------------------
     # Things related to the running of code
     #-------------------------------------------------------------------------
 
     def ex(self, cmd):
         """Execute a normal python statement in user namespace."""
         with self.builtin_trap:
             exec cmd in self.user_global_ns, self.user_ns
 
     def ev(self, expr):
         """Evaluate python expression expr in user namespace.
 
         Returns the result of evaluation
         """
         with self.builtin_trap:
             return eval(expr, self.user_global_ns, self.user_ns)
 
     def safe_execfile(self, fname, *where, **kw):
         """A safe version of the builtin execfile().
 
         This version will never throw an exception, but instead print
         helpful error messages to the screen.  This only works on pure
         Python files with the .py extension.
 
         Parameters
         ----------
         fname : string
             The name of the file to be executed.
         where : tuple
             One or two namespaces, passed to execfile() as (globals,locals).
             If only one is given, it is passed as both.
         exit_ignore : bool (False)
             If True, then silence SystemExit for non-zero status (it is always
             silenced for zero status, as it is so common).
         raise_exceptions : bool (False)
             If True raise exceptions everywhere. Meant for testing.
 
         """
         kw.setdefault('exit_ignore', False)
         kw.setdefault('raise_exceptions', False)
 
         fname = os.path.abspath(os.path.expanduser(fname))
 
         # Make sure we can open the file
         try:
             with open(fname) as thefile:
                 pass
         except:
             warn('Could not open file <%s> for safe execution.' % fname)
             return
 
         # Find things also in current directory.  This is needed to mimic the
         # behavior of running a script from the system command line, where
         # Python inserts the script's directory into sys.path
         dname = os.path.dirname(fname)
 
         with prepended_to_syspath(dname):
             try:
                 py3compat.execfile(fname,*where)
             except SystemExit, status:
                 # If the call was made with 0 or None exit status (sys.exit(0)
                 # or sys.exit() ), don't bother showing a traceback, as both of
                 # these are considered normal by the OS:
                 # > python -c'import sys;sys.exit(0)'; echo $?
                 # 0
                 # > python -c'import sys;sys.exit()'; echo $?
                 # 0
                 # For other exit status, we show the exception unless
                 # explicitly silenced, but only in short form.
                 if kw['raise_exceptions']:
                     raise
                 if status.code not in (0, None) and not kw['exit_ignore']:
                     self.showtraceback(exception_only=True)
             except:
                 if kw['raise_exceptions']:
                     raise
                 self.showtraceback()
 
     def safe_execfile_ipy(self, fname):
         """Like safe_execfile, but for .ipy files with IPython syntax.
 
         Parameters
         ----------
         fname : str
             The name of the file to execute.  The filename must have a
             .ipy extension.
         """
         fname = os.path.abspath(os.path.expanduser(fname))
 
         # Make sure we can open the file
         try:
             with open(fname) as thefile:
                 pass
         except:
             warn('Could not open file <%s> for safe execution.' % fname)
             return
 
         # Find things also in current directory.  This is needed to mimic the
         # behavior of running a script from the system command line, where
         # Python inserts the script's directory into sys.path
         dname = os.path.dirname(fname)
 
         with prepended_to_syspath(dname):
             try:
                 with open(fname) as thefile:
                     # self.run_cell currently captures all exceptions
                     # raised in user code.  It would be nice if there were
                     # versions of runlines, execfile that did raise, so
                     # we could catch the errors.
                     self.run_cell(thefile.read(), store_history=False)
             except:
                 self.showtraceback()
                 warn('Unknown failure executing file: <%s>' % fname)
 
     def safe_run_module(self, mod_name, where):
         """A safe version of runpy.run_module().
 
         This version will never throw an exception, but instead print
         helpful error messages to the screen.
 
         Parameters
         ----------
         mod_name : string
             The name of the module to be executed.
         where : dict
             The globals namespace.
         """
         try:
             where.update(
                 runpy.run_module(str(mod_name), run_name="__main__",
                                  alter_sys=True)
                 )
         except:
             self.showtraceback()
             warn('Unknown failure executing module: <%s>' % mod_name)
 
-    def run_cell(self, raw_cell, store_history=False):
+    def run_cell(self, raw_cell, store_history=False, silent=False):
         """Run a complete IPython cell.
 
         Parameters
         ----------
         raw_cell : str
           The code (including IPython code such as %magic functions) to run.
         store_history : bool
           If True, the raw and translated cell will be stored in IPython's
           history. For user code calling back into IPython's machinery, this
           should be set to False.
+        silent : bool
+          If True, avoid side-effets, such as implicit displayhooks, history,
+          and logging.  silent=True forces store_history=False.
         """
         if (not raw_cell) or raw_cell.isspace():
             return
+        
+        if silent:
+            store_history = False
 
         for line in raw_cell.splitlines():
             self.input_splitter.push(line)
         cell = self.input_splitter.source_reset()
 
         with self.builtin_trap:
             prefilter_failed = False
             if len(cell.splitlines()) == 1:
                 try:
                     # use prefilter_lines to handle trailing newlines
                     # restore trailing newline for ast.parse
                     cell = self.prefilter_manager.prefilter_lines(cell) + '\n'
                 except AliasError as e:
                     error(e)
                     prefilter_failed = True
                 except Exception:
                     # don't allow prefilter errors to crash IPython
                     self.showtraceback()
                     prefilter_failed = True
 
             # Store raw and processed history
             if store_history:
                 self.history_manager.store_inputs(self.execution_count,
                                                   cell, raw_cell)
-
-            self.logger.log(cell, raw_cell)
+            if not silent:
+                self.logger.log(cell, raw_cell)
 
             if not prefilter_failed:
                 # don't run if prefilter failed
                 cell_name = self.compile.cache(cell, self.execution_count)
 
                 with self.display_trap:
                     try:
                         code_ast = self.compile.ast_parse(cell, filename=cell_name)
                     except IndentationError:
                         self.showindentationerror()
                         if store_history:
                             self.execution_count += 1
                         return None
                     except (OverflowError, SyntaxError, ValueError, TypeError,
                             MemoryError):
                         self.showsyntaxerror()
                         if store_history:
                             self.execution_count += 1
                         return None
-
+                    
+                    interactivity = "none" if silent else "last_expr"
                     self.run_ast_nodes(code_ast.body, cell_name,
-                                       interactivity="last_expr")
-
+                                       interactivity=interactivity)
+                    
                     # Execute any registered post-execution functions.
-                    for func, status in self._post_execute.iteritems():
+                    # unless we are silent
+                    post_exec = [] if silent else self._post_execute.iteritems()
+                    
+                    for func, status in post_exec:
                         if self.disable_failing_post_execute and not status:
                             continue
                         try:
                             func()
                         except KeyboardInterrupt:
                             print >> io.stderr, "\nKeyboardInterrupt"
                         except Exception:
                             # register as failing:
                             self._post_execute[func] = False
                             self.showtraceback()
                             print >> io.stderr, '\n'.join([
                                 "post-execution function %r produced an error." % func,
                                 "If this problem persists, you can disable failing post-exec functions with:",
                                 "",
                                 "    get_ipython().disable_failing_post_execute = True"
                             ])
 
         if store_history:
             # Write output to the database. Does nothing unless
             # history output logging is enabled.
             self.history_manager.store_output(self.execution_count)
             # Each cell is a *single* input, regardless of how many lines it has
             self.execution_count += 1
 
     def run_ast_nodes(self, nodelist, cell_name, interactivity='last_expr'):
         """Run a sequence of AST nodes. The execution mode depends on the
         interactivity parameter.
 
         Parameters
         ----------
         nodelist : list
           A sequence of AST nodes to run.
         cell_name : str
           Will be passed to the compiler as the filename of the cell. Typically
           the value returned by ip.compile.cache(cell).
         interactivity : str
           'all', 'last', 'last_expr' or 'none', specifying which nodes should be
           run interactively (displaying output from expressions). 'last_expr'
           will run the last node interactively only if it is an expression (i.e.
           expressions in loops or other blocks are not displayed. Other values
           for this parameter will raise a ValueError.
         """
         if not nodelist:
             return
 
         if interactivity == 'last_expr':
             if isinstance(nodelist[-1], ast.Expr):
                 interactivity = "last"
             else:
                 interactivity = "none"
 
         if interactivity == 'none':
             to_run_exec, to_run_interactive = nodelist, []
         elif interactivity == 'last':
             to_run_exec, to_run_interactive = nodelist[:-1], nodelist[-1:]
         elif interactivity == 'all':
             to_run_exec, to_run_interactive = [], nodelist
         else:
             raise ValueError("Interactivity was %r" % interactivity)
 
         exec_count = self.execution_count
 
         try:
             for i, node in enumerate(to_run_exec):
                 mod = ast.Module([node])
                 code = self.compile(mod, cell_name, "exec")
                 if self.run_code(code):
                     return True
 
             for i, node in enumerate(to_run_interactive):
                 mod = ast.Interactive([node])
                 code = self.compile(mod, cell_name, "single")
                 if self.run_code(code):
                     return True
 
             # Flush softspace
             if softspace(sys.stdout, 0):
                 print
 
         except:
             # It's possible to have exceptions raised here, typically by
             # compilation of odd code (such as a naked 'return' outside a
             # function) that did parse but isn't valid. Typically the exception
             # is a SyntaxError, but it's safest just to catch anything and show
             # the user a traceback.
 
             # We do only one try/except outside the loop to minimize the impact
             # on runtime, and also because if any node in the node list is
             # broken, we should stop execution completely.
             self.showtraceback()
 
         return False
 
     def run_code(self, code_obj):
         """Execute a code object.
 
         When an exception occurs, self.showtraceback() is called to display a
         traceback.
 
         Parameters
         ----------
         code_obj : code object
           A compiled code object, to be executed
 
         Returns
         -------
         False : successful execution.
         True : an error occurred.
         """
 
         # Set our own excepthook in case the user code tries to call it
         # directly, so that the IPython crash handler doesn't get triggered
         old_excepthook,sys.excepthook = sys.excepthook, self.excepthook
 
         # we save the original sys.excepthook in the instance, in case config
         # code (such as magics) needs access to it.
         self.sys_excepthook = old_excepthook
         outflag = 1  # happens in more places, so it's easier as default
         try:
             try:
                 self.hooks.pre_run_code_hook()
                 #rprint('Running code', repr(code_obj)) # dbg
                 exec code_obj in self.user_global_ns, self.user_ns
             finally:
                 # Reset our crash handler in place
                 sys.excepthook = old_excepthook
         except SystemExit:
             self.showtraceback(exception_only=True)
             warn("To exit: use 'exit', 'quit', or Ctrl-D.", level=1)
         except self.custom_exceptions:
             etype,value,tb = sys.exc_info()
             self.CustomTB(etype,value,tb)
         except:
             self.showtraceback()
         else:
             outflag = 0
         return outflag
 
     # For backwards compatibility
     runcode = run_code
 
     #-------------------------------------------------------------------------
     # Things related to GUI support and pylab
     #-------------------------------------------------------------------------
 
     def enable_gui(self, gui=None):
         raise NotImplementedError('Implement enable_gui in a subclass')
 
     def enable_pylab(self, gui=None, import_all=True):
         """Activate pylab support at runtime.
 
         This turns on support for matplotlib, preloads into the interactive
         namespace all of numpy and pylab, and configures IPython to correctly
         interact with the GUI event loop.  The GUI backend to be used can be
         optionally selected with the optional :param:`gui` argument.
 
         Parameters
         ----------
         gui : optional, string
 
           If given, dictates the choice of matplotlib GUI backend to use
           (should be one of IPython's supported backends, 'qt', 'osx', 'tk',
           'gtk', 'wx' or 'inline'), otherwise we use the default chosen by
           matplotlib (as dictated by the matplotlib build-time options plus the
           user's matplotlibrc configuration file).  Note that not all backends
           make sense in all contexts, for example a terminal ipython can't
           display figures inline.
         """
 
         # We want to prevent the loading of pylab to pollute the user's
         # namespace as shown by the %who* magics, so we execute the activation
         # code in an empty namespace, and we update *both* user_ns and
         # user_ns_hidden with this information.
         ns = {}
         try:
             gui = pylab_activate(ns, gui, import_all, self)
         except KeyError:
             error("Backend %r not supported" % gui)
             return
         self.user_ns.update(ns)
         self.user_ns_hidden.update(ns)
         # Now we must activate the gui pylab wants to use, and fix %run to take
         # plot updates into account
         self.enable_gui(gui)
         self.magic_run = self._pylab_magic_run
 
     #-------------------------------------------------------------------------
     # Utilities
     #-------------------------------------------------------------------------
 
     def var_expand(self, cmd, depth=0, formatter=DollarFormatter()):
         """Expand python variables in a string.
 
         The depth argument indicates how many frames above the caller should
         be walked to look for the local namespace where to expand variables.
 
         The global namespace for expansion is always the user's interactive
         namespace.
         """
         ns = self.user_ns.copy()
         ns.update(sys._getframe(depth+1).f_locals)
         ns.pop('self', None)
         try:
             cmd = formatter.format(cmd, **ns)
         except Exception:
             # if formatter couldn't format, just let it go untransformed
             pass
         return cmd
 
     def mktempfile(self, data=None, prefix='ipython_edit_'):
         """Make a new tempfile and return its filename.
 
         This makes a call to tempfile.mktemp, but it registers the created
         filename internally so ipython cleans it up at exit time.
 
         Optional inputs:
 
           - data(None): if data is given, it gets written out to the temp file
           immediately, and the file is closed again."""
 
         filename = tempfile.mktemp('.py', prefix)
         self.tempfiles.append(filename)
 
         if data:
             tmp_file = open(filename,'w')
             tmp_file.write(data)
             tmp_file.close()
         return filename
 
     # TODO:  This should be removed when Term is refactored.
     def write(self,data):
         """Write a string to the default output"""
         io.stdout.write(data)
 
     # TODO:  This should be removed when Term is refactored.
     def write_err(self,data):
         """Write a string to the default error output"""
         io.stderr.write(data)
 
     def ask_yes_no(self, prompt, default=None):
         if self.quiet:
             return True
         return ask_yes_no(prompt,default)
 
     def show_usage(self):
         """Show a usage message"""
         page.page(IPython.core.usage.interactive_usage)
 
     def find_user_code(self, target, raw=True, py_only=False):
         """Get a code string from history, file, url, or a string or macro.
 
         This is mainly used by magic functions.
 
         Parameters
         ----------
 
         target : str
 
           A string specifying code to retrieve. This will be tried respectively
           as: ranges of input history (see %history for syntax), url,
           correspnding .py file, filename, or an expression evaluating to a
           string or Macro in the user namespace.
 
         raw : bool
           If true (default), retrieve raw history. Has no effect on the other
           retrieval mechanisms.
 
         py_only : bool (default False)
           Only try to fetch python code, do not try alternative methods to decode file
           if unicode fails.
 
         Returns
         -------
         A string of code.
 
         ValueError is raised if nothing is found, and TypeError if it evaluates
         to an object of another type. In each case, .args[0] is a printable
         message.
         """
         code = self.extract_input_lines(target, raw=raw)  # Grab history
         if code:
             return code
         utarget = unquote_filename(target)
         try:
             if utarget.startswith(('http://', 'https://')):
                 return openpy.read_py_url(utarget, skip_encoding_cookie=True)
         except UnicodeDecodeError:
             if not py_only :
                 response = urllib.urlopen(target)
                 return response.read().decode('latin1')
             raise ValueError(("'%s' seem to be unreadable.") % utarget)
 
         potential_target = [target]
         try :
             potential_target.insert(0,get_py_filename(target))
         except IOError:
             pass
 
         for tgt in potential_target :
             if os.path.isfile(tgt):                        # Read file
                 try :
                     return openpy.read_py_file(tgt, skip_encoding_cookie=True)
                 except UnicodeDecodeError :
                     if not py_only :
                         with io_open(tgt,'r', encoding='latin1') as f :
                             return f.read()
                     raise ValueError(("'%s' seem to be unreadable.") % target)
 
         try:                                              # User namespace
             codeobj = eval(target, self.user_ns)
         except Exception:
             raise ValueError(("'%s' was not found in history, as a file, url, "
                                 "nor in the user namespace.") % target)
         if isinstance(codeobj, basestring):
             return codeobj
         elif isinstance(codeobj, Macro):
             return codeobj.value
 
         raise TypeError("%s is neither a string nor a macro." % target,
                         codeobj)
 
     #-------------------------------------------------------------------------
     # Things related to IPython exiting
     #-------------------------------------------------------------------------
     def atexit_operations(self):
         """This will be executed at the time of exit.
 
         Cleanup operations and saving of persistent data that is done
         unconditionally by IPython should be performed here.
 
         For things that may depend on startup flags or platform specifics (such
         as having readline or not), register a separate atexit function in the
         code that has the appropriate information, rather than trying to
         clutter
         """
         # Close the history session (this stores the end time and line count)
         # this must be *before* the tempfile cleanup, in case of temporary
         # history db
         self.history_manager.end_session()
 
         # Cleanup all tempfiles left around
         for tfile in self.tempfiles:
             try:
                 os.unlink(tfile)
             except OSError:
                 pass
 
         # Clear all user namespaces to release all references cleanly.
         self.reset(new_session=False)
 
         # Run user hooks
         self.hooks.shutdown_hook()
 
     def cleanup(self):
         self.restore_sys_module_state()
 
 
 class InteractiveShellABC(object):
     """An abstract base class for InteractiveShell."""
     __metaclass__ = abc.ABCMeta
 
 InteractiveShellABC.register(InteractiveShell)

IPython/core/tests/test_interactiveshell.py

@@ -1,304 +1,359 b''
 # -*- coding: utf-8 -*-
 """Tests for the key interactiveshell module.
 
 Historically the main classes in interactiveshell have been under-tested.  This
 module should grow as many single-method tests as possible to trap many of the
 recurring bugs we seem to encounter with high-level interaction.
 
 Authors
 -------
 * Fernando Perez
 """
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 # stdlib
 import os
 import shutil
 import tempfile
 import unittest
 from os.path import join
 import sys
 from StringIO import StringIO
 
 from IPython.testing.decorators import skipif
 from IPython.utils import io
 
 #-----------------------------------------------------------------------------
 # Tests
 #-----------------------------------------------------------------------------
 
 class InteractiveShellTestCase(unittest.TestCase):
     def test_naked_string_cells(self):
         """Test that cells with only naked strings are fully executed"""
         ip = get_ipython()
         # First, single-line inputs
         ip.run_cell('"a"\n')
         self.assertEquals(ip.user_ns['_'], 'a')
         # And also multi-line cells
         ip.run_cell('"""a\nb"""\n')
         self.assertEquals(ip.user_ns['_'], 'a\nb')
 
     def test_run_empty_cell(self):
         """Just make sure we don't get a horrible error with a blank
         cell of input. Yes, I did overlook that."""
         ip = get_ipython()
         old_xc = ip.execution_count
         ip.run_cell('')
         self.assertEquals(ip.execution_count, old_xc)
 
     def test_run_cell_multiline(self):
         """Multi-block, multi-line cells must execute correctly.
         """
         ip = get_ipython()
         src = '\n'.join(["x=1",
                          "y=2",
                          "if 1:",
                          "    x += 1",
                          "    y += 1",])
         ip.run_cell(src)
         self.assertEquals(ip.user_ns['x'], 2)
         self.assertEquals(ip.user_ns['y'], 3)
 
     def test_multiline_string_cells(self):
         "Code sprinkled with multiline strings should execute (GH-306)"
         ip = get_ipython()
         ip.run_cell('tmp=0')
         self.assertEquals(ip.user_ns['tmp'], 0)
         ip.run_cell('tmp=1;"""a\nb"""\n')
         self.assertEquals(ip.user_ns['tmp'], 1)
 
     def test_dont_cache_with_semicolon(self):
         "Ending a line with semicolon should not cache the returned object (GH-307)"
         ip = get_ipython()
         oldlen = len(ip.user_ns['Out'])
         a = ip.run_cell('1;', store_history=True)
         newlen = len(ip.user_ns['Out'])
         self.assertEquals(oldlen, newlen)
         #also test the default caching behavior
         ip.run_cell('1', store_history=True)
         newlen = len(ip.user_ns['Out'])
         self.assertEquals(oldlen+1, newlen)
 
     def test_In_variable(self):
         "Verify that In variable grows with user input (GH-284)"
         ip = get_ipython()
         oldlen = len(ip.user_ns['In'])
         ip.run_cell('1;', store_history=True)
         newlen = len(ip.user_ns['In'])
         self.assertEquals(oldlen+1, newlen)
         self.assertEquals(ip.user_ns['In'][-1],'1;')
         
     def test_magic_names_in_string(self):
         ip = get_ipython()
         ip.run_cell('a = """\n%exit\n"""')
         self.assertEquals(ip.user_ns['a'], '\n%exit\n')
     
     def test_alias_crash(self):
         """Errors in prefilter can't crash IPython"""
         ip = get_ipython()
         ip.run_cell('%alias parts echo first %s second %s')
         # capture stderr:
         save_err = io.stderr
         io.stderr = StringIO()
         ip.run_cell('parts 1')
         err = io.stderr.getvalue()
         io.stderr = save_err
         self.assertEquals(err.split(':')[0], 'ERROR')
     
     def test_trailing_newline(self):
         """test that running !(command) does not raise a SyntaxError"""
         ip = get_ipython()
         ip.run_cell('!(true)\n', False)
         ip.run_cell('!(true)\n\n\n', False)
     
     def test_gh_597(self):
         """Pretty-printing lists of objects with non-ascii reprs may cause
         problems."""
         class Spam(object):
           def __repr__(self):
             return "\xe9"*50
         import IPython.core.formatters
         f = IPython.core.formatters.PlainTextFormatter()
         f([Spam(),Spam()])
     
 
     def test_future_flags(self):
         """Check that future flags are used for parsing code (gh-777)"""
         ip = get_ipython()
         ip.run_cell('from __future__ import print_function')
         try:
             ip.run_cell('prfunc_return_val = print(1,2, sep=" ")')
             assert 'prfunc_return_val' in ip.user_ns
         finally:
             # Reset compiler flags so we don't mess up other tests.
             ip.compile.reset_compiler_flags()
 
     def test_future_unicode(self):
         """Check that unicode_literals is imported from __future__ (gh #786)"""
         ip = get_ipython()
         try:
             ip.run_cell(u'byte_str = "a"')
             assert isinstance(ip.user_ns['byte_str'], str) # string literals are byte strings by default
             ip.run_cell('from __future__ import unicode_literals')
             ip.run_cell(u'unicode_str = "a"')
             assert isinstance(ip.user_ns['unicode_str'], unicode) # strings literals are now unicode
         finally:
             # Reset compiler flags so we don't mess up other tests.
             ip.compile.reset_compiler_flags()
     
     def test_can_pickle(self):
         "Can we pickle objects defined interactively (GH-29)"
         ip = get_ipython()
         ip.reset()
         ip.run_cell(("class Mylist(list):\n"
                      "    def __init__(self,x=[]):\n"
                      "        list.__init__(self,x)"))
         ip.run_cell("w=Mylist([1,2,3])")
         
         from cPickle import dumps
         
         # We need to swap in our main module - this is only necessary
         # inside the test framework, because IPython puts the interactive module
         # in place (but the test framework undoes this).
         _main = sys.modules['__main__']
         sys.modules['__main__'] = ip.user_module
         try:
             res = dumps(ip.user_ns["w"])
         finally:
             sys.modules['__main__'] = _main
         self.assertTrue(isinstance(res, bytes))
         
     def test_global_ns(self):
         "Code in functions must be able to access variables outside them."
         ip = get_ipython()
         ip.run_cell("a = 10")
         ip.run_cell(("def f(x):\n"
                      "    return x + a"))
         ip.run_cell("b = f(12)")
         self.assertEqual(ip.user_ns["b"], 22)
 
     def test_bad_custom_tb(self):
         """Check that InteractiveShell is protected from bad custom exception handlers"""
         ip = get_ipython()
         from IPython.utils import io
         save_stderr = io.stderr
         try:
             # capture stderr
             io.stderr = StringIO()
             ip.set_custom_exc((IOError,), lambda etype,value,tb: 1/0)
             self.assertEquals(ip.custom_exceptions, (IOError,))
             ip.run_cell(u'raise IOError("foo")')
             self.assertEquals(ip.custom_exceptions, ())
             self.assertTrue("Custom TB Handler failed" in io.stderr.getvalue())
         finally:
             io.stderr = save_stderr
 
     def test_bad_custom_tb_return(self):
         """Check that InteractiveShell is protected from bad return types in custom exception handlers"""
         ip = get_ipython()
         from IPython.utils import io
         save_stderr = io.stderr
         try:
             # capture stderr
             io.stderr = StringIO()
             ip.set_custom_exc((NameError,),lambda etype,value,tb, tb_offset=None: 1)
             self.assertEquals(ip.custom_exceptions, (NameError,))
             ip.run_cell(u'a=abracadabra')
             self.assertEquals(ip.custom_exceptions, ())
             self.assertTrue("Custom TB Handler failed" in io.stderr.getvalue())
         finally:
             io.stderr = save_stderr
 
     def test_drop_by_id(self):
         ip = get_ipython()
         myvars = {"a":object(), "b":object(), "c": object()}
         ip.push(myvars, interactive=False)
         for name in myvars:
             assert name in ip.user_ns, name
             assert name in ip.user_ns_hidden, name
         ip.user_ns['b'] = 12
         ip.drop_by_id(myvars)
         for name in ["a", "c"]:
             assert name not in ip.user_ns, name
             assert name not in ip.user_ns_hidden, name
         assert ip.user_ns['b'] == 12
         ip.reset()
 
     def test_var_expand(self):
         ip = get_ipython()
         ip.user_ns['f'] = u'Ca\xf1o'
         self.assertEqual(ip.var_expand(u'echo $f'), u'echo Ca\xf1o')
         self.assertEqual(ip.var_expand(u'echo {f}'), u'echo Ca\xf1o')
         self.assertEqual(ip.var_expand(u'echo {f[:-1]}'), u'echo Ca\xf1')
         self.assertEqual(ip.var_expand(u'echo {1*2}'), u'echo 2')
 
         ip.user_ns['f'] = b'Ca\xc3\xb1o'
         # This should not raise any exception:
         ip.var_expand(u'echo $f')
     
     def test_bad_var_expand(self):
         """var_expand on invalid formats shouldn't raise"""
         ip = get_ipython()
         
         # SyntaxError
         self.assertEqual(ip.var_expand(u"{'a':5}"), u"{'a':5}")
         # NameError
         self.assertEqual(ip.var_expand(u"{asdf}"), u"{asdf}")
         # ZeroDivisionError
         self.assertEqual(ip.var_expand(u"{1/0}"), u"{1/0}")
+    
+    def test_silent_nopostexec(self):
+        """run_cell(silent=True) doesn't invoke post-exec funcs"""
+        ip = get_ipython()
+        
+        d = dict(called=False)
+        def set_called():
+            d['called'] = True
+        
+        ip.register_post_execute(set_called)
+        ip.run_cell("1", silent=True)
+        self.assertFalse(d['called'])
+        # double-check that non-silent exec did what we expected
+        # silent to avoid
+        ip.run_cell("1")
+        self.assertTrue(d['called'])
+        # remove post-exec
+        ip._post_execute.pop(set_called)
+    
+    def test_silent_noadvance(self):
+        """run_cell(silent=True) doesn't advance execution_count"""
+        ip = get_ipython()
+        
+        ec = ip.execution_count
+        # silent should force store_history=False
+        ip.run_cell("1", store_history=True, silent=True)
+        
+        self.assertEquals(ec, ip.execution_count)
+        # double-check that non-silent exec did what we expected
+        # silent to avoid
+        ip.run_cell("1", store_history=True)
+        self.assertEquals(ec+1, ip.execution_count)
+    
+    def test_silent_nodisplayhook(self):
+        """run_cell(silent=True) doesn't trigger displayhook"""
+        ip = get_ipython()
+        
+        d = dict(called=False)
+        
+        trap = ip.display_trap
+        save_hook = trap.hook
+        
+        def failing_hook(*args, **kwargs):
+            d['called'] = True
+        
+        try:
+            trap.hook = failing_hook
+            ip.run_cell("1", silent=True)
+            self.assertFalse(d['called'])
+            # double-check that non-silent exec did what we expected
+            # silent to avoid
+            ip.run_cell("1")
+            self.assertTrue(d['called'])
+        finally:
+            trap.hook = save_hook
 
     @skipif(sys.version_info[0] >= 3, "softspace removed in py3")
     def test_print_softspace(self):
         """Verify that softspace is handled correctly when executing multiple
         statements.
 
         In [1]: print 1; print 2
         1
         2
 
         In [2]: print 1,; print 2
         1 2
         """
 
 class TestSafeExecfileNonAsciiPath(unittest.TestCase):
 
     def setUp(self):
         self.BASETESTDIR = tempfile.mkdtemp()
         self.TESTDIR = join(self.BASETESTDIR, u"åäö")
         os.mkdir(self.TESTDIR)
         with open(join(self.TESTDIR, u"åäötestscript.py"), "w") as sfile:
             sfile.write("pass\n")
         self.oldpath = os.getcwdu()
         os.chdir(self.TESTDIR)
         self.fname = u"åäötestscript.py"
 
 
     def tearDown(self):
         os.chdir(self.oldpath)
         shutil.rmtree(self.BASETESTDIR)
 
     def test_1(self):
         """Test safe_execfile with non-ascii path
         """
         _ip.shell.safe_execfile(self.fname, {}, raise_exceptions=True)
 
 
 class TestSystemRaw(unittest.TestCase):
     def test_1(self):
         """Test system_raw with non-ascii cmd
         """
         cmd = ur'''python -c "'åäö'"   '''
         _ip.shell.system_raw(cmd)
 
 
 def test__IPYTHON__():
     # This shouldn't raise a NameError, that's all
     __IPYTHON__

IPython/extensions/parallelmagic.py

@@ -1,325 +1,325 b''
 # encoding: utf-8
 """
 =============
 parallelmagic
 =============
 
 Magic command interface for interactive parallel work.
 
 Usage
 =====
 
 ``%autopx``
 
 @AUTOPX_DOC@
 
 ``%px``
 
 @PX_DOC@
 
 ``%result``
 
 @RESULT_DOC@
 
 """
 
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2008-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 import ast
 import re
 
 from IPython.core.plugin import Plugin
 from IPython.utils.traitlets import Bool, Any, Instance
 from IPython.testing.skipdoctest import skip_doctest
 
 #-----------------------------------------------------------------------------
 # Definitions of magic functions for use with IPython
 #-----------------------------------------------------------------------------
 
 
 NO_ACTIVE_VIEW = """
 Use activate() on a DirectView object to activate it for magics.
 """
 
 
 class ParalleMagic(Plugin):
     """A component to manage the %result, %px and %autopx magics."""
 
     active_view = Instance('IPython.parallel.client.view.DirectView')
     verbose = Bool(False, config=True)
     shell = Instance('IPython.core.interactiveshell.InteractiveShellABC')
 
     def __init__(self, shell=None, config=None):
         super(ParalleMagic, self).__init__(shell=shell, config=config)
         self._define_magics()
         # A flag showing if autopx is activated or not
         self.autopx = False
 
     def _define_magics(self):
         """Define the magic functions."""
         self.shell.define_magic('result', self.magic_result)
         self.shell.define_magic('px', self.magic_px)
         self.shell.define_magic('autopx', self.magic_autopx)
 
     @skip_doctest
     def magic_result(self, ipself, parameter_s=''):
         """Print the result of command i on all engines..
 
         To use this a :class:`DirectView` instance must be created
         and then activated by calling its :meth:`activate` method.
 
         Then you can do the following::
 
             In [23]: %result
             Out[23]:
             <Results List>
             [0] In [6]: a = 10
             [1] In [6]: a = 10
 
             In [22]: %result 6
             Out[22]:
             <Results List>
             [0] In [6]: a = 10
             [1] In [6]: a = 10
         """
         if self.active_view is None:
             print NO_ACTIVE_VIEW
             return
 
         try:
             index = int(parameter_s)
         except:
             index = None
         result = self.active_view.get_result(index)
         return result
 
     @skip_doctest
     def magic_px(self, ipself, parameter_s=''):
         """Executes the given python command in parallel.
 
         To use this a :class:`DirectView` instance must be created
         and then activated by calling its :meth:`activate` method.
 
         Then you can do the following::
 
             In [24]: %px a = 5
             Parallel execution on engine(s): all
             Out[24]:
             <Results List>
             [0] In [7]: a = 5
             [1] In [7]: a = 5
         """
 
         if self.active_view is None:
             print NO_ACTIVE_VIEW
             return
         print "Parallel execution on engine(s): %s" % self.active_view.targets
         result = self.active_view.execute(parameter_s, block=False)
         if self.active_view.block:
             result.get()
             self._maybe_display_output(result)
 
     @skip_doctest
     def magic_autopx(self, ipself, parameter_s=''):
         """Toggles auto parallel mode.
 
         To use this a :class:`DirectView` instance must be created
         and then activated by calling its :meth:`activate` method. Once this
         is called, all commands typed at the command line are send to
         the engines to be executed in parallel. To control which engine
         are used, set the ``targets`` attributed of the multiengine client
         before entering ``%autopx`` mode.
 
         Then you can do the following::
 
             In [25]: %autopx
             %autopx to enabled
 
             In [26]: a = 10
             Parallel execution on engine(s): [0,1,2,3]
             In [27]: print a
             Parallel execution on engine(s): [0,1,2,3]
             [stdout:0] 10
             [stdout:1] 10
             [stdout:2] 10
             [stdout:3] 10
 
 
             In [27]: %autopx
             %autopx disabled
         """
         if self.autopx:
             self._disable_autopx()
         else:
             self._enable_autopx()
 
     def _enable_autopx(self):
         """Enable %autopx mode by saving the original run_cell and installing
         pxrun_cell.
         """
         if self.active_view is None:
             print NO_ACTIVE_VIEW
             return
 
         # override run_cell and run_code
         self._original_run_cell = self.shell.run_cell
         self.shell.run_cell = self.pxrun_cell
         self._original_run_code = self.shell.run_code
         self.shell.run_code = self.pxrun_code
 
         self.autopx = True
         print "%autopx enabled"
 
     def _disable_autopx(self):
         """Disable %autopx by restoring the original InteractiveShell.run_cell.
         """
         if self.autopx:
             self.shell.run_cell = self._original_run_cell
             self.shell.run_code = self._original_run_code
             self.autopx = False
             print "%autopx disabled"
 
     def _maybe_display_output(self, result):
         """Maybe display the output of a parallel result.
 
         If self.active_view.block is True, wait for the result
         and display the result.  Otherwise, this is a noop.
         """
         if isinstance(result.stdout, basestring):
             # single result
             stdouts = [result.stdout.rstrip()]
         else:
             stdouts = [s.rstrip() for s in result.stdout]
 
         targets = self.active_view.targets
         if isinstance(targets, int):
             targets = [targets]
         elif targets == 'all':
             targets = self.active_view.client.ids
 
         if any(stdouts):
             for eid,stdout in zip(targets, stdouts):
                 print '[stdout:%i]'%eid, stdout
 
 
-    def pxrun_cell(self, raw_cell, store_history=True):
+    def pxrun_cell(self, raw_cell, store_history=False, silent=False):
         """drop-in replacement for InteractiveShell.run_cell.
 
         This executes code remotely, instead of in the local namespace.
 
         See InteractiveShell.run_cell for details.
         """
 
         if (not raw_cell) or raw_cell.isspace():
             return
 
         ipself = self.shell
 
         with ipself.builtin_trap:
             cell = ipself.prefilter_manager.prefilter_lines(raw_cell)
 
             # Store raw and processed history
             if store_history:
                 ipself.history_manager.store_inputs(ipself.execution_count,
                                                   cell, raw_cell)
 
             # ipself.logger.log(cell, raw_cell)
 
             cell_name = ipself.compile.cache(cell, ipself.execution_count)
 
             try:
                 code_ast = ast.parse(cell, filename=cell_name)
             except (OverflowError, SyntaxError, ValueError, TypeError, MemoryError):
                 # Case 1
                 ipself.showsyntaxerror()
                 ipself.execution_count += 1
                 return None
             except NameError:
                 # ignore name errors, because we don't know the remote keys
                 pass
 
         if store_history:
             # Write output to the database. Does nothing unless
             # history output logging is enabled.
             ipself.history_manager.store_output(ipself.execution_count)
             # Each cell is a *single* input, regardless of how many lines it has
             ipself.execution_count += 1
         if re.search(r'get_ipython\(\)\.magic\(u?["\']%?autopx', cell):
             self._disable_autopx()
             return False
         else:
             try:
-                result = self.active_view.execute(cell, block=False)
+                result = self.active_view.execute(cell, silent=False, block=False)
             except:
                 ipself.showtraceback()
                 return True
             else:
                 if self.active_view.block:
                     try:
                         result.get()
                     except:
                         self.shell.showtraceback()
                         return True
                     else:
                         self._maybe_display_output(result)
                 return False
 
     def pxrun_code(self, code_obj):
         """drop-in replacement for InteractiveShell.run_code.
 
         This executes code remotely, instead of in the local namespace.
 
         See InteractiveShell.run_code for details.
         """
         ipself = self.shell
         # check code object for the autopx magic
         if 'get_ipython' in code_obj.co_names and 'magic' in code_obj.co_names and \
             any( [ isinstance(c, basestring) and 'autopx' in c for c in code_obj.co_consts ]):
             self._disable_autopx()
             return False
         else:
             try:
                 result = self.active_view.execute(code_obj, block=False)
             except:
                 ipself.showtraceback()
                 return True
             else:
                 if self.active_view.block:
                     try:
                         result.get()
                     except:
                         self.shell.showtraceback()
                         return True
                     else:
                         self._maybe_display_output(result)
                 return False
 
 
 __doc__ = __doc__.replace('@AUTOPX_DOC@',
                           "        " + ParalleMagic.magic_autopx.__doc__)
 __doc__ = __doc__.replace('@PX_DOC@',
                           "        " + ParalleMagic.magic_px.__doc__)
 __doc__ = __doc__.replace('@RESULT_DOC@',
                           "        " + ParalleMagic.magic_result.__doc__)
 
 
 _loaded = False
 
 
 def load_ipython_extension(ip):
     """Load the extension in IPython."""
     global _loaded
     if not _loaded:
         plugin = ParalleMagic(shell=ip, config=ip.config)
         ip.plugin_manager.register_plugin('parallelmagic', plugin)
         _loaded = True
 

IPython/frontend/consoleapp.py

@@ -1,352 +1,348 b''
 """ A minimal application base mixin for all ZMQ based IPython frontends.
 
 This is not a complete console app, as subprocess will not be able to receive
 input, there is no real readline support, among other limitations. This is a
 refactoring of what used to be the IPython/frontend/qt/console/qtconsoleapp.py
 
 Authors:
 
 * Evan Patterson
 * Min RK
 * Erik Tollerud
 * Fernando Perez
 * Bussonnier Matthias
 * Thomas Kluyver
 * Paul Ivanov
 
 """
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 # stdlib imports
 import atexit
 import json
 import os
 import signal
 import sys
 import uuid
 
 
 # Local imports
 from IPython.config.application import boolean_flag
 from IPython.config.configurable import Configurable
 from IPython.core.profiledir import ProfileDir
 from IPython.lib.kernel import tunnel_to_kernel, find_connection_file, swallow_argv
 from IPython.zmq.blockingkernelmanager import BlockingKernelManager
 from IPython.utils.path import filefind
 from IPython.utils.py3compat import str_to_bytes
 from IPython.utils.traitlets import (
     Dict, List, Unicode, CUnicode, Int, CBool, Any
 )
 from IPython.zmq.ipkernel import (
     flags as ipkernel_flags,
     aliases as ipkernel_aliases,
     IPKernelApp
 )
 from IPython.zmq.session import Session, default_secure
 from IPython.zmq.zmqshell import ZMQInteractiveShell
 
 #-----------------------------------------------------------------------------
 # Network Constants
 #-----------------------------------------------------------------------------
 
 from IPython.utils.localinterfaces import LOCALHOST, LOCAL_IPS
 
 #-----------------------------------------------------------------------------
 # Globals
 #-----------------------------------------------------------------------------
 
 
 #-----------------------------------------------------------------------------
 # Aliases and Flags
 #-----------------------------------------------------------------------------
 
 flags = dict(ipkernel_flags)
 
 # the flags that are specific to the frontend
 # these must be scrubbed before being passed to the kernel,
 # or it will raise an error on unrecognized flags
 app_flags = {
     'existing' : ({'IPythonConsoleApp' : {'existing' : 'kernel*.json'}},
             "Connect to an existing kernel. If no argument specified, guess most recent"),
 }
 app_flags.update(boolean_flag(
     'confirm-exit', 'IPythonConsoleApp.confirm_exit',
     """Set to display confirmation dialog on exit. You can always use 'exit' or 'quit',
        to force a direct exit without any confirmation.
     """,
     """Don't prompt the user when exiting. This will terminate the kernel
        if it is owned by the frontend, and leave it alive if it is external.
     """
 ))
 flags.update(app_flags)
 
 aliases = dict(ipkernel_aliases)
 
 # also scrub aliases from the frontend
 app_aliases = dict(
     hb = 'IPythonConsoleApp.hb_port',
     shell = 'IPythonConsoleApp.shell_port',
     iopub = 'IPythonConsoleApp.iopub_port',
     stdin = 'IPythonConsoleApp.stdin_port',
     ip = 'IPythonConsoleApp.ip',
     existing = 'IPythonConsoleApp.existing',
     f = 'IPythonConsoleApp.connection_file',
 
 
     ssh = 'IPythonConsoleApp.sshserver',
 )
 aliases.update(app_aliases)
 
 #-----------------------------------------------------------------------------
 # Classes
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # IPythonConsole
 #-----------------------------------------------------------------------------
 
 
 class IPythonConsoleApp(Configurable):
     name = 'ipython-console-mixin'
     default_config_file_name='ipython_config.py'
 
     description = """
         The IPython Mixin Console.
         
         This class contains the common portions of console client (QtConsole,
         ZMQ-based terminal console, etc).  It is not a full console, in that
         launched terminal subprocesses will not be able to accept input.
         
         The Console using this mixing supports various extra features beyond
         the single-process Terminal IPython shell, such as connecting to
         existing kernel, via:
         
             ipython <appname> --existing
         
         as well as tunnel via SSH
         
     """
 
     classes = [IPKernelApp, ZMQInteractiveShell, ProfileDir, Session]
     flags = Dict(flags)
     aliases = Dict(aliases)
     kernel_manager_class = BlockingKernelManager
 
     kernel_argv = List(Unicode)
     # frontend flags&aliases to be stripped when building kernel_argv
     frontend_flags = Any(app_flags)
     frontend_aliases = Any(app_aliases)
 
-    pure = CBool(False, config=True,
-        help="Use a pure Python kernel instead of an IPython kernel.")
     # create requested profiles by default, if they don't exist:
     auto_create = CBool(True)
     # connection info:
     ip = Unicode(LOCALHOST, config=True,
         help="""Set the kernel\'s IP address [default localhost].
         If the IP address is something other than localhost, then
         Consoles on other machines will be able to connect
         to the Kernel, so be careful!"""
     )
     
     sshserver = Unicode('', config=True,
         help="""The SSH server to use to connect to the kernel.""")
     sshkey = Unicode('', config=True,
         help="""Path to the ssh key to use for logging in to the ssh server.""")
     
     hb_port = Int(0, config=True,
         help="set the heartbeat port [default: random]")
     shell_port = Int(0, config=True,
         help="set the shell (XREP) port [default: random]")
     iopub_port = Int(0, config=True,
         help="set the iopub (PUB) port [default: random]")
     stdin_port = Int(0, config=True,
         help="set the stdin (XREQ) port [default: random]")
     connection_file = Unicode('', config=True,
         help="""JSON file in which to store connection info [default: kernel-<pid>.json]
 
         This file will contain the IP, ports, and authentication key needed to connect
         clients to this kernel. By default, this file will be created in the security-dir
         of the current profile, but can be specified by absolute path.
         """)
     def _connection_file_default(self):
         return 'kernel-%i.json' % os.getpid()
 
     existing = CUnicode('', config=True,
         help="""Connect to an already running kernel""")
 
     confirm_exit = CBool(True, config=True,
         help="""
         Set to display confirmation dialog on exit. You can always use 'exit' or 'quit',
         to force a direct exit without any confirmation.""",
     )
 
 
     def build_kernel_argv(self, argv=None):
         """build argv to be passed to kernel subprocess"""
         if argv is None:
             argv = sys.argv[1:]
         self.kernel_argv = swallow_argv(argv, self.frontend_aliases, self.frontend_flags)
         # kernel should inherit default config file from frontend
         self.kernel_argv.append("--KernelApp.parent_appname='%s'"%self.name)
     
     def init_connection_file(self):
         """find the connection file, and load the info if found.
         
         The current working directory and the current profile's security
         directory will be searched for the file if it is not given by
         absolute path.
         
         When attempting to connect to an existing kernel and the `--existing`
         argument does not match an existing file, it will be interpreted as a
         fileglob, and the matching file in the current profile's security dir
         with the latest access time will be used.
         
         After this method is called, self.connection_file contains the *full path*
         to the connection file, never just its name.
         """
         if self.existing:
             try:
                 cf = find_connection_file(self.existing)
             except Exception:
                 self.log.critical("Could not find existing kernel connection file %s", self.existing)
                 self.exit(1)
             self.log.info("Connecting to existing kernel: %s" % cf)
             self.connection_file = cf
         else:
             # not existing, check if we are going to write the file
             # and ensure that self.connection_file is a full path, not just the shortname
             try:
                 cf = find_connection_file(self.connection_file)
             except Exception:
                 # file might not exist
                 if self.connection_file == os.path.basename(self.connection_file):
                     # just shortname, put it in security dir
                     cf = os.path.join(self.profile_dir.security_dir, self.connection_file)
                 else:
                     cf = self.connection_file
                 self.connection_file = cf
         
         # should load_connection_file only be used for existing?
         # as it is now, this allows reusing ports if an existing
         # file is requested
         try:
             self.load_connection_file()
         except Exception:
             self.log.error("Failed to load connection file: %r", self.connection_file, exc_info=True)
             self.exit(1)
     
     def load_connection_file(self):
         """load ip/port/hmac config from JSON connection file"""
         # this is identical to KernelApp.load_connection_file
         # perhaps it can be centralized somewhere?
         try:
             fname = filefind(self.connection_file, ['.', self.profile_dir.security_dir])
         except IOError:
             self.log.debug("Connection File not found: %s", self.connection_file)
             return
         self.log.debug(u"Loading connection file %s", fname)
         with open(fname) as f:
             s = f.read()
         cfg = json.loads(s)
         if self.ip == LOCALHOST and 'ip' in cfg:
             # not overridden by config or cl_args
             self.ip = cfg['ip']
         for channel in ('hb', 'shell', 'iopub', 'stdin'):
             name = channel + '_port'
             if getattr(self, name) == 0 and name in cfg:
                 # not overridden by config or cl_args
                 setattr(self, name, cfg[name])
         if 'key' in cfg:
             self.config.Session.key = str_to_bytes(cfg['key'])
     
     def init_ssh(self):
         """set up ssh tunnels, if needed."""
         if not self.sshserver and not self.sshkey:
             return
         
         if self.sshkey and not self.sshserver:
             # specifying just the key implies that we are connecting directly
             self.sshserver = self.ip
             self.ip = LOCALHOST
         
         # build connection dict for tunnels:
         info = dict(ip=self.ip,
                     shell_port=self.shell_port,
                     iopub_port=self.iopub_port,
                     stdin_port=self.stdin_port,
                     hb_port=self.hb_port
         )
         
         self.log.info("Forwarding connections to %s via %s"%(self.ip, self.sshserver))
         
         # tunnels return a new set of ports, which will be on localhost:
         self.ip = LOCALHOST
         try:
             newports = tunnel_to_kernel(info, self.sshserver, self.sshkey)
         except:
             # even catch KeyboardInterrupt
             self.log.error("Could not setup tunnels", exc_info=True)
             self.exit(1)
         
         self.shell_port, self.iopub_port, self.stdin_port, self.hb_port = newports
         
         cf = self.connection_file
         base,ext = os.path.splitext(cf)
         base = os.path.basename(base)
         self.connection_file = os.path.basename(base)+'-ssh'+ext
         self.log.critical("To connect another client via this tunnel, use:")
         self.log.critical("--existing %s" % self.connection_file)
     
     def _new_connection_file(self):
         cf = ''
         while not cf:
             # we don't need a 128b id to distinguish kernels, use more readable
             # 48b node segment (12 hex chars).  Users running more than 32k simultaneous
             # kernels can subclass.
             ident = str(uuid.uuid4()).split('-')[-1]
             cf = os.path.join(self.profile_dir.security_dir, 'kernel-%s.json' % ident)
             # only keep if it's actually new.  Protect against unlikely collision
             # in 48b random search space
             cf = cf if not os.path.exists(cf) else ''
         return cf
 
     def init_kernel_manager(self):
         # Don't let Qt or ZMQ swallow KeyboardInterupts.
         signal.signal(signal.SIGINT, signal.SIG_DFL)
 
         # Create a KernelManager and start a kernel.
         self.kernel_manager = self.kernel_manager_class(
                                 ip=self.ip,
                                 shell_port=self.shell_port,
                                 iopub_port=self.iopub_port,
                                 stdin_port=self.stdin_port,
                                 hb_port=self.hb_port,
                                 connection_file=self.connection_file,
                                 config=self.config,
         )
         # start the kernel
         if not self.existing:
-            kwargs = dict(ipython=not self.pure)
-            kwargs['extra_arguments'] = self.kernel_argv
-            self.kernel_manager.start_kernel(**kwargs)
+            self.kernel_manager.start_kernel(extra_arguments=self.kernel_argv)
         elif self.sshserver:
             # ssh, write new connection file
             self.kernel_manager.write_connection_file()
         atexit.register(self.kernel_manager.cleanup_connection_file)
         self.kernel_manager.start_channels()
 
 
     def initialize(self, argv=None):
         """
         Classes which mix this class in should call:
                IPythonConsoleApp.initialize(self,argv)
         """
         self.init_connection_file()
         default_secure(self.config)
         self.init_ssh()
         self.init_kernel_manager()
 

IPython/frontend/html/notebook/clustermanager.py

@@ -1,174 +1,172 b''
 """Manage IPython.parallel clusters in the notebook.
 
 Authors:
 
 * Brian Granger
 """
 
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2008-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 import os
 
 from tornado import web
 from zmq.eventloop import ioloop
 
 from IPython.config.configurable import LoggingConfigurable
 from IPython.config.loader import load_pyconfig_files
 from IPython.utils.traitlets import Dict, Instance, CFloat
 from IPython.parallel.apps.ipclusterapp import IPClusterStart
 from IPython.core.profileapp import list_profiles_in
 from IPython.core.profiledir import ProfileDir
 from IPython.utils.path import get_ipython_dir
 from IPython.utils.sysinfo import num_cpus
 
 
 #-----------------------------------------------------------------------------
 # Classes
 #-----------------------------------------------------------------------------
 
 
 class DummyIPClusterStart(IPClusterStart):
     """Dummy subclass to skip init steps that conflict with global app.
     
     Instantiating and initializing this class should result in fully configured
     launchers, but no other side effects or state.
     """
 
     def init_signal(self):
         pass
-    def init_logging(self):
-        pass
     def reinit_logging(self):
         pass
 
 
 class ClusterManager(LoggingConfigurable):
 
     profiles = Dict()
 
     delay = CFloat(1., config=True,
         help="delay (in s) between starting the controller and the engines")
 
     loop = Instance('zmq.eventloop.ioloop.IOLoop')
     def _loop_default(self):
         from zmq.eventloop.ioloop import IOLoop
         return IOLoop.instance()
 
     def build_launchers(self, profile_dir):
         starter = DummyIPClusterStart(log=self.log)
         starter.initialize(['--profile-dir', profile_dir])
         cl = starter.controller_launcher
         esl = starter.engine_launcher
         n = starter.n
         return cl, esl, n
 
     def get_profile_dir(self, name, path):
         p = ProfileDir.find_profile_dir_by_name(path,name=name)
         return p.location
 
     def update_profiles(self):
         """List all profiles in the ipython_dir and cwd.
         """
         for path in [get_ipython_dir(), os.getcwdu()]:
             for profile in list_profiles_in(path):
                 pd = self.get_profile_dir(profile, path)
                 if profile not in self.profiles:
                     self.log.debug("Overwriting profile %s" % profile)
                     self.profiles[profile] = {
                         'profile': profile,
                         'profile_dir': pd,
                         'status': 'stopped'
                     }
 
     def list_profiles(self):
         self.update_profiles()
         result = [self.profile_info(p) for p in sorted(self.profiles.keys())]
         return result
 
     def check_profile(self, profile):
         if profile not in self.profiles:
             raise web.HTTPError(404, u'profile not found')
 
     def profile_info(self, profile):
         self.check_profile(profile)
         result = {}
         data = self.profiles.get(profile)
         result['profile'] = profile
         result['profile_dir'] = data['profile_dir']
         result['status'] = data['status']
         if 'n' in data:
             result['n'] = data['n']
         return result
 
     def start_cluster(self, profile, n=None):
         """Start a cluster for a given profile."""
         self.check_profile(profile)
         data = self.profiles[profile]
         if data['status'] == 'running':
             raise web.HTTPError(409, u'cluster already running')
         cl, esl, default_n = self.build_launchers(data['profile_dir'])
         n = n if n is not None else default_n
         def clean_data():
             data.pop('controller_launcher',None)
             data.pop('engine_set_launcher',None)
             data.pop('n',None)
             data['status'] = 'stopped'
         def engines_stopped(r):
             self.log.debug('Engines stopped')
             if cl.running:
                 cl.stop()
             clean_data()
         esl.on_stop(engines_stopped)
         def controller_stopped(r):
             self.log.debug('Controller stopped')
             if esl.running:
                 esl.stop()
             clean_data()
         cl.on_stop(controller_stopped)
 
         dc = ioloop.DelayedCallback(lambda: cl.start(), 0, self.loop)
         dc.start()
         dc = ioloop.DelayedCallback(lambda: esl.start(n), 1000*self.delay, self.loop)
         dc.start()
 
         self.log.debug('Cluster started')
         data['controller_launcher'] = cl
         data['engine_set_launcher'] = esl
         data['n'] = n
         data['status'] = 'running'
         return self.profile_info(profile)
 
     def stop_cluster(self, profile):
         """Stop a cluster for a given profile."""
         self.check_profile(profile)
         data = self.profiles[profile]
         if data['status'] == 'stopped':
             raise web.HTTPError(409, u'cluster not running')
         data = self.profiles[profile]
         cl = data['controller_launcher']
         esl = data['engine_set_launcher']
         if cl.running:
             cl.stop()
         if esl.running:
             esl.stop()
         # Return a temp info dict, the real one is updated in the on_stop
         # logic above.
         result = {
             'profile': data['profile'],
             'profile_dir': data['profile_dir'],
             'status': 'stopped'
         }
         return result
 
     def stop_all_clusters(self):
         for p in self.profiles.keys():
             self.stop_cluster(p)

IPython/frontend/html/notebook/handlers.py

@@ -1,733 +1,734 b''
 """Tornado handlers for the notebook.
 
 Authors:
 
 * Brian Granger
 """
 
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2008-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 import logging
 import Cookie
 import time
 import uuid
 
 from tornado import web
 from tornado import websocket
 
 from zmq.eventloop import ioloop
 from zmq.utils import jsonapi
 
 from IPython.external.decorator import decorator
 from IPython.zmq.session import Session
 from IPython.lib.security import passwd_check
+from IPython.utils.jsonutil import date_default
 
 try:
     from docutils.core import publish_string
 except ImportError:
     publish_string = None
 
 #-----------------------------------------------------------------------------
 # Monkeypatch for Tornado <= 2.1.1 - Remove when no longer necessary!
 #-----------------------------------------------------------------------------
 
 # Google Chrome, as of release 16, changed its websocket protocol number.  The
 # parts tornado cares about haven't really changed, so it's OK to continue
 # accepting Chrome connections, but as of Tornado 2.1.1 (the currently released
 # version as of Oct 30/2011) the version check fails, see the issue report:
 
 # https://github.com/facebook/tornado/issues/385
 
 # This issue has been fixed in Tornado post 2.1.1:
 
 # https://github.com/facebook/tornado/commit/84d7b458f956727c3b0d6710
 
 # Here we manually apply the same patch as above so that users of IPython can
 # continue to work with an officially released Tornado.  We make the
 # monkeypatch version check as narrow as possible to limit its effects; once
 # Tornado 2.1.1 is no longer found in the wild we'll delete this code.
 
 import tornado
 
 if tornado.version_info <= (2,1,1):
 
     def _execute(self, transforms, *args, **kwargs):
         from tornado.websocket import WebSocketProtocol8, WebSocketProtocol76
         
         self.open_args = args
         self.open_kwargs = kwargs
 
         # The difference between version 8 and 13 is that in 8 the
         # client sends a "Sec-Websocket-Origin" header and in 13 it's
         # simply "Origin".
         if self.request.headers.get("Sec-WebSocket-Version") in ("7", "8", "13"):
             self.ws_connection = WebSocketProtocol8(self)
             self.ws_connection.accept_connection()
             
         elif self.request.headers.get("Sec-WebSocket-Version"):
             self.stream.write(tornado.escape.utf8(
                 "HTTP/1.1 426 Upgrade Required\r\n"
                 "Sec-WebSocket-Version: 8\r\n\r\n"))
             self.stream.close()
             
         else:
             self.ws_connection = WebSocketProtocol76(self)
             self.ws_connection.accept_connection()
 
     websocket.WebSocketHandler._execute = _execute
     del _execute
     
 #-----------------------------------------------------------------------------
 # Decorator for disabling read-only handlers
 #-----------------------------------------------------------------------------
 
 @decorator
 def not_if_readonly(f, self, *args, **kwargs):
     if self.application.read_only:
         raise web.HTTPError(403, "Notebook server is read-only")
     else:
         return f(self, *args, **kwargs)
 
 @decorator
 def authenticate_unless_readonly(f, self, *args, **kwargs):
     """authenticate this page *unless* readonly view is active.
     
     In read-only mode, the notebook list and print view should
     be accessible without authentication.
     """
     
     @web.authenticated
     def auth_f(self, *args, **kwargs):
         return f(self, *args, **kwargs)
 
     if self.application.read_only:
         return f(self, *args, **kwargs)
     else:
         return auth_f(self, *args, **kwargs)
 
 #-----------------------------------------------------------------------------
 # Top-level handlers
 #-----------------------------------------------------------------------------
 
 class RequestHandler(web.RequestHandler):
     """RequestHandler with default variable setting."""
 
     def render(*args, **kwargs):
         kwargs.setdefault('message', '')
         return web.RequestHandler.render(*args, **kwargs)
 
 class AuthenticatedHandler(RequestHandler):
     """A RequestHandler with an authenticated user."""
 
     def get_current_user(self):
         user_id = self.get_secure_cookie("username")
         # For now the user_id should not return empty, but it could eventually
         if user_id == '':
             user_id = 'anonymous'
         if user_id is None:
             # prevent extra Invalid cookie sig warnings:
             self.clear_cookie('username')
             if not self.application.password and not self.application.read_only:
                 user_id = 'anonymous'
         return user_id
 
     @property
     def logged_in(self):
         """Is a user currently logged in?
 
         """
         user = self.get_current_user()
         return (user and not user == 'anonymous')
 
     @property
     def login_available(self):
         """May a user proceed to log in?
 
         This returns True if login capability is available, irrespective of
         whether the user is already logged in or not.
 
         """
         return bool(self.application.password)
 
     @property
     def read_only(self):
         """Is the notebook read-only?
 
         """
         return self.application.read_only
 
     @property
     def ws_url(self):
         """websocket url matching the current request
 
         turns http[s]://host[:port] into
                 ws[s]://host[:port]
         """
         proto = self.request.protocol.replace('http', 'ws')
         host = self.application.ipython_app.websocket_host # default to config value
         if host == '':
             host = self.request.host # get from request
         return "%s://%s" % (proto, host)
 
 
 class AuthenticatedFileHandler(AuthenticatedHandler, web.StaticFileHandler):
     """static files should only be accessible when logged in"""
 
     @authenticate_unless_readonly
     def get(self, path):
         return web.StaticFileHandler.get(self, path)
 
 
 class ProjectDashboardHandler(AuthenticatedHandler):
 
     @authenticate_unless_readonly
     def get(self):
         nbm = self.application.notebook_manager
         project = nbm.notebook_dir
         self.render(
             'projectdashboard.html', project=project,
             base_project_url=self.application.ipython_app.base_project_url,
             base_kernel_url=self.application.ipython_app.base_kernel_url,
             read_only=self.read_only,
             logged_in=self.logged_in,
             login_available=self.login_available
         )
 
 
 class LoginHandler(AuthenticatedHandler):
 
     def _render(self, message=None):
         self.render('login.html',
                 next=self.get_argument('next', default='/'),
                 read_only=self.read_only,
                 logged_in=self.logged_in,
                 login_available=self.login_available,
                 base_project_url=self.application.ipython_app.base_project_url,
                 message=message
         )
 
     def get(self):
         if self.current_user:
             self.redirect(self.get_argument('next', default='/'))
         else:
             self._render()
 
     def post(self):
         pwd = self.get_argument('password', default=u'')
         if self.application.password:
             if passwd_check(self.application.password, pwd):
                 self.set_secure_cookie('username', str(uuid.uuid4()))
             else:
                 self._render(message={'error': 'Invalid password'})
                 return
 
         self.redirect(self.get_argument('next', default='/'))
 
 
 class LogoutHandler(AuthenticatedHandler):
 
     def get(self):
         self.clear_cookie('username')
         if self.login_available:
             message = {'info': 'Successfully logged out.'}
         else:
             message = {'warning': 'Cannot log out.  Notebook authentication '
                        'is disabled.'}
 
         self.render('logout.html',
                     read_only=self.read_only,
                     logged_in=self.logged_in,
                     login_available=self.login_available,
                     base_project_url=self.application.ipython_app.base_project_url,
                     message=message)
 
 
 class NewHandler(AuthenticatedHandler):
 
     @web.authenticated
     def get(self):
         nbm = self.application.notebook_manager
         project = nbm.notebook_dir
         notebook_id = nbm.new_notebook()
         self.render(
             'notebook.html', project=project,
             notebook_id=notebook_id,
             base_project_url=self.application.ipython_app.base_project_url,
             base_kernel_url=self.application.ipython_app.base_kernel_url,
             kill_kernel=False,
             read_only=False,
             logged_in=self.logged_in,
             login_available=self.login_available,
             mathjax_url=self.application.ipython_app.mathjax_url,
         )
 
 
 class NamedNotebookHandler(AuthenticatedHandler):
 
     @authenticate_unless_readonly
     def get(self, notebook_id):
         nbm = self.application.notebook_manager
         project = nbm.notebook_dir
         if not nbm.notebook_exists(notebook_id):
             raise web.HTTPError(404, u'Notebook does not exist: %s' % notebook_id)
         
         self.render(
             'notebook.html', project=project,
             notebook_id=notebook_id,
             base_project_url=self.application.ipython_app.base_project_url,
             base_kernel_url=self.application.ipython_app.base_kernel_url,
             kill_kernel=False,
             read_only=self.read_only,
             logged_in=self.logged_in,
             login_available=self.login_available,
             mathjax_url=self.application.ipython_app.mathjax_url,
         )
 
 
 class PrintNotebookHandler(AuthenticatedHandler):
 
     @authenticate_unless_readonly
     def get(self, notebook_id):
         nbm = self.application.notebook_manager
         project = nbm.notebook_dir
         if not nbm.notebook_exists(notebook_id):
             raise web.HTTPError(404, u'Notebook does not exist: %s' % notebook_id)
         
         self.render(
             'printnotebook.html', project=project,
             notebook_id=notebook_id,
             base_project_url=self.application.ipython_app.base_project_url,
             base_kernel_url=self.application.ipython_app.base_kernel_url,
             kill_kernel=False,
             read_only=self.read_only,
             logged_in=self.logged_in,
             login_available=self.login_available,
             mathjax_url=self.application.ipython_app.mathjax_url,
         )
 
 #-----------------------------------------------------------------------------
 # Kernel handlers
 #-----------------------------------------------------------------------------
 
 
 class MainKernelHandler(AuthenticatedHandler):
 
     @web.authenticated
     def get(self):
         km = self.application.kernel_manager
         self.finish(jsonapi.dumps(km.kernel_ids))
 
     @web.authenticated
     def post(self):
         km = self.application.kernel_manager
         notebook_id = self.get_argument('notebook', default=None)
         kernel_id = km.start_kernel(notebook_id)
         data = {'ws_url':self.ws_url,'kernel_id':kernel_id}
         self.set_header('Location', '/'+kernel_id)
         self.finish(jsonapi.dumps(data))
 
 
 class KernelHandler(AuthenticatedHandler):
 
     SUPPORTED_METHODS = ('DELETE')
 
     @web.authenticated
     def delete(self, kernel_id):
         km = self.application.kernel_manager
         km.kill_kernel(kernel_id)
         self.set_status(204)
         self.finish()
 
 
 class KernelActionHandler(AuthenticatedHandler):
 
     @web.authenticated
     def post(self, kernel_id, action):
         km = self.application.kernel_manager
         if action == 'interrupt':
             km.interrupt_kernel(kernel_id)
             self.set_status(204)
         if action == 'restart':
             new_kernel_id = km.restart_kernel(kernel_id)
             data = {'ws_url':self.ws_url,'kernel_id':new_kernel_id}
             self.set_header('Location', '/'+new_kernel_id)
             self.write(jsonapi.dumps(data))
         self.finish()
 
 
 class ZMQStreamHandler(websocket.WebSocketHandler):
 
     def _reserialize_reply(self, msg_list):
         """Reserialize a reply message using JSON.
 
         This takes the msg list from the ZMQ socket, unserializes it using
         self.session and then serializes the result using JSON. This method
         should be used by self._on_zmq_reply to build messages that can
         be sent back to the browser.
         """
         idents, msg_list = self.session.feed_identities(msg_list)
         msg = self.session.unserialize(msg_list)
         try:
             msg['header'].pop('date')
         except KeyError:
             pass
         try:
             msg['parent_header'].pop('date')
         except KeyError:
             pass
         msg.pop('buffers')
-        return jsonapi.dumps(msg)
+        return jsonapi.dumps(msg, default=date_default)
 
     def _on_zmq_reply(self, msg_list):
         try:
             msg = self._reserialize_reply(msg_list)
-        except:
-            self.application.log.critical("Malformed message: %r" % msg_list)
+        except Exception:
+            self.application.log.critical("Malformed message: %r" % msg_list, exc_info=True)
         else:
             self.write_message(msg)
 
     def allow_draft76(self):
         """Allow draft 76, until browsers such as Safari update to RFC 6455.
         
         This has been disabled by default in tornado in release 2.2.0, and
         support will be removed in later versions.
         """
         return True
 
 
 class AuthenticatedZMQStreamHandler(ZMQStreamHandler):
 
     def open(self, kernel_id):
         self.kernel_id = kernel_id.decode('ascii')
         try:
             cfg = self.application.ipython_app.config
         except AttributeError:
             # protect from the case where this is run from something other than
             # the notebook app:
             cfg = None
         self.session = Session(config=cfg)
         self.save_on_message = self.on_message
         self.on_message = self.on_first_message
 
     def get_current_user(self):
         user_id = self.get_secure_cookie("username")
         if user_id == '' or (user_id is None and not self.application.password):
             user_id = 'anonymous'
         return user_id
 
     def _inject_cookie_message(self, msg):
         """Inject the first message, which is the document cookie,
         for authentication."""
         if isinstance(msg, unicode):
             # Cookie can't constructor doesn't accept unicode strings for some reason
             msg = msg.encode('utf8', 'replace')
         try:
             self.request._cookies = Cookie.SimpleCookie(msg)
         except:
             logging.warn("couldn't parse cookie string: %s",msg, exc_info=True)
 
     def on_first_message(self, msg):
         self._inject_cookie_message(msg)
         if self.get_current_user() is None:
             logging.warn("Couldn't authenticate WebSocket connection")
             raise web.HTTPError(403)
         self.on_message = self.save_on_message
 
 
 class IOPubHandler(AuthenticatedZMQStreamHandler):
 
     def initialize(self, *args, **kwargs):
         self._kernel_alive = True
         self._beating = False
         self.iopub_stream = None
         self.hb_stream = None
 
     def on_first_message(self, msg):
         try:
             super(IOPubHandler, self).on_first_message(msg)
         except web.HTTPError:
             self.close()
             return
         km = self.application.kernel_manager
         self.time_to_dead = km.time_to_dead
         self.first_beat = km.first_beat
         kernel_id = self.kernel_id
         try:
             self.iopub_stream = km.create_iopub_stream(kernel_id)
             self.hb_stream = km.create_hb_stream(kernel_id)
         except web.HTTPError:
             # WebSockets don't response to traditional error codes so we
             # close the connection.
             if not self.stream.closed():
                 self.stream.close()
             self.close()
         else:
             self.iopub_stream.on_recv(self._on_zmq_reply)
             self.start_hb(self.kernel_died)
 
     def on_message(self, msg):
         pass
 
     def on_close(self):
         # This method can be called twice, once by self.kernel_died and once
         # from the WebSocket close event. If the WebSocket connection is
         # closed before the ZMQ streams are setup, they could be None.
         self.stop_hb()
         if self.iopub_stream is not None and not self.iopub_stream.closed():
             self.iopub_stream.on_recv(None)
             self.iopub_stream.close()
         if self.hb_stream is not None and not self.hb_stream.closed():
             self.hb_stream.close()
 
     def start_hb(self, callback):
         """Start the heartbeating and call the callback if the kernel dies."""
         if not self._beating:
             self._kernel_alive = True
 
             def ping_or_dead():
                 self.hb_stream.flush()
                 if self._kernel_alive:
                     self._kernel_alive = False
                     self.hb_stream.send(b'ping')
                     # flush stream to force immediate socket send
                     self.hb_stream.flush()
                 else:
                     try:
                         callback()
                     except:
                         pass
                     finally:
                         self.stop_hb()
 
             def beat_received(msg):
                 self._kernel_alive = True
 
             self.hb_stream.on_recv(beat_received)
             loop = ioloop.IOLoop.instance()
             self._hb_periodic_callback = ioloop.PeriodicCallback(ping_or_dead, self.time_to_dead*1000, loop)
             loop.add_timeout(time.time()+self.first_beat, self._really_start_hb)
             self._beating= True
     
     def _really_start_hb(self):
         """callback for delayed heartbeat start
         
         Only start the hb loop if we haven't been closed during the wait.
         """
         if self._beating and not self.hb_stream.closed():
             self._hb_periodic_callback.start()
 
     def stop_hb(self):
         """Stop the heartbeating and cancel all related callbacks."""
         if self._beating:
             self._beating = False
             self._hb_periodic_callback.stop()
             if not self.hb_stream.closed():
                 self.hb_stream.on_recv(None)
 
     def kernel_died(self):
         self.application.kernel_manager.delete_mapping_for_kernel(self.kernel_id)
         self.application.log.error("Kernel %s failed to respond to heartbeat", self.kernel_id)
         self.write_message(
             {'header': {'msg_type': 'status'},
              'parent_header': {},
              'content': {'execution_state':'dead'}
             }
         )
         self.on_close()
 
 
 class ShellHandler(AuthenticatedZMQStreamHandler):
 
     def initialize(self, *args, **kwargs):
         self.shell_stream = None
 
     def on_first_message(self, msg):
         try:
             super(ShellHandler, self).on_first_message(msg)
         except web.HTTPError:
             self.close()
             return
         km = self.application.kernel_manager
         self.max_msg_size = km.max_msg_size
         kernel_id = self.kernel_id
         try:
             self.shell_stream = km.create_shell_stream(kernel_id)
         except web.HTTPError:
             # WebSockets don't response to traditional error codes so we
             # close the connection.
             if not self.stream.closed():
                 self.stream.close()
             self.close()
         else:
             self.shell_stream.on_recv(self._on_zmq_reply)
 
     def on_message(self, msg):
         if len(msg) < self.max_msg_size:
             msg = jsonapi.loads(msg)
             self.session.send(self.shell_stream, msg)
 
     def on_close(self):
         # Make sure the stream exists and is not already closed.
         if self.shell_stream is not None and not self.shell_stream.closed():
             self.shell_stream.close()
 
 
 #-----------------------------------------------------------------------------
 # Notebook web service handlers
 #-----------------------------------------------------------------------------
 
 class NotebookRootHandler(AuthenticatedHandler):
 
     @authenticate_unless_readonly
     def get(self):
         nbm = self.application.notebook_manager
         files = nbm.list_notebooks()
         self.finish(jsonapi.dumps(files))
 
     @web.authenticated
     def post(self):
         nbm = self.application.notebook_manager
         body = self.request.body.strip()
         format = self.get_argument('format', default='json')
         name = self.get_argument('name', default=None)
         if body:
             notebook_id = nbm.save_new_notebook(body, name=name, format=format)
         else:
             notebook_id = nbm.new_notebook()
         self.set_header('Location', '/'+notebook_id)
         self.finish(jsonapi.dumps(notebook_id))
 
 
 class NotebookHandler(AuthenticatedHandler):
 
     SUPPORTED_METHODS = ('GET', 'PUT', 'DELETE')
 
     @authenticate_unless_readonly
     def get(self, notebook_id):
         nbm = self.application.notebook_manager
         format = self.get_argument('format', default='json')
         last_mod, name, data = nbm.get_notebook(notebook_id, format)
         
         if format == u'json':
             self.set_header('Content-Type', 'application/json')
             self.set_header('Content-Disposition','attachment; filename="%s.ipynb"' % name)
         elif format == u'py':
             self.set_header('Content-Type', 'application/x-python')
             self.set_header('Content-Disposition','attachment; filename="%s.py"' % name)
         self.set_header('Last-Modified', last_mod)
         self.finish(data)
 
     @web.authenticated
     def put(self, notebook_id):
         nbm = self.application.notebook_manager
         format = self.get_argument('format', default='json')
         name = self.get_argument('name', default=None)
         nbm.save_notebook(notebook_id, self.request.body, name=name, format=format)
         self.set_status(204)
         self.finish()
 
     @web.authenticated
     def delete(self, notebook_id):
         nbm = self.application.notebook_manager
         nbm.delete_notebook(notebook_id)
         self.set_status(204)
         self.finish()
 
 
 class NotebookCopyHandler(AuthenticatedHandler):
 
     @web.authenticated
     def get(self, notebook_id):
         nbm = self.application.notebook_manager
         project = nbm.notebook_dir
         notebook_id = nbm.copy_notebook(notebook_id)
         self.render(
             'notebook.html', project=project,
             notebook_id=notebook_id,
             base_project_url=self.application.ipython_app.base_project_url,
             base_kernel_url=self.application.ipython_app.base_kernel_url,
             kill_kernel=False,
             read_only=False,
             logged_in=self.logged_in,
             login_available=self.login_available,
             mathjax_url=self.application.ipython_app.mathjax_url,
         )
 
 
 #-----------------------------------------------------------------------------
 # Cluster handlers
 #-----------------------------------------------------------------------------
 
 
 class MainClusterHandler(AuthenticatedHandler):
 
     @web.authenticated
     def get(self):
         cm = self.application.cluster_manager
         self.finish(jsonapi.dumps(cm.list_profiles()))
 
 
 class ClusterProfileHandler(AuthenticatedHandler):
 
     @web.authenticated
     def get(self, profile):
         cm = self.application.cluster_manager
         self.finish(jsonapi.dumps(cm.profile_info(profile)))
 
 
 class ClusterActionHandler(AuthenticatedHandler):
 
     @web.authenticated
     def post(self, profile, action):
         cm = self.application.cluster_manager
         if action == 'start':
             n = self.get_argument('n',default=None)
             if n is None:
                 data = cm.start_cluster(profile)
             else:
                 data = cm.start_cluster(profile,int(n))
         if action == 'stop':
             data = cm.stop_cluster(profile)
         self.finish(jsonapi.dumps(data))
 
 
 #-----------------------------------------------------------------------------
 # RST web service handlers
 #-----------------------------------------------------------------------------
 
 
 class RSTHandler(AuthenticatedHandler):
 
     @web.authenticated
     def post(self):
         if publish_string is None:
             raise web.HTTPError(503, u'docutils not available')
         body = self.request.body.strip()
         source = body
         # template_path=os.path.join(os.path.dirname(__file__), u'templates', u'rst_template.html')
         defaults = {'file_insertion_enabled': 0,
                     'raw_enabled': 0,
                     '_disable_config': 1,
                     'stylesheet_path': 0
                     # 'template': template_path
         }
         try:
             html = publish_string(source, writer_name='html',
                                   settings_overrides=defaults
             )
         except:
             raise web.HTTPError(400, u'Invalid RST')
         print html
         self.set_header('Content-Type', 'text/html')
         self.finish(html)
 
 

IPython/frontend/html/notebook/notebookapp.py

@@ -1,565 +1,565 b''
 # coding: utf-8
 """A tornado based IPython notebook server.
 
 Authors:
 
 * Brian Granger
 """
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2008-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 # stdlib
 import errno
 import logging
 import os
 import re
 import select
 import signal
 import socket
 import sys
 import threading
 import time
 import webbrowser
 
 # Third party
 import zmq
 
 # Install the pyzmq ioloop. This has to be done before anything else from
 # tornado is imported.
 from zmq.eventloop import ioloop
 ioloop.install()
 
 from tornado import httpserver
 from tornado import web
 
 # Our own libraries
 from .kernelmanager import MappingKernelManager
 from .handlers import (LoginHandler, LogoutHandler,
     ProjectDashboardHandler, NewHandler, NamedNotebookHandler,
     MainKernelHandler, KernelHandler, KernelActionHandler, IOPubHandler,
     ShellHandler, NotebookRootHandler, NotebookHandler, NotebookCopyHandler,
     RSTHandler, AuthenticatedFileHandler, PrintNotebookHandler,
     MainClusterHandler, ClusterProfileHandler, ClusterActionHandler
 )
 from .notebookmanager import NotebookManager
 from .clustermanager import ClusterManager
 
 from IPython.config.application import catch_config_error, boolean_flag
 from IPython.core.application import BaseIPythonApplication
 from IPython.core.profiledir import ProfileDir
 from IPython.lib.kernel import swallow_argv
 from IPython.zmq.session import Session, default_secure
 from IPython.zmq.zmqshell import ZMQInteractiveShell
 from IPython.zmq.ipkernel import (
     flags as ipkernel_flags,
     aliases as ipkernel_aliases,
     IPKernelApp
 )
 from IPython.utils.traitlets import Dict, Unicode, Integer, List, Enum, Bool
 from IPython.utils import py3compat
 
 #-----------------------------------------------------------------------------
 # Module globals
 #-----------------------------------------------------------------------------
 
 _kernel_id_regex = r"(?P<kernel_id>\w+-\w+-\w+-\w+-\w+)"
 _kernel_action_regex = r"(?P<action>restart|interrupt)"
 _notebook_id_regex = r"(?P<notebook_id>\w+-\w+-\w+-\w+-\w+)"
 _profile_regex = r"(?P<profile>[a-zA-Z0-9]+)"
 _cluster_action_regex = r"(?P<action>start|stop)"
 
 
 LOCALHOST = '127.0.0.1'
 
 _examples = """
 ipython notebook                       # start the notebook
 ipython notebook --profile=sympy       # use the sympy profile
 ipython notebook --pylab=inline        # pylab in inline plotting mode
 ipython notebook --certfile=mycert.pem # use SSL/TLS certificate
 ipython notebook --port=5555 --ip=*    # Listen on port 5555, all interfaces
 """
 
 #-----------------------------------------------------------------------------
 # Helper functions
 #-----------------------------------------------------------------------------
 
 def url_path_join(a,b):
     if a.endswith('/') and b.startswith('/'):
         return a[:-1]+b
     else:
         return a+b
 
 #-----------------------------------------------------------------------------
 # The Tornado web application
 #-----------------------------------------------------------------------------
 
 class NotebookWebApplication(web.Application):
 
     def __init__(self, ipython_app, kernel_manager, notebook_manager, 
                  cluster_manager, log,
                  base_project_url, settings_overrides):
         handlers = [
             (r"/", ProjectDashboardHandler),
             (r"/login", LoginHandler),
             (r"/logout", LogoutHandler),
             (r"/new", NewHandler),
             (r"/%s" % _notebook_id_regex, NamedNotebookHandler),
             (r"/%s/copy" % _notebook_id_regex, NotebookCopyHandler),
             (r"/%s/print" % _notebook_id_regex, PrintNotebookHandler),
             (r"/kernels", MainKernelHandler),
             (r"/kernels/%s" % _kernel_id_regex, KernelHandler),
             (r"/kernels/%s/%s" % (_kernel_id_regex, _kernel_action_regex), KernelActionHandler),
             (r"/kernels/%s/iopub" % _kernel_id_regex, IOPubHandler),
             (r"/kernels/%s/shell" % _kernel_id_regex, ShellHandler),
             (r"/notebooks", NotebookRootHandler),
             (r"/notebooks/%s" % _notebook_id_regex, NotebookHandler),
             (r"/rstservice/render", RSTHandler),
             (r"/files/(.*)", AuthenticatedFileHandler, {'path' : notebook_manager.notebook_dir}),
             (r"/clusters", MainClusterHandler),
             (r"/clusters/%s/%s" % (_profile_regex, _cluster_action_regex), ClusterActionHandler),
             (r"/clusters/%s" % _profile_regex, ClusterProfileHandler),
         ]
         settings = dict(
             template_path=os.path.join(os.path.dirname(__file__), "templates"),
             static_path=os.path.join(os.path.dirname(__file__), "static"),
             cookie_secret=os.urandom(1024),
             login_url="/login",
         )
 
         # allow custom overrides for the tornado web app.
         settings.update(settings_overrides)
 
         # Python < 2.6.5 doesn't accept unicode keys in f(**kwargs), and
         # base_project_url will always be unicode, which will in turn
         # make the patterns unicode, and ultimately result in unicode
         # keys in kwargs to handler._execute(**kwargs) in tornado.
         # This enforces that base_project_url be ascii in that situation.
         # 
         # Note that the URLs these patterns check against are escaped,
         # and thus guaranteed to be ASCII: 'héllo' is really 'h%C3%A9llo'.
         base_project_url = py3compat.unicode_to_str(base_project_url, 'ascii')
         
         # prepend base_project_url onto the patterns that we match
         new_handlers = []
         for handler in handlers:
             pattern = url_path_join(base_project_url, handler[0])
             new_handler = tuple([pattern]+list(handler[1:]))
             new_handlers.append( new_handler )
 
         super(NotebookWebApplication, self).__init__(new_handlers, **settings)
 
         self.kernel_manager = kernel_manager
         self.notebook_manager = notebook_manager
         self.cluster_manager = cluster_manager
         self.ipython_app = ipython_app
         self.read_only = self.ipython_app.read_only
         self.log = log
 
 
 #-----------------------------------------------------------------------------
 # Aliases and Flags
 #-----------------------------------------------------------------------------
 
 flags = dict(ipkernel_flags)
 flags['no-browser']=(
     {'NotebookApp' : {'open_browser' : False}},
     "Don't open the notebook in a browser after startup."
 )
 flags['no-mathjax']=(
     {'NotebookApp' : {'enable_mathjax' : False}},
     """Disable MathJax
     
     MathJax is the javascript library IPython uses to render math/LaTeX. It is
     very large, so you may want to disable it if you have a slow internet
     connection, or for offline use of the notebook.
     
     When disabled, equations etc. will appear as their untransformed TeX source.
     """
 )
 flags['read-only'] = (
     {'NotebookApp' : {'read_only' : True}},
     """Allow read-only access to notebooks.
     
     When using a password to protect the notebook server, this flag
     allows unauthenticated clients to view the notebook list, and
     individual notebooks, but not edit them, start kernels, or run
     code.
     
     If no password is set, the server will be entirely read-only.
     """
 )
 
 # Add notebook manager flags
 flags.update(boolean_flag('script', 'NotebookManager.save_script',
                'Auto-save a .py script everytime the .ipynb notebook is saved',
                'Do not auto-save .py scripts for every notebook'))
 
 # the flags that are specific to the frontend
 # these must be scrubbed before being passed to the kernel,
 # or it will raise an error on unrecognized flags
 notebook_flags = ['no-browser', 'no-mathjax', 'read-only', 'script', 'no-script']
 
 aliases = dict(ipkernel_aliases)
 
 aliases.update({
     'ip': 'NotebookApp.ip',
     'port': 'NotebookApp.port',
     'keyfile': 'NotebookApp.keyfile',
     'certfile': 'NotebookApp.certfile',
     'notebook-dir': 'NotebookManager.notebook_dir',
     'browser': 'NotebookApp.browser',
 })
 
 # remove ipkernel flags that are singletons, and don't make sense in
 # multi-kernel evironment:
 aliases.pop('f', None)
 
 notebook_aliases = [u'port', u'ip', u'keyfile', u'certfile',
                     u'notebook-dir']
 
 #-----------------------------------------------------------------------------
 # NotebookApp
 #-----------------------------------------------------------------------------
 
 class NotebookApp(BaseIPythonApplication):
 
     name = 'ipython-notebook'
     default_config_file_name='ipython_notebook_config.py'
     
     description = """
         The IPython HTML Notebook.
         
         This launches a Tornado based HTML Notebook Server that serves up an
         HTML5/Javascript Notebook client.
     """
     examples = _examples
     
     classes = [IPKernelApp, ZMQInteractiveShell, ProfileDir, Session,
                MappingKernelManager, NotebookManager]
     flags = Dict(flags)
     aliases = Dict(aliases)
 
     kernel_argv = List(Unicode)
 
     log_level = Enum((0,10,20,30,40,50,'DEBUG','INFO','WARN','ERROR','CRITICAL'),
                     default_value=logging.INFO,
                     config=True,
                     help="Set the log level by value or name.")
 
     # create requested profiles by default, if they don't exist:
     auto_create = Bool(True)
 
     # file to be opened in the notebook server
     file_to_run = Unicode('')
 
     # Network related information.
 
     ip = Unicode(LOCALHOST, config=True,
         help="The IP address the notebook server will listen on."
     )
 
     def _ip_changed(self, name, old, new):
         if new == u'*': self.ip = u''
 
     port = Integer(8888, config=True,
         help="The port the notebook server will listen on."
     )
 
     certfile = Unicode(u'', config=True, 
         help="""The full path to an SSL/TLS certificate file."""
     )
     
     keyfile = Unicode(u'', config=True, 
         help="""The full path to a private key file for usage with SSL/TLS."""
     )
 
     password = Unicode(u'', config=True,
                       help="""Hashed password to use for web authentication.
 
                       To generate, type in a python/IPython shell:
 
                         from IPython.lib import passwd; passwd()
 
                       The string should be of the form type:salt:hashed-password.
                       """
     )
 
     open_browser = Bool(True, config=True,
                         help="""Whether to open in a browser after starting.
                         The specific browser used is platform dependent and
                         determined by the python standard library `webbrowser`
                         module, unless it is overridden using the --browser
                         (NotebookApp.browser) configuration option.
                         """)
 
     browser = Unicode(u'', config=True,
                       help="""Specify what command to use to invoke a web
                       browser when opening the notebook. If not specified, the
                       default browser will be determined by the `webbrowser`
                       standard library module, which allows setting of the
                       BROWSER environment variable to override it.
                       """)
     
     read_only = Bool(False, config=True,
         help="Whether to prevent editing/execution of notebooks."
     )
     
     webapp_settings = Dict(config=True,
             help="Supply overrides for the tornado.web.Application that the "
                  "IPython notebook uses.")
     
     enable_mathjax = Bool(True, config=True,
         help="""Whether to enable MathJax for typesetting math/TeX
 
         MathJax is the javascript library IPython uses to render math/LaTeX. It is
         very large, so you may want to disable it if you have a slow internet
         connection, or for offline use of the notebook.
 
         When disabled, equations etc. will appear as their untransformed TeX source.
         """
     )
     def _enable_mathjax_changed(self, name, old, new):
         """set mathjax url to empty if mathjax is disabled"""
         if not new:
             self.mathjax_url = u''
 
     base_project_url = Unicode('/', config=True,
                                help='''The base URL for the notebook server''')
     base_kernel_url = Unicode('/', config=True,
                                help='''The base URL for the kernel server''')
     websocket_host = Unicode("", config=True,
         help="""The hostname for the websocket server."""
     )
 
     mathjax_url = Unicode("", config=True,
         help="""The url for MathJax.js."""
     )
     def _mathjax_url_default(self):
         if not self.enable_mathjax:
             return u''
         static_path = self.webapp_settings.get("static_path", os.path.join(os.path.dirname(__file__), "static"))
         static_url_prefix = self.webapp_settings.get("static_url_prefix",
                                                      "/static/")
         if os.path.exists(os.path.join(static_path, 'mathjax', "MathJax.js")):
             self.log.info("Using local MathJax")
             return static_url_prefix+u"mathjax/MathJax.js"
         else:
             if self.certfile:
                 # HTTPS: load from Rackspace CDN, because SSL certificate requires it
                 base = u"https://c328740.ssl.cf1.rackcdn.com"
             else:
                 base = u"http://cdn.mathjax.org"
             
             url = base + u"/mathjax/latest/MathJax.js"
             self.log.info("Using MathJax from CDN: %s", url)
             return url
     
     def _mathjax_url_changed(self, name, old, new):
         if new and not self.enable_mathjax:
             # enable_mathjax=False overrides mathjax_url
             self.mathjax_url = u''
         else:
             self.log.info("Using MathJax: %s", new)
 
     def parse_command_line(self, argv=None):
         super(NotebookApp, self).parse_command_line(argv)
         if argv is None:
             argv = sys.argv[1:]
 
         # Scrub frontend-specific flags
         self.kernel_argv = swallow_argv(argv, notebook_aliases, notebook_flags)
         # Kernel should inherit default config file from frontend
         self.kernel_argv.append("--KernelApp.parent_appname='%s'"%self.name)
 
         if self.extra_args:
             self.file_to_run = os.path.abspath(self.extra_args[0])
             self.config.NotebookManager.notebook_dir = os.path.dirname(self.file_to_run)
 
     def init_configurables(self):
         # force Session default to be secure
         default_secure(self.config)
         # Create a KernelManager and start a kernel.
         self.kernel_manager = MappingKernelManager(
             config=self.config, log=self.log, kernel_argv=self.kernel_argv,
             connection_dir = self.profile_dir.security_dir,
         )
         self.notebook_manager = NotebookManager(config=self.config, log=self.log)
         self.notebook_manager.list_notebooks()
         self.cluster_manager = ClusterManager(config=self.config, log=self.log)
         self.cluster_manager.update_profiles()
 
     def init_logging(self):
-        super(NotebookApp, self).init_logging()
         # This prevents double log messages because tornado use a root logger that
         # self.log is a child of. The logging module dipatches log messages to a log
         # and all of its ancenstors until propagate is set to False.
         self.log.propagate = False
     
     def init_webapp(self):
         """initialize tornado webapp and httpserver"""
         self.web_app = NotebookWebApplication(
             self, self.kernel_manager, self.notebook_manager, 
             self.cluster_manager, self.log,
             self.base_project_url, self.webapp_settings
         )
         if self.certfile:
             ssl_options = dict(certfile=self.certfile)
             if self.keyfile:
                 ssl_options['keyfile'] = self.keyfile
         else:
             ssl_options = None
         self.web_app.password = self.password
         self.http_server = httpserver.HTTPServer(self.web_app, ssl_options=ssl_options)
         if ssl_options is None and not self.ip and not (self.read_only and not self.password):
             self.log.critical('WARNING: the notebook server is listening on all IP addresses '
                               'but not using any encryption or authentication. This is highly '
                               'insecure and not recommended.')
 
         # Try random ports centered around the default.
         from random import randint
         n = 50  # Max number of attempts, keep reasonably large.
         for port in range(self.port, self.port+5) + [self.port + randint(-2*n, 2*n) for i in range(n-5)]:
             try:
                 self.http_server.listen(port, self.ip)
             except socket.error, e:
                 if e.errno != errno.EADDRINUSE:
                     raise
                 self.log.info('The port %i is already in use, trying another random port.' % port)
             else:
                 self.port = port
                 break
     
     def init_signal(self):
         # FIXME: remove this check when pyzmq dependency is >= 2.1.11
         # safely extract zmq version info:
         try:
             zmq_v = zmq.pyzmq_version_info()
         except AttributeError:
             zmq_v = [ int(n) for n in re.findall(r'\d+', zmq.__version__) ]
             if 'dev' in zmq.__version__:
                 zmq_v.append(999)
             zmq_v = tuple(zmq_v)
         if zmq_v >= (2,1,9):
             # This won't work with 2.1.7 and
             # 2.1.9-10 will log ugly 'Interrupted system call' messages,
             # but it will work
             signal.signal(signal.SIGINT, self._handle_sigint)
         signal.signal(signal.SIGTERM, self._signal_stop)
     
     def _handle_sigint(self, sig, frame):
         """SIGINT handler spawns confirmation dialog"""
         # register more forceful signal handler for ^C^C case
         signal.signal(signal.SIGINT, self._signal_stop)
         # request confirmation dialog in bg thread, to avoid
         # blocking the App
         thread = threading.Thread(target=self._confirm_exit)
         thread.daemon = True
         thread.start()
     
     def _restore_sigint_handler(self):
         """callback for restoring original SIGINT handler"""
         signal.signal(signal.SIGINT, self._handle_sigint)
     
     def _confirm_exit(self):
         """confirm shutdown on ^C
         
         A second ^C, or answering 'y' within 5s will cause shutdown,
         otherwise original SIGINT handler will be restored.
         """
         # FIXME: remove this delay when pyzmq dependency is >= 2.1.11
         time.sleep(0.1)
         sys.stdout.write("Shutdown Notebook Server (y/[n])? ")
         sys.stdout.flush()
         r,w,x = select.select([sys.stdin], [], [], 5)
         if r:
             line = sys.stdin.readline()
             if line.lower().startswith('y'):
                 self.log.critical("Shutdown confirmed")
                 ioloop.IOLoop.instance().stop()
                 return
         else:
             print "No answer for 5s:",
         print "resuming operation..."
         # no answer, or answer is no:
         # set it back to original SIGINT handler
         # use IOLoop.add_callback because signal.signal must be called
         # from main thread
         ioloop.IOLoop.instance().add_callback(self._restore_sigint_handler)
     
     def _signal_stop(self, sig, frame):
         self.log.critical("received signal %s, stopping", sig)
         ioloop.IOLoop.instance().stop()
     
     @catch_config_error
     def initialize(self, argv=None):
+        self.init_logging()
         super(NotebookApp, self).initialize(argv)
         self.init_configurables()
         self.init_webapp()
         self.init_signal()
 
     def cleanup_kernels(self):
         """shutdown all kernels
         
         The kernels will shutdown themselves when this process no longer exists,
         but explicit shutdown allows the KernelManagers to cleanup the connection files.
         """
         self.log.info('Shutting down kernels')
         km = self.kernel_manager
         # copy list, since kill_kernel deletes keys
         for kid in list(km.kernel_ids):
             km.kill_kernel(kid)
 
     def start(self):
         ip = self.ip if self.ip else '[all ip addresses on your system]'
         proto = 'https' if self.certfile else 'http'
         info = self.log.info
         info("The IPython Notebook is running at: %s://%s:%i%s" %
              (proto, ip, self.port,self.base_project_url) )
         info("Use Control-C to stop this server and shut down all kernels.")
 
         if self.open_browser:
             ip = self.ip or '127.0.0.1'
             if self.browser:
                 browser = webbrowser.get(self.browser)
             else:
                 browser = webbrowser.get()
 
             if self.file_to_run:
                 filename, _ = os.path.splitext(os.path.basename(self.file_to_run))
                 for nb in self.notebook_manager.list_notebooks():
                     if filename == nb['name']:
                         url = nb['notebook_id']
                         break
                 else:
                     url = ''
             else:
                 url = ''
             b = lambda : browser.open("%s://%s:%i%s%s" % (proto, ip,
                                         self.port, self.base_project_url, url),
                                     new=2)
             threading.Thread(target=b).start()
         try:
             ioloop.IOLoop.instance().start()
         except KeyboardInterrupt:
             info("Interrupted...")
         finally:
             self.cleanup_kernels()
     
 
 #-----------------------------------------------------------------------------
 # Main entry point
 #-----------------------------------------------------------------------------
 
 def launch_new_instance():
     app = NotebookApp.instance()
     app.initialize()
     app.start()
 

IPython/frontend/qt/console/qtconsoleapp.py

@@ -1,382 +1,372 b''
 """ A minimal application using the Qt console-style IPython frontend.
 
 This is not a complete console app, as subprocess will not be able to receive
 input, there is no real readline support, among other limitations.
 
 Authors:
 
 * Evan Patterson
 * Min RK
 * Erik Tollerud
 * Fernando Perez
 * Bussonnier Matthias
 * Thomas Kluyver
 * Paul Ivanov
 
 """
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 # stdlib imports
 import json
 import os
 import signal
 import sys
 import uuid
 
 # If run on Windows, install an exception hook which pops up a
 # message box. Pythonw.exe hides the console, so without this
 # the application silently fails to load.
 #
 # We always install this handler, because the expectation is for
 # qtconsole to bring up a GUI even if called from the console.
 # The old handler is called, so the exception is printed as well.
 # If desired, check for pythonw with an additional condition
 # (sys.executable.lower().find('pythonw.exe') >= 0).
 if os.name == 'nt':
     old_excepthook = sys.excepthook
 
     def gui_excepthook(exctype, value, tb):
         try:
             import ctypes, traceback
             MB_ICONERROR = 0x00000010L
             title = u'Error starting IPython QtConsole'
             msg = u''.join(traceback.format_exception(exctype, value, tb))
             ctypes.windll.user32.MessageBoxW(0, msg, title, MB_ICONERROR)
         finally:
             # Also call the old exception hook to let it do
             # its thing too.
             old_excepthook(exctype, value, tb)
 
     sys.excepthook = gui_excepthook
 
 # System library imports
 from IPython.external.qt import QtCore, QtGui
 
 # Local imports
 from IPython.config.application import boolean_flag, catch_config_error
 from IPython.core.application import BaseIPythonApplication
 from IPython.core.profiledir import ProfileDir
 from IPython.lib.kernel import tunnel_to_kernel, find_connection_file
 from IPython.frontend.qt.console.frontend_widget import FrontendWidget
 from IPython.frontend.qt.console.ipython_widget import IPythonWidget
 from IPython.frontend.qt.console.rich_ipython_widget import RichIPythonWidget
 from IPython.frontend.qt.console import styles
 from IPython.frontend.qt.console.mainwindow import MainWindow
 from IPython.frontend.qt.kernelmanager import QtKernelManager
 from IPython.utils.path import filefind
 from IPython.utils.py3compat import str_to_bytes
 from IPython.utils.traitlets import (
     Dict, List, Unicode, Integer, CaselessStrEnum, CBool, Any
 )
 from IPython.zmq.ipkernel import IPKernelApp
 from IPython.zmq.session import Session, default_secure
 from IPython.zmq.zmqshell import ZMQInteractiveShell
 
 from IPython.frontend.consoleapp import (
         IPythonConsoleApp, app_aliases, app_flags, flags, aliases
     )
 
 #-----------------------------------------------------------------------------
 # Network Constants
 #-----------------------------------------------------------------------------
 
 from IPython.utils.localinterfaces import LOCALHOST, LOCAL_IPS
 
 #-----------------------------------------------------------------------------
 # Globals
 #-----------------------------------------------------------------------------
 
 _examples = """
 ipython qtconsole                 # start the qtconsole
 ipython qtconsole --pylab=inline  # start with pylab in inline plotting mode
 """
 
 #-----------------------------------------------------------------------------
 # Aliases and Flags
 #-----------------------------------------------------------------------------
 
 # start with copy of flags
 flags = dict(flags)
 qt_flags = {
-    'pure' : ({'IPythonQtConsoleApp' : {'pure' : True}},
-            "Use a pure Python kernel instead of an IPython kernel."),
-    'plain' : ({'ConsoleWidget' : {'kind' : 'plain'}},
+    'plain' : ({'IPythonQtConsoleApp' : {'plain' : True}},
             "Disable rich text support."),
 }
 qt_flags.update(boolean_flag(
     'gui-completion', 'ConsoleWidget.gui_completion',
     "use a GUI widget for tab completion",
     "use plaintext output for completion"
 ))
 # and app_flags from the Console Mixin
 qt_flags.update(app_flags)
 # add frontend flags to the full set
 flags.update(qt_flags)
 
 # start with copy of front&backend aliases list
 aliases = dict(aliases)
 qt_aliases = dict(
 
     style = 'IPythonWidget.syntax_style',
     stylesheet = 'IPythonQtConsoleApp.stylesheet',
     colors = 'ZMQInteractiveShell.colors',
 
     editor = 'IPythonWidget.editor',
     paging = 'ConsoleWidget.paging',
 )
 # and app_aliases from the Console Mixin
 qt_aliases.update(app_aliases)
 # add frontend aliases to the full set
 aliases.update(qt_aliases)
 
 # get flags&aliases into sets, and remove a couple that
 # shouldn't be scrubbed from backend flags:
 qt_aliases = set(qt_aliases.keys())
 qt_aliases.remove('colors')
 qt_flags = set(qt_flags.keys())
 
 #-----------------------------------------------------------------------------
 # Classes
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # IPythonQtConsole
 #-----------------------------------------------------------------------------
 
 
 class IPythonQtConsoleApp(BaseIPythonApplication, IPythonConsoleApp):
     name = 'ipython-qtconsole'
 
     description = """
         The IPython QtConsole.
         
         This launches a Console-style application using Qt.  It is not a full
         console, in that launched terminal subprocesses will not be able to accept
         input.
         
         The QtConsole supports various extra features beyond the Terminal IPython
         shell, such as inline plotting with matplotlib, via:
         
             ipython qtconsole --pylab=inline
         
         as well as saving your session as HTML, and printing the output.
         
     """
     examples = _examples
 
     classes = [IPKernelApp, IPythonWidget, ZMQInteractiveShell, ProfileDir, Session]
     flags = Dict(flags)
     aliases = Dict(aliases)
     frontend_flags = Any(qt_flags)
     frontend_aliases = Any(qt_aliases)
     kernel_manager_class = QtKernelManager
 
     stylesheet = Unicode('', config=True,
         help="path to a custom CSS stylesheet")
 
     plain = CBool(False, config=True,
         help="Use a plaintext widget instead of rich text (plain can't print/save).")
 
-    def _pure_changed(self, name, old, new):
-        kind = 'plain' if self.plain else 'rich'
+    def _plain_changed(self, name, old, new):
+        kind = 'plain' if new else 'rich'
         self.config.ConsoleWidget.kind = kind
-        if self.pure:
-            self.widget_factory = FrontendWidget
-        elif self.plain:
+        if new:
             self.widget_factory = IPythonWidget
         else:
             self.widget_factory = RichIPythonWidget
 
-    _plain_changed = _pure_changed
-
     # the factory for creating a widget
     widget_factory = Any(RichIPythonWidget)
 
     def parse_command_line(self, argv=None):
         super(IPythonQtConsoleApp, self).parse_command_line(argv)
         self.build_kernel_argv(argv)
 
 
     def new_frontend_master(self):
         """ Create and return new frontend attached to new kernel, launched on localhost.
         """
         ip = self.ip if self.ip in LOCAL_IPS else LOCALHOST
         kernel_manager = self.kernel_manager_class(
                                 ip=ip,
                                 connection_file=self._new_connection_file(),
                                 config=self.config,
         )
         # start the kernel
-        kwargs = dict(ipython=not self.pure)
+        kwargs = dict()
         kwargs['extra_arguments'] = self.kernel_argv
         kernel_manager.start_kernel(**kwargs)
         kernel_manager.start_channels()
         widget = self.widget_factory(config=self.config,
                                    local_kernel=True)
         self.init_colors(widget)
         widget.kernel_manager = kernel_manager
         widget._existing = False
         widget._may_close = True
         widget._confirm_exit = self.confirm_exit
         return widget
 
     def new_frontend_slave(self, current_widget):
         """Create and return a new frontend attached to an existing kernel.
         
         Parameters
         ----------
         current_widget : IPythonWidget
             The IPythonWidget whose kernel this frontend is to share
         """
         kernel_manager = self.kernel_manager_class(
                                 connection_file=current_widget.kernel_manager.connection_file,
                                 config = self.config,
         )
         kernel_manager.load_connection_file()
         kernel_manager.start_channels()
         widget = self.widget_factory(config=self.config,
                                 local_kernel=False)
         self.init_colors(widget)
         widget._existing = True
         widget._may_close = False
         widget._confirm_exit = False
         widget.kernel_manager = kernel_manager
         return widget
 
     def init_qt_elements(self):
         # Create the widget.
         self.app = QtGui.QApplication([])
 
         base_path = os.path.abspath(os.path.dirname(__file__))
         icon_path = os.path.join(base_path, 'resources', 'icon', 'IPythonConsole.svg')
         self.app.icon = QtGui.QIcon(icon_path)
         QtGui.QApplication.setWindowIcon(self.app.icon)
 
         local_kernel = (not self.existing) or self.ip in LOCAL_IPS
         self.widget = self.widget_factory(config=self.config,
                                         local_kernel=local_kernel)
         self.init_colors(self.widget)
         self.widget._existing = self.existing
         self.widget._may_close = not self.existing
         self.widget._confirm_exit = self.confirm_exit
 
         self.widget.kernel_manager = self.kernel_manager
         self.window = MainWindow(self.app,
                                 confirm_exit=self.confirm_exit,
                                 new_frontend_factory=self.new_frontend_master,
                                 slave_frontend_factory=self.new_frontend_slave,
                                 )
         self.window.log = self.log
         self.window.add_tab_with_frontend(self.widget)
         self.window.init_menu_bar()
 
-        self.window.setWindowTitle('Python' if self.pure else 'IPython')
+        self.window.setWindowTitle('IPython')
 
     def init_colors(self, widget):
         """Configure the coloring of the widget"""
         # Note: This will be dramatically simplified when colors
         # are removed from the backend.
 
-        if self.pure:
-            # only IPythonWidget supports styling
-            return
-
         # parse the colors arg down to current known labels
         try:
             colors = self.config.ZMQInteractiveShell.colors
         except AttributeError:
             colors = None
         try:
             style = self.config.IPythonWidget.syntax_style
         except AttributeError:
             style = None
         try:
             sheet = self.config.IPythonWidget.style_sheet
         except AttributeError:
             sheet = None
 
         # find the value for colors:
         if colors:
             colors=colors.lower()
             if colors in ('lightbg', 'light'):
                 colors='lightbg'
             elif colors in ('dark', 'linux'):
                 colors='linux'
             else:
                 colors='nocolor'
         elif style:
             if style=='bw':
                 colors='nocolor'
             elif styles.dark_style(style):
                 colors='linux'
             else:
                 colors='lightbg'
         else:
             colors=None
 
         # Configure the style
         if style:
             widget.style_sheet = styles.sheet_from_template(style, colors)
             widget.syntax_style = style
             widget._syntax_style_changed()
             widget._style_sheet_changed()
         elif colors:
             # use a default dark/light/bw style
             widget.set_default_style(colors=colors)
 
         if self.stylesheet:
             # we got an explicit stylesheet
             if os.path.isfile(self.stylesheet):
                 with open(self.stylesheet) as f:
                     sheet = f.read()
             else:
                 raise IOError("Stylesheet %r not found." % self.stylesheet)
         if sheet:
             widget.style_sheet = sheet
             widget._style_sheet_changed()
             
 
     def init_signal(self):
         """allow clean shutdown on sigint"""
         signal.signal(signal.SIGINT, lambda sig, frame: self.exit(-2))
         # need a timer, so that QApplication doesn't block until a real
         # Qt event fires (can require mouse movement)
         # timer trick from http://stackoverflow.com/q/4938723/938949
         timer = QtCore.QTimer()
          # Let the interpreter run each 200 ms:
         timer.timeout.connect(lambda: None)
         timer.start(200)
         # hold onto ref, so the timer doesn't get cleaned up
         self._sigint_timer = timer
 
     @catch_config_error
     def initialize(self, argv=None):
         super(IPythonQtConsoleApp, self).initialize(argv)
         IPythonConsoleApp.initialize(self,argv)
         self.init_qt_elements()
         self.init_signal()
 
     def start(self):
 
         # draw the window
         self.window.show()
         self.window.raise_()
 
         # Start the application main loop.
         self.app.exec_()
 
 #-----------------------------------------------------------------------------
 # Main entry point
 #-----------------------------------------------------------------------------
 
 def main():
     app = IPythonQtConsoleApp()
     app.initialize()
     app.start()
 
 
 if __name__ == '__main__':
     main()

IPython/lib/kernel.py

@@ -1,315 +1,315 b''
 """Utilities for connecting to kernels
 
 Authors:
 
 * Min Ragan-Kelley
 
 """
 
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 import glob
 import json
 import os
 import sys
 from getpass import getpass
 from subprocess import Popen, PIPE
 
 # external imports
 from IPython.external.ssh import tunnel
 
 # IPython imports
 from IPython.core.profiledir import ProfileDir
 from IPython.utils.path import filefind, get_ipython_dir
 from IPython.utils.py3compat import str_to_bytes
 
 
 #-----------------------------------------------------------------------------
 # Functions
 #-----------------------------------------------------------------------------
 
 def get_connection_file(app=None):
     """Return the path to the connection file of an app
     
     Parameters
     ----------
     app : KernelApp instance [optional]
         If unspecified, the currently running app will be used
     """
     if app is None:
-        from IPython.zmq.kernelapp import KernelApp
-        if not KernelApp.initialized():
+        from IPython.zmq.ipkernel import IPKernelApp
+        if not IPKernelApp.initialized():
             raise RuntimeError("app not specified, and not in a running Kernel")
 
-        app = KernelApp.instance()
+        app = IPKernelApp.instance()
     return filefind(app.connection_file, ['.', app.profile_dir.security_dir])
 
 def find_connection_file(filename, profile=None):
     """find a connection file, and return its absolute path.
     
     The current working directory and the profile's security
     directory will be searched for the file if it is not given by
     absolute path.
     
     If profile is unspecified, then the current running application's
     profile will be used, or 'default', if not run from IPython.
     
     If the argument does not match an existing file, it will be interpreted as a
     fileglob, and the matching file in the profile's security dir with
     the latest access time will be used.
     
     Parameters
     ----------
     filename : str
         The connection file or fileglob to search for.
     profile : str [optional]
         The name of the profile to use when searching for the connection file,
         if different from the current IPython session or 'default'.
     
     Returns
     -------
     str : The absolute path of the connection file.
     """
     from IPython.core.application import BaseIPythonApplication as IPApp
     try:
         # quick check for absolute path, before going through logic
         return filefind(filename)
     except IOError:
         pass
     
     if profile is None:
         # profile unspecified, check if running from an IPython app
         if IPApp.initialized():
             app = IPApp.instance()
             profile_dir = app.profile_dir
         else:
             # not running in IPython, use default profile
             profile_dir = ProfileDir.find_profile_dir_by_name(get_ipython_dir(), 'default')
     else:
         # find profiledir by profile name:
         profile_dir = ProfileDir.find_profile_dir_by_name(get_ipython_dir(), profile)
     security_dir = profile_dir.security_dir
     
     try:
         # first, try explicit name
         return filefind(filename, ['.', security_dir])
     except IOError:
         pass
     
     # not found by full name
     
     if '*' in filename:
         # given as a glob already
         pat = filename
     else:
         # accept any substring match
         pat = '*%s*' % filename
     matches = glob.glob( os.path.join(security_dir, pat) )
     if not matches:
         raise IOError("Could not find %r in %r" % (filename, security_dir))
     elif len(matches) == 1:
         return matches[0]
     else:
         # get most recent match, by access time:
         return sorted(matches, key=lambda f: os.stat(f).st_atime)[-1]
 
 def get_connection_info(connection_file=None, unpack=False, profile=None):
     """Return the connection information for the current Kernel.
     
     Parameters
     ----------
     connection_file : str [optional]
         The connection file to be used. Can be given by absolute path, or
         IPython will search in the security directory of a given profile.
         If run from IPython, 
         
         If unspecified, the connection file for the currently running
         IPython Kernel will be used, which is only allowed from inside a kernel.
     unpack : bool [default: False]
         if True, return the unpacked dict, otherwise just the string contents
         of the file.
     profile : str [optional]
         The name of the profile to use when searching for the connection file,
         if different from the current IPython session or 'default'.
         
     
     Returns
     -------
     The connection dictionary of the current kernel, as string or dict,
     depending on `unpack`.
     """
     if connection_file is None:
         # get connection file from current kernel
         cf = get_connection_file()
     else:
         # connection file specified, allow shortnames:
         cf = find_connection_file(connection_file, profile=profile)
     
     with open(cf) as f:
         info = f.read()
     
     if unpack:
         info = json.loads(info)
         # ensure key is bytes:
         info['key'] = str_to_bytes(info.get('key', ''))
     return info
 
 def connect_qtconsole(connection_file=None, argv=None, profile=None):
     """Connect a qtconsole to the current kernel.
     
     This is useful for connecting a second qtconsole to a kernel, or to a
     local notebook.
     
     Parameters
     ----------
     connection_file : str [optional]
         The connection file to be used. Can be given by absolute path, or
         IPython will search in the security directory of a given profile.
         If run from IPython, 
         
         If unspecified, the connection file for the currently running
         IPython Kernel will be used, which is only allowed from inside a kernel.
     argv : list [optional]
         Any extra args to be passed to the console.
     profile : str [optional]
         The name of the profile to use when searching for the connection file,
         if different from the current IPython session or 'default'.
     
     
     Returns
     -------
     subprocess.Popen instance running the qtconsole frontend
     """
     argv = [] if argv is None else argv
     
     if connection_file is None:
         # get connection file from current kernel
         cf = get_connection_file()
     else:
         cf = find_connection_file(connection_file, profile=profile)
     
     cmd = ';'.join([
         "from IPython.frontend.qt.console import qtconsoleapp",
         "qtconsoleapp.main()"
     ])
     
     return Popen([sys.executable, '-c', cmd, '--existing', cf] + argv, stdout=PIPE, stderr=PIPE)
 
 def tunnel_to_kernel(connection_info, sshserver, sshkey=None):
     """tunnel connections to a kernel via ssh
     
     This will open four SSH tunnels from localhost on this machine to the
     ports associated with the kernel.  They can be either direct
     localhost-localhost tunnels, or if an intermediate server is necessary,
     the kernel must be listening on a public IP.
     
     Parameters
     ----------
     connection_info : dict or str (path)
         Either a connection dict, or the path to a JSON connection file
     sshserver : str
         The ssh sever to use to tunnel to the kernel. Can be a full
         `user@server:port` string. ssh config aliases are respected.
     sshkey : str [optional]
         Path to file containing ssh key to use for authentication.
         Only necessary if your ssh config does not already associate
         a keyfile with the host.
     
     Returns
     -------
     
     (shell, iopub, stdin, hb) : ints
         The four ports on localhost that have been forwarded to the kernel.
     """
     if isinstance(connection_info, basestring):
         # it's a path, unpack it
         with open(connection_info) as f:
             connection_info = json.loads(f.read())
     
     cf = connection_info
     
     lports = tunnel.select_random_ports(4)
     rports = cf['shell_port'], cf['iopub_port'], cf['stdin_port'], cf['hb_port']
     
     remote_ip = cf['ip']
     
     if tunnel.try_passwordless_ssh(sshserver, sshkey):
         password=False
     else:
         password = getpass("SSH Password for %s: "%sshserver)
     
     for lp,rp in zip(lports, rports):
         tunnel.ssh_tunnel(lp, rp, sshserver, remote_ip, sshkey, password)
     
     return tuple(lports)
     
 
 def swallow_argv(argv, aliases=None, flags=None):
     """strip frontend-specific aliases and flags from an argument list
     
     For use primarily in frontend apps that want to pass a subset of command-line
     arguments through to a subprocess, where frontend-specific flags and aliases
     should be removed from the list.
     
     Parameters
     ----------
     
     argv : list(str)
         The starting argv, to be filtered
     aliases : container of aliases (dict, list, set, etc.)
         The frontend-specific aliases to be removed
     flags : container of flags (dict, list, set, etc.)
         The frontend-specific flags to be removed
     
     Returns
     -------
     
     argv : list(str)
         The argv list, excluding flags and aliases that have been stripped
     """
     
     if aliases is None:
         aliases = set()
     if flags is None:
         flags = set()
     
     stripped = list(argv) # copy
     
     swallow_next = False
     was_flag = False
     for a in argv:
         if swallow_next:
             swallow_next = False
             # last arg was an alias, remove the next one
             # *unless* the last alias has a no-arg flag version, in which
             # case, don't swallow the next arg if it's also a flag:
             if not (was_flag and a.startswith('-')):
                 stripped.remove(a)
                 continue
         if a.startswith('-'):
             split = a.lstrip('-').split('=')
             alias = split[0]
             if alias in aliases:
                 stripped.remove(a)
                 if len(split) == 1:
                     # alias passed with arg via space
                     swallow_next = True
                     # could have been a flag that matches an alias, e.g. `existing`
                     # in which case, we might not swallow the next arg
                     was_flag = alias in flags
             elif alias in flags and len(split) == 1:
                 # strip flag, but don't swallow next, as flags don't take args
                 stripped.remove(a)
     
     # return shortened list
     return stripped
 

IPython/parallel/__init__.py

@@ -1,40 +1,62 b''
 """The IPython ZMQ-based parallel computing interface.
 
 Authors:
 
 * MinRK
 """
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2011 The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 import os
 import warnings
 
 import zmq
 
 from IPython.zmq import check_for_zmq
 
 if os.name == 'nt':
     min_pyzmq = '2.1.7'
 else:
     min_pyzmq = '2.1.4'
 
 check_for_zmq(min_pyzmq, 'IPython.parallel')
 
 from IPython.utils.pickleutil import Reference
 
 from .client.asyncresult import *
 from .client.client import Client
 from .client.remotefunction import *
 from .client.view import *
+from .util import interactive
 from .controller.dependency import *
 
+#-----------------------------------------------------------------------------
+# Functions
+#-----------------------------------------------------------------------------
+
+
+def bind_kernel(**kwargs):
+    """Bind an Engine's Kernel to be used as a full IPython kernel.
+    
+    This allows a running Engine to be used simultaneously as a full IPython kernel
+    with the QtConsole or other frontends.
+    
+    This function returns immediately.
+    """
+    from IPython.parallel.apps.ipengineapp import IPEngineApp
+    if IPEngineApp.initialized():
+        app = IPEngineApp.instance()
+    else:
+        raise RuntimeError("Must be called from an IPEngineApp instance")
+    
+    return app.bind_kernel(**kwargs)
+
 

IPython/parallel/apps/baseapp.py

@@ -1,265 +1,272 b''
 # encoding: utf-8
 """
 The Base Application class for IPython.parallel apps
 
 Authors:
 
 * Brian Granger
 * Min RK
 
 """
 
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2008-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 from __future__ import with_statement
 
 import os
 import logging
 import re
 import sys
 
 from subprocess import Popen, PIPE
 
 from IPython.config.application import catch_config_error
 from IPython.core import release
 from IPython.core.crashhandler import CrashHandler
 from IPython.core.application import (
     BaseIPythonApplication,
     base_aliases as base_ip_aliases,
     base_flags as base_ip_flags
 )
 from IPython.utils.path import expand_path
 
 from IPython.utils.traitlets import Unicode, Bool, Instance, Dict, List
 
 #-----------------------------------------------------------------------------
 # Module errors
 #-----------------------------------------------------------------------------
 
 class PIDFileError(Exception):
     pass
 
 
 #-----------------------------------------------------------------------------
 # Crash handler for this application
 #-----------------------------------------------------------------------------
 
 class ParallelCrashHandler(CrashHandler):
     """sys.excepthook for IPython itself, leaves a detailed report on disk."""
 
     def __init__(self, app):
         contact_name = release.authors['Min'][0]
         contact_email = release.author_email
         bug_tracker = 'https://github.com/ipython/ipython/issues'
         super(ParallelCrashHandler,self).__init__(
             app, contact_name, contact_email, bug_tracker
         )
 
 
 #-----------------------------------------------------------------------------
 # Main application
 #-----------------------------------------------------------------------------
 base_aliases = {}
 base_aliases.update(base_ip_aliases)
 base_aliases.update({
     'profile-dir' : 'ProfileDir.location',
     'work-dir' : 'BaseParallelApplication.work_dir',
     'log-to-file' : 'BaseParallelApplication.log_to_file',
     'clean-logs' : 'BaseParallelApplication.clean_logs',
     'log-url' : 'BaseParallelApplication.log_url',
     'cluster-id' : 'BaseParallelApplication.cluster_id',
 })
 
 base_flags = {
     'log-to-file' : (
         {'BaseParallelApplication' : {'log_to_file' : True}},
         "send log output to a file"
     )
 }
 base_flags.update(base_ip_flags)
 
 class BaseParallelApplication(BaseIPythonApplication):
     """The base Application for IPython.parallel apps
     
     Principle extensions to BaseIPyythonApplication:
     
     * work_dir
     * remote logging via pyzmq
     * IOLoop instance
     """
 
     crash_handler_class = ParallelCrashHandler
     
     def _log_level_default(self):
         # temporarily override default_log_level to INFO
         return logging.INFO
+    
+    def _log_format_default(self):
+        """override default log format to include time"""
+        return u"%(asctime)s.%(msecs).03d [%(name)s] %(message)s"
 
     work_dir = Unicode(os.getcwdu(), config=True,
         help='Set the working dir for the process.'
     )
     def _work_dir_changed(self, name, old, new):
         self.work_dir = unicode(expand_path(new))
 
     log_to_file = Bool(config=True,
         help="whether to log to a file")
 
     clean_logs = Bool(False, config=True,
         help="whether to cleanup old logfiles before starting")
 
     log_url = Unicode('', config=True,
         help="The ZMQ URL of the iplogger to aggregate logging.")
 
     cluster_id = Unicode('', config=True,
         help="""String id to add to runtime files, to prevent name collisions when
         using multiple clusters with a single profile simultaneously.
         
         When set, files will be named like: 'ipcontroller-<cluster_id>-engine.json'
         
         Since this is text inserted into filenames, typical recommendations apply:
         Simple character strings are ideal, and spaces are not recommended (but should
         generally work).
         """
     )
     def _cluster_id_changed(self, name, old, new):
         self.name = self.__class__.name
         if new:
             self.name += '-%s'%new
     
     def _config_files_default(self):
         return ['ipcontroller_config.py', 'ipengine_config.py', 'ipcluster_config.py']
     
     loop = Instance('zmq.eventloop.ioloop.IOLoop')
     def _loop_default(self):
         from zmq.eventloop.ioloop import IOLoop
         return IOLoop.instance()
 
     aliases = Dict(base_aliases)
     flags = Dict(base_flags)
     
     @catch_config_error
     def initialize(self, argv=None):
         """initialize the app"""
         super(BaseParallelApplication, self).initialize(argv)
         self.to_work_dir()
         self.reinit_logging()
         
     def to_work_dir(self):
         wd = self.work_dir
         if unicode(wd) != os.getcwdu():
             os.chdir(wd)
             self.log.info("Changing to working dir: %s" % wd)
         # This is the working dir by now.
         sys.path.insert(0, '')
 
     def reinit_logging(self):
         # Remove old log files
         log_dir = self.profile_dir.log_dir
         if self.clean_logs:
             for f in os.listdir(log_dir):
                 if re.match(r'%s-\d+\.(log|err|out)'%self.name,f):
                     os.remove(os.path.join(log_dir, f))
         if self.log_to_file:
             # Start logging to the new log file
             log_filename = self.name + u'-' + str(os.getpid()) + u'.log'
             logfile = os.path.join(log_dir, log_filename)
             open_log_file = open(logfile, 'w')
         else:
             open_log_file = None
         if open_log_file is not None:
-            self.log.removeHandler(self._log_handler)
+            while self.log.handlers:
+                self.log.removeHandler(self.log.handlers[0])
             self._log_handler = logging.StreamHandler(open_log_file)
             self.log.addHandler(self._log_handler)
+        else:
+            self._log_handler = self.log.handlers[0]
         # Add timestamps to log format:
-        self._log_formatter = logging.Formatter("%(asctime)s.%(msecs).03d [%(name)s] %(message)s",
+        self._log_formatter = logging.Formatter(self.log_format,
                                                 datefmt="%Y-%m-%d %H:%M:%S")
         self._log_handler.setFormatter(self._log_formatter)
         # do not propagate log messages to root logger
         # ipcluster app will sometimes print duplicate messages during shutdown
         # if this is 1 (default):
         self.log.propagate = False
 
     def write_pid_file(self, overwrite=False):
         """Create a .pid file in the pid_dir with my pid.
 
         This must be called after pre_construct, which sets `self.pid_dir`.
         This raises :exc:`PIDFileError` if the pid file exists already.
         """
         pid_file = os.path.join(self.profile_dir.pid_dir, self.name + u'.pid')
         if os.path.isfile(pid_file):
             pid = self.get_pid_from_file()
             if not overwrite:
                 raise PIDFileError(
                     'The pid file [%s] already exists. \nThis could mean that this '
                     'server is already running with [pid=%s].' % (pid_file, pid)
                 )
         with open(pid_file, 'w') as f:
             self.log.info("Creating pid file: %s" % pid_file)
             f.write(repr(os.getpid())+'\n')
 
     def remove_pid_file(self):
         """Remove the pid file.
 
         This should be called at shutdown by registering a callback with
         :func:`reactor.addSystemEventTrigger`. This needs to return
         ``None``.
         """
         pid_file = os.path.join(self.profile_dir.pid_dir, self.name + u'.pid')
         if os.path.isfile(pid_file):
             try:
                 self.log.info("Removing pid file: %s" % pid_file)
                 os.remove(pid_file)
             except:
                 self.log.warn("Error removing the pid file: %s" % pid_file)
 
     def get_pid_from_file(self):
         """Get the pid from the pid file.
 
         If the  pid file doesn't exist a :exc:`PIDFileError` is raised.
         """
         pid_file = os.path.join(self.profile_dir.pid_dir, self.name + u'.pid')
         if os.path.isfile(pid_file):
             with open(pid_file, 'r') as f:
                 s = f.read().strip()
                 try:
                     pid = int(s)
                 except:
                     raise PIDFileError("invalid pid file: %s (contents: %r)"%(pid_file, s))
                 return pid
         else:
             raise PIDFileError('pid file not found: %s' % pid_file)
     
     def check_pid(self, pid):
         if os.name == 'nt':
             try:
                 import ctypes
                 # returns 0 if no such process (of ours) exists
                 # positive int otherwise
                 p = ctypes.windll.kernel32.OpenProcess(1,0,pid)
             except Exception:
                 self.log.warn(
                     "Could not determine whether pid %i is running via `OpenProcess`. "
                     " Making the likely assumption that it is."%pid
                 )
                 return True
             return bool(p)
         else:
             try:
                 p = Popen(['ps','x'], stdout=PIPE, stderr=PIPE)
                 output,_ = p.communicate()
             except OSError:
                 self.log.warn(
                     "Could not determine whether pid %i is running via `ps x`. "
                     " Making the likely assumption that it is."%pid
                 )
                 return True
             pids = map(int, re.findall(r'^\W*\d+', output, re.MULTILINE))
             return pid in pids

IPython/parallel/apps/ipcontrollerapp.py

@@ -1,493 +1,491 b''
 #!/usr/bin/env python
 # encoding: utf-8
 """
 The IPython controller application.
 
 Authors:
 
 * Brian Granger
 * MinRK
 
 """
 
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2008-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 from __future__ import with_statement
 
 import json
 import os
 import socket
 import stat
 import sys
 
 from multiprocessing import Process
 from signal import signal, SIGINT, SIGABRT, SIGTERM
 
 import zmq
 from zmq.devices import ProcessMonitoredQueue
 from zmq.log.handlers import PUBHandler
 
 from IPython.core.profiledir import ProfileDir
 
 from IPython.parallel.apps.baseapp import (
     BaseParallelApplication,
     base_aliases,
     base_flags,
     catch_config_error,
 )
 from IPython.utils.importstring import import_item
 from IPython.utils.traitlets import Instance, Unicode, Bool, List, Dict, TraitError
 
 from IPython.zmq.session import (
     Session, session_aliases, session_flags, default_secure
 )
 
 from IPython.parallel.controller.heartmonitor import HeartMonitor
 from IPython.parallel.controller.hub import HubFactory
 from IPython.parallel.controller.scheduler import TaskScheduler,launch_scheduler
 from IPython.parallel.controller.sqlitedb import SQLiteDB
 
 from IPython.parallel.util import split_url, disambiguate_url
 
 # conditional import of MongoDB backend class
 
 try:
     from IPython.parallel.controller.mongodb import MongoDB
 except ImportError:
     maybe_mongo = []
 else:
     maybe_mongo = [MongoDB]
 
 
 #-----------------------------------------------------------------------------
 # Module level variables
 #-----------------------------------------------------------------------------
 
 
 #: The default config file name for this application
 default_config_file_name = u'ipcontroller_config.py'
 
 
 _description = """Start the IPython controller for parallel computing.
 
 The IPython controller provides a gateway between the IPython engines and
 clients. The controller needs to be started before the engines and can be
 configured using command line options or using a cluster directory. Cluster
 directories contain config, log and security files and are usually located in
 your ipython directory and named as "profile_name". See the `profile`
 and `profile-dir` options for details.
 """
 
 _examples = """
 ipcontroller --ip=192.168.0.1 --port=1000  # listen on ip, port for engines
 ipcontroller --scheme=pure  # use the pure zeromq scheduler
 """
 
 
 #-----------------------------------------------------------------------------
 # The main application
 #-----------------------------------------------------------------------------
 flags = {}
 flags.update(base_flags)
 flags.update({
     'usethreads' : ( {'IPControllerApp' : {'use_threads' : True}},
                     'Use threads instead of processes for the schedulers'),
     'sqlitedb' : ({'HubFactory' : {'db_class' : 'IPython.parallel.controller.sqlitedb.SQLiteDB'}},
                     'use the SQLiteDB backend'),
     'mongodb' : ({'HubFactory' : {'db_class' : 'IPython.parallel.controller.mongodb.MongoDB'}},
                     'use the MongoDB backend'),
     'dictdb' : ({'HubFactory' : {'db_class' : 'IPython.parallel.controller.dictdb.DictDB'}},
                     'use the in-memory DictDB backend'),
     'nodb' : ({'HubFactory' : {'db_class' : 'IPython.parallel.controller.dictdb.NoDB'}},
                     """use dummy DB backend, which doesn't store any information.
                     
                     This can be used to prevent growth of the memory footprint of the Hub
                     in cases where its record-keeping is not required.  Requesting results
                     of tasks submitted by other clients, db_queries, and task resubmission
                     will not be available."""),
     'reuse' : ({'IPControllerApp' : {'reuse_files' : True}},
                     'reuse existing json connection files')
 })
 
 flags.update(session_flags)
 
 aliases = dict(
     ssh = 'IPControllerApp.ssh_server',
     enginessh = 'IPControllerApp.engine_ssh_server',
     location = 'IPControllerApp.location',
 
     url = 'HubFactory.url',
     ip = 'HubFactory.ip',
     transport = 'HubFactory.transport',
     port = 'HubFactory.regport',
 
     ping = 'HeartMonitor.period',
 
     scheme = 'TaskScheduler.scheme_name',
     hwm = 'TaskScheduler.hwm',
 )
 aliases.update(base_aliases)
 aliases.update(session_aliases)
 
 
 class IPControllerApp(BaseParallelApplication):
 
     name = u'ipcontroller'
     description = _description
     examples = _examples
     config_file_name = Unicode(default_config_file_name)
     classes = [ProfileDir, Session, HubFactory, TaskScheduler, HeartMonitor, SQLiteDB] + maybe_mongo
     
     # change default to True
     auto_create = Bool(True, config=True,
         help="""Whether to create profile dir if it doesn't exist.""")
     
     reuse_files = Bool(False, config=True,
         help="""Whether to reuse existing json connection files.
         If False, connection files will be removed on a clean exit.
         """
     )
     ssh_server = Unicode(u'', config=True,
         help="""ssh url for clients to use when connecting to the Controller
         processes. It should be of the form: [user@]server[:port]. The
         Controller's listening addresses must be accessible from the ssh server""",
     )
     engine_ssh_server = Unicode(u'', config=True,
         help="""ssh url for engines to use when connecting to the Controller
         processes. It should be of the form: [user@]server[:port]. The
         Controller's listening addresses must be accessible from the ssh server""",
     )
     location = Unicode(u'', config=True,
         help="""The external IP or domain name of the Controller, used for disambiguating
         engine and client connections.""",
     )
     import_statements = List([], config=True,
         help="import statements to be run at startup.  Necessary in some environments"
     )
 
     use_threads = Bool(False, config=True,
         help='Use threads instead of processes for the schedulers',
     )
 
     engine_json_file = Unicode('ipcontroller-engine.json', config=True,
         help="JSON filename where engine connection info will be stored.")
     client_json_file = Unicode('ipcontroller-client.json', config=True,
         help="JSON filename where client connection info will be stored.")
     
     def _cluster_id_changed(self, name, old, new):
         super(IPControllerApp, self)._cluster_id_changed(name, old, new)
         self.engine_json_file = "%s-engine.json" % self.name
         self.client_json_file = "%s-client.json" % self.name
 
 
     # internal
     children = List()
     mq_class = Unicode('zmq.devices.ProcessMonitoredQueue')
 
     def _use_threads_changed(self, name, old, new):
         self.mq_class = 'zmq.devices.%sMonitoredQueue'%('Thread' if new else 'Process')
     
     write_connection_files = Bool(True,
         help="""Whether to write connection files to disk.
         True in all cases other than runs with `reuse_files=True` *after the first*
         """
     )
 
     aliases = Dict(aliases)
     flags = Dict(flags)
     
 
     def save_connection_dict(self, fname, cdict):
         """save a connection dict to json file."""
         c = self.config
         url = cdict['url']
         location = cdict['location']
         if not location:
             try:
                 proto,ip,port = split_url(url)
             except AssertionError:
                 pass
             else:
                 try:
                     location = socket.gethostbyname_ex(socket.gethostname())[2][-1]
                 except (socket.gaierror, IndexError):
                     self.log.warn("Could not identify this machine's IP, assuming 127.0.0.1."
                     " You may need to specify '--location=<external_ip_address>' to help"
                     " IPython decide when to connect via loopback.")
                     location = '127.0.0.1'
             cdict['location'] = location
         fname = os.path.join(self.profile_dir.security_dir, fname)
         self.log.info("writing connection info to %s", fname)
         with open(fname, 'w') as f:
             f.write(json.dumps(cdict, indent=2))
         os.chmod(fname, stat.S_IRUSR|stat.S_IWUSR)
     
     def load_config_from_json(self):
         """load config from existing json connector files."""
         c = self.config
         self.log.debug("loading config from JSON")
         # load from engine config
         fname = os.path.join(self.profile_dir.security_dir, self.engine_json_file)
         self.log.info("loading connection info from %s", fname)
         with open(fname) as f:
             cfg = json.loads(f.read())
         key = cfg['exec_key']
         # json gives unicode, Session.key wants bytes
         c.Session.key = key.encode('ascii')
         xport,addr = cfg['url'].split('://')
         c.HubFactory.engine_transport = xport
         ip,ports = addr.split(':')
         c.HubFactory.engine_ip = ip
         c.HubFactory.regport = int(ports)
         self.location = cfg['location']
         if not self.engine_ssh_server:
             self.engine_ssh_server = cfg['ssh']
         # load client config
         fname = os.path.join(self.profile_dir.security_dir, self.client_json_file)
         self.log.info("loading connection info from %s", fname)
         with open(fname) as f:
             cfg = json.loads(f.read())
         assert key == cfg['exec_key'], "exec_key mismatch between engine and client keys"
         xport,addr = cfg['url'].split('://')
         c.HubFactory.client_transport = xport
         ip,ports = addr.split(':')
         c.HubFactory.client_ip = ip
         if not self.ssh_server:
             self.ssh_server = cfg['ssh']
         assert int(ports) == c.HubFactory.regport, "regport mismatch"
     
     def cleanup_connection_files(self):
         if self.reuse_files:
             self.log.debug("leaving JSON connection files for reuse")
             return
         self.log.debug("cleaning up JSON connection files")
         for f in (self.client_json_file, self.engine_json_file):
             f = os.path.join(self.profile_dir.security_dir, f)
             try:
                 os.remove(f)
             except Exception as e:
                 self.log.error("Failed to cleanup connection file: %s", e)
             else:
                 self.log.debug(u"removed %s", f)
     
     def load_secondary_config(self):
         """secondary config, loading from JSON and setting defaults"""
         if self.reuse_files:
             try:
                 self.load_config_from_json()
             except (AssertionError,IOError) as e:
                 self.log.error("Could not load config from JSON: %s" % e)
             else:
                 # successfully loaded config from JSON, and reuse=True
                 # no need to wite back the same file
                 self.write_connection_files = False
                 
         # switch Session.key default to secure
         default_secure(self.config)
         self.log.debug("Config changed")
         self.log.debug(repr(self.config))
         
     def init_hub(self):
         c = self.config
         
         self.do_import_statements()
         
         try:
             self.factory = HubFactory(config=c, log=self.log)
             # self.start_logging()
             self.factory.init_hub()
         except TraitError:
             raise
         except Exception:
             self.log.error("Couldn't construct the Controller", exc_info=True)
             self.exit(1)
         
         if self.write_connection_files:
             # save to new json config files
             f = self.factory
             cdict = {'exec_key' : f.session.key.decode('ascii'),
                     'ssh' : self.ssh_server,
                     'url' : "%s://%s:%s"%(f.client_transport, f.client_ip, f.regport),
                     'location' : self.location
                     }
             self.save_connection_dict(self.client_json_file, cdict)
             edict = cdict
             edict['url']="%s://%s:%s"%((f.client_transport, f.client_ip, f.regport))
             edict['ssh'] = self.engine_ssh_server
             self.save_connection_dict(self.engine_json_file, edict)
 
     def init_schedulers(self):
         children = self.children
         mq = import_item(str(self.mq_class))
         
         hub = self.factory
         # disambiguate url, in case of *
         monitor_url = disambiguate_url(hub.monitor_url)
         # maybe_inproc = 'inproc://monitor' if self.use_threads else monitor_url
         # IOPub relay (in a Process)
         q = mq(zmq.PUB, zmq.SUB, zmq.PUB, b'N/A',b'iopub')
         q.bind_in(hub.client_info['iopub'])
         q.bind_out(hub.engine_info['iopub'])
         q.setsockopt_out(zmq.SUBSCRIBE, b'')
         q.connect_mon(monitor_url)
         q.daemon=True
         children.append(q)
 
         # Multiplexer Queue (in a Process)
         q = mq(zmq.ROUTER, zmq.ROUTER, zmq.PUB, b'in', b'out')
         q.bind_in(hub.client_info['mux'])
         q.setsockopt_in(zmq.IDENTITY, b'mux')
         q.bind_out(hub.engine_info['mux'])
         q.connect_mon(monitor_url)
         q.daemon=True
         children.append(q)
 
         # Control Queue (in a Process)
         q = mq(zmq.ROUTER, zmq.ROUTER, zmq.PUB, b'incontrol', b'outcontrol')
         q.bind_in(hub.client_info['control'])
         q.setsockopt_in(zmq.IDENTITY, b'control')
         q.bind_out(hub.engine_info['control'])
         q.connect_mon(monitor_url)
         q.daemon=True
         children.append(q)
         try:
             scheme = self.config.TaskScheduler.scheme_name
         except AttributeError:
             scheme = TaskScheduler.scheme_name.get_default_value()
         # Task Queue (in a Process)
         if scheme == 'pure':
             self.log.warn("task::using pure XREQ Task scheduler")
             q = mq(zmq.ROUTER, zmq.DEALER, zmq.PUB, b'intask', b'outtask')
             # q.setsockopt_out(zmq.HWM, hub.hwm)
             q.bind_in(hub.client_info['task'][1])
             q.setsockopt_in(zmq.IDENTITY, b'task')
             q.bind_out(hub.engine_info['task'])
             q.connect_mon(monitor_url)
             q.daemon=True
             children.append(q)
         elif scheme == 'none':
             self.log.warn("task::using no Task scheduler")
 
         else:
             self.log.info("task::using Python %s Task scheduler"%scheme)
             sargs = (hub.client_info['task'][1], hub.engine_info['task'],
                                 monitor_url, disambiguate_url(hub.client_info['notification']))
             kwargs = dict(logname='scheduler', loglevel=self.log_level,
                             log_url = self.log_url, config=dict(self.config))
             if 'Process' in self.mq_class:
                 # run the Python scheduler in a Process
                 q = Process(target=launch_scheduler, args=sargs, kwargs=kwargs)
                 q.daemon=True
                 children.append(q)
             else:
                 # single-threaded Controller
                 kwargs['in_thread'] = True
                 launch_scheduler(*sargs, **kwargs)
 
     def terminate_children(self):
         child_procs = []
         for child in self.children:
             if isinstance(child, ProcessMonitoredQueue):
                 child_procs.append(child.launcher)
             elif isinstance(child, Process):
                 child_procs.append(child)
         if child_procs:
             self.log.critical("terminating children...")
             for child in child_procs:
                 try:
                     child.terminate()
                 except OSError:
                     # already dead
                     pass
     
     def handle_signal(self, sig, frame):
         self.log.critical("Received signal %i, shutting down", sig)
         self.terminate_children()
         self.loop.stop()
     
     def init_signal(self):
         for sig in (SIGINT, SIGABRT, SIGTERM):
             signal(sig, self.handle_signal)
 
     def do_import_statements(self):
         statements = self.import_statements
         for s in statements:
             try:
                 self.log.msg("Executing statement: '%s'" % s)
                 exec s in globals(), locals()
             except:
                 self.log.msg("Error running statement: %s" % s)
 
     def forward_logging(self):
         if self.log_url:
             self.log.info("Forwarding logging to %s"%self.log_url)
             context = zmq.Context.instance()
             lsock = context.socket(zmq.PUB)
             lsock.connect(self.log_url)
             handler = PUBHandler(lsock)
-            self.log.removeHandler(self._log_handler)
             handler.root_topic = 'controller'
             handler.setLevel(self.log_level)
             self.log.addHandler(handler)
-            self._log_handler = handler
     
     @catch_config_error
     def initialize(self, argv=None):
         super(IPControllerApp, self).initialize(argv)
         self.forward_logging()
         self.load_secondary_config()
         self.init_hub()
         self.init_schedulers()
     
     def start(self):
         # Start the subprocesses:
         self.factory.start()
         # children must be started before signals are setup,
         # otherwise signal-handling will fire multiple times
         for child in self.children:
             child.start()
         self.init_signal()
 
         self.write_pid_file(overwrite=True)
 
         try:
             self.factory.loop.start()
         except KeyboardInterrupt:
             self.log.critical("Interrupted, Exiting...\n")
         finally:
             self.cleanup_connection_files()
             
 
 
 def launch_new_instance():
     """Create and run the IPython controller"""
     if sys.platform == 'win32':
         # make sure we don't get called from a multiprocessing subprocess
         # this can result in infinite Controllers being started on Windows
         # which doesn't have a proper fork, so multiprocessing is wonky
         
         # this only comes up when IPython has been installed using vanilla
         # setuptools, and *not* distribute.
         import multiprocessing
         p = multiprocessing.current_process()
         # the main process has name 'MainProcess'
         # subprocesses will have names like 'Process-1'
         if p.name != 'MainProcess':
             # we are a subprocess, don't start another Controller!
             return
     app = IPControllerApp.instance()
     app.initialize()
     app.start()
 
 
 if __name__ == '__main__':
     launch_new_instance()

IPython/parallel/apps/ipengineapp.py

@@ -1,330 +1,377 b''
 #!/usr/bin/env python
 # encoding: utf-8
 """
 The IPython engine application
 
 Authors:
 
 * Brian Granger
 * MinRK
 
 """
 
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2008-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 import json
 import os
 import sys
 import time
 
 import zmq
 from zmq.eventloop import ioloop
 
 from IPython.core.profiledir import ProfileDir
 from IPython.parallel.apps.baseapp import (
     BaseParallelApplication,
     base_aliases,
     base_flags,
     catch_config_error,
 )
 from IPython.zmq.log import EnginePUBHandler
+from IPython.zmq.ipkernel import Kernel, IPKernelApp
 from IPython.zmq.session import (
     Session, session_aliases, session_flags
 )
 
 from IPython.config.configurable import Configurable
 
 from IPython.parallel.engine.engine import EngineFactory
-from IPython.parallel.engine.streamkernel import Kernel
-from IPython.parallel.util import disambiguate_url, asbytes
+from IPython.parallel.util import disambiguate_url
 
 from IPython.utils.importstring import import_item
-from IPython.utils.traitlets import Bool, Unicode, Dict, List, Float
+from IPython.utils.py3compat import cast_bytes
+from IPython.utils.traitlets import Bool, Unicode, Dict, List, Float, Instance
 
 
 #-----------------------------------------------------------------------------
 # Module level variables
 #-----------------------------------------------------------------------------
 
 #: The default config file name for this application
 default_config_file_name = u'ipengine_config.py'
 
 _description = """Start an IPython engine for parallel computing.
 
 IPython engines run in parallel and perform computations on behalf of a client
 and controller. A controller needs to be started before the engines. The
 engine can be configured using command line options or using a cluster
 directory. Cluster directories contain config, log and security files and are
 usually located in your ipython directory and named as "profile_name".
 See the `profile` and `profile-dir` options for details.
 """
 
 _examples = """
 ipengine --ip=192.168.0.1 --port=1000     # connect to hub at ip and port
 ipengine --log-to-file --log-level=DEBUG  # log to a file with DEBUG verbosity
 """
 
 #-----------------------------------------------------------------------------
 # MPI configuration
 #-----------------------------------------------------------------------------
 
 mpi4py_init = """from mpi4py import MPI as mpi
 mpi.size = mpi.COMM_WORLD.Get_size()
 mpi.rank = mpi.COMM_WORLD.Get_rank()
 """
 
 
 pytrilinos_init = """from PyTrilinos import Epetra
 class SimpleStruct:
 pass
 mpi = SimpleStruct()
 mpi.rank = 0
 mpi.size = 0
 """
 
 class MPI(Configurable):
     """Configurable for MPI initialization"""
     use = Unicode('', config=True,
         help='How to enable MPI (mpi4py, pytrilinos, or empty string to disable).'
         )
 
     def _use_changed(self, name, old, new):
         # load default init script if it's not set
         if not self.init_script:
             self.init_script = self.default_inits.get(new, '')
 
     init_script = Unicode('', config=True,
         help="Initialization code for MPI")
 
     default_inits = Dict({'mpi4py' : mpi4py_init, 'pytrilinos':pytrilinos_init},
         config=True)
 
 
 #-----------------------------------------------------------------------------
 # Main application
 #-----------------------------------------------------------------------------
 aliases = dict(
     file = 'IPEngineApp.url_file',
     c = 'IPEngineApp.startup_command',
     s = 'IPEngineApp.startup_script',
 
     url = 'EngineFactory.url',
     ssh = 'EngineFactory.sshserver',
     sshkey = 'EngineFactory.sshkey',
     ip = 'EngineFactory.ip',
     transport = 'EngineFactory.transport',
     port = 'EngineFactory.regport',
     location = 'EngineFactory.location',
 
     timeout = 'EngineFactory.timeout',
 
     mpi = 'MPI.use',
 
 )
 aliases.update(base_aliases)
 aliases.update(session_aliases)
 flags = {}
 flags.update(base_flags)
 flags.update(session_flags)
 
 class IPEngineApp(BaseParallelApplication):
 
     name = 'ipengine'
     description = _description
     examples = _examples
     config_file_name = Unicode(default_config_file_name)
     classes = List([ProfileDir, Session, EngineFactory, Kernel, MPI])
 
     startup_script = Unicode(u'', config=True,
         help='specify a script to be run at startup')
     startup_command = Unicode('', config=True,
             help='specify a command to be run at startup')
 
     url_file = Unicode(u'', config=True,
         help="""The full location of the file containing the connection information for
         the controller. If this is not given, the file must be in the
         security directory of the cluster directory.  This location is
         resolved using the `profile` or `profile_dir` options.""",
         )
     wait_for_url_file = Float(5, config=True,
         help="""The maximum number of seconds to wait for url_file to exist.
         This is useful for batch-systems and shared-filesystems where the
         controller and engine are started at the same time and it
         may take a moment for the controller to write the connector files.""")
 
     url_file_name = Unicode(u'ipcontroller-engine.json', config=True)
 
     def _cluster_id_changed(self, name, old, new):
         if new:
             base = 'ipcontroller-%s' % new
         else:
             base = 'ipcontroller'
         self.url_file_name = "%s-engine.json" % base
 
     log_url = Unicode('', config=True,
         help="""The URL for the iploggerapp instance, for forwarding
         logging to a central location.""")
+    
+    # an IPKernelApp instance, used to setup listening for shell frontends
+    kernel_app = Instance(IPKernelApp)
 
     aliases = Dict(aliases)
     flags = Dict(flags)
+    
+    @property
+    def kernel(self):
+        """allow access to the Kernel object, so I look like IPKernelApp"""
+        return self.engine.kernel
 
     def find_url_file(self):
         """Set the url file.
 
         Here we don't try to actually see if it exists for is valid as that
         is hadled by the connection logic.
         """
         config = self.config
         # Find the actual controller key file
         if not self.url_file:
             self.url_file = os.path.join(
                 self.profile_dir.security_dir,
                 self.url_file_name
             )
     
     def load_connector_file(self):
         """load config from a JSON connector file,
         at a *lower* priority than command-line/config files.
         """
         
         self.log.info("Loading url_file %r", self.url_file)
         config = self.config
         
         with open(self.url_file) as f:
             d = json.loads(f.read())
         
         if 'exec_key' in d:
-            config.Session.key = asbytes(d['exec_key'])
+            config.Session.key = cast_bytes(d['exec_key'])
         
         try:
             config.EngineFactory.location
         except AttributeError:
             config.EngineFactory.location = d['location']
         
         d['url'] = disambiguate_url(d['url'], config.EngineFactory.location)
         try:
             config.EngineFactory.url
         except AttributeError:
             config.EngineFactory.url = d['url']
         
         try:
             config.EngineFactory.sshserver
         except AttributeError:
             config.EngineFactory.sshserver = d['ssh']
+    
+    def bind_kernel(self, **kwargs):
+        """Promote engine to listening kernel, accessible to frontends."""
+        if self.kernel_app is not None:
+            return
+        
+        self.log.info("Opening ports for direct connections as an IPython kernel")
+        
+        kernel = self.kernel
+        
+        kwargs.setdefault('config', self.config)
+        kwargs.setdefault('log', self.log)
+        kwargs.setdefault('profile_dir', self.profile_dir)
+        kwargs.setdefault('session', self.engine.session)
+        
+        app = self.kernel_app = IPKernelApp(**kwargs)
+        
+        # allow IPKernelApp.instance():
+        IPKernelApp._instance = app
         
+        app.init_connection_file()
+        # relevant contents of init_sockets:
+        
+        app.shell_port = app._bind_socket(kernel.shell_streams[0], app.shell_port)
+        app.log.debug("shell ROUTER Channel on port: %i", app.shell_port)
+        
+        app.iopub_port = app._bind_socket(kernel.iopub_socket, app.iopub_port)
+        app.log.debug("iopub PUB Channel on port: %i", app.iopub_port)
+        
+        kernel.stdin_socket = self.engine.context.socket(zmq.ROUTER)
+        app.stdin_port = app._bind_socket(kernel.stdin_socket, app.stdin_port)
+        app.log.debug("stdin ROUTER Channel on port: %i", app.stdin_port)
+        
+        # start the heartbeat, and log connection info:
+        
+        app.init_heartbeat()
+        
+        app.log_connection_info()
+        app.write_connection_file()
+        
+    
     def init_engine(self):
         # This is the working dir by now.
         sys.path.insert(0, '')
         config = self.config
         # print config
         self.find_url_file()
         
         # was the url manually specified?
         keys = set(self.config.EngineFactory.keys())
         keys = keys.union(set(self.config.RegistrationFactory.keys()))
         
         if keys.intersection(set(['ip', 'url', 'port'])):
             # Connection info was specified, don't wait for the file
             url_specified = True
             self.wait_for_url_file = 0
         else:
             url_specified = False
 
         if self.wait_for_url_file and not os.path.exists(self.url_file):
             self.log.warn("url_file %r not found", self.url_file)
             self.log.warn("Waiting up to %.1f seconds for it to arrive.", self.wait_for_url_file)
             tic = time.time()
             while not os.path.exists(self.url_file) and (time.time()-tic < self.wait_for_url_file):
                 # wait for url_file to exist, or until time limit
                 time.sleep(0.1)
             
         if os.path.exists(self.url_file):
             self.load_connector_file()
         elif not url_specified:
             self.log.fatal("Fatal: url file never arrived: %s", self.url_file)
             self.exit(1)
         
         
         try:
             exec_lines = config.Kernel.exec_lines
         except AttributeError:
             config.Kernel.exec_lines = []
             exec_lines = config.Kernel.exec_lines
         
         if self.startup_script:
             enc = sys.getfilesystemencoding() or 'utf8'
             cmd="execfile(%r)" % self.startup_script.encode(enc)
             exec_lines.append(cmd)
         if self.startup_command:
             exec_lines.append(self.startup_command)
 
         # Create the underlying shell class and Engine
         # shell_class = import_item(self.master_config.Global.shell_class)
         # print self.config
         try:
             self.engine = EngineFactory(config=config, log=self.log)
         except:
             self.log.error("Couldn't start the Engine", exc_info=True)
             self.exit(1)
     
     def forward_logging(self):
         if self.log_url:
             self.log.info("Forwarding logging to %s", self.log_url)
             context = self.engine.context
             lsock = context.socket(zmq.PUB)
             lsock.connect(self.log_url)
-            self.log.removeHandler(self._log_handler)
             handler = EnginePUBHandler(self.engine, lsock)
             handler.setLevel(self.log_level)
             self.log.addHandler(handler)
-            self._log_handler = handler
     
     def init_mpi(self):
         global mpi
         self.mpi = MPI(config=self.config)
 
         mpi_import_statement = self.mpi.init_script
         if mpi_import_statement:
             try:
                 self.log.info("Initializing MPI:")
                 self.log.info(mpi_import_statement)
                 exec mpi_import_statement in globals()
             except:
                 mpi = None
         else:
             mpi = None
 
     @catch_config_error
     def initialize(self, argv=None):
         super(IPEngineApp, self).initialize(argv)
         self.init_mpi()
         self.init_engine()
         self.forward_logging()
     
     def start(self):
         self.engine.start()
         try:
             self.engine.loop.start()
         except KeyboardInterrupt:
             self.log.critical("Engine Interrupted, shutting down...\n")
 
 
 def launch_new_instance():
     """Create and run the IPython engine"""
     app = IPEngineApp.instance()
     app.initialize()
     app.start()
 
 
 if __name__ == '__main__':
     launch_new_instance()
 

IPython/parallel/apps/launcher.py

@@ -1,1345 +1,1341 b''
 # encoding: utf-8
 """
 Facilities for launching IPython processes asynchronously.
 
 Authors:
 
 * Brian Granger
 * MinRK
 """
 
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2008-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 import copy
 import logging
 import os
 import stat
+import sys
 import time
 
 # signal imports, handling various platforms, versions
 
 from signal import SIGINT, SIGTERM
 try:
     from signal import SIGKILL
 except ImportError:
     # Windows
     SIGKILL=SIGTERM
 
 try:
     # Windows >= 2.7, 3.2
     from signal import CTRL_C_EVENT as SIGINT
 except ImportError:
     pass
 
 from subprocess import Popen, PIPE, STDOUT
 try:
     from subprocess import check_output
 except ImportError:
     # pre-2.7, define check_output with Popen
     def check_output(*args, **kwargs):
         kwargs.update(dict(stdout=PIPE))
         p = Popen(*args, **kwargs)
         out,err = p.communicate()
         return out
 
 from zmq.eventloop import ioloop
 
 from IPython.config.application import Application
 from IPython.config.configurable import LoggingConfigurable
 from IPython.utils.text import EvalFormatter
 from IPython.utils.traitlets import (
     Any, Integer, CFloat, List, Unicode, Dict, Instance, HasTraits, CRegExp
 )
-from IPython.utils.path import get_ipython_module_path, get_home_dir
-from IPython.utils.process import find_cmd, pycmd2argv, FindCmdError
+from IPython.utils.path import get_home_dir
+from IPython.utils.process import find_cmd, FindCmdError
 
 from .win32support import forward_read_events
 
 from .winhpcjob import IPControllerTask, IPEngineTask, IPControllerJob, IPEngineSetJob
 
 WINDOWS = os.name == 'nt'
 
 #-----------------------------------------------------------------------------
 # Paths to the kernel apps
 #-----------------------------------------------------------------------------
 
+cmd = "from IPython.parallel.apps.%s import launch_new_instance; launch_new_instance()"
 
-ipcluster_cmd_argv = pycmd2argv(get_ipython_module_path(
-    'IPython.parallel.apps.ipclusterapp'
-))
+ipcluster_cmd_argv = [sys.executable, "-c", cmd % "ipclusterapp"]
 
-ipengine_cmd_argv = pycmd2argv(get_ipython_module_path(
-    'IPython.parallel.apps.ipengineapp'
-))
+ipengine_cmd_argv = [sys.executable, "-c", cmd % "ipengineapp"]
 
-ipcontroller_cmd_argv = pycmd2argv(get_ipython_module_path(
-    'IPython.parallel.apps.ipcontrollerapp'
-))
+ipcontroller_cmd_argv = [sys.executable, "-c", cmd % "ipcontrollerapp"]
 
 #-----------------------------------------------------------------------------
 # Base launchers and errors
 #-----------------------------------------------------------------------------
 
 
 class LauncherError(Exception):
     pass
 
 
 class ProcessStateError(LauncherError):
     pass
 
 
 class UnknownStatus(LauncherError):
     pass
 
 
 class BaseLauncher(LoggingConfigurable):
     """An asbtraction for starting, stopping and signaling a process."""
 
     # In all of the launchers, the work_dir is where child processes will be
     # run. This will usually be the profile_dir, but may not be. any work_dir
     # passed into the __init__ method will override the config value.
     # This should not be used to set the work_dir for the actual engine
     # and controller. Instead, use their own config files or the
     # controller_args, engine_args attributes of the launchers to add
     # the work_dir option.
     work_dir = Unicode(u'.')
     loop = Instance('zmq.eventloop.ioloop.IOLoop')
 
     start_data = Any()
     stop_data = Any()
 
     def _loop_default(self):
         return ioloop.IOLoop.instance()
 
     def __init__(self, work_dir=u'.', config=None, **kwargs):
         super(BaseLauncher, self).__init__(work_dir=work_dir, config=config, **kwargs)
         self.state = 'before' # can be before, running, after
         self.stop_callbacks = []
         self.start_data = None
         self.stop_data = None
 
     @property
     def args(self):
         """A list of cmd and args that will be used to start the process.
 
         This is what is passed to :func:`spawnProcess` and the first element
         will be the process name.
         """
         return self.find_args()
 
     def find_args(self):
         """The ``.args`` property calls this to find the args list.
 
         Subcommand should implement this to construct the cmd and args.
         """
         raise NotImplementedError('find_args must be implemented in a subclass')
 
     @property
     def arg_str(self):
         """The string form of the program arguments."""
         return ' '.join(self.args)
 
     @property
     def running(self):
         """Am I running."""
         if self.state == 'running':
             return True
         else:
             return False
 
     def start(self):
         """Start the process."""
         raise NotImplementedError('start must be implemented in a subclass')
 
     def stop(self):
         """Stop the process and notify observers of stopping.
 
         This method will return None immediately.
         To observe the actual process stopping, see :meth:`on_stop`.
         """
         raise NotImplementedError('stop must be implemented in a subclass')
 
     def on_stop(self, f):
         """Register a callback to be called with this Launcher's stop_data
         when the process actually finishes.
         """
         if self.state=='after':
             return f(self.stop_data)
         else:
             self.stop_callbacks.append(f)
 
     def notify_start(self, data):
         """Call this to trigger startup actions.
 
         This logs the process startup and sets the state to 'running'.  It is
         a pass-through so it can be used as a callback.
         """
 
         self.log.debug('Process %r started: %r', self.args[0], data)
         self.start_data = data
         self.state = 'running'
         return data
 
     def notify_stop(self, data):
         """Call this to trigger process stop actions.
 
         This logs the process stopping and sets the state to 'after'. Call
         this to trigger callbacks registered via :meth:`on_stop`."""
 
         self.log.debug('Process %r stopped: %r', self.args[0], data)
         self.stop_data = data
         self.state = 'after'
         for i in range(len(self.stop_callbacks)):
             d = self.stop_callbacks.pop()
             d(data)
         return data
 
     def signal(self, sig):
         """Signal the process.
 
         Parameters
         ----------
         sig : str or int
             'KILL', 'INT', etc., or any signal number
         """
         raise NotImplementedError('signal must be implemented in a subclass')
 
 class ClusterAppMixin(HasTraits):
     """MixIn for cluster args as traits"""
     profile_dir=Unicode('')
     cluster_id=Unicode('')
 
     @property
     def cluster_args(self):
         return ['--profile-dir', self.profile_dir, '--cluster-id', self.cluster_id]
 
 class ControllerMixin(ClusterAppMixin):
     controller_cmd = List(ipcontroller_cmd_argv, config=True,
         help="""Popen command to launch ipcontroller.""")
     # Command line arguments to ipcontroller.
     controller_args = List(['--log-to-file','--log-level=%i' % logging.INFO], config=True,
         help="""command-line args to pass to ipcontroller""")
 
 class EngineMixin(ClusterAppMixin):
     engine_cmd = List(ipengine_cmd_argv, config=True,
         help="""command to launch the Engine.""")
     # Command line arguments for ipengine.
     engine_args = List(['--log-to-file','--log-level=%i' % logging.INFO], config=True,
         help="command-line arguments to pass to ipengine"
     )
 
 
 #-----------------------------------------------------------------------------
 # Local process launchers
 #-----------------------------------------------------------------------------
 
 
 class LocalProcessLauncher(BaseLauncher):
     """Start and stop an external process in an asynchronous manner.
 
     This will launch the external process with a working directory of
     ``self.work_dir``.
     """
 
     # This is used to to construct self.args, which is passed to
     # spawnProcess.
     cmd_and_args = List([])
     poll_frequency = Integer(100) # in ms
 
     def __init__(self, work_dir=u'.', config=None, **kwargs):
         super(LocalProcessLauncher, self).__init__(
             work_dir=work_dir, config=config, **kwargs
         )
         self.process = None
         self.poller = None
 
     def find_args(self):
         return self.cmd_and_args
 
     def start(self):
         self.log.debug("Starting %s: %r", self.__class__.__name__, self.args)
         if self.state == 'before':
             self.process = Popen(self.args,
                 stdout=PIPE,stderr=PIPE,stdin=PIPE,
                 env=os.environ,
                 cwd=self.work_dir
             )
             if WINDOWS:
                 self.stdout = forward_read_events(self.process.stdout)
                 self.stderr = forward_read_events(self.process.stderr)
             else:
                 self.stdout = self.process.stdout.fileno()
                 self.stderr = self.process.stderr.fileno()
             self.loop.add_handler(self.stdout, self.handle_stdout, self.loop.READ)
             self.loop.add_handler(self.stderr, self.handle_stderr, self.loop.READ)
             self.poller = ioloop.PeriodicCallback(self.poll, self.poll_frequency, self.loop)
             self.poller.start()
             self.notify_start(self.process.pid)
         else:
             s = 'The process was already started and has state: %r' % self.state
             raise ProcessStateError(s)
 
     def stop(self):
         return self.interrupt_then_kill()
 
     def signal(self, sig):
         if self.state == 'running':
             if WINDOWS and sig != SIGINT:
                 # use Windows tree-kill for better child cleanup
                 check_output(['taskkill', '-pid', str(self.process.pid), '-t', '-f'])
             else:
                 self.process.send_signal(sig)
 
     def interrupt_then_kill(self, delay=2.0):
         """Send INT, wait a delay and then send KILL."""
         try:
             self.signal(SIGINT)
         except Exception:
             self.log.debug("interrupt failed")
             pass
         self.killer  = ioloop.DelayedCallback(lambda : self.signal(SIGKILL), delay*1000, self.loop)
         self.killer.start()
 
     # callbacks, etc:
 
     def handle_stdout(self, fd, events):
         if WINDOWS:
             line = self.stdout.recv()
         else:
             line = self.process.stdout.readline()
         # a stopped process will be readable but return empty strings
         if line:
             self.log.debug(line[:-1])
         else:
             self.poll()
 
     def handle_stderr(self, fd, events):
         if WINDOWS:
             line = self.stderr.recv()
         else:
             line = self.process.stderr.readline()
         # a stopped process will be readable but return empty strings
         if line:
             self.log.debug(line[:-1])
         else:
             self.poll()
 
     def poll(self):
         status = self.process.poll()
         if status is not None:
             self.poller.stop()
             self.loop.remove_handler(self.stdout)
             self.loop.remove_handler(self.stderr)
             self.notify_stop(dict(exit_code=status, pid=self.process.pid))
         return status
 
 class LocalControllerLauncher(LocalProcessLauncher, ControllerMixin):
     """Launch a controller as a regular external process."""
 
     def find_args(self):
         return self.controller_cmd + self.cluster_args + self.controller_args
 
     def start(self):
         """Start the controller by profile_dir."""
         return super(LocalControllerLauncher, self).start()
 
 
 class LocalEngineLauncher(LocalProcessLauncher, EngineMixin):
     """Launch a single engine as a regular externall process."""
 
     def find_args(self):
         return self.engine_cmd + self.cluster_args + self.engine_args
 
 
 class LocalEngineSetLauncher(LocalEngineLauncher):
     """Launch a set of engines as regular external processes."""
 
     delay = CFloat(0.1, config=True,
         help="""delay (in seconds) between starting each engine after the first.
         This can help force the engines to get their ids in order, or limit
         process flood when starting many engines."""
     )
 
     # launcher class
     launcher_class = LocalEngineLauncher
 
     launchers = Dict()
     stop_data = Dict()
 
     def __init__(self, work_dir=u'.', config=None, **kwargs):
         super(LocalEngineSetLauncher, self).__init__(
             work_dir=work_dir, config=config, **kwargs
         )
         self.stop_data = {}
 
     def start(self, n):
         """Start n engines by profile or profile_dir."""
         dlist = []
         for i in range(n):
             if i > 0:
                 time.sleep(self.delay)
             el = self.launcher_class(work_dir=self.work_dir, config=self.config, log=self.log,
                                     profile_dir=self.profile_dir, cluster_id=self.cluster_id,
             )
 
             # Copy the engine args over to each engine launcher.
             el.engine_cmd = copy.deepcopy(self.engine_cmd)
             el.engine_args = copy.deepcopy(self.engine_args)
             el.on_stop(self._notice_engine_stopped)
             d = el.start()
             self.launchers[i] = el
             dlist.append(d)
         self.notify_start(dlist)
         return dlist
 
     def find_args(self):
         return ['engine set']
 
     def signal(self, sig):
         dlist = []
         for el in self.launchers.itervalues():
             d = el.signal(sig)
             dlist.append(d)
         return dlist
 
     def interrupt_then_kill(self, delay=1.0):
         dlist = []
         for el in self.launchers.itervalues():
             d = el.interrupt_then_kill(delay)
             dlist.append(d)
         return dlist
 
     def stop(self):
         return self.interrupt_then_kill()
 
     def _notice_engine_stopped(self, data):
         pid = data['pid']
         for idx,el in self.launchers.iteritems():
             if el.process.pid == pid:
                 break
         self.launchers.pop(idx)
         self.stop_data[idx] = data
         if not self.launchers:
             self.notify_stop(self.stop_data)
 
 
 #-----------------------------------------------------------------------------
 # MPI launchers
 #-----------------------------------------------------------------------------
 
 
 class MPILauncher(LocalProcessLauncher):
     """Launch an external process using mpiexec."""
 
     mpi_cmd = List(['mpiexec'], config=True,
         help="The mpiexec command to use in starting the process."
     )
     mpi_args = List([], config=True,
         help="The command line arguments to pass to mpiexec."
     )
     program = List(['date'],
         help="The program to start via mpiexec.")
     program_args = List([],
         help="The command line argument to the program."
     )
     n = Integer(1)
     
     def __init__(self, *args, **kwargs):
         # deprecation for old MPIExec names:
         config = kwargs.get('config', {})
         for oldname in ('MPIExecLauncher', 'MPIExecControllerLauncher', 'MPIExecEngineSetLauncher'):
             deprecated = config.get(oldname)
             if deprecated:
                 newname = oldname.replace('MPIExec', 'MPI')
                 config[newname].update(deprecated)
                 self.log.warn("WARNING: %s name has been deprecated, use %s", oldname, newname)
 
         super(MPILauncher, self).__init__(*args, **kwargs)
 
     def find_args(self):
         """Build self.args using all the fields."""
         return self.mpi_cmd + ['-n', str(self.n)] + self.mpi_args + \
                self.program + self.program_args
 
     def start(self, n):
         """Start n instances of the program using mpiexec."""
         self.n = n
         return super(MPILauncher, self).start()
 
 
 class MPIControllerLauncher(MPILauncher, ControllerMixin):
     """Launch a controller using mpiexec."""
 
     # alias back to *non-configurable* program[_args] for use in find_args()
     # this way all Controller/EngineSetLaunchers have the same form, rather
     # than *some* having `program_args` and others `controller_args`
     @property
     def program(self):
         return self.controller_cmd
 
     @property
     def program_args(self):
         return self.cluster_args + self.controller_args
 
     def start(self):
         """Start the controller by profile_dir."""
         return super(MPIControllerLauncher, self).start(1)
 
 
 class MPIEngineSetLauncher(MPILauncher, EngineMixin):
     """Launch engines using mpiexec"""
 
     # alias back to *non-configurable* program[_args] for use in find_args()
     # this way all Controller/EngineSetLaunchers have the same form, rather
     # than *some* having `program_args` and others `controller_args`
     @property
     def program(self):
         return self.engine_cmd
 
     @property
     def program_args(self):
         return self.cluster_args + self.engine_args
 
     def start(self, n):
         """Start n engines by profile or profile_dir."""
         self.n = n
         return super(MPIEngineSetLauncher, self).start(n)
 
 # deprecated MPIExec names
 class DeprecatedMPILauncher(object):
     def warn(self):
         oldname = self.__class__.__name__
         newname = oldname.replace('MPIExec', 'MPI')
         self.log.warn("WARNING: %s name is deprecated, use %s", oldname, newname)
     
 class MPIExecLauncher(MPILauncher, DeprecatedMPILauncher):
     """Deprecated, use MPILauncher"""
     def __init__(self, *args, **kwargs):
         super(MPIExecLauncher, self).__init__(*args, **kwargs)
         self.warn()
 
 class MPIExecControllerLauncher(MPIControllerLauncher, DeprecatedMPILauncher):
     """Deprecated, use MPIControllerLauncher"""
     def __init__(self, *args, **kwargs):
         super(MPIExecControllerLauncher, self).__init__(*args, **kwargs)
         self.warn()
 
 class MPIExecEngineSetLauncher(MPIEngineSetLauncher, DeprecatedMPILauncher):
     """Deprecated, use MPIEngineSetLauncher"""
     def __init__(self, *args, **kwargs):
         super(MPIExecEngineSetLauncher, self).__init__(*args, **kwargs)
         self.warn()
 
 
 #-----------------------------------------------------------------------------
 # SSH launchers
 #-----------------------------------------------------------------------------
 
 # TODO: Get SSH Launcher back to level of sshx in 0.10.2
 
 class SSHLauncher(LocalProcessLauncher):
     """A minimal launcher for ssh.
 
     To be useful this will probably have to be extended to use the ``sshx``
     idea for environment variables.  There could be other things this needs
     as well.
     """
 
     ssh_cmd = List(['ssh'], config=True,
         help="command for starting ssh")
     ssh_args = List(['-tt'], config=True,
         help="args to pass to ssh")
     scp_cmd = List(['scp'], config=True,
         help="command for sending files")
     program = List(['date'],
         help="Program to launch via ssh")
     program_args = List([],
         help="args to pass to remote program")
     hostname = Unicode('', config=True,
         help="hostname on which to launch the program")
     user = Unicode('', config=True,
         help="username for ssh")
     location = Unicode('', config=True,
         help="user@hostname location for ssh in one setting")
     to_fetch = List([], config=True,
         help="List of (remote, local) files to fetch after starting")
     to_send = List([], config=True,
         help="List of (local, remote) files to send before starting")
 
     def _hostname_changed(self, name, old, new):
         if self.user:
             self.location = u'%s@%s' % (self.user, new)
         else:
             self.location = new
 
     def _user_changed(self, name, old, new):
         self.location = u'%s@%s' % (new, self.hostname)
 
     def find_args(self):
         return self.ssh_cmd + self.ssh_args + [self.location] + \
                self.program + self.program_args
     
     def _send_file(self, local, remote):
         """send a single file"""
         remote = "%s:%s" % (self.location, remote)
         for i in range(10):
             if not os.path.exists(local):
                 self.log.debug("waiting for %s" % local)
                 time.sleep(1)
             else:
                 break
         self.log.info("sending %s to %s", local, remote)
         check_output(self.scp_cmd + [local, remote])
     
     def send_files(self):
         """send our files (called before start)"""
         if not self.to_send:
             return
         for local_file, remote_file in self.to_send:
             self._send_file(local_file, remote_file)
 
     def _fetch_file(self, remote, local):
         """fetch a single file"""
         full_remote = "%s:%s" % (self.location, remote)
         self.log.info("fetching %s from %s", local, full_remote)
         for i in range(10):
             # wait up to 10s for remote file to exist
             check = check_output(self.ssh_cmd + self.ssh_args + \
                 [self.location, 'test -e', remote, "&& echo 'yes' || echo 'no'"])
             check = check.strip()
             if check == 'no':
                 time.sleep(1)
             elif check == 'yes':
                 break
         check_output(self.scp_cmd + [full_remote, local])
     
     def fetch_files(self):
         """fetch remote files (called after start)"""
         if not self.to_fetch:
             return
         for remote_file, local_file in self.to_fetch:
             self._fetch_file(remote_file, local_file)
 
     def start(self, hostname=None, user=None):
         if hostname is not None:
             self.hostname = hostname
         if user is not None:
             self.user = user
         
         self.send_files()
         super(SSHLauncher, self).start()
         self.fetch_files()
 
     def signal(self, sig):
         if self.state == 'running':
             # send escaped ssh connection-closer
             self.process.stdin.write('~.')
             self.process.stdin.flush()
 
 class SSHClusterLauncher(SSHLauncher):
     
     remote_profile_dir = Unicode('', config=True,
         help="""The remote profile_dir to use.  
         
         If not specified, use calling profile, stripping out possible leading homedir.
         """)
 
     def _remote_profie_dir_default(self):
         """turns /home/you/.ipython/profile_foo into .ipython/profile_foo
         """
         home = get_home_dir()
         if not home.endswith('/'):
             home = home+'/'
         
         if self.profile_dir.startswith(home):
             return self.profile_dir[len(home):]
         else:
             return self.profile_dir
     
     def _cluster_id_changed(self, name, old, new):
         if new:
             raise ValueError("cluster id not supported by SSH launchers")
     
     @property
     def cluster_args(self):
         return ['--profile-dir', self.remote_profile_dir]
 
 class SSHControllerLauncher(SSHClusterLauncher, ControllerMixin):
 
     # alias back to *non-configurable* program[_args] for use in find_args()
     # this way all Controller/EngineSetLaunchers have the same form, rather
     # than *some* having `program_args` and others `controller_args`
 
     def _controller_cmd_default(self):
         return ['ipcontroller']
     
     @property
     def program(self):
         return self.controller_cmd
     
     @property
     def program_args(self):
         return self.cluster_args + self.controller_args
 
     def _to_fetch_default(self):
         return [
             (os.path.join(self.remote_profile_dir, 'security', cf),
              os.path.join(self.profile_dir, 'security', cf),)
             for cf in ('ipcontroller-client.json', 'ipcontroller-engine.json')
         ]
 
 class SSHEngineLauncher(SSHClusterLauncher, EngineMixin):
 
     # alias back to *non-configurable* program[_args] for use in find_args()
     # this way all Controller/EngineSetLaunchers have the same form, rather
     # than *some* having `program_args` and others `controller_args`
     
     def _engine_cmd_default(self):
         return ['ipengine']
     
     @property
     def program(self):
         return self.engine_cmd
 
     @property
     def program_args(self):
         return self.cluster_args + self.engine_args
     
     def _to_send_default(self):
         return [
             (os.path.join(self.profile_dir, 'security', cf), 
              os.path.join(self.remote_profile_dir, 'security', cf))
             for cf in ('ipcontroller-client.json', 'ipcontroller-engine.json')
         ]
 
 
 class SSHEngineSetLauncher(LocalEngineSetLauncher):
     launcher_class = SSHEngineLauncher
     engines = Dict(config=True,
         help="""dict of engines to launch.  This is a dict by hostname of ints,
         corresponding to the number of engines to start on that host.""")
     
     @property
     def engine_count(self):
         """determine engine count from `engines` dict"""
         count = 0
         for n in self.engines.itervalues():
             if isinstance(n, (tuple,list)):
                 n,args = n
             count += n
         return count
     
     def start(self, n):
         """Start engines by profile or profile_dir.
         `n` is ignored, and the `engines` config property is used instead.
         """
 
         dlist = []
         for host, n in self.engines.iteritems():
             if isinstance(n, (tuple, list)):
                 n, args = n
             else:
                 args = copy.deepcopy(self.engine_args)
 
             if '@' in host:
                 user,host = host.split('@',1)
             else:
                 user=None
             for i in range(n):
                 if i > 0:
                     time.sleep(self.delay)
                 el = self.launcher_class(work_dir=self.work_dir, config=self.config, log=self.log,
                                         profile_dir=self.profile_dir, cluster_id=self.cluster_id,
                 )
                 if i > 0:
                     # only send files for the first engine on each host
                     el.to_send = []
 
                 # Copy the engine args over to each engine launcher.
                 el.engine_cmd = self.engine_cmd
                 el.engine_args = args
                 el.on_stop(self._notice_engine_stopped)
                 d = el.start(user=user, hostname=host)
                 self.launchers[ "%s/%i" % (host,i) ] = el
                 dlist.append(d)
         self.notify_start(dlist)
         return dlist
 
 
 class SSHProxyEngineSetLauncher(SSHClusterLauncher):
     """Launcher for calling
     `ipcluster engines` on a remote machine.
     
     Requires that remote profile is already configured.
     """
     
     n = Integer()
     ipcluster_cmd = List(['ipcluster'], config=True)
     
     @property
     def program(self):
         return self.ipcluster_cmd + ['engines']
     
     @property
     def program_args(self):
         return ['-n', str(self.n), '--profile-dir', self.remote_profile_dir]
     
     def _to_send_default(self):
         return [
             (os.path.join(self.profile_dir, 'security', cf), 
              os.path.join(self.remote_profile_dir, 'security', cf))
             for cf in ('ipcontroller-client.json', 'ipcontroller-engine.json')
     ]
     
     def start(self, n):
         self.n = n
         super(SSHProxyEngineSetLauncher, self).start()
 
 
 #-----------------------------------------------------------------------------
 # Windows HPC Server 2008 scheduler launchers
 #-----------------------------------------------------------------------------
 
 
 # This is only used on Windows.
 def find_job_cmd():
     if WINDOWS:
         try:
             return find_cmd('job')
         except (FindCmdError, ImportError):
             # ImportError will be raised if win32api is not installed
             return 'job'
     else:
         return 'job'
 
 
 class WindowsHPCLauncher(BaseLauncher):
 
     job_id_regexp = CRegExp(r'\d+', config=True,
         help="""A regular expression used to get the job id from the output of the
         submit_command. """
         )
     job_file_name = Unicode(u'ipython_job.xml', config=True,
         help="The filename of the instantiated job script.")
     # The full path to the instantiated job script. This gets made dynamically
     # by combining the work_dir with the job_file_name.
     job_file = Unicode(u'')
     scheduler = Unicode('', config=True,
         help="The hostname of the scheduler to submit the job to.")
     job_cmd = Unicode(find_job_cmd(), config=True,
         help="The command for submitting jobs.")
 
     def __init__(self, work_dir=u'.', config=None, **kwargs):
         super(WindowsHPCLauncher, self).__init__(
             work_dir=work_dir, config=config, **kwargs
         )
 
     @property
     def job_file(self):
         return os.path.join(self.work_dir, self.job_file_name)
 
     def write_job_file(self, n):
         raise NotImplementedError("Implement write_job_file in a subclass.")
 
     def find_args(self):
         return [u'job.exe']
 
     def parse_job_id(self, output):
         """Take the output of the submit command and return the job id."""
         m = self.job_id_regexp.search(output)
         if m is not None:
             job_id = m.group()
         else:
             raise LauncherError("Job id couldn't be determined: %s" % output)
         self.job_id = job_id
         self.log.info('Job started with id: %r', job_id)
         return job_id
 
     def start(self, n):
         """Start n copies of the process using the Win HPC job scheduler."""
         self.write_job_file(n)
         args = [
             'submit',
             '/jobfile:%s' % self.job_file,
             '/scheduler:%s' % self.scheduler
         ]
         self.log.debug("Starting Win HPC Job: %s" % (self.job_cmd + ' ' + ' '.join(args),))
 
         output = check_output([self.job_cmd]+args,
             env=os.environ,
             cwd=self.work_dir,
             stderr=STDOUT
         )
         job_id = self.parse_job_id(output)
         self.notify_start(job_id)
         return job_id
 
     def stop(self):
         args = [
             'cancel',
             self.job_id,
             '/scheduler:%s' % self.scheduler
         ]
         self.log.info("Stopping Win HPC Job: %s" % (self.job_cmd + ' ' + ' '.join(args),))
         try:
             output = check_output([self.job_cmd]+args,
                 env=os.environ,
                 cwd=self.work_dir,
                 stderr=STDOUT
             )
         except:
             output = 'The job already appears to be stoppped: %r' % self.job_id
         self.notify_stop(dict(job_id=self.job_id, output=output))  # Pass the output of the kill cmd
         return output
 
 
 class WindowsHPCControllerLauncher(WindowsHPCLauncher, ClusterAppMixin):
 
     job_file_name = Unicode(u'ipcontroller_job.xml', config=True,
         help="WinHPC xml job file.")
     controller_args = List([], config=False,
         help="extra args to pass to ipcontroller")
 
     def write_job_file(self, n):
         job = IPControllerJob(config=self.config)
 
         t = IPControllerTask(config=self.config)
         # The tasks work directory is *not* the actual work directory of
         # the controller. It is used as the base path for the stdout/stderr
         # files that the scheduler redirects to.
         t.work_directory = self.profile_dir
         # Add the profile_dir and from self.start().
         t.controller_args.extend(self.cluster_args)
         t.controller_args.extend(self.controller_args)
         job.add_task(t)
 
         self.log.debug("Writing job description file: %s", self.job_file)
         job.write(self.job_file)
 
     @property
     def job_file(self):
         return os.path.join(self.profile_dir, self.job_file_name)
 
     def start(self):
         """Start the controller by profile_dir."""
         return super(WindowsHPCControllerLauncher, self).start(1)
 
 
 class WindowsHPCEngineSetLauncher(WindowsHPCLauncher, ClusterAppMixin):
 
     job_file_name = Unicode(u'ipengineset_job.xml', config=True,
         help="jobfile for ipengines job")
     engine_args = List([], config=False,
         help="extra args to pas to ipengine")
 
     def write_job_file(self, n):
         job = IPEngineSetJob(config=self.config)
 
         for i in range(n):
             t = IPEngineTask(config=self.config)
             # The tasks work directory is *not* the actual work directory of
             # the engine. It is used as the base path for the stdout/stderr
             # files that the scheduler redirects to.
             t.work_directory = self.profile_dir
             # Add the profile_dir and from self.start().
             t.engine_args.extend(self.cluster_args)
             t.engine_args.extend(self.engine_args)
             job.add_task(t)
 
         self.log.debug("Writing job description file: %s", self.job_file)
         job.write(self.job_file)
 
     @property
     def job_file(self):
         return os.path.join(self.profile_dir, self.job_file_name)
 
     def start(self, n):
         """Start the controller by profile_dir."""
         return super(WindowsHPCEngineSetLauncher, self).start(n)
 
 
 #-----------------------------------------------------------------------------
 # Batch (PBS) system launchers
 #-----------------------------------------------------------------------------
 
 class BatchClusterAppMixin(ClusterAppMixin):
     """ClusterApp mixin that updates the self.context dict, rather than cl-args."""
     def _profile_dir_changed(self, name, old, new):
         self.context[name] = new
     _cluster_id_changed = _profile_dir_changed
 
     def _profile_dir_default(self):
         self.context['profile_dir'] = ''
         return ''
     def _cluster_id_default(self):
         self.context['cluster_id'] = ''
         return ''
 
 
 class BatchSystemLauncher(BaseLauncher):
     """Launch an external process using a batch system.
 
     This class is designed to work with UNIX batch systems like PBS, LSF,
     GridEngine, etc.  The overall model is that there are different commands
     like qsub, qdel, etc. that handle the starting and stopping of the process.
 
     This class also has the notion of a batch script. The ``batch_template``
     attribute can be set to a string that is a template for the batch script.
     This template is instantiated using string formatting. Thus the template can
     use {n} fot the number of instances. Subclasses can add additional variables
     to the template dict.
     """
 
     # Subclasses must fill these in.  See PBSEngineSet
     submit_command = List([''], config=True,
         help="The name of the command line program used to submit jobs.")
     delete_command = List([''], config=True,
         help="The name of the command line program used to delete jobs.")
     job_id_regexp = CRegExp('', config=True,
         help="""A regular expression used to get the job id from the output of the
         submit_command.""")
     batch_template = Unicode('', config=True,
         help="The string that is the batch script template itself.")
     batch_template_file = Unicode(u'', config=True,
         help="The file that contains the batch template.")
     batch_file_name = Unicode(u'batch_script', config=True,
         help="The filename of the instantiated batch script.")
     queue = Unicode(u'', config=True,
         help="The PBS Queue.")
 
     def _queue_changed(self, name, old, new):
         self.context[name] = new
 
     n = Integer(1)
     _n_changed = _queue_changed
 
     # not configurable, override in subclasses
     # PBS Job Array regex
     job_array_regexp = CRegExp('')
     job_array_template = Unicode('')
     # PBS Queue regex
     queue_regexp = CRegExp('')
     queue_template = Unicode('')
     # The default batch template, override in subclasses
     default_template = Unicode('')
     # The full path to the instantiated batch script.
     batch_file = Unicode(u'')
     # the format dict used with batch_template:
     context = Dict()
     def _context_default(self):
         """load the default context with the default values for the basic keys
 
         because the _trait_changed methods only load the context if they
         are set to something other than the default value.
         """
         return dict(n=1, queue=u'', profile_dir=u'', cluster_id=u'')
     
     # the Formatter instance for rendering the templates:
     formatter = Instance(EvalFormatter, (), {})
 
 
     def find_args(self):
         return self.submit_command + [self.batch_file]
 
     def __init__(self, work_dir=u'.', config=None, **kwargs):
         super(BatchSystemLauncher, self).__init__(
             work_dir=work_dir, config=config, **kwargs
         )
         self.batch_file = os.path.join(self.work_dir, self.batch_file_name)
 
     def parse_job_id(self, output):
         """Take the output of the submit command and return the job id."""
         m = self.job_id_regexp.search(output)
         if m is not None:
             job_id = m.group()
         else:
             raise LauncherError("Job id couldn't be determined: %s" % output)
         self.job_id = job_id
         self.log.info('Job submitted with job id: %r', job_id)
         return job_id
 
     def write_batch_script(self, n):
         """Instantiate and write the batch script to the work_dir."""
         self.n = n
         # first priority is batch_template if set
         if self.batch_template_file and not self.batch_template:
             # second priority is batch_template_file
             with open(self.batch_template_file) as f:
                 self.batch_template = f.read()
         if not self.batch_template:
             # third (last) priority is default_template
             self.batch_template = self.default_template
 
             # add jobarray or queue lines to user-specified template
             # note that this is *only* when user did not specify a template.
             # print self.job_array_regexp.search(self.batch_template)
             if not self.job_array_regexp.search(self.batch_template):
                 self.log.debug("adding job array settings to batch script")
                 firstline, rest = self.batch_template.split('\n',1)
                 self.batch_template = u'\n'.join([firstline, self.job_array_template, rest])
 
             # print self.queue_regexp.search(self.batch_template)
             if self.queue and not self.queue_regexp.search(self.batch_template):
                 self.log.debug("adding PBS queue settings to batch script")
                 firstline, rest = self.batch_template.split('\n',1)
                 self.batch_template = u'\n'.join([firstline, self.queue_template, rest])
 
         script_as_string = self.formatter.format(self.batch_template, **self.context)
         self.log.debug('Writing batch script: %s', self.batch_file)
 
         with open(self.batch_file, 'w') as f:
             f.write(script_as_string)
         os.chmod(self.batch_file, stat.S_IRUSR | stat.S_IWUSR | stat.S_IXUSR)
 
     def start(self, n):
         """Start n copies of the process using a batch system."""
         self.log.debug("Starting %s: %r", self.__class__.__name__, self.args)
         # Here we save profile_dir in the context so they
         # can be used in the batch script template as {profile_dir}
         self.write_batch_script(n)
         output = check_output(self.args, env=os.environ)
 
         job_id = self.parse_job_id(output)
         self.notify_start(job_id)
         return job_id
 
     def stop(self):
         output = check_output(self.delete_command+[self.job_id], env=os.environ)
         self.notify_stop(dict(job_id=self.job_id, output=output)) # Pass the output of the kill cmd
         return output
 
 
 class PBSLauncher(BatchSystemLauncher):
     """A BatchSystemLauncher subclass for PBS."""
 
     submit_command = List(['qsub'], config=True,
         help="The PBS submit command ['qsub']")
     delete_command = List(['qdel'], config=True,
         help="The PBS delete command ['qsub']")
     job_id_regexp = CRegExp(r'\d+', config=True,
         help="Regular expresion for identifying the job ID [r'\d+']")
 
     batch_file = Unicode(u'')
     job_array_regexp = CRegExp('#PBS\W+-t\W+[\w\d\-\$]+')
     job_array_template = Unicode('#PBS -t 1-{n}')
     queue_regexp = CRegExp('#PBS\W+-q\W+\$?\w+')
     queue_template = Unicode('#PBS -q {queue}')
 
 
 class PBSControllerLauncher(PBSLauncher, BatchClusterAppMixin):
     """Launch a controller using PBS."""
 
     batch_file_name = Unicode(u'pbs_controller', config=True,
         help="batch file name for the controller job.")
     default_template= Unicode("""#!/bin/sh
 #PBS -V
 #PBS -N ipcontroller
 %s --log-to-file --profile-dir="{profile_dir}" --cluster-id="{cluster_id}"
 """%(' '.join(ipcontroller_cmd_argv)))
 
 
     def start(self):
         """Start the controller by profile or profile_dir."""
         return super(PBSControllerLauncher, self).start(1)
 
 
 class PBSEngineSetLauncher(PBSLauncher, BatchClusterAppMixin):
     """Launch Engines using PBS"""
     batch_file_name = Unicode(u'pbs_engines', config=True,
         help="batch file name for the engine(s) job.")
     default_template= Unicode(u"""#!/bin/sh
 #PBS -V
 #PBS -N ipengine
 %s --profile-dir="{profile_dir}" --cluster-id="{cluster_id}"
 """%(' '.join(ipengine_cmd_argv)))
 
     def start(self, n):
         """Start n engines by profile or profile_dir."""
         return super(PBSEngineSetLauncher, self).start(n)
 
 #SGE is very similar to PBS
 
 class SGELauncher(PBSLauncher):
     """Sun GridEngine is a PBS clone with slightly different syntax"""
     job_array_regexp = CRegExp('#\$\W+\-t')
     job_array_template = Unicode('#$ -t 1-{n}')
     queue_regexp = CRegExp('#\$\W+-q\W+\$?\w+')
     queue_template = Unicode('#$ -q {queue}')
 
 class SGEControllerLauncher(SGELauncher, BatchClusterAppMixin):
     """Launch a controller using SGE."""
 
     batch_file_name = Unicode(u'sge_controller', config=True,
         help="batch file name for the ipontroller job.")
     default_template= Unicode(u"""#$ -V
 #$ -S /bin/sh
 #$ -N ipcontroller
 %s --log-to-file --profile-dir="{profile_dir}" --cluster-id="{cluster_id}"
 """%(' '.join(ipcontroller_cmd_argv)))
 
     def start(self):
         """Start the controller by profile or profile_dir."""
         return super(SGEControllerLauncher, self).start(1)
 
 class SGEEngineSetLauncher(SGELauncher, BatchClusterAppMixin):
     """Launch Engines with SGE"""
     batch_file_name = Unicode(u'sge_engines', config=True,
         help="batch file name for the engine(s) job.")
     default_template = Unicode("""#$ -V
 #$ -S /bin/sh
 #$ -N ipengine
 %s --profile-dir="{profile_dir}" --cluster-id="{cluster_id}"
 """%(' '.join(ipengine_cmd_argv)))
 
     def start(self, n):
         """Start n engines by profile or profile_dir."""
         return super(SGEEngineSetLauncher, self).start(n)
 
 
 # LSF launchers
 
 class LSFLauncher(BatchSystemLauncher):
     """A BatchSystemLauncher subclass for LSF."""
 
     submit_command = List(['bsub'], config=True,
                           help="The PBS submit command ['bsub']")
     delete_command = List(['bkill'], config=True,
                           help="The PBS delete command ['bkill']")
     job_id_regexp = CRegExp(r'\d+', config=True,
                             help="Regular expresion for identifying the job ID [r'\d+']")
 
     batch_file = Unicode(u'')
     job_array_regexp = CRegExp('#BSUB[ \t]-J+\w+\[\d+-\d+\]')
     job_array_template = Unicode('#BSUB -J ipengine[1-{n}]')
     queue_regexp = CRegExp('#BSUB[ \t]+-q[ \t]+\w+')
     queue_template = Unicode('#BSUB -q {queue}')
 
     def start(self, n):
         """Start n copies of the process using LSF batch system.
         This cant inherit from the base class because bsub expects
         to be piped a shell script in order to honor the #BSUB directives :
         bsub < script
         """
         # Here we save profile_dir in the context so they
         # can be used in the batch script template as {profile_dir}
         self.write_batch_script(n)
         #output = check_output(self.args, env=os.environ)
         piped_cmd = self.args[0]+'<\"'+self.args[1]+'\"'
         self.log.debug("Starting %s: %s", self.__class__.__name__, piped_cmd)
         p = Popen(piped_cmd, shell=True,env=os.environ,stdout=PIPE)
         output,err = p.communicate()
         job_id = self.parse_job_id(output)
         self.notify_start(job_id)
         return job_id
 
 
 class LSFControllerLauncher(LSFLauncher, BatchClusterAppMixin):
     """Launch a controller using LSF."""
 
     batch_file_name = Unicode(u'lsf_controller', config=True,
                               help="batch file name for the controller job.")
     default_template= Unicode("""#!/bin/sh
     #BSUB -J ipcontroller
     #BSUB -oo ipcontroller.o.%%J
     #BSUB -eo ipcontroller.e.%%J
     %s --log-to-file --profile-dir="{profile_dir}" --cluster-id="{cluster_id}"
     """%(' '.join(ipcontroller_cmd_argv)))
 
     def start(self):
         """Start the controller by profile or profile_dir."""
         return super(LSFControllerLauncher, self).start(1)
 
 
 class LSFEngineSetLauncher(LSFLauncher, BatchClusterAppMixin):
     """Launch Engines using LSF"""
     batch_file_name = Unicode(u'lsf_engines', config=True,
                               help="batch file name for the engine(s) job.")
     default_template= Unicode(u"""#!/bin/sh
     #BSUB -oo ipengine.o.%%J
     #BSUB -eo ipengine.e.%%J
     %s --profile-dir="{profile_dir}" --cluster-id="{cluster_id}"
     """%(' '.join(ipengine_cmd_argv)))
 
     def start(self, n):
         """Start n engines by profile or profile_dir."""
         return super(LSFEngineSetLauncher, self).start(n)
 
 
 #-----------------------------------------------------------------------------
 # A launcher for ipcluster itself!
 #-----------------------------------------------------------------------------
 
 
 class IPClusterLauncher(LocalProcessLauncher):
     """Launch the ipcluster program in an external process."""
 
     ipcluster_cmd = List(ipcluster_cmd_argv, config=True,
         help="Popen command for ipcluster")
     ipcluster_args = List(
         ['--clean-logs=True', '--log-to-file', '--log-level=%i'%logging.INFO], config=True,
         help="Command line arguments to pass to ipcluster.")
     ipcluster_subcommand = Unicode('start')
     profile = Unicode('default')
     n = Integer(2)
 
     def find_args(self):
         return self.ipcluster_cmd + [self.ipcluster_subcommand] + \
             ['--n=%i'%self.n, '--profile=%s'%self.profile] + \
             self.ipcluster_args
 
     def start(self):
         return super(IPClusterLauncher, self).start()
 
 #-----------------------------------------------------------------------------
 # Collections of launchers
 #-----------------------------------------------------------------------------
 
 local_launchers = [
     LocalControllerLauncher,
     LocalEngineLauncher,
     LocalEngineSetLauncher,
 ]
 mpi_launchers = [
     MPILauncher,
     MPIControllerLauncher,
     MPIEngineSetLauncher,
 ]
 ssh_launchers = [
     SSHLauncher,
     SSHControllerLauncher,
     SSHEngineLauncher,
     SSHEngineSetLauncher,
 ]
 winhpc_launchers = [
     WindowsHPCLauncher,
     WindowsHPCControllerLauncher,
     WindowsHPCEngineSetLauncher,
 ]
 pbs_launchers = [
     PBSLauncher,
     PBSControllerLauncher,
     PBSEngineSetLauncher,
 ]
 sge_launchers = [
     SGELauncher,
     SGEControllerLauncher,
     SGEEngineSetLauncher,
 ]
 lsf_launchers = [
     LSFLauncher,
     LSFControllerLauncher,
     LSFEngineSetLauncher,
 ]
 all_launchers = local_launchers + mpi_launchers + ssh_launchers + winhpc_launchers\
                 + pbs_launchers + sge_launchers + lsf_launchers
 

IPython/parallel/client/client.py

@@ -1,1505 +1,1628 b''
 """A semi-synchronous Client for the ZMQ cluster
 
 Authors:
 
 * MinRK
 """
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2010-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 import os
 import json
 import sys
 from threading import Thread, Event
 import time
 import warnings
 from datetime import datetime
 from getpass import getpass
 from pprint import pprint
 
 pjoin = os.path.join
 
 import zmq
 # from zmq.eventloop import ioloop, zmqstream
 
 from IPython.config.configurable import MultipleInstanceError
 from IPython.core.application import BaseIPythonApplication
 
 from IPython.utils.jsonutil import rekey
 from IPython.utils.localinterfaces import LOCAL_IPS
 from IPython.utils.path import get_ipython_dir
+from IPython.utils.py3compat import cast_bytes
 from IPython.utils.traitlets import (HasTraits, Integer, Instance, Unicode,
                                     Dict, List, Bool, Set, Any)
 from IPython.external.decorator import decorator
 from IPython.external.ssh import tunnel
 
 from IPython.parallel import Reference
 from IPython.parallel import error
 from IPython.parallel import util
 
 from IPython.zmq.session import Session, Message
 
 from .asyncresult import AsyncResult, AsyncHubResult
 from IPython.core.profiledir import ProfileDir, ProfileDirError
 from .view import DirectView, LoadBalancedView
 
 if sys.version_info[0] >= 3:
     # xrange is used in a couple 'isinstance' tests in py2
     # should be just 'range' in 3k
     xrange = range
 
 #--------------------------------------------------------------------------
 # Decorators for Client methods
 #--------------------------------------------------------------------------
 
 @decorator
 def spin_first(f, self, *args, **kwargs):
     """Call spin() to sync state prior to calling the method."""
     self.spin()
     return f(self, *args, **kwargs)
 
 
 #--------------------------------------------------------------------------
 # Classes
 #--------------------------------------------------------------------------
 
+
+class ExecuteReply(object):
+    """wrapper for finished Execute results"""
+    def __init__(self, msg_id, content, metadata):
+        self.msg_id = msg_id
+        self._content = content
+        self.execution_count = content['execution_count']
+        self.metadata = metadata
+    
+    def __getitem__(self, key):
+        return self.metadata[key]
+    
+    def __getattr__(self, key):
+        if key not in self.metadata:
+            raise AttributeError(key)
+        return self.metadata[key]
+    
+    def __repr__(self):
+        pyout = self.metadata['pyout'] or {}
+        text_out = pyout.get('data', {}).get('text/plain', '')
+        if len(text_out) > 32:
+            text_out = text_out[:29] + '...'
+        
+        return "<ExecuteReply[%i]: %s>" % (self.execution_count, text_out)
+    
+    def _repr_html_(self):
+        pyout = self.metadata['pyout'] or {'data':{}}
+        return pyout['data'].get("text/html")
+    
+    def _repr_latex_(self):
+        pyout = self.metadata['pyout'] or {'data':{}}
+        return pyout['data'].get("text/latex")
+    
+    def _repr_json_(self):
+        pyout = self.metadata['pyout'] or {'data':{}}
+        return pyout['data'].get("application/json")
+    
+    def _repr_javascript_(self):
+        pyout = self.metadata['pyout'] or {'data':{}}
+        return pyout['data'].get("application/javascript")
+    
+    def _repr_png_(self):
+        pyout = self.metadata['pyout'] or {'data':{}}
+        return pyout['data'].get("image/png")
+    
+    def _repr_jpeg_(self):
+        pyout = self.metadata['pyout'] or {'data':{}}
+        return pyout['data'].get("image/jpeg")
+    
+    def _repr_svg_(self):
+        pyout = self.metadata['pyout'] or {'data':{}}
+        return pyout['data'].get("image/svg+xml")
+
+
 class Metadata(dict):
     """Subclass of dict for initializing metadata values.
 
     Attribute access works on keys.
 
     These objects have a strict set of keys - errors will raise if you try
     to add new keys.
     """
     def __init__(self, *args, **kwargs):
         dict.__init__(self)
         md = {'msg_id' : None,
               'submitted' : None,
               'started' : None,
               'completed' : None,
               'received' : None,
               'engine_uuid' : None,
               'engine_id' : None,
               'follow' : None,
               'after' : None,
               'status' : None,
 
               'pyin' : None,
               'pyout' : None,
               'pyerr' : None,
               'stdout' : '',
               'stderr' : '',
+              'outputs' : [],
             }
         self.update(md)
         self.update(dict(*args, **kwargs))
 
     def __getattr__(self, key):
         """getattr aliased to getitem"""
         if key in self.iterkeys():
             return self[key]
         else:
             raise AttributeError(key)
 
     def __setattr__(self, key, value):
         """setattr aliased to setitem, with strict"""
         if key in self.iterkeys():
             self[key] = value
         else:
             raise AttributeError(key)
 
     def __setitem__(self, key, value):
         """strict static key enforcement"""
         if key in self.iterkeys():
             dict.__setitem__(self, key, value)
         else:
             raise KeyError(key)
 
 
 class Client(HasTraits):
     """A semi-synchronous client to the IPython ZMQ cluster
 
     Parameters
     ----------
 
     url_or_file : bytes or unicode; zmq url or path to ipcontroller-client.json
         Connection information for the Hub's registration.  If a json connector
         file is given, then likely no further configuration is necessary.
         [Default: use profile]
     profile : bytes
         The name of the Cluster profile to be used to find connector information.
         If run from an IPython application, the default profile will be the same
         as the running application, otherwise it will be 'default'.
     context : zmq.Context
         Pass an existing zmq.Context instance, otherwise the client will create its own.
     debug : bool
         flag for lots of message printing for debug purposes
     timeout : int/float
         time (in seconds) to wait for connection replies from the Hub
         [Default: 10]
 
     #-------------- session related args ----------------
 
     config : Config object
         If specified, this will be relayed to the Session for configuration
     username : str
         set username for the session object
     packer : str (import_string) or callable
         Can be either the simple keyword 'json' or 'pickle', or an import_string to a
         function to serialize messages. Must support same input as
         JSON, and output must be bytes.
         You can pass a callable directly as `pack`
     unpacker : str (import_string) or callable
         The inverse of packer.  Only necessary if packer is specified as *not* one
         of 'json' or 'pickle'.
 
     #-------------- ssh related args ----------------
     # These are args for configuring the ssh tunnel to be used
     # credentials are used to forward connections over ssh to the Controller
     # Note that the ip given in `addr` needs to be relative to sshserver
     # The most basic case is to leave addr as pointing to localhost (127.0.0.1),
     # and set sshserver as the same machine the Controller is on. However,
     # the only requirement is that sshserver is able to see the Controller
     # (i.e. is within the same trusted network).
 
     sshserver : str
         A string of the form passed to ssh, i.e. 'server.tld' or '[e-mail]:port'
         If keyfile or password is specified, and this is not, it will default to
         the ip given in addr.
     sshkey : str; path to ssh private key file
         This specifies a key to be used in ssh login, default None.
         Regular default ssh keys will be used without specifying this argument.
     password : str
         Your ssh password to sshserver. Note that if this is left None,
         you will be prompted for it if passwordless key based login is unavailable.
     paramiko : bool
         flag for whether to use paramiko instead of shell ssh for tunneling.
         [default: True on win32, False else]
 
     ------- exec authentication args -------
     If even localhost is untrusted, you can have some protection against
     unauthorized execution by signing messages with HMAC digests.
     Messages are still sent as cleartext, so if someone can snoop your
     loopback traffic this will not protect your privacy, but will prevent
     unauthorized execution.
 
     exec_key : str
         an authentication key or file containing a key
         default: None
 
 
     Attributes
     ----------
 
     ids : list of int engine IDs
         requesting the ids attribute always synchronizes
         the registration state. To request ids without synchronization,
         use semi-private _ids attributes.
 
     history : list of msg_ids
         a list of msg_ids, keeping track of all the execution
         messages you have submitted in order.
 
     outstanding : set of msg_ids
         a set of msg_ids that have been submitted, but whose
         results have not yet been received.
 
     results : dict
         a dict of all our results, keyed by msg_id
 
     block : bool
         determines default behavior when block not specified
         in execution methods
 
     Methods
     -------
 
     spin
         flushes incoming results and registration state changes
         control methods spin, and requesting `ids` also ensures up to date
 
     wait
         wait on one or more msg_ids
 
     execution methods
         apply
         legacy: execute, run
 
     data movement
         push, pull, scatter, gather
 
     query methods
         queue_status, get_result, purge, result_status
 
     control methods
         abort, shutdown
 
     """
 
 
     block = Bool(False)
     outstanding = Set()
     results = Instance('collections.defaultdict', (dict,))
     metadata = Instance('collections.defaultdict', (Metadata,))
     history = List()
     debug = Bool(False)
     _spin_thread = Any()
     _stop_spinning = Any()
 
     profile=Unicode()
     def _profile_default(self):
         if BaseIPythonApplication.initialized():
             # an IPython app *might* be running, try to get its profile
             try:
                 return BaseIPythonApplication.instance().profile
             except (AttributeError, MultipleInstanceError):
                 # could be a *different* subclass of config.Application,
                 # which would raise one of these two errors.
                 return u'default'
         else:
             return u'default'
 
 
     _outstanding_dict = Instance('collections.defaultdict', (set,))
     _ids = List()
     _connected=Bool(False)
     _ssh=Bool(False)
     _context = Instance('zmq.Context')
     _config = Dict()
     _engines=Instance(util.ReverseDict, (), {})
     # _hub_socket=Instance('zmq.Socket')
     _query_socket=Instance('zmq.Socket')
     _control_socket=Instance('zmq.Socket')
     _iopub_socket=Instance('zmq.Socket')
     _notification_socket=Instance('zmq.Socket')
     _mux_socket=Instance('zmq.Socket')
     _task_socket=Instance('zmq.Socket')
     _task_scheme=Unicode()
     _closed = False
     _ignored_control_replies=Integer(0)
     _ignored_hub_replies=Integer(0)
 
     def __new__(self, *args, **kw):
         # don't raise on positional args
         return HasTraits.__new__(self, **kw)
 
     def __init__(self, url_or_file=None, profile=None, profile_dir=None, ipython_dir=None,
             context=None, debug=False, exec_key=None,
             sshserver=None, sshkey=None, password=None, paramiko=None,
             timeout=10, **extra_args
             ):
         if profile:
             super(Client, self).__init__(debug=debug, profile=profile)
         else:
             super(Client, self).__init__(debug=debug)
         if context is None:
             context = zmq.Context.instance()
         self._context = context
         self._stop_spinning = Event()
 
         self._setup_profile_dir(self.profile, profile_dir, ipython_dir)
         if self._cd is not None:
             if url_or_file is None:
                 url_or_file = pjoin(self._cd.security_dir, 'ipcontroller-client.json')
-        assert url_or_file is not None, "I can't find enough information to connect to a hub!"\
-            " Please specify at least one of url_or_file or profile."
+        if url_or_file is None:
+            raise ValueError(
+                "I can't find enough information to connect to a hub!"
+                " Please specify at least one of url_or_file or profile."
+            )
 
         if not util.is_url(url_or_file):
             # it's not a url, try for a file
             if not os.path.exists(url_or_file):
                 if self._cd:
                     url_or_file = os.path.join(self._cd.security_dir, url_or_file)
-                assert os.path.exists(url_or_file), "Not a valid connection file or url: %r"%url_or_file
+                if not os.path.exists(url_or_file):
+                    raise IOError("Connection file not found: %r" % url_or_file)
             with open(url_or_file) as f:
                 cfg = json.loads(f.read())
         else:
             cfg = {'url':url_or_file}
 
         # sync defaults from args, json:
         if sshserver:
             cfg['ssh'] = sshserver
         if exec_key:
             cfg['exec_key'] = exec_key
         exec_key = cfg['exec_key']
         location = cfg.setdefault('location', None)
         cfg['url'] = util.disambiguate_url(cfg['url'], location)
         url = cfg['url']
         proto,addr,port = util.split_url(url)
         if location is not None and addr == '127.0.0.1':
             # location specified, and connection is expected to be local
             if location not in LOCAL_IPS and not sshserver:
                 # load ssh from JSON *only* if the controller is not on
                 # this machine
                 sshserver=cfg['ssh']
             if location not in LOCAL_IPS and not sshserver:
                 # warn if no ssh specified, but SSH is probably needed
                 # This is only a warning, because the most likely cause
                 # is a local Controller on a laptop whose IP is dynamic
                 warnings.warn("""
             Controller appears to be listening on localhost, but not on this machine.
             If this is true, you should specify Client(...,sshserver='you@%s')
             or instruct your controller to listen on an external IP."""%location,
                 RuntimeWarning)
         elif not sshserver:
             # otherwise sync with cfg
             sshserver = cfg['ssh']
 
         self._config = cfg
 
         self._ssh = bool(sshserver or sshkey or password)
         if self._ssh and sshserver is None:
             # default to ssh via localhost
             sshserver = url.split('://')[1].split(':')[0]
         if self._ssh and password is None:
             if tunnel.try_passwordless_ssh(sshserver, sshkey, paramiko):
                 password=False
             else:
                 password = getpass("SSH Password for %s: "%sshserver)
         ssh_kwargs = dict(keyfile=sshkey, password=password, paramiko=paramiko)
 
         # configure and construct the session
         if exec_key is not None:
             if os.path.isfile(exec_key):
                 extra_args['keyfile'] = exec_key
             else:
-                exec_key = util.asbytes(exec_key)
+                exec_key = cast_bytes(exec_key)
                 extra_args['key'] = exec_key
         self.session = Session(**extra_args)
 
         self._query_socket = self._context.socket(zmq.DEALER)
         self._query_socket.setsockopt(zmq.IDENTITY, self.session.bsession)
         if self._ssh:
             tunnel.tunnel_connection(self._query_socket, url, sshserver, **ssh_kwargs)
         else:
             self._query_socket.connect(url)
 
         self.session.debug = self.debug
 
         self._notification_handlers = {'registration_notification' : self._register_engine,
                                     'unregistration_notification' : self._unregister_engine,
                                     'shutdown_notification' : lambda msg: self.close(),
                                     }
         self._queue_handlers = {'execute_reply' : self._handle_execute_reply,
                                 'apply_reply' : self._handle_apply_reply}
         self._connect(sshserver, ssh_kwargs, timeout)
 
     def __del__(self):
         """cleanup sockets, but _not_ context."""
         self.close()
 
     def _setup_profile_dir(self, profile, profile_dir, ipython_dir):
         if ipython_dir is None:
             ipython_dir = get_ipython_dir()
         if profile_dir is not None:
             try:
                 self._cd = ProfileDir.find_profile_dir(profile_dir)
                 return
             except ProfileDirError:
                 pass
         elif profile is not None:
             try:
                 self._cd = ProfileDir.find_profile_dir_by_name(
                     ipython_dir, profile)
                 return
             except ProfileDirError:
                 pass
         self._cd = None
 
     def _update_engines(self, engines):
         """Update our engines dict and _ids from a dict of the form: {id:uuid}."""
         for k,v in engines.iteritems():
             eid = int(k)
             self._engines[eid] = v
             self._ids.append(eid)
         self._ids = sorted(self._ids)
         if sorted(self._engines.keys()) != range(len(self._engines)) and \
                         self._task_scheme == 'pure' and self._task_socket:
             self._stop_scheduling_tasks()
 
     def _stop_scheduling_tasks(self):
         """Stop scheduling tasks because an engine has been unregistered
         from a pure ZMQ scheduler.
         """
         self._task_socket.close()
         self._task_socket = None
         msg = "An engine has been unregistered, and we are using pure " +\
               "ZMQ task scheduling.  Task farming will be disabled."
         if self.outstanding:
             msg += " If you were running tasks when this happened, " +\
                    "some `outstanding` msg_ids may never resolve."
         warnings.warn(msg, RuntimeWarning)
 
     def _build_targets(self, targets):
         """Turn valid target IDs or 'all' into two lists:
         (int_ids, uuids).
         """
         if not self._ids:
             # flush notification socket if no engines yet, just in case
             if not self.ids:
                 raise error.NoEnginesRegistered("Can't build targets without any engines")
 
         if targets is None:
             targets = self._ids
         elif isinstance(targets, basestring):
             if targets.lower() == 'all':
                 targets = self._ids
             else:
                 raise TypeError("%r not valid str target, must be 'all'"%(targets))
         elif isinstance(targets, int):
             if targets < 0:
                 targets = self.ids[targets]
             if targets not in self._ids:
                 raise IndexError("No such engine: %i"%targets)
             targets = [targets]
 
         if isinstance(targets, slice):
             indices = range(len(self._ids))[targets]
             ids = self.ids
             targets = [ ids[i] for i in indices ]
 
         if not isinstance(targets, (tuple, list, xrange)):
             raise TypeError("targets by int/slice/collection of ints only, not %s"%(type(targets)))
 
-        return [util.asbytes(self._engines[t]) for t in targets], list(targets)
+        return [cast_bytes(self._engines[t]) for t in targets], list(targets)
 
     def _connect(self, sshserver, ssh_kwargs, timeout):
         """setup all our socket connections to the cluster. This is called from
         __init__."""
 
         # Maybe allow reconnecting?
         if self._connected:
             return
         self._connected=True
 
         def connect_socket(s, url):
             url = util.disambiguate_url(url, self._config['location'])
             if self._ssh:
                 return tunnel.tunnel_connection(s, url, sshserver, **ssh_kwargs)
             else:
                 return s.connect(url)
 
         self.session.send(self._query_socket, 'connection_request')
         # use Poller because zmq.select has wrong units in pyzmq 2.1.7
         poller = zmq.Poller()
         poller.register(self._query_socket, zmq.POLLIN)
         # poll expects milliseconds, timeout is seconds
         evts = poller.poll(timeout*1000)
         if not evts:
             raise error.TimeoutError("Hub connection request timed out")
         idents,msg = self.session.recv(self._query_socket,mode=0)
         if self.debug:
             pprint(msg)
         msg = Message(msg)
         content = msg.content
         self._config['registration'] = dict(content)
         if content.status == 'ok':
             ident = self.session.bsession
             if content.mux:
                 self._mux_socket = self._context.socket(zmq.DEALER)
                 self._mux_socket.setsockopt(zmq.IDENTITY, ident)
                 connect_socket(self._mux_socket, content.mux)
             if content.task:
                 self._task_scheme, task_addr = content.task
                 self._task_socket = self._context.socket(zmq.DEALER)
                 self._task_socket.setsockopt(zmq.IDENTITY, ident)
                 connect_socket(self._task_socket, task_addr)
             if content.notification:
                 self._notification_socket = self._context.socket(zmq.SUB)
                 connect_socket(self._notification_socket, content.notification)
                 self._notification_socket.setsockopt(zmq.SUBSCRIBE, b'')
             # if content.query:
             #     self._query_socket = self._context.socket(zmq.DEALER)
             #     self._query_socket.setsockopt(zmq.IDENTITY, self.session.bsession)
             #     connect_socket(self._query_socket, content.query)
             if content.control:
                 self._control_socket = self._context.socket(zmq.DEALER)
                 self._control_socket.setsockopt(zmq.IDENTITY, ident)
                 connect_socket(self._control_socket, content.control)
             if content.iopub:
                 self._iopub_socket = self._context.socket(zmq.SUB)
                 self._iopub_socket.setsockopt(zmq.SUBSCRIBE, b'')
                 self._iopub_socket.setsockopt(zmq.IDENTITY, ident)
                 connect_socket(self._iopub_socket, content.iopub)
             self._update_engines(dict(content.engines))
         else:
             self._connected = False
             raise Exception("Failed to connect!")
 
     #--------------------------------------------------------------------------
     # handlers and callbacks for incoming messages
     #--------------------------------------------------------------------------
 
     def _unwrap_exception(self, content):
         """unwrap exception, and remap engine_id to int."""
         e = error.unwrap_exception(content)
         # print e.traceback
         if e.engine_info:
             e_uuid = e.engine_info['engine_uuid']
             eid = self._engines[e_uuid]
             e.engine_info['engine_id'] = eid
         return e
 
     def _extract_metadata(self, header, parent, content):
         md = {'msg_id' : parent['msg_id'],
               'received' : datetime.now(),
               'engine_uuid' : header.get('engine', None),
               'follow' : parent.get('follow', []),
               'after' : parent.get('after', []),
               'status' : content['status'],
             }
 
         if md['engine_uuid'] is not None:
             md['engine_id'] = self._engines.get(md['engine_uuid'], None)
 
         if 'date' in parent:
             md['submitted'] = parent['date']
         if 'started' in header:
             md['started'] = header['started']
         if 'date' in header:
             md['completed'] = header['date']
         return md
 
     def _register_engine(self, msg):
         """Register a new engine, and update our connection info."""
         content = msg['content']
         eid = content['id']
         d = {eid : content['queue']}
         self._update_engines(d)
 
     def _unregister_engine(self, msg):
         """Unregister an engine that has died."""
         content = msg['content']
         eid = int(content['id'])
         if eid in self._ids:
             self._ids.remove(eid)
             uuid = self._engines.pop(eid)
 
             self._handle_stranded_msgs(eid, uuid)
 
         if self._task_socket and self._task_scheme == 'pure':
             self._stop_scheduling_tasks()
 
     def _handle_stranded_msgs(self, eid, uuid):
         """Handle messages known to be on an engine when the engine unregisters.
 
         It is possible that this will fire prematurely - that is, an engine will
         go down after completing a result, and the client will be notified
         of the unregistration and later receive the successful result.
         """
 
         outstanding = self._outstanding_dict[uuid]
 
         for msg_id in list(outstanding):
             if msg_id in self.results:
                 # we already
                 continue
             try:
                 raise error.EngineError("Engine %r died while running task %r"%(eid, msg_id))
             except:
                 content = error.wrap_exception()
             # build a fake message:
             parent = {}
             header = {}
             parent['msg_id'] = msg_id
             header['engine'] = uuid
             header['date'] = datetime.now()
             msg = dict(parent_header=parent, header=header, content=content)
             self._handle_apply_reply(msg)
 
     def _handle_execute_reply(self, msg):
         """Save the reply to an execute_request into our results.
 
         execute messages are never actually used. apply is used instead.
         """
 
         parent = msg['parent_header']
         msg_id = parent['msg_id']
         if msg_id not in self.outstanding:
             if msg_id in self.history:
                 print ("got stale result: %s"%msg_id)
             else:
                 print ("got unknown result: %s"%msg_id)
         else:
             self.outstanding.remove(msg_id)
-        self.results[msg_id] = self._unwrap_exception(msg['content'])
+
+        content = msg['content']
+        header = msg['header']
+
+        # construct metadata:
+        md = self.metadata[msg_id]
+        md.update(self._extract_metadata(header, parent, content))
+        # is this redundant?
+        self.metadata[msg_id] = md
+        
+        e_outstanding = self._outstanding_dict[md['engine_uuid']]
+        if msg_id in e_outstanding:
+            e_outstanding.remove(msg_id)
+
+        # construct result:
+        if content['status'] == 'ok':
+            self.results[msg_id] = ExecuteReply(msg_id, content, md)
+        elif content['status'] == 'aborted':
+            self.results[msg_id] = error.TaskAborted(msg_id)
+        elif content['status'] == 'resubmitted':
+            # TODO: handle resubmission
+            pass
+        else:
+            self.results[msg_id] = self._unwrap_exception(content)
 
     def _handle_apply_reply(self, msg):
         """Save the reply to an apply_request into our results."""
         parent = msg['parent_header']
         msg_id = parent['msg_id']
         if msg_id not in self.outstanding:
             if msg_id in self.history:
                 print ("got stale result: %s"%msg_id)
                 print self.results[msg_id]
                 print msg
             else:
                 print ("got unknown result: %s"%msg_id)
         else:
             self.outstanding.remove(msg_id)
         content = msg['content']
         header = msg['header']
 
         # construct metadata:
         md = self.metadata[msg_id]
         md.update(self._extract_metadata(header, parent, content))
         # is this redundant?
         self.metadata[msg_id] = md
 
         e_outstanding = self._outstanding_dict[md['engine_uuid']]
         if msg_id in e_outstanding:
             e_outstanding.remove(msg_id)
 
         # construct result:
         if content['status'] == 'ok':
             self.results[msg_id] = util.unserialize_object(msg['buffers'])[0]
         elif content['status'] == 'aborted':
             self.results[msg_id] = error.TaskAborted(msg_id)
         elif content['status'] == 'resubmitted':
             # TODO: handle resubmission
             pass
         else:
             self.results[msg_id] = self._unwrap_exception(content)
 
     def _flush_notifications(self):
         """Flush notifications of engine registrations waiting
         in ZMQ queue."""
         idents,msg = self.session.recv(self._notification_socket, mode=zmq.NOBLOCK)
         while msg is not None:
             if self.debug:
                 pprint(msg)
             msg_type = msg['header']['msg_type']
             handler = self._notification_handlers.get(msg_type, None)
             if handler is None:
                 raise Exception("Unhandled message type: %s"%msg.msg_type)
             else:
                 handler(msg)
             idents,msg = self.session.recv(self._notification_socket, mode=zmq.NOBLOCK)
 
     def _flush_results(self, sock):
         """Flush task or queue results waiting in ZMQ queue."""
         idents,msg = self.session.recv(sock, mode=zmq.NOBLOCK)
         while msg is not None:
             if self.debug:
                 pprint(msg)
             msg_type = msg['header']['msg_type']
             handler = self._queue_handlers.get(msg_type, None)
             if handler is None:
                 raise Exception("Unhandled message type: %s"%msg.msg_type)
             else:
                 handler(msg)
             idents,msg = self.session.recv(sock, mode=zmq.NOBLOCK)
 
     def _flush_control(self, sock):
         """Flush replies from the control channel waiting
         in the ZMQ queue.
 
         Currently: ignore them."""
         if self._ignored_control_replies <= 0:
             return
         idents,msg = self.session.recv(sock, mode=zmq.NOBLOCK)
         while msg is not None:
             self._ignored_control_replies -= 1
             if self.debug:
                 pprint(msg)
             idents,msg = self.session.recv(sock, mode=zmq.NOBLOCK)
 
     def _flush_ignored_control(self):
         """flush ignored control replies"""
         while self._ignored_control_replies > 0:
             self.session.recv(self._control_socket)
             self._ignored_control_replies -= 1
 
     def _flush_ignored_hub_replies(self):
         ident,msg = self.session.recv(self._query_socket, mode=zmq.NOBLOCK)
         while msg is not None:
             ident,msg = self.session.recv(self._query_socket, mode=zmq.NOBLOCK)
 
     def _flush_iopub(self, sock):
         """Flush replies from the iopub channel waiting
         in the ZMQ queue.
         """
         idents,msg = self.session.recv(sock, mode=zmq.NOBLOCK)
         while msg is not None:
             if self.debug:
                 pprint(msg)
             parent = msg['parent_header']
             # ignore IOPub messages with no parent.
             # Caused by print statements or warnings from before the first execution.
             if not parent:
                 continue
             msg_id = parent['msg_id']
             content = msg['content']
             header = msg['header']
             msg_type = msg['header']['msg_type']
 
             # init metadata:
             md = self.metadata[msg_id]
 
             if msg_type == 'stream':
                 name = content['name']
                 s = md[name] or ''
                 md[name] = s + content['data']
             elif msg_type == 'pyerr':
                 md.update({'pyerr' : self._unwrap_exception(content)})
             elif msg_type == 'pyin':
                 md.update({'pyin' : content['code']})
+            elif msg_type == 'display_data':
+                md['outputs'].append(content)
+            elif msg_type == 'pyout':
+                md['pyout'] = content
             else:
-                md.update({msg_type : content.get('data', '')})
+                # unhandled msg_type (status, etc.)
+                pass
 
             # reduntant?
             self.metadata[msg_id] = md
 
             idents,msg = self.session.recv(sock, mode=zmq.NOBLOCK)
 
     #--------------------------------------------------------------------------
     # len, getitem
     #--------------------------------------------------------------------------
 
     def __len__(self):
         """len(client) returns # of engines."""
         return len(self.ids)
 
     def __getitem__(self, key):
         """index access returns DirectView multiplexer objects
 
         Must be int, slice, or list/tuple/xrange of ints"""
         if not isinstance(key, (int, slice, tuple, list, xrange)):
             raise TypeError("key by int/slice/iterable of ints only, not %s"%(type(key)))
         else:
             return self.direct_view(key)
 
     #--------------------------------------------------------------------------
     # Begin public methods
     #--------------------------------------------------------------------------
 
     @property
     def ids(self):
         """Always up-to-date ids property."""
         self._flush_notifications()
         # always copy:
         return list(self._ids)
 
     def close(self):
         if self._closed:
             return
         self.stop_spin_thread()
         snames = filter(lambda n: n.endswith('socket'), dir(self))
         for socket in map(lambda name: getattr(self, name), snames):
             if isinstance(socket, zmq.Socket) and not socket.closed:
                 socket.close()
         self._closed = True
 
     def _spin_every(self, interval=1):
         """target func for use in spin_thread"""
         while True:
             if self._stop_spinning.is_set():
                 return
             time.sleep(interval)
             self.spin()
 
     def spin_thread(self, interval=1):
         """call Client.spin() in a background thread on some regular interval
         
         This helps ensure that messages don't pile up too much in the zmq queue
         while you are working on other things, or just leaving an idle terminal.
         
         It also helps limit potential padding of the `received` timestamp
         on AsyncResult objects, used for timings.
         
         Parameters
         ----------
         
         interval : float, optional
             The interval on which to spin the client in the background thread
             (simply passed to time.sleep).
         
         Notes
         -----
         
         For precision timing, you may want to use this method to put a bound
         on the jitter (in seconds) in `received` timestamps used
         in AsyncResult.wall_time.
         
         """
         if self._spin_thread is not None:
             self.stop_spin_thread()
         self._stop_spinning.clear()
         self._spin_thread = Thread(target=self._spin_every, args=(interval,))
         self._spin_thread.daemon = True
         self._spin_thread.start()
     
     def stop_spin_thread(self):
         """stop background spin_thread, if any"""
         if self._spin_thread is not None:
             self._stop_spinning.set()
             self._spin_thread.join()
             self._spin_thread = None
 
     def spin(self):
         """Flush any registration notifications and execution results
         waiting in the ZMQ queue.
         """
         if self._notification_socket:
             self._flush_notifications()
+        if self._iopub_socket:
+            self._flush_iopub(self._iopub_socket)
         if self._mux_socket:
             self._flush_results(self._mux_socket)
         if self._task_socket:
             self._flush_results(self._task_socket)
         if self._control_socket:
             self._flush_control(self._control_socket)
-        if self._iopub_socket:
-            self._flush_iopub(self._iopub_socket)
         if self._query_socket:
             self._flush_ignored_hub_replies()
 
     def wait(self, jobs=None, timeout=-1):
         """waits on one or more `jobs`, for up to `timeout` seconds.
 
         Parameters
         ----------
 
         jobs : int, str, or list of ints and/or strs, or one or more AsyncResult objects
                 ints are indices to self.history
                 strs are msg_ids
                 default: wait on all outstanding messages
         timeout : float
                 a time in seconds, after which to give up.
                 default is -1, which means no timeout
 
         Returns
         -------
 
         True : when all msg_ids are done
         False : timeout reached, some msg_ids still outstanding
         """
         tic = time.time()
         if jobs is None:
             theids = self.outstanding
         else:
             if isinstance(jobs, (int, basestring, AsyncResult)):
                 jobs = [jobs]
             theids = set()
             for job in jobs:
                 if isinstance(job, int):
                     # index access
                     job = self.history[job]
                 elif isinstance(job, AsyncResult):
                     map(theids.add, job.msg_ids)
                     continue
                 theids.add(job)
         if not theids.intersection(self.outstanding):
             return True
         self.spin()
         while theids.intersection(self.outstanding):
             if timeout >= 0 and ( time.time()-tic ) > timeout:
                 break
             time.sleep(1e-3)
             self.spin()
         return len(theids.intersection(self.outstanding)) == 0
 
     #--------------------------------------------------------------------------
     # Control methods
     #--------------------------------------------------------------------------
 
     @spin_first
     def clear(self, targets=None, block=None):
         """Clear the namespace in target(s)."""
         block = self.block if block is None else block
         targets = self._build_targets(targets)[0]
         for t in targets:
             self.session.send(self._control_socket, 'clear_request', content={}, ident=t)
         error = False
         if block:
             self._flush_ignored_control()
             for i in range(len(targets)):
                 idents,msg = self.session.recv(self._control_socket,0)
                 if self.debug:
                     pprint(msg)
                 if msg['content']['status'] != 'ok':
                     error = self._unwrap_exception(msg['content'])
         else:
             self._ignored_control_replies += len(targets)
         if error:
             raise error
 
 
     @spin_first
     def abort(self, jobs=None, targets=None, block=None):
         """Abort specific jobs from the execution queues of target(s).
 
         This is a mechanism to prevent jobs that have already been submitted
         from executing.
 
         Parameters
         ----------
 
         jobs : msg_id, list of msg_ids, or AsyncResult
             The jobs to be aborted
             
             If unspecified/None: abort all outstanding jobs.
 
         """
         block = self.block if block is None else block
         jobs = jobs if jobs is not None else list(self.outstanding)
         targets = self._build_targets(targets)[0]
         
         msg_ids = []
         if isinstance(jobs, (basestring,AsyncResult)):
             jobs = [jobs]
         bad_ids = filter(lambda obj: not isinstance(obj, (basestring, AsyncResult)), jobs)
         if bad_ids:
             raise TypeError("Invalid msg_id type %r, expected str or AsyncResult"%bad_ids[0])
         for j in jobs:
             if isinstance(j, AsyncResult):
                 msg_ids.extend(j.msg_ids)
             else:
                 msg_ids.append(j)
         content = dict(msg_ids=msg_ids)
         for t in targets:
             self.session.send(self._control_socket, 'abort_request',
                     content=content, ident=t)
         error = False
         if block:
             self._flush_ignored_control()
             for i in range(len(targets)):
                 idents,msg = self.session.recv(self._control_socket,0)
                 if self.debug:
                     pprint(msg)
                 if msg['content']['status'] != 'ok':
                     error = self._unwrap_exception(msg['content'])
         else:
             self._ignored_control_replies += len(targets)
         if error:
             raise error
 
     @spin_first
     def shutdown(self, targets=None, restart=False, hub=False, block=None):
         """Terminates one or more engine processes, optionally including the hub."""
         block = self.block if block is None else block
         if hub:
             targets = 'all'
         targets = self._build_targets(targets)[0]
         for t in targets:
             self.session.send(self._control_socket, 'shutdown_request',
                         content={'restart':restart},ident=t)
         error = False
         if block or hub:
             self._flush_ignored_control()
             for i in range(len(targets)):
                 idents,msg = self.session.recv(self._control_socket, 0)
                 if self.debug:
                     pprint(msg)
                 if msg['content']['status'] != 'ok':
                     error = self._unwrap_exception(msg['content'])
         else:
             self._ignored_control_replies += len(targets)
 
         if hub:
             time.sleep(0.25)
             self.session.send(self._query_socket, 'shutdown_request')
             idents,msg = self.session.recv(self._query_socket, 0)
             if self.debug:
                 pprint(msg)
             if msg['content']['status'] != 'ok':
                 error = self._unwrap_exception(msg['content'])
 
         if error:
             raise error
 
     #--------------------------------------------------------------------------
     # Execution related methods
     #--------------------------------------------------------------------------
 
     def _maybe_raise(self, result):
         """wrapper for maybe raising an exception if apply failed."""
         if isinstance(result, error.RemoteError):
             raise result
 
         return result
 
-    def send_apply_message(self, socket, f, args=None, kwargs=None, subheader=None, track=False,
+    def send_apply_request(self, socket, f, args=None, kwargs=None, subheader=None, track=False,
                             ident=None):
         """construct and send an apply message via a socket.
 
         This is the principal method with which all engine execution is performed by views.
         """
 
-        assert not self._closed, "cannot use me anymore, I'm closed!"
+        if self._closed:
+            raise RuntimeError("Client cannot be used after its sockets have been closed")
+        
         # defaults:
         args = args if args is not None else []
         kwargs = kwargs if kwargs is not None else {}
         subheader = subheader if subheader is not None else {}
 
         # validate arguments
         if not callable(f) and not isinstance(f, Reference):
             raise TypeError("f must be callable, not %s"%type(f))
         if not isinstance(args, (tuple, list)):
             raise TypeError("args must be tuple or list, not %s"%type(args))
         if not isinstance(kwargs, dict):
             raise TypeError("kwargs must be dict, not %s"%type(kwargs))
         if not isinstance(subheader, dict):
             raise TypeError("subheader must be dict, not %s"%type(subheader))
 
         bufs = util.pack_apply_message(f,args,kwargs)
 
         msg = self.session.send(socket, "apply_request", buffers=bufs, ident=ident,
                             subheader=subheader, track=track)
 
         msg_id = msg['header']['msg_id']
         self.outstanding.add(msg_id)
         if ident:
             # possibly routed to a specific engine
             if isinstance(ident, list):
                 ident = ident[-1]
             if ident in self._engines.values():
                 # save for later, in case of engine death
                 self._outstanding_dict[ident].add(msg_id)
         self.history.append(msg_id)
         self.metadata[msg_id]['submitted'] = datetime.now()
 
         return msg
 
+    def send_execute_request(self, socket, code, silent=True, subheader=None, ident=None):
+        """construct and send an execute request via a socket.
+
+        """
+
+        if self._closed:
+            raise RuntimeError("Client cannot be used after its sockets have been closed")
+        
+        # defaults:
+        subheader = subheader if subheader is not None else {}
+
+        # validate arguments
+        if not isinstance(code, basestring):
+            raise TypeError("code must be text, not %s" % type(code))
+        if not isinstance(subheader, dict):
+            raise TypeError("subheader must be dict, not %s" % type(subheader))
+        
+        content = dict(code=code, silent=bool(silent), user_variables=[], user_expressions={})
+
+
+        msg = self.session.send(socket, "execute_request", content=content, ident=ident,
+                            subheader=subheader)
+
+        msg_id = msg['header']['msg_id']
+        self.outstanding.add(msg_id)
+        if ident:
+            # possibly routed to a specific engine
+            if isinstance(ident, list):
+                ident = ident[-1]
+            if ident in self._engines.values():
+                # save for later, in case of engine death
+                self._outstanding_dict[ident].add(msg_id)
+        self.history.append(msg_id)
+        self.metadata[msg_id]['submitted'] = datetime.now()
+
+        return msg
+
     #--------------------------------------------------------------------------
     # construct a View object
     #--------------------------------------------------------------------------
 
     def load_balanced_view(self, targets=None):
         """construct a DirectView object.
 
         If no arguments are specified, create a LoadBalancedView
         using all engines.
 
         Parameters
         ----------
 
         targets: list,slice,int,etc. [default: use all engines]
             The subset of engines across which to load-balance
         """
         if targets == 'all':
             targets = None
         if targets is not None:
             targets = self._build_targets(targets)[1]
         return LoadBalancedView(client=self, socket=self._task_socket, targets=targets)
 
     def direct_view(self, targets='all'):
         """construct a DirectView object.
 
         If no targets are specified, create a DirectView using all engines.
         
         rc.direct_view('all') is distinguished from rc[:] in that 'all' will
         evaluate the target engines at each execution, whereas rc[:] will connect to
         all *current* engines, and that list will not change.
         
         That is, 'all' will always use all engines, whereas rc[:] will not use
         engines added after the DirectView is constructed.
 
         Parameters
         ----------
 
         targets: list,slice,int,etc. [default: use all engines]
             The engines to use for the View
         """
         single = isinstance(targets, int)
         # allow 'all' to be lazily evaluated at each execution
         if targets != 'all':
             targets = self._build_targets(targets)[1]
         if single:
             targets = targets[0]
         return DirectView(client=self, socket=self._mux_socket, targets=targets)
 
     #--------------------------------------------------------------------------
     # Query methods
     #--------------------------------------------------------------------------
 
     @spin_first
     def get_result(self, indices_or_msg_ids=None, block=None):
         """Retrieve a result by msg_id or history index, wrapped in an AsyncResult object.
 
         If the client already has the results, no request to the Hub will be made.
 
         This is a convenient way to construct AsyncResult objects, which are wrappers
         that include metadata about execution, and allow for awaiting results that
         were not submitted by this Client.
 
         It can also be a convenient way to retrieve the metadata associated with
         blocking execution, since it always retrieves
 
         Examples
         --------
         ::
 
             In [10]: r = client.apply()
 
         Parameters
         ----------
 
         indices_or_msg_ids : integer history index, str msg_id, or list of either
             The indices or msg_ids of indices to be retrieved
 
         block : bool
             Whether to wait for the result to be done
 
         Returns
         -------
 
         AsyncResult
             A single AsyncResult object will always be returned.
 
         AsyncHubResult
             A subclass of AsyncResult that retrieves results from the Hub
 
         """
         block = self.block if block is None else block
         if indices_or_msg_ids is None:
             indices_or_msg_ids = -1
 
         if not isinstance(indices_or_msg_ids, (list,tuple)):
             indices_or_msg_ids = [indices_or_msg_ids]
 
         theids = []
         for id in indices_or_msg_ids:
             if isinstance(id, int):
                 id = self.history[id]
             if not isinstance(id, basestring):
                 raise TypeError("indices must be str or int, not %r"%id)
             theids.append(id)
 
         local_ids = filter(lambda msg_id: msg_id in self.history or msg_id in self.results, theids)
         remote_ids = filter(lambda msg_id: msg_id not in local_ids, theids)
 
         if remote_ids:
             ar = AsyncHubResult(self, msg_ids=theids)
         else:
             ar = AsyncResult(self, msg_ids=theids)
 
         if block:
             ar.wait()
 
         return ar
 
     @spin_first
     def resubmit(self, indices_or_msg_ids=None, subheader=None, block=None):
         """Resubmit one or more tasks.
 
         in-flight tasks may not be resubmitted.
 
         Parameters
         ----------
 
         indices_or_msg_ids : integer history index, str msg_id, or list of either
             The indices or msg_ids of indices to be retrieved
 
         block : bool
             Whether to wait for the result to be done
 
         Returns
         -------
 
         AsyncHubResult
             A subclass of AsyncResult that retrieves results from the Hub
 
         """
         block = self.block if block is None else block
         if indices_or_msg_ids is None:
             indices_or_msg_ids = -1
 
         if not isinstance(indices_or_msg_ids, (list,tuple)):
             indices_or_msg_ids = [indices_or_msg_ids]
 
         theids = []
         for id in indices_or_msg_ids:
             if isinstance(id, int):
                 id = self.history[id]
             if not isinstance(id, basestring):
                 raise TypeError("indices must be str or int, not %r"%id)
             theids.append(id)
 
-        for msg_id in theids:
-            self.outstanding.discard(msg_id)
-            if msg_id in self.history:
-                self.history.remove(msg_id)
-            self.results.pop(msg_id, None)
-            self.metadata.pop(msg_id, None)
         content = dict(msg_ids = theids)
 
         self.session.send(self._query_socket, 'resubmit_request', content)
 
         zmq.select([self._query_socket], [], [])
         idents,msg = self.session.recv(self._query_socket, zmq.NOBLOCK)
         if self.debug:
             pprint(msg)
         content = msg['content']
         if content['status'] != 'ok':
             raise self._unwrap_exception(content)
+        mapping = content['resubmitted']
+        new_ids = [ mapping[msg_id] for msg_id in theids ]
 
-        ar = AsyncHubResult(self, msg_ids=theids)
+        ar = AsyncHubResult(self, msg_ids=new_ids)
 
         if block:
             ar.wait()
 
         return ar
 
     @spin_first
     def result_status(self, msg_ids, status_only=True):
         """Check on the status of the result(s) of the apply request with `msg_ids`.
 
         If status_only is False, then the actual results will be retrieved, else
         only the status of the results will be checked.
 
         Parameters
         ----------
 
         msg_ids : list of msg_ids
             if int:
                 Passed as index to self.history for convenience.
         status_only : bool (default: True)
             if False:
                 Retrieve the actual results of completed tasks.
 
         Returns
         -------
 
         results : dict
             There will always be the keys 'pending' and 'completed', which will
             be lists of msg_ids that are incomplete or complete. If `status_only`
             is False, then completed results will be keyed by their `msg_id`.
         """
         if not isinstance(msg_ids, (list,tuple)):
             msg_ids = [msg_ids]
 
         theids = []
         for msg_id in msg_ids:
             if isinstance(msg_id, int):
                 msg_id = self.history[msg_id]
             if not isinstance(msg_id, basestring):
                 raise TypeError("msg_ids must be str, not %r"%msg_id)
             theids.append(msg_id)
 
         completed = []
         local_results = {}
 
         # comment this block out to temporarily disable local shortcut:
         for msg_id in theids:
             if msg_id in self.results:
                 completed.append(msg_id)
                 local_results[msg_id] = self.results[msg_id]
                 theids.remove(msg_id)
 
         if theids: # some not locally cached
             content = dict(msg_ids=theids, status_only=status_only)
             msg = self.session.send(self._query_socket, "result_request", content=content)
             zmq.select([self._query_socket], [], [])
             idents,msg = self.session.recv(self._query_socket, zmq.NOBLOCK)
             if self.debug:
                 pprint(msg)
             content = msg['content']
             if content['status'] != 'ok':
                 raise self._unwrap_exception(content)
             buffers = msg['buffers']
         else:
             content = dict(completed=[],pending=[])
 
         content['completed'].extend(completed)
 
         if status_only:
             return content
 
         failures = []
         # load cached results into result:
         content.update(local_results)
 
         # update cache with results:
         for msg_id in sorted(theids):
             if msg_id in content['completed']:
                 rec = content[msg_id]
                 parent = rec['header']
                 header = rec['result_header']
                 rcontent = rec['result_content']
                 iodict = rec['io']
                 if isinstance(rcontent, str):
                     rcontent = self.session.unpack(rcontent)
 
                 md = self.metadata[msg_id]
                 md.update(self._extract_metadata(header, parent, rcontent))
                 if rec.get('received'):
                     md['received'] = rec['received']
                 md.update(iodict)
 
                 if rcontent['status'] == 'ok':
                     res,buffers = util.unserialize_object(buffers)
                 else:
                     print rcontent
                     res = self._unwrap_exception(rcontent)
                     failures.append(res)
 
                 self.results[msg_id] = res
                 content[msg_id] = res
 
         if len(theids) == 1 and failures:
                 raise failures[0]
 
         error.collect_exceptions(failures, "result_status")
         return content
 
     @spin_first
     def queue_status(self, targets='all', verbose=False):
         """Fetch the status of engine queues.
 
         Parameters
         ----------
 
         targets : int/str/list of ints/strs
                 the engines whose states are to be queried.
                 default : all
         verbose : bool
                 Whether to return lengths only, or lists of ids for each element
         """
         if targets == 'all':
             # allow 'all' to be evaluated on the engine
             engine_ids = None
         else:
             engine_ids = self._build_targets(targets)[1]
         content = dict(targets=engine_ids, verbose=verbose)
         self.session.send(self._query_socket, "queue_request", content=content)
         idents,msg = self.session.recv(self._query_socket, 0)
         if self.debug:
             pprint(msg)
         content = msg['content']
         status = content.pop('status')
         if status != 'ok':
             raise self._unwrap_exception(content)
         content = rekey(content)
         if isinstance(targets, int):
             return content[targets]
         else:
             return content
 
     @spin_first
     def purge_results(self, jobs=[], targets=[]):
         """Tell the Hub to forget results.
 
         Individual results can be purged by msg_id, or the entire
         history of specific targets can be purged.
 
         Use `purge_results('all')` to scrub everything from the Hub's db.
 
         Parameters
         ----------
 
         jobs : str or list of str or AsyncResult objects
                 the msg_ids whose results should be forgotten.
         targets : int/str/list of ints/strs
                 The targets, by int_id, whose entire history is to be purged.
 
                 default : None
         """
         if not targets and not jobs:
             raise ValueError("Must specify at least one of `targets` and `jobs`")
         if targets:
             targets = self._build_targets(targets)[1]
 
         # construct msg_ids from jobs
         if jobs == 'all':
             msg_ids = jobs
         else:
             msg_ids = []
             if isinstance(jobs, (basestring,AsyncResult)):
                 jobs = [jobs]
             bad_ids = filter(lambda obj: not isinstance(obj, (basestring, AsyncResult)), jobs)
             if bad_ids:
                 raise TypeError("Invalid msg_id type %r, expected str or AsyncResult"%bad_ids[0])
             for j in jobs:
                 if isinstance(j, AsyncResult):
                     msg_ids.extend(j.msg_ids)
                 else:
                     msg_ids.append(j)
 
         content = dict(engine_ids=targets, msg_ids=msg_ids)
         self.session.send(self._query_socket, "purge_request", content=content)
         idents, msg = self.session.recv(self._query_socket, 0)
         if self.debug:
             pprint(msg)
         content = msg['content']
         if content['status'] != 'ok':
             raise self._unwrap_exception(content)
 
     @spin_first
     def hub_history(self):
         """Get the Hub's history
 
         Just like the Client, the Hub has a history, which is a list of msg_ids.
         This will contain the history of all clients, and, depending on configuration,
         may contain history across multiple cluster sessions.
 
         Any msg_id returned here is a valid argument to `get_result`.
 
         Returns
         -------
 
         msg_ids : list of strs
                 list of all msg_ids, ordered by task submission time.
         """
 
         self.session.send(self._query_socket, "history_request", content={})
         idents, msg = self.session.recv(self._query_socket, 0)
 
         if self.debug:
             pprint(msg)
         content = msg['content']
         if content['status'] != 'ok':
             raise self._unwrap_exception(content)
         else:
             return content['history']
 
     @spin_first
     def db_query(self, query, keys=None):
         """Query the Hub's TaskRecord database
 
         This will return a list of task record dicts that match `query`
 
         Parameters
         ----------
 
         query : mongodb query dict
             The search dict. See mongodb query docs for details.
         keys : list of strs [optional]
             The subset of keys to be returned.  The default is to fetch everything but buffers.
             'msg_id' will *always* be included.
         """
         if isinstance(keys, basestring):
             keys = [keys]
         content = dict(query=query, keys=keys)
         self.session.send(self._query_socket, "db_request", content=content)
         idents, msg = self.session.recv(self._query_socket, 0)
         if self.debug:
             pprint(msg)
         content = msg['content']
         if content['status'] != 'ok':
             raise self._unwrap_exception(content)
 
         records = content['records']
 
         buffer_lens = content['buffer_lens']
         result_buffer_lens = content['result_buffer_lens']
         buffers = msg['buffers']
         has_bufs = buffer_lens is not None
         has_rbufs = result_buffer_lens is not None
         for i,rec in enumerate(records):
             # relink buffers
             if has_bufs:
                 blen = buffer_lens[i]
                 rec['buffers'], buffers = buffers[:blen],buffers[blen:]
             if has_rbufs:
                 blen = result_buffer_lens[i]
                 rec['result_buffers'], buffers = buffers[:blen],buffers[blen:]
 
         return records
 
 __all__ = [ 'Client' ]

IPython/parallel/client/view.py

@@ -1,1075 +1,1100 b''
 """Views of remote engines.
 
 Authors:
 
 * Min RK
 """
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2010-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 import imp
 import sys
 import warnings
 from contextlib import contextmanager
 from types import ModuleType
 
 import zmq
 
 from IPython.testing.skipdoctest import skip_doctest
 from IPython.utils.traitlets import (
     HasTraits, Any, Bool, List, Dict, Set, Instance, CFloat, Integer
 )
 from IPython.external.decorator import decorator
 
 from IPython.parallel import util
 from IPython.parallel.controller.dependency import Dependency, dependent
 
 from . import map as Map
 from .asyncresult import AsyncResult, AsyncMapResult
 from .remotefunction import ParallelFunction, parallel, remote, getname
 
 #-----------------------------------------------------------------------------
 # Decorators
 #-----------------------------------------------------------------------------
 
 @decorator
 def save_ids(f, self, *args, **kwargs):
     """Keep our history and outstanding attributes up to date after a method call."""
     n_previous = len(self.client.history)
     try:
         ret = f(self, *args, **kwargs)
     finally:
         nmsgs = len(self.client.history) - n_previous
         msg_ids = self.client.history[-nmsgs:]
         self.history.extend(msg_ids)
         map(self.outstanding.add, msg_ids)
     return ret
 
 @decorator
 def sync_results(f, self, *args, **kwargs):
     """sync relevant results from self.client to our results attribute."""
     ret = f(self, *args, **kwargs)
     delta = self.outstanding.difference(self.client.outstanding)
     completed = self.outstanding.intersection(delta)
     self.outstanding = self.outstanding.difference(completed)
     for msg_id in completed:
         self.results[msg_id] = self.client.results[msg_id]
     return ret
 
 @decorator
 def spin_after(f, self, *args, **kwargs):
     """call spin after the method."""
     ret = f(self, *args, **kwargs)
     self.spin()
     return ret
 
 #-----------------------------------------------------------------------------
 # Classes
 #-----------------------------------------------------------------------------
 
 @skip_doctest
 class View(HasTraits):
     """Base View class for more convenint apply(f,*args,**kwargs) syntax via attributes.
 
     Don't use this class, use subclasses.
 
     Methods
     -------
 
     spin
         flushes incoming results and registration state changes
         control methods spin, and requesting `ids` also ensures up to date
 
     wait
         wait on one or more msg_ids
 
     execution methods
         apply
         legacy: execute, run
 
     data movement
         push, pull, scatter, gather
 
     query methods
         get_result, queue_status, purge_results, result_status
 
     control methods
         abort, shutdown
 
     """
     # flags
     block=Bool(False)
     track=Bool(True)
     targets = Any()
 
     history=List()
     outstanding = Set()
     results = Dict()
     client = Instance('IPython.parallel.Client')
 
     _socket = Instance('zmq.Socket')
     _flag_names = List(['targets', 'block', 'track'])
     _targets = Any()
     _idents = Any()
 
     def __init__(self, client=None, socket=None, **flags):
         super(View, self).__init__(client=client, _socket=socket)
         self.block = client.block
 
         self.set_flags(**flags)
 
         assert not self.__class__ is View, "Don't use base View objects, use subclasses"
 
-
     def __repr__(self):
         strtargets = str(self.targets)
         if len(strtargets) > 16:
             strtargets = strtargets[:12]+'...]'
         return "<%s %s>"%(self.__class__.__name__, strtargets)
-
+    
+    def __len__(self):
+        if isinstance(self.targets, list):
+            return len(self.targets)
+        elif isinstance(self.targets, int):
+            return 1
+        else:
+            return len(self.client)
+    
     def set_flags(self, **kwargs):
         """set my attribute flags by keyword.
 
         Views determine behavior with a few attributes (`block`, `track`, etc.).
         These attributes can be set all at once by name with this method.
 
         Parameters
         ----------
 
         block : bool
             whether to wait for results
         track : bool
             whether to create a MessageTracker to allow the user to
             safely edit after arrays and buffers during non-copying
             sends.
         """
         for name, value in kwargs.iteritems():
             if name not in self._flag_names:
                 raise KeyError("Invalid name: %r"%name)
             else:
                 setattr(self, name, value)
 
     @contextmanager
     def temp_flags(self, **kwargs):
         """temporarily set flags, for use in `with` statements.
 
         See set_flags for permanent setting of flags
 
         Examples
         --------
 
         >>> view.track=False
         ...
         >>> with view.temp_flags(track=True):
         ...    ar = view.apply(dostuff, my_big_array)
         ...    ar.tracker.wait() # wait for send to finish
         >>> view.track
         False
 
         """
         # preflight: save flags, and set temporaries
         saved_flags = {}
         for f in self._flag_names:
             saved_flags[f] = getattr(self, f)
         self.set_flags(**kwargs)
         # yield to the with-statement block
         try:
             yield
         finally:
             # postflight: restore saved flags
             self.set_flags(**saved_flags)
 
 
     #----------------------------------------------------------------
     # apply
     #----------------------------------------------------------------
 
     @sync_results
     @save_ids
     def _really_apply(self, f, args, kwargs, block=None, **options):
-        """wrapper for client.send_apply_message"""
+        """wrapper for client.send_apply_request"""
         raise NotImplementedError("Implement in subclasses")
 
     def apply(self, f, *args, **kwargs):
         """calls f(*args, **kwargs) on remote engines, returning the result.
 
         This method sets all apply flags via this View's attributes.
 
         if self.block is False:
             returns AsyncResult
         else:
             returns actual result of f(*args, **kwargs)
         """
         return self._really_apply(f, args, kwargs)
 
     def apply_async(self, f, *args, **kwargs):
         """calls f(*args, **kwargs) on remote engines in a nonblocking manner.
 
         returns AsyncResult
         """
         return self._really_apply(f, args, kwargs, block=False)
 
     @spin_after
     def apply_sync(self, f, *args, **kwargs):
         """calls f(*args, **kwargs) on remote engines in a blocking manner,
          returning the result.
 
         returns: actual result of f(*args, **kwargs)
         """
         return self._really_apply(f, args, kwargs, block=True)
 
     #----------------------------------------------------------------
     # wrappers for client and control methods
     #----------------------------------------------------------------
     @sync_results
     def spin(self):
         """spin the client, and sync"""
         self.client.spin()
 
     @sync_results
     def wait(self, jobs=None, timeout=-1):
         """waits on one or more `jobs`, for up to `timeout` seconds.
 
         Parameters
         ----------
 
         jobs : int, str, or list of ints and/or strs, or one or more AsyncResult objects
                 ints are indices to self.history
                 strs are msg_ids
                 default: wait on all outstanding messages
         timeout : float
                 a time in seconds, after which to give up.
                 default is -1, which means no timeout
 
         Returns
         -------
 
         True : when all msg_ids are done
         False : timeout reached, some msg_ids still outstanding
         """
         if jobs is None:
             jobs = self.history
         return self.client.wait(jobs, timeout)
 
     def abort(self, jobs=None, targets=None, block=None):
         """Abort jobs on my engines.
 
         Parameters
         ----------
 
         jobs : None, str, list of strs, optional
             if None: abort all jobs.
             else: abort specific msg_id(s).
         """
         block = block if block is not None else self.block
         targets = targets if targets is not None else self.targets
         jobs = jobs if jobs is not None else list(self.outstanding)
         
         return self.client.abort(jobs=jobs, targets=targets, block=block)
 
     def queue_status(self, targets=None, verbose=False):
         """Fetch the Queue status of my engines"""
         targets = targets if targets is not None else self.targets
         return self.client.queue_status(targets=targets, verbose=verbose)
 
     def purge_results(self, jobs=[], targets=[]):
         """Instruct the controller to forget specific results."""
         if targets is None or targets == 'all':
             targets = self.targets
         return self.client.purge_results(jobs=jobs, targets=targets)
 
     def shutdown(self, targets=None, restart=False, hub=False, block=None):
         """Terminates one or more engine processes, optionally including the hub.
         """
         block = self.block if block is None else block
         if targets is None or targets == 'all':
             targets = self.targets
         return self.client.shutdown(targets=targets, restart=restart, hub=hub, block=block)
 
     @spin_after
     def get_result(self, indices_or_msg_ids=None):
         """return one or more results, specified by history index or msg_id.
 
         See client.get_result for details.
 
         """
 
         if indices_or_msg_ids is None:
             indices_or_msg_ids = -1
         if isinstance(indices_or_msg_ids, int):
             indices_or_msg_ids = self.history[indices_or_msg_ids]
         elif isinstance(indices_or_msg_ids, (list,tuple,set)):
             indices_or_msg_ids = list(indices_or_msg_ids)
             for i,index in enumerate(indices_or_msg_ids):
                 if isinstance(index, int):
                     indices_or_msg_ids[i] = self.history[index]
         return self.client.get_result(indices_or_msg_ids)
 
     #-------------------------------------------------------------------
     # Map
     #-------------------------------------------------------------------
 
     def map(self, f, *sequences, **kwargs):
         """override in subclasses"""
         raise NotImplementedError
 
     def map_async(self, f, *sequences, **kwargs):
         """Parallel version of builtin `map`, using this view's engines.
 
         This is equivalent to map(...block=False)
 
         See `self.map` for details.
         """
         if 'block' in kwargs:
             raise TypeError("map_async doesn't take a `block` keyword argument.")
         kwargs['block'] = False
         return self.map(f,*sequences,**kwargs)
 
     def map_sync(self, f, *sequences, **kwargs):
         """Parallel version of builtin `map`, using this view's engines.
 
         This is equivalent to map(...block=True)
 
         See `self.map` for details.
         """
         if 'block' in kwargs:
             raise TypeError("map_sync doesn't take a `block` keyword argument.")
         kwargs['block'] = True
         return self.map(f,*sequences,**kwargs)
 
     def imap(self, f, *sequences, **kwargs):
         """Parallel version of `itertools.imap`.
 
         See `self.map` for details.
 
         """
 
         return iter(self.map_async(f,*sequences, **kwargs))
 
     #-------------------------------------------------------------------
     # Decorators
     #-------------------------------------------------------------------
 
     def remote(self, block=True, **flags):
         """Decorator for making a RemoteFunction"""
         block = self.block if block is None else block
         return remote(self, block=block, **flags)
 
     def parallel(self, dist='b', block=None, **flags):
         """Decorator for making a ParallelFunction"""
         block = self.block if block is None else block
         return parallel(self, dist=dist, block=block, **flags)
 
 @skip_doctest
 class DirectView(View):
     """Direct Multiplexer View of one or more engines.
 
     These are created via indexed access to a client:
 
     >>> dv_1 = client[1]
     >>> dv_all = client[:]
     >>> dv_even = client[::2]
     >>> dv_some = client[1:3]
 
     This object provides dictionary access to engine namespaces:
 
     # push a=5:
     >>> dv['a'] = 5
     # pull 'foo':
     >>> db['foo']
 
     """
 
     def __init__(self, client=None, socket=None, targets=None):
         super(DirectView, self).__init__(client=client, socket=socket, targets=targets)
 
     @property
     def importer(self):
         """sync_imports(local=True) as a property.
 
         See sync_imports for details.
 
         """
         return self.sync_imports(True)
 
     @contextmanager
     def sync_imports(self, local=True, quiet=False):
         """Context Manager for performing simultaneous local and remote imports.
 
         'import x as y' will *not* work.  The 'as y' part will simply be ignored.
         
         If `local=True`, then the package will also be imported locally.
         
         If `quiet=True`, no output will be produced when attempting remote 
         imports. 
         
         Note that remote-only (`local=False`) imports have not been implemented.
         
         >>> with view.sync_imports():
         ...    from numpy import recarray
         importing recarray from numpy on engine(s)
 
         """
         import __builtin__
         local_import = __builtin__.__import__
         modules = set()
         results = []
         @util.interactive
         def remote_import(name, fromlist, level):
             """the function to be passed to apply, that actually performs the import
             on the engine, and loads up the user namespace.
             """
             import sys
             user_ns = globals()
             mod = __import__(name, fromlist=fromlist, level=level)
             if fromlist:
                 for key in fromlist:
                     user_ns[key] = getattr(mod, key)
             else:
                 user_ns[name] = sys.modules[name]
 
         def view_import(name, globals={}, locals={}, fromlist=[], level=-1):
             """the drop-in replacement for __import__, that optionally imports
             locally as well.
             """
             # don't override nested imports
             save_import = __builtin__.__import__
             __builtin__.__import__ = local_import
 
             if imp.lock_held():
                 # this is a side-effect import, don't do it remotely, or even
                 # ignore the local effects
                 return local_import(name, globals, locals, fromlist, level)
 
             imp.acquire_lock()
             if local:
                 mod = local_import(name, globals, locals, fromlist, level)
             else:
                 raise NotImplementedError("remote-only imports not yet implemented")
             imp.release_lock()
 
             key = name+':'+','.join(fromlist or [])
             if level == -1 and key not in modules:
                 modules.add(key)
                 if not quiet:
                     if fromlist:
                         print "importing %s from %s on engine(s)"%(','.join(fromlist), name)
                     else:
                         print "importing %s on engine(s)"%name
                 results.append(self.apply_async(remote_import, name, fromlist, level))
             # restore override
             __builtin__.__import__ = save_import
 
             return mod
 
         # override __import__
         __builtin__.__import__ = view_import
         try:
             # enter the block
             yield
         except ImportError:
             if local:
                 raise
             else:
                 # ignore import errors if not doing local imports
                 pass
         finally:
             # always restore __import__
             __builtin__.__import__ = local_import
 
         for r in results:
             # raise possible remote ImportErrors here
             r.get()
 
 
     @sync_results
     @save_ids
     def _really_apply(self, f, args=None, kwargs=None, targets=None, block=None, track=None):
         """calls f(*args, **kwargs) on remote engines, returning the result.
 
         This method sets all of `apply`'s flags via this View's attributes.
 
         Parameters
         ----------
 
         f : callable
 
         args : list [default: empty]
 
         kwargs : dict [default: empty]
 
         targets : target list [default: self.targets]
             where to run
         block : bool [default: self.block]
             whether to block
         track : bool [default: self.track]
             whether to ask zmq to track the message, for safe non-copying sends
 
         Returns
         -------
 
         if self.block is False:
             returns AsyncResult
         else:
             returns actual result of f(*args, **kwargs) on the engine(s)
             This will be a list of self.targets is also a list (even length 1), or
             the single result if self.targets is an integer engine id
         """
         args = [] if args is None else args
         kwargs = {} if kwargs is None else kwargs
         block = self.block if block is None else block
         track = self.track if track is None else track
         targets = self.targets if targets is None else targets
 
         _idents = self.client._build_targets(targets)[0]
         msg_ids = []
         trackers = []
         for ident in _idents:
-            msg = self.client.send_apply_message(self._socket, f, args, kwargs, track=track,
+            msg = self.client.send_apply_request(self._socket, f, args, kwargs, track=track,
                                     ident=ident)
             if track:
                 trackers.append(msg['tracker'])
             msg_ids.append(msg['header']['msg_id'])
         tracker = None if track is False else zmq.MessageTracker(*trackers)
         ar = AsyncResult(self.client, msg_ids, fname=getname(f), targets=targets, tracker=tracker)
         if block:
             try:
                 return ar.get()
             except KeyboardInterrupt:
                 pass
         return ar
 
+
     @spin_after
     def map(self, f, *sequences, **kwargs):
         """view.map(f, *sequences, block=self.block) => list|AsyncMapResult
 
         Parallel version of builtin `map`, using this View's `targets`.
 
         There will be one task per target, so work will be chunked
         if the sequences are longer than `targets`.
 
         Results can be iterated as they are ready, but will become available in chunks.
 
         Parameters
         ----------
 
         f : callable
             function to be mapped
         *sequences: one or more sequences of matching length
             the sequences to be distributed and passed to `f`
         block : bool
             whether to wait for the result or not [default self.block]
 
         Returns
         -------
 
         if block=False:
             AsyncMapResult
                 An object like AsyncResult, but which reassembles the sequence of results
                 into a single list. AsyncMapResults can be iterated through before all
                 results are complete.
         else:
             list
                 the result of map(f,*sequences)
         """
 
         block = kwargs.pop('block', self.block)
         for k in kwargs.keys():
             if k not in ['block', 'track']:
                 raise TypeError("invalid keyword arg, %r"%k)
 
         assert len(sequences) > 0, "must have some sequences to map onto!"
         pf = ParallelFunction(self, f, block=block, **kwargs)
         return pf.map(*sequences)
 
-    def execute(self, code, targets=None, block=None):
+    @sync_results
+    @save_ids
+    def execute(self, code, silent=True, targets=None, block=None):
         """Executes `code` on `targets` in blocking or nonblocking manner.
 
         ``execute`` is always `bound` (affects engine namespace)
 
         Parameters
         ----------
 
         code : str
                 the code string to be executed
         block : bool
                 whether or not to wait until done to return
                 default: self.block
         """
-        return self._really_apply(util._execute, args=(code,), block=block, targets=targets)
+        block = self.block if block is None else block
+        targets = self.targets if targets is None else targets
+
+        _idents = self.client._build_targets(targets)[0]
+        msg_ids = []
+        trackers = []
+        for ident in _idents:
+            msg = self.client.send_execute_request(self._socket, code, silent=silent, ident=ident)
+            msg_ids.append(msg['header']['msg_id'])
+        ar = AsyncResult(self.client, msg_ids, fname='execute', targets=targets)
+        if block:
+            try:
+                ar.get()
+            except KeyboardInterrupt:
+                pass
+        return ar
 
     def run(self, filename, targets=None, block=None):
         """Execute contents of `filename` on my engine(s).
 
         This simply reads the contents of the file and calls `execute`.
 
         Parameters
         ----------
 
         filename : str
                 The path to the file
         targets : int/str/list of ints/strs
                 the engines on which to execute
                 default : all
         block : bool
                 whether or not to wait until done
                 default: self.block
 
         """
         with open(filename, 'r') as f:
             # add newline in case of trailing indented whitespace
             # which will cause SyntaxError
             code = f.read()+'\n'
         return self.execute(code, block=block, targets=targets)
 
     def update(self, ns):
         """update remote namespace with dict `ns`
 
         See `push` for details.
         """
         return self.push(ns, block=self.block, track=self.track)
 
     def push(self, ns, targets=None, block=None, track=None):
         """update remote namespace with dict `ns`
 
         Parameters
         ----------
 
         ns : dict
             dict of keys with which to update engine namespace(s)
         block : bool [default : self.block]
             whether to wait to be notified of engine receipt
 
         """
 
         block = block if block is not None else self.block
         track = track if track is not None else self.track
         targets = targets if targets is not None else self.targets
         # applier = self.apply_sync if block else self.apply_async
         if not isinstance(ns, dict):
             raise TypeError("Must be a dict, not %s"%type(ns))
         return self._really_apply(util._push, kwargs=ns, block=block, track=track, targets=targets)
 
     def get(self, key_s):
         """get object(s) by `key_s` from remote namespace
 
         see `pull` for details.
         """
         # block = block if block is not None else self.block
         return self.pull(key_s, block=True)
 
     def pull(self, names, targets=None, block=None):
         """get object(s) by `name` from remote namespace
 
         will return one object if it is a key.
         can also take a list of keys, in which case it will return a list of objects.
         """
         block = block if block is not None else self.block
         targets = targets if targets is not None else self.targets
         applier = self.apply_sync if block else self.apply_async
         if isinstance(names, basestring):
             pass
         elif isinstance(names, (list,tuple,set)):
             for key in names:
                 if not isinstance(key, basestring):
                     raise TypeError("keys must be str, not type %r"%type(key))
         else:
             raise TypeError("names must be strs, not %r"%names)
         return self._really_apply(util._pull, (names,), block=block, targets=targets)
 
     def scatter(self, key, seq, dist='b', flatten=False, targets=None, block=None, track=None):
         """
         Partition a Python sequence and send the partitions to a set of engines.
         """
         block = block if block is not None else self.block
         track = track if track is not None else self.track
         targets = targets if targets is not None else self.targets
         
         # construct integer ID list:
         targets = self.client._build_targets(targets)[1]
 
         mapObject = Map.dists[dist]()
         nparts = len(targets)
         msg_ids = []
         trackers = []
         for index, engineid in enumerate(targets):
             partition = mapObject.getPartition(seq, index, nparts)
             if flatten and len(partition) == 1:
                 ns = {key: partition[0]}
             else:
                 ns = {key: partition}
             r = self.push(ns, block=False, track=track, targets=engineid)
             msg_ids.extend(r.msg_ids)
             if track:
                 trackers.append(r._tracker)
 
         if track:
             tracker = zmq.MessageTracker(*trackers)
         else:
             tracker = None
 
         r = AsyncResult(self.client, msg_ids, fname='scatter', targets=targets, tracker=tracker)
         if block:
             r.wait()
         else:
             return r
 
     @sync_results
     @save_ids
     def gather(self, key, dist='b', targets=None, block=None):
         """
         Gather a partitioned sequence on a set of engines as a single local seq.
         """
         block = block if block is not None else self.block
         targets = targets if targets is not None else self.targets
         mapObject = Map.dists[dist]()
         msg_ids = []
 
         # construct integer ID list:
         targets = self.client._build_targets(targets)[1]
         
         for index, engineid in enumerate(targets):
             msg_ids.extend(self.pull(key, block=False, targets=engineid).msg_ids)
 
         r = AsyncMapResult(self.client, msg_ids, mapObject, fname='gather')
 
         if block:
             try:
                 return r.get()
             except KeyboardInterrupt:
                 pass
         return r
 
     def __getitem__(self, key):
         return self.get(key)
 
     def __setitem__(self,key, value):
         self.update({key:value})
 
     def clear(self, targets=None, block=False):
         """Clear the remote namespaces on my engines."""
         block = block if block is not None else self.block
         targets = targets if targets is not None else self.targets
         return self.client.clear(targets=targets, block=block)
 
     def kill(self, targets=None, block=True):
         """Kill my engines."""
         block = block if block is not None else self.block
         targets = targets if targets is not None else self.targets
         return self.client.kill(targets=targets, block=block)
 
     #----------------------------------------
     # activate for %px,%autopx magics
     #----------------------------------------
     def activate(self):
         """Make this `View` active for parallel magic commands.
 
         IPython has a magic command syntax to work with `MultiEngineClient` objects.
         In a given IPython session there is a single active one.  While
         there can be many `Views` created and used by the user,
         there is only one active one.  The active `View` is used whenever
         the magic commands %px and %autopx are used.
 
         The activate() method is called on a given `View` to make it
         active.  Once this has been done, the magic commands can be used.
         """
 
         try:
             # This is injected into __builtins__.
             ip = get_ipython()
         except NameError:
             print "The IPython parallel magics (%result, %px, %autopx) only work within IPython."
         else:
             pmagic = ip.plugin_manager.get_plugin('parallelmagic')
             if pmagic is None:
                 ip.magic_load_ext('parallelmagic')
                 pmagic = ip.plugin_manager.get_plugin('parallelmagic')
 
             pmagic.active_view = self
 
 
 @skip_doctest
 class LoadBalancedView(View):
     """An load-balancing View that only executes via the Task scheduler.
 
     Load-balanced views can be created with the client's `view` method:
 
     >>> v = client.load_balanced_view()
 
     or targets can be specified, to restrict the potential destinations:
 
     >>> v = client.client.load_balanced_view([1,3])
 
     which would restrict loadbalancing to between engines 1 and 3.
 
     """
 
     follow=Any()
     after=Any()
     timeout=CFloat()
     retries = Integer(0)
 
     _task_scheme = Any()
     _flag_names = List(['targets', 'block', 'track', 'follow', 'after', 'timeout', 'retries'])
 
     def __init__(self, client=None, socket=None, **flags):
         super(LoadBalancedView, self).__init__(client=client, socket=socket, **flags)
         self._task_scheme=client._task_scheme
 
     def _validate_dependency(self, dep):
         """validate a dependency.
 
         For use in `set_flags`.
         """
         if dep is None or isinstance(dep, (basestring, AsyncResult, Dependency)):
             return True
         elif isinstance(dep, (list,set, tuple)):
             for d in dep:
                 if not isinstance(d, (basestring, AsyncResult)):
                     return False
         elif isinstance(dep, dict):
             if set(dep.keys()) != set(Dependency().as_dict().keys()):
                 return False
             if not isinstance(dep['msg_ids'], list):
                 return False
             for d in dep['msg_ids']:
                 if not isinstance(d, basestring):
                     return False
         else:
             return False
 
         return True
 
     def _render_dependency(self, dep):
         """helper for building jsonable dependencies from various input forms."""
         if isinstance(dep, Dependency):
             return dep.as_dict()
         elif isinstance(dep, AsyncResult):
             return dep.msg_ids
         elif dep is None:
             return []
         else:
             # pass to Dependency constructor
             return list(Dependency(dep))
 
     def set_flags(self, **kwargs):
         """set my attribute flags by keyword.
 
         A View is a wrapper for the Client's apply method, but with attributes
         that specify keyword arguments, those attributes can be set by keyword
         argument with this method.
 
         Parameters
         ----------
 
         block : bool
             whether to wait for results
         track : bool
             whether to create a MessageTracker to allow the user to
             safely edit after arrays and buffers during non-copying
             sends.
 
         after : Dependency or collection of msg_ids
             Only for load-balanced execution (targets=None)
             Specify a list of msg_ids as a time-based dependency.
             This job will only be run *after* the dependencies
             have been met.
 
         follow : Dependency or collection of msg_ids
             Only for load-balanced execution (targets=None)
             Specify a list of msg_ids as a location-based dependency.
             This job will only be run on an engine where this dependency
             is met.
 
         timeout : float/int or None
             Only for load-balanced execution (targets=None)
             Specify an amount of time (in seconds) for the scheduler to
             wait for dependencies to be met before failing with a
             DependencyTimeout.
 
         retries : int
             Number of times a task will be retried on failure.
         """
 
         super(LoadBalancedView, self).set_flags(**kwargs)
         for name in ('follow', 'after'):
             if name in kwargs:
                 value = kwargs[name]
                 if self._validate_dependency(value):
                     setattr(self, name, value)
                 else:
                     raise ValueError("Invalid dependency: %r"%value)
         if 'timeout' in kwargs:
             t = kwargs['timeout']
             if not isinstance(t, (int, long, float, type(None))):
                 raise TypeError("Invalid type for timeout: %r"%type(t))
             if t is not None:
                 if t < 0:
                     raise ValueError("Invalid timeout: %s"%t)
             self.timeout = t
 
     @sync_results
     @save_ids
     def _really_apply(self, f, args=None, kwargs=None, block=None, track=None,
                                         after=None, follow=None, timeout=None,
                                         targets=None, retries=None):
         """calls f(*args, **kwargs) on a remote engine, returning the result.
 
         This method temporarily sets all of `apply`'s flags for a single call.
 
         Parameters
         ----------
 
         f : callable
 
         args : list [default: empty]
 
         kwargs : dict [default: empty]
 
         block : bool [default: self.block]
             whether to block
         track : bool [default: self.track]
             whether to ask zmq to track the message, for safe non-copying sends
 
         !!!!!! TODO: THE REST HERE  !!!!
 
         Returns
         -------
 
         if self.block is False:
             returns AsyncResult
         else:
             returns actual result of f(*args, **kwargs) on the engine(s)
             This will be a list of self.targets is also a list (even length 1), or
             the single result if self.targets is an integer engine id
         """
 
         # validate whether we can run
         if self._socket.closed:
             msg = "Task farming is disabled"
             if self._task_scheme == 'pure':
                 msg += " because the pure ZMQ scheduler cannot handle"
                 msg += " disappearing engines."
             raise RuntimeError(msg)
 
         if self._task_scheme == 'pure':
             # pure zmq scheme doesn't support extra features
             msg = "Pure ZMQ scheduler doesn't support the following flags:"
             "follow, after, retries, targets, timeout"
             if (follow or after or retries or targets or timeout):
                 # hard fail on Scheduler flags
                 raise RuntimeError(msg)
             if isinstance(f, dependent):
                 # soft warn on functional dependencies
                 warnings.warn(msg, RuntimeWarning)
 
         # build args
         args = [] if args is None else args
         kwargs = {} if kwargs is None else kwargs
         block = self.block if block is None else block
         track = self.track if track is None else track
         after = self.after if after is None else after
         retries = self.retries if retries is None else retries
         follow = self.follow if follow is None else follow
         timeout = self.timeout if timeout is None else timeout
         targets = self.targets if targets is None else targets
 
         if not isinstance(retries, int):
             raise TypeError('retries must be int, not %r'%type(retries))
 
         if targets is None:
             idents = []
         else:
             idents = self.client._build_targets(targets)[0]
             # ensure *not* bytes
             idents = [ ident.decode() for ident in idents ]
 
         after = self._render_dependency(after)
         follow = self._render_dependency(follow)
         subheader = dict(after=after, follow=follow, timeout=timeout, targets=idents, retries=retries)
 
-        msg = self.client.send_apply_message(self._socket, f, args, kwargs, track=track,
+        msg = self.client.send_apply_request(self._socket, f, args, kwargs, track=track,
                                 subheader=subheader)
         tracker = None if track is False else msg['tracker']
 
         ar = AsyncResult(self.client, msg['header']['msg_id'], fname=getname(f), targets=None, tracker=tracker)
 
         if block:
             try:
                 return ar.get()
             except KeyboardInterrupt:
                 pass
         return ar
 
     @spin_after
     @save_ids
     def map(self, f, *sequences, **kwargs):
         """view.map(f, *sequences, block=self.block, chunksize=1, ordered=True) => list|AsyncMapResult
 
         Parallel version of builtin `map`, load-balanced by this View.
 
         `block`, and `chunksize` can be specified by keyword only.
 
         Each `chunksize` elements will be a separate task, and will be
         load-balanced. This lets individual elements be available for iteration
         as soon as they arrive.
 
         Parameters
         ----------
 
         f : callable
             function to be mapped
         *sequences: one or more sequences of matching length
             the sequences to be distributed and passed to `f`
         block : bool [default self.block]
             whether to wait for the result or not
         track : bool
             whether to create a MessageTracker to allow the user to
             safely edit after arrays and buffers during non-copying
             sends.
         chunksize : int [default 1]
             how many elements should be in each task.
         ordered : bool [default True]
             Whether the results should be gathered as they arrive, or enforce
             the order of submission.
             
             Only applies when iterating through AsyncMapResult as results arrive.
             Has no effect when block=True.
 
         Returns
         -------
 
         if block=False:
             AsyncMapResult
                 An object like AsyncResult, but which reassembles the sequence of results
                 into a single list. AsyncMapResults can be iterated through before all
                 results are complete.
             else:
                 the result of map(f,*sequences)
 
         """
 
         # default
         block = kwargs.get('block', self.block)
         chunksize = kwargs.get('chunksize', 1)
         ordered = kwargs.get('ordered', True)
 
         keyset = set(kwargs.keys())
         extra_keys = keyset.difference_update(set(['block', 'chunksize']))
         if extra_keys:
             raise TypeError("Invalid kwargs: %s"%list(extra_keys))
 
         assert len(sequences) > 0, "must have some sequences to map onto!"
 
         pf = ParallelFunction(self, f, block=block, chunksize=chunksize, ordered=ordered)
         return pf.map(*sequences)
 
 __all__ = ['LoadBalancedView', 'DirectView']

IPython/parallel/controller/heartmonitor.py

@@ -1,181 +1,182 b''
 #!/usr/bin/env python
 """
 A multi-heart Heartbeat system using PUB and XREP sockets. pings are sent out on the PUB,
 and hearts are tracked based on their XREQ identities.
 
 Authors:
 
 * Min RK
 """
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2010-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 from __future__ import print_function
 import time
 import uuid
 
 import zmq
 from zmq.devices import ThreadDevice
 from zmq.eventloop import ioloop, zmqstream
 
 from IPython.config.configurable import LoggingConfigurable
+from IPython.utils.py3compat import str_to_bytes
 from IPython.utils.traitlets import Set, Instance, CFloat, Integer
 
-from IPython.parallel.util import asbytes, log_errors
+from IPython.parallel.util import log_errors
 
 class Heart(object):
     """A basic heart object for responding to a HeartMonitor.
     This is a simple wrapper with defaults for the most common
     Device model for responding to heartbeats.
 
     It simply builds a threadsafe zmq.FORWARDER Device, defaulting to using
     SUB/XREQ for in/out.
 
     You can specify the XREQ's IDENTITY via the optional heart_id argument."""
     device=None
     id=None
     def __init__(self, in_addr, out_addr, in_type=zmq.SUB, out_type=zmq.DEALER, heart_id=None):
         self.device = ThreadDevice(zmq.FORWARDER, in_type, out_type)
         # do not allow the device to share global Context.instance,
         # which is the default behavior in pyzmq > 2.1.10
         self.device.context_factory = zmq.Context
         
         self.device.daemon=True
         self.device.connect_in(in_addr)
         self.device.connect_out(out_addr)
         if in_type == zmq.SUB:
             self.device.setsockopt_in(zmq.SUBSCRIBE, b"")
         if heart_id is None:
             heart_id = uuid.uuid4().bytes
         self.device.setsockopt_out(zmq.IDENTITY, heart_id)
         self.id = heart_id
 
     def start(self):
         return self.device.start()
 
 
 class HeartMonitor(LoggingConfigurable):
     """A basic HeartMonitor class
     pingstream: a PUB stream
     pongstream: an XREP stream
     period: the period of the heartbeat in milliseconds"""
 
     period = Integer(3000, config=True,
         help='The frequency at which the Hub pings the engines for heartbeats '
         '(in ms)',
     )
 
     pingstream=Instance('zmq.eventloop.zmqstream.ZMQStream')
     pongstream=Instance('zmq.eventloop.zmqstream.ZMQStream')
     loop = Instance('zmq.eventloop.ioloop.IOLoop')
     def _loop_default(self):
         return ioloop.IOLoop.instance()
 
     # not settable:
     hearts=Set()
     responses=Set()
     on_probation=Set()
     last_ping=CFloat(0)
     _new_handlers = Set()
     _failure_handlers = Set()
     lifetime = CFloat(0)
     tic = CFloat(0)
 
     def __init__(self, **kwargs):
         super(HeartMonitor, self).__init__(**kwargs)
 
         self.pongstream.on_recv(self.handle_pong)
 
     def start(self):
         self.tic = time.time()
         self.caller = ioloop.PeriodicCallback(self.beat, self.period, self.loop)
         self.caller.start()
 
     def add_new_heart_handler(self, handler):
         """add a new handler for new hearts"""
         self.log.debug("heartbeat::new_heart_handler: %s", handler)
         self._new_handlers.add(handler)
 
     def add_heart_failure_handler(self, handler):
         """add a new handler for heart failure"""
         self.log.debug("heartbeat::new heart failure handler: %s", handler)
         self._failure_handlers.add(handler)
 
     def beat(self):
         self.pongstream.flush()
         self.last_ping = self.lifetime
 
         toc = time.time()
         self.lifetime += toc-self.tic
         self.tic = toc
         self.log.debug("heartbeat::sending %s", self.lifetime)
         goodhearts = self.hearts.intersection(self.responses)
         missed_beats = self.hearts.difference(goodhearts)
         heartfailures = self.on_probation.intersection(missed_beats)
         newhearts = self.responses.difference(goodhearts)
         map(self.handle_new_heart, newhearts)
         map(self.handle_heart_failure, heartfailures)
         self.on_probation = missed_beats.intersection(self.hearts)
         self.responses = set()
         # print self.on_probation, self.hearts
         # self.log.debug("heartbeat::beat %.3f, %i beating hearts", self.lifetime, len(self.hearts))
-        self.pingstream.send(asbytes(str(self.lifetime)))
+        self.pingstream.send(str_to_bytes(str(self.lifetime)))
         # flush stream to force immediate socket send
         self.pingstream.flush()
 
     def handle_new_heart(self, heart):
         if self._new_handlers:
             for handler in self._new_handlers:
                 handler(heart)
         else:
             self.log.info("heartbeat::yay, got new heart %s!", heart)
         self.hearts.add(heart)
 
     def handle_heart_failure(self, heart):
         if self._failure_handlers:
             for handler in self._failure_handlers:
                 try:
                     handler(heart)
                 except Exception as e:
                     self.log.error("heartbeat::Bad Handler! %s", handler, exc_info=True)
                     pass
         else:
             self.log.info("heartbeat::Heart %s failed :(", heart)
         self.hearts.remove(heart)
 
 
     @log_errors
     def handle_pong(self, msg):
         "a heart just beat"
-        current = asbytes(str(self.lifetime))
-        last = asbytes(str(self.last_ping))
+        current = str_to_bytes(str(self.lifetime))
+        last = str_to_bytes(str(self.last_ping))
         if msg[1] == current:
             delta = time.time()-self.tic
             # self.log.debug("heartbeat::heart %r took %.2f ms to respond"%(msg[0], 1000*delta))
             self.responses.add(msg[0])
         elif msg[1] == last:
             delta = time.time()-self.tic + (self.lifetime-self.last_ping)
             self.log.warn("heartbeat::heart %r missed a beat, and took %.2f ms to respond", msg[0], 1000*delta)
             self.responses.add(msg[0])
         else:
             self.log.warn("heartbeat::got bad heartbeat (possibly old?): %s (current=%.3f)", msg[1], self.lifetime)
 
 
 if __name__ == '__main__':
     loop = ioloop.IOLoop.instance()
     context = zmq.Context()
     pub = context.socket(zmq.PUB)
     pub.bind('tcp://127.0.0.1:5555')
     xrep = context.socket(zmq.ROUTER)
     xrep.bind('tcp://127.0.0.1:5556')
 
     outstream = zmqstream.ZMQStream(pub, loop)
     instream = zmqstream.ZMQStream(xrep, loop)
 
     hb = HeartMonitor(loop, outstream, instream)
 
     loop.start()

IPython/parallel/controller/hub.py

@@ -1,1300 +1,1310 b''
 """The IPython Controller Hub with 0MQ
 This is the master object that handles connections from engines and clients,
 and monitors traffic through the various queues.
 
 Authors:
 
 * Min RK
 """
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2010-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 from __future__ import print_function
 
 import sys
 import time
 from datetime import datetime
 
 import zmq
 from zmq.eventloop import ioloop
 from zmq.eventloop.zmqstream import ZMQStream
 
 # internal:
 from IPython.utils.importstring import import_item
+from IPython.utils.py3compat import cast_bytes
 from IPython.utils.traitlets import (
         HasTraits, Instance, Integer, Unicode, Dict, Set, Tuple, CBytes, DottedObjectName
         )
 
 from IPython.parallel import error, util
 from IPython.parallel.factory import RegistrationFactory
 
 from IPython.zmq.session import SessionFactory
 
 from .heartmonitor import HeartMonitor
 
 #-----------------------------------------------------------------------------
 # Code
 #-----------------------------------------------------------------------------
 
 def _passer(*args, **kwargs):
     return
 
 def _printer(*args, **kwargs):
     print (args)
     print (kwargs)
 
 def empty_record():
     """Return an empty dict with all record keys."""
     return {
         'msg_id' : None,
         'header' : None,
         'content': None,
         'buffers': None,
         'submitted': None,
         'client_uuid' : None,
         'engine_uuid' : None,
         'started': None,
         'completed': None,
         'resubmitted': None,
         'received': None,
         'result_header' : None,
         'result_content' : None,
         'result_buffers' : None,
         'queue' : None,
         'pyin' : None,
         'pyout': None,
         'pyerr': None,
         'stdout': '',
         'stderr': '',
     }
 
 def init_record(msg):
     """Initialize a TaskRecord based on a request."""
     header = msg['header']
     return {
         'msg_id' : header['msg_id'],
         'header' : header,
         'content': msg['content'],
         'buffers': msg['buffers'],
         'submitted': header['date'],
         'client_uuid' : None,
         'engine_uuid' : None,
         'started': None,
         'completed': None,
         'resubmitted': None,
         'received': None,
         'result_header' : None,
         'result_content' : None,
         'result_buffers' : None,
         'queue' : None,
         'pyin' : None,
         'pyout': None,
         'pyerr': None,
         'stdout': '',
         'stderr': '',
     }
 
 
 class EngineConnector(HasTraits):
     """A simple object for accessing the various zmq connections of an object.
     Attributes are:
     id (int): engine ID
     uuid (str): uuid (unused?)
     queue (str): identity of queue's XREQ socket
     registration (str): identity of registration XREQ socket
     heartbeat (str): identity of heartbeat XREQ socket
     """
     id=Integer(0)
     queue=CBytes()
     control=CBytes()
     registration=CBytes()
     heartbeat=CBytes()
     pending=Set()
 
 class HubFactory(RegistrationFactory):
     """The Configurable for setting up a Hub."""
 
     # port-pairs for monitoredqueues:
     hb = Tuple(Integer,Integer,config=True,
         help="""XREQ/SUB Port pair for Engine heartbeats""")
     def _hb_default(self):
         return tuple(util.select_random_ports(2))
 
     mux = Tuple(Integer,Integer,config=True,
         help="""Engine/Client Port pair for MUX queue""")
 
     def _mux_default(self):
         return tuple(util.select_random_ports(2))
 
     task = Tuple(Integer,Integer,config=True,
         help="""Engine/Client Port pair for Task queue""")
     def _task_default(self):
         return tuple(util.select_random_ports(2))
 
     control = Tuple(Integer,Integer,config=True,
         help="""Engine/Client Port pair for Control queue""")
 
     def _control_default(self):
         return tuple(util.select_random_ports(2))
 
     iopub = Tuple(Integer,Integer,config=True,
         help="""Engine/Client Port pair for IOPub relay""")
 
     def _iopub_default(self):
         return tuple(util.select_random_ports(2))
 
     # single ports:
     mon_port = Integer(config=True,
         help="""Monitor (SUB) port for queue traffic""")
 
     def _mon_port_default(self):
         return util.select_random_ports(1)[0]
 
     notifier_port = Integer(config=True,
         help="""PUB port for sending engine status notifications""")
 
     def _notifier_port_default(self):
         return util.select_random_ports(1)[0]
 
     engine_ip = Unicode('127.0.0.1', config=True,
         help="IP on which to listen for engine connections. [default: loopback]")
     engine_transport = Unicode('tcp', config=True,
         help="0MQ transport for engine connections. [default: tcp]")
 
     client_ip = Unicode('127.0.0.1', config=True,
         help="IP on which to listen for client connections. [default: loopback]")
     client_transport = Unicode('tcp', config=True,
         help="0MQ transport for client connections. [default : tcp]")
 
     monitor_ip = Unicode('127.0.0.1', config=True,
         help="IP on which to listen for monitor messages. [default: loopback]")
     monitor_transport = Unicode('tcp', config=True,
         help="0MQ transport for monitor messages. [default : tcp]")
 
     monitor_url = Unicode('')
 
     db_class = DottedObjectName('IPython.parallel.controller.dictdb.DictDB',
         config=True, help="""The class to use for the DB backend""")
 
     # not configurable
     db = Instance('IPython.parallel.controller.dictdb.BaseDB')
     heartmonitor = Instance('IPython.parallel.controller.heartmonitor.HeartMonitor')
 
     def _ip_changed(self, name, old, new):
         self.engine_ip = new
         self.client_ip = new
         self.monitor_ip = new
         self._update_monitor_url()
 
     def _update_monitor_url(self):
         self.monitor_url = "%s://%s:%i" % (self.monitor_transport, self.monitor_ip, self.mon_port)
 
     def _transport_changed(self, name, old, new):
         self.engine_transport = new
         self.client_transport = new
         self.monitor_transport = new
         self._update_monitor_url()
 
     def __init__(self, **kwargs):
         super(HubFactory, self).__init__(**kwargs)
         self._update_monitor_url()
 
 
     def construct(self):
         self.init_hub()
 
     def start(self):
         self.heartmonitor.start()
         self.log.info("Heartmonitor started")
 
     def init_hub(self):
         """construct"""
         client_iface = "%s://%s:" % (self.client_transport, self.client_ip) + "%i"
         engine_iface = "%s://%s:" % (self.engine_transport, self.engine_ip) + "%i"
 
         ctx = self.context
         loop = self.loop
 
         # Registrar socket
         q = ZMQStream(ctx.socket(zmq.ROUTER), loop)
         q.bind(client_iface % self.regport)
         self.log.info("Hub listening on %s for registration.", client_iface % self.regport)
         if self.client_ip != self.engine_ip:
             q.bind(engine_iface % self.regport)
             self.log.info("Hub listening on %s for registration.", engine_iface % self.regport)
 
         ### Engine connections ###
 
         # heartbeat
         hpub = ctx.socket(zmq.PUB)
         hpub.bind(engine_iface % self.hb[0])
         hrep = ctx.socket(zmq.ROUTER)
         hrep.bind(engine_iface % self.hb[1])
         self.heartmonitor = HeartMonitor(loop=loop, config=self.config, log=self.log,
                                 pingstream=ZMQStream(hpub,loop),
                                 pongstream=ZMQStream(hrep,loop)
                             )
 
         ### Client connections ###
         # Notifier socket
         n = ZMQStream(ctx.socket(zmq.PUB), loop)
         n.bind(client_iface%self.notifier_port)
 
         ### build and launch the queues ###
 
         # monitor socket
         sub = ctx.socket(zmq.SUB)
         sub.setsockopt(zmq.SUBSCRIBE, b"")
         sub.bind(self.monitor_url)
         sub.bind('inproc://monitor')
         sub = ZMQStream(sub, loop)
 
         # connect the db
         self.log.info('Hub using DB backend: %r'%(self.db_class.split()[-1]))
         # cdir = self.config.Global.cluster_dir
         self.db = import_item(str(self.db_class))(session=self.session.session,
                                             config=self.config, log=self.log)
         time.sleep(.25)
         try:
             scheme = self.config.TaskScheduler.scheme_name
         except AttributeError:
             from .scheduler import TaskScheduler
             scheme = TaskScheduler.scheme_name.get_default_value()
         # build connection dicts
         self.engine_info = {
             'control' : engine_iface%self.control[1],
             'mux': engine_iface%self.mux[1],
             'heartbeat': (engine_iface%self.hb[0], engine_iface%self.hb[1]),
             'task' : engine_iface%self.task[1],
             'iopub' : engine_iface%self.iopub[1],
             # 'monitor' : engine_iface%self.mon_port,
             }
 
         self.client_info = {
             'control' : client_iface%self.control[0],
             'mux': client_iface%self.mux[0],
             'task' : (scheme, client_iface%self.task[0]),
             'iopub' : client_iface%self.iopub[0],
             'notification': client_iface%self.notifier_port
             }
         self.log.debug("Hub engine addrs: %s", self.engine_info)
         self.log.debug("Hub client addrs: %s", self.client_info)
 
         # resubmit stream
         r = ZMQStream(ctx.socket(zmq.DEALER), loop)
         url = util.disambiguate_url(self.client_info['task'][-1])
         r.setsockopt(zmq.IDENTITY, self.session.bsession)
         r.connect(url)
 
         self.hub = Hub(loop=loop, session=self.session, monitor=sub, heartmonitor=self.heartmonitor,
                 query=q, notifier=n, resubmit=r, db=self.db,
                 engine_info=self.engine_info, client_info=self.client_info,
                 log=self.log)
 
 
 class Hub(SessionFactory):
     """The IPython Controller Hub with 0MQ connections
 
     Parameters
     ==========
     loop: zmq IOLoop instance
     session: Session object
     <removed> context: zmq context for creating new connections (?)
     queue: ZMQStream for monitoring the command queue (SUB)
     query: ZMQStream for engine registration and client queries requests (XREP)
     heartbeat: HeartMonitor object checking the pulse of the engines
     notifier: ZMQStream for broadcasting engine registration changes (PUB)
     db: connection to db for out of memory logging of commands
                 NotImplemented
     engine_info: dict of zmq connection information for engines to connect
                 to the queues.
     client_info: dict of zmq connection information for engines to connect
                 to the queues.
     """
     # internal data structures:
     ids=Set() # engine IDs
     keytable=Dict()
     by_ident=Dict()
     engines=Dict()
     clients=Dict()
     hearts=Dict()
     pending=Set()
     queues=Dict()  # pending msg_ids keyed by engine_id
     tasks=Dict() # pending msg_ids submitted as tasks, keyed by client_id
     completed=Dict() # completed msg_ids keyed by engine_id
     all_completed=Set() # completed msg_ids keyed by engine_id
     dead_engines=Set() # completed msg_ids keyed by engine_id
     unassigned=Set() # set of task msg_ds not yet assigned a destination
     incoming_registrations=Dict()
     registration_timeout=Integer()
     _idcounter=Integer(0)
 
     # objects from constructor:
     query=Instance(ZMQStream)
     monitor=Instance(ZMQStream)
     notifier=Instance(ZMQStream)
     resubmit=Instance(ZMQStream)
     heartmonitor=Instance(HeartMonitor)
     db=Instance(object)
     client_info=Dict()
     engine_info=Dict()
 
 
     def __init__(self, **kwargs):
         """
         # universal:
         loop: IOLoop for creating future connections
         session: streamsession for sending serialized data
         # engine:
         queue: ZMQStream for monitoring queue messages
         query: ZMQStream for engine+client registration and client requests
         heartbeat: HeartMonitor object for tracking engines
         # extra:
         db: ZMQStream for db connection (NotImplemented)
         engine_info: zmq address/protocol dict for engine connections
         client_info: zmq address/protocol dict for client connections
         """
 
         super(Hub, self).__init__(**kwargs)
         self.registration_timeout = max(5000, 2*self.heartmonitor.period)
 
         # validate connection dicts:
         for k,v in self.client_info.iteritems():
             if k == 'task':
                 util.validate_url_container(v[1])
             else:
                 util.validate_url_container(v)
         # util.validate_url_container(self.client_info)
         util.validate_url_container(self.engine_info)
 
         # register our callbacks
         self.query.on_recv(self.dispatch_query)
         self.monitor.on_recv(self.dispatch_monitor_traffic)
 
         self.heartmonitor.add_heart_failure_handler(self.handle_heart_failure)
         self.heartmonitor.add_new_heart_handler(self.handle_new_heart)
 
         self.monitor_handlers = {b'in' : self.save_queue_request,
                                 b'out': self.save_queue_result,
                                 b'intask': self.save_task_request,
                                 b'outtask': self.save_task_result,
                                 b'tracktask': self.save_task_destination,
                                 b'incontrol': _passer,
                                 b'outcontrol': _passer,
                                 b'iopub': self.save_iopub_message,
         }
 
         self.query_handlers = {'queue_request': self.queue_status,
                                 'result_request': self.get_results,
                                 'history_request': self.get_history,
                                 'db_request': self.db_query,
                                 'purge_request': self.purge_results,
                                 'load_request': self.check_load,
                                 'resubmit_request': self.resubmit_task,
                                 'shutdown_request': self.shutdown_request,
                                 'registration_request' : self.register_engine,
                                 'unregistration_request' : self.unregister_engine,
                                 'connection_request': self.connection_request,
         }
 
         # ignore resubmit replies
         self.resubmit.on_recv(lambda msg: None, copy=False)
 
         self.log.info("hub::created hub")
 
     @property
     def _next_id(self):
         """gemerate a new ID.
 
         No longer reuse old ids, just count from 0."""
         newid = self._idcounter
         self._idcounter += 1
         return newid
         # newid = 0
         # incoming = [id[0] for id in self.incoming_registrations.itervalues()]
         # # print newid, self.ids, self.incoming_registrations
         # while newid in self.ids or newid in incoming:
         #     newid += 1
         # return newid
 
     #-----------------------------------------------------------------------------
     # message validation
     #-----------------------------------------------------------------------------
 
     def _validate_targets(self, targets):
         """turn any valid targets argument into a list of integer ids"""
         if targets is None:
             # default to all
             return self.ids
 
         if isinstance(targets, (int,str,unicode)):
             # only one target specified
             targets = [targets]
         _targets = []
         for t in targets:
             # map raw identities to ids
             if isinstance(t, (str,unicode)):
-                t = self.by_ident.get(t, t)
+                t = self.by_ident.get(cast_bytes(t), t)
             _targets.append(t)
         targets = _targets
         bad_targets = [ t for t in targets if t not in self.ids ]
         if bad_targets:
             raise IndexError("No Such Engine: %r" % bad_targets)
         if not targets:
             raise IndexError("No Engines Registered")
         return targets
 
     #-----------------------------------------------------------------------------
     # dispatch methods (1 per stream)
     #-----------------------------------------------------------------------------
 
 
     @util.log_errors
     def dispatch_monitor_traffic(self, msg):
         """all ME and Task queue messages come through here, as well as
         IOPub traffic."""
         self.log.debug("monitor traffic: %r", msg[0])
         switch = msg[0]
         try:
             idents, msg = self.session.feed_identities(msg[1:])
         except ValueError:
             idents=[]
         if not idents:
-            self.log.error("Bad Monitor Message: %r", msg)
+            self.log.error("Monitor message without topic: %r", msg)
             return
         handler = self.monitor_handlers.get(switch, None)
         if handler is not None:
             handler(idents, msg)
         else:
-            self.log.error("Invalid monitor topic: %r", switch)
+            self.log.error("Unrecognized monitor topic: %r", switch)
 
 
     @util.log_errors
     def dispatch_query(self, msg):
         """Route registration requests and queries from clients."""
         try:
             idents, msg = self.session.feed_identities(msg)
         except ValueError:
             idents = []
         if not idents:
             self.log.error("Bad Query Message: %r", msg)
             return
         client_id = idents[0]
         try:
             msg = self.session.unserialize(msg, content=True)
         except Exception:
             content = error.wrap_exception()
             self.log.error("Bad Query Message: %r", msg, exc_info=True)
             self.session.send(self.query, "hub_error", ident=client_id,
                     content=content)
             return
         # print client_id, header, parent, content
         #switch on message type:
         msg_type = msg['header']['msg_type']
         self.log.info("client::client %r requested %r", client_id, msg_type)
         handler = self.query_handlers.get(msg_type, None)
         try:
             assert handler is not None, "Bad Message Type: %r" % msg_type
         except:
             content = error.wrap_exception()
             self.log.error("Bad Message Type: %r", msg_type, exc_info=True)
             self.session.send(self.query, "hub_error", ident=client_id,
                     content=content)
             return
 
         else:
             handler(idents, msg)
 
     def dispatch_db(self, msg):
         """"""
         raise NotImplementedError
 
     #---------------------------------------------------------------------------
     # handler methods (1 per event)
     #---------------------------------------------------------------------------
 
     #----------------------- Heartbeat --------------------------------------
 
     def handle_new_heart(self, heart):
         """handler to attach to heartbeater.
         Called when a new heart starts to beat.
         Triggers completion of registration."""
         self.log.debug("heartbeat::handle_new_heart(%r)", heart)
         if heart not in self.incoming_registrations:
             self.log.info("heartbeat::ignoring new heart: %r", heart)
         else:
             self.finish_registration(heart)
 
 
     def handle_heart_failure(self, heart):
         """handler to attach to heartbeater.
         called when a previously registered heart fails to respond to beat request.
         triggers unregistration"""
         self.log.debug("heartbeat::handle_heart_failure(%r)", heart)
         eid = self.hearts.get(heart, None)
         queue = self.engines[eid].queue
         if eid is None or self.keytable[eid] in self.dead_engines:
             self.log.info("heartbeat::ignoring heart failure %r (not an engine or already dead)", heart)
         else:
             self.unregister_engine(heart, dict(content=dict(id=eid, queue=queue)))
 
     #----------------------- MUX Queue Traffic ------------------------------
 
     def save_queue_request(self, idents, msg):
         if len(idents) < 2:
             self.log.error("invalid identity prefix: %r", idents)
             return
         queue_id, client_id = idents[:2]
         try:
             msg = self.session.unserialize(msg)
         except Exception:
             self.log.error("queue::client %r sent invalid message to %r: %r", client_id, queue_id, msg, exc_info=True)
             return
 
         eid = self.by_ident.get(queue_id, None)
         if eid is None:
             self.log.error("queue::target %r not registered", queue_id)
             self.log.debug("queue::    valid are: %r", self.by_ident.keys())
             return
         record = init_record(msg)
         msg_id = record['msg_id']
         self.log.info("queue::client %r submitted request %r to %s", client_id, msg_id, eid)
         # Unicode in records
         record['engine_uuid'] = queue_id.decode('ascii')
         record['client_uuid'] = client_id.decode('ascii')
         record['queue'] = 'mux'
 
         try:
             # it's posible iopub arrived first:
             existing = self.db.get_record(msg_id)
             for key,evalue in existing.iteritems():
                 rvalue = record.get(key, None)
                 if evalue and rvalue and evalue != rvalue:
                     self.log.warn("conflicting initial state for record: %r:%r <%r> %r", msg_id, rvalue, key, evalue)
                 elif evalue and not rvalue:
                     record[key] = evalue
             try:
                 self.db.update_record(msg_id, record)
             except Exception:
                 self.log.error("DB Error updating record %r", msg_id, exc_info=True)
         except KeyError:
             try:
                 self.db.add_record(msg_id, record)
             except Exception:
                 self.log.error("DB Error adding record %r", msg_id, exc_info=True)
 
 
         self.pending.add(msg_id)
         self.queues[eid].append(msg_id)
 
     def save_queue_result(self, idents, msg):
         if len(idents) < 2:
             self.log.error("invalid identity prefix: %r", idents)
             return
 
         client_id, queue_id = idents[:2]
         try:
             msg = self.session.unserialize(msg)
         except Exception:
             self.log.error("queue::engine %r sent invalid message to %r: %r",
                     queue_id, client_id, msg, exc_info=True)
             return
 
         eid = self.by_ident.get(queue_id, None)
         if eid is None:
             self.log.error("queue::unknown engine %r is sending a reply: ", queue_id)
             return
 
         parent = msg['parent_header']
         if not parent:
             return
         msg_id = parent['msg_id']
         if msg_id in self.pending:
             self.pending.remove(msg_id)
             self.all_completed.add(msg_id)
             self.queues[eid].remove(msg_id)
             self.completed[eid].append(msg_id)
             self.log.info("queue::request %r completed on %s", msg_id, eid)
         elif msg_id not in self.all_completed:
             # it could be a result from a dead engine that died before delivering the
             # result
             self.log.warn("queue:: unknown msg finished %r", msg_id)
             return
         # update record anyway, because the unregistration could have been premature
         rheader = msg['header']
         completed = rheader['date']
         started = rheader.get('started', None)
         result = {
             'result_header' : rheader,
             'result_content': msg['content'],
             'received': datetime.now(),
             'started' : started,
             'completed' : completed
         }
 
         result['result_buffers'] = msg['buffers']
         try:
             self.db.update_record(msg_id, result)
         except Exception:
             self.log.error("DB Error updating record %r", msg_id, exc_info=True)
 
 
     #--------------------- Task Queue Traffic ------------------------------
 
     def save_task_request(self, idents, msg):
         """Save the submission of a task."""
         client_id = idents[0]
 
         try:
             msg = self.session.unserialize(msg)
         except Exception:
             self.log.error("task::client %r sent invalid task message: %r",
                     client_id, msg, exc_info=True)
             return
         record = init_record(msg)
 
         record['client_uuid'] = client_id.decode('ascii')
         record['queue'] = 'task'
         header = msg['header']
         msg_id = header['msg_id']
         self.pending.add(msg_id)
         self.unassigned.add(msg_id)
         try:
             # it's posible iopub arrived first:
             existing = self.db.get_record(msg_id)
             if existing['resubmitted']:
                 for key in ('submitted', 'client_uuid', 'buffers'):
                     # don't clobber these keys on resubmit
                     # submitted and client_uuid should be different
                     # and buffers might be big, and shouldn't have changed
                     record.pop(key)
                     # still check content,header which should not change
                     # but are not expensive to compare as buffers
 
             for key,evalue in existing.iteritems():
                 if key.endswith('buffers'):
                     # don't compare buffers
                     continue
                 rvalue = record.get(key, None)
                 if evalue and rvalue and evalue != rvalue:
                     self.log.warn("conflicting initial state for record: %r:%r <%r> %r", msg_id, rvalue, key, evalue)
                 elif evalue and not rvalue:
                     record[key] = evalue
             try:
                 self.db.update_record(msg_id, record)
             except Exception:
                 self.log.error("DB Error updating record %r", msg_id, exc_info=True)
         except KeyError:
             try:
                 self.db.add_record(msg_id, record)
             except Exception:
                 self.log.error("DB Error adding record %r", msg_id, exc_info=True)
         except Exception:
             self.log.error("DB Error saving task request %r", msg_id, exc_info=True)
 
     def save_task_result(self, idents, msg):
         """save the result of a completed task."""
         client_id = idents[0]
         try:
             msg = self.session.unserialize(msg)
         except Exception:
             self.log.error("task::invalid task result message send to %r: %r",
                     client_id, msg, exc_info=True)
             return
 
         parent = msg['parent_header']
         if not parent:
             # print msg
             self.log.warn("Task %r had no parent!", msg)
             return
         msg_id = parent['msg_id']
         if msg_id in self.unassigned:
             self.unassigned.remove(msg_id)
 
         header = msg['header']
-        engine_uuid = header.get('engine', None)
-        eid = self.by_ident.get(engine_uuid, None)
+        engine_uuid = header.get('engine', u'')
+        eid = self.by_ident.get(cast_bytes(engine_uuid), None)
+        
+        status = header.get('status', None)
 
         if msg_id in self.pending:
             self.log.info("task::task %r finished on %s", msg_id, eid)
             self.pending.remove(msg_id)
             self.all_completed.add(msg_id)
             if eid is not None:
-                self.completed[eid].append(msg_id)
+                if status != 'aborted':
+                    self.completed[eid].append(msg_id)
                 if msg_id in self.tasks[eid]:
                     self.tasks[eid].remove(msg_id)
             completed = header['date']
             started = header.get('started', None)
             result = {
                 'result_header' : header,
                 'result_content': msg['content'],
                 'started' : started,
                 'completed' : completed,
                 'received' : datetime.now(),
                 'engine_uuid': engine_uuid,
             }
 
             result['result_buffers'] = msg['buffers']
             try:
                 self.db.update_record(msg_id, result)
             except Exception:
                 self.log.error("DB Error saving task request %r", msg_id, exc_info=True)
 
         else:
             self.log.debug("task::unknown task %r finished", msg_id)
 
     def save_task_destination(self, idents, msg):
         try:
             msg = self.session.unserialize(msg, content=True)
         except Exception:
             self.log.error("task::invalid task tracking message", exc_info=True)
             return
         content = msg['content']
         # print (content)
         msg_id = content['msg_id']
         engine_uuid = content['engine_id']
-        eid = self.by_ident[util.asbytes(engine_uuid)]
+        eid = self.by_ident[cast_bytes(engine_uuid)]
 
         self.log.info("task::task %r arrived on %r", msg_id, eid)
         if msg_id in self.unassigned:
             self.unassigned.remove(msg_id)
         # else:
         #     self.log.debug("task::task %r not listed as MIA?!"%(msg_id))
 
         self.tasks[eid].append(msg_id)
         # self.pending[msg_id][1].update(received=datetime.now(),engine=(eid,engine_uuid))
         try:
             self.db.update_record(msg_id, dict(engine_uuid=engine_uuid))
         except Exception:
             self.log.error("DB Error saving task destination %r", msg_id, exc_info=True)
 
 
     def mia_task_request(self, idents, msg):
         raise NotImplementedError
         client_id = idents[0]
         # content = dict(mia=self.mia,status='ok')
         # self.session.send('mia_reply', content=content, idents=client_id)
 
 
     #--------------------- IOPub Traffic ------------------------------
 
     def save_iopub_message(self, topics, msg):
         """save an iopub message into the db"""
         # print (topics)
         try:
             msg = self.session.unserialize(msg, content=True)
         except Exception:
             self.log.error("iopub::invalid IOPub message", exc_info=True)
             return
 
         parent = msg['parent_header']
         if not parent:
-            self.log.error("iopub::invalid IOPub message: %r", msg)
+            self.log.warn("iopub::IOPub message lacks parent: %r", msg)
             return
         msg_id = parent['msg_id']
         msg_type = msg['header']['msg_type']
         content = msg['content']
 
         # ensure msg_id is in db
         try:
             rec = self.db.get_record(msg_id)
         except KeyError:
             rec = empty_record()
             rec['msg_id'] = msg_id
             self.db.add_record(msg_id, rec)
         # stream
         d = {}
         if msg_type == 'stream':
             name = content['name']
             s = rec[name] or ''
             d[name] = s + content['data']
 
         elif msg_type == 'pyerr':
             d['pyerr'] = content
         elif msg_type == 'pyin':
             d['pyin'] = content['code']
         else:
             d[msg_type] = content.get('data', '')
 
         try:
             self.db.update_record(msg_id, d)
         except Exception:
             self.log.error("DB Error saving iopub message %r", msg_id, exc_info=True)
 
 
 
     #-------------------------------------------------------------------------
     # Registration requests
     #-------------------------------------------------------------------------
 
     def connection_request(self, client_id, msg):
         """Reply with connection addresses for clients."""
         self.log.info("client::client %r connected", client_id)
         content = dict(status='ok')
         content.update(self.client_info)
         jsonable = {}
         for k,v in self.keytable.iteritems():
             if v not in self.dead_engines:
                 jsonable[str(k)] = v.decode('ascii')
         content['engines'] = jsonable
         self.session.send(self.query, 'connection_reply', content, parent=msg, ident=client_id)
 
     def register_engine(self, reg, msg):
         """Register a new engine."""
         content = msg['content']
         try:
-            queue = util.asbytes(content['queue'])
+            queue = cast_bytes(content['queue'])
         except KeyError:
             self.log.error("registration::queue not specified", exc_info=True)
             return
         heart = content.get('heartbeat', None)
         if heart:
-            heart = util.asbytes(heart)
+            heart = cast_bytes(heart)
         """register a new engine, and create the socket(s) necessary"""
         eid = self._next_id
         # print (eid, queue, reg, heart)
 
         self.log.debug("registration::register_engine(%i, %r, %r, %r)", eid, queue, reg, heart)
 
         content = dict(id=eid,status='ok')
         content.update(self.engine_info)
         # check if requesting available IDs:
         if queue in self.by_ident:
             try:
                 raise KeyError("queue_id %r in use" % queue)
             except:
                 content = error.wrap_exception()
                 self.log.error("queue_id %r in use", queue, exc_info=True)
         elif heart in self.hearts: # need to check unique hearts?
             try:
                 raise KeyError("heart_id %r in use" % heart)
             except:
                 self.log.error("heart_id %r in use", heart, exc_info=True)
                 content = error.wrap_exception()
         else:
             for h, pack in self.incoming_registrations.iteritems():
                 if heart == h:
                     try:
                         raise KeyError("heart_id %r in use" % heart)
                     except:
                         self.log.error("heart_id %r in use", heart, exc_info=True)
                         content = error.wrap_exception()
                     break
                 elif queue == pack[1]:
                     try:
                         raise KeyError("queue_id %r in use" % queue)
                     except:
                         self.log.error("queue_id %r in use", queue, exc_info=True)
                         content = error.wrap_exception()
                     break
 
         msg = self.session.send(self.query, "registration_reply",
                 content=content,
                 ident=reg)
 
         if content['status'] == 'ok':
             if heart in self.heartmonitor.hearts:
                 # already beating
                 self.incoming_registrations[heart] = (eid,queue,reg[0],None)
                 self.finish_registration(heart)
             else:
                 purge = lambda : self._purge_stalled_registration(heart)
                 dc = ioloop.DelayedCallback(purge, self.registration_timeout, self.loop)
                 dc.start()
                 self.incoming_registrations[heart] = (eid,queue,reg[0],dc)
         else:
             self.log.error("registration::registration %i failed: %r", eid, content['evalue'])
         return eid
 
     def unregister_engine(self, ident, msg):
         """Unregister an engine that explicitly requested to leave."""
         try:
             eid = msg['content']['id']
         except:
             self.log.error("registration::bad engine id for unregistration: %r", ident, exc_info=True)
             return
         self.log.info("registration::unregister_engine(%r)", eid)
         # print (eid)
         uuid = self.keytable[eid]
         content=dict(id=eid, queue=uuid.decode('ascii'))
         self.dead_engines.add(uuid)
         # self.ids.remove(eid)
         # uuid = self.keytable.pop(eid)
         #
         # ec = self.engines.pop(eid)
         # self.hearts.pop(ec.heartbeat)
         # self.by_ident.pop(ec.queue)
         # self.completed.pop(eid)
         handleit = lambda : self._handle_stranded_msgs(eid, uuid)
         dc = ioloop.DelayedCallback(handleit, self.registration_timeout, self.loop)
         dc.start()
         ############## TODO: HANDLE IT ################
 
         if self.notifier:
             self.session.send(self.notifier, "unregistration_notification", content=content)
 
     def _handle_stranded_msgs(self, eid, uuid):
         """Handle messages known to be on an engine when the engine unregisters.
 
         It is possible that this will fire prematurely - that is, an engine will
         go down after completing a result, and the client will be notified
         that the result failed and later receive the actual result.
         """
 
         outstanding = self.queues[eid]
 
         for msg_id in outstanding:
             self.pending.remove(msg_id)
             self.all_completed.add(msg_id)
             try:
                 raise error.EngineError("Engine %r died while running task %r" % (eid, msg_id))
             except:
                 content = error.wrap_exception()
             # build a fake header:
             header = {}
             header['engine'] = uuid
             header['date'] = datetime.now()
             rec = dict(result_content=content, result_header=header, result_buffers=[])
             rec['completed'] = header['date']
             rec['engine_uuid'] = uuid
             try:
                 self.db.update_record(msg_id, rec)
             except Exception:
                 self.log.error("DB Error handling stranded msg %r", msg_id, exc_info=True)
 
 
     def finish_registration(self, heart):
         """Second half of engine registration, called after our HeartMonitor
         has received a beat from the Engine's Heart."""
         try:
             (eid,queue,reg,purge) = self.incoming_registrations.pop(heart)
         except KeyError:
             self.log.error("registration::tried to finish nonexistant registration", exc_info=True)
             return
         self.log.info("registration::finished registering engine %i:%r", eid, queue)
         if purge is not None:
             purge.stop()
         control = queue
         self.ids.add(eid)
         self.keytable[eid] = queue
         self.engines[eid] = EngineConnector(id=eid, queue=queue, registration=reg,
                                     control=control, heartbeat=heart)
         self.by_ident[queue] = eid
         self.queues[eid] = list()
         self.tasks[eid] = list()
         self.completed[eid] = list()
         self.hearts[heart] = eid
         content = dict(id=eid, queue=self.engines[eid].queue.decode('ascii'))
         if self.notifier:
             self.session.send(self.notifier, "registration_notification", content=content)
         self.log.info("engine::Engine Connected: %i", eid)
 
     def _purge_stalled_registration(self, heart):
         if heart in self.incoming_registrations:
             eid = self.incoming_registrations.pop(heart)[0]
             self.log.info("registration::purging stalled registration: %i", eid)
         else:
             pass
 
     #-------------------------------------------------------------------------
     # Client Requests
     #-------------------------------------------------------------------------
 
     def shutdown_request(self, client_id, msg):
         """handle shutdown request."""
         self.session.send(self.query, 'shutdown_reply', content={'status': 'ok'}, ident=client_id)
         # also notify other clients of shutdown
         self.session.send(self.notifier, 'shutdown_notice', content={'status': 'ok'})
         dc = ioloop.DelayedCallback(lambda : self._shutdown(), 1000, self.loop)
         dc.start()
 
     def _shutdown(self):
         self.log.info("hub::hub shutting down.")
         time.sleep(0.1)
         sys.exit(0)
 
 
     def check_load(self, client_id, msg):
         content = msg['content']
         try:
             targets = content['targets']
             targets = self._validate_targets(targets)
         except:
             content = error.wrap_exception()
             self.session.send(self.query, "hub_error",
                     content=content, ident=client_id)
             return
 
         content = dict(status='ok')
         # loads = {}
         for t in targets:
             content[bytes(t)] = len(self.queues[t])+len(self.tasks[t])
         self.session.send(self.query, "load_reply", content=content, ident=client_id)
 
 
     def queue_status(self, client_id, msg):
         """Return the Queue status of one or more targets.
         if verbose: return the msg_ids
         else: return len of each type.
         keys: queue (pending MUX jobs)
             tasks (pending Task jobs)
             completed (finished jobs from both queues)"""
         content = msg['content']
         targets = content['targets']
         try:
             targets = self._validate_targets(targets)
         except:
             content = error.wrap_exception()
             self.session.send(self.query, "hub_error",
                     content=content, ident=client_id)
             return
         verbose = content.get('verbose', False)
         content = dict(status='ok')
         for t in targets:
             queue = self.queues[t]
             completed = self.completed[t]
             tasks = self.tasks[t]
             if not verbose:
                 queue = len(queue)
                 completed = len(completed)
                 tasks = len(tasks)
             content[str(t)] = {'queue': queue, 'completed': completed , 'tasks': tasks}
         content['unassigned'] = list(self.unassigned) if verbose else len(self.unassigned)
         # print (content)
         self.session.send(self.query, "queue_reply", content=content, ident=client_id)
 
     def purge_results(self, client_id, msg):
         """Purge results from memory. This method is more valuable before we move
         to a DB based message storage mechanism."""
         content = msg['content']
         self.log.info("Dropping records with %s", content)
         msg_ids = content.get('msg_ids', [])
         reply = dict(status='ok')
         if msg_ids == 'all':
             try:
                 self.db.drop_matching_records(dict(completed={'$ne':None}))
             except Exception:
                 reply = error.wrap_exception()
         else:
             pending = filter(lambda m: m in self.pending, msg_ids)
             if pending:
                 try:
                     raise IndexError("msg pending: %r" % pending[0])
                 except:
                     reply = error.wrap_exception()
             else:
                 try:
                     self.db.drop_matching_records(dict(msg_id={'$in':msg_ids}))
                 except Exception:
                     reply = error.wrap_exception()
 
             if reply['status'] == 'ok':
                 eids = content.get('engine_ids', [])
                 for eid in eids:
                     if eid not in self.engines:
                         try:
                             raise IndexError("No such engine: %i" % eid)
                         except:
                             reply = error.wrap_exception()
                         break
                     uid = self.engines[eid].queue
                     try:
                         self.db.drop_matching_records(dict(engine_uuid=uid, completed={'$ne':None}))
                     except Exception:
                         reply = error.wrap_exception()
                         break
 
         self.session.send(self.query, 'purge_reply', content=reply, ident=client_id)
 
     def resubmit_task(self, client_id, msg):
         """Resubmit one or more tasks."""
         def finish(reply):
             self.session.send(self.query, 'resubmit_reply', content=reply, ident=client_id)
 
         content = msg['content']
         msg_ids = content['msg_ids']
         reply = dict(status='ok')
         try:
             records = self.db.find_records({'msg_id' : {'$in' : msg_ids}}, keys=[
                 'header', 'content', 'buffers'])
         except Exception:
             self.log.error('db::db error finding tasks to resubmit', exc_info=True)
             return finish(error.wrap_exception())
 
         # validate msg_ids
         found_ids = [ rec['msg_id'] for rec in records ]
-        invalid_ids = filter(lambda m: m in self.pending, found_ids)
+        pending_ids = [ msg_id for msg_id in found_ids if msg_id in self.pending ]
         if len(records) > len(msg_ids):
             try:
                 raise RuntimeError("DB appears to be in an inconsistent state."
                     "More matching records were found than should exist")
             except Exception:
                 return finish(error.wrap_exception())
         elif len(records) < len(msg_ids):
             missing = [ m for m in msg_ids if m not in found_ids ]
             try:
                 raise KeyError("No such msg(s): %r" % missing)
             except KeyError:
                 return finish(error.wrap_exception())
-        elif invalid_ids:
-            msg_id = invalid_ids[0]
+        elif pending_ids:
+            pass
+            # no need to raise on resubmit of pending task, now that we
+            # resubmit under new ID, but do we want to raise anyway?
+            # msg_id = invalid_ids[0]
+            # try:
+            #     raise ValueError("Task(s) %r appears to be inflight" % )
+            # except Exception:
+            #     return finish(error.wrap_exception())
+
+        # mapping of original IDs to resubmitted IDs
+        resubmitted = {}
+
+        # send the messages
+        for rec in records:
+            header = rec['header']
+            msg = self.session.msg(header['msg_type'])
+            msg_id = msg['msg_id']
+            msg['content'] = rec['content']
+            header.update(msg['header'])
+            msg['header'] = header
+
+            self.session.send(self.resubmit, msg, buffers=rec['buffers'])
+
+            resubmitted[rec['msg_id']] = msg_id
+            self.pending.add(msg_id)
+            msg['buffers'] = []
             try:
-                raise ValueError("Task %r appears to be inflight" % msg_id)
+                self.db.add_record(msg_id, init_record(msg))
             except Exception:
-                return finish(error.wrap_exception())
+                self.log.error("db::DB Error updating record: %s", msg_id, exc_info=True)
 
-        # clear the existing records
-        now = datetime.now()
-        rec = empty_record()
-        map(rec.pop, ['msg_id', 'header', 'content', 'buffers', 'submitted'])
-        rec['resubmitted'] = now
-        rec['queue'] = 'task'
-        rec['client_uuid'] = client_id[0]
-        try:
-            for msg_id in msg_ids:
-                self.all_completed.discard(msg_id)
-                self.db.update_record(msg_id, rec)
-        except Exception:
-            self.log.error('db::db error upating record', exc_info=True)
-            reply = error.wrap_exception()
-        else:
-            # send the messages
-            for rec in records:
-                header = rec['header']
-                # include resubmitted in header to prevent digest collision
-                header['resubmitted'] = now
-                msg = self.session.msg(header['msg_type'])
-                msg['content'] = rec['content']
-                msg['header'] = header
-                msg['header']['msg_id'] = rec['msg_id']
-                self.session.send(self.resubmit, msg, buffers=rec['buffers'])
-
-        finish(dict(status='ok'))
+        finish(dict(status='ok', resubmitted=resubmitted))
+        
+        # store the new IDs in the Task DB
+        for msg_id, resubmit_id in resubmitted.iteritems():
+            try:
+                self.db.update_record(msg_id, {'resubmitted' : resubmit_id})
+            except Exception:
+                self.log.error("db::DB Error updating record: %s", msg_id, exc_info=True)
 
 
     def _extract_record(self, rec):
         """decompose a TaskRecord dict into subsection of reply for get_result"""
         io_dict = {}
         for key in ('pyin', 'pyout', 'pyerr', 'stdout', 'stderr'):
                 io_dict[key] = rec[key]
         content = { 'result_content': rec['result_content'],
                             'header': rec['header'],
                             'result_header' : rec['result_header'],
                             'received' : rec['received'],
                             'io' : io_dict,
                           }
         if rec['result_buffers']:
             buffers = map(bytes, rec['result_buffers'])
         else:
             buffers = []
 
         return content, buffers
 
     def get_results(self, client_id, msg):
         """Get the result of 1 or more messages."""
         content = msg['content']
         msg_ids = sorted(set(content['msg_ids']))
         statusonly = content.get('status_only', False)
         pending = []
         completed = []
         content = dict(status='ok')
         content['pending'] = pending
         content['completed'] = completed
         buffers = []
         if not statusonly:
             try:
                 matches = self.db.find_records(dict(msg_id={'$in':msg_ids}))
                 # turn match list into dict, for faster lookup
                 records = {}
                 for rec in matches:
                     records[rec['msg_id']] = rec
             except Exception:
                 content = error.wrap_exception()
                 self.session.send(self.query, "result_reply", content=content,
                                                     parent=msg, ident=client_id)
                 return
         else:
             records = {}
         for msg_id in msg_ids:
             if msg_id in self.pending:
                 pending.append(msg_id)
             elif msg_id in self.all_completed:
                 completed.append(msg_id)
                 if not statusonly:
                     c,bufs = self._extract_record(records[msg_id])
                     content[msg_id] = c
                     buffers.extend(bufs)
             elif msg_id in records:
                 if rec['completed']:
                     completed.append(msg_id)
                     c,bufs = self._extract_record(records[msg_id])
                     content[msg_id] = c
                     buffers.extend(bufs)
                 else:
                     pending.append(msg_id)
             else:
                 try:
                     raise KeyError('No such message: '+msg_id)
                 except:
                     content = error.wrap_exception()
                 break
         self.session.send(self.query, "result_reply", content=content,
                                             parent=msg, ident=client_id,
                                             buffers=buffers)
 
     def get_history(self, client_id, msg):
         """Get a list of all msg_ids in our DB records"""
         try:
             msg_ids = self.db.get_history()
         except Exception as e:
             content = error.wrap_exception()
         else:
             content = dict(status='ok', history=msg_ids)
 
         self.session.send(self.query, "history_reply", content=content,
                                             parent=msg, ident=client_id)
 
     def db_query(self, client_id, msg):
         """Perform a raw query on the task record database."""
         content = msg['content']
         query = content.get('query', {})
         keys = content.get('keys', None)
         buffers = []
         empty = list()
         try:
             records = self.db.find_records(query, keys)
         except Exception as e:
             content = error.wrap_exception()
         else:
             # extract buffers from reply content:
             if keys is not None:
                 buffer_lens = [] if 'buffers' in keys else None
                 result_buffer_lens = [] if 'result_buffers' in keys else None
             else:
                 buffer_lens = None
                 result_buffer_lens = None
 
             for rec in records:
                 # buffers may be None, so double check
                 b = rec.pop('buffers', empty) or empty
                 if buffer_lens is not None:
                     buffer_lens.append(len(b))
                     buffers.extend(b)
                 rb = rec.pop('result_buffers', empty) or empty
                 if result_buffer_lens is not None:
                     result_buffer_lens.append(len(rb))
                     buffers.extend(rb)
             content = dict(status='ok', records=records, buffer_lens=buffer_lens,
                                     result_buffer_lens=result_buffer_lens)
         # self.log.debug (content)
         self.session.send(self.query, "db_reply", content=content,
                                             parent=msg, ident=client_id,
                                             buffers=buffers)
 

IPython/parallel/controller/scheduler.py

@@ -1,767 +1,768 b''
 """The Python scheduler for rich scheduling.
 
 The Pure ZMQ scheduler does not allow routing schemes other than LRU,
 nor does it check msg_id DAG dependencies. For those, a slightly slower
 Python Scheduler exists.
 
 Authors:
 
 * Min RK
 """
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2010-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #----------------------------------------------------------------------
 # Imports
 #----------------------------------------------------------------------
 
 from __future__ import print_function
 
 import logging
 import sys
 import time
 
 from datetime import datetime, timedelta
 from random import randint, random
 from types import FunctionType
 
 try:
     import numpy
 except ImportError:
     numpy = None
 
 import zmq
 from zmq.eventloop import ioloop, zmqstream
 
 # local imports
 from IPython.external.decorator import decorator
 from IPython.config.application import Application
 from IPython.config.loader import Config
 from IPython.utils.traitlets import Instance, Dict, List, Set, Integer, Enum, CBytes
+from IPython.utils.py3compat import cast_bytes
 
 from IPython.parallel import error, util
 from IPython.parallel.factory import SessionFactory
-from IPython.parallel.util import connect_logger, local_logger, asbytes
+from IPython.parallel.util import connect_logger, local_logger
 
 from .dependency import Dependency
 
 @decorator
 def logged(f,self,*args,**kwargs):
     # print ("#--------------------")
     self.log.debug("scheduler::%s(*%s,**%s)", f.func_name, args, kwargs)
     # print ("#--")
     return f(self,*args, **kwargs)
 
 #----------------------------------------------------------------------
 # Chooser functions
 #----------------------------------------------------------------------
 
 def plainrandom(loads):
     """Plain random pick."""
     n = len(loads)
     return randint(0,n-1)
 
 def lru(loads):
     """Always pick the front of the line.
 
     The content of `loads` is ignored.
 
     Assumes LRU ordering of loads, with oldest first.
     """
     return 0
 
 def twobin(loads):
     """Pick two at random, use the LRU of the two.
 
     The content of loads is ignored.
 
     Assumes LRU ordering of loads, with oldest first.
     """
     n = len(loads)
     a = randint(0,n-1)
     b = randint(0,n-1)
     return min(a,b)
 
 def weighted(loads):
     """Pick two at random using inverse load as weight.
 
     Return the less loaded of the two.
     """
     # weight 0 a million times more than 1:
     weights = 1./(1e-6+numpy.array(loads))
     sums = weights.cumsum()
     t = sums[-1]
     x = random()*t
     y = random()*t
     idx = 0
     idy = 0
     while sums[idx] < x:
         idx += 1
     while sums[idy] < y:
         idy += 1
     if weights[idy] > weights[idx]:
         return idy
     else:
         return idx
 
 def leastload(loads):
     """Always choose the lowest load.
 
     If the lowest load occurs more than once, the first
     occurance will be used.  If loads has LRU ordering, this means
     the LRU of those with the lowest load is chosen.
     """
     return loads.index(min(loads))
 
 #---------------------------------------------------------------------
 # Classes
 #---------------------------------------------------------------------
 
 
 # store empty default dependency:
 MET = Dependency([])
 
 
 class Job(object):
     """Simple container for a job"""
     def __init__(self, msg_id, raw_msg, idents, msg, header, targets, after, follow, timeout):
         self.msg_id = msg_id
         self.raw_msg = raw_msg
         self.idents = idents
         self.msg = msg
         self.header = header
         self.targets = targets
         self.after = after
         self.follow = follow
         self.timeout = timeout
         
         
         self.timestamp = time.time()
         self.blacklist = set()
 
     @property
     def dependents(self):
         return self.follow.union(self.after)
 
 class TaskScheduler(SessionFactory):
     """Python TaskScheduler object.
 
     This is the simplest object that supports msg_id based
     DAG dependencies. *Only* task msg_ids are checked, not
     msg_ids of jobs submitted via the MUX queue.
 
     """
 
     hwm = Integer(1, config=True,
         help="""specify the High Water Mark (HWM) for the downstream
         socket in the Task scheduler. This is the maximum number
         of allowed outstanding tasks on each engine.
         
         The default (1) means that only one task can be outstanding on each
         engine.  Setting TaskScheduler.hwm=0 means there is no limit, and the
         engines continue to be assigned tasks while they are working,
         effectively hiding network latency behind computation, but can result
         in an imbalance of work when submitting many heterogenous tasks all at
         once.  Any positive value greater than one is a compromise between the
         two.
 
         """
     )
     scheme_name = Enum(('leastload', 'pure', 'lru', 'plainrandom', 'weighted', 'twobin'),
         'leastload', config=True, allow_none=False,
         help="""select the task scheduler scheme  [default: Python LRU]
         Options are: 'pure', 'lru', 'plainrandom', 'weighted', 'twobin','leastload'"""
     )
     def _scheme_name_changed(self, old, new):
         self.log.debug("Using scheme %r"%new)
         self.scheme = globals()[new]
 
     # input arguments:
     scheme = Instance(FunctionType) # function for determining the destination
     def _scheme_default(self):
         return leastload
     client_stream = Instance(zmqstream.ZMQStream) # client-facing stream
     engine_stream = Instance(zmqstream.ZMQStream) # engine-facing stream
     notifier_stream = Instance(zmqstream.ZMQStream) # hub-facing sub stream
     mon_stream = Instance(zmqstream.ZMQStream) # hub-facing pub stream
 
     # internals:
     graph = Dict() # dict by msg_id of [ msg_ids that depend on key ]
     retries = Dict() # dict by msg_id of retries remaining (non-neg ints)
     # waiting = List() # list of msg_ids ready to run, but haven't due to HWM
     depending = Dict() # dict by msg_id of Jobs
     pending = Dict() # dict by engine_uuid of submitted tasks
     completed = Dict() # dict by engine_uuid of completed tasks
     failed = Dict() # dict by engine_uuid of failed tasks
     destinations = Dict() # dict by msg_id of engine_uuids where jobs ran (reverse of completed+failed)
     clients = Dict() # dict by msg_id for who submitted the task
     targets = List() # list of target IDENTs
     loads = List() # list of engine loads
     # full = Set() # set of IDENTs that have HWM outstanding tasks
     all_completed = Set() # set of all completed tasks
     all_failed = Set() # set of all failed tasks
     all_done = Set() # set of all finished tasks=union(completed,failed)
     all_ids = Set() # set of all submitted task IDs
 
     auditor = Instance('zmq.eventloop.ioloop.PeriodicCallback')
 
     ident = CBytes() # ZMQ identity. This should just be self.session.session
                      # but ensure Bytes
     def _ident_default(self):
         return self.session.bsession
 
     def start(self):
         self.engine_stream.on_recv(self.dispatch_result, copy=False)
         self.client_stream.on_recv(self.dispatch_submission, copy=False)
 
         self._notification_handlers = dict(
             registration_notification = self._register_engine,
             unregistration_notification = self._unregister_engine
         )
         self.notifier_stream.on_recv(self.dispatch_notification)
         self.auditor = ioloop.PeriodicCallback(self.audit_timeouts, 2e3, self.loop) # 1 Hz
         self.auditor.start()
         self.log.info("Scheduler started [%s]"%self.scheme_name)
 
     def resume_receiving(self):
         """Resume accepting jobs."""
         self.client_stream.on_recv(self.dispatch_submission, copy=False)
 
     def stop_receiving(self):
         """Stop accepting jobs while there are no engines.
         Leave them in the ZMQ queue."""
         self.client_stream.on_recv(None)
 
     #-----------------------------------------------------------------------
     # [Un]Registration Handling
     #-----------------------------------------------------------------------
 
     
     @util.log_errors
     def dispatch_notification(self, msg):
         """dispatch register/unregister events."""
         try:
             idents,msg = self.session.feed_identities(msg)
         except ValueError:
             self.log.warn("task::Invalid Message: %r",msg)
             return
         try:
             msg = self.session.unserialize(msg)
         except ValueError:
             self.log.warn("task::Unauthorized message from: %r"%idents)
             return
 
         msg_type = msg['header']['msg_type']
 
         handler = self._notification_handlers.get(msg_type, None)
         if handler is None:
             self.log.error("Unhandled message type: %r"%msg_type)
         else:
             try:
-                handler(asbytes(msg['content']['queue']))
+                handler(cast_bytes(msg['content']['queue']))
             except Exception:
                 self.log.error("task::Invalid notification msg: %r", msg, exc_info=True)
 
     def _register_engine(self, uid):
         """New engine with ident `uid` became available."""
         # head of the line:
         self.targets.insert(0,uid)
         self.loads.insert(0,0)
 
         # initialize sets
         self.completed[uid] = set()
         self.failed[uid] = set()
         self.pending[uid] = {}
 
         # rescan the graph:
         self.update_graph(None)
 
     def _unregister_engine(self, uid):
         """Existing engine with ident `uid` became unavailable."""
         if len(self.targets) == 1:
             # this was our only engine
             pass
 
         # handle any potentially finished tasks:
         self.engine_stream.flush()
 
         # don't pop destinations, because they might be used later
         # map(self.destinations.pop, self.completed.pop(uid))
         # map(self.destinations.pop, self.failed.pop(uid))
 
         # prevent this engine from receiving work
         idx = self.targets.index(uid)
         self.targets.pop(idx)
         self.loads.pop(idx)
 
         # wait 5 seconds before cleaning up pending jobs, since the results might
         # still be incoming
         if self.pending[uid]:
             dc = ioloop.DelayedCallback(lambda : self.handle_stranded_tasks(uid), 5000, self.loop)
             dc.start()
         else:
             self.completed.pop(uid)
             self.failed.pop(uid)
 
 
     def handle_stranded_tasks(self, engine):
         """Deal with jobs resident in an engine that died."""
         lost = self.pending[engine]
         for msg_id in lost.keys():
             if msg_id not in self.pending[engine]:
                 # prevent double-handling of messages
                 continue
 
-            raw_msg = lost[msg_id][0]
+            raw_msg = lost[msg_id].raw_msg
             idents,msg = self.session.feed_identities(raw_msg, copy=False)
             parent = self.session.unpack(msg[1].bytes)
             idents = [engine, idents[0]]
 
             # build fake error reply
             try:
                 raise error.EngineError("Engine %r died while running task %r"%(engine, msg_id))
             except:
                 content = error.wrap_exception()
             # build fake header
             header = dict(
                 status='error',
                 engine=engine,
                 date=datetime.now(),
             )
             msg = self.session.msg('apply_reply', content, parent=parent, subheader=header)
             raw_reply = map(zmq.Message, self.session.serialize(msg, ident=idents))
             # and dispatch it
             self.dispatch_result(raw_reply)
 
         # finally scrub completed/failed lists
         self.completed.pop(engine)
         self.failed.pop(engine)
 
 
     #-----------------------------------------------------------------------
     # Job Submission
     #-----------------------------------------------------------------------
     
     
     @util.log_errors
     def dispatch_submission(self, raw_msg):
         """Dispatch job submission to appropriate handlers."""
         # ensure targets up to date:
         self.notifier_stream.flush()
         try:
             idents, msg = self.session.feed_identities(raw_msg, copy=False)
             msg = self.session.unserialize(msg, content=False, copy=False)
         except Exception:
             self.log.error("task::Invaid task msg: %r"%raw_msg, exc_info=True)
             return
 
 
         # send to monitor
         self.mon_stream.send_multipart([b'intask']+raw_msg, copy=False)
 
         header = msg['header']
         msg_id = header['msg_id']
         self.all_ids.add(msg_id)
 
         # get targets as a set of bytes objects
         # from a list of unicode objects
         targets = header.get('targets', [])
-        targets = map(asbytes, targets)
+        targets = map(cast_bytes, targets)
         targets = set(targets)
 
         retries = header.get('retries', 0)
         self.retries[msg_id] = retries
 
         # time dependencies
         after = header.get('after', None)
         if after:
             after = Dependency(after)
             if after.all:
                 if after.success:
                     after = Dependency(after.difference(self.all_completed),
                                 success=after.success,
                                 failure=after.failure,
                                 all=after.all,
                     )
                 if after.failure:
                     after = Dependency(after.difference(self.all_failed),
                                 success=after.success,
                                 failure=after.failure,
                                 all=after.all,
                     )
             if after.check(self.all_completed, self.all_failed):
                 # recast as empty set, if `after` already met,
                 # to prevent unnecessary set comparisons
                 after = MET
         else:
             after = MET
 
         # location dependencies
         follow = Dependency(header.get('follow', []))
 
         # turn timeouts into datetime objects:
         timeout = header.get('timeout', None)
         if timeout:
             # cast to float, because jsonlib returns floats as decimal.Decimal,
             # which timedelta does not accept
             timeout = datetime.now() + timedelta(0,float(timeout),0)
 
         job = Job(msg_id=msg_id, raw_msg=raw_msg, idents=idents, msg=msg,
                  header=header, targets=targets, after=after, follow=follow,
                  timeout=timeout,
         )
 
         # validate and reduce dependencies:
         for dep in after,follow:
             if not dep: # empty dependency
                 continue
             # check valid:
             if msg_id in dep or dep.difference(self.all_ids):
                 self.depending[msg_id] = job
                 return self.fail_unreachable(msg_id, error.InvalidDependency)
             # check if unreachable:
             if dep.unreachable(self.all_completed, self.all_failed):
                 self.depending[msg_id] = job
                 return self.fail_unreachable(msg_id)
 
         if after.check(self.all_completed, self.all_failed):
             # time deps already met, try to run
             if not self.maybe_run(job):
                 # can't run yet
                 if msg_id not in self.all_failed:
                     # could have failed as unreachable
                     self.save_unmet(job)
         else:
             self.save_unmet(job)
 
     def audit_timeouts(self):
         """Audit all waiting tasks for expired timeouts."""
         now = datetime.now()
         for msg_id in self.depending.keys():
             # must recheck, in case one failure cascaded to another:
             if msg_id in self.depending:
                 job = self.depending[msg_id]
                 if job.timeout and job.timeout < now:
                     self.fail_unreachable(msg_id, error.TaskTimeout)
 
     def fail_unreachable(self, msg_id, why=error.ImpossibleDependency):
         """a task has become unreachable, send a reply with an ImpossibleDependency
         error."""
         if msg_id not in self.depending:
             self.log.error("msg %r already failed!", msg_id)
             return
         job = self.depending.pop(msg_id)
         for mid in job.dependents:
             if mid in self.graph:
                 self.graph[mid].remove(msg_id)
 
         try:
             raise why()
         except:
             content = error.wrap_exception()
 
         self.all_done.add(msg_id)
         self.all_failed.add(msg_id)
 
         msg = self.session.send(self.client_stream, 'apply_reply', content,
                                                 parent=job.header, ident=job.idents)
         self.session.send(self.mon_stream, msg, ident=[b'outtask']+job.idents)
 
         self.update_graph(msg_id, success=False)
 
     def maybe_run(self, job):
         """check location dependencies, and run if they are met."""
         msg_id = job.msg_id
         self.log.debug("Attempting to assign task %s", msg_id)
         if not self.targets:
             # no engines, definitely can't run
             return False
         
         if job.follow or job.targets or job.blacklist or self.hwm:
             # we need a can_run filter
             def can_run(idx):
                 # check hwm
                 if self.hwm and self.loads[idx] == self.hwm:
                     return False
                 target = self.targets[idx]
                 # check blacklist
                 if target in job.blacklist:
                     return False
                 # check targets
                 if job.targets and target not in job.targets:
                     return False
                 # check follow
                 return job.follow.check(self.completed[target], self.failed[target])
 
             indices = filter(can_run, range(len(self.targets)))
 
             if not indices:
                 # couldn't run
                 if job.follow.all:
                     # check follow for impossibility
                     dests = set()
                     relevant = set()
                     if job.follow.success:
                         relevant = self.all_completed
                     if job.follow.failure:
                         relevant = relevant.union(self.all_failed)
                     for m in job.follow.intersection(relevant):
                         dests.add(self.destinations[m])
                     if len(dests) > 1:
                         self.depending[msg_id] = job
                         self.fail_unreachable(msg_id)
                         return False
                 if job.targets:
                     # check blacklist+targets for impossibility
                     job.targets.difference_update(job.blacklist)
                     if not job.targets or not job.targets.intersection(self.targets):
                         self.depending[msg_id] = job
                         self.fail_unreachable(msg_id)
                         return False
                 return False
         else:
             indices = None
 
         self.submit_task(job, indices)
         return True
 
     def save_unmet(self, job):
         """Save a message for later submission when its dependencies are met."""
         msg_id = job.msg_id
         self.depending[msg_id] = job
         # track the ids in follow or after, but not those already finished
         for dep_id in job.after.union(job.follow).difference(self.all_done):
             if dep_id not in self.graph:
                 self.graph[dep_id] = set()
             self.graph[dep_id].add(msg_id)
 
     def submit_task(self, job, indices=None):
         """Submit a task to any of a subset of our targets."""
         if indices:
             loads = [self.loads[i] for i in indices]
         else:
             loads = self.loads
         idx = self.scheme(loads)
         if indices:
             idx = indices[idx]
         target = self.targets[idx]
         # print (target, map(str, msg[:3]))
         # send job to the engine
         self.engine_stream.send(target, flags=zmq.SNDMORE, copy=False)
         self.engine_stream.send_multipart(job.raw_msg, copy=False)
         # update load
         self.add_job(idx)
         self.pending[target][job.msg_id] = job
         # notify Hub
         content = dict(msg_id=job.msg_id, engine_id=target.decode('ascii'))
         self.session.send(self.mon_stream, 'task_destination', content=content,
                         ident=[b'tracktask',self.ident])
 
 
     #-----------------------------------------------------------------------
     # Result Handling
     #-----------------------------------------------------------------------
     
     
     @util.log_errors
     def dispatch_result(self, raw_msg):
         """dispatch method for result replies"""
         try:
             idents,msg = self.session.feed_identities(raw_msg, copy=False)
             msg = self.session.unserialize(msg, content=False, copy=False)
             engine = idents[0]
             try:
                 idx = self.targets.index(engine)
             except ValueError:
                 pass # skip load-update for dead engines
             else:
                 self.finish_job(idx)
         except Exception:
             self.log.error("task::Invaid result: %r", raw_msg, exc_info=True)
             return
 
         header = msg['header']
         parent = msg['parent_header']
         if header.get('dependencies_met', True):
             success = (header['status'] == 'ok')
             msg_id = parent['msg_id']
             retries = self.retries[msg_id]
             if not success and retries > 0:
                 # failed
                 self.retries[msg_id] = retries - 1
                 self.handle_unmet_dependency(idents, parent)
             else:
                 del self.retries[msg_id]
                 # relay to client and update graph
                 self.handle_result(idents, parent, raw_msg, success)
                 # send to Hub monitor
                 self.mon_stream.send_multipart([b'outtask']+raw_msg, copy=False)
         else:
             self.handle_unmet_dependency(idents, parent)
 
     def handle_result(self, idents, parent, raw_msg, success=True):
         """handle a real task result, either success or failure"""
         # first, relay result to client
         engine = idents[0]
         client = idents[1]
         # swap_ids for XREP-XREP mirror
         raw_msg[:2] = [client,engine]
         # print (map(str, raw_msg[:4]))
         self.client_stream.send_multipart(raw_msg, copy=False)
         # now, update our data structures
         msg_id = parent['msg_id']
         self.pending[engine].pop(msg_id)
         if success:
             self.completed[engine].add(msg_id)
             self.all_completed.add(msg_id)
         else:
             self.failed[engine].add(msg_id)
             self.all_failed.add(msg_id)
         self.all_done.add(msg_id)
         self.destinations[msg_id] = engine
 
         self.update_graph(msg_id, success)
 
     def handle_unmet_dependency(self, idents, parent):
         """handle an unmet dependency"""
         engine = idents[0]
         msg_id = parent['msg_id']
 
         job = self.pending[engine].pop(msg_id)
         job.blacklist.add(engine)
 
         if job.blacklist == job.targets:
             self.depending[msg_id] = job
             self.fail_unreachable(msg_id)
         elif not self.maybe_run(job):
             # resubmit failed
             if msg_id not in self.all_failed:
                 # put it back in our dependency tree
                 self.save_unmet(job)
 
         if self.hwm:
             try:
                 idx = self.targets.index(engine)
             except ValueError:
                 pass # skip load-update for dead engines
             else:
                 if self.loads[idx] == self.hwm-1:
                     self.update_graph(None)
 
 
 
     def update_graph(self, dep_id=None, success=True):
         """dep_id just finished. Update our dependency
         graph and submit any jobs that just became runable.
 
         Called with dep_id=None to update entire graph for hwm, but without finishing
         a task.
         """
         # print ("\n\n***********")
         # pprint (dep_id)
         # pprint (self.graph)
         # pprint (self.depending)
         # pprint (self.all_completed)
         # pprint (self.all_failed)
         # print ("\n\n***********\n\n")
         # update any jobs that depended on the dependency
         jobs = self.graph.pop(dep_id, [])
 
         # recheck *all* jobs if
         # a) we have HWM and an engine just become no longer full
         # or b) dep_id was given as None
         
         if dep_id is None or self.hwm and any( [ load==self.hwm-1 for load in self.loads ]):
             jobs = self.depending.keys()
         
         for msg_id in sorted(jobs, key=lambda msg_id: self.depending[msg_id].timestamp):
             job = self.depending[msg_id]
 
             if job.after.unreachable(self.all_completed, self.all_failed)\
                     or job.follow.unreachable(self.all_completed, self.all_failed):
                 self.fail_unreachable(msg_id)
 
             elif job.after.check(self.all_completed, self.all_failed): # time deps met, maybe run
                 if self.maybe_run(job):
 
                     self.depending.pop(msg_id)
                     for mid in job.dependents:
                         if mid in self.graph:
                             self.graph[mid].remove(msg_id)
 
     #----------------------------------------------------------------------
     # methods to be overridden by subclasses
     #----------------------------------------------------------------------
 
     def add_job(self, idx):
         """Called after self.targets[idx] just got the job with header.
         Override with subclasses.  The default ordering is simple LRU.
         The default loads are the number of outstanding jobs."""
         self.loads[idx] += 1
         for lis in (self.targets, self.loads):
             lis.append(lis.pop(idx))
 
 
     def finish_job(self, idx):
         """Called after self.targets[idx] just finished a job.
         Override with subclasses."""
         self.loads[idx] -= 1
 
 
 
 def launch_scheduler(in_addr, out_addr, mon_addr, not_addr, config=None,
                         logname='root', log_url=None, loglevel=logging.DEBUG,
                         identity=b'task', in_thread=False):
 
     ZMQStream = zmqstream.ZMQStream
 
     if config:
         # unwrap dict back into Config
         config = Config(config)
 
     if in_thread:
         # use instance() to get the same Context/Loop as our parent
         ctx = zmq.Context.instance()
         loop = ioloop.IOLoop.instance()
     else:
         # in a process, don't use instance()
         # for safety with multiprocessing
         ctx = zmq.Context()
         loop = ioloop.IOLoop()
     ins = ZMQStream(ctx.socket(zmq.ROUTER),loop)
     ins.setsockopt(zmq.IDENTITY, identity)
     ins.bind(in_addr)
 
     outs = ZMQStream(ctx.socket(zmq.ROUTER),loop)
     outs.setsockopt(zmq.IDENTITY, identity)
     outs.bind(out_addr)
     mons = zmqstream.ZMQStream(ctx.socket(zmq.PUB),loop)
     mons.connect(mon_addr)
     nots = zmqstream.ZMQStream(ctx.socket(zmq.SUB),loop)
     nots.setsockopt(zmq.SUBSCRIBE, b'')
     nots.connect(not_addr)
 
     # setup logging.
     if in_thread:
         log = Application.instance().log
     else:
         if log_url:
             log = connect_logger(logname, ctx, log_url, root="scheduler", loglevel=loglevel)
         else:
             log = local_logger(logname, loglevel)
 
     scheduler = TaskScheduler(client_stream=ins, engine_stream=outs,
                             mon_stream=mons, notifier_stream=nots,
                             loop=loop, log=log,
                             config=config)
     scheduler.start()
     if not in_thread:
         try:
             loop.start()
         except KeyboardInterrupt:
             scheduler.log.critical("Interrupted, exiting...")
 

IPython/parallel/controller/sqlitedb.py

@@ -1,411 +1,412 b''
 """A TaskRecord backend using sqlite3
 
 Authors:
 
 * Min RK
 """
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 import json
 import os
 import cPickle as pickle
 from datetime import datetime
 
 try:
     import sqlite3
 except ImportError:
     sqlite3 = None
 
 from zmq.eventloop import ioloop
 
 from IPython.utils.traitlets import Unicode, Instance, List, Dict
 from .dictdb import BaseDB
 from IPython.utils.jsonutil import date_default, extract_dates, squash_dates
 
 #-----------------------------------------------------------------------------
 # SQLite operators, adapters, and converters
 #-----------------------------------------------------------------------------
 
 try:
     buffer
 except NameError:
     # py3k
     buffer = memoryview
 
 operators = {
  '$lt' : "<",
  '$gt' : ">",
  # null is handled weird with ==,!=
  '$eq' : "=",
  '$ne' : "!=",
  '$lte': "<=",
  '$gte': ">=",
  '$in' : ('=', ' OR '),
  '$nin': ('!=', ' AND '),
  # '$all': None,
  # '$mod': None,
  # '$exists' : None
 }
 null_operators = {
 '=' : "IS NULL",
 '!=' : "IS NOT NULL",
 }
 
 def _adapt_dict(d):
     return json.dumps(d, default=date_default)
 
 def _convert_dict(ds):
     if ds is None:
         return ds
     else:
         if isinstance(ds, bytes):
             # If I understand the sqlite doc correctly, this will always be utf8
             ds = ds.decode('utf8')
         return extract_dates(json.loads(ds))
 
 def _adapt_bufs(bufs):
     # this is *horrible*
     # copy buffers into single list and pickle it:
     if bufs and isinstance(bufs[0], (bytes, buffer)):
         return sqlite3.Binary(pickle.dumps(map(bytes, bufs),-1))
     elif bufs:
         return bufs
     else:
         return None
 
 def _convert_bufs(bs):
     if bs is None:
         return []
     else:
         return pickle.loads(bytes(bs))
 
 #-----------------------------------------------------------------------------
 # SQLiteDB class
 #-----------------------------------------------------------------------------
 
 class SQLiteDB(BaseDB):
     """SQLite3 TaskRecord backend."""
 
     filename = Unicode('tasks.db', config=True,
         help="""The filename of the sqlite task database. [default: 'tasks.db']""")
     location = Unicode('', config=True,
         help="""The directory containing the sqlite task database.  The default
         is to use the cluster_dir location.""")
     table = Unicode("", config=True,
         help="""The SQLite Table to use for storing tasks for this session. If unspecified,
         a new table will be created with the Hub's IDENT.  Specifying the table will result
         in tasks from previous sessions being available via Clients' db_query and
         get_result methods.""")
 
     if sqlite3 is not None:
         _db = Instance('sqlite3.Connection')
     else:
         _db = None
     # the ordered list of column names
     _keys = List(['msg_id' ,
             'header' ,
             'content',
             'buffers',
             'submitted',
             'client_uuid' ,
             'engine_uuid' ,
             'started',
             'completed',
             'resubmitted',
             'received',
             'result_header' ,
             'result_content' ,
             'result_buffers' ,
             'queue' ,
             'pyin' ,
             'pyout',
             'pyerr',
             'stdout',
             'stderr',
         ])
     # sqlite datatypes for checking that db is current format
     _types = Dict({'msg_id' : 'text' ,
             'header' : 'dict text',
             'content' : 'dict text',
             'buffers' : 'bufs blob',
             'submitted' : 'timestamp',
             'client_uuid' : 'text',
             'engine_uuid' : 'text',
             'started' : 'timestamp',
             'completed' : 'timestamp',
-            'resubmitted' : 'timestamp',
+            'resubmitted' : 'text',
             'received' : 'timestamp',
             'result_header' : 'dict text',
             'result_content' : 'dict text',
             'result_buffers' : 'bufs blob',
             'queue' : 'text',
             'pyin' : 'text',
             'pyout' : 'text',
             'pyerr' : 'text',
             'stdout' : 'text',
             'stderr' : 'text',
         })
 
     def __init__(self, **kwargs):
         super(SQLiteDB, self).__init__(**kwargs)
         if sqlite3 is None:
             raise ImportError("SQLiteDB requires sqlite3")
         if not self.table:
             # use session, and prefix _, since starting with # is illegal
             self.table = '_'+self.session.replace('-','_')
         if not self.location:
             # get current profile
             from IPython.core.application import BaseIPythonApplication
             if BaseIPythonApplication.initialized():
                 app = BaseIPythonApplication.instance()
                 if app.profile_dir is not None:
                     self.location = app.profile_dir.location
                 else:
                     self.location = u'.'
             else:
                 self.location = u'.'
         self._init_db()
 
         # register db commit as 2s periodic callback
         # to prevent clogging pipes
         # assumes we are being run in a zmq ioloop app
         loop = ioloop.IOLoop.instance()
         pc = ioloop.PeriodicCallback(self._db.commit, 2000, loop)
         pc.start()
 
     def _defaults(self, keys=None):
         """create an empty record"""
         d = {}
         keys = self._keys if keys is None else keys
         for key in keys:
             d[key] = None
         return d
 
     def _check_table(self):
         """Ensure that an incorrect table doesn't exist
 
         If a bad (old) table does exist, return False
         """
         cursor = self._db.execute("PRAGMA table_info(%s)"%self.table)
         lines = cursor.fetchall()
         if not lines:
             # table does not exist
             return True
         types = {}
         keys = []
         for line in lines:
             keys.append(line[1])
             types[line[1]] = line[2]
         if self._keys != keys:
             # key mismatch
             self.log.warn('keys mismatch')
             return False
         for key in self._keys:
             if types[key] != self._types[key]:
                 self.log.warn(
                     'type mismatch: %s: %s != %s'%(key,types[key],self._types[key])
                 )
                 return False
         return True
 
     def _init_db(self):
         """Connect to the database and get new session number."""
         # register adapters
         sqlite3.register_adapter(dict, _adapt_dict)
         sqlite3.register_converter('dict', _convert_dict)
         sqlite3.register_adapter(list, _adapt_bufs)
         sqlite3.register_converter('bufs', _convert_bufs)
         # connect to the db
         dbfile = os.path.join(self.location, self.filename)
         self._db = sqlite3.connect(dbfile, detect_types=sqlite3.PARSE_DECLTYPES,
             # isolation_level = None)#,
              cached_statements=64)
         # print dir(self._db)
-        first_table = self.table
+        first_table = previous_table = self.table
         i=0
         while not self._check_table():
             i+=1
             self.table = first_table+'_%i'%i
             self.log.warn(
                 "Table %s exists and doesn't match db format, trying %s"%
-                (first_table,self.table)
+                (previous_table, self.table)
             )
+            previous_table = self.table
 
         self._db.execute("""CREATE TABLE IF NOT EXISTS %s
                 (msg_id text PRIMARY KEY,
                 header dict text,
                 content dict text,
                 buffers bufs blob,
                 submitted timestamp,
                 client_uuid text,
                 engine_uuid text,
                 started timestamp,
                 completed timestamp,
-                resubmitted timestamp,
+                resubmitted text,
                 received timestamp,
                 result_header dict text,
                 result_content dict text,
                 result_buffers bufs blob,
                 queue text,
                 pyin text,
                 pyout text,
                 pyerr text,
                 stdout text,
                 stderr text)
                 """%self.table)
         self._db.commit()
 
     def _dict_to_list(self, d):
         """turn a mongodb-style record dict into a list."""
 
         return [ d[key] for key in self._keys ]
 
     def _list_to_dict(self, line, keys=None):
         """Inverse of dict_to_list"""
         keys = self._keys if keys is None else keys
         d = self._defaults(keys)
         for key,value in zip(keys, line):
             d[key] = value
 
         return d
 
     def _render_expression(self, check):
         """Turn a mongodb-style search dict into an SQL query."""
         expressions = []
         args = []
 
         skeys = set(check.keys())
         skeys.difference_update(set(self._keys))
         skeys.difference_update(set(['buffers', 'result_buffers']))
         if skeys:
             raise KeyError("Illegal testing key(s): %s"%skeys)
 
         for name,sub_check in check.iteritems():
             if isinstance(sub_check, dict):
                 for test,value in sub_check.iteritems():
                     try:
                         op = operators[test]
                     except KeyError:
                         raise KeyError("Unsupported operator: %r"%test)
                     if isinstance(op, tuple):
                         op, join = op
 
                     if value is None and op in null_operators:
                         expr = "%s %s" % (name, null_operators[op])
                     else:
                         expr = "%s %s ?"%(name, op)
                         if isinstance(value, (tuple,list)):
                             if op in null_operators and any([v is None for v in value]):
                                 # equality tests don't work with NULL
                                 raise ValueError("Cannot use %r test with NULL values on SQLite backend"%test)
                             expr = '( %s )'%( join.join([expr]*len(value)) )
                             args.extend(value)
                         else:
                             args.append(value)
                     expressions.append(expr)
             else:
                 # it's an equality check
                 if sub_check is None:
                     expressions.append("%s IS NULL" % name)
                 else:
                     expressions.append("%s = ?"%name)
                     args.append(sub_check)
 
         expr = " AND ".join(expressions)
         return expr, args
 
     def add_record(self, msg_id, rec):
         """Add a new Task Record, by msg_id."""
         d = self._defaults()
         d.update(rec)
         d['msg_id'] = msg_id
         line = self._dict_to_list(d)
         tups = '(%s)'%(','.join(['?']*len(line)))
         self._db.execute("INSERT INTO %s VALUES %s"%(self.table, tups), line)
         # self._db.commit()
 
     def get_record(self, msg_id):
         """Get a specific Task Record, by msg_id."""
         cursor = self._db.execute("""SELECT * FROM %s WHERE msg_id==?"""%self.table, (msg_id,))
         line = cursor.fetchone()
         if line is None:
             raise KeyError("No such msg: %r"%msg_id)
         return self._list_to_dict(line)
 
     def update_record(self, msg_id, rec):
         """Update the data in an existing record."""
         query = "UPDATE %s SET "%self.table
         sets = []
         keys = sorted(rec.keys())
         values = []
         for key in keys:
             sets.append('%s = ?'%key)
             values.append(rec[key])
         query += ', '.join(sets)
         query += ' WHERE msg_id == ?'
         values.append(msg_id)
         self._db.execute(query, values)
         # self._db.commit()
 
     def drop_record(self, msg_id):
         """Remove a record from the DB."""
         self._db.execute("""DELETE FROM %s WHERE msg_id==?"""%self.table, (msg_id,))
         # self._db.commit()
 
     def drop_matching_records(self, check):
         """Remove a record from the DB."""
         expr,args = self._render_expression(check)
         query = "DELETE FROM %s WHERE %s"%(self.table, expr)
         self._db.execute(query,args)
         # self._db.commit()
 
     def find_records(self, check, keys=None):
         """Find records matching a query dict, optionally extracting subset of keys.
 
         Returns list of matching records.
 
         Parameters
         ----------
 
         check: dict
             mongodb-style query argument
         keys: list of strs [optional]
             if specified, the subset of keys to extract.  msg_id will *always* be
             included.
         """
         if keys:
             bad_keys = [ key for key in keys if key not in self._keys ]
             if bad_keys:
                 raise KeyError("Bad record key(s): %s"%bad_keys)
 
         if keys:
             # ensure msg_id is present and first:
             if 'msg_id' in keys:
                 keys.remove('msg_id')
             keys.insert(0, 'msg_id')
             req = ', '.join(keys)
         else:
             req = '*'
         expr,args = self._render_expression(check)
         query = """SELECT %s FROM %s WHERE %s"""%(req, self.table, expr)
         cursor = self._db.execute(query, args)
         matches = cursor.fetchall()
         records = []
         for line in matches:
             rec = self._list_to_dict(line, keys)
             records.append(rec)
         return records
 
     def get_history(self):
         """get all msg_ids, ordered by time submitted."""
         query = """SELECT msg_id FROM %s ORDER by submitted ASC"""%self.table
         cursor = self._db.execute(query)
         # will be a list of length 1 tuples
         return [ tup[0] for tup in cursor.fetchall()]
 
 __all__ = ['SQLiteDB']
  No newline at end of file

IPython/parallel/engine/engine.py

@@ -1,234 +1,237 b''
 """A simple engine that talks to a controller over 0MQ.
 it handles registration, etc. and launches a kernel
 connected to the Controller's Schedulers.
 
 Authors:
 
 * Min RK
 """
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2010-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 from __future__ import print_function
 
 import sys
 import time
 from getpass import getpass
 
 import zmq
 from zmq.eventloop import ioloop, zmqstream
 
 from IPython.external.ssh import tunnel
 # internal
 from IPython.utils.traitlets import (
     Instance, Dict, Integer, Type, CFloat, Unicode, CBytes, Bool
 )
-from IPython.utils import py3compat
+from IPython.utils.py3compat import cast_bytes
 
 from IPython.parallel.controller.heartmonitor import Heart
 from IPython.parallel.factory import RegistrationFactory
-from IPython.parallel.util import disambiguate_url, asbytes
+from IPython.parallel.util import disambiguate_url
 
 from IPython.zmq.session import Message
-
-from .streamkernel import Kernel
+from IPython.zmq.ipkernel import Kernel
 
 class EngineFactory(RegistrationFactory):
     """IPython engine"""
 
     # configurables:
     out_stream_factory=Type('IPython.zmq.iostream.OutStream', config=True,
         help="""The OutStream for handling stdout/err.
         Typically 'IPython.zmq.iostream.OutStream'""")
     display_hook_factory=Type('IPython.zmq.displayhook.ZMQDisplayHook', config=True,
         help="""The class for handling displayhook.
         Typically 'IPython.zmq.displayhook.ZMQDisplayHook'""")
     location=Unicode(config=True,
         help="""The location (an IP address) of the controller.  This is
         used for disambiguating URLs, to determine whether
         loopback should be used to connect or the public address.""")
     timeout=CFloat(2,config=True,
         help="""The time (in seconds) to wait for the Controller to respond
         to registration requests before giving up.""")
     sshserver=Unicode(config=True,
         help="""The SSH server to use for tunneling connections to the Controller.""")
     sshkey=Unicode(config=True,
         help="""The SSH private key file to use when tunneling connections to the Controller.""")
     paramiko=Bool(sys.platform == 'win32', config=True,
         help="""Whether to use paramiko instead of openssh for tunnels.""")
 
     # not configurable:
     user_ns=Dict()
     id=Integer(allow_none=True)
     registrar=Instance('zmq.eventloop.zmqstream.ZMQStream')
     kernel=Instance(Kernel)
 
     bident = CBytes()
     ident = Unicode()
     def _ident_changed(self, name, old, new):
-        self.bident = asbytes(new)
+        self.bident = cast_bytes(new)
     using_ssh=Bool(False)
 
 
     def __init__(self, **kwargs):
         super(EngineFactory, self).__init__(**kwargs)
         self.ident = self.session.session
 
     def init_connector(self):
         """construct connection function, which handles tunnels."""
         self.using_ssh = bool(self.sshkey or self.sshserver)
 
         if self.sshkey and not self.sshserver:
             # We are using ssh directly to the controller, tunneling localhost to localhost
             self.sshserver = self.url.split('://')[1].split(':')[0]
 
         if self.using_ssh:
             if tunnel.try_passwordless_ssh(self.sshserver, self.sshkey, self.paramiko):
                 password=False
             else:
                 password = getpass("SSH Password for %s: "%self.sshserver)
         else:
             password = False
 
         def connect(s, url):
             url = disambiguate_url(url, self.location)
             if self.using_ssh:
                 self.log.debug("Tunneling connection to %s via %s"%(url, self.sshserver))
                 return tunnel.tunnel_connection(s, url, self.sshserver,
                             keyfile=self.sshkey, paramiko=self.paramiko,
                             password=password,
                 )
             else:
                 return s.connect(url)
 
         def maybe_tunnel(url):
             """like connect, but don't complete the connection (for use by heartbeat)"""
             url = disambiguate_url(url, self.location)
             if self.using_ssh:
                 self.log.debug("Tunneling connection to %s via %s"%(url, self.sshserver))
                 url,tunnelobj = tunnel.open_tunnel(url, self.sshserver,
                             keyfile=self.sshkey, paramiko=self.paramiko,
                             password=password,
                 )
             return url
         return connect, maybe_tunnel
 
     def register(self):
         """send the registration_request"""
 
         self.log.info("Registering with controller at %s"%self.url)
         ctx = self.context
         connect,maybe_tunnel = self.init_connector()
         reg = ctx.socket(zmq.DEALER)
         reg.setsockopt(zmq.IDENTITY, self.bident)
         connect(reg, self.url)
         self.registrar = zmqstream.ZMQStream(reg, self.loop)
 
 
         content = dict(queue=self.ident, heartbeat=self.ident, control=self.ident)
         self.registrar.on_recv(lambda msg: self.complete_registration(msg, connect, maybe_tunnel))
         # print (self.session.key)
         self.session.send(self.registrar, "registration_request",content=content)
 
     def complete_registration(self, msg, connect, maybe_tunnel):
         # print msg
         self._abort_dc.stop()
         ctx = self.context
         loop = self.loop
         identity = self.bident
         idents,msg = self.session.feed_identities(msg)
         msg = Message(self.session.unserialize(msg))
 
         if msg.content.status == 'ok':
             self.id = int(msg.content.id)
 
             # launch heartbeat
             hb_addrs = msg.content.heartbeat
 
             # possibly forward hb ports with tunnels
             hb_addrs = [ maybe_tunnel(addr) for addr in hb_addrs ]
             heart = Heart(*map(str, hb_addrs), heart_id=identity)
             heart.start()
 
             # create Shell Streams (MUX, Task, etc.):
             queue_addr = msg.content.mux
             shell_addrs = [ str(queue_addr) ]
             task_addr = msg.content.task
             if task_addr:
                 shell_addrs.append(str(task_addr))
 
             # Uncomment this to go back to two-socket model
             # shell_streams = []
             # for addr in shell_addrs:
             #     stream = zmqstream.ZMQStream(ctx.socket(zmq.ROUTER), loop)
             #     stream.setsockopt(zmq.IDENTITY, identity)
             #     stream.connect(disambiguate_url(addr, self.location))
             #     shell_streams.append(stream)
 
             # Now use only one shell stream for mux and tasks
             stream = zmqstream.ZMQStream(ctx.socket(zmq.ROUTER), loop)
             stream.setsockopt(zmq.IDENTITY, identity)
             shell_streams = [stream]
             for addr in shell_addrs:
                 connect(stream, addr)
             # end single stream-socket
 
             # control stream:
             control_addr = str(msg.content.control)
             control_stream = zmqstream.ZMQStream(ctx.socket(zmq.ROUTER), loop)
             control_stream.setsockopt(zmq.IDENTITY, identity)
             connect(control_stream, control_addr)
 
             # create iopub stream:
             iopub_addr = msg.content.iopub
-            iopub_stream = zmqstream.ZMQStream(ctx.socket(zmq.PUB), loop)
-            iopub_stream.setsockopt(zmq.IDENTITY, identity)
-            connect(iopub_stream, iopub_addr)
-
-            # # Redirect input streams and set a display hook.
+            iopub_socket = ctx.socket(zmq.PUB)
+            iopub_socket.setsockopt(zmq.IDENTITY, identity)
+            connect(iopub_socket, iopub_addr)
+
+            # disable history:
+            self.config.HistoryManager.hist_file = ':memory:'
+            
+            # Redirect input streams and set a display hook.
             if self.out_stream_factory:
-                sys.stdout = self.out_stream_factory(self.session, iopub_stream, u'stdout')
-                sys.stdout.topic = py3compat.cast_bytes('engine.%i.stdout' % self.id)
-                sys.stderr = self.out_stream_factory(self.session, iopub_stream, u'stderr')
-                sys.stderr.topic = py3compat.cast_bytes('engine.%i.stderr' % self.id)
+                sys.stdout = self.out_stream_factory(self.session, iopub_socket, u'stdout')
+                sys.stdout.topic = cast_bytes('engine.%i.stdout' % self.id)
+                sys.stderr = self.out_stream_factory(self.session, iopub_socket, u'stderr')
+                sys.stderr.topic = cast_bytes('engine.%i.stderr' % self.id)
             if self.display_hook_factory:
-                sys.displayhook = self.display_hook_factory(self.session, iopub_stream)
-                sys.displayhook.topic = py3compat.cast_bytes('engine.%i.pyout' % self.id)
+                sys.displayhook = self.display_hook_factory(self.session, iopub_socket)
+                sys.displayhook.topic = cast_bytes('engine.%i.pyout' % self.id)
 
             self.kernel = Kernel(config=self.config, int_id=self.id, ident=self.ident, session=self.session,
-                    control_stream=control_stream, shell_streams=shell_streams, iopub_stream=iopub_stream,
-                    loop=loop, user_ns = self.user_ns, log=self.log)
+                    control_stream=control_stream, shell_streams=shell_streams, iopub_socket=iopub_socket,
+                    loop=loop, user_ns=self.user_ns, log=self.log)
+            self.kernel.shell.display_pub.topic = cast_bytes('engine.%i.displaypub' % self.id)
             self.kernel.start()
 
 
         else:
             self.log.fatal("Registration Failed: %s"%msg)
             raise Exception("Registration Failed: %s"%msg)
 
         self.log.info("Completed registration with id %i"%self.id)
 
 
     def abort(self):
         self.log.fatal("Registration timed out after %.1f seconds"%self.timeout)
         if self.url.startswith('127.'):
             self.log.fatal("""
             If the controller and engines are not on the same machine,
             you will have to instruct the controller to listen on an external IP (in ipcontroller_config.py):
                 c.HubFactory.ip='*' # for all interfaces, internal and external
                 c.HubFactory.ip='192.168.1.101' # or any interface that the engines can see
             or tunnel connections via ssh.
             """)
         self.session.send(self.registrar, "unregistration_request", content=dict(id=self.id))
         time.sleep(1)
         sys.exit(255)
 
     def start(self):
         dc = ioloop.DelayedCallback(self.register, 0, self.loop)
         dc.start()
         self._abort_dc = ioloop.DelayedCallback(self.abort, self.timeout*1000, self.loop)
         self._abort_dc.start()
 

IPython/parallel/tests/__init__.py

@@ -1,122 +1,122 b''
 """toplevel setup/teardown for parallel tests."""
 
 #-------------------------------------------------------------------------------
 #  Copyright (C) 2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-------------------------------------------------------------------------------
 
 #-------------------------------------------------------------------------------
 # Imports
 #-------------------------------------------------------------------------------
 
 import os
 import tempfile
 import time
 from subprocess import Popen
 
 from IPython.utils.path import get_ipython_dir
 from IPython.parallel import Client
 from IPython.parallel.apps.launcher import (LocalProcessLauncher,
                                                   ipengine_cmd_argv,
                                                   ipcontroller_cmd_argv,
                                                   SIGKILL,
                                                   ProcessStateError,
 )
 
 # globals
 launchers = []
 blackhole = open(os.devnull, 'w')
 
 # Launcher class
 class TestProcessLauncher(LocalProcessLauncher):
     """subclass LocalProcessLauncher, to prevent extra sockets and threads being created on Windows"""
     def start(self):
         if self.state == 'before':
             self.process = Popen(self.args,
                 stdout=blackhole, stderr=blackhole,
                 env=os.environ,
                 cwd=self.work_dir
             )
             self.notify_start(self.process.pid)
             self.poll = self.process.poll
         else:
             s = 'The process was already started and has state: %r' % self.state
             raise ProcessStateError(s)
 
 # nose setup/teardown
 
 def setup():
     cluster_dir = os.path.join(get_ipython_dir(), 'profile_iptest')
     engine_json = os.path.join(cluster_dir, 'security', 'ipcontroller-engine.json')
     client_json = os.path.join(cluster_dir, 'security', 'ipcontroller-client.json')
     for json in (engine_json, client_json):
         if os.path.exists(json):
             os.remove(json)
     
     cp = TestProcessLauncher()
     cp.cmd_and_args = ipcontroller_cmd_argv + \
                 ['--profile=iptest', '--log-level=50', '--ping=250']
     cp.start()
     launchers.append(cp)
     tic = time.time()
     while not os.path.exists(engine_json) or not os.path.exists(client_json):
         if cp.poll() is not None:
             print cp.poll()
             raise RuntimeError("The test controller failed to start.")
-        elif time.time()-tic > 10:
+        elif time.time()-tic > 15:
             raise RuntimeError("Timeout waiting for the test controller to start.")
         time.sleep(0.1)
     add_engines(1)
 
 def add_engines(n=1, profile='iptest', total=False):
     """add a number of engines to a given profile.
     
     If total is True, then already running engines are counted, and only
     the additional engines necessary (if any) are started.
     """
     rc = Client(profile=profile)
     base = len(rc)
     
     if total:
         n = max(n - base, 0)
     
     eps = []
     for i in range(n):
         ep = TestProcessLauncher()
         ep.cmd_and_args = ipengine_cmd_argv + ['--profile=%s'%profile, '--log-level=50']
         ep.start()
         launchers.append(ep)
         eps.append(ep)
     tic = time.time()
     while len(rc) < base+n:
         if any([ ep.poll() is not None for ep in eps ]):
             raise RuntimeError("A test engine failed to start.")
-        elif time.time()-tic > 10:
+        elif time.time()-tic > 15:
             raise RuntimeError("Timeout waiting for engines to connect.")
         time.sleep(.1)
         rc.spin()
     rc.close()
     return eps
 
 def teardown():
     time.sleep(1)
     while launchers:
         p = launchers.pop()
         if p.poll() is None:
             try:
                 p.stop()
             except Exception, e:
                 print e
                 pass
         if p.poll() is None:
             time.sleep(.25)
         if p.poll() is None:
             try:
                 print 'cleaning up test process...'
                 p.signal(SIGKILL)
             except:
                 print "couldn't shutdown process: ", p
     blackhole.close()
     

IPython/parallel/tests/test_asyncresult.py

@@ -1,205 +1,205 b''
 """Tests for asyncresult.py
 
 Authors:
 
 * Min RK
 """
 
 #-------------------------------------------------------------------------------
 #  Copyright (C) 2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-------------------------------------------------------------------------------
 
 #-------------------------------------------------------------------------------
 # Imports
 #-------------------------------------------------------------------------------
 
 import time
 
 from IPython.parallel.error import TimeoutError
 
 from IPython.parallel import error, Client
 from IPython.parallel.tests import add_engines
 from .clienttest import ClusterTestCase
 
 def setup():
     add_engines(2, total=True)
 
 def wait(n):
     import time
     time.sleep(n)
     return n
 
 class AsyncResultTest(ClusterTestCase):
 
     def test_single_result_view(self):
         """various one-target views get the right value for single_result"""
         eid = self.client.ids[-1]
         ar = self.client[eid].apply_async(lambda : 42)
         self.assertEquals(ar.get(), 42)
         ar = self.client[[eid]].apply_async(lambda : 42)
         self.assertEquals(ar.get(), [42])
         ar = self.client[-1:].apply_async(lambda : 42)
         self.assertEquals(ar.get(), [42])
 
     def test_get_after_done(self):
         ar = self.client[-1].apply_async(lambda : 42)
         ar.wait()
         self.assertTrue(ar.ready())
         self.assertEquals(ar.get(), 42)
         self.assertEquals(ar.get(), 42)
 
     def test_get_before_done(self):
         ar = self.client[-1].apply_async(wait, 0.1)
         self.assertRaises(TimeoutError, ar.get, 0)
         ar.wait(0)
         self.assertFalse(ar.ready())
         self.assertEquals(ar.get(), 0.1)
 
     def test_get_after_error(self):
         ar = self.client[-1].apply_async(lambda : 1/0)
         ar.wait(10)
         self.assertRaisesRemote(ZeroDivisionError, ar.get)
         self.assertRaisesRemote(ZeroDivisionError, ar.get)
         self.assertRaisesRemote(ZeroDivisionError, ar.get_dict)
     
     def test_get_dict(self):
         n = len(self.client)
         ar = self.client[:].apply_async(lambda : 5)
         self.assertEquals(ar.get(), [5]*n)
         d = ar.get_dict()
         self.assertEquals(sorted(d.keys()), sorted(self.client.ids))
         for eid,r in d.iteritems():
             self.assertEquals(r, 5)
     
     def test_list_amr(self):
         ar = self.client.load_balanced_view().map_async(wait, [0.1]*5)
         rlist = list(ar)
     
     def test_getattr(self):
         ar = self.client[:].apply_async(wait, 0.5)
         self.assertRaises(AttributeError, lambda : ar._foo)
         self.assertRaises(AttributeError, lambda : ar.__length_hint__())
         self.assertRaises(AttributeError, lambda : ar.foo)
         self.assertRaises(AttributeError, lambda : ar.engine_id)
         self.assertFalse(hasattr(ar, '__length_hint__'))
         self.assertFalse(hasattr(ar, 'foo'))
         self.assertFalse(hasattr(ar, 'engine_id'))
         ar.get(5)
         self.assertRaises(AttributeError, lambda : ar._foo)
         self.assertRaises(AttributeError, lambda : ar.__length_hint__())
         self.assertRaises(AttributeError, lambda : ar.foo)
         self.assertTrue(isinstance(ar.engine_id, list))
         self.assertEquals(ar.engine_id, ar['engine_id'])
         self.assertFalse(hasattr(ar, '__length_hint__'))
         self.assertFalse(hasattr(ar, 'foo'))
         self.assertTrue(hasattr(ar, 'engine_id'))
 
     def test_getitem(self):
         ar = self.client[:].apply_async(wait, 0.5)
         self.assertRaises(TimeoutError, lambda : ar['foo'])
         self.assertRaises(TimeoutError, lambda : ar['engine_id'])
         ar.get(5)
         self.assertRaises(KeyError, lambda : ar['foo'])
         self.assertTrue(isinstance(ar['engine_id'], list))
         self.assertEquals(ar.engine_id, ar['engine_id'])
     
     def test_single_result(self):
         ar = self.client[-1].apply_async(wait, 0.5)
         self.assertRaises(TimeoutError, lambda : ar['foo'])
         self.assertRaises(TimeoutError, lambda : ar['engine_id'])
         self.assertTrue(ar.get(5) == 0.5)
         self.assertTrue(isinstance(ar['engine_id'], int))
         self.assertTrue(isinstance(ar.engine_id, int))
         self.assertEquals(ar.engine_id, ar['engine_id'])
     
     def test_abort(self):
         e = self.client[-1]
         ar = e.execute('import time; time.sleep(1)', block=False)
         ar2 = e.apply_async(lambda : 2)
         ar2.abort()
         self.assertRaises(error.TaskAborted, ar2.get)
         ar.get()
     
     def test_len(self):
         v = self.client.load_balanced_view()
         ar = v.map_async(lambda x: x, range(10))
         self.assertEquals(len(ar), 10)
         ar = v.apply_async(lambda x: x, range(10))
         self.assertEquals(len(ar), 1)
         ar = self.client[:].apply_async(lambda x: x, range(10))
         self.assertEquals(len(ar), len(self.client.ids))
     
     def test_wall_time_single(self):
         v = self.client.load_balanced_view()
         ar = v.apply_async(time.sleep, 0.25)
         self.assertRaises(TimeoutError, getattr, ar, 'wall_time')
         ar.get(2)
         self.assertTrue(ar.wall_time < 1.)
         self.assertTrue(ar.wall_time > 0.2)
 
     def test_wall_time_multi(self):
         self.minimum_engines(4)
         v = self.client[:]
         ar = v.apply_async(time.sleep, 0.25)
         self.assertRaises(TimeoutError, getattr, ar, 'wall_time')
         ar.get(2)
         self.assertTrue(ar.wall_time < 1.)
         self.assertTrue(ar.wall_time > 0.2)
 
     def test_serial_time_single(self):
         v = self.client.load_balanced_view()
         ar = v.apply_async(time.sleep, 0.25)
         self.assertRaises(TimeoutError, getattr, ar, 'serial_time')
         ar.get(2)
-        self.assertTrue(ar.serial_time < 0.5)
+        self.assertTrue(ar.serial_time < 1.)
         self.assertTrue(ar.serial_time > 0.2)
 
     def test_serial_time_multi(self):
         self.minimum_engines(4)
         v = self.client[:]
         ar = v.apply_async(time.sleep, 0.25)
         self.assertRaises(TimeoutError, getattr, ar, 'serial_time')
         ar.get(2)
         self.assertTrue(ar.serial_time < 2.)
         self.assertTrue(ar.serial_time > 0.8)
 
     def test_elapsed_single(self):
         v = self.client.load_balanced_view()
         ar = v.apply_async(time.sleep, 0.25)
         while not ar.ready():
             time.sleep(0.01)
-            self.assertTrue(ar.elapsed < 0.3)
-        self.assertTrue(ar.elapsed < 0.3)
+            self.assertTrue(ar.elapsed < 1)
+        self.assertTrue(ar.elapsed < 1)
         ar.get(2)
 
     def test_elapsed_multi(self):
         v = self.client[:]
         ar = v.apply_async(time.sleep, 0.25)
         while not ar.ready():
             time.sleep(0.01)
-            self.assertTrue(ar.elapsed < 0.3)
-        self.assertTrue(ar.elapsed < 0.3)
+            self.assertTrue(ar.elapsed < 1)
+        self.assertTrue(ar.elapsed < 1)
         ar.get(2)
 
     def test_hubresult_timestamps(self):
         self.minimum_engines(4)
         v = self.client[:]
         ar = v.apply_async(time.sleep, 0.25)
         ar.get(2)
         rc2 = Client(profile='iptest')
         # must have try/finally to close second Client, otherwise
         # will have dangling sockets causing problems
         try:
             time.sleep(0.25)
             hr = rc2.get_result(ar.msg_ids)
             self.assertTrue(hr.elapsed > 0., "got bad elapsed: %s" % hr.elapsed)
             hr.get(1)
             self.assertTrue(hr.wall_time < ar.wall_time + 0.2, "got bad wall_time: %s > %s" % (hr.wall_time, ar.wall_time))
             self.assertEquals(hr.serial_time, ar.serial_time)
         finally:
             rc2.close()
 
 

IPython/parallel/tests/test_client.py

@@ -1,349 +1,387 b''
 """Tests for parallel client.py
 
 Authors:
 
 * Min RK
 """
 
 #-------------------------------------------------------------------------------
 #  Copyright (C) 2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-------------------------------------------------------------------------------
 
 #-------------------------------------------------------------------------------
 # Imports
 #-------------------------------------------------------------------------------
 
 from __future__ import division
 
 import time
 from datetime import datetime
 from tempfile import mktemp
 
 import zmq
 
 from IPython.parallel.client import client as clientmod
 from IPython.parallel import error
 from IPython.parallel import AsyncResult, AsyncHubResult
 from IPython.parallel import LoadBalancedView, DirectView
 
 from clienttest import ClusterTestCase, segfault, wait, add_engines
 
 def setup():
     add_engines(4, total=True)
 
 class TestClient(ClusterTestCase):
     
     def test_ids(self):
         n = len(self.client.ids)
         self.add_engines(2)
         self.assertEquals(len(self.client.ids), n+2)
     
     def test_view_indexing(self):
         """test index access for views"""
         self.minimum_engines(4)
         targets = self.client._build_targets('all')[-1]
         v = self.client[:]
         self.assertEquals(v.targets, targets)
         t = self.client.ids[2]
         v = self.client[t]
         self.assert_(isinstance(v, DirectView))
         self.assertEquals(v.targets, t)
         t = self.client.ids[2:4]
         v = self.client[t]
         self.assert_(isinstance(v, DirectView))
         self.assertEquals(v.targets, t)
         v = self.client[::2]
         self.assert_(isinstance(v, DirectView))
         self.assertEquals(v.targets, targets[::2])
         v = self.client[1::3]
         self.assert_(isinstance(v, DirectView))
         self.assertEquals(v.targets, targets[1::3])
         v = self.client[:-3]
         self.assert_(isinstance(v, DirectView))
         self.assertEquals(v.targets, targets[:-3])
         v = self.client[-1]
         self.assert_(isinstance(v, DirectView))
         self.assertEquals(v.targets, targets[-1])
         self.assertRaises(TypeError, lambda : self.client[None])
     
     def test_lbview_targets(self):
         """test load_balanced_view targets"""
         v = self.client.load_balanced_view()
         self.assertEquals(v.targets, None)
         v = self.client.load_balanced_view(-1)
         self.assertEquals(v.targets, [self.client.ids[-1]])
         v = self.client.load_balanced_view('all')
         self.assertEquals(v.targets, None)
     
     def test_dview_targets(self):
         """test direct_view targets"""
         v = self.client.direct_view()
         self.assertEquals(v.targets, 'all')
         v = self.client.direct_view('all')
         self.assertEquals(v.targets, 'all')
         v = self.client.direct_view(-1)
         self.assertEquals(v.targets, self.client.ids[-1])
     
     def test_lazy_all_targets(self):
         """test lazy evaluation of rc.direct_view('all')"""
         v = self.client.direct_view()
         self.assertEquals(v.targets, 'all')
         
         def double(x):
             return x*2
         seq = range(100)
         ref = [ double(x) for x in seq ]
         
         # add some engines, which should be used
         self.add_engines(1)
         n1 = len(self.client.ids)
         
         # simple apply
         r = v.apply_sync(lambda : 1)
         self.assertEquals(r, [1] * n1)
         
         # map goes through remotefunction
         r = v.map_sync(double, seq)
         self.assertEquals(r, ref)
 
         # add a couple more engines, and try again
         self.add_engines(2)
         n2 = len(self.client.ids)
         self.assertNotEquals(n2, n1)
         
         # apply
         r = v.apply_sync(lambda : 1)
         self.assertEquals(r, [1] * n2)
         
         # map
         r = v.map_sync(double, seq)
         self.assertEquals(r, ref)
     
     def test_targets(self):
         """test various valid targets arguments"""
         build = self.client._build_targets
         ids = self.client.ids
         idents,targets = build(None)
         self.assertEquals(ids, targets)
     
     def test_clear(self):
         """test clear behavior"""
         self.minimum_engines(2)
         v = self.client[:]
         v.block=True
         v.push(dict(a=5))
         v.pull('a')
         id0 = self.client.ids[-1]
         self.client.clear(targets=id0, block=True)
         a = self.client[:-1].get('a')
         self.assertRaisesRemote(NameError, self.client[id0].get, 'a')
         self.client.clear(block=True)
         for i in self.client.ids:
             self.assertRaisesRemote(NameError, self.client[i].get, 'a')
     
     def test_get_result(self):
         """test getting results from the Hub."""
         c = clientmod.Client(profile='iptest')
         t = c.ids[-1]
         ar = c[t].apply_async(wait, 1)
         # give the monitor time to notice the message
         time.sleep(.25)
         ahr = self.client.get_result(ar.msg_ids)
         self.assertTrue(isinstance(ahr, AsyncHubResult))
         self.assertEquals(ahr.get(), ar.get())
         ar2 = self.client.get_result(ar.msg_ids)
         self.assertFalse(isinstance(ar2, AsyncHubResult))
         c.close()
     
     def test_ids_list(self):
         """test client.ids"""
         ids = self.client.ids
         self.assertEquals(ids, self.client._ids)
         self.assertFalse(ids is self.client._ids)
         ids.remove(ids[-1])
         self.assertNotEquals(ids, self.client._ids)
     
     def test_queue_status(self):
         ids = self.client.ids
         id0 = ids[0]
         qs = self.client.queue_status(targets=id0)
         self.assertTrue(isinstance(qs, dict))
         self.assertEquals(sorted(qs.keys()), ['completed', 'queue', 'tasks'])
         allqs = self.client.queue_status()
         self.assertTrue(isinstance(allqs, dict))
         intkeys = list(allqs.keys())
         intkeys.remove('unassigned')
         self.assertEquals(sorted(intkeys), sorted(self.client.ids))
         unassigned = allqs.pop('unassigned')
         for eid,qs in allqs.items():
             self.assertTrue(isinstance(qs, dict))
             self.assertEquals(sorted(qs.keys()), ['completed', 'queue', 'tasks'])
 
     def test_shutdown(self):
         ids = self.client.ids
         id0 = ids[0]
         self.client.shutdown(id0, block=True)
         while id0 in self.client.ids:
             time.sleep(0.1)
             self.client.spin()
         
         self.assertRaises(IndexError, lambda : self.client[id0])
         
     def test_result_status(self):
         pass
         # to be written
     
     def test_db_query_dt(self):
         """test db query by date"""
         hist = self.client.hub_history()
         middle = self.client.db_query({'msg_id' : hist[len(hist)//2]})[0]
         tic = middle['submitted']
         before = self.client.db_query({'submitted' : {'$lt' : tic}})
         after = self.client.db_query({'submitted' : {'$gte' : tic}})
         self.assertEquals(len(before)+len(after),len(hist))
         for b in before:
             self.assertTrue(b['submitted'] < tic)
         for a in after:
             self.assertTrue(a['submitted'] >= tic)
         same = self.client.db_query({'submitted' : tic})
         for s in same:
             self.assertTrue(s['submitted'] == tic)
     
     def test_db_query_keys(self):
         """test extracting subset of record keys"""
         found = self.client.db_query({'msg_id': {'$ne' : ''}},keys=['submitted', 'completed'])
         for rec in found:
             self.assertEquals(set(rec.keys()), set(['msg_id', 'submitted', 'completed']))
     
     def test_db_query_default_keys(self):
         """default db_query excludes buffers"""
         found = self.client.db_query({'msg_id': {'$ne' : ''}})
         for rec in found:
             keys = set(rec.keys())
             self.assertFalse('buffers' in keys, "'buffers' should not be in: %s" % keys)
             self.assertFalse('result_buffers' in keys, "'result_buffers' should not be in: %s" % keys)
     
     def test_db_query_msg_id(self):
         """ensure msg_id is always in db queries"""
         found = self.client.db_query({'msg_id': {'$ne' : ''}},keys=['submitted', 'completed'])
         for rec in found:
             self.assertTrue('msg_id' in rec.keys())
         found = self.client.db_query({'msg_id': {'$ne' : ''}},keys=['submitted'])
         for rec in found:
             self.assertTrue('msg_id' in rec.keys())
         found = self.client.db_query({'msg_id': {'$ne' : ''}},keys=['msg_id'])
         for rec in found:
             self.assertTrue('msg_id' in rec.keys())
     
     def test_db_query_get_result(self):
         """pop in db_query shouldn't pop from result itself"""
         self.client[:].apply_sync(lambda : 1)
         found = self.client.db_query({'msg_id': {'$ne' : ''}})
         rc2 = clientmod.Client(profile='iptest')
         # If this bug is not fixed, this call will hang:
         ar = rc2.get_result(self.client.history[-1])
         ar.wait(2)
         self.assertTrue(ar.ready())
         ar.get()
         rc2.close()
     
     def test_db_query_in(self):
         """test db query with '$in','$nin' operators"""
         hist = self.client.hub_history()
         even = hist[::2]
         odd = hist[1::2]
         recs = self.client.db_query({ 'msg_id' : {'$in' : even}})
         found = [ r['msg_id'] for r in recs ]
         self.assertEquals(set(even), set(found))
         recs = self.client.db_query({ 'msg_id' : {'$nin' : even}})
         found = [ r['msg_id'] for r in recs ]
         self.assertEquals(set(odd), set(found))
     
     def test_hub_history(self):
         hist = self.client.hub_history()
         recs = self.client.db_query({ 'msg_id' : {"$ne":''}})
         recdict = {}
         for rec in recs:
             recdict[rec['msg_id']] = rec
         
         latest = datetime(1984,1,1)
         for msg_id in hist:
             rec = recdict[msg_id]
             newt = rec['submitted']
             self.assertTrue(newt >= latest)
             latest = newt
         ar = self.client[-1].apply_async(lambda : 1)
         ar.get()
         time.sleep(0.25)
         self.assertEquals(self.client.hub_history()[-1:],ar.msg_ids)
     
+    def _wait_for_idle(self):
+        """wait for an engine to become idle, according to the Hub"""
+        rc = self.client
+        
+        # timeout 2s, polling every 100ms
+        for i in range(20):
+            qs = rc.queue_status()
+            if qs['unassigned'] or any(qs[eid]['tasks'] for eid in rc.ids):
+                time.sleep(0.1)
+            else:
+                break
+        
+        # ensure Hub up to date:
+        qs = rc.queue_status()
+        self.assertEquals(qs['unassigned'], 0)
+        for eid in rc.ids:
+            self.assertEquals(qs[eid]['tasks'], 0)
+    
+    
     def test_resubmit(self):
         def f():
             import random
             return random.random()
         v = self.client.load_balanced_view()
         ar = v.apply_async(f)
         r1 = ar.get(1)
         # give the Hub a chance to notice:
-        time.sleep(0.5)
+        self._wait_for_idle()
         ahr = self.client.resubmit(ar.msg_ids)
         r2 = ahr.get(1)
         self.assertFalse(r1 == r2)
 
+    def test_resubmit_aborted(self):
+        def f():
+            import random
+            return random.random()
+        v = self.client.load_balanced_view()
+        # restrict to one engine, so we can put a sleep
+        # ahead of the task, so it will get aborted
+        eid = self.client.ids[-1]
+        v.targets = [eid]
+        sleep = v.apply_async(time.sleep, 0.5)
+        ar = v.apply_async(f)
+        ar.abort()
+        self.assertRaises(error.TaskAborted, ar.get)
+        # Give the Hub a chance to get up to date:
+        self._wait_for_idle()
+        ahr = self.client.resubmit(ar.msg_ids)
+        r2 = ahr.get(1)
+
     def test_resubmit_inflight(self):
-        """ensure ValueError on resubmit of inflight task"""
+        """resubmit of inflight task"""
         v = self.client.load_balanced_view()
         ar = v.apply_async(time.sleep,1)
         # give the message a chance to arrive
         time.sleep(0.2)
-        self.assertRaisesRemote(ValueError, self.client.resubmit, ar.msg_ids)
+        ahr = self.client.resubmit(ar.msg_ids)
         ar.get(2)
+        ahr.get(2)
 
     def test_resubmit_badkey(self):
         """ensure KeyError on resubmit of nonexistant task"""
         self.assertRaisesRemote(KeyError, self.client.resubmit, ['invalid'])
 
     def test_purge_results(self):
         # ensure there are some tasks
         for i in range(5):
             self.client[:].apply_sync(lambda : 1)
         # Wait for the Hub to realise the result is done:
         # This prevents a race condition, where we
         # might purge a result the Hub still thinks is pending.
         time.sleep(0.1)
         rc2 = clientmod.Client(profile='iptest')
         hist = self.client.hub_history()
         ahr = rc2.get_result([hist[-1]])
         ahr.wait(10)
         self.client.purge_results(hist[-1])
         newhist = self.client.hub_history()
         self.assertEquals(len(newhist)+1,len(hist))
         rc2.spin()
         rc2.close()
         
     def test_purge_all_results(self):
         self.client.purge_results('all')
         hist = self.client.hub_history()
         self.assertEquals(len(hist), 0)
     
     def test_spin_thread(self):
         self.client.spin_thread(0.01)
         ar = self.client[-1].apply_async(lambda : 1)
         time.sleep(0.1)
         self.assertTrue(ar.wall_time < 0.1,
             "spin should have kept wall_time < 0.1, but got %f" % ar.wall_time
         )
     
     def test_stop_spin_thread(self):
         self.client.spin_thread(0.01)
         self.client.stop_spin_thread()
         ar = self.client[-1].apply_async(lambda : 1)
         time.sleep(0.15)
         self.assertTrue(ar.wall_time > 0.1,
             "Shouldn't be spinning, but got wall_time=%f" % ar.wall_time
         )
     
 

IPython/parallel/tests/test_db.py

@@ -1,240 +1,243 b''
 """Tests for db backends
 
 Authors:
 
 * Min RK
 """
 
 #-------------------------------------------------------------------------------
 #  Copyright (C) 2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-------------------------------------------------------------------------------
 
 #-------------------------------------------------------------------------------
 # Imports
 #-------------------------------------------------------------------------------
 
 from __future__ import division
 
+import logging
 import os
 import tempfile
 import time
 
 from datetime import datetime, timedelta
 from unittest import TestCase
 
 from IPython.parallel import error
 from IPython.parallel.controller.dictdb import DictDB
 from IPython.parallel.controller.sqlitedb import SQLiteDB
 from IPython.parallel.controller.hub import init_record, empty_record
 
 from IPython.testing import decorators as dec
 from IPython.zmq.session import Session
 
 
 #-------------------------------------------------------------------------------
 # TestCases
 #-------------------------------------------------------------------------------
 
 
 def setup():
     global temp_db
     temp_db = tempfile.NamedTemporaryFile(suffix='.db').name
 
 
 class TestDictBackend(TestCase):
     def setUp(self):
         self.session = Session()
         self.db = self.create_db()
         self.load_records(16)
     
     def create_db(self):
         return DictDB()
     
     def load_records(self, n=1):
         """load n records for testing"""
         #sleep 1/10 s, to ensure timestamp is different to previous calls
         time.sleep(0.1)
         msg_ids = []
         for i in range(n):
             msg = self.session.msg('apply_request', content=dict(a=5))
             msg['buffers'] = []
             rec = init_record(msg)
             msg_id = msg['header']['msg_id']
             msg_ids.append(msg_id)
             self.db.add_record(msg_id, rec)
         return msg_ids
     
     def test_add_record(self):
         before = self.db.get_history()
         self.load_records(5)
         after = self.db.get_history()
         self.assertEquals(len(after), len(before)+5)
         self.assertEquals(after[:-5],before)
         
     def test_drop_record(self):
         msg_id = self.load_records()[-1]
         rec = self.db.get_record(msg_id)
         self.db.drop_record(msg_id)
         self.assertRaises(KeyError,self.db.get_record, msg_id)
     
     def _round_to_millisecond(self, dt):
         """necessary because mongodb rounds microseconds"""
         micro = dt.microsecond
         extra = int(str(micro)[-3:])
         return dt - timedelta(microseconds=extra)
     
     def test_update_record(self):
         now = self._round_to_millisecond(datetime.now())
         # 
         msg_id = self.db.get_history()[-1]
         rec1 = self.db.get_record(msg_id)
         data = {'stdout': 'hello there', 'completed' : now}
         self.db.update_record(msg_id, data)
         rec2 = self.db.get_record(msg_id)
         self.assertEquals(rec2['stdout'], 'hello there')
         self.assertEquals(rec2['completed'], now)
         rec1.update(data)
         self.assertEquals(rec1, rec2)
     
     # def test_update_record_bad(self):
     #     """test updating nonexistant records"""
     #     msg_id = str(uuid.uuid4())
     #     data = {'stdout': 'hello there'}
     #     self.assertRaises(KeyError, self.db.update_record, msg_id, data)
 
     def test_find_records_dt(self):
         """test finding records by date"""
         hist = self.db.get_history()
         middle = self.db.get_record(hist[len(hist)//2])
         tic = middle['submitted']
         before = self.db.find_records({'submitted' : {'$lt' : tic}})
         after = self.db.find_records({'submitted' : {'$gte' : tic}})
         self.assertEquals(len(before)+len(after),len(hist))
         for b in before:
             self.assertTrue(b['submitted'] < tic)
         for a in after:
             self.assertTrue(a['submitted'] >= tic)
         same = self.db.find_records({'submitted' : tic})
         for s in same:
             self.assertTrue(s['submitted'] == tic)
     
     def test_find_records_keys(self):
         """test extracting subset of record keys"""
         found = self.db.find_records({'msg_id': {'$ne' : ''}},keys=['submitted', 'completed'])
         for rec in found:
             self.assertEquals(set(rec.keys()), set(['msg_id', 'submitted', 'completed']))
     
     def test_find_records_msg_id(self):
         """ensure msg_id is always in found records"""
         found = self.db.find_records({'msg_id': {'$ne' : ''}},keys=['submitted', 'completed'])
         for rec in found:
             self.assertTrue('msg_id' in rec.keys())
         found = self.db.find_records({'msg_id': {'$ne' : ''}},keys=['submitted'])
         for rec in found:
             self.assertTrue('msg_id' in rec.keys())
         found = self.db.find_records({'msg_id': {'$ne' : ''}},keys=['msg_id'])
         for rec in found:
             self.assertTrue('msg_id' in rec.keys())
     
     def test_find_records_in(self):
         """test finding records with '$in','$nin' operators"""
         hist = self.db.get_history()
         even = hist[::2]
         odd = hist[1::2]
         recs = self.db.find_records({ 'msg_id' : {'$in' : even}})
         found = [ r['msg_id'] for r in recs ]
         self.assertEquals(set(even), set(found))
         recs = self.db.find_records({ 'msg_id' : {'$nin' : even}})
         found = [ r['msg_id'] for r in recs ]
         self.assertEquals(set(odd), set(found))
     
     def test_get_history(self):
         msg_ids = self.db.get_history()
         latest = datetime(1984,1,1)
         for msg_id in msg_ids:
             rec = self.db.get_record(msg_id)
             newt = rec['submitted']
             self.assertTrue(newt >= latest)
             latest = newt
         msg_id = self.load_records(1)[-1]
         self.assertEquals(self.db.get_history()[-1],msg_id)
     
     def test_datetime(self):
         """get/set timestamps with datetime objects"""
         msg_id = self.db.get_history()[-1]
         rec = self.db.get_record(msg_id)
         self.assertTrue(isinstance(rec['submitted'], datetime))
         self.db.update_record(msg_id, dict(completed=datetime.now()))
         rec = self.db.get_record(msg_id)
         self.assertTrue(isinstance(rec['completed'], datetime))
 
     def test_drop_matching(self):
         msg_ids = self.load_records(10)
         query = {'msg_id' : {'$in':msg_ids}}
         self.db.drop_matching_records(query)
         recs = self.db.find_records(query)
         self.assertEquals(len(recs), 0)
     
     def test_null(self):
         """test None comparison queries"""
         msg_ids = self.load_records(10)
 
         query = {'msg_id' : None}
         recs = self.db.find_records(query)
         self.assertEquals(len(recs), 0)
 
         query = {'msg_id' : {'$ne' : None}}
         recs = self.db.find_records(query)
         self.assertTrue(len(recs) >= 10)
     
     def test_pop_safe_get(self):
         """editing query results shouldn't affect record [get]"""
         msg_id = self.db.get_history()[-1]
         rec = self.db.get_record(msg_id)
         rec.pop('buffers')
         rec['garbage'] = 'hello'
         rec2 = self.db.get_record(msg_id)
         self.assertTrue('buffers' in rec2)
         self.assertFalse('garbage' in rec2)
     
     def test_pop_safe_find(self):
         """editing query results shouldn't affect record [find]"""
         msg_id = self.db.get_history()[-1]
         rec = self.db.find_records({'msg_id' : msg_id})[0]
         rec.pop('buffers')
         rec['garbage'] = 'hello'
         rec2 = self.db.find_records({'msg_id' : msg_id})[0]
         self.assertTrue('buffers' in rec2)
         self.assertFalse('garbage' in rec2)
 
     def test_pop_safe_find_keys(self):
         """editing query results shouldn't affect record [find+keys]"""
         msg_id = self.db.get_history()[-1]
         rec = self.db.find_records({'msg_id' : msg_id}, keys=['buffers'])[0]
         rec.pop('buffers')
         rec['garbage'] = 'hello'
         rec2 = self.db.find_records({'msg_id' : msg_id})[0]
         self.assertTrue('buffers' in rec2)
         self.assertFalse('garbage' in rec2)
 
 
 class TestSQLiteBackend(TestDictBackend):
 
     @dec.skip_without('sqlite3')
     def create_db(self):
         location, fname = os.path.split(temp_db)
-        return SQLiteDB(location=location, fname=fname)
+        log = logging.getLogger('test')
+        log.setLevel(logging.CRITICAL)
+        return SQLiteDB(location=location, fname=fname, log=log)
     
     def tearDown(self):
         self.db._db.close()
 
 
 def teardown():
     """cleanup task db file after all tests have run"""
     try:
         os.remove(temp_db)
     except:
         pass

IPython/parallel/tests/test_view.py

@@ -1,553 +1,694 b''
 # -*- coding: utf-8 -*-
 """test View objects
 
 Authors:
 
 * Min RK
 """
 #-------------------------------------------------------------------------------
 #  Copyright (C) 2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-------------------------------------------------------------------------------
 
 #-------------------------------------------------------------------------------
 # Imports
 #-------------------------------------------------------------------------------
 
 import sys
 import time
 from tempfile import mktemp
 from StringIO import StringIO
 
 import zmq
 from nose import SkipTest
 
 from IPython.testing import decorators as dec
+from IPython.testing.ipunittest import ParametricTestCase
 
 from IPython import parallel  as pmod
 from IPython.parallel import error
 from IPython.parallel import AsyncResult, AsyncHubResult, AsyncMapResult
 from IPython.parallel import DirectView
 from IPython.parallel.util import interactive
 
 from IPython.parallel.tests import add_engines
 
 from .clienttest import ClusterTestCase, crash, wait, skip_without
 
 def setup():
     add_engines(3, total=True)
 
-class TestView(ClusterTestCase):
+class TestView(ClusterTestCase, ParametricTestCase):
     
     def test_z_crash_mux(self):
         """test graceful handling of engine death (direct)"""
         raise SkipTest("crash tests disabled, due to undesirable crash reports")
         # self.add_engines(1)
         eid = self.client.ids[-1]
         ar = self.client[eid].apply_async(crash)
         self.assertRaisesRemote(error.EngineError, ar.get, 10)
         eid = ar.engine_id
         tic = time.time()
         while eid in self.client.ids and time.time()-tic < 5:
             time.sleep(.01)
             self.client.spin()
         self.assertFalse(eid in self.client.ids, "Engine should have died")
     
     def test_push_pull(self):
         """test pushing and pulling"""
         data = dict(a=10, b=1.05, c=range(10), d={'e':(1,2),'f':'hi'})
         t = self.client.ids[-1]
         v = self.client[t]
         push = v.push
         pull = v.pull
         v.block=True
         nengines = len(self.client)
         push({'data':data})
         d = pull('data')
         self.assertEquals(d, data)
         self.client[:].push({'data':data})
         d = self.client[:].pull('data', block=True)
         self.assertEquals(d, nengines*[data])
         ar = push({'data':data}, block=False)
         self.assertTrue(isinstance(ar, AsyncResult))
         r = ar.get()
         ar = self.client[:].pull('data', block=False)
         self.assertTrue(isinstance(ar, AsyncResult))
         r = ar.get()
         self.assertEquals(r, nengines*[data])
         self.client[:].push(dict(a=10,b=20))
         r = self.client[:].pull(('a','b'), block=True)
         self.assertEquals(r, nengines*[[10,20]])
     
     def test_push_pull_function(self):
         "test pushing and pulling functions"
         def testf(x):
             return 2.0*x
         
         t = self.client.ids[-1]
         v = self.client[t]
         v.block=True
         push = v.push
         pull = v.pull
         execute = v.execute
         push({'testf':testf})
         r = pull('testf')
         self.assertEqual(r(1.0), testf(1.0))
         execute('r = testf(10)')
         r = pull('r')
         self.assertEquals(r, testf(10))
         ar = self.client[:].push({'testf':testf}, block=False)
         ar.get()
         ar = self.client[:].pull('testf', block=False)
         rlist = ar.get()
         for r in rlist:
             self.assertEqual(r(1.0), testf(1.0))
         execute("def g(x): return x*x")
         r = pull(('testf','g'))
         self.assertEquals((r[0](10),r[1](10)), (testf(10), 100))
     
     def test_push_function_globals(self):
         """test that pushed functions have access to globals"""
         @interactive
         def geta():
             return a
         # self.add_engines(1)
         v = self.client[-1]
         v.block=True
         v['f'] = geta
         self.assertRaisesRemote(NameError, v.execute, 'b=f()')
         v.execute('a=5')
         v.execute('b=f()')
         self.assertEquals(v['b'], 5)
     
     def test_push_function_defaults(self):
         """test that pushed functions preserve default args"""
         def echo(a=10):
             return a
         v = self.client[-1]
         v.block=True
         v['f'] = echo
         v.execute('b=f()')
         self.assertEquals(v['b'], 10)
 
     def test_get_result(self):
         """test getting results from the Hub."""
         c = pmod.Client(profile='iptest')
         # self.add_engines(1)
         t = c.ids[-1]
         v = c[t]
         v2 = self.client[t]
         ar = v.apply_async(wait, 1)
         # give the monitor time to notice the message
         time.sleep(.25)
         ahr = v2.get_result(ar.msg_ids)
         self.assertTrue(isinstance(ahr, AsyncHubResult))
         self.assertEquals(ahr.get(), ar.get())
         ar2 = v2.get_result(ar.msg_ids)
         self.assertFalse(isinstance(ar2, AsyncHubResult))
         c.spin()
         c.close()
     
     def test_run_newline(self):
         """test that run appends newline to files"""
         tmpfile = mktemp()
         with open(tmpfile, 'w') as f:
             f.write("""def g():
                 return 5
                 """)
         v = self.client[-1]
         v.run(tmpfile, block=True)
         self.assertEquals(v.apply_sync(lambda f: f(), pmod.Reference('g')), 5)
 
     def test_apply_tracked(self):
         """test tracking for apply"""
         # self.add_engines(1)
         t = self.client.ids[-1]
         v = self.client[t]
         v.block=False
         def echo(n=1024*1024, **kwargs):
             with v.temp_flags(**kwargs):
                 return v.apply(lambda x: x, 'x'*n)
         ar = echo(1, track=False)
         self.assertTrue(isinstance(ar._tracker, zmq.MessageTracker))
         self.assertTrue(ar.sent)
         ar = echo(track=True)
         self.assertTrue(isinstance(ar._tracker, zmq.MessageTracker))
         self.assertEquals(ar.sent, ar._tracker.done)
         ar._tracker.wait()
         self.assertTrue(ar.sent)
         
     def test_push_tracked(self):
         t = self.client.ids[-1]
         ns = dict(x='x'*1024*1024)
         v = self.client[t]
         ar = v.push(ns, block=False, track=False)
         self.assertTrue(isinstance(ar._tracker, zmq.MessageTracker))
         self.assertTrue(ar.sent)
         
         ar = v.push(ns, block=False, track=True)
         self.assertTrue(isinstance(ar._tracker, zmq.MessageTracker))
         ar._tracker.wait()
         self.assertEquals(ar.sent, ar._tracker.done)
         self.assertTrue(ar.sent)
         ar.get()
         
     def test_scatter_tracked(self):
         t = self.client.ids
         x='x'*1024*1024
         ar = self.client[t].scatter('x', x, block=False, track=False)
         self.assertTrue(isinstance(ar._tracker, zmq.MessageTracker))
         self.assertTrue(ar.sent)
         
         ar = self.client[t].scatter('x', x, block=False, track=True)
         self.assertTrue(isinstance(ar._tracker, zmq.MessageTracker))
         self.assertEquals(ar.sent, ar._tracker.done)
         ar._tracker.wait()
         self.assertTrue(ar.sent)
         ar.get()
     
     def test_remote_reference(self):
         v = self.client[-1]
         v['a'] = 123
         ra = pmod.Reference('a')
         b = v.apply_sync(lambda x: x, ra)
         self.assertEquals(b, 123)
 
 
     def test_scatter_gather(self):
         view = self.client[:]
         seq1 = range(16)
         view.scatter('a', seq1)
         seq2 = view.gather('a', block=True)
         self.assertEquals(seq2, seq1)
         self.assertRaisesRemote(NameError, view.gather, 'asdf', block=True)
     
     @skip_without('numpy')
     def test_scatter_gather_numpy(self):
         import numpy
         from numpy.testing.utils import assert_array_equal, assert_array_almost_equal
         view = self.client[:]
         a = numpy.arange(64)
         view.scatter('a', a)
         b = view.gather('a', block=True)
         assert_array_equal(b, a)
     
     def test_scatter_gather_lazy(self):
         """scatter/gather with targets='all'"""
         view = self.client.direct_view(targets='all')
         x = range(64)
         view.scatter('x', x)
         gathered = view.gather('x', block=True)
         self.assertEquals(gathered, x)
         
 
     @dec.known_failure_py3
     @skip_without('numpy')
     def test_push_numpy_nocopy(self):
         import numpy
         view = self.client[:]
         a = numpy.arange(64)
         view['A'] = a
         @interactive
         def check_writeable(x):
             return x.flags.writeable
         
         for flag in view.apply_sync(check_writeable, pmod.Reference('A')):
             self.assertFalse(flag, "array is writeable, push shouldn't have pickled it")
         
         view.push(dict(B=a))
         for flag in view.apply_sync(check_writeable, pmod.Reference('B')):
             self.assertFalse(flag, "array is writeable, push shouldn't have pickled it")
     
     @skip_without('numpy')
     def test_apply_numpy(self):
         """view.apply(f, ndarray)"""
         import numpy
         from numpy.testing.utils import assert_array_equal, assert_array_almost_equal
         
         A = numpy.random.random((100,100))
         view = self.client[-1]
         for dt in [ 'int32', 'uint8', 'float32', 'float64' ]:
             B = A.astype(dt)
             C = view.apply_sync(lambda x:x, B)
             assert_array_equal(B,C)
     
     def test_map(self):
         view = self.client[:]
         def f(x):
             return x**2
         data = range(16)
         r = view.map_sync(f, data)
         self.assertEquals(r, map(f, data))
     
     def test_map_iterable(self):
         """test map on iterables (direct)"""
         view = self.client[:]
         # 101 is prime, so it won't be evenly distributed
         arr = range(101)
         # ensure it will be an iterator, even in Python 3
         it = iter(arr)
         r = view.map_sync(lambda x:x, arr)
         self.assertEquals(r, list(arr))
     
     def test_scatterGatherNonblocking(self):
         data = range(16)
         view = self.client[:]
         view.scatter('a', data, block=False)
         ar = view.gather('a', block=False)
         self.assertEquals(ar.get(), data)
     
     @skip_without('numpy')
     def test_scatter_gather_numpy_nonblocking(self):
         import numpy
         from numpy.testing.utils import assert_array_equal, assert_array_almost_equal
         a = numpy.arange(64)
         view = self.client[:]
         ar = view.scatter('a', a, block=False)
         self.assertTrue(isinstance(ar, AsyncResult))
         amr = view.gather('a', block=False)
         self.assertTrue(isinstance(amr, AsyncMapResult))
         assert_array_equal(amr.get(), a)
 
     def test_execute(self):
         view = self.client[:]
         # self.client.debug=True
         execute = view.execute
         ar = execute('c=30', block=False)
         self.assertTrue(isinstance(ar, AsyncResult))
         ar = execute('d=[0,1,2]', block=False)
         self.client.wait(ar, 1)
         self.assertEquals(len(ar.get()), len(self.client))
         for c in view['c']:
             self.assertEquals(c, 30)
     
     def test_abort(self):
         view = self.client[-1]
         ar = view.execute('import time; time.sleep(1)', block=False)
         ar2 = view.apply_async(lambda : 2)
         ar3 = view.apply_async(lambda : 3)
         view.abort(ar2)
         view.abort(ar3.msg_ids)
         self.assertRaises(error.TaskAborted, ar2.get)
         self.assertRaises(error.TaskAborted, ar3.get)
     
     def test_abort_all(self):
         """view.abort() aborts all outstanding tasks"""
         view = self.client[-1]
         ars = [ view.apply_async(time.sleep, 0.25) for i in range(10) ]
         view.abort()
         view.wait(timeout=5)
         for ar in ars[5:]:
             self.assertRaises(error.TaskAborted, ar.get)
     
     def test_temp_flags(self):
         view = self.client[-1]
         view.block=True
         with view.temp_flags(block=False):
             self.assertFalse(view.block)
         self.assertTrue(view.block)
     
     @dec.known_failure_py3
     def test_importer(self):
         view = self.client[-1]
         view.clear(block=True)
         with view.importer:
             import re
         
         @interactive
         def findall(pat, s):
             # this globals() step isn't necessary in real code
             # only to prevent a closure in the test
             re = globals()['re']
             return re.findall(pat, s)
         
         self.assertEquals(view.apply_sync(findall, '\w+', 'hello world'), 'hello world'.split())
     
     # parallel magic tests
     
     def test_magic_px_blocking(self):
         ip = get_ipython()
         v = self.client[-1]
         v.activate()
         v.block=True
 
         ip.magic_px('a=5')
         self.assertEquals(v['a'], 5)
         ip.magic_px('a=10')
         self.assertEquals(v['a'], 10)
         sio = StringIO()
         savestdout = sys.stdout
         sys.stdout = sio
         # just 'print a' worst ~99% of the time, but this ensures that
         # the stdout message has arrived when the result is finished:
         ip.magic_px('import sys,time;print (a); sys.stdout.flush();time.sleep(0.2)')
         sys.stdout = savestdout
         buf = sio.getvalue()
         self.assertTrue('[stdout:' in buf, buf)
         self.assertTrue(buf.rstrip().endswith('10'))
         self.assertRaisesRemote(ZeroDivisionError, ip.magic_px, '1/0')
 
     def test_magic_px_nonblocking(self):
         ip = get_ipython()
         v = self.client[-1]
         v.activate()
         v.block=False
 
         ip.magic_px('a=5')
         self.assertEquals(v['a'], 5)
         ip.magic_px('a=10')
         self.assertEquals(v['a'], 10)
         sio = StringIO()
         savestdout = sys.stdout
         sys.stdout = sio
         ip.magic_px('print a')
         sys.stdout = savestdout
         buf = sio.getvalue()
         self.assertFalse('[stdout:%i]'%v.targets in buf)
         ip.magic_px('1/0')
         ar = v.get_result(-1)
         self.assertRaisesRemote(ZeroDivisionError, ar.get)
     
     def test_magic_autopx_blocking(self):
         ip = get_ipython()
         v = self.client[-1]
         v.activate()
         v.block=True
 
         sio = StringIO()
         savestdout = sys.stdout
         sys.stdout = sio
         ip.magic_autopx()
         ip.run_cell('\n'.join(('a=5','b=10','c=0')))
-        ip.run_cell('print b')
+        ip.run_cell('b*=2')
+        ip.run_cell('print (b)')
         ip.run_cell("b/c")
-        ip.run_code(compile('b*=2', '', 'single'))
         ip.magic_autopx()
         sys.stdout = savestdout
         output = sio.getvalue().strip()
         self.assertTrue(output.startswith('%autopx enabled'))
         self.assertTrue(output.endswith('%autopx disabled'))
         self.assertTrue('RemoteError: ZeroDivisionError' in output)
-        ar = v.get_result(-2)
+        ar = v.get_result(-1)
         self.assertEquals(v['a'], 5)
         self.assertEquals(v['b'], 20)
         self.assertRaisesRemote(ZeroDivisionError, ar.get)
 
     def test_magic_autopx_nonblocking(self):
         ip = get_ipython()
         v = self.client[-1]
         v.activate()
         v.block=False
 
         sio = StringIO()
         savestdout = sys.stdout
         sys.stdout = sio
         ip.magic_autopx()
         ip.run_cell('\n'.join(('a=5','b=10','c=0')))
-        ip.run_cell('print b')
+        ip.run_cell('print (b)')
+        ip.run_cell('import time; time.sleep(0.1)')
         ip.run_cell("b/c")
-        ip.run_code(compile('b*=2', '', 'single'))
+        ip.run_cell('b*=2')
         ip.magic_autopx()
         sys.stdout = savestdout
         output = sio.getvalue().strip()
         self.assertTrue(output.startswith('%autopx enabled'))
         self.assertTrue(output.endswith('%autopx disabled'))
         self.assertFalse('ZeroDivisionError' in output)
         ar = v.get_result(-2)
-        self.assertEquals(v['a'], 5)
-        self.assertEquals(v['b'], 20)
         self.assertRaisesRemote(ZeroDivisionError, ar.get)
+        # prevent TaskAborted on pulls, due to ZeroDivisionError
+        time.sleep(0.5)
+        self.assertEquals(v['a'], 5)
+        # b*=2 will not fire, due to abort
+        self.assertEquals(v['b'], 10)
     
     def test_magic_result(self):
         ip = get_ipython()
         v = self.client[-1]
         v.activate()
         v['a'] = 111
         ra = v['a']
         
         ar = ip.magic_result()
         self.assertEquals(ar.msg_ids, [v.history[-1]])
         self.assertEquals(ar.get(), 111)
         ar = ip.magic_result('-2')
         self.assertEquals(ar.msg_ids, [v.history[-2]])
     
     def test_unicode_execute(self):
         """test executing unicode strings"""
         v = self.client[-1]
         v.block=True
         if sys.version_info[0] >= 3:
             code="a='é'"
         else:
             code=u"a=u'é'"
         v.execute(code)
         self.assertEquals(v['a'], u'é')
         
     def test_unicode_apply_result(self):
         """test unicode apply results"""
         v = self.client[-1]
         r = v.apply_sync(lambda : u'é')
         self.assertEquals(r, u'é')
     
     def test_unicode_apply_arg(self):
         """test passing unicode arguments to apply"""
         v = self.client[-1]
         
         @interactive
         def check_unicode(a, check):
             assert isinstance(a, unicode), "%r is not unicode"%a
             assert isinstance(check, bytes), "%r is not bytes"%check
             assert a.encode('utf8') == check, "%s != %s"%(a,check)
         
         for s in [ u'é', u'ßø®∫',u'asdf' ]:
             try:
                 v.apply_sync(check_unicode, s, s.encode('utf8'))
             except error.RemoteError as e:
                 if e.ename == 'AssertionError':
                     self.fail(e.evalue)
                 else:
                     raise e
     
     def test_map_reference(self):
         """view.map(<Reference>, *seqs) should work"""
         v = self.client[:]
         v.scatter('n', self.client.ids, flatten=True)
         v.execute("f = lambda x,y: x*y")
         rf = pmod.Reference('f')
         nlist = list(range(10))
         mlist = nlist[::-1]
         expected = [ m*n for m,n in zip(mlist, nlist) ]
         result = v.map_sync(rf, mlist, nlist)
         self.assertEquals(result, expected)
 
     def test_apply_reference(self):
         """view.apply(<Reference>, *args) should work"""
         v = self.client[:]
         v.scatter('n', self.client.ids, flatten=True)
         v.execute("f = lambda x: n*x")
         rf = pmod.Reference('f')
         result = v.apply_sync(rf, 5)
         expected = [ 5*id for id in self.client.ids ]
         self.assertEquals(result, expected)
     
     def test_eval_reference(self):
         v = self.client[self.client.ids[0]]
         v['g'] = range(5)
         rg = pmod.Reference('g[0]')
         echo = lambda x:x
         self.assertEquals(v.apply_sync(echo, rg), 0)
     
     def test_reference_nameerror(self):
         v = self.client[self.client.ids[0]]
         r = pmod.Reference('elvis_has_left')
         echo = lambda x:x
         self.assertRaisesRemote(NameError, v.apply_sync, echo, r)
 
     def test_single_engine_map(self):
         e0 = self.client[self.client.ids[0]]
         r = range(5)
         check = [ -1*i for i in r ]
         result = e0.map_sync(lambda x: -1*x, r)
         self.assertEquals(result, check)
+    
+    def test_len(self):
+        """len(view) makes sense"""
+        e0 = self.client[self.client.ids[0]]
+        yield self.assertEquals(len(e0), 1)
+        v = self.client[:]
+        yield self.assertEquals(len(v), len(self.client.ids))
+        v = self.client.direct_view('all')
+        yield self.assertEquals(len(v), len(self.client.ids))
+        v = self.client[:2]
+        yield self.assertEquals(len(v), 2)
+        v = self.client[:1]
+        yield self.assertEquals(len(v), 1)
+        v = self.client.load_balanced_view()
+        yield self.assertEquals(len(v), len(self.client.ids))
+        # parametric tests seem to require manual closing?
+        self.client.close()
+
+    
+    # begin execute tests
+    def _wait_for(self, f, timeout=10):
+        tic = time.time()
+        while time.time() <= tic + timeout:
+            if f():
+                return
+            time.sleep(0.1)
+            self.client.spin()
+        if not f():
+            print "Warning: Awaited condition never arrived"
+            
+    
+    def test_execute_reply(self):
+        e0 = self.client[self.client.ids[0]]
+        e0.block = True
+        ar = e0.execute("5", silent=False)
+        er = ar.get()
+        self._wait_for(lambda : bool(er.pyout))
+        self.assertEquals(str(er), "<ExecuteReply[%i]: 5>" % er.execution_count)
+        self.assertEquals(er.pyout['data']['text/plain'], '5')
+
+    def test_execute_reply_stdout(self):
+        e0 = self.client[self.client.ids[0]]
+        e0.block = True
+        ar = e0.execute("print (5)", silent=False)
+        er = ar.get()
+        self._wait_for(lambda : bool(er.stdout))
+        self.assertEquals(er.stdout.strip(), '5')
+        
+    def test_execute_pyout(self):
+        """execute triggers pyout with silent=False"""
+        view = self.client[:]
+        ar = view.execute("5", silent=False, block=True)
+        self._wait_for(lambda : all(ar.pyout))
+        
+        expected = [{'text/plain' : '5'}] * len(view)
+        mimes = [ out['data'] for out in ar.pyout ]
+        self.assertEquals(mimes, expected)
+    
+    def test_execute_silent(self):
+        """execute does not trigger pyout with silent=True"""
+        view = self.client[:]
+        ar = view.execute("5", block=True)
+        expected = [None] * len(view)
+        self.assertEquals(ar.pyout, expected)
+    
+    def test_execute_magic(self):
+        """execute accepts IPython commands"""
+        view = self.client[:]
+        view.execute("a = 5")
+        ar = view.execute("%whos", block=True)
+        # this will raise, if that failed
+        ar.get(5)
+        self._wait_for(lambda : all(ar.stdout))
+        for stdout in ar.stdout:
+            lines = stdout.splitlines()
+            self.assertEquals(lines[0].split(), ['Variable', 'Type', 'Data/Info'])
+            found = False
+            for line in lines[2:]:
+                split = line.split()
+                if split == ['a', 'int', '5']:
+                    found = True
+                    break
+            self.assertTrue(found, "whos output wrong: %s" % stdout)
+    
+    def test_execute_displaypub(self):
+        """execute tracks display_pub output"""
+        view = self.client[:]
+        view.execute("from IPython.core.display import *")
+        ar = view.execute("[ display(i) for i in range(5) ]", block=True)
+        
+        self._wait_for(lambda : all(len(er.outputs) >= 5 for er in ar))
+        expected = [ {u'text/plain' : unicode(j)} for j in range(5) ]
+        for outputs in ar.outputs:
+            mimes = [ out['data'] for out in outputs ]
+            self.assertEquals(mimes, expected)
+    
+    def test_apply_displaypub(self):
+        """apply tracks display_pub output"""
+        view = self.client[:]
+        view.execute("from IPython.core.display import *")
+        
+        @interactive
+        def publish():
+            [ display(i) for i in range(5) ]
+        
+        ar = view.apply_async(publish)
+        ar.get(5)
+        self._wait_for(lambda : all(len(out) >= 5 for out in ar.outputs))
+        expected = [ {u'text/plain' : unicode(j)} for j in range(5) ]
+        for outputs in ar.outputs:
+            mimes = [ out['data'] for out in outputs ]
+            self.assertEquals(mimes, expected)
+    
+    def test_execute_raises(self):
+        """exceptions in execute requests raise appropriately"""
+        view = self.client[-1]
+        ar = view.execute("1/0")
+        self.assertRaisesRemote(ZeroDivisionError, ar.get, 2)
+    
+    @dec.skipif_not_matplotlib
+    def test_magic_pylab(self):
+        """%pylab works on engines"""
+        view = self.client[-1]
+        ar = view.execute("%pylab inline")
+        # at least check if this raised:
+        reply = ar.get(5)
+        # include imports, in case user config
+        ar = view.execute("plot(rand(100))", silent=False)
+        reply = ar.get(5)
+        self._wait_for(lambda : all(ar.outputs))
+        self.assertEquals(len(reply.outputs), 1)
+        output = reply.outputs[0]
+        self.assertTrue("data" in output)
+        data = output['data']
+        self.assertTrue("image/png" in data)
+        
 

IPython/parallel/util.py

@@ -1,495 +1,358 b''
 """some generic utilities for dealing with classes, urls, and serialization
 
 Authors:
 
 * Min RK
 """
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2010-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 # Standard library imports.
 import logging
 import os
 import re
 import stat
 import socket
 import sys
 from signal import signal, SIGINT, SIGABRT, SIGTERM
 try:
     from signal import SIGKILL
 except ImportError:
     SIGKILL=None
 
 try:
     import cPickle
     pickle = cPickle
 except:
     cPickle = None
     import pickle
 
 # System library imports
 import zmq
 from zmq.log import handlers
 
 from IPython.external.decorator import decorator
 
 # IPython imports
 from IPython.config.application import Application
 from IPython.utils import py3compat
 from IPython.utils.pickleutil import can, uncan, canSequence, uncanSequence
 from IPython.utils.newserialized import serialize, unserialize
 from IPython.zmq.log import EnginePUBHandler
+from IPython.zmq.serialize import (
+    unserialize_object, serialize_object, pack_apply_message, unpack_apply_message
+)
 
 if py3compat.PY3:
     buffer = memoryview
 
 #-----------------------------------------------------------------------------
 # Classes
 #-----------------------------------------------------------------------------
 
 class Namespace(dict):
     """Subclass of dict for attribute access to keys."""
     
     def __getattr__(self, key):
         """getattr aliased to getitem"""
         if key in self.iterkeys():
             return self[key]
         else:
             raise NameError(key)
 
     def __setattr__(self, key, value):
         """setattr aliased to setitem, with strict"""
         if hasattr(dict, key):
             raise KeyError("Cannot override dict keys %r"%key)
         self[key] = value
     
 
 class ReverseDict(dict):
     """simple double-keyed subset of dict methods."""
     
     def __init__(self, *args, **kwargs):
         dict.__init__(self, *args, **kwargs)
         self._reverse = dict()
         for key, value in self.iteritems():
             self._reverse[value] = key
     
     def __getitem__(self, key):
         try:
             return dict.__getitem__(self, key)
         except KeyError:
             return self._reverse[key]
     
     def __setitem__(self, key, value):
         if key in self._reverse:
             raise KeyError("Can't have key %r on both sides!"%key)
         dict.__setitem__(self, key, value)
         self._reverse[value] = key
     
     def pop(self, key):
         value = dict.pop(self, key)
         self._reverse.pop(value)
         return value
     
     def get(self, key, default=None):
         try:
             return self[key]
         except KeyError:
             return default
 
 #-----------------------------------------------------------------------------
 # Functions
 #-----------------------------------------------------------------------------
 
 @decorator
 def log_errors(f, self, *args, **kwargs):
     """decorator to log unhandled exceptions raised in a method.
     
     For use wrapping on_recv callbacks, so that exceptions
     do not cause the stream to be closed.
     """
     try:
         return f(self, *args, **kwargs)
     except Exception:
         self.log.error("Uncaught exception in %r" % f, exc_info=True)
     
 
-def asbytes(s):
-    """ensure that an object is ascii bytes"""
-    if isinstance(s, unicode):
-        s = s.encode('ascii')
-    return s
-
 def is_url(url):
     """boolean check for whether a string is a zmq url"""
     if '://' not in url:
         return False
     proto, addr = url.split('://', 1)
     if proto.lower() not in ['tcp','pgm','epgm','ipc','inproc']:
         return False
     return True
 
 def validate_url(url):
     """validate a url for zeromq"""
     if not isinstance(url, basestring):
         raise TypeError("url must be a string, not %r"%type(url))
     url = url.lower()
     
     proto_addr = url.split('://')
     assert len(proto_addr) == 2, 'Invalid url: %r'%url
     proto, addr = proto_addr
     assert proto in ['tcp','pgm','epgm','ipc','inproc'], "Invalid protocol: %r"%proto
     
     # domain pattern adapted from http://www.regexlib.com/REDetails.aspx?regexp_id=391
     # author: Remi Sabourin
     pat = re.compile(r'^([\w\d]([\w\d\-]{0,61}[\w\d])?\.)*[\w\d]([\w\d\-]{0,61}[\w\d])?$')
     
     if proto == 'tcp':
         lis = addr.split(':')
         assert len(lis) == 2, 'Invalid url: %r'%url
         addr,s_port = lis
         try:
             port = int(s_port)
         except ValueError:
             raise AssertionError("Invalid port %r in url: %r"%(port, url))
         
         assert addr == '*' or pat.match(addr) is not None, 'Invalid url: %r'%url
         
     else:
         # only validate tcp urls currently
         pass
     
     return True
 
 
 def validate_url_container(container):
     """validate a potentially nested collection of urls."""
     if isinstance(container, basestring):
         url = container
         return validate_url(url)
     elif isinstance(container, dict):
         container = container.itervalues()
     
     for element in container:
         validate_url_container(element)
 
 
 def split_url(url):
     """split a zmq url (tcp://ip:port) into ('tcp','ip','port')."""
     proto_addr = url.split('://')
     assert len(proto_addr) == 2, 'Invalid url: %r'%url
     proto, addr = proto_addr
     lis = addr.split(':')
     assert len(lis) == 2, 'Invalid url: %r'%url
     addr,s_port = lis
     return proto,addr,s_port
     
 def disambiguate_ip_address(ip, location=None):
     """turn multi-ip interfaces '0.0.0.0' and '*' into connectable
     ones, based on the location (default interpretation of location is localhost)."""
     if ip in ('0.0.0.0', '*'):
         try:
             external_ips = socket.gethostbyname_ex(socket.gethostname())[2]
         except (socket.gaierror, IndexError):
             # couldn't identify this machine, assume localhost
             external_ips = []
         if location is None or location in external_ips or not external_ips:
             # If location is unspecified or cannot be determined, assume local
             ip='127.0.0.1'
         elif location:
             return location
     return ip
 
 def disambiguate_url(url, location=None):
     """turn multi-ip interfaces '0.0.0.0' and '*' into connectable
     ones, based on the location (default interpretation is localhost).
     
     This is for zeromq urls, such as tcp://*:10101."""
     try:
         proto,ip,port = split_url(url)
     except AssertionError:
         # probably not tcp url; could be ipc, etc.
         return url
     
     ip = disambiguate_ip_address(ip,location)
     
     return "%s://%s:%s"%(proto,ip,port)
 
-def serialize_object(obj, threshold=64e-6):
-    """Serialize an object into a list of sendable buffers.
-    
-    Parameters
-    ----------
-    
-    obj : object
-        The object to be serialized
-    threshold : float
-        The threshold for not double-pickling the content.
-        
-    
-    Returns
-    -------
-    ('pmd', [bufs]) :
-        where pmd is the pickled metadata wrapper,
-        bufs is a list of data buffers
-    """
-    databuffers = []
-    if isinstance(obj, (list, tuple)):
-        clist = canSequence(obj)
-        slist = map(serialize, clist)
-        for s in slist:
-            if s.typeDescriptor in ('buffer', 'ndarray') or s.getDataSize() > threshold:
-                databuffers.append(s.getData())
-                s.data = None
-        return pickle.dumps(slist,-1), databuffers
-    elif isinstance(obj, dict):
-        sobj = {}
-        for k in sorted(obj.iterkeys()):
-            s = serialize(can(obj[k]))
-            if s.typeDescriptor in ('buffer', 'ndarray') or s.getDataSize() > threshold:
-                databuffers.append(s.getData())
-                s.data = None
-            sobj[k] = s
-        return pickle.dumps(sobj,-1),databuffers
-    else:
-        s = serialize(can(obj))
-        if s.typeDescriptor in ('buffer', 'ndarray') or s.getDataSize() > threshold:
-            databuffers.append(s.getData())
-            s.data = None
-        return pickle.dumps(s,-1),databuffers
-            
-        
-def unserialize_object(bufs):
-    """reconstruct an object serialized by serialize_object from data buffers."""
-    bufs = list(bufs)
-    sobj = pickle.loads(bufs.pop(0))
-    if isinstance(sobj, (list, tuple)):
-        for s in sobj:
-            if s.data is None:
-                s.data = bufs.pop(0)
-        return uncanSequence(map(unserialize, sobj)), bufs
-    elif isinstance(sobj, dict):
-        newobj = {}
-        for k in sorted(sobj.iterkeys()):
-            s = sobj[k]
-            if s.data is None:
-                s.data = bufs.pop(0)
-            newobj[k] = uncan(unserialize(s))
-        return newobj, bufs
-    else:
-        if sobj.data is None:
-            sobj.data = bufs.pop(0)
-        return uncan(unserialize(sobj)), bufs
-
-def pack_apply_message(f, args, kwargs, threshold=64e-6):
-    """pack up a function, args, and kwargs to be sent over the wire
-    as a series of buffers. Any object whose data is larger than `threshold`
-    will not have their data copied (currently only numpy arrays support zero-copy)"""
-    msg = [pickle.dumps(can(f),-1)]
-    databuffers = [] # for large objects
-    sargs, bufs = serialize_object(args,threshold)
-    msg.append(sargs)
-    databuffers.extend(bufs)
-    skwargs, bufs = serialize_object(kwargs,threshold)
-    msg.append(skwargs)
-    databuffers.extend(bufs)
-    msg.extend(databuffers)
-    return msg
-
-def unpack_apply_message(bufs, g=None, copy=True):
-    """unpack f,args,kwargs from buffers packed by pack_apply_message()
-    Returns: original f,args,kwargs"""
-    bufs = list(bufs) # allow us to pop
-    assert len(bufs) >= 3, "not enough buffers!"
-    if not copy:
-        for i in range(3):
-            bufs[i] = bufs[i].bytes
-    cf = pickle.loads(bufs.pop(0))
-    sargs = list(pickle.loads(bufs.pop(0)))
-    skwargs = dict(pickle.loads(bufs.pop(0)))
-    # print sargs, skwargs
-    f = uncan(cf, g)
-    for sa in sargs:
-        if sa.data is None:
-            m = bufs.pop(0)
-            if sa.getTypeDescriptor() in ('buffer', 'ndarray'):
-                # always use a buffer, until memoryviews get sorted out
-                sa.data = buffer(m)
-                # disable memoryview support
-                # if copy:
-                #     sa.data = buffer(m)
-                # else:
-                #     sa.data = m.buffer
-            else:
-                if copy:
-                    sa.data = m
-                else:
-                    sa.data = m.bytes
-    
-    args = uncanSequence(map(unserialize, sargs), g)
-    kwargs = {}
-    for k in sorted(skwargs.iterkeys()):
-        sa = skwargs[k]
-        if sa.data is None:
-            m = bufs.pop(0)
-            if sa.getTypeDescriptor() in ('buffer', 'ndarray'):
-                # always use a buffer, until memoryviews get sorted out
-                sa.data = buffer(m)
-                # disable memoryview support
-                # if copy:
-                #     sa.data = buffer(m)
-                # else:
-                #     sa.data = m.buffer
-            else:
-                if copy:
-                    sa.data = m
-                else:
-                    sa.data = m.bytes
-
-        kwargs[k] = uncan(unserialize(sa), g)
-    
-    return f,args,kwargs
 
 #--------------------------------------------------------------------------
 # helpers for implementing old MEC API via view.apply
 #--------------------------------------------------------------------------
 
 def interactive(f):
     """decorator for making functions appear as interactively defined.
     This results in the function being linked to the user_ns as globals()
     instead of the module globals().
     """
     f.__module__ = '__main__'
     return f
 
 @interactive
 def _push(**ns):
     """helper method for implementing `client.push` via `client.apply`"""
     globals().update(ns)
 
 @interactive
 def _pull(keys):
     """helper method for implementing `client.pull` via `client.apply`"""
     user_ns = globals()
     if isinstance(keys, (list,tuple, set)):
         for key in keys:
             if not user_ns.has_key(key):
                 raise NameError("name '%s' is not defined"%key)
         return map(user_ns.get, keys)
     else:
         if not user_ns.has_key(keys):
             raise NameError("name '%s' is not defined"%keys)
         return user_ns.get(keys)
 
 @interactive
 def _execute(code):
     """helper method for implementing `client.execute` via `client.apply`"""
     exec code in globals()
 
 #--------------------------------------------------------------------------
 # extra process management utilities
 #--------------------------------------------------------------------------
 
 _random_ports = set()
 
 def select_random_ports(n):
     """Selects and return n random ports that are available."""
     ports = []
     for i in xrange(n):
         sock = socket.socket()
         sock.bind(('', 0))
         while sock.getsockname()[1] in _random_ports:
             sock.close()
             sock = socket.socket()
             sock.bind(('', 0))
         ports.append(sock)
     for i, sock in enumerate(ports):
         port = sock.getsockname()[1]
         sock.close()
         ports[i] = port
         _random_ports.add(port)
     return ports
 
 def signal_children(children):
     """Relay interupt/term signals to children, for more solid process cleanup."""
     def terminate_children(sig, frame):
         log = Application.instance().log
         log.critical("Got signal %i, terminating children..."%sig)
         for child in children:
             child.terminate()
         
         sys.exit(sig != SIGINT)
         # sys.exit(sig)
     for sig in (SIGINT, SIGABRT, SIGTERM):
         signal(sig, terminate_children)
 
 def generate_exec_key(keyfile):
     import uuid
     newkey = str(uuid.uuid4())
     with open(keyfile, 'w') as f:
         # f.write('ipython-key ')
         f.write(newkey+'\n')
     # set user-only RW permissions (0600)
     # this will have no effect on Windows
     os.chmod(keyfile, stat.S_IRUSR|stat.S_IWUSR)
 
 
 def integer_loglevel(loglevel):
     try:
         loglevel = int(loglevel)
     except ValueError:
         if isinstance(loglevel, str):
             loglevel = getattr(logging, loglevel)
     return loglevel
 
 def connect_logger(logname, context, iface, root="ip", loglevel=logging.DEBUG):
     logger = logging.getLogger(logname)
     if any([isinstance(h, handlers.PUBHandler) for h in logger.handlers]):
         # don't add a second PUBHandler
         return
     loglevel = integer_loglevel(loglevel)
     lsock = context.socket(zmq.PUB)
     lsock.connect(iface)
     handler = handlers.PUBHandler(lsock)
     handler.setLevel(loglevel)
     handler.root_topic = root
     logger.addHandler(handler)
     logger.setLevel(loglevel)
 
 def connect_engine_logger(context, iface, engine, loglevel=logging.DEBUG):
     logger = logging.getLogger()
     if any([isinstance(h, handlers.PUBHandler) for h in logger.handlers]):
         # don't add a second PUBHandler
         return
     loglevel = integer_loglevel(loglevel)
     lsock = context.socket(zmq.PUB)
     lsock.connect(iface)
     handler = EnginePUBHandler(engine, lsock)
     handler.setLevel(loglevel)
     logger.addHandler(handler)
     logger.setLevel(loglevel)
     return logger
 
 def local_logger(logname, loglevel=logging.DEBUG):
     loglevel = integer_loglevel(loglevel)
     logger = logging.getLogger(logname)
     if any([isinstance(h, logging.StreamHandler) for h in logger.handlers]):
         # don't add a second StreamHandler
         return
     handler = logging.StreamHandler()
     handler.setLevel(loglevel)
     formatter = logging.Formatter("%(asctime)s.%(msecs).03d [%(name)s] %(message)s",
                 datefmt="%Y-%m-%d %H:%M:%S")
     handler.setFormatter(formatter)
 
     logger.addHandler(handler)
     logger.setLevel(loglevel)
     return logger
 

IPython/zmq/displayhook.py

@@ -1,72 +1,73 b''
 import __builtin__
 import sys
 from base64 import encodestring
 
 from IPython.core.displayhook import DisplayHook
 from IPython.utils.traitlets import Instance, Dict
 from session import extract_header, Session
 
 class ZMQDisplayHook(object):
     """A simple displayhook that publishes the object's repr over a ZeroMQ
     socket."""
     topic=None
 
     def __init__(self, session, pub_socket):
         self.session = session
         self.pub_socket = pub_socket
         self.parent_header = {}
 
     def __call__(self, obj):
         if obj is None:
             return
 
         __builtin__._ = obj
         sys.stdout.flush()
         sys.stderr.flush()
         msg = self.session.send(self.pub_socket, u'pyout', {u'data':repr(obj)},
                                parent=self.parent_header, ident=self.topic)
 
     def set_parent(self, parent):
         self.parent_header = extract_header(parent)
 
 
 def _encode_binary(format_dict):
     pngdata = format_dict.get('image/png')
     if pngdata is not None:
         format_dict['image/png'] = encodestring(pngdata).decode('ascii')
     jpegdata = format_dict.get('image/jpeg')
     if jpegdata is not None:
         format_dict['image/jpeg'] = encodestring(jpegdata).decode('ascii')
 
 
 class ZMQShellDisplayHook(DisplayHook):
     """A displayhook subclass that publishes data using ZeroMQ. This is intended
     to work with an InteractiveShell instance. It sends a dict of different
     representations of the object."""
+    topic=None
 
     session = Instance(Session)
     pub_socket = Instance('zmq.Socket')
     parent_header = Dict({})
 
     def set_parent(self, parent):
         """Set the parent for outbound messages."""
         self.parent_header = extract_header(parent)
 
     def start_displayhook(self):
         self.msg = self.session.msg(u'pyout', {}, parent=self.parent_header)
 
     def write_output_prompt(self):
         """Write the output prompt."""
         self.msg['content']['execution_count'] = self.prompt_count
 
     def write_format_data(self, format_dict):
         _encode_binary(format_dict)
         self.msg['content']['data'] = format_dict
 
     def finish_displayhook(self):
         """Finish up all displayhook activities."""
         sys.stdout.flush()
         sys.stderr.flush()
-        self.session.send(self.pub_socket, self.msg)
+        self.session.send(self.pub_socket, self.msg, ident=self.topic)
         self.msg = None
 

IPython/zmq/eventloops.py

@@ -1,214 +1,223 b''
 # encoding: utf-8
 """Event loop integration for the ZeroMQ-based kernels.
 """
 
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2011  The IPython Development Team
 
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 import sys
 
 # System library imports.
 import zmq
 
 # Local imports.
+from IPython.config.application import Application
 from IPython.utils import io
 
+
 #------------------------------------------------------------------------------
 # Eventloops for integrating the Kernel into different GUIs
 #------------------------------------------------------------------------------
 
 def loop_qt4(kernel):
     """Start a kernel with PyQt4 event loop integration."""
 
     from IPython.external.qt_for_kernel import QtCore
     from IPython.lib.guisupport import get_app_qt4, start_event_loop_qt4
 
     kernel.app = get_app_qt4([" "])
     kernel.app.setQuitOnLastWindowClosed(False)
     kernel.timer = QtCore.QTimer()
     kernel.timer.timeout.connect(kernel.do_one_iteration)
     # Units for the timer are in milliseconds
     kernel.timer.start(1000*kernel._poll_interval)
     start_event_loop_qt4(kernel.app)
 
 
 def loop_wx(kernel):
     """Start a kernel with wx event loop support."""
 
     import wx
     from IPython.lib.guisupport import start_event_loop_wx
 
     doi = kernel.do_one_iteration
      # Wx uses milliseconds
     poll_interval = int(1000*kernel._poll_interval)
 
     # We have to put the wx.Timer in a wx.Frame for it to fire properly.
     # We make the Frame hidden when we create it in the main app below.
     class TimerFrame(wx.Frame):
         def __init__(self, func):
             wx.Frame.__init__(self, None, -1)
             self.timer = wx.Timer(self)
             # Units for the timer are in milliseconds
             self.timer.Start(poll_interval)
             self.Bind(wx.EVT_TIMER, self.on_timer)
             self.func = func
 
         def on_timer(self, event):
             self.func()
 
     # We need a custom wx.App to create our Frame subclass that has the
     # wx.Timer to drive the ZMQ event loop.
     class IPWxApp(wx.App):
         def OnInit(self):
             self.frame = TimerFrame(doi)
             self.frame.Show(False)
             return True
 
     # The redirect=False here makes sure that wx doesn't replace
     # sys.stdout/stderr with its own classes.
     kernel.app = IPWxApp(redirect=False)
 
     # The import of wx on Linux sets the handler for signal.SIGINT
     # to 0.  This is a bug in wx or gtk.  We fix by just setting it
     # back to the Python default.
     import signal
     if not callable(signal.getsignal(signal.SIGINT)):
         signal.signal(signal.SIGINT, signal.default_int_handler)
 
     start_event_loop_wx(kernel.app)
 
 
 def loop_tk(kernel):
     """Start a kernel with the Tk event loop."""
 
     import Tkinter
     doi = kernel.do_one_iteration
     # Tk uses milliseconds
     poll_interval = int(1000*kernel._poll_interval)
     # For Tkinter, we create a Tk object and call its withdraw method.
     class Timer(object):
         def __init__(self, func):
             self.app = Tkinter.Tk()
             self.app.withdraw()
             self.func = func
 
         def on_timer(self):
             self.func()
             self.app.after(poll_interval, self.on_timer)
 
         def start(self):
             self.on_timer()  # Call it once to get things going.
             self.app.mainloop()
 
     kernel.timer = Timer(doi)
     kernel.timer.start()
 
 
 def loop_gtk(kernel):
     """Start the kernel, coordinating with the GTK event loop"""
     from .gui.gtkembed import GTKEmbed
 
     gtk_kernel = GTKEmbed(kernel)
     gtk_kernel.start()
 
 
 def loop_cocoa(kernel):
     """Start the kernel, coordinating with the Cocoa CFRunLoop event loop
     via the matplotlib MacOSX backend.
     """
     import matplotlib
     if matplotlib.__version__ < '1.1.0':
         kernel.log.warn(
         "MacOSX backend in matplotlib %s doesn't have a Timer, "
         "falling back on Tk for CFRunLoop integration.  Note that "
         "even this won't work if Tk is linked against X11 instead of "
         "Cocoa (e.g. EPD).  To use the MacOSX backend in the kernel, "
         "you must use matplotlib >= 1.1.0, or a native libtk."
         )
         return loop_tk(kernel)
     
     from matplotlib.backends.backend_macosx import TimerMac, show
 
     # scale interval for sec->ms
     poll_interval = int(1000*kernel._poll_interval)
 
     real_excepthook = sys.excepthook
     def handle_int(etype, value, tb):
         """don't let KeyboardInterrupts look like crashes"""
         if etype is KeyboardInterrupt:
             io.raw_print("KeyboardInterrupt caught in CFRunLoop")
         else:
             real_excepthook(etype, value, tb)
 
     # add doi() as a Timer to the CFRunLoop
     def doi():
         # restore excepthook during IPython code
         sys.excepthook = real_excepthook
         kernel.do_one_iteration()
         # and back:
         sys.excepthook = handle_int
 
     t = TimerMac(poll_interval)
     t.add_callback(doi)
     t.start()
 
     # but still need a Poller for when there are no active windows,
     # during which time mainloop() returns immediately
     poller = zmq.Poller()
-    poller.register(kernel.shell_socket, zmq.POLLIN)
+    if kernel.control_stream:
+        poller.register(kernel.control_stream.socket, zmq.POLLIN)
+    for stream in kernel.shell_streams:
+        poller.register(stream.socket, zmq.POLLIN)
 
     while True:
         try:
             # double nested try/except, to properly catch KeyboardInterrupt
             # due to pyzmq Issue #130
             try:
                 # don't let interrupts during mainloop invoke crash_handler:
                 sys.excepthook = handle_int
                 show.mainloop()
                 sys.excepthook = real_excepthook
                 # use poller if mainloop returned (no windows)
                 # scale by extra factor of 10, since it's a real poll
                 poller.poll(10*poll_interval)
                 kernel.do_one_iteration()
             except:
                 raise
         except KeyboardInterrupt:
             # Ctrl-C shouldn't crash the kernel
             io.raw_print("KeyboardInterrupt caught in kernel")
         finally:
             # ensure excepthook is restored
             sys.excepthook = real_excepthook
 
 # mapping of keys to loop functions
 loop_map = {
     'qt' : loop_qt4,
     'qt4': loop_qt4,
     'inline': None,
     'osx': loop_cocoa,
     'wx' : loop_wx,
     'tk' : loop_tk,
     'gtk': loop_gtk,
     None : None,
 }
 
 
 def enable_gui(gui, kernel=None):
     """Enable integration with a given GUI"""
-    if kernel is None:
-        from .ipkernel import IPKernelApp
-        kernel = IPKernelApp.instance().kernel
     if gui not in loop_map:
         raise ValueError("GUI %r not supported" % gui)
+    if kernel is None:
+        if Application.initialized():
+            kernel = getattr(Application.instance(), 'kernel', None)
+        if kernel is None:
+            raise RuntimeError("You didn't specify a kernel,"
+                " and no IPython Application with a kernel appears to be running."
+            )
     loop = loop_map[gui]
     if kernel.eventloop is not None and kernel.eventloop is not loop:
         raise RuntimeError("Cannot activate multiple GUI eventloops")
     kernel.eventloop = loop

IPython/zmq/ipkernel.py

@@ -1,710 +1,919 b''
 #!/usr/bin/env python
 """A simple interactive kernel that talks to a frontend over 0MQ.
 
 Things to do:
 
 * Implement `set_parent` logic. Right before doing exec, the Kernel should
   call set_parent on all the PUB objects with the message about to be executed.
 * Implement random port and security key logic.
 * Implement control messages.
 * Implement event loop and poll version.
 """
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 from __future__ import print_function
 
-# Standard library imports.
+# Standard library imports
 import __builtin__
 import atexit
 import sys
 import time
 import traceback
 import logging
+import uuid
+
+from datetime import datetime
 from signal import (
-        signal, default_int_handler, SIGINT, SIG_IGN
+        signal, getsignal, default_int_handler, SIGINT, SIG_IGN
 )
-# System library imports.
+
+# System library imports
 import zmq
+from zmq.eventloop import ioloop
+from zmq.eventloop.zmqstream import ZMQStream
 
-# Local imports.
+# Local imports
 from IPython.core import pylabtools
 from IPython.config.configurable import Configurable
 from IPython.config.application import boolean_flag, catch_config_error
 from IPython.core.application import ProfileDir
 from IPython.core.error import StdinNotImplementedError
 from IPython.core.shellapp import (
     InteractiveShellApp, shell_flags, shell_aliases
 )
 from IPython.utils import io
 from IPython.utils import py3compat
 from IPython.utils.frame import extract_module_locals
 from IPython.utils.jsonutil import json_clean
 from IPython.utils.traitlets import (
-    Any, Instance, Float, Dict, CaselessStrEnum
+    Any, Instance, Float, Dict, CaselessStrEnum, List, Set, Integer, Unicode
 )
 
 from entry_point import base_launch_kernel
 from kernelapp import KernelApp, kernel_flags, kernel_aliases
+from serialize import serialize_object, unpack_apply_message
 from session import Session, Message
 from zmqshell import ZMQInteractiveShell
 
 
 #-----------------------------------------------------------------------------
 # Main kernel class
 #-----------------------------------------------------------------------------
 
 class Kernel(Configurable):
 
     #---------------------------------------------------------------------------
     # Kernel interface
     #---------------------------------------------------------------------------
 
     # attribute to override with a GUI
     eventloop = Any(None)
+    def _eventloop_changed(self, name, old, new):
+        """schedule call to eventloop from IOLoop"""
+        loop = ioloop.IOLoop.instance()
+        loop.add_timeout(time.time()+0.1, self.enter_eventloop)
 
     shell = Instance('IPython.core.interactiveshell.InteractiveShellABC')
     session = Instance(Session)
     profile_dir = Instance('IPython.core.profiledir.ProfileDir')
-    shell_socket = Instance('zmq.Socket')
-    iopub_socket = Instance('zmq.Socket')
-    stdin_socket = Instance('zmq.Socket')
+    shell_streams = List()
+    control_stream = Instance(ZMQStream)
+    iopub_socket = Instance(zmq.Socket)
+    stdin_socket = Instance(zmq.Socket)
     log = Instance(logging.Logger)
     
-    user_module = Instance('types.ModuleType')
+    user_module = Any()
     def _user_module_changed(self, name, old, new):
         if self.shell is not None:
             self.shell.user_module = new
     
     user_ns = Dict(default_value=None)
     def _user_ns_changed(self, name, old, new):
         if self.shell is not None:
             self.shell.user_ns = new
             self.shell.init_user_ns()
 
-    # Private interface
+    # identities:
+    int_id = Integer(-1)
+    ident = Unicode()
+
+    def _ident_default(self):
+        return unicode(uuid.uuid4())
 
+
+    # Private interface
+    
     # Time to sleep after flushing the stdout/err buffers in each execute
     # cycle.  While this introduces a hard limit on the minimal latency of the
     # execute cycle, it helps prevent output synchronization problems for
     # clients.
     # Units are in seconds.  The minimum zmq latency on local host is probably
     # ~150 microseconds, set this to 500us for now.  We may need to increase it
     # a little if it's not enough after more interactive testing.
     _execute_sleep = Float(0.0005, config=True)
 
     # Frequency of the kernel's event loop.
     # Units are in seconds, kernel subclasses for GUI toolkits may need to
     # adapt to milliseconds.
     _poll_interval = Float(0.05, config=True)
 
     # If the shutdown was requested over the network, we leave here the
     # necessary reply message so it can be sent by our registered atexit
     # handler.  This ensures that the reply is only sent to clients truly at
     # the end of our shutdown process (which happens after the underlying
     # IPython shell's own shutdown).
     _shutdown_message = None
 
     # This is a dict of port number that the kernel is listening on. It is set
     # by record_ports and used by connect_request.
     _recorded_ports = Dict()
-
+    
+    # set of aborted msg_ids
+    aborted = Set()
 
 
     def __init__(self, **kwargs):
         super(Kernel, self).__init__(**kwargs)
 
-        # Before we even start up the shell, register *first* our exit handlers
-        # so they come before the shell's
-        atexit.register(self._at_shutdown)
-
         # Initialize the InteractiveShell subclass
         self.shell = ZMQInteractiveShell.instance(config=self.config,
             profile_dir = self.profile_dir,
             user_module = self.user_module,
             user_ns     = self.user_ns,
         )
         self.shell.displayhook.session = self.session
         self.shell.displayhook.pub_socket = self.iopub_socket
+        self.shell.displayhook.topic = self._topic('pyout')
         self.shell.display_pub.session = self.session
         self.shell.display_pub.pub_socket = self.iopub_socket
 
         # TMP - hack while developing
         self.shell._reply_content = None
 
         # Build dict of handlers for message types
         msg_types = [ 'execute_request', 'complete_request',
                       'object_info_request', 'history_request',
-                      'connect_request', 'shutdown_request']
-        self.handlers = {}
+                      'connect_request', 'shutdown_request',
+                      'apply_request',
+                    ]
+        self.shell_handlers = {}
         for msg_type in msg_types:
-            self.handlers[msg_type] = getattr(self, msg_type)
-
-    def do_one_iteration(self):
-        """Do one iteration of the kernel's evaluation loop.
-        """
+            self.shell_handlers[msg_type] = getattr(self, msg_type)
+        
+        control_msg_types = msg_types + [ 'clear_request', 'abort_request' ]
+        self.control_handlers = {}
+        for msg_type in control_msg_types:
+            self.control_handlers[msg_type] = getattr(self, msg_type)
+
+    def dispatch_control(self, msg):
+        """dispatch control requests"""
+        idents,msg = self.session.feed_identities(msg, copy=False)
         try:
-            ident,msg = self.session.recv(self.shell_socket, zmq.NOBLOCK)
-        except Exception:
-            self.log.warn("Invalid Message:", exc_info=True)
-            return
-        if msg is None:
+            msg = self.session.unserialize(msg, content=True, copy=False)
+        except:
+            self.log.error("Invalid Control Message", exc_info=True)
             return
 
-        msg_type = msg['header']['msg_type']
+        self.log.debug("Control received: %s", msg)
+
+        header = msg['header']
+        msg_id = header['msg_id']
+        msg_type = header['msg_type']
 
-        # This assert will raise in versions of zeromq 2.0.7 and lesser.
-        # We now require 2.0.8 or above, so we can uncomment for safety.
-        # print(ident,msg, file=sys.__stdout__)
-        assert ident is not None, "Missing message part."
+        handler = self.control_handlers.get(msg_type, None)
+        if handler is None:
+            self.log.error("UNKNOWN CONTROL MESSAGE TYPE: %r", msg_type)
+        else:
+            try:
+                handler(self.control_stream, idents, msg)
+            except Exception:
+                self.log.error("Exception in control handler:", exc_info=True)
+    
+    def dispatch_shell(self, stream, msg):
+        """dispatch shell requests"""
+        # flush control requests first
+        if self.control_stream:
+            self.control_stream.flush()
+        
+        idents,msg = self.session.feed_identities(msg, copy=False)
+        try:
+            msg = self.session.unserialize(msg, content=True, copy=False)
+        except:
+            self.log.error("Invalid Message", exc_info=True)
+            return
 
+        header = msg['header']
+        msg_id = header['msg_id']
+        msg_type = msg['header']['msg_type']
+        
         # Print some info about this message and leave a '--->' marker, so it's
         # easier to trace visually the message chain when debugging.  Each
         # handler prints its message at the end.
-        self.log.debug('\n*** MESSAGE TYPE:'+str(msg_type)+'***')
-        self.log.debug('   Content: '+str(msg['content'])+'\n   --->\n   ')
+        self.log.debug('\n*** MESSAGE TYPE:%s***', msg_type)
+        self.log.debug('   Content: %s\n   --->\n   ', msg['content'])
 
-        # Find and call actual handler for message
-        handler = self.handlers.get(msg_type, None)
+        if msg_id in self.aborted:
+            self.aborted.remove(msg_id)
+            # is it safe to assume a msg_id will not be resubmitted?
+            reply_type = msg_type.split('_')[0] + '_reply'
+            status = {'status' : 'aborted'}
+            sub = {'engine' : self.ident}
+            sub.update(status)
+            reply_msg = self.session.send(stream, reply_type, subheader=sub,
+                        content=status, parent=msg, ident=idents)
+            return
+        
+        handler = self.shell_handlers.get(msg_type, None)
         if handler is None:
-            self.log.error("UNKNOWN MESSAGE TYPE:" +str(msg))
+            self.log.error("UNKNOWN MESSAGE TYPE: %r", msg_type)
         else:
-            handler(ident, msg)
-
-        # Check whether we should exit, in case the incoming message set the
-        # exit flag on
-        if self.shell.exit_now:
-            self.log.debug('\nExiting IPython kernel...')
-            # We do a normal, clean exit, which allows any actions registered
-            # via atexit (such as history saving) to take place.
-            sys.exit(0)
-
-
-    def start(self):
-        """ Start the kernel main loop.
-        """
-        # a KeyboardInterrupt (SIGINT) can occur on any python statement, so
-        # let's ignore (SIG_IGN) them until we're in a place to handle them properly
-        signal(SIGINT,SIG_IGN)
-        poller = zmq.Poller()
-        poller.register(self.shell_socket, zmq.POLLIN)
-        # loop while self.eventloop has not been overridden
-        while self.eventloop is None:
+            # ensure default_int_handler during handler call
+            sig = signal(SIGINT, default_int_handler)
             try:
-                # scale by extra factor of 10, because there is no
-                # reason for this to be anything less than ~ 0.1s
-                # since it is a real poller and will respond
-                # to events immediately
-
-                # double nested try/except, to properly catch KeyboardInterrupt
-                # due to pyzmq Issue #130
-                try:
-                    poller.poll(10*1000*self._poll_interval)
-                    # restore raising of KeyboardInterrupt
-                    signal(SIGINT, default_int_handler)
-                    self.do_one_iteration()
-                except:
-                    raise
-                finally:
-                    # prevent raising of KeyboardInterrupt
-                    signal(SIGINT,SIG_IGN)
-            except KeyboardInterrupt:
-                # Ctrl-C shouldn't crash the kernel
-                io.raw_print("KeyboardInterrupt caught in kernel")
-        # stop ignoring sigint, now that we are out of our own loop,
-        # we don't want to prevent future code from handling it
+                handler(stream, idents, msg)
+            except Exception:
+                self.log.error("Exception in message handler:", exc_info=True)
+            finally:
+                signal(SIGINT, sig)
+    
+    def enter_eventloop(self):
+        """enter eventloop"""
+        self.log.info("entering eventloop")
+        # restore default_int_handler
         signal(SIGINT, default_int_handler)
         while self.eventloop is not None:
             try:
                 self.eventloop(self)
             except KeyboardInterrupt:
                 # Ctrl-C shouldn't crash the kernel
-                io.raw_print("KeyboardInterrupt caught in kernel")
+                self.log.error("KeyboardInterrupt caught in kernel")
                 continue
             else:
                 # eventloop exited cleanly, this means we should stop (right?)
                 self.eventloop = None
                 break
+        self.log.info("exiting eventloop")
+        # if eventloop exits, IOLoop should stop
+        ioloop.IOLoop.instance().stop()
+
+    def start(self):
+        """register dispatchers for streams"""
+        self.shell.exit_now = False
+        if self.control_stream:
+            self.control_stream.on_recv(self.dispatch_control, copy=False)
+
+        def make_dispatcher(stream):
+            def dispatcher(msg):
+                return self.dispatch_shell(stream, msg)
+            return dispatcher
+
+        for s in self.shell_streams:
+            s.on_recv(make_dispatcher(s), copy=False)
+    
+    def do_one_iteration(self):
+        """step eventloop just once"""
+        if self.control_stream:
+            self.control_stream.flush()
+        for stream in self.shell_streams:
+            # handle at most one request per iteration
+            stream.flush(zmq.POLLIN, 1)
+            stream.flush(zmq.POLLOUT)
 
 
     def record_ports(self, ports):
         """Record the ports that this kernel is using.
 
         The creator of the Kernel instance must call this methods if they
         want the :meth:`connect_request` method to return the port numbers.
         """
         self._recorded_ports = ports
 
     #---------------------------------------------------------------------------
     # Kernel request handlers
     #---------------------------------------------------------------------------
-
+    
+    def _make_subheader(self):
+        """init subheader dict, for execute/apply_reply"""
+        return {
+            'dependencies_met' : True,
+            'engine' : self.ident,
+            'started': datetime.now(),
+        }
+    
     def _publish_pyin(self, code, parent, execution_count):
         """Publish the code request on the pyin stream."""
 
-        self.session.send(self.iopub_socket, u'pyin', {u'code':code,
-                          u'execution_count': execution_count}, parent=parent)
+        self.session.send(self.iopub_socket, u'pyin',
+                            {u'code':code, u'execution_count': execution_count},
+                            parent=parent, ident=self._topic('pyin')
+        )
 
-    def execute_request(self, ident, parent):
+    def execute_request(self, stream, ident, parent):
 
         self.session.send(self.iopub_socket,
                           u'status',
                           {u'execution_state':u'busy'},
-                          parent=parent )
+                          parent=parent,
+                          ident=self._topic('status'),
+                          )
         
         try:
             content = parent[u'content']
             code = content[u'code']
             silent = content[u'silent']
         except:
             self.log.error("Got bad msg: ")
-            self.log.error(str(Message(parent)))
+            self.log.error("%s", parent)
             return
+        
+        sub = self._make_subheader()
 
         shell = self.shell # we'll need this a lot here
 
         # Replace raw_input. Note that is not sufficient to replace
         # raw_input in the user namespace.
         if content.get('allow_stdin', False):
             raw_input = lambda prompt='': self._raw_input(prompt, ident, parent)
         else:
             raw_input = lambda prompt='' : self._no_raw_input()
 
         if py3compat.PY3:
             __builtin__.input = raw_input
         else:
             __builtin__.raw_input = raw_input
 
         # Set the parent message of the display hook and out streams.
         shell.displayhook.set_parent(parent)
         shell.display_pub.set_parent(parent)
         sys.stdout.set_parent(parent)
         sys.stderr.set_parent(parent)
 
         # Re-broadcast our input for the benefit of listening clients, and
         # start computing output
         if not silent:
             self._publish_pyin(code, parent, shell.execution_count)
 
         reply_content = {}
         try:
-            if silent:
-                # run_code uses 'exec' mode, so no displayhook will fire, and it
-                # doesn't call logging or history manipulations.  Print
-                # statements in that code will obviously still execute.
-                shell.run_code(code)
-            else:
-                # FIXME: the shell calls the exception handler itself.
-                shell.run_cell(code, store_history=True)
+            # FIXME: the shell calls the exception handler itself.
+            shell.run_cell(code, store_history=not silent, silent=silent)
         except:
             status = u'error'
             # FIXME: this code right now isn't being used yet by default,
             # because the run_cell() call above directly fires off exception
             # reporting.  This code, therefore, is only active in the scenario
             # where runlines itself has an unhandled exception.  We need to
             # uniformize this, for all exception construction to come from a
             # single location in the codbase.
             etype, evalue, tb = sys.exc_info()
             tb_list = traceback.format_exception(etype, evalue, tb)
             reply_content.update(shell._showtraceback(etype, evalue, tb_list))
         else:
             status = u'ok'
 
         reply_content[u'status'] = status
 
         # Return the execution counter so clients can display prompts
-        reply_content['execution_count'] = shell.execution_count -1
+        reply_content['execution_count'] = shell.execution_count - 1
 
         # FIXME - fish exception info out of shell, possibly left there by
         # runlines.  We'll need to clean up this logic later.
         if shell._reply_content is not None:
             reply_content.update(shell._reply_content)
             # reset after use
             shell._reply_content = None
 
         # At this point, we can tell whether the main code execution succeeded
         # or not.  If it did, we proceed to evaluate user_variables/expressions
         if reply_content['status'] == 'ok':
             reply_content[u'user_variables'] = \
-                         shell.user_variables(content[u'user_variables'])
+                         shell.user_variables(content.get(u'user_variables', []))
             reply_content[u'user_expressions'] = \
-                         shell.user_expressions(content[u'user_expressions'])
+                         shell.user_expressions(content.get(u'user_expressions', {}))
         else:
             # If there was an error, don't even try to compute variables or
             # expressions
             reply_content[u'user_variables'] = {}
             reply_content[u'user_expressions'] = {}
 
         # Payloads should be retrieved regardless of outcome, so we can both
         # recover partial output (that could have been generated early in a
         # block, before an error) and clear the payload system always.
         reply_content[u'payload'] = shell.payload_manager.read_payload()
         # Be agressive about clearing the payload because we don't want
         # it to sit in memory until the next execute_request comes in.
         shell.payload_manager.clear_payload()
 
         # Flush output before sending the reply.
         sys.stdout.flush()
         sys.stderr.flush()
         # FIXME: on rare occasions, the flush doesn't seem to make it to the
         # clients... This seems to mitigate the problem, but we definitely need
         # to better understand what's going on.
         if self._execute_sleep:
             time.sleep(self._execute_sleep)
 
         # Send the reply.
         reply_content = json_clean(reply_content)
-        reply_msg = self.session.send(self.shell_socket, u'execute_reply',
-                                      reply_content, parent, ident=ident)
-        self.log.debug(str(reply_msg))
+        
+        sub['status'] = reply_content['status']
+        if reply_content['status'] == 'error' and \
+                        reply_content['ename'] == 'UnmetDependency':
+                sub['dependencies_met'] = False
+
+        reply_msg = self.session.send(stream, u'execute_reply',
+                                      reply_content, parent, subheader=sub,
+                                      ident=ident)
+        
+        self.log.debug("%s", reply_msg)
 
-        if reply_msg['content']['status'] == u'error':
-            self._abort_queue()
+        if not silent and reply_msg['content']['status'] == u'error':
+            self._abort_queues()
 
         self.session.send(self.iopub_socket,
                           u'status',
                           {u'execution_state':u'idle'},
-                          parent=parent )
+                          parent=parent,
+                          ident=self._topic('status'))
 
-    def complete_request(self, ident, parent):
+    def complete_request(self, stream, ident, parent):
         txt, matches = self._complete(parent)
         matches = {'matches' : matches,
                    'matched_text' : txt,
                    'status' : 'ok'}
         matches = json_clean(matches)
-        completion_msg = self.session.send(self.shell_socket, 'complete_reply',
+        completion_msg = self.session.send(stream, 'complete_reply',
                                            matches, parent, ident)
-        self.log.debug(str(completion_msg))
+        self.log.debug("%s", completion_msg)
 
-    def object_info_request(self, ident, parent):
+    def object_info_request(self, stream, ident, parent):
         content = parent['content']
         object_info = self.shell.object_inspect(content['oname'],
                         detail_level = content.get('detail_level', 0)
         )
         # Before we send this object over, we scrub it for JSON usage
         oinfo = json_clean(object_info)
-        msg = self.session.send(self.shell_socket, 'object_info_reply',
+        msg = self.session.send(stream, 'object_info_reply',
                                 oinfo, parent, ident)
-        self.log.debug(msg)
+        self.log.debug("%s", msg)
 
-    def history_request(self, ident, parent):
+    def history_request(self, stream, ident, parent):
         # We need to pull these out, as passing **kwargs doesn't work with
         # unicode keys before Python 2.6.5.
         hist_access_type = parent['content']['hist_access_type']
         raw = parent['content']['raw']
         output = parent['content']['output']
         if hist_access_type == 'tail':
             n = parent['content']['n']
             hist = self.shell.history_manager.get_tail(n, raw=raw, output=output,
                                                             include_latest=True)
 
         elif hist_access_type == 'range':
             session = parent['content']['session']
             start = parent['content']['start']
             stop = parent['content']['stop']
             hist = self.shell.history_manager.get_range(session, start, stop,
                                                         raw=raw, output=output)
 
         elif hist_access_type == 'search':
             pattern = parent['content']['pattern']
             hist = self.shell.history_manager.search(pattern, raw=raw,
                                                      output=output) 
 
         else:
             hist = []
         hist = list(hist)
         content = {'history' : hist}
         content = json_clean(content)
-        msg = self.session.send(self.shell_socket, 'history_reply',
+        msg = self.session.send(stream, 'history_reply',
                                 content, parent, ident)
         self.log.debug("Sending history reply with %i entries", len(hist))
 
-    def connect_request(self, ident, parent):
+    def connect_request(self, stream, ident, parent):
         if self._recorded_ports is not None:
             content = self._recorded_ports.copy()
         else:
             content = {}
-        msg = self.session.send(self.shell_socket, 'connect_reply',
+        msg = self.session.send(stream, 'connect_reply',
                                 content, parent, ident)
-        self.log.debug(msg)
+        self.log.debug("%s", msg)
 
-    def shutdown_request(self, ident, parent):
+    def shutdown_request(self, stream, ident, parent):
         self.shell.exit_now = True
+        content = dict(status='ok')
+        content.update(parent['content'])
+        self.session.send(stream, u'shutdown_reply', content, parent, ident=ident)
+        # same content, but different msg_id for broadcasting on IOPub
         self._shutdown_message = self.session.msg(u'shutdown_reply',
-                                                  parent['content'], parent)
-        sys.exit(0)
+                                                  content, parent
+        )
+
+        self._at_shutdown()
+        # call sys.exit after a short delay
+        loop = ioloop.IOLoop.instance()
+        loop.add_timeout(time.time()+0.1, loop.stop)
+
+    #---------------------------------------------------------------------------
+    # Engine methods
+    #---------------------------------------------------------------------------
+
+    def apply_request(self, stream, ident, parent):
+        try:
+            content = parent[u'content']
+            bufs = parent[u'buffers']
+            msg_id = parent['header']['msg_id']
+        except:
+            self.log.error("Got bad msg: %s", parent, exc_info=True)
+            return
+
+        # Set the parent message of the display hook and out streams.
+        self.shell.displayhook.set_parent(parent)
+        self.shell.display_pub.set_parent(parent)
+        sys.stdout.set_parent(parent)
+        sys.stderr.set_parent(parent)
+
+        # pyin_msg = self.session.msg(u'pyin',{u'code':code}, parent=parent)
+        # self.iopub_socket.send(pyin_msg)
+        # self.session.send(self.iopub_socket, u'pyin', {u'code':code},parent=parent)
+        sub = self._make_subheader()
+        try:
+            working = self.shell.user_ns
+
+            prefix = "_"+str(msg_id).replace("-","")+"_"
+
+            f,args,kwargs = unpack_apply_message(bufs, working, copy=False)
+
+            fname = getattr(f, '__name__', 'f')
+
+            fname = prefix+"f"
+            argname = prefix+"args"
+            kwargname = prefix+"kwargs"
+            resultname = prefix+"result"
+
+            ns = { fname : f, argname : args, kwargname : kwargs , resultname : None }
+            # print ns
+            working.update(ns)
+            code = "%s = %s(*%s,**%s)" % (resultname, fname, argname, kwargname)
+            try:
+                exec code in self.shell.user_global_ns, self.shell.user_ns
+                result = working.get(resultname)
+            finally:
+                for key in ns.iterkeys():
+                    working.pop(key)
+
+            packed_result,buf = serialize_object(result)
+            result_buf = [packed_result]+buf
+        except:
+            exc_content = self._wrap_exception('apply')
+            # exc_msg = self.session.msg(u'pyerr', exc_content, parent)
+            self.session.send(self.iopub_socket, u'pyerr', exc_content, parent=parent,
+                                ident=self._topic('pyerr'))
+            reply_content = exc_content
+            result_buf = []
+
+            if exc_content['ename'] == 'UnmetDependency':
+                sub['dependencies_met'] = False
+        else:
+            reply_content = {'status' : 'ok'}
+
+        # put 'ok'/'error' status in header, for scheduler introspection:
+        sub['status'] = reply_content['status']
+
+        # flush i/o
+        sys.stdout.flush()
+        sys.stderr.flush()
+        
+        reply_msg = self.session.send(stream, u'apply_reply', reply_content,
+                    parent=parent, ident=ident,buffers=result_buf, subheader=sub)
+
+    #---------------------------------------------------------------------------
+    # Control messages
+    #---------------------------------------------------------------------------
+
+    def abort_request(self, stream, ident, parent):
+        """abort a specifig msg by id"""
+        msg_ids = parent['content'].get('msg_ids', None)
+        if isinstance(msg_ids, basestring):
+            msg_ids = [msg_ids]
+        if not msg_ids:
+            self.abort_queues()
+        for mid in msg_ids:
+            self.aborted.add(str(mid))
+
+        content = dict(status='ok')
+        reply_msg = self.session.send(stream, 'abort_reply', content=content,
+                parent=parent, ident=ident)
+        self.log.debug("%s", reply_msg)
+
+    def clear_request(self, stream, idents, parent):
+        """Clear our namespace."""
+        self.shell.reset(False)
+        msg = self.session.send(stream, 'clear_reply', ident=idents, parent=parent,
+                content = dict(status='ok'))
+
 
     #---------------------------------------------------------------------------
     # Protected interface
     #---------------------------------------------------------------------------
 
-    def _abort_queue(self):
+
+    def _wrap_exception(self, method=None):
+        # import here, because _wrap_exception is only used in parallel,
+        # and parallel has higher min pyzmq version
+        from IPython.parallel.error import wrap_exception
+        e_info = dict(engine_uuid=self.ident, engine_id=self.int_id, method=method)
+        content = wrap_exception(e_info)
+        return content
+
+    def _topic(self, topic):
+        """prefixed topic for IOPub messages"""
+        if self.int_id >= 0:
+            base = "engine.%i" % self.int_id
+        else:
+            base = "kernel.%s" % self.ident
+        
+        return py3compat.cast_bytes("%s.%s" % (base, topic))
+    
+    def _abort_queues(self):
+        for stream in self.shell_streams:
+            if stream:
+                self._abort_queue(stream)
+
+    def _abort_queue(self, stream):
+        poller = zmq.Poller()
+        poller.register(stream.socket, zmq.POLLIN)
         while True:
-            try:
-                ident,msg = self.session.recv(self.shell_socket, zmq.NOBLOCK)
-            except Exception:
-                self.log.warn("Invalid Message:", exc_info=True)
-                continue
+            idents,msg = self.session.recv(stream, zmq.NOBLOCK, content=True)
             if msg is None:
-                break
-            else:
-                assert ident is not None, \
-                       "Unexpected missing message part."
+                return
 
-            self.log.debug("Aborting:\n"+str(Message(msg)))
+            self.log.info("Aborting:")
+            self.log.info("%s", msg)
             msg_type = msg['header']['msg_type']
             reply_type = msg_type.split('_')[0] + '_reply'
-            reply_msg = self.session.send(self.shell_socket, reply_type,
-                    {'status' : 'aborted'}, msg, ident=ident)
-            self.log.debug(reply_msg)
+
+            status = {'status' : 'aborted'}
+            sub = {'engine' : self.ident}
+            sub.update(status)
+            reply_msg = self.session.send(stream, reply_type, subheader=sub,
+                        content=status, parent=msg, ident=idents)
+            self.log.debug("%s", reply_msg)
             # We need to wait a bit for requests to come in. This can probably
             # be set shorter for true asynchronous clients.
-            time.sleep(0.1)
+            poller.poll(50)
+
 
     def _no_raw_input(self):
         """Raise StdinNotImplentedError if active frontend doesn't support
         stdin."""
         raise StdinNotImplementedError("raw_input was called, but this "
                                        "frontend does not support stdin.") 
         
     def _raw_input(self, prompt, ident, parent):
         # Flush output before making the request.
         sys.stderr.flush()
         sys.stdout.flush()
 
         # Send the input request.
         content = json_clean(dict(prompt=prompt))
         self.session.send(self.stdin_socket, u'input_request', content, parent,
                           ident=ident)
 
         # Await a response.
         while True:
             try:
                 ident, reply = self.session.recv(self.stdin_socket, 0)
             except Exception:
                 self.log.warn("Invalid Message:", exc_info=True)
             else:
                 break
         try:
             value = reply['content']['value']
         except:
             self.log.error("Got bad raw_input reply: ")
-            self.log.error(str(Message(parent)))
+            self.log.error("%s", parent)
             value = ''
         if value == '\x04':
             # EOF
             raise EOFError
         return value
 
     def _complete(self, msg):
         c = msg['content']
         try:
             cpos = int(c['cursor_pos'])
         except:
             # If we don't get something that we can convert to an integer, at
             # least attempt the completion guessing the cursor is at the end of
             # the text, if there's any, and otherwise of the line
             cpos = len(c['text'])
             if cpos==0:
                 cpos = len(c['line'])
         return self.shell.complete(c['text'], c['line'], cpos)
 
     def _object_info(self, context):
         symbol, leftover = self._symbol_from_context(context)
         if symbol is not None and not leftover:
             doc = getattr(symbol, '__doc__', '')
         else:
             doc = ''
         object_info = dict(docstring = doc)
         return object_info
 
     def _symbol_from_context(self, context):
         if not context:
             return None, context
 
         base_symbol_string = context[0]
         symbol = self.shell.user_ns.get(base_symbol_string, None)
         if symbol is None:
             symbol = __builtin__.__dict__.get(base_symbol_string, None)
         if symbol is None:
             return None, context
 
         context = context[1:]
         for i, name in enumerate(context):
             new_symbol = getattr(symbol, name, None)
             if new_symbol is None:
                 return symbol, context[i:]
             else:
                 symbol = new_symbol
 
         return symbol, []
 
     def _at_shutdown(self):
         """Actions taken at shutdown by the kernel, called by python's atexit.
         """
         # io.rprint("Kernel at_shutdown") # dbg
         if self._shutdown_message is not None:
-            self.session.send(self.shell_socket, self._shutdown_message)
-            self.session.send(self.iopub_socket, self._shutdown_message)
-            self.log.debug(str(self._shutdown_message))
-            # A very short sleep to give zmq time to flush its message buffers
-            # before Python truly shuts down.
-            time.sleep(0.01)
+            self.session.send(self.iopub_socket, self._shutdown_message, ident=self._topic('shutdown'))
+            self.log.debug("%s", self._shutdown_message)
+        [ s.flush(zmq.POLLOUT) for s in self.shell_streams ]
 
 #-----------------------------------------------------------------------------
 # Aliases and Flags for the IPKernelApp
 #-----------------------------------------------------------------------------
 
 flags = dict(kernel_flags)
 flags.update(shell_flags)
 
 addflag = lambda *args: flags.update(boolean_flag(*args))
 
 flags['pylab'] = (
     {'IPKernelApp' : {'pylab' : 'auto'}},
     """Pre-load matplotlib and numpy for interactive use with
     the default matplotlib backend."""
 )
 
 aliases = dict(kernel_aliases)
 aliases.update(shell_aliases)
 
 # it's possible we don't want short aliases for *all* of these:
 aliases.update(dict(
     pylab='IPKernelApp.pylab',
 ))
 
 #-----------------------------------------------------------------------------
 # The IPKernelApp class
 #-----------------------------------------------------------------------------
 
 class IPKernelApp(KernelApp, InteractiveShellApp):
     name = 'ipkernel'
 
     aliases = Dict(aliases)
     flags = Dict(flags)
     classes = [Kernel, ZMQInteractiveShell, ProfileDir, Session]
     
     # configurables
     pylab = CaselessStrEnum(['tk', 'qt', 'wx', 'gtk', 'osx', 'inline', 'auto'],
         config=True,
         help="""Pre-load matplotlib and numpy for interactive use,
         selecting a particular matplotlib backend and loop integration.
         """
     )
     
     @catch_config_error
     def initialize(self, argv=None):
         super(IPKernelApp, self).initialize(argv)
         self.init_path()
         self.init_shell()
         self.init_extensions()
         self.init_code()
 
     def init_kernel(self):
+        
+        shell_stream = ZMQStream(self.shell_socket)
 
         kernel = Kernel(config=self.config, session=self.session,
-                                shell_socket=self.shell_socket,
+                                shell_streams=[shell_stream],
                                 iopub_socket=self.iopub_socket,
                                 stdin_socket=self.stdin_socket,
                                 log=self.log,
                                 profile_dir=self.profile_dir,
         )
         self.kernel = kernel
         kernel.record_ports(self.ports)
         shell = kernel.shell
         if self.pylab:
             try:
                 gui, backend = pylabtools.find_gui_and_backend(self.pylab)
                 shell.enable_pylab(gui, import_all=self.pylab_import_all)
             except Exception:
                 self.log.error("Pylab initialization failed", exc_info=True)
                 # print exception straight to stdout, because normally 
                 # _showtraceback associates the reply with an execution, 
                 # which means frontends will never draw it, as this exception 
                 # is not associated with any execute request.
                 
                 # replace pyerr-sending traceback with stdout
                 _showtraceback = shell._showtraceback
                 def print_tb(etype, evalue, stb):
                     print ("Error initializing pylab, pylab mode will not "
                            "be active", file=io.stderr)
                     print (shell.InteractiveTB.stb2text(stb), file=io.stdout)
                 shell._showtraceback = print_tb
                 
                 # send the traceback over stdout
                 shell.showtraceback(tb_offset=0)
                 
                 # restore proper _showtraceback method
                 shell._showtraceback = _showtraceback
                 
 
     def init_shell(self):
         self.shell = self.kernel.shell
         self.shell.configurables.append(self)
 
 
 #-----------------------------------------------------------------------------
 # Kernel main and launch functions
 #-----------------------------------------------------------------------------
 
 def launch_kernel(*args, **kwargs):
     """Launches a localhost IPython kernel, binding to the specified ports.
 
     This function simply calls entry_point.base_launch_kernel with the right
     first command to start an ipkernel.  See base_launch_kernel for arguments.
 
     Returns
     -------
     A tuple of form:
         (kernel_process, shell_port, iopub_port, stdin_port, hb_port)
     where kernel_process is a Popen object and the ports are integers.
     """
     return base_launch_kernel('from IPython.zmq.ipkernel import main; main()',
                               *args, **kwargs)
 
 
 def embed_kernel(module=None, local_ns=None, **kwargs):
     """Embed and start an IPython kernel in a given scope.
     
     Parameters
     ----------
     module : ModuleType, optional
         The module to load into IPython globals (default: caller)
     local_ns : dict, optional
         The namespace to load into IPython user namespace (default: caller)
     
     kwargs : various, optional
         Further keyword args are relayed to the KernelApp constructor,
         allowing configuration of the Kernel.  Will only have an effect
         on the first embed_kernel call for a given process.
     
     """
     # get the app if it exists, or set it up if it doesn't
     if IPKernelApp.initialized():
         app = IPKernelApp.instance()
     else:
         app = IPKernelApp.instance(**kwargs)
         app.initialize([])
+        # Undo unnecessary sys module mangling from init_sys_modules.
+        # This would not be necessary if we could prevent it
+        # in the first place by using a different InteractiveShell
+        # subclass, as in the regular embed case.
+        main = app.kernel.shell._orig_sys_modules_main_mod
+        if main is not None:
+            sys.modules[app.kernel.shell._orig_sys_modules_main_name] = main
 
     # load the calling scope if not given
     (caller_module, caller_locals) = extract_module_locals(1)
     if module is None:
         module = caller_module
     if local_ns is None:
         local_ns = caller_locals
     
     app.kernel.user_module = module
     app.kernel.user_ns = local_ns
     app.start()
 
 def main():
     """Run an IPKernel as an application"""
     app = IPKernelApp.instance()
     app.initialize()
     app.start()
 
 
 if __name__ == '__main__':
     main()

IPython/zmq/kernelapp.py

@@ -1,306 +1,333 b''
 """An Application for launching a kernel
 
 Authors
 -------
 * MinRK
 """
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING.txt, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
-# Standard library imports.
+# Standard library imports
+import atexit
 import json
 import os
 import sys
+import signal
 
-# System library imports.
+# System library imports
 import zmq
+from zmq.eventloop import ioloop
 
-# IPython imports.
+# IPython imports
 from IPython.core.ultratb import FormattedTB
 from IPython.core.application import (
     BaseIPythonApplication, base_flags, base_aliases, catch_config_error
 )
 from IPython.utils import io
 from IPython.utils.localinterfaces import LOCALHOST
 from IPython.utils.path import filefind
 from IPython.utils.py3compat import str_to_bytes
 from IPython.utils.traitlets import (Any, Instance, Dict, Unicode, Integer, Bool,
                                         DottedObjectName)
 from IPython.utils.importstring import import_item
 # local imports
 from IPython.zmq.entry_point import write_connection_file
 from IPython.zmq.heartbeat import Heartbeat
 from IPython.zmq.parentpoller import ParentPollerUnix, ParentPollerWindows
 from IPython.zmq.session import (
     Session, session_flags, session_aliases, default_secure,
 )
 
 
 #-----------------------------------------------------------------------------
 # Flags and Aliases
 #-----------------------------------------------------------------------------
 
 kernel_aliases = dict(base_aliases)
 kernel_aliases.update({
     'ip' : 'KernelApp.ip',
     'hb' : 'KernelApp.hb_port',
     'shell' : 'KernelApp.shell_port',
     'iopub' : 'KernelApp.iopub_port',
     'stdin' : 'KernelApp.stdin_port',
     'f' : 'KernelApp.connection_file',
     'parent': 'KernelApp.parent',
 })
 if sys.platform.startswith('win'):
     kernel_aliases['interrupt'] = 'KernelApp.interrupt'
 
 kernel_flags = dict(base_flags)
 kernel_flags.update({
     'no-stdout' : (
             {'KernelApp' : {'no_stdout' : True}},
             "redirect stdout to the null device"),
     'no-stderr' : (
             {'KernelApp' : {'no_stderr' : True}},
             "redirect stderr to the null device"),
 })
 
 # inherit flags&aliases for Sessions
 kernel_aliases.update(session_aliases)
 kernel_flags.update(session_flags)
 
 
 
 #-----------------------------------------------------------------------------
 # Application class for starting a Kernel
 #-----------------------------------------------------------------------------
 
 class KernelApp(BaseIPythonApplication):
-    name='pykernel'
+    name='ipkernel'
     aliases = Dict(kernel_aliases)
     flags = Dict(kernel_flags)
     classes = [Session]
     # the kernel class, as an importstring
-    kernel_class = DottedObjectName('IPython.zmq.pykernel.Kernel')
+    kernel_class = DottedObjectName('IPython.zmq.ipkernel.Kernel')
     kernel = Any()
     poller = Any() # don't restrict this even though current pollers are all Threads
     heartbeat = Instance(Heartbeat)
     session = Instance('IPython.zmq.session.Session')
     ports = Dict()
+    _full_connection_file = Unicode()
     
     # inherit config file name from parent:
     parent_appname = Unicode(config=True)
     def _parent_appname_changed(self, name, old, new):
         if self.config_file_specified:
             # it was manually specified, ignore
             return
         self.config_file_name = new.replace('-','_') + u'_config.py'
         # don't let this count as specifying the config file
         self.config_file_specified = False
         
     # connection info:
     ip = Unicode(LOCALHOST, config=True,
         help="Set the IP or interface on which the kernel will listen.")
     hb_port = Integer(0, config=True, help="set the heartbeat port [default: random]")
     shell_port = Integer(0, config=True, help="set the shell (XREP) port [default: random]")
     iopub_port = Integer(0, config=True, help="set the iopub (PUB) port [default: random]")
     stdin_port = Integer(0, config=True, help="set the stdin (XREQ) port [default: random]")
     connection_file = Unicode('', config=True, 
     help="""JSON file in which to store connection info [default: kernel-<pid>.json]
     
     This file will contain the IP, ports, and authentication key needed to connect
     clients to this kernel. By default, this file will be created in the security-dir
     of the current profile, but can be specified by absolute path.
     """)
 
     # streams, etc.
     no_stdout = Bool(False, config=True, help="redirect stdout to the null device")
     no_stderr = Bool(False, config=True, help="redirect stderr to the null device")
     outstream_class = DottedObjectName('IPython.zmq.iostream.OutStream',
         config=True, help="The importstring for the OutStream factory")
     displayhook_class = DottedObjectName('IPython.zmq.displayhook.ZMQDisplayHook',
         config=True, help="The importstring for the DisplayHook factory")
 
     # polling
     parent = Integer(0, config=True,
         help="""kill this process if its parent dies.  On Windows, the argument
         specifies the HANDLE of the parent process, otherwise it is simply boolean.
         """)
     interrupt = Integer(0, config=True,
         help="""ONLY USED ON WINDOWS
         Interrupt this process when the parent is signalled.
         """)
 
     def init_crash_handler(self):
         # Install minimal exception handling
         sys.excepthook = FormattedTB(mode='Verbose', color_scheme='NoColor',
                                      ostream=sys.__stdout__)
 
     def init_poller(self):
         if sys.platform == 'win32':
             if self.interrupt or self.parent:
                 self.poller = ParentPollerWindows(self.interrupt, self.parent)
         elif self.parent:
             self.poller = ParentPollerUnix()
 
     def _bind_socket(self, s, port):
         iface = 'tcp://%s' % self.ip
         if port <= 0:
             port = s.bind_to_random_port(iface)
         else:
             s.bind(iface + ':%i'%port)
         return port
 
     def load_connection_file(self):
         """load ip/port/hmac config from JSON connection file"""
         try:
             fname = filefind(self.connection_file, ['.', self.profile_dir.security_dir])
         except IOError:
             self.log.debug("Connection file not found: %s", self.connection_file)
+            # This means I own it, so I will clean it up:
+            atexit.register(self.cleanup_connection_file)
             return
         self.log.debug(u"Loading connection file %s", fname)
         with open(fname) as f:
             s = f.read()
         cfg = json.loads(s)
         if self.ip == LOCALHOST and 'ip' in cfg:
             # not overridden by config or cl_args
             self.ip = cfg['ip']
         for channel in ('hb', 'shell', 'iopub', 'stdin'):
             name = channel + '_port'
             if getattr(self, name) == 0 and name in cfg:
                 # not overridden by config or cl_args
                 setattr(self, name, cfg[name])
         if 'key' in cfg:
             self.config.Session.key = str_to_bytes(cfg['key'])
     
     def write_connection_file(self):
         """write connection info to JSON file"""
         if os.path.basename(self.connection_file) == self.connection_file:
             cf = os.path.join(self.profile_dir.security_dir, self.connection_file)
         else:
             cf = self.connection_file
         write_connection_file(cf, ip=self.ip, key=self.session.key,
         shell_port=self.shell_port, stdin_port=self.stdin_port, hb_port=self.hb_port,
         iopub_port=self.iopub_port)
+        
+        self._full_connection_file = cf
+    
+    def cleanup_connection_file(self):
+        cf = self._full_connection_file
+        self.log.debug("cleaning up connection file: %r", cf)
+        try:
+            os.remove(cf)
+        except (IOError, OSError):
+            pass
     
     def init_connection_file(self):
         if not self.connection_file:
             self.connection_file = "kernel-%s.json"%os.getpid()
         try:
             self.load_connection_file()
         except Exception:
             self.log.error("Failed to load connection file: %r", self.connection_file, exc_info=True)
             self.exit(1)
     
     def init_sockets(self):
         # Create a context, a session, and the kernel sockets.
         self.log.info("Starting the kernel at pid: %i", os.getpid())
         context = zmq.Context.instance()
         # Uncomment this to try closing the context.
         # atexit.register(context.term)
 
         self.shell_socket = context.socket(zmq.ROUTER)
         self.shell_port = self._bind_socket(self.shell_socket, self.shell_port)
         self.log.debug("shell ROUTER Channel on port: %i"%self.shell_port)
 
         self.iopub_socket = context.socket(zmq.PUB)
         self.iopub_port = self._bind_socket(self.iopub_socket, self.iopub_port)
         self.log.debug("iopub PUB Channel on port: %i"%self.iopub_port)
 
         self.stdin_socket = context.socket(zmq.ROUTER)
         self.stdin_port = self._bind_socket(self.stdin_socket, self.stdin_port)
         self.log.debug("stdin ROUTER Channel on port: %i"%self.stdin_port)
-        
+    
+    def init_heartbeat(self):
+        """start the heart beating"""
         # heartbeat doesn't share context, because it mustn't be blocked
         # by the GIL, which is accessed by libzmq when freeing zero-copy messages
         hb_ctx = zmq.Context()
         self.heartbeat = Heartbeat(hb_ctx, (self.ip, self.hb_port))
         self.hb_port = self.heartbeat.port
         self.log.debug("Heartbeat REP Channel on port: %i"%self.hb_port)
+        self.heartbeat.start()
 
         # Helper to make it easier to connect to an existing kernel.
         # set log-level to critical, to make sure it is output
         self.log.critical("To connect another client to this kernel, use:")
-        
+    
+    def log_connection_info(self):
+        """display connection info, and store ports"""
         basename = os.path.basename(self.connection_file)
         if basename == self.connection_file or \
             os.path.dirname(self.connection_file) == self.profile_dir.security_dir:
             # use shortname
             tail = basename
             if self.profile != 'default':
                 tail += " --profile %s" % self.profile
         else:
             tail = self.connection_file
         self.log.critical("--existing %s", tail)
 
 
         self.ports = dict(shell=self.shell_port, iopub=self.iopub_port,
                                 stdin=self.stdin_port, hb=self.hb_port)
 
     def init_session(self):
         """create our session object"""
         default_secure(self.config)
         self.session = Session(config=self.config, username=u'kernel')
 
     def init_blackhole(self):
         """redirects stdout/stderr to devnull if necessary"""
         if self.no_stdout or self.no_stderr:
             blackhole = open(os.devnull, 'w')
             if self.no_stdout:
                 sys.stdout = sys.__stdout__ = blackhole
             if self.no_stderr:
                 sys.stderr = sys.__stderr__ = blackhole
     
     def init_io(self):
         """Redirect input streams and set a display hook."""
         if self.outstream_class:
             outstream_factory = import_item(str(self.outstream_class))
             sys.stdout = outstream_factory(self.session, self.iopub_socket, u'stdout')
             sys.stderr = outstream_factory(self.session, self.iopub_socket, u'stderr')
         if self.displayhook_class:
             displayhook_factory = import_item(str(self.displayhook_class))
             sys.displayhook = displayhook_factory(self.session, self.iopub_socket)
 
+    def init_signal(self):
+        signal.signal(signal.SIGINT, signal.SIG_IGN)
+
     def init_kernel(self):
         """Create the Kernel object itself"""
         kernel_factory = import_item(str(self.kernel_class))
         self.kernel = kernel_factory(config=self.config, session=self.session,
                                 shell_socket=self.shell_socket,
                                 iopub_socket=self.iopub_socket,
                                 stdin_socket=self.stdin_socket,
                                 log=self.log
         )
         self.kernel.record_ports(self.ports)
 
     @catch_config_error
     def initialize(self, argv=None):
         super(KernelApp, self).initialize(argv)
         self.init_blackhole()
         self.init_connection_file()
         self.init_session()
         self.init_poller()
         self.init_sockets()
-        # writing connection file must be *after* init_sockets
+        self.init_heartbeat()
+        # writing/displaying connection info must be *after* init_sockets/heartbeat
+        self.log_connection_info()
         self.write_connection_file()
         self.init_io()
+        self.init_signal()
         self.init_kernel()
         # flush stdout/stderr, so that anything written to these streams during
         # initialization do not get associated with the first execution request
         sys.stdout.flush()
         sys.stderr.flush()
 
     def start(self):
-        self.heartbeat.start()
         if self.poller is not None:
             self.poller.start()
+        self.kernel.start()
         try:
-            self.kernel.start()
+            ioloop.IOLoop.instance().start()
         except KeyboardInterrupt:
             pass
 

IPython/zmq/kernelmanager.py

@@ -1,1000 +1,994 b''
 """Base classes to manage the interaction with a running kernel.
 
 TODO
 * Create logger to handle debugging and console messages.
 """
 
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2008-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 # Standard library imports.
 import atexit
 import errno
 import json
 from subprocess import Popen
 import os
 import signal
 import sys
 from threading import Thread
 import time
 
 # System library imports.
 import zmq
 # import ZMQError in top-level namespace, to avoid ugly attribute-error messages
 # during garbage collection of threads at exit:
 from zmq import ZMQError
 from zmq.eventloop import ioloop, zmqstream
 
 # Local imports.
 from IPython.config.loader import Config
 from IPython.utils.localinterfaces import LOCALHOST, LOCAL_IPS
 from IPython.utils.traitlets import (
     HasTraits, Any, Instance, Type, Unicode, Integer, Bool
 )
 from IPython.utils.py3compat import str_to_bytes
 from IPython.zmq.entry_point import write_connection_file
 from session import Session
 
 #-----------------------------------------------------------------------------
 # Constants and exceptions
 #-----------------------------------------------------------------------------
 
 class InvalidPortNumber(Exception):
     pass
 
 #-----------------------------------------------------------------------------
 # Utility functions
 #-----------------------------------------------------------------------------
 
 # some utilities to validate message structure, these might get moved elsewhere
 # if they prove to have more generic utility
 
 def validate_string_list(lst):
     """Validate that the input is a list of strings.
 
     Raises ValueError if not."""
     if not isinstance(lst, list):
         raise ValueError('input %r must be a list' % lst)
     for x in lst:
         if not isinstance(x, basestring):
             raise ValueError('element %r in list must be a string' % x)
 
 
 def validate_string_dict(dct):
     """Validate that the input is a dict with string keys and values.
 
     Raises ValueError if not."""
     for k,v in dct.iteritems():
         if not isinstance(k, basestring):
             raise ValueError('key %r in dict must be a string' % k)
         if not isinstance(v, basestring):
             raise ValueError('value %r in dict must be a string' % v)
 
 
 #-----------------------------------------------------------------------------
 # ZMQ Socket Channel classes
 #-----------------------------------------------------------------------------
 
 class ZMQSocketChannel(Thread):
     """The base class for the channels that use ZMQ sockets.
     """
     context = None
     session = None
     socket = None
     ioloop = None
     stream = None
     _address = None
     _exiting = False
 
     def __init__(self, context, session, address):
         """Create a channel
 
         Parameters
         ----------
         context : :class:`zmq.Context`
             The ZMQ context to use.
         session : :class:`session.Session`
             The session to use.
         address : tuple
             Standard (ip, port) tuple that the kernel is listening on.
         """
         super(ZMQSocketChannel, self).__init__()
         self.daemon = True
 
         self.context = context
         self.session = session
         if address[1] == 0:
             message = 'The port number for a channel cannot be 0.'
             raise InvalidPortNumber(message)
         self._address = address
         atexit.register(self._notice_exit)
     
     def _notice_exit(self):
         self._exiting = True
 
     def _run_loop(self):
         """Run my loop, ignoring EINTR events in the poller"""
         while True:
             try:
                 self.ioloop.start()
             except ZMQError as e:
                 if e.errno == errno.EINTR:
                     continue
                 else:
                     raise
             except Exception:
                 if self._exiting:
                     break
                 else:
                     raise
             else:
                 break
 
     def stop(self):
         """Stop the channel's activity.
 
         This calls :method:`Thread.join` and returns when the thread
         terminates. :class:`RuntimeError` will be raised if
         :method:`self.start` is called again.
         """
         self.join()
 
     @property
     def address(self):
         """Get the channel's address as an (ip, port) tuple.
 
         By the default, the address is (localhost, 0), where 0 means a random
         port.
         """
         return self._address
 
     def _queue_send(self, msg):
         """Queue a message to be sent from the IOLoop's thread.
         
         Parameters
         ----------
         msg : message to send
         
         This is threadsafe, as it uses IOLoop.add_callback to give the loop's
         thread control of the action.
         """
         def thread_send():
             self.session.send(self.stream, msg)
         self.ioloop.add_callback(thread_send)
 
     def _handle_recv(self, msg):
         """callback for stream.on_recv
         
         unpacks message, and calls handlers with it.
         """
         ident,smsg = self.session.feed_identities(msg)
         self.call_handlers(self.session.unserialize(smsg))
     
 
 
 class ShellSocketChannel(ZMQSocketChannel):
     """The XREQ channel for issues request/replies to the kernel.
     """
 
     command_queue = None
     # flag for whether execute requests should be allowed to call raw_input:
     allow_stdin = True
 
     def __init__(self, context, session, address):
         super(ShellSocketChannel, self).__init__(context, session, address)
         self.ioloop = ioloop.IOLoop()
 
     def run(self):
         """The thread's main activity.  Call start() instead."""
         self.socket = self.context.socket(zmq.DEALER)
         self.socket.setsockopt(zmq.IDENTITY, self.session.bsession)
         self.socket.connect('tcp://%s:%i' % self.address)
         self.stream = zmqstream.ZMQStream(self.socket, self.ioloop)
         self.stream.on_recv(self._handle_recv)
         self._run_loop()
         try:
             self.socket.close()
         except:
             pass
 
     def stop(self):
         self.ioloop.stop()
         super(ShellSocketChannel, self).stop()
 
     def call_handlers(self, msg):
         """This method is called in the ioloop thread when a message arrives.
 
         Subclasses should override this method to handle incoming messages.
         It is important to remember that this method is called in the thread
         so that some logic must be done to ensure that the application leve
         handlers are called in the application thread.
         """
         raise NotImplementedError('call_handlers must be defined in a subclass.')
 
     def execute(self, code, silent=False,
                 user_variables=None, user_expressions=None, allow_stdin=None):
         """Execute code in the kernel.
 
         Parameters
         ----------
         code : str
             A string of Python code.
 
         silent : bool, optional (default False)
             If set, the kernel will execute the code as quietly possible.
 
         user_variables : list, optional
             A list of variable names to pull from the user's namespace.  They
             will come back as a dict with these names as keys and their
             :func:`repr` as values.
 
         user_expressions : dict, optional
             A dict with string keys and  to pull from the user's
             namespace.  They will come back as a dict with these names as keys
             and their :func:`repr` as values.
 
         allow_stdin : bool, optional
             Flag for 
             A dict with string keys and  to pull from the user's
             namespace.  They will come back as a dict with these names as keys
             and their :func:`repr` as values.
 
         Returns
         -------
         The msg_id of the message sent.
         """
         if user_variables is None:
             user_variables = []
         if user_expressions is None:
             user_expressions = {}
         if allow_stdin is None:
             allow_stdin = self.allow_stdin
         
         
         # Don't waste network traffic if inputs are invalid
         if not isinstance(code, basestring):
             raise ValueError('code %r must be a string' % code)
         validate_string_list(user_variables)
         validate_string_dict(user_expressions)
 
         # Create class for content/msg creation. Related to, but possibly
         # not in Session.
         content = dict(code=code, silent=silent,
                        user_variables=user_variables,
                        user_expressions=user_expressions,
                        allow_stdin=allow_stdin,
                        )
         msg = self.session.msg('execute_request', content)
         self._queue_send(msg)
         return msg['header']['msg_id']
 
     def complete(self, text, line, cursor_pos, block=None):
         """Tab complete text in the kernel's namespace.
 
         Parameters
         ----------
         text : str
             The text to complete.
         line : str
             The full line of text that is the surrounding context for the
             text to complete.
         cursor_pos : int
             The position of the cursor in the line where the completion was
             requested.
         block : str, optional
             The full block of code in which the completion is being requested.
 
         Returns
         -------
         The msg_id of the message sent.
         """
         content = dict(text=text, line=line, block=block, cursor_pos=cursor_pos)
         msg = self.session.msg('complete_request', content)
         self._queue_send(msg)
         return msg['header']['msg_id']
 
     def object_info(self, oname, detail_level=0):
         """Get metadata information about an object.
 
         Parameters
         ----------
         oname : str
             A string specifying the object name.
         detail_level : int, optional
             The level of detail for the introspection (0-2)
 
         Returns
         -------
         The msg_id of the message sent.
         """
         content = dict(oname=oname, detail_level=detail_level)
         msg = self.session.msg('object_info_request', content)
         self._queue_send(msg)
         return msg['header']['msg_id']
 
     def history(self, raw=True, output=False, hist_access_type='range', **kwargs):
         """Get entries from the history list.
 
         Parameters
         ----------
         raw : bool
             If True, return the raw input.
         output : bool
             If True, then return the output as well.
         hist_access_type : str
             'range' (fill in session, start and stop params), 'tail' (fill in n)
              or 'search' (fill in pattern param).
 
         session : int
             For a range request, the session from which to get lines. Session
             numbers are positive integers; negative ones count back from the
             current session.
         start : int
             The first line number of a history range.
         stop : int
             The final (excluded) line number of a history range.
 
         n : int
             The number of lines of history to get for a tail request.
 
         pattern : str
             The glob-syntax pattern for a search request.
 
         Returns
         -------
         The msg_id of the message sent.
         """
         content = dict(raw=raw, output=output, hist_access_type=hist_access_type,
                                                                     **kwargs)
         msg = self.session.msg('history_request', content)
         self._queue_send(msg)
         return msg['header']['msg_id']
 
     def shutdown(self, restart=False):
         """Request an immediate kernel shutdown.
 
         Upon receipt of the (empty) reply, client code can safely assume that
         the kernel has shut down and it's safe to forcefully terminate it if
         it's still alive.
 
         The kernel will send the reply via a function registered with Python's
         atexit module, ensuring it's truly done as the kernel is done with all
         normal operation.
         """
         # Send quit message to kernel. Once we implement kernel-side setattr,
         # this should probably be done that way, but for now this will do.
         msg = self.session.msg('shutdown_request', {'restart':restart})
         self._queue_send(msg)
         return msg['header']['msg_id']
 
 
 
 class SubSocketChannel(ZMQSocketChannel):
     """The SUB channel which listens for messages that the kernel publishes.
     """
 
     def __init__(self, context, session, address):
         super(SubSocketChannel, self).__init__(context, session, address)
         self.ioloop = ioloop.IOLoop()
 
     def run(self):
         """The thread's main activity.  Call start() instead."""
         self.socket = self.context.socket(zmq.SUB)
         self.socket.setsockopt(zmq.SUBSCRIBE,b'')
         self.socket.setsockopt(zmq.IDENTITY, self.session.bsession)
         self.socket.connect('tcp://%s:%i' % self.address)
         self.stream = zmqstream.ZMQStream(self.socket, self.ioloop)
         self.stream.on_recv(self._handle_recv)
         self._run_loop()
         try:
             self.socket.close()
         except:
             pass
 
     def stop(self):
         self.ioloop.stop()
         super(SubSocketChannel, self).stop()
 
     def call_handlers(self, msg):
         """This method is called in the ioloop thread when a message arrives.
 
         Subclasses should override this method to handle incoming messages.
         It is important to remember that this method is called in the thread
         so that some logic must be done to ensure that the application leve
         handlers are called in the application thread.
         """
         raise NotImplementedError('call_handlers must be defined in a subclass.')
 
     def flush(self, timeout=1.0):
         """Immediately processes all pending messages on the SUB channel.
 
         Callers should use this method to ensure that :method:`call_handlers`
         has been called for all messages that have been received on the
         0MQ SUB socket of this channel.
 
         This method is thread safe.
 
         Parameters
         ----------
         timeout : float, optional
             The maximum amount of time to spend flushing, in seconds. The
             default is one second.
         """
         # We do the IOLoop callback process twice to ensure that the IOLoop
         # gets to perform at least one full poll.
         stop_time = time.time() + timeout
         for i in xrange(2):
             self._flushed = False
             self.ioloop.add_callback(self._flush)
             while not self._flushed and time.time() < stop_time:
                 time.sleep(0.01)
 
     def _flush(self):
         """Callback for :method:`self.flush`."""
         self.stream.flush()
         self._flushed = True
 
 
 class StdInSocketChannel(ZMQSocketChannel):
     """A reply channel to handle raw_input requests that the kernel makes."""
 
     msg_queue = None
 
     def __init__(self, context, session, address):
         super(StdInSocketChannel, self).__init__(context, session, address)
         self.ioloop = ioloop.IOLoop()
 
     def run(self):
         """The thread's main activity.  Call start() instead."""
         self.socket = self.context.socket(zmq.DEALER)
         self.socket.setsockopt(zmq.IDENTITY, self.session.bsession)
         self.socket.connect('tcp://%s:%i' % self.address)
         self.stream = zmqstream.ZMQStream(self.socket, self.ioloop)
         self.stream.on_recv(self._handle_recv)
         self._run_loop()
         try:
             self.socket.close()
         except:
             pass
         
 
     def stop(self):
         self.ioloop.stop()
         super(StdInSocketChannel, self).stop()
 
     def call_handlers(self, msg):
         """This method is called in the ioloop thread when a message arrives.
 
         Subclasses should override this method to handle incoming messages.
         It is important to remember that this method is called in the thread
         so that some logic must be done to ensure that the application leve
         handlers are called in the application thread.
         """
         raise NotImplementedError('call_handlers must be defined in a subclass.')
 
     def input(self, string):
         """Send a string of raw input to the kernel."""
         content = dict(value=string)
         msg = self.session.msg('input_reply', content)
         self._queue_send(msg)
 
 
 class HBSocketChannel(ZMQSocketChannel):
     """The heartbeat channel which monitors the kernel heartbeat.
 
     Note that the heartbeat channel is paused by default. As long as you start
     this channel, the kernel manager will ensure that it is paused and un-paused
     as appropriate.
     """
 
     time_to_dead = 3.0
     socket = None
     poller = None
     _running = None
     _pause = None
     _beating = None
 
     def __init__(self, context, session, address):
         super(HBSocketChannel, self).__init__(context, session, address)
         self._running = False
         self._pause =True
         self.poller = zmq.Poller()
 
     def _create_socket(self):
         if self.socket is not None:
             # close previous socket, before opening a new one
             self.poller.unregister(self.socket)
             self.socket.close()
         self.socket = self.context.socket(zmq.REQ)
         self.socket.setsockopt(zmq.LINGER, 0)
         self.socket.connect('tcp://%s:%i' % self.address)
         
         self.poller.register(self.socket, zmq.POLLIN)
     
     def _poll(self, start_time):
         """poll for heartbeat replies until we reach self.time_to_dead
         
         Ignores interrupts, and returns the result of poll(), which
         will be an empty list if no messages arrived before the timeout,
         or the event tuple if there is a message to receive.
         """
         
         until_dead = self.time_to_dead - (time.time() - start_time)
         # ensure poll at least once
         until_dead = max(until_dead, 1e-3)
         events = []
         while True:
             try:
                 events = self.poller.poll(1000 * until_dead)
             except ZMQError as e:
                 if e.errno == errno.EINTR:
                     # ignore interrupts during heartbeat
                     # this may never actually happen
                     until_dead = self.time_to_dead - (time.time() - start_time)
                     until_dead = max(until_dead, 1e-3)
                     pass
                 else:
                     raise
             except Exception:
                 if self._exiting:
                     break
                 else:
                     raise
             else:
                 break
         return events
 
     def run(self):
         """The thread's main activity.  Call start() instead."""
         self._create_socket()
         self._running = True
         self._beating = True
         
         while self._running:
             if self._pause:
                 # just sleep, and skip the rest of the loop
                 time.sleep(self.time_to_dead)
                 continue
             
             since_last_heartbeat = 0.0
             # io.rprint('Ping from HB channel') # dbg
             # no need to catch EFSM here, because the previous event was
             # either a recv or connect, which cannot be followed by EFSM
             self.socket.send(b'ping')
             request_time = time.time()
             ready = self._poll(request_time)
             if ready:
                 self._beating = True
                 # the poll above guarantees we have something to recv
                 self.socket.recv()
                 # sleep the remainder of the cycle
                 remainder = self.time_to_dead - (time.time() - request_time)
                 if remainder > 0:
                     time.sleep(remainder)
                 continue
             else:
                 # nothing was received within the time limit, signal heart failure
                 self._beating = False
                 since_last_heartbeat = time.time() - request_time
                 self.call_handlers(since_last_heartbeat)
                 # and close/reopen the socket, because the REQ/REP cycle has been broken
                 self._create_socket()
                 continue
         try:
             self.socket.close()
         except:
             pass
 
     def pause(self):
         """Pause the heartbeat."""
         self._pause = True
 
     def unpause(self):
         """Unpause the heartbeat."""
         self._pause = False
 
     def is_beating(self):
         """Is the heartbeat running and responsive (and not paused)."""
         if self.is_alive() and not self._pause and self._beating:
             return True
         else:
             return False
 
     def stop(self):
         self._running = False
         super(HBSocketChannel, self).stop()
 
     def call_handlers(self, since_last_heartbeat):
         """This method is called in the ioloop thread when a message arrives.
 
         Subclasses should override this method to handle incoming messages.
         It is important to remember that this method is called in the thread
         so that some logic must be done to ensure that the application level
         handlers are called in the application thread.
         """
         raise NotImplementedError('call_handlers must be defined in a subclass.')
 
 
 #-----------------------------------------------------------------------------
 # Main kernel manager class
 #-----------------------------------------------------------------------------
 
 class KernelManager(HasTraits):
     """ Manages a kernel for a frontend.
 
     The SUB channel is for the frontend to receive messages published by the
     kernel.
 
     The REQ channel is for the frontend to make requests of the kernel.
 
     The REP channel is for the kernel to request stdin (raw_input) from the
     frontend.
     """
     # config object for passing to child configurables
     config = Instance(Config)
 
     # The PyZMQ Context to use for communication with the kernel.
     context = Instance(zmq.Context)
     def _context_default(self):
         return zmq.Context.instance()
 
     # The Session to use for communication with the kernel.
     session = Instance(Session)
 
     # The kernel process with which the KernelManager is communicating.
     kernel = Instance(Popen)
 
     # The addresses for the communication channels.
     connection_file = Unicode('')
     ip = Unicode(LOCALHOST)
     def _ip_changed(self, name, old, new):
         if new == '*':
             self.ip = '0.0.0.0'
     shell_port = Integer(0)
     iopub_port = Integer(0)
     stdin_port = Integer(0)
     hb_port = Integer(0)
 
     # The classes to use for the various channels.
     shell_channel_class = Type(ShellSocketChannel)
     sub_channel_class = Type(SubSocketChannel)
     stdin_channel_class = Type(StdInSocketChannel)
     hb_channel_class = Type(HBSocketChannel)
 
     # Protected traits.
     _launch_args = Any
     _shell_channel = Any
     _sub_channel = Any
     _stdin_channel = Any
     _hb_channel = Any
     _connection_file_written=Bool(False)
 
     def __init__(self, **kwargs):
         super(KernelManager, self).__init__(**kwargs)
         if self.session is None:
             self.session = Session(config=self.config)
     
     def __del__(self):
         self.cleanup_connection_file()
     
 
     #--------------------------------------------------------------------------
     # Channel management methods:
     #--------------------------------------------------------------------------
 
     def start_channels(self, shell=True, sub=True, stdin=True, hb=True):
         """Starts the channels for this kernel.
 
         This will create the channels if they do not exist and then start
         them. If port numbers of 0 are being used (random ports) then you
         must first call :method:`start_kernel`. If the channels have been
         stopped and you call this, :class:`RuntimeError` will be raised.
         """
         if shell:
             self.shell_channel.start()
         if sub:
             self.sub_channel.start()
         if stdin:
             self.stdin_channel.start()
             self.shell_channel.allow_stdin = True
         else:
             self.shell_channel.allow_stdin = False
         if hb:
             self.hb_channel.start()
 
     def stop_channels(self):
         """Stops all the running channels for this kernel.
         """
         if self.shell_channel.is_alive():
             self.shell_channel.stop()
         if self.sub_channel.is_alive():
             self.sub_channel.stop()
         if self.stdin_channel.is_alive():
             self.stdin_channel.stop()
         if self.hb_channel.is_alive():
             self.hb_channel.stop()
 
     @property
     def channels_running(self):
         """Are any of the channels created and running?"""
         return (self.shell_channel.is_alive() or self.sub_channel.is_alive() or
                 self.stdin_channel.is_alive() or self.hb_channel.is_alive())
 
     #--------------------------------------------------------------------------
     # Kernel process management methods:
     #--------------------------------------------------------------------------
     
     def cleanup_connection_file(self):
         """cleanup connection file *if we wrote it*
         
         Will not raise if the connection file was already removed somehow.
         """
         if self._connection_file_written:
             # cleanup connection files on full shutdown of kernel we started
             self._connection_file_written = False
             try:
                 os.remove(self.connection_file)
             except OSError:
                 pass
     
     def load_connection_file(self):
         """load connection info from JSON dict in self.connection_file"""
         with open(self.connection_file) as f:
             cfg = json.loads(f.read())
         
         self.ip = cfg['ip']
         self.shell_port = cfg['shell_port']
         self.stdin_port = cfg['stdin_port']
         self.iopub_port = cfg['iopub_port']
         self.hb_port = cfg['hb_port']
         self.session.key = str_to_bytes(cfg['key'])
     
     def write_connection_file(self):
         """write connection info to JSON dict in self.connection_file"""
         if self._connection_file_written:
             return
         self.connection_file,cfg = write_connection_file(self.connection_file,
             ip=self.ip, key=self.session.key,
             stdin_port=self.stdin_port, iopub_port=self.iopub_port,
             shell_port=self.shell_port, hb_port=self.hb_port)
         # write_connection_file also sets default ports:
         self.shell_port = cfg['shell_port']
         self.stdin_port = cfg['stdin_port']
         self.iopub_port = cfg['iopub_port']
         self.hb_port = cfg['hb_port']
         
         self._connection_file_written = True
     
     def start_kernel(self, **kw):
         """Starts a kernel process and configures the manager to use it.
 
         If random ports (port=0) are being used, this method must be called
         before the channels are created.
 
         Parameters:
         -----------
-        ipython : bool, optional (default True)
-             Whether to use an IPython kernel instead of a plain Python kernel.
-
         launcher : callable, optional (default None)
              A custom function for launching the kernel process (generally a
              wrapper around ``entry_point.base_launch_kernel``). In most cases,
              it should not be necessary to use this parameter.
 
         **kw : optional
              See respective options for IPython and Python kernels.
         """
         if self.ip not in LOCAL_IPS:
             raise RuntimeError("Can only launch a kernel on a local interface. "
                                "Make sure that the '*_address' attributes are "
                                "configured properly. "
                                "Currently valid addresses are: %s"%LOCAL_IPS
                                )
         
         # write connection file / get default ports
         self.write_connection_file()
 
         self._launch_args = kw.copy()
         launch_kernel = kw.pop('launcher', None)
         if launch_kernel is None:
-            if kw.pop('ipython', True):
-                from ipkernel import launch_kernel
-            else:
-                from pykernel import launch_kernel
+            from ipkernel import launch_kernel
         self.kernel = launch_kernel(fname=self.connection_file, **kw)
 
     def shutdown_kernel(self, restart=False):
         """ Attempts to the stop the kernel process cleanly. If the kernel
         cannot be stopped, it is killed, if possible.
         """
         # FIXME: Shutdown does not work on Windows due to ZMQ errors!
         if sys.platform == 'win32':
             self.kill_kernel()
             return
 
         # Pause the heart beat channel if it exists.
         if self._hb_channel is not None:
             self._hb_channel.pause()
 
         # Don't send any additional kernel kill messages immediately, to give
         # the kernel a chance to properly execute shutdown actions. Wait for at
         # most 1s, checking every 0.1s.
         self.shell_channel.shutdown(restart=restart)
         for i in range(10):
             if self.is_alive:
                 time.sleep(0.1)
             else:
                 break
         else:
             # OK, we've waited long enough.
             if self.has_kernel:
                 self.kill_kernel()
 
         if not restart and self._connection_file_written:
             # cleanup connection files on full shutdown of kernel we started
             self._connection_file_written = False
             try:
                 os.remove(self.connection_file)
             except IOError:
                 pass
 
     def restart_kernel(self, now=False, **kw):
         """Restarts a kernel with the arguments that were used to launch it.
 
         If the old kernel was launched with random ports, the same ports will be
         used for the new kernel.
 
         Parameters
         ----------
         now : bool, optional
             If True, the kernel is forcefully restarted *immediately*, without
             having a chance to do any cleanup action.  Otherwise the kernel is
             given 1s to clean up before a forceful restart is issued.
 
             In all cases the kernel is restarted, the only difference is whether
             it is given a chance to perform a clean shutdown or not.
 
         **kw : optional
             Any options specified here will replace those used to launch the
             kernel.
         """
         if self._launch_args is None:
             raise RuntimeError("Cannot restart the kernel. "
                                "No previous call to 'start_kernel'.")
         else:
             # Stop currently running kernel.
             if self.has_kernel:
                 if now:
                     self.kill_kernel()
                 else:
                     self.shutdown_kernel(restart=True)
 
             # Start new kernel.
             self._launch_args.update(kw)
             self.start_kernel(**self._launch_args)
 
             # FIXME: Messages get dropped in Windows due to probable ZMQ bug
             # unless there is some delay here.
             if sys.platform == 'win32':
                 time.sleep(0.2)
 
     @property
     def has_kernel(self):
         """Returns whether a kernel process has been specified for the kernel
         manager.
         """
         return self.kernel is not None
 
     def kill_kernel(self):
         """ Kill the running kernel. """
         if self.has_kernel:
             # Pause the heart beat channel if it exists.
             if self._hb_channel is not None:
                 self._hb_channel.pause()
 
             # Attempt to kill the kernel.
             try:
                 self.kernel.kill()
             except OSError, e:
                 # In Windows, we will get an Access Denied error if the process
                 # has already terminated. Ignore it.
                 if sys.platform == 'win32':
                     if e.winerror != 5:
                         raise
                 # On Unix, we may get an ESRCH error if the process has already
                 # terminated. Ignore it.
                 else:
                     from errno import ESRCH
                     if e.errno != ESRCH:
                         raise
             self.kernel = None
         else:
             raise RuntimeError("Cannot kill kernel. No kernel is running!")
 
     def interrupt_kernel(self):
         """ Interrupts the kernel. Unlike ``signal_kernel``, this operation is
         well supported on all platforms.
         """
         if self.has_kernel:
             if sys.platform == 'win32':
                 from parentpoller import ParentPollerWindows as Poller
                 Poller.send_interrupt(self.kernel.win32_interrupt_event)
             else:
                 self.kernel.send_signal(signal.SIGINT)
         else:
             raise RuntimeError("Cannot interrupt kernel. No kernel is running!")
 
     def signal_kernel(self, signum):
         """ Sends a signal to the kernel. Note that since only SIGTERM is
         supported on Windows, this function is only useful on Unix systems.
         """
         if self.has_kernel:
             self.kernel.send_signal(signum)
         else:
             raise RuntimeError("Cannot signal kernel. No kernel is running!")
 
     @property
     def is_alive(self):
         """Is the kernel process still running?"""
         if self.has_kernel:
             if self.kernel.poll() is None:
                 return True
             else:
                 return False
         elif self._hb_channel is not None:
             # We didn't start the kernel with this KernelManager so we
             # use the heartbeat.
             return self._hb_channel.is_beating()
         else:
             # no heartbeat and not local, we can't tell if it's running,
             # so naively return True
             return True
 
     #--------------------------------------------------------------------------
     # Channels used for communication with the kernel:
     #--------------------------------------------------------------------------
 
     @property
     def shell_channel(self):
         """Get the REQ socket channel object to make requests of the kernel."""
         if self._shell_channel is None:
             self._shell_channel = self.shell_channel_class(self.context,
                                                          self.session,
                                                          (self.ip, self.shell_port))
         return self._shell_channel
 
     @property
     def sub_channel(self):
         """Get the SUB socket channel object."""
         if self._sub_channel is None:
             self._sub_channel = self.sub_channel_class(self.context,
                                                        self.session,
                                                        (self.ip, self.iopub_port))
         return self._sub_channel
 
     @property
     def stdin_channel(self):
         """Get the REP socket channel object to handle stdin (raw_input)."""
         if self._stdin_channel is None:
             self._stdin_channel = self.stdin_channel_class(self.context,
                                                        self.session,
                                                        (self.ip, self.stdin_port))
         return self._stdin_channel
 
     @property
     def hb_channel(self):
         """Get the heartbeat socket channel object to check that the
         kernel is alive."""
         if self._hb_channel is None:
             self._hb_channel = self.hb_channel_class(self.context,
                                                        self.session,
                                                        (self.ip, self.hb_port))
         return self._hb_channel

IPython/zmq/session.py

@@ -1,756 +1,756 b''
 """Session object for building, serializing, sending, and receiving messages in
 IPython. The Session object supports serialization, HMAC signatures, and
 metadata on messages.
 
 Also defined here are utilities for working with Sessions:
 * A SessionFactory to be used as a base class for configurables that work with
 Sessions.
 * A Message object for convenience that allows attribute-access to the msg dict.
 
 Authors:
 
 * Min RK
 * Brian Granger
 * Fernando Perez
 """
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2010-2011  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-----------------------------------------------------------------------------
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 import hmac
 import logging
 import os
 import pprint
 import uuid
 from datetime import datetime
 
 try:
     import cPickle
     pickle = cPickle
 except:
     cPickle = None
     import pickle
 
 import zmq
 from zmq.utils import jsonapi
 from zmq.eventloop.ioloop import IOLoop
 from zmq.eventloop.zmqstream import ZMQStream
 
 from IPython.config.application import Application, boolean_flag
 from IPython.config.configurable import Configurable, LoggingConfigurable
 from IPython.utils.importstring import import_item
 from IPython.utils.jsonutil import extract_dates, squash_dates, date_default
 from IPython.utils.py3compat import str_to_bytes
 from IPython.utils.traitlets import (CBytes, Unicode, Bool, Any, Instance, Set,
                                         DottedObjectName, CUnicode)
 
 #-----------------------------------------------------------------------------
 # utility functions
 #-----------------------------------------------------------------------------
 
 def squash_unicode(obj):
     """coerce unicode back to bytestrings."""
     if isinstance(obj,dict):
         for key in obj.keys():
             obj[key] = squash_unicode(obj[key])
             if isinstance(key, unicode):
                 obj[squash_unicode(key)] = obj.pop(key)
     elif isinstance(obj, list):
         for i,v in enumerate(obj):
             obj[i] = squash_unicode(v)
     elif isinstance(obj, unicode):
         obj = obj.encode('utf8')
     return obj
 
 #-----------------------------------------------------------------------------
 # globals and defaults
 #-----------------------------------------------------------------------------
 
 
 # ISO8601-ify datetime objects
 json_packer = lambda obj: jsonapi.dumps(obj, default=date_default)
 json_unpacker = lambda s: extract_dates(jsonapi.loads(s))
 
 pickle_packer = lambda o: pickle.dumps(o,-1)
 pickle_unpacker = pickle.loads
 
 default_packer = json_packer
 default_unpacker = json_unpacker
 
 DELIM=b"<IDS|MSG>"
 
 
 #-----------------------------------------------------------------------------
 # Mixin tools for apps that use Sessions
 #-----------------------------------------------------------------------------
 
 session_aliases = dict(
     ident = 'Session.session',
     user = 'Session.username',
     keyfile = 'Session.keyfile',
 )
 
 session_flags  = {
     'secure' : ({'Session' : { 'key' : str_to_bytes(str(uuid.uuid4())),
                             'keyfile' : '' }},
         """Use HMAC digests for authentication of messages.
         Setting this flag will generate a new UUID to use as the HMAC key.
         """),
     'no-secure' : ({'Session' : { 'key' : b'', 'keyfile' : '' }},
         """Don't authenticate messages."""),
 }
 
 def default_secure(cfg):
     """Set the default behavior for a config environment to be secure.
     
     If Session.key/keyfile have not been set, set Session.key to
     a new random UUID.
     """
     
     if 'Session' in cfg:
         if 'key' in cfg.Session or 'keyfile' in cfg.Session:
             return
     # key/keyfile not specified, generate new UUID:
     cfg.Session.key = str_to_bytes(str(uuid.uuid4()))
 
 
 #-----------------------------------------------------------------------------
 # Classes
 #-----------------------------------------------------------------------------
 
 class SessionFactory(LoggingConfigurable):
     """The Base class for configurables that have a Session, Context, logger,
     and IOLoop.
     """
 
     logname = Unicode('')
     def _logname_changed(self, name, old, new):
         self.log = logging.getLogger(new)
 
     # not configurable:
     context = Instance('zmq.Context')
     def _context_default(self):
         return zmq.Context.instance()
 
     session = Instance('IPython.zmq.session.Session')
 
     loop = Instance('zmq.eventloop.ioloop.IOLoop', allow_none=False)
     def _loop_default(self):
         return IOLoop.instance()
 
     def __init__(self, **kwargs):
         super(SessionFactory, self).__init__(**kwargs)
 
         if self.session is None:
             # construct the session
             self.session = Session(**kwargs)
 
 
 class Message(object):
     """A simple message object that maps dict keys to attributes.
 
     A Message can be created from a dict and a dict from a Message instance
     simply by calling dict(msg_obj)."""
 
     def __init__(self, msg_dict):
         dct = self.__dict__
         for k, v in dict(msg_dict).iteritems():
             if isinstance(v, dict):
                 v = Message(v)
             dct[k] = v
 
     # Having this iterator lets dict(msg_obj) work out of the box.
     def __iter__(self):
         return iter(self.__dict__.iteritems())
 
     def __repr__(self):
         return repr(self.__dict__)
 
     def __str__(self):
         return pprint.pformat(self.__dict__)
 
     def __contains__(self, k):
         return k in self.__dict__
 
     def __getitem__(self, k):
         return self.__dict__[k]
 
 
 def msg_header(msg_id, msg_type, username, session):
     date = datetime.now()
     return locals()
 
 def extract_header(msg_or_header):
     """Given a message or header, return the header."""
     if not msg_or_header:
         return {}
     try:
         # See if msg_or_header is the entire message.
         h = msg_or_header['header']
     except KeyError:
         try:
             # See if msg_or_header is just the header
             h = msg_or_header['msg_id']
         except KeyError:
             raise
         else:
             h = msg_or_header
     if not isinstance(h, dict):
         h = dict(h)
     return h
 
 class Session(Configurable):
     """Object for handling serialization and sending of messages.
 
     The Session object handles building messages and sending them
     with ZMQ sockets or ZMQStream objects.  Objects can communicate with each
     other over the network via Session objects, and only need to work with the
     dict-based IPython message spec. The Session will handle
     serialization/deserialization, security, and metadata.
 
     Sessions support configurable serialiization via packer/unpacker traits,
     and signing with HMAC digests via the key/keyfile traits.
 
     Parameters
     ----------
 
     debug : bool
         whether to trigger extra debugging statements
     packer/unpacker : str : 'json', 'pickle' or import_string
         importstrings for methods to serialize message parts.  If just
         'json' or 'pickle', predefined JSON and pickle packers will be used.
         Otherwise, the entire importstring must be used.
 
         The functions must accept at least valid JSON input, and output *bytes*.
 
         For example, to use msgpack:
         packer = 'msgpack.packb', unpacker='msgpack.unpackb'
     pack/unpack : callables
         You can also set the pack/unpack callables for serialization directly.
     session : bytes
         the ID of this Session object.  The default is to generate a new UUID.
     username : unicode
         username added to message headers.  The default is to ask the OS.
     key : bytes
         The key used to initialize an HMAC signature.  If unset, messages
         will not be signed or checked.
     keyfile : filepath
         The file containing a key.  If this is set, `key` will be initialized
         to the contents of the file.
 
     """
 
     debug=Bool(False, config=True, help="""Debug output in the Session""")
 
     packer = DottedObjectName('json',config=True,
             help="""The name of the packer for serializing messages.
             Should be one of 'json', 'pickle', or an import name
             for a custom callable serializer.""")
     def _packer_changed(self, name, old, new):
         if new.lower() == 'json':
             self.pack = json_packer
             self.unpack = json_unpacker
         elif new.lower() == 'pickle':
             self.pack = pickle_packer
             self.unpack = pickle_unpacker
         else:
             self.pack = import_item(str(new))
 
     unpacker = DottedObjectName('json', config=True,
         help="""The name of the unpacker for unserializing messages.
         Only used with custom functions for `packer`.""")
     def _unpacker_changed(self, name, old, new):
         if new.lower() == 'json':
             self.pack = json_packer
             self.unpack = json_unpacker
         elif new.lower() == 'pickle':
             self.pack = pickle_packer
             self.unpack = pickle_unpacker
         else:
             self.unpack = import_item(str(new))
 
     session = CUnicode(u'', config=True,
         help="""The UUID identifying this session.""")
     def _session_default(self):
         u = unicode(uuid.uuid4())
         self.bsession = u.encode('ascii')
         return u
 
     def _session_changed(self, name, old, new):
         self.bsession = self.session.encode('ascii')
 
     # bsession is the session as bytes
     bsession = CBytes(b'')
 
     username = Unicode(os.environ.get('USER',u'username'), config=True,
         help="""Username for the Session. Default is your system username.""")
 
     # message signature related traits:
     
     key = CBytes(b'', config=True,
         help="""execution key, for extra authentication.""")
     def _key_changed(self, name, old, new):
         if new:
             self.auth = hmac.HMAC(new)
         else:
             self.auth = None
     auth = Instance(hmac.HMAC)
     digest_history = Set()
 
     keyfile = Unicode('', config=True,
         help="""path to file containing execution key.""")
     def _keyfile_changed(self, name, old, new):
         with open(new, 'rb') as f:
             self.key = f.read().strip()
 
     # serialization traits:
     
     pack = Any(default_packer) # the actual packer function
     def _pack_changed(self, name, old, new):
         if not callable(new):
             raise TypeError("packer must be callable, not %s"%type(new))
 
     unpack = Any(default_unpacker) # the actual packer function
     def _unpack_changed(self, name, old, new):
         # unpacker is not checked - it is assumed to be
         if not callable(new):
             raise TypeError("unpacker must be callable, not %s"%type(new))
 
     def __init__(self, **kwargs):
         """create a Session object
 
         Parameters
         ----------
 
         debug : bool
             whether to trigger extra debugging statements
         packer/unpacker : str : 'json', 'pickle' or import_string
             importstrings for methods to serialize message parts.  If just
             'json' or 'pickle', predefined JSON and pickle packers will be used.
             Otherwise, the entire importstring must be used.
 
             The functions must accept at least valid JSON input, and output
             *bytes*.
 
             For example, to use msgpack:
             packer = 'msgpack.packb', unpacker='msgpack.unpackb'
         pack/unpack : callables
             You can also set the pack/unpack callables for serialization
             directly.
         session : unicode (must be ascii)
             the ID of this Session object.  The default is to generate a new
             UUID.
         bsession : bytes
             The session as bytes
         username : unicode
             username added to message headers.  The default is to ask the OS.
         key : bytes
             The key used to initialize an HMAC signature.  If unset, messages
             will not be signed or checked.
         keyfile : filepath
             The file containing a key.  If this is set, `key` will be
             initialized to the contents of the file.
         """
         super(Session, self).__init__(**kwargs)
         self._check_packers()
         self.none = self.pack({})
         # ensure self._session_default() if necessary, so bsession is defined:
         self.session
 
     @property
     def msg_id(self):
         """always return new uuid"""
         return str(uuid.uuid4())
 
     def _check_packers(self):
         """check packers for binary data and datetime support."""
         pack = self.pack
         unpack = self.unpack
 
         # check simple serialization
         msg = dict(a=[1,'hi'])
         try:
             packed = pack(msg)
         except Exception:
             raise ValueError("packer could not serialize a simple message")
 
         # ensure packed message is bytes
         if not isinstance(packed, bytes):
             raise ValueError("message packed to %r, but bytes are required"%type(packed))
 
         # check that unpack is pack's inverse
         try:
             unpacked = unpack(packed)
         except Exception:
             raise ValueError("unpacker could not handle the packer's output")
 
         # check datetime support
         msg = dict(t=datetime.now())
         try:
             unpacked = unpack(pack(msg))
         except Exception:
             self.pack = lambda o: pack(squash_dates(o))
             self.unpack = lambda s: extract_dates(unpack(s))
 
     def msg_header(self, msg_type):
         return msg_header(self.msg_id, msg_type, self.username, self.session)
 
     def msg(self, msg_type, content=None, parent=None, subheader=None, header=None):
         """Return the nested message dict.
 
         This format is different from what is sent over the wire. The
         serialize/unserialize methods converts this nested message dict to the wire
         format, which is a list of message parts.
         """
         msg = {}
         header = self.msg_header(msg_type) if header is None else header
         msg['header'] = header
         msg['msg_id'] = header['msg_id']
         msg['msg_type'] = header['msg_type']
         msg['parent_header'] = {} if parent is None else extract_header(parent)
         msg['content'] = {} if content is None else content
         sub = {} if subheader is None else subheader
         msg['header'].update(sub)
         return msg
 
     def sign(self, msg_list):
         """Sign a message with HMAC digest. If no auth, return b''.
 
         Parameters
         ----------
         msg_list : list
             The [p_header,p_parent,p_content] part of the message list.
         """
         if self.auth is None:
             return b''
         h = self.auth.copy()
         for m in msg_list:
             h.update(m)
         return str_to_bytes(h.hexdigest())
 
     def serialize(self, msg, ident=None):
         """Serialize the message components to bytes.
 
         This is roughly the inverse of unserialize. The serialize/unserialize
         methods work with full message lists, whereas pack/unpack work with
         the individual message parts in the message list.
 
         Parameters
         ----------
         msg : dict or Message
             The nexted message dict as returned by the self.msg method.
 
         Returns
         -------
         msg_list : list
             The list of bytes objects to be sent with the format:
             [ident1,ident2,...,DELIM,HMAC,p_header,p_parent,p_content,
              buffer1,buffer2,...]. In this list, the p_* entities are
             the packed or serialized versions, so if JSON is used, these
             are utf8 encoded JSON strings.
         """
         content = msg.get('content', {})
         if content is None:
             content = self.none
         elif isinstance(content, dict):
             content = self.pack(content)
         elif isinstance(content, bytes):
             # content is already packed, as in a relayed message
             pass
         elif isinstance(content, unicode):
             # should be bytes, but JSON often spits out unicode
             content = content.encode('utf8')
         else:
             raise TypeError("Content incorrect type: %s"%type(content))
 
         real_message = [self.pack(msg['header']),
                         self.pack(msg['parent_header']),
                         content
         ]
 
         to_send = []
 
         if isinstance(ident, list):
             # accept list of idents
             to_send.extend(ident)
         elif ident is not None:
             to_send.append(ident)
         to_send.append(DELIM)
 
         signature = self.sign(real_message)
         to_send.append(signature)
 
         to_send.extend(real_message)
 
         return to_send
 
     def send(self, stream, msg_or_type, content=None, parent=None, ident=None,
              buffers=None, subheader=None, track=False, header=None):
         """Build and send a message via stream or socket.
 
         The message format used by this function internally is as follows:
 
         [ident1,ident2,...,DELIM,HMAC,p_header,p_parent,p_content,
          buffer1,buffer2,...]
 
         The serialize/unserialize methods convert the nested message dict into this
         format.
 
         Parameters
         ----------
 
         stream : zmq.Socket or ZMQStream
             The socket-like object used to send the data.
         msg_or_type : str or Message/dict
             Normally, msg_or_type will be a msg_type unless a message is being
             sent more than once. If a header is supplied, this can be set to
             None and the msg_type will be pulled from the header.
 
         content : dict or None
             The content of the message (ignored if msg_or_type is a message).
         header : dict or None
             The header dict for the message (ignores if msg_to_type is a message).
         parent : Message or dict or None
             The parent or parent header describing the parent of this message
             (ignored if msg_or_type is a message).
         ident : bytes or list of bytes
             The zmq.IDENTITY routing path.
         subheader : dict or None
             Extra header keys for this message's header (ignored if msg_or_type
             is a message).
         buffers : list or None
             The already-serialized buffers to be appended to the message.
         track : bool
             Whether to track.  Only for use with Sockets, because ZMQStream
             objects cannot track messages.
 
         Returns
         -------
         msg : dict
             The constructed message.
         (msg,tracker) : (dict, MessageTracker)
             if track=True, then a 2-tuple will be returned,
             the first element being the constructed
             message, and the second being the MessageTracker
 
         """
 
         if not isinstance(stream, (zmq.Socket, ZMQStream)):
             raise TypeError("stream must be Socket or ZMQStream, not %r"%type(stream))
         elif track and isinstance(stream, ZMQStream):
             raise TypeError("ZMQStream cannot track messages")
 
         if isinstance(msg_or_type, (Message, dict)):
             # We got a Message or message dict, not a msg_type so don't
             # build a new Message.
             msg = msg_or_type
         else:
             msg = self.msg(msg_or_type, content=content, parent=parent,
                            subheader=subheader, header=header)
 
         buffers = [] if buffers is None else buffers
         to_send = self.serialize(msg, ident)
         flag = 0
         if buffers:
             flag = zmq.SNDMORE
             _track = False
         else:
             _track=track
         if track:
             tracker = stream.send_multipart(to_send, flag, copy=False, track=_track)
         else:
             tracker = stream.send_multipart(to_send, flag, copy=False)
         for b in buffers[:-1]:
             stream.send(b, flag, copy=False)
         if buffers:
             if track:
                 tracker = stream.send(buffers[-1], copy=False, track=track)
             else:
                 tracker = stream.send(buffers[-1], copy=False)
 
         # omsg = Message(msg)
         if self.debug:
             pprint.pprint(msg)
             pprint.pprint(to_send)
             pprint.pprint(buffers)
 
         msg['tracker'] = tracker
 
         return msg
 
     def send_raw(self, stream, msg_list, flags=0, copy=True, ident=None):
         """Send a raw message via ident path.
 
         This method is used to send a already serialized message.
 
         Parameters
         ----------
         stream : ZMQStream or Socket
             The ZMQ stream or socket to use for sending the message.
         msg_list : list
             The serialized list of messages to send. This only includes the
             [p_header,p_parent,p_content,buffer1,buffer2,...] portion of
             the message.
         ident : ident or list
             A single ident or a list of idents to use in sending.
         """
         to_send = []
         if isinstance(ident, bytes):
             ident = [ident]
         if ident is not None:
             to_send.extend(ident)
 
         to_send.append(DELIM)
         to_send.append(self.sign(msg_list))
         to_send.extend(msg_list)
         stream.send_multipart(msg_list, flags, copy=copy)
 
     def recv(self, socket, mode=zmq.NOBLOCK, content=True, copy=True):
         """Receive and unpack a message.
 
         Parameters
         ----------
         socket : ZMQStream or Socket
             The socket or stream to use in receiving.
 
         Returns
         -------
         [idents], msg
             [idents] is a list of idents and msg is a nested message dict of
             same format as self.msg returns.
         """
         if isinstance(socket, ZMQStream):
             socket = socket.socket
         try:
-            msg_list = socket.recv_multipart(mode)
+            msg_list = socket.recv_multipart(mode, copy=copy)
         except zmq.ZMQError as e:
             if e.errno == zmq.EAGAIN:
                 # We can convert EAGAIN to None as we know in this case
                 # recv_multipart won't return None.
                 return None,None
             else:
                 raise
         # split multipart message into identity list and message dict
         # invalid large messages can cause very expensive string comparisons
         idents, msg_list = self.feed_identities(msg_list, copy)
         try:
             return idents, self.unserialize(msg_list, content=content, copy=copy)
         except Exception as e:
             # TODO: handle it
             raise e
 
     def feed_identities(self, msg_list, copy=True):
         """Split the identities from the rest of the message.
 
         Feed until DELIM is reached, then return the prefix as idents and
         remainder as msg_list. This is easily broken by setting an IDENT to DELIM,
         but that would be silly.
 
         Parameters
         ----------
         msg_list : a list of Message or bytes objects
             The message to be split.
         copy : bool
             flag determining whether the arguments are bytes or Messages
 
         Returns
         -------
         (idents, msg_list) : two lists
             idents will always be a list of bytes, each of which is a ZMQ
             identity. msg_list will be a list of bytes or zmq.Messages of the
             form [HMAC,p_header,p_parent,p_content,buffer1,buffer2,...] and
             should be unpackable/unserializable via self.unserialize at this
             point.
         """
         if copy:
             idx = msg_list.index(DELIM)
             return msg_list[:idx], msg_list[idx+1:]
         else:
             failed = True
             for idx,m in enumerate(msg_list):
                 if m.bytes == DELIM:
                     failed = False
                     break
             if failed:
                 raise ValueError("DELIM not in msg_list")
             idents, msg_list = msg_list[:idx], msg_list[idx+1:]
             return [m.bytes for m in idents], msg_list
 
     def unserialize(self, msg_list, content=True, copy=True):
         """Unserialize a msg_list to a nested message dict.
 
         This is roughly the inverse of serialize. The serialize/unserialize
         methods work with full message lists, whereas pack/unpack work with
         the individual message parts in the message list.
 
         Parameters:
         -----------
         msg_list : list of bytes or Message objects
             The list of message parts of the form [HMAC,p_header,p_parent,
             p_content,buffer1,buffer2,...].
         content : bool (True)
             Whether to unpack the content dict (True), or leave it packed
             (False).
         copy : bool (True)
             Whether to return the bytes (True), or the non-copying Message
             object in each place (False).
 
         Returns
         -------
         msg : dict
             The nested message dict with top-level keys [header, parent_header,
             content, buffers].
         """
         minlen = 4
         message = {}
         if not copy:
             for i in range(minlen):
                 msg_list[i] = msg_list[i].bytes
         if self.auth is not None:
             signature = msg_list[0]
             if not signature:
                 raise ValueError("Unsigned Message")
             if signature in self.digest_history:
                 raise ValueError("Duplicate Signature: %r"%signature)
             self.digest_history.add(signature)
             check = self.sign(msg_list[1:4])
             if not signature == check:
                 raise ValueError("Invalid Signature: %r"%signature)
         if not len(msg_list) >= minlen:
             raise TypeError("malformed message, must have at least %i elements"%minlen)
         header = self.unpack(msg_list[1])
         message['header'] = header
         message['msg_id'] = header['msg_id']
         message['msg_type'] = header['msg_type']
         message['parent_header'] = self.unpack(msg_list[2])
         if content:
             message['content'] = self.unpack(msg_list[3])
         else:
             message['content'] = msg_list[3]
 
         message['buffers'] = msg_list[4:]
         return message
 
 def test_msg2obj():
     am = dict(x=1)
     ao = Message(am)
     assert ao.x == am['x']
 
     am['y'] = dict(z=1)
     ao = Message(am)
     assert ao.y.z == am['y']['z']
 
     k1, k2 = 'y', 'z'
     assert ao[k1][k2] == am[k1][k2]
 
     am2 = dict(ao)
     assert am['x'] == am2['x']
     assert am['y']['z'] == am2['y']['z']
 

IPython/zmq/tests/test_embed_kernel.py

@@ -1,159 +1,191 b''
 """test IPython.embed_kernel()"""
 
 #-------------------------------------------------------------------------------
 #  Copyright (C) 2012  The IPython Development Team
 #
 #  Distributed under the terms of the BSD License.  The full license is in
 #  the file COPYING, distributed as part of this software.
 #-------------------------------------------------------------------------------
 
 #-------------------------------------------------------------------------------
 # Imports
 #-------------------------------------------------------------------------------
 
 import os
 import shutil
 import sys
 import tempfile
 import time
 
 from contextlib import contextmanager
 from subprocess import Popen, PIPE
 
 import nose.tools as nt
 
 from IPython.zmq.blockingkernelmanager import BlockingKernelManager
-from IPython.utils import path
+from IPython.utils import path, py3compat
 
 
 #-------------------------------------------------------------------------------
 # Tests
 #-------------------------------------------------------------------------------
 
 def setup():
     """setup temporary IPYTHONDIR for tests"""
     global IPYTHONDIR
     global env
     global save_get_ipython_dir
     
     IPYTHONDIR = tempfile.mkdtemp()
     env = dict(IPYTHONDIR=IPYTHONDIR)
     save_get_ipython_dir = path.get_ipython_dir
     path.get_ipython_dir = lambda : IPYTHONDIR
 
 
 def teardown():
     path.get_ipython_dir = save_get_ipython_dir
     
     try:
         shutil.rmtree(IPYTHONDIR)
     except (OSError, IOError):
         # no such file
         pass
 
 
 @contextmanager
 def setup_kernel(cmd):
     """start an embedded kernel in a subprocess, and wait for it to be ready
     
     Returns
     -------
     kernel_manager: connected KernelManager instance
     """
     kernel = Popen([sys.executable, '-c', cmd], stdout=PIPE, stderr=PIPE, env=env)
     connection_file = os.path.join(IPYTHONDIR,
                                     'profile_default',
                                     'security',
                                     'kernel-%i.json' % kernel.pid
     )
     # wait for connection file to exist, timeout after 5s
     tic = time.time()
-    while not os.path.exists(connection_file) and kernel.poll() is None and time.time() < tic + 5:
+    while not os.path.exists(connection_file) and kernel.poll() is None and time.time() < tic + 10:
         time.sleep(0.1)
     
+    if kernel.poll() is not None:
+        o,e = kernel.communicate()
+        e = py3compat.cast_unicode(e)
+        raise IOError("Kernel failed to start:\n%s" % e)
+    
     if not os.path.exists(connection_file):
         if kernel.poll() is None:
             kernel.terminate()
         raise IOError("Connection file %r never arrived" % connection_file)
     
-    if kernel.poll() is not None:
-        raise IOError("Kernel failed to start")
-    
     km = BlockingKernelManager(connection_file=connection_file)
     km.load_connection_file()
     km.start_channels()
     
     try:
         yield km
     finally:
         km.stop_channels()
         kernel.terminate()
 
 def test_embed_kernel_basic():
     """IPython.embed_kernel() is basically functional"""
     cmd = '\n'.join([
         'from IPython import embed_kernel',
         'def go():',
         '    a=5',
         '    b="hi there"',
         '    embed_kernel()',
         'go()',
         '',
     ])
     
     with setup_kernel(cmd) as km:
         shell = km.shell_channel
     
         # oinfo a (int)
         msg_id = shell.object_info('a')
         msg = shell.get_msg(block=True, timeout=2)
         content = msg['content']
         nt.assert_true(content['found'])
     
         msg_id = shell.execute("c=a*2")
         msg = shell.get_msg(block=True, timeout=2)
         content = msg['content']
         nt.assert_equals(content['status'], u'ok')
 
         # oinfo c (should be 10)
         msg_id = shell.object_info('c')
         msg = shell.get_msg(block=True, timeout=2)
         content = msg['content']
         nt.assert_true(content['found'])
         nt.assert_equals(content['string_form'], u'10')
 
 def test_embed_kernel_namespace():
     """IPython.embed_kernel() inherits calling namespace"""
     cmd = '\n'.join([
         'from IPython import embed_kernel',
         'def go():',
         '    a=5',
         '    b="hi there"',
         '    embed_kernel()',
         'go()',
         '',
     ])
     
     with setup_kernel(cmd) as km:
         shell = km.shell_channel
     
         # oinfo a (int)
         msg_id = shell.object_info('a')
         msg = shell.get_msg(block=True, timeout=2)
         content = msg['content']
         nt.assert_true(content['found'])
         nt.assert_equals(content['string_form'], u'5')
 
         # oinfo b (str)
         msg_id = shell.object_info('b')
         msg = shell.get_msg(block=True, timeout=2)
         content = msg['content']
         nt.assert_true(content['found'])
         nt.assert_equals(content['string_form'], u'hi there')
 
         # oinfo c (undefined)
         msg_id = shell.object_info('c')
         msg = shell.get_msg(block=True, timeout=2)
         content = msg['content']
         nt.assert_false(content['found'])
 
+def test_embed_kernel_reentrant():
+    """IPython.embed_kernel() can be called multiple times"""
+    cmd = '\n'.join([
+        'from IPython import embed_kernel',
+        'count = 0',
+        'def go():',
+        '    global count',
+        '    embed_kernel()',
+        '    count = count + 1',
+        '',
+        'while True:'
+        '    go()',
+        '',
+    ])
+    
+    with setup_kernel(cmd) as km:
+        shell = km.shell_channel
+        for i in range(5):
+            msg_id = shell.object_info('count')
+            msg = shell.get_msg(block=True, timeout=2)
+            content = msg['content']
+            nt.assert_true(content['found'])
+            nt.assert_equals(content['string_form'], unicode(i))
+            
+            # exit from embed_kernel
+            shell.execute("get_ipython().exit_now = True")
+            msg = shell.get_msg(block=True, timeout=2)
+            time.sleep(0.2)
+
+

IPython/zmq/zmqshell.py

@@ -1,525 +1,541 b''
 """A ZMQ-based subclass of InteractiveShell.
 
 This code is meant to ease the refactoring of the base InteractiveShell into
 something with a cleaner architecture for 2-process use, without actually
 breaking InteractiveShell itself.  So we're doing something a bit ugly, where
 we subclass and override what we want to fix.  Once this is working well, we
 can go back to the base class and refactor the code for a cleaner inheritance
 implementation that doesn't rely on so much monkeypatching.
 
 But this lets us maintain a fully working IPython as we develop the new
 machinery.  This should thus be thought of as scaffolding.
 """
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 from __future__ import print_function
 
 # Stdlib
 import inspect
 import os
 import sys
+import time
 from subprocess import Popen, PIPE
 
+# System library imports
+from zmq.eventloop import ioloop
+
 # Our own
 from IPython.core.interactiveshell import (
     InteractiveShell, InteractiveShellABC
 )
 from IPython.core import page, pylabtools
 from IPython.core.autocall import ZMQExitAutocall
 from IPython.core.displaypub import DisplayPublisher
 from IPython.core.macro import Macro
 from IPython.core.magic import MacroToEdit
 from IPython.core.payloadpage import install_payload_page
 from IPython.lib.kernel import (
     get_connection_file, get_connection_info, connect_qtconsole
 )
 from IPython.testing.skipdoctest import skip_doctest
 from IPython.utils import io
 from IPython.utils.jsonutil import json_clean
 from IPython.utils.path import get_py_filename
 from IPython.utils.process import arg_split
-from IPython.utils.traitlets import Instance, Type, Dict, CBool
+from IPython.utils.traitlets import Instance, Type, Dict, CBool, CBytes
 from IPython.utils.warn import warn, error
 from IPython.zmq.displayhook import ZMQShellDisplayHook, _encode_binary
 from IPython.zmq.session import extract_header
 from session import Session
 
 
 #-----------------------------------------------------------------------------
 # Functions and classes
 #-----------------------------------------------------------------------------
 
 class ZMQDisplayPublisher(DisplayPublisher):
     """A display publisher that publishes data using a ZeroMQ PUB socket."""
 
     session = Instance(Session)
     pub_socket = Instance('zmq.Socket')
     parent_header = Dict({})
+    topic = CBytes(b'displaypub')
 
     def set_parent(self, parent):
         """Set the parent for outbound messages."""
         self.parent_header = extract_header(parent)
     
     def _flush_streams(self):
         """flush IO Streams prior to display"""
         sys.stdout.flush()
         sys.stderr.flush()
 
     def publish(self, source, data, metadata=None):
         self._flush_streams()
         if metadata is None:
             metadata = {}
         self._validate_data(source, data, metadata)
         content = {}
         content['source'] = source
         _encode_binary(data)
         content['data'] = data
         content['metadata'] = metadata
         self.session.send(
             self.pub_socket, u'display_data', json_clean(content),
-            parent=self.parent_header
+            parent=self.parent_header, ident=self.topic,
         )
 
     def clear_output(self, stdout=True, stderr=True, other=True):
         content = dict(stdout=stdout, stderr=stderr, other=other)
         
         if stdout:
             print('\r', file=sys.stdout, end='')
         if stderr:
             print('\r', file=sys.stderr, end='')
         
         self._flush_streams()
         
         self.session.send(
             self.pub_socket, u'clear_output', content,
-            parent=self.parent_header
+            parent=self.parent_header, ident=self.topic,
         )
 
 class ZMQInteractiveShell(InteractiveShell):
     """A subclass of InteractiveShell for ZMQ."""
 
     displayhook_class = Type(ZMQShellDisplayHook)
     display_pub_class = Type(ZMQDisplayPublisher)
 
     # Override the traitlet in the parent class, because there's no point using
     # readline for the kernel. Can be removed when the readline code is moved
     # to the terminal frontend.
     colors_force = CBool(True)
     readline_use = CBool(False)
     # autoindent has no meaning in a zmqshell, and attempting to enable it
     # will print a warning in the absence of readline.
     autoindent = CBool(False)
 
     exiter = Instance(ZMQExitAutocall)
     def _exiter_default(self):
         return ZMQExitAutocall(self)
+    
+    def _exit_now_changed(self, name, old, new):
+        """stop eventloop when exit_now fires"""
+        if new:
+            loop = ioloop.IOLoop.instance()
+            loop.add_timeout(time.time()+0.1, loop.stop)
 
     keepkernel_on_exit = None
 
     # Over ZeroMQ, GUI control isn't done with PyOS_InputHook as there is no
     # interactive input being read; we provide event loop support in ipkernel
     from .eventloops import enable_gui
     enable_gui = staticmethod(enable_gui)
 
     def init_environment(self):
         """Configure the user's environment.
 
         """
         env = os.environ
         # These two ensure 'ls' produces nice coloring on BSD-derived systems
         env['TERM'] = 'xterm-color'
         env['CLICOLOR'] = '1'
         # Since normal pagers don't work at all (over pexpect we don't have
         # single-key control of the subprocess), try to disable paging in
         # subprocesses as much as possible.
         env['PAGER'] = 'cat'
         env['GIT_PAGER'] = 'cat'
         
         # And install the payload version of page.
         install_payload_page()
 
     def auto_rewrite_input(self, cmd):
         """Called to show the auto-rewritten input for autocall and friends.
 
         FIXME: this payload is currently not correctly processed by the
         frontend.
         """
         new = self.prompt_manager.render('rewrite') + cmd
         payload = dict(
             source='IPython.zmq.zmqshell.ZMQInteractiveShell.auto_rewrite_input',
             transformed_input=new,
             )
         self.payload_manager.write_payload(payload)
 
     def ask_exit(self):
         """Engage the exit actions."""
+        self.exit_now = True
         payload = dict(
             source='IPython.zmq.zmqshell.ZMQInteractiveShell.ask_exit',
             exit=True,
             keepkernel=self.keepkernel_on_exit,
             )
         self.payload_manager.write_payload(payload)
 
     def _showtraceback(self, etype, evalue, stb):
 
         exc_content = {
             u'traceback' : stb,
             u'ename' : unicode(etype.__name__),
             u'evalue' : unicode(evalue)
         }
 
         dh = self.displayhook
         # Send exception info over pub socket for other clients than the caller
         # to pick up
-        exc_msg = dh.session.send(dh.pub_socket, u'pyerr', json_clean(exc_content), dh.parent_header)
+        topic = None
+        if dh.topic:
+            topic = dh.topic.replace(b'pyout', b'pyerr')
+        
+        exc_msg = dh.session.send(dh.pub_socket, u'pyerr', json_clean(exc_content), dh.parent_header, ident=topic)
 
         # FIXME - Hack: store exception info in shell object.  Right now, the
         # caller is reading this info after the fact, we need to fix this logic
         # to remove this hack.  Even uglier, we need to store the error status
         # here, because in the main loop, the logic that sets it is being
         # skipped because runlines swallows the exceptions.
         exc_content[u'status'] = u'error'
         self._reply_content = exc_content
         # /FIXME
 
         return exc_content
 
     #------------------------------------------------------------------------
     # Magic overrides
     #------------------------------------------------------------------------
     # Once the base class stops inheriting from magic, this code needs to be
     # moved into a separate machinery as well.  For now, at least isolate here
     # the magics which this class needs to implement differently from the base
     # class, or that are unique to it.
 
     def magic_doctest_mode(self,parameter_s=''):
         """Toggle doctest mode on and off.
 
         This mode is intended to make IPython behave as much as possible like a
         plain Python shell, from the perspective of how its prompts, exceptions
         and output look.  This makes it easy to copy and paste parts of a
         session into doctests.  It does so by:
 
         - Changing the prompts to the classic ``>>>`` ones.
         - Changing the exception reporting mode to 'Plain'.
         - Disabling pretty-printing of output.
 
         Note that IPython also supports the pasting of code snippets that have
         leading '>>>' and '...' prompts in them.  This means that you can paste
         doctests from files or docstrings (even if they have leading
         whitespace), and the code will execute correctly.  You can then use
         '%history -t' to see the translated history; this will give you the
         input after removal of all the leading prompts and whitespace, which
         can be pasted back into an editor.
 
         With these features, you can switch into this mode easily whenever you
         need to do testing and changes to doctests, without having to leave
         your existing IPython session.
         """
 
         from IPython.utils.ipstruct import Struct
 
         # Shorthands
         shell = self.shell
         disp_formatter = self.shell.display_formatter
         ptformatter = disp_formatter.formatters['text/plain']
         # dstore is a data store kept in the instance metadata bag to track any
         # changes we make, so we can undo them later.
         dstore = shell.meta.setdefault('doctest_mode', Struct())
         save_dstore = dstore.setdefault
 
         # save a few values we'll need to recover later
         mode = save_dstore('mode', False)
         save_dstore('rc_pprint', ptformatter.pprint)
         save_dstore('rc_plain_text_only',disp_formatter.plain_text_only)
         save_dstore('xmode', shell.InteractiveTB.mode)
 
         if mode == False:
             # turn on
             ptformatter.pprint = False
             disp_formatter.plain_text_only = True
             shell.magic_xmode('Plain')
         else:
             # turn off
             ptformatter.pprint = dstore.rc_pprint
             disp_formatter.plain_text_only = dstore.rc_plain_text_only
             shell.magic_xmode(dstore.xmode)
 
         # Store new mode and inform on console
         dstore.mode = bool(1-int(mode))
         mode_label = ['OFF','ON'][dstore.mode]
         print('Doctest mode is:', mode_label)
 
         # Send the payload back so that clients can modify their prompt display
         payload = dict(
             source='IPython.zmq.zmqshell.ZMQInteractiveShell.magic_doctest_mode',
             mode=dstore.mode)
         self.payload_manager.write_payload(payload)
 
     @skip_doctest
     def magic_edit(self,parameter_s='',last_call=['','']):
         """Bring up an editor and execute the resulting code.
 
         Usage:
           %edit [options] [args]
 
         %edit runs an external text editor. You will need to set the command for
         this editor via the ``TerminalInteractiveShell.editor`` option in your
         configuration file before it will work.
 
         This command allows you to conveniently edit multi-line code right in
         your IPython session.
 
         If called without arguments, %edit opens up an empty editor with a
         temporary file and will execute the contents of this file when you
         close it (don't forget to save it!).
 
 
         Options:
 
         -n <number>: open the editor at a specified line number.  By default,
         the IPython editor hook uses the unix syntax 'editor +N filename', but
         you can configure this by providing your own modified hook if your
         favorite editor supports line-number specifications with a different
         syntax.
 
         -p: this will call the editor with the same data as the previous time
         it was used, regardless of how long ago (in your current session) it
         was.
 
         -r: use 'raw' input.  This option only applies to input taken from the
         user's history.  By default, the 'processed' history is used, so that
         magics are loaded in their transformed version to valid Python.  If
         this option is given, the raw input as typed as the command line is
         used instead.  When you exit the editor, it will be executed by
         IPython's own processor.
 
         -x: do not execute the edited code immediately upon exit. This is
         mainly useful if you are editing programs which need to be called with
         command line arguments, which you can then do using %run.
 
 
         Arguments:
 
         If arguments are given, the following possibilites exist:
 
         - The arguments are numbers or pairs of colon-separated numbers (like
         1 4:8 9). These are interpreted as lines of previous input to be
         loaded into the editor. The syntax is the same of the %macro command.
 
         - If the argument doesn't start with a number, it is evaluated as a
         variable and its contents loaded into the editor. You can thus edit
         any string which contains python code (including the result of
         previous edits).
 
         - If the argument is the name of an object (other than a string),
         IPython will try to locate the file where it was defined and open the
         editor at the point where it is defined. You can use `%edit function`
         to load an editor exactly at the point where 'function' is defined,
         edit it and have the file be executed automatically.
 
         If the object is a macro (see %macro for details), this opens up your
         specified editor with a temporary file containing the macro's data.
         Upon exit, the macro is reloaded with the contents of the file.
 
         Note: opening at an exact line is only supported under Unix, and some
         editors (like kedit and gedit up to Gnome 2.8) do not understand the
         '+NUMBER' parameter necessary for this feature. Good editors like
         (X)Emacs, vi, jed, pico and joe all do.
 
         - If the argument is not found as a variable, IPython will look for a
         file with that name (adding .py if necessary) and load it into the
         editor. It will execute its contents with execfile() when you exit,
         loading any code in the file into your interactive namespace.
 
         After executing your code, %edit will return as output the code you
         typed in the editor (except when it was an existing file). This way
         you can reload the code in further invocations of %edit as a variable,
         via _<NUMBER> or Out[<NUMBER>], where <NUMBER> is the prompt number of
         the output.
 
         Note that %edit is also available through the alias %ed.
 
         This is an example of creating a simple function inside the editor and
         then modifying it. First, start up the editor:
 
         In [1]: ed
         Editing... done. Executing edited code...
         Out[1]: 'def foo():n    print "foo() was defined in an editing session"n'
 
         We can then call the function foo():
 
         In [2]: foo()
         foo() was defined in an editing session
 
         Now we edit foo.  IPython automatically loads the editor with the
         (temporary) file where foo() was previously defined:
 
         In [3]: ed foo
         Editing... done. Executing edited code...
 
         And if we call foo() again we get the modified version:
 
         In [4]: foo()
         foo() has now been changed!
 
         Here is an example of how to edit a code snippet successive
         times. First we call the editor:
 
         In [5]: ed
         Editing... done. Executing edited code...
         hello
         Out[5]: "print 'hello'n"
 
         Now we call it again with the previous output (stored in _):
 
         In [6]: ed _
         Editing... done. Executing edited code...
         hello world
         Out[6]: "print 'hello world'n"
 
         Now we call it with the output #8 (stored in _8, also as Out[8]):
 
         In [7]: ed _8
         Editing... done. Executing edited code...
         hello again
         Out[7]: "print 'hello again'n"
         """
 
         opts,args = self.parse_options(parameter_s,'prn:')
 
         try:
             filename, lineno, _ = self._find_edit_target(args, opts, last_call)
         except MacroToEdit as e:
             # TODO: Implement macro editing over 2 processes.
             print("Macro editing not yet implemented in 2-process model.")
             return
 
         # Make sure we send to the client an absolute path, in case the working
         # directory of client and kernel don't match
         filename = os.path.abspath(filename)
 
         payload = {
             'source' : 'IPython.zmq.zmqshell.ZMQInteractiveShell.edit_magic',
             'filename' : filename,
             'line_number' : lineno
         }
         self.payload_manager.write_payload(payload)
 
     # A few magics that are adapted to the specifics of using pexpect and a
     # remote terminal
 
     def magic_clear(self, arg_s):
         """Clear the terminal."""
         if os.name == 'posix':
             self.shell.system("clear")
         else:
             self.shell.system("cls")
 
     if os.name == 'nt':
         # This is the usual name in windows
         magic_cls = magic_clear
 
     # Terminal pagers won't work over pexpect, but we do have our own pager
 
     def magic_less(self, arg_s):
         """Show a file through the pager.
 
         Files ending in .py are syntax-highlighted."""
         cont = open(arg_s).read()
         if arg_s.endswith('.py'):
             cont = self.shell.pycolorize(cont)
         page.page(cont)
 
     magic_more = magic_less
 
     # Man calls a pager, so we also need to redefine it
     if os.name == 'posix':
         def magic_man(self, arg_s):
             """Find the man page for the given command and display in pager."""
             page.page(self.shell.getoutput('man %s | col -b' % arg_s,
                                            split=False))
 
     # FIXME: this is specific to the GUI, so we should let the gui app load
     # magics at startup that are only for the gui.  Once the gui app has proper
     # profile and configuration management, we can have it initialize a kernel
     # with a special config file that provides these.
     def magic_guiref(self, arg_s):
         """Show a basic reference about the GUI console."""
         from IPython.core.usage import gui_reference
         page.page(gui_reference, auto_html=True)
     
     def magic_connect_info(self, arg_s):
         """Print information for connecting other clients to this kernel
         
         It will print the contents of this session's connection file, as well as
         shortcuts for local clients.
         
         In the simplest case, when called from the most recently launched kernel,
         secondary clients can be connected, simply with:
         
         $> ipython <app> --existing
         
         """
         
         from IPython.core.application import BaseIPythonApplication as BaseIPApp
         
         if BaseIPApp.initialized():
             app = BaseIPApp.instance()
             security_dir = app.profile_dir.security_dir
             profile = app.profile
         else:
             profile = 'default'
             security_dir = ''
         
         try:
             connection_file = get_connection_file()
             info = get_connection_info(unpack=False)
         except Exception as e:
             error("Could not get connection info: %r" % e)
             return
         
         # add profile flag for non-default profile
         profile_flag = "--profile %s" % profile if profile != 'default' else ""
         
         # if it's in the security dir, truncate to basename
         if security_dir == os.path.dirname(connection_file):
             connection_file = os.path.basename(connection_file)
         
         
         print (info + '\n')
         print ("Paste the above JSON into a file, and connect with:\n"
             "    $> ipython <app> --existing <file>\n"
             "or, if you are local, you can connect with just:\n"
             "    $> ipython <app> --existing {0} {1}\n"
             "or even just:\n"
             "    $> ipython <app> --existing {1}\n"
             "if this is the most recent IPython session you have started.".format(
             connection_file, profile_flag
             )
         )
 
     def magic_qtconsole(self, arg_s):
         """Open a qtconsole connected to this kernel.
         
         Useful for connecting a qtconsole to running notebooks, for better
         debugging.
         """
         try:
             p = connect_qtconsole(argv=arg_split(arg_s, os.name=='posix'))
         except Exception as e:
             error("Could not start qtconsole: %r" % e)
             return
 
     def set_next_input(self, text):
         """Send the specified text to the frontend to be presented at the next
         input cell."""
         payload = dict(
             source='IPython.zmq.zmqshell.ZMQInteractiveShell.set_next_input',
             text=text
         )
         self.payload_manager.write_payload(payload)
 
 
 InteractiveShellABC.register(ZMQInteractiveShell)

docs/source/parallel/parallel_db.txt

@@ -1,137 +1,137 b''
 .. _parallel_db:
 
 =======================
 IPython's Task Database
 =======================
 
 The IPython Hub stores all task requests and results in a database. Currently supported backends
 are: MongoDB, SQLite (the default), and an in-memory DictDB.  The most common use case for
 this is clients requesting results for tasks they did not submit, via:
 
 .. sourcecode:: ipython
 
     In [1]: rc.get_result(task_id)
 
 However, since we have this DB backend, we provide a direct query method in the :class:`client`
 for users who want deeper introspection into their task history. The :meth:`db_query` method of
 the Client is modeled after MongoDB queries, so if you have used MongoDB it should look
 familiar.  In fact, when the MongoDB backend is in use, the query is relayed directly.  However,
 when using other backends, the interface is emulated and only a subset of queries is possible.
 
 .. seealso::
 
     MongoDB query docs: http://www.mongodb.org/display/DOCS/Querying
 
 :meth:`Client.db_query` takes a dictionary query object, with keys from the TaskRecord key list,
 and values of either exact values to test, or MongoDB queries, which are dicts of The form:
 ``{'operator' : 'argument(s)'}``. There is also an optional `keys` argument, that specifies
 which subset of keys should be retrieved. The default is to retrieve all keys excluding the
 request and result buffers. :meth:`db_query` returns a list of TaskRecord dicts. Also like
 MongoDB, the `msg_id` key will always be included, whether requested or not.
 
 TaskRecord keys:
 
 =============== =============== =============
 Key             Type            Description
 =============== =============== =============
-msg_id          uuid(bytes)     The msg ID
+msg_id          uuid(ascii)     The msg ID
 header          dict            The request header
 content         dict            The request content (likely empty)
 buffers         list(bytes)     buffers containing serialized request objects
 submitted       datetime        timestamp for time of submission (set by client)
 client_uuid     uuid(bytes)     IDENT of client's socket
 engine_uuid     uuid(bytes)     IDENT of engine's socket
 started         datetime        time task began execution on engine
 completed       datetime        time task finished execution (success or failure) on engine
-resubmitted     datetime        time of resubmission (if applicable)
+resubmitted     uuid(ascii)     msg_id of resubmitted task (if applicable)
 result_header   dict            header for result
 result_content  dict            content for result
 result_buffers  list(bytes)     buffers containing serialized request objects
 queue           bytes           The name of the queue for the task ('mux' or 'task')
 pyin            <unused>        Python input (unused)
 pyout           <unused>        Python output (unused)
 pyerr           <unused>        Python traceback (unused)
 stdout          str             Stream of stdout data
 stderr          str             Stream of stderr data
 
 =============== =============== =============
 
 MongoDB operators we emulate on all backends:
 
 ==========  =================
 Operator    Python equivalent
 ==========  =================
   '$in'       in
   '$nin'      not in
   '$eq'       ==
   '$ne'       !=
   '$ge'       >
   '$gte'      >=
   '$le'       <
   '$lte'      <=
 ==========  =================
 
 
 The DB Query is useful for two primary cases:
 
 1. deep polling of task status or metadata
 2. selecting a subset of tasks, on which to perform a later operation (e.g. wait on result, purge records, resubmit,...)
 
 Example Queries
 ===============
 
 
 To get all msg_ids that are not completed, only retrieving their ID and start time:
 
 .. sourcecode:: ipython
 
     In [1]: incomplete = rc.db_query({'complete' : None}, keys=['msg_id', 'started'])
 
 All jobs started in the last hour by me:
 
 .. sourcecode:: ipython
 
     In [1]: from datetime import datetime, timedelta
 
     In [2]: hourago = datetime.now() - timedelta(1./24)
 
     In [3]: recent = rc.db_query({'started' : {'$gte' : hourago },
                                     'client_uuid' : rc.session.session})
 
 All jobs started more than an hour ago, by clients *other than me*:
 
 .. sourcecode:: ipython
 
     In [3]: recent = rc.db_query({'started' : {'$le' : hourago },
                                     'client_uuid' : {'$ne' : rc.session.session}})
 
 Result headers for all jobs on engine 3 or 4:
 
 .. sourcecode:: ipython
 
     In [1]: uuids = map(rc._engines.get, (3,4))
 
     In [2]: hist34 = rc.db_query({'engine_uuid' : {'$in' : uuids }, keys='result_header')
 
 
 Cost
 ====
 
 The advantage of the database backends is, of course, that large amounts of
 data can be stored that won't fit in memory.  The default 'backend' is actually
 to just store all of this information in a Python dictionary.  This is very fast,
 but will run out of memory quickly if you move a lot of data around, or your
 cluster is to run for a long time.
 
 Unfortunately, the DB backends (SQLite and MongoDB) right now are rather slow,
 and can still consume large amounts of resources, particularly if large tasks
 or results are being created at a high frequency.
 
 For this reason, we have added :class:`~.NoDB`,a dummy backend that doesn't
 actually store any information. When you use this database, nothing is stored,
 and any request for results will result in a KeyError.  This obviously prevents
 later requests for results and task resubmission from functioning, but
 sometimes those nice features are not as useful as keeping Hub memory under
 control.
 
 

IPython/parallel/engine/streamkernel.py

@@ -1,414 +0,0 b''
-"""
-Kernel adapted from kernel.py to use ZMQ Streams
-
-Authors:
-
-* Min RK
-* Brian Granger
-* Fernando Perez
-* Evan Patterson
-"""
-#-----------------------------------------------------------------------------
-#  Copyright (C) 2010-2011  The IPython Development Team
-#
-#  Distributed under the terms of the BSD License.  The full license is in
-#  the file COPYING, distributed as part of this software.
-#-----------------------------------------------------------------------------
-
-#-----------------------------------------------------------------------------
-# Imports
-#-----------------------------------------------------------------------------
-
-# Standard library imports.
-from __future__ import print_function
-
-import sys
-import time
-
-from code import CommandCompiler
-from datetime import datetime
-from pprint import pprint
-
-# System library imports.
-import zmq
-from zmq.eventloop import ioloop, zmqstream
-
-# Local imports.
-from IPython.utils.traitlets import Instance, List, Integer, Dict, Set, Unicode, CBytes
-from IPython.zmq.completer import KernelCompleter
-
-from IPython.parallel.error import wrap_exception
-from IPython.parallel.factory import SessionFactory
-from IPython.parallel.util import serialize_object, unpack_apply_message, asbytes
-
-def printer(*args):
-    pprint(args, stream=sys.__stdout__)
-
-
-class _Passer(zmqstream.ZMQStream):
-    """Empty class that implements `send()` that does nothing.
-
-    Subclass ZMQStream for Session typechecking
-
-    """
-    def __init__(self, *args, **kwargs):
-        pass
-
-    def send(self, *args, **kwargs):
-        pass
-    send_multipart = send
-
-
-#-----------------------------------------------------------------------------
-# Main kernel class
-#-----------------------------------------------------------------------------
-
-class Kernel(SessionFactory):
-
-    #---------------------------------------------------------------------------
-    # Kernel interface
-    #---------------------------------------------------------------------------
-
-    # kwargs:
-    exec_lines = List(Unicode, config=True,
-        help="List of lines to execute")
-
-    # identities:
-    int_id = Integer(-1)
-    bident = CBytes()
-    ident = Unicode()
-    def _ident_changed(self, name, old, new):
-        self.bident = asbytes(new)
-
-    user_ns = Dict(config=True,  help="""Set the user's namespace of the Kernel""")
-
-    control_stream = Instance(zmqstream.ZMQStream)
-    task_stream = Instance(zmqstream.ZMQStream)
-    iopub_stream = Instance(zmqstream.ZMQStream)
-    client = Instance('IPython.parallel.Client')
-
-    # internals
-    shell_streams = List()
-    compiler = Instance(CommandCompiler, (), {})
-    completer = Instance(KernelCompleter)
-
-    aborted = Set()
-    shell_handlers = Dict()
-    control_handlers = Dict()
-
-    def _set_prefix(self):
-        self.prefix = "engine.%s"%self.int_id
-
-    def _connect_completer(self):
-        self.completer = KernelCompleter(self.user_ns)
-
-    def __init__(self, **kwargs):
-        super(Kernel, self).__init__(**kwargs)
-        self._set_prefix()
-        self._connect_completer()
-
-        self.on_trait_change(self._set_prefix, 'id')
-        self.on_trait_change(self._connect_completer, 'user_ns')
-
-        # Build dict of handlers for message types
-        for msg_type in ['execute_request', 'complete_request', 'apply_request',
-                'clear_request']:
-            self.shell_handlers[msg_type] = getattr(self, msg_type)
-
-        for msg_type in ['shutdown_request', 'abort_request']+self.shell_handlers.keys():
-            self.control_handlers[msg_type] = getattr(self, msg_type)
-
-        self._initial_exec_lines()
-
-    def _wrap_exception(self, method=None):
-        e_info = dict(engine_uuid=self.ident, engine_id=self.int_id, method=method)
-        content=wrap_exception(e_info)
-        return content
-
-    def _initial_exec_lines(self):
-        s = _Passer()
-        content = dict(silent=True, user_variable=[],user_expressions=[])
-        for line in self.exec_lines:
-            self.log.debug("executing initialization: %s"%line)
-            content.update({'code':line})
-            msg = self.session.msg('execute_request', content)
-            self.execute_request(s, [], msg)
-
-
-    #-------------------- control handlers -----------------------------
-    def abort_queues(self):
-        for stream in self.shell_streams:
-            if stream:
-                self.abort_queue(stream)
-
-    def abort_queue(self, stream):
-        while True:
-            idents,msg = self.session.recv(stream, zmq.NOBLOCK, content=True)
-            if msg is None:
-                return
-
-            self.log.info("Aborting:")
-            self.log.info(str(msg))
-            msg_type = msg['header']['msg_type']
-            reply_type = msg_type.split('_')[0] + '_reply'
-            # reply_msg = self.session.msg(reply_type, {'status' : 'aborted'}, msg)
-            # self.reply_socket.send(ident,zmq.SNDMORE)
-            # self.reply_socket.send_json(reply_msg)
-            reply_msg = self.session.send(stream, reply_type,
-                        content={'status' : 'aborted'}, parent=msg, ident=idents)
-            self.log.debug(str(reply_msg))
-            # We need to wait a bit for requests to come in. This can probably
-            # be set shorter for true asynchronous clients.
-            time.sleep(0.05)
-
-    def abort_request(self, stream, ident, parent):
-        """abort a specifig msg by id"""
-        msg_ids = parent['content'].get('msg_ids', None)
-        if isinstance(msg_ids, basestring):
-            msg_ids = [msg_ids]
-        if not msg_ids:
-            self.abort_queues()
-        for mid in msg_ids:
-            self.aborted.add(str(mid))
-
-        content = dict(status='ok')
-        reply_msg = self.session.send(stream, 'abort_reply', content=content,
-                parent=parent, ident=ident)
-        self.log.debug(str(reply_msg))
-
-    def shutdown_request(self, stream, ident, parent):
-        """kill ourself.  This should really be handled in an external process"""
-        try:
-            self.abort_queues()
-        except:
-            content = self._wrap_exception('shutdown')
-        else:
-            content = dict(parent['content'])
-            content['status'] = 'ok'
-        msg = self.session.send(stream, 'shutdown_reply',
-                                content=content, parent=parent, ident=ident)
-        self.log.debug(str(msg))
-        dc = ioloop.DelayedCallback(lambda : sys.exit(0), 1000, self.loop)
-        dc.start()
-
-    def dispatch_control(self, msg):
-        idents,msg = self.session.feed_identities(msg, copy=False)
-        try:
-            msg = self.session.unserialize(msg, content=True, copy=False)
-        except:
-            self.log.error("Invalid Message", exc_info=True)
-            return
-        else:
-            self.log.debug("Control received, %s", msg)
-
-        header = msg['header']
-        msg_id = header['msg_id']
-        msg_type = header['msg_type']
-
-        handler = self.control_handlers.get(msg_type, None)
-        if handler is None:
-            self.log.error("UNKNOWN CONTROL MESSAGE TYPE: %r"%msg_type)
-        else:
-            handler(self.control_stream, idents, msg)
-
-
-    #-------------------- queue helpers ------------------------------
-
-    def check_dependencies(self, dependencies):
-        if not dependencies:
-            return True
-        if len(dependencies) == 2 and dependencies[0] in 'any all'.split():
-            anyorall = dependencies[0]
-            dependencies = dependencies[1]
-        else:
-            anyorall = 'all'
-        results = self.client.get_results(dependencies,status_only=True)
-        if results['status'] != 'ok':
-            return False
-
-        if anyorall == 'any':
-            if not results['completed']:
-                return False
-        else:
-            if results['pending']:
-                return False
-
-        return True
-
-    def check_aborted(self, msg_id):
-        return msg_id in self.aborted
-
-    #-------------------- queue handlers -----------------------------
-
-    def clear_request(self, stream, idents, parent):
-        """Clear our namespace."""
-        self.user_ns = {}
-        msg = self.session.send(stream, 'clear_reply', ident=idents, parent=parent,
-                content = dict(status='ok'))
-        self._initial_exec_lines()
-
-    def execute_request(self, stream, ident, parent):
-        self.log.debug('execute request %s'%parent)
-        try:
-            code = parent[u'content'][u'code']
-        except:
-            self.log.error("Got bad msg: %s"%parent, exc_info=True)
-            return
-        self.session.send(self.iopub_stream, u'pyin', {u'code':code},parent=parent,
-                            ident=asbytes('%s.pyin'%self.prefix))
-        started = datetime.now()
-        try:
-            comp_code = self.compiler(code, '<zmq-kernel>')
-            # allow for not overriding displayhook
-            if hasattr(sys.displayhook, 'set_parent'):
-                sys.displayhook.set_parent(parent)
-                sys.stdout.set_parent(parent)
-                sys.stderr.set_parent(parent)
-            exec comp_code in self.user_ns, self.user_ns
-        except:
-            exc_content = self._wrap_exception('execute')
-            # exc_msg = self.session.msg(u'pyerr', exc_content, parent)
-            self.session.send(self.iopub_stream, u'pyerr', exc_content, parent=parent,
-                            ident=asbytes('%s.pyerr'%self.prefix))
-            reply_content = exc_content
-        else:
-            reply_content = {'status' : 'ok'}
-
-        reply_msg = self.session.send(stream, u'execute_reply', reply_content, parent=parent,
-                    ident=ident, subheader = dict(started=started))
-        self.log.debug(str(reply_msg))
-        if reply_msg['content']['status'] == u'error':
-            self.abort_queues()
-
-    def complete_request(self, stream, ident, parent):
-        matches = {'matches' : self.complete(parent),
-                   'status' : 'ok'}
-        completion_msg = self.session.send(stream, 'complete_reply',
-                                           matches, parent, ident)
-        # print >> sys.__stdout__, completion_msg
-
-    def complete(self, msg):
-        return self.completer.complete(msg.content.line, msg.content.text)
-
-    def apply_request(self, stream, ident, parent):
-        # flush previous reply, so this request won't block it
-        stream.flush(zmq.POLLOUT)
-        try:
-            content = parent[u'content']
-            bufs = parent[u'buffers']
-            msg_id = parent['header']['msg_id']
-            # bound = parent['header'].get('bound', False)
-        except:
-            self.log.error("Got bad msg: %s"%parent, exc_info=True)
-            return
-        # pyin_msg = self.session.msg(u'pyin',{u'code':code}, parent=parent)
-        # self.iopub_stream.send(pyin_msg)
-        # self.session.send(self.iopub_stream, u'pyin', {u'code':code},parent=parent)
-        sub = {'dependencies_met' : True, 'engine' : self.ident,
-                'started': datetime.now()}
-        try:
-            # allow for not overriding displayhook
-            if hasattr(sys.displayhook, 'set_parent'):
-                sys.displayhook.set_parent(parent)
-                sys.stdout.set_parent(parent)
-                sys.stderr.set_parent(parent)
-            # exec "f(*args,**kwargs)" in self.user_ns, self.user_ns
-            working = self.user_ns
-            # suffix =
-            prefix = "_"+str(msg_id).replace("-","")+"_"
-
-            f,args,kwargs = unpack_apply_message(bufs, working, copy=False)
-            # if bound:
-            #     bound_ns = Namespace(working)
-            #     args = [bound_ns]+list(args)
-
-            fname = getattr(f, '__name__', 'f')
-
-            fname = prefix+"f"
-            argname = prefix+"args"
-            kwargname = prefix+"kwargs"
-            resultname = prefix+"result"
-
-            ns = { fname : f, argname : args, kwargname : kwargs , resultname : None }
-            # print ns
-            working.update(ns)
-            code = "%s=%s(*%s,**%s)"%(resultname, fname, argname, kwargname)
-            try:
-                exec code in working,working
-                result = working.get(resultname)
-            finally:
-                for key in ns.iterkeys():
-                    working.pop(key)
-            # if bound:
-            #     working.update(bound_ns)
-
-            packed_result,buf = serialize_object(result)
-            result_buf = [packed_result]+buf
-        except:
-            exc_content = self._wrap_exception('apply')
-            # exc_msg = self.session.msg(u'pyerr', exc_content, parent)
-            self.session.send(self.iopub_stream, u'pyerr', exc_content, parent=parent,
-                                ident=asbytes('%s.pyerr'%self.prefix))
-            reply_content = exc_content
-            result_buf = []
-
-            if exc_content['ename'] == 'UnmetDependency':
-                sub['dependencies_met'] = False
-        else:
-            reply_content = {'status' : 'ok'}
-
-        # put 'ok'/'error' status in header, for scheduler introspection:
-        sub['status'] = reply_content['status']
-
-        reply_msg = self.session.send(stream, u'apply_reply', reply_content,
-                    parent=parent, ident=ident,buffers=result_buf, subheader=sub)
-
-        # flush i/o
-        # should this be before reply_msg is sent, like in the single-kernel code,
-        # or should nothing get in the way of real results?
-        sys.stdout.flush()
-        sys.stderr.flush()
-
-    def dispatch_queue(self, stream, msg):
-        self.control_stream.flush()
-        idents,msg = self.session.feed_identities(msg, copy=False)
-        try:
-            msg = self.session.unserialize(msg, content=True, copy=False)
-        except:
-            self.log.error("Invalid Message", exc_info=True)
-            return
-        else:
-            self.log.debug("Message received, %s", msg)
-
-
-        header = msg['header']
-        msg_id = header['msg_id']
-        msg_type = msg['header']['msg_type']
-        if self.check_aborted(msg_id):
-            self.aborted.remove(msg_id)
-            # is it safe to assume a msg_id will not be resubmitted?
-            reply_type = msg_type.split('_')[0] + '_reply'
-            status = {'status' : 'aborted'}
-            reply_msg = self.session.send(stream, reply_type, subheader=status,
-                        content=status, parent=msg, ident=idents)
-            return
-        handler = self.shell_handlers.get(msg_type, None)
-        if handler is None:
-            self.log.error("UNKNOWN MESSAGE TYPE: %r"%msg_type)
-        else:
-            handler(stream, idents, msg)
-
-    def start(self):
-        #### stream mode:
-        if self.control_stream:
-            self.control_stream.on_recv(self.dispatch_control, copy=False)
-
-        def make_dispatcher(stream):
-            def dispatcher(msg):
-                return self.dispatch_queue(stream, msg)
-            return dispatcher
-
-        for s in self.shell_streams:
-            s.on_recv(make_dispatcher(s), copy=False)
-
-

IPython/zmq/pykernel.py

@@ -1,282 +0,0 b''
-#!/usr/bin/env python
-"""A simple interactive kernel that talks to a frontend over 0MQ.
-
-Things to do:
-
-* Implement `set_parent` logic. Right before doing exec, the Kernel should
-  call set_parent on all the PUB objects with the message about to be executed.
-* Implement random port and security key logic.
-* Implement control messages.
-* Implement event loop and poll version.
-"""
-
-#-----------------------------------------------------------------------------
-# Imports
-#-----------------------------------------------------------------------------
-
-# Standard library imports.
-import __builtin__
-from code import CommandCompiler
-import sys
-import time
-import traceback
-
-# System library imports.
-import zmq
-
-# Local imports.
-from IPython.utils import py3compat
-from IPython.utils.traitlets import HasTraits, Instance, Dict, Float
-from completer import KernelCompleter
-from entry_point import base_launch_kernel
-from session import Session, Message
-from kernelapp import KernelApp
-
-#-----------------------------------------------------------------------------
-# Main kernel class
-#-----------------------------------------------------------------------------
-
-class Kernel(HasTraits):
-
-    # Private interface
-
-    # Time to sleep after flushing the stdout/err buffers in each execute
-    # cycle.  While this introduces a hard limit on the minimal latency of the
-    # execute cycle, it helps prevent output synchronization problems for
-    # clients.
-    # Units are in seconds.  The minimum zmq latency on local host is probably
-    # ~150 microseconds, set this to 500us for now.  We may need to increase it
-    # a little if it's not enough after more interactive testing.
-    _execute_sleep = Float(0.0005, config=True)
-
-    # This is a dict of port number that the kernel is listening on. It is set
-    # by record_ports and used by connect_request.
-    _recorded_ports = Dict()
-
-    #---------------------------------------------------------------------------
-    # Kernel interface
-    #---------------------------------------------------------------------------
-
-    session = Instance(Session)
-    shell_socket = Instance('zmq.Socket')
-    iopub_socket = Instance('zmq.Socket')
-    stdin_socket = Instance('zmq.Socket')
-    log = Instance('logging.Logger')
-
-    def __init__(self, **kwargs):
-        super(Kernel, self).__init__(**kwargs)
-        self.user_ns = {}
-        self.history = []
-        self.compiler = CommandCompiler()
-        self.completer = KernelCompleter(self.user_ns)
-
-        # Build dict of handlers for message types
-        msg_types = [ 'execute_request', 'complete_request',
-                      'object_info_request', 'shutdown_request' ]
-        self.handlers = {}
-        for msg_type in msg_types:
-            self.handlers[msg_type] = getattr(self, msg_type)
-
-    def start(self):
-        """ Start the kernel main loop.
-        """
-        while True:
-            ident,msg = self.session.recv(self.shell_socket,0)
-            assert ident is not None, "Missing message part."
-            omsg = Message(msg)
-            self.log.debug(str(omsg))
-            handler = self.handlers.get(omsg.msg_type, None)
-            if handler is None:
-                self.log.error("UNKNOWN MESSAGE TYPE: %s"%omsg)
-            else:
-                handler(ident, omsg)
-
-    def record_ports(self, ports):
-        """Record the ports that this kernel is using.
-
-        The creator of the Kernel instance must call this methods if they
-        want the :meth:`connect_request` method to return the port numbers.
-        """
-        self._recorded_ports = ports
-
-    #---------------------------------------------------------------------------
-    # Kernel request handlers
-    #---------------------------------------------------------------------------
-
-    def execute_request(self, ident, parent):
-        try:
-            code = parent[u'content'][u'code']
-        except:
-            self.log.error("Got bad msg: %s"%Message(parent))
-            return
-        pyin_msg = self.session.send(self.iopub_socket, u'pyin',{u'code':code}, parent=parent)
-
-        try:
-            comp_code = self.compiler(code, '<zmq-kernel>')
-
-            # Replace raw_input. Note that is not sufficient to replace
-            # raw_input in the user namespace.
-            raw_input = lambda prompt='': self._raw_input(prompt, ident, parent)
-            if py3compat.PY3:
-                __builtin__.input = raw_input
-            else:
-                __builtin__.raw_input = raw_input
-
-            # Set the parent message of the display hook and out streams.
-            sys.displayhook.set_parent(parent)
-            sys.stdout.set_parent(parent)
-            sys.stderr.set_parent(parent)
-
-            exec comp_code in self.user_ns, self.user_ns
-        except:
-            etype, evalue, tb = sys.exc_info()
-            tb = traceback.format_exception(etype, evalue, tb)
-            exc_content = {
-                u'status' : u'error',
-                u'traceback' : tb,
-                u'ename' : unicode(etype.__name__),
-                u'evalue' : unicode(evalue)
-            }
-            exc_msg = self.session.send(self.iopub_socket, u'pyerr', exc_content, parent)
-            reply_content = exc_content
-        else:
-            reply_content = { 'status' : 'ok', 'payload' : {} }
-
-        # Flush output before sending the reply.
-        sys.stderr.flush()
-        sys.stdout.flush()
-        # FIXME: on rare occasions, the flush doesn't seem to make it to the
-        # clients... This seems to mitigate the problem, but we definitely need
-        # to better understand what's going on.
-        if self._execute_sleep:
-            time.sleep(self._execute_sleep)
-
-        # Send the reply.
-        reply_msg = self.session.send(self.shell_socket, u'execute_reply', reply_content, parent, ident=ident)
-        self.log.debug(Message(reply_msg))
-        if reply_msg['content']['status'] == u'error':
-            self._abort_queue()
-
-    def complete_request(self, ident, parent):
-        matches = {'matches' : self._complete(parent),
-                   'status' : 'ok'}
-        completion_msg = self.session.send(self.shell_socket, 'complete_reply',
-                                           matches, parent, ident)
-        self.log.debug(completion_msg)
-
-    def object_info_request(self, ident, parent):
-        context = parent['content']['oname'].split('.')
-        object_info = self._object_info(context)
-        msg = self.session.send(self.shell_socket, 'object_info_reply',
-                                object_info, parent, ident)
-        self.log.debug(msg)
-
-    def shutdown_request(self, ident, parent):
-        content = dict(parent['content'])
-        msg = self.session.send(self.shell_socket, 'shutdown_reply',
-                                content, parent, ident)
-        msg = self.session.send(self.iopub_socket, 'shutdown_reply',
-                                content, parent, ident)
-        self.log.debug(msg)
-        time.sleep(0.1)
-        sys.exit(0)
-
-    #---------------------------------------------------------------------------
-    # Protected interface
-    #---------------------------------------------------------------------------
-
-    def _abort_queue(self):
-        while True:
-            ident,msg = self.session.recv(self.shell_socket, zmq.NOBLOCK)
-            if msg is None:
-                # msg=None on EAGAIN
-                break
-            else:
-                assert ident is not None, "Missing message part."
-            self.log.debug("Aborting: %s"%Message(msg))
-            msg_type = msg['header']['msg_type']
-            reply_type = msg_type.split('_')[0] + '_reply'
-            reply_msg = self.session.send(self.shell_socket, reply_type, {'status':'aborted'}, msg, ident=ident)
-            self.log.debug(Message(reply_msg))
-            # We need to wait a bit for requests to come in. This can probably
-            # be set shorter for true asynchronous clients.
-            time.sleep(0.1)
-
-    def _raw_input(self, prompt, ident, parent):
-        # Flush output before making the request.
-        sys.stderr.flush()
-        sys.stdout.flush()
-
-        # Send the input request.
-        content = dict(prompt=prompt)
-        msg = self.session.send(self.stdin_socket, u'input_request', content, parent, ident=ident)
-
-        # Await a response.
-        ident,reply = self.session.recv(self.stdin_socket, 0)
-        try:
-            value = reply['content']['value']
-        except:
-            self.log.error("Got bad raw_input reply: %s"%Message(parent))
-            value = ''
-        return value
-
-    def _complete(self, msg):
-        return self.completer.complete(msg.content.line, msg.content.text)
-
-    def _object_info(self, context):
-        symbol, leftover = self._symbol_from_context(context)
-        if symbol is not None and not leftover:
-            doc = getattr(symbol, '__doc__', '')
-        else:
-            doc = ''
-        object_info = dict(docstring = doc)
-        return object_info
-
-    def _symbol_from_context(self, context):
-        if not context:
-            return None, context
-
-        base_symbol_string = context[0]
-        symbol = self.user_ns.get(base_symbol_string, None)
-        if symbol is None:
-            symbol = __builtin__.__dict__.get(base_symbol_string, None)
-        if symbol is None:
-            return None, context
-
-        context = context[1:]
-        for i, name in enumerate(context):
-            new_symbol = getattr(symbol, name, None)
-            if new_symbol is None:
-                return symbol, context[i:]
-            else:
-                symbol = new_symbol
-
-        return symbol, []
-
-#-----------------------------------------------------------------------------
-# Kernel main and launch functions
-#-----------------------------------------------------------------------------
-
-def launch_kernel(*args, **kwargs):
-    """ Launches a simple Python kernel, binding to the specified ports.
-
-    This function simply calls entry_point.base_launch_kernel with the right first
-    command to start a pykernel.  See base_launch_kernel for arguments.
-
-    Returns
-    -------
-    A tuple of form:
-        (kernel_process, xrep_port, pub_port, req_port, hb_port)
-    where kernel_process is a Popen object and the ports are integers.
-    """
-    return base_launch_kernel('from IPython.zmq.pykernel import main; main()',
-                                *args, **kwargs)
-
-def main():
-    """Run a PyKernel as an application"""
-    app = KernelApp.instance()
-    app.initialize()
-    app.start()
-
-if __name__ == '__main__':
-    main()