Fixing minor bugs in IPython.kernel.

* Regular expressions now using raw strings in launcher.py
* Directory permissions on cluster directory handled properly.
* A raw ipcluster start will now create cluster default and add the
  default config files.
* Config files updated to use raw strings where appropriate.

IPython/config/default/ipcluster_config.py

@@ -1,184 +1,184 b''
 import os
 
 c = get_config()
 
 #-----------------------------------------------------------------------------
 # Select which launchers to use
 #-----------------------------------------------------------------------------
 
 # This allows you to control what method is used to start the controller
 # and engines.  The following methods are currently supported:
 # - Start as a regular process on localhost.
 # - Start using mpiexec.
 # - Start using the Windows HPC Server 2008 scheduler
 # - Start using PBS
 # - Start using SSH (currently broken)
 
 
 # The selected launchers can be configured below.
 
 # Options are:
 # - LocalControllerLauncher
 # - MPIExecControllerLauncher
 # - PBSControllerLauncher
 # - WindowsHPCControllerLauncher
 # c.Global.controller_launcher = 'IPython.kernel.launcher.LocalControllerLauncher'
 
 # Options are:
 # - LocalEngineSetLauncher
 # - MPIExecEngineSetLauncher
 # - PBSEngineSetLauncher
 # - WindowsHPCEngineSetLauncher
 # c.Global.engine_launcher = 'IPython.kernel.launcher.LocalEngineSetLauncher'
 
 #-----------------------------------------------------------------------------
 # Global configuration
 #-----------------------------------------------------------------------------
 
 # The default number of engines that will be started. This is overridden by
 # the -n command line option: "ipcluster start -n 4"
 # c.Global.n = 2
 
 # Log to a file in cluster_dir/log, otherwise just log to sys.stdout.
 # c.Global.log_to_file = False
 
 # Remove old logs from cluster_dir/log before starting.
 # c.Global.clean_logs = True
 
 # The working directory for the process. The application will use os.chdir
 # to change to this directory before starting.
 # c.Global.work_dir = os.getcwd()
 
 
 #-----------------------------------------------------------------------------
 # Local process launchers
 #-----------------------------------------------------------------------------
 
 # The command line arguments to call the controller with.
 # c.LocalControllerLauncher.controller_args = \
 #    ['--log-to-file','--log-level', '40']
 
 # The working directory for the controller
 # c.LocalEngineSetLauncher.work_dir = u''
 
 # Command line argument passed to the engines.
 # c.LocalEngineSetLauncher.engine_args = ['--log-to-file','--log-level', '40']
 
 #-----------------------------------------------------------------------------
 # MPIExec launchers
 #-----------------------------------------------------------------------------
 
 # The mpiexec/mpirun command to use in started the controller.
 # c.MPIExecControllerLauncher.mpi_cmd = ['mpiexec']
 
 # Additional arguments to pass to the actual mpiexec command.
 # c.MPIExecControllerLauncher.mpi_args = []
 
 # The command line argument to call the controller with.
 # c.MPIExecControllerLauncher.controller_args = \
 #     ['--log-to-file','--log-level', '40']
 
 
 # The mpiexec/mpirun command to use in started the controller.
 # c.MPIExecEngineSetLauncher.mpi_cmd = ['mpiexec']
 
 # Additional arguments to pass to the actual mpiexec command.
 # c.MPIExecEngineSetLauncher.mpi_args = []
 
 # Command line argument passed to the engines.
 # c.MPIExecEngineSetLauncher.engine_args = ['--log-to-file','--log-level', '40']
 
 # The default number of engines to start if not given elsewhere.
 # c.MPIExecEngineSetLauncher.n = 1
 
 #-----------------------------------------------------------------------------
 # SSH launchers
 #-----------------------------------------------------------------------------
 
 # Todo
 
 
 #-----------------------------------------------------------------------------
 # Unix batch (PBS) schedulers launchers
 #-----------------------------------------------------------------------------
 
 # The command line program to use to submit a PBS job.
 # c.PBSControllerLauncher.submit_command = 'qsub'
 
 # The command line program to use to delete a PBS job.
 # c.PBSControllerLauncher.delete_command = 'qdel'
 
 # A regular expression that takes the output of qsub and find the job id.
-# c.PBSControllerLauncher.job_id_regexp = '\d+'
+# c.PBSControllerLauncher.job_id_regexp = r'\d+'
 
 # The batch submission script used to start the controller. This is where
 # environment variables would be setup, etc. This string is interpolated using
 # the Itpl module in IPython.external. Basically, you can use ${n} for the
 # number of engine and ${cluster_dir} for the cluster_dir.
 # c.PBSControllerLauncher.batch_template = """"""
 
 # The name of the instantiated batch script that will actually be used to
 # submit the job. This will be written to the cluster directory.
 # c.PBSControllerLauncher.batch_file_name = u'pbs_batch_script_controller'
 
 
 # The command line program to use to submit a PBS job.
 # c.PBSEngineSetLauncher.submit_command = 'qsub'
 
 # The command line program to use to delete a PBS job.
 # c.PBSEngineSetLauncher.delete_command = 'qdel'
 
 # A regular expression that takes the output of qsub and find the job id.
-# c.PBSEngineSetLauncher.job_id_regexp = '\d+'
+# c.PBSEngineSetLauncher.job_id_regexp = r'\d+'
 
 # The batch submission script used to start the engines. This is where
 # environment variables would be setup, etc. This string is interpolated using
 # the Itpl module in IPython.external. Basically, you can use ${n} for the
 # number of engine and ${cluster_dir} for the cluster_dir.
 # c.PBSEngineSetLauncher.batch_template = """"""
 
 # The name of the instantiated batch script that will actually be used to
 # submit the job. This will be written to the cluster directory.
 # c.PBSEngineSetLauncher.batch_file_name = u'pbs_batch_script_engines'
 
 #-----------------------------------------------------------------------------
 # Windows HPC Server 2008 launcher configuration
 #-----------------------------------------------------------------------------
 
 # c.IPControllerJob.job_name = 'IPController'
 # c.IPControllerJob.is_exclusive = False
-# c.IPControllerJob.username = 'USERDOMAIN\\USERNAME'
+# c.IPControllerJob.username = r'USERDOMAIN\USERNAME'
 # c.IPControllerJob.priority = 'Highest'
 # c.IPControllerJob.requested_nodes = ''
 # c.IPControllerJob.project = 'MyProject'
 
 # c.IPControllerTask.task_name = 'IPController'
 # c.IPControllerTask.controller_cmd = [u'ipcontroller.exe']
 # c.IPControllerTask.controller_args = ['--log-to-file', '--log-level', '40']
 # c.IPControllerTask.environment_variables = {}
 
 # c.WindowsHPCControllerLauncher.scheduler = 'HEADNODE'
 # c.WindowsHPCControllerLauncher.job_file_name = u'ipcontroller_job.xml'
 
 
 # c.IPEngineSetJob.job_name = 'IPEngineSet'
 # c.IPEngineSetJob.is_exclusive = False
-# c.IPEngineSetJob.username = 'USERDOMAIN\\USERNAME'
+# c.IPEngineSetJob.username = r'USERDOMAIN\USERNAME'
 # c.IPEngineSetJob.priority = 'Highest'
 # c.IPEngineSetJob.requested_nodes = ''
 # c.IPEngineSetJob.project = 'MyProject'
 
 # c.IPEngineTask.task_name = 'IPEngine'
 # c.IPEngineTask.engine_cmd = [u'ipengine.exe']
 # c.IPEngineTask.engine_args = ['--log-to-file', '--log-level', '40']
 # c.IPEngineTask.environment_variables = {}
 
 # c.WindowsHPCEngineSetLauncher.scheduler = 'HEADNODE'
 # c.WindowsHPCEngineSetLauncher.job_file_name = u'ipengineset_job.xml'
 
 
 
 
 
 
 

IPython/kernel/clusterdir.py

@@ -1,481 +1,475 b''
 #!/usr/bin/env python
 # encoding: utf-8
 """
 The IPython cluster directory
 """
 
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2008-2009  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 shutil
 import sys
 
 from twisted.python import log
 
 from IPython.core import release
 from IPython.config.loader import PyFileConfigLoader
 from IPython.core.application import Application
 from IPython.core.component import Component
 from IPython.config.loader import ArgParseConfigLoader, NoConfigDefault
 from IPython.utils.traitlets import Unicode, Bool
 from IPython.utils import genutils
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 
 
 class ClusterDirError(Exception):
     pass
 
 
 class PIDFileError(Exception):
     pass
 
 
 class ClusterDir(Component):
     """An object to manage the cluster directory and its resources.
 
     The cluster directory is used by :command:`ipcontroller`, 
     :command:`ipcontroller` and :command:`ipcontroller` to manage the
     configuration, logging and security of these applications.
 
     This object knows how to find, create and manage these directories. This
     should be used by any code that want's to handle cluster directories.
     """
 
     security_dir_name = Unicode('security')
     log_dir_name = Unicode('log')
     pid_dir_name = Unicode('pid')
     security_dir = Unicode(u'')
     log_dir = Unicode(u'')
     pid_dir = Unicode(u'')
     location = Unicode(u'')
 
     def __init__(self, location):
         super(ClusterDir, self).__init__(None)
         self.location = location
 
     def _location_changed(self, name, old, new):
         if not os.path.isdir(new):
-            os.makedirs(new, mode=0777)
-        else:
-            os.chmod(new, 0777)
+            os.makedirs(new)
         self.security_dir = os.path.join(new, self.security_dir_name)
         self.log_dir = os.path.join(new, self.log_dir_name)
         self.pid_dir = os.path.join(new, self.pid_dir_name)
         self.check_dirs()
 
     def _log_dir_changed(self, name, old, new):
         self.check_log_dir()
 
     def check_log_dir(self):
         if not os.path.isdir(self.log_dir):
-            os.mkdir(self.log_dir, 0777)
-        else:
-            os.chmod(self.log_dir, 0777)
+            os.mkdir(self.log_dir)
 
     def _security_dir_changed(self, name, old, new):
         self.check_security_dir()
 
     def check_security_dir(self):
         if not os.path.isdir(self.security_dir):
             os.mkdir(self.security_dir, 0700)
-        else:
             os.chmod(self.security_dir, 0700)
 
     def _pid_dir_changed(self, name, old, new):
         self.check_pid_dir()
 
     def check_pid_dir(self):
         if not os.path.isdir(self.pid_dir):
             os.mkdir(self.pid_dir, 0700)
-        else:
             os.chmod(self.pid_dir, 0700)
 
     def check_dirs(self):
         self.check_security_dir()
         self.check_log_dir()
         self.check_pid_dir()
 
     def load_config_file(self, filename):
         """Load a config file from the top level of the cluster dir.
 
         Parameters
         ----------
         filename : unicode or str
             The filename only of the config file that must be located in
             the top-level of the cluster directory.
         """
         loader = PyFileConfigLoader(filename, self.location)
         return loader.load_config()
 
     def copy_config_file(self, config_file, path=None, overwrite=False):
         """Copy a default config file into the active cluster directory.
 
         Default configuration files are kept in :mod:`IPython.config.default`.
         This function moves these from that location to the working cluster
         directory.
         """
         if path is None:
             import IPython.config.default
             path = IPython.config.default.__file__.split(os.path.sep)[:-1]
             path = os.path.sep.join(path)
         src = os.path.join(path, config_file)
         dst = os.path.join(self.location, config_file)
         if not os.path.isfile(dst) or overwrite:
             shutil.copy(src, dst)
 
     def copy_all_config_files(self, path=None, overwrite=False):
         """Copy all config files into the active cluster directory."""
         for f in [u'ipcontroller_config.py', u'ipengine_config.py',
                   u'ipcluster_config.py']:
             self.copy_config_file(f, path=path, overwrite=overwrite)
 
     @classmethod
     def create_cluster_dir(csl, cluster_dir):
         """Create a new cluster directory given a full path.
 
         Parameters
         ----------
         cluster_dir : str
             The full path to the cluster directory.  If it does exist, it will
             be used.  If not, it will be created.
         """
         return ClusterDir(cluster_dir)
 
     @classmethod
     def create_cluster_dir_by_profile(cls, path, profile=u'default'):
         """Create a cluster dir by profile name and path.
 
         Parameters
         ----------
         path : str
             The path (directory) to put the cluster directory in.
         profile : str
             The name of the profile.  The name of the cluster directory will
             be "cluster_<profile>".
         """
         if not os.path.isdir(path):
             raise ClusterDirError('Directory not found: %s' % path)
         cluster_dir = os.path.join(path, u'cluster_' + profile)
         return ClusterDir(cluster_dir)
 
     @classmethod
     def find_cluster_dir_by_profile(cls, ipython_dir, profile=u'default'):
         """Find an existing cluster dir by profile name, return its ClusterDir.
 
         This searches through a sequence of paths for a cluster dir.  If it
         is not found, a :class:`ClusterDirError` exception will be raised.
 
         The search path algorithm is:
         1. ``os.getcwd()``
         2. ``ipython_dir``
         3. The directories found in the ":" separated 
            :env:`IPCLUSTER_DIR_PATH` environment variable.
 
         Parameters
         ----------
         ipython_dir : unicode or str
             The IPython directory to use.
         profile : unicode or str
             The name of the profile.  The name of the cluster directory
             will be "cluster_<profile>".
         """
         dirname = u'cluster_' + profile
         cluster_dir_paths = os.environ.get('IPCLUSTER_DIR_PATH','')
         if cluster_dir_paths:
             cluster_dir_paths = cluster_dir_paths.split(':')
         else:
             cluster_dir_paths = []
         paths = [os.getcwd(), ipython_dir] + cluster_dir_paths
         for p in paths:
             cluster_dir = os.path.join(p, dirname)
             if os.path.isdir(cluster_dir):
                 return ClusterDir(cluster_dir)
         else:
             raise ClusterDirError('Cluster directory not found in paths: %s' % dirname)
 
     @classmethod
     def find_cluster_dir(cls, cluster_dir):
         """Find/create a cluster dir and return its ClusterDir.
 
         This will create the cluster directory if it doesn't exist.
 
         Parameters
         ----------
         cluster_dir : unicode or str
             The path of the cluster directory.  This is expanded using
             :func:`IPython.utils.genutils.expand_path`.
         """
         cluster_dir = genutils.expand_path(cluster_dir)
         if not os.path.isdir(cluster_dir):
             raise ClusterDirError('Cluster directory not found: %s' % cluster_dir)
         return ClusterDir(cluster_dir)
 
 
 class AppWithClusterDirArgParseConfigLoader(ArgParseConfigLoader):
     """Default command line options for IPython cluster applications."""
 
     def _add_other_arguments(self):
         self.parser.add_argument('--ipython-dir', 
             dest='Global.ipython_dir',type=unicode,
             help='Set to override default location of Global.ipython_dir.',
             default=NoConfigDefault,
             metavar='Global.ipython_dir'
         )
         self.parser.add_argument('-p', '--profile',
             dest='Global.profile',type=unicode,
             help='The string name of the profile to be used. This determines '
             'the name of the cluster dir as: cluster_<profile>. The default profile '
             'is named "default".  The cluster directory is resolve this way '
             'if the --cluster-dir option is not used.',
             default=NoConfigDefault,
             metavar='Global.profile'
         )
         self.parser.add_argument('--log-level',
             dest="Global.log_level",type=int,
             help='Set the log level (0,10,20,30,40,50).  Default is 30.',
             default=NoConfigDefault,
             metavar="Global.log_level"
         )
         self.parser.add_argument('--cluster-dir',
             dest='Global.cluster_dir',type=unicode,
             help='Set the cluster dir. This overrides the logic used by the '
             '--profile option.',
             default=NoConfigDefault,
             metavar='Global.cluster_dir'
         ),
         self.parser.add_argument('--work-dir',
             dest='Global.work_dir',type=unicode,
             help='Set the working dir for the process.',
             default=NoConfigDefault,
             metavar='Global.work_dir'
         )
         self.parser.add_argument('--clean-logs',
             dest='Global.clean_logs', action='store_true',
             help='Delete old log flies before starting.',
             default=NoConfigDefault
         )
         self.parser.add_argument('--no-clean-logs',
             dest='Global.clean_logs', action='store_false',
             help="Don't Delete old log flies before starting.",
             default=NoConfigDefault
         )
 
 class ApplicationWithClusterDir(Application):
     """An application that puts everything into a cluster directory.
 
     Instead of looking for things in the ipython_dir, this type of application
     will use its own private directory called the "cluster directory"
     for things like config files, log files, etc.
 
     The cluster directory is resolved as follows:
 
     * If the ``--cluster-dir`` option is given, it is used.
     * If ``--cluster-dir`` is not given, the application directory is 
       resolve using the profile name as ``cluster_<profile>``. The search 
       path for this directory is then i) cwd if it is found there
       and ii) in ipython_dir otherwise.
 
     The config file for the application is to be put in the cluster
     dir and named the value of the ``config_file_name`` class attribute.
     """
 
     auto_create_cluster_dir = True
 
     def create_default_config(self):
         super(ApplicationWithClusterDir, self).create_default_config()
         self.default_config.Global.profile = u'default'
         self.default_config.Global.cluster_dir = u''
         self.default_config.Global.work_dir = os.getcwd()
         self.default_config.Global.log_to_file = False
         self.default_config.Global.clean_logs = False
 
     def create_command_line_config(self):
         """Create and return a command line config loader."""
         return AppWithClusterDirArgParseConfigLoader(
             description=self.description, 
             version=release.version
         )
 
     def find_resources(self):
         """This resolves the cluster directory.
 
         This tries to find the cluster directory and if successful, it will
         have done:
         * Sets ``self.cluster_dir_obj`` to the :class:`ClusterDir` object for 
           the application.
         * Sets ``self.cluster_dir`` attribute of the application and config
           objects.
 
         The algorithm used for this is as follows:
         1. Try ``Global.cluster_dir``.
         2. Try using ``Global.profile``.
         3. If both of these fail and ``self.auto_create_cluster_dir`` is
            ``True``, then create the new cluster dir in the IPython directory.
         4. If all fails, then raise :class:`ClusterDirError`.
         """
 
         try:
             cluster_dir = self.command_line_config.Global.cluster_dir
         except AttributeError:
             cluster_dir = self.default_config.Global.cluster_dir
         cluster_dir = genutils.expand_path(cluster_dir)
         try:
             self.cluster_dir_obj = ClusterDir.find_cluster_dir(cluster_dir)
         except ClusterDirError:
             pass
         else:
             self.log.info('Using existing cluster dir: %s' % \
                 self.cluster_dir_obj.location
             )
             self.finish_cluster_dir()
             return
 
         try:
             self.profile = self.command_line_config.Global.profile
         except AttributeError:
             self.profile = self.default_config.Global.profile
         try:
             self.cluster_dir_obj = ClusterDir.find_cluster_dir_by_profile(
                 self.ipython_dir, self.profile)
         except ClusterDirError:
             pass
         else:
             self.log.info('Using existing cluster dir: %s' % \
                 self.cluster_dir_obj.location
             )
             self.finish_cluster_dir()
             return
 
         if self.auto_create_cluster_dir:
             self.cluster_dir_obj = ClusterDir.create_cluster_dir_by_profile(
                 self.ipython_dir, self.profile
             )
             self.log.info('Creating new cluster dir: %s' % \
                 self.cluster_dir_obj.location
             )
             self.finish_cluster_dir()
         else:
             raise ClusterDirError('Could not find a valid cluster directory.')
 
     def finish_cluster_dir(self):
         # Set the cluster directory
         self.cluster_dir = self.cluster_dir_obj.location
         
         # These have to be set because they could be different from the one
         # that we just computed.  Because command line has the highest
         # priority, this will always end up in the master_config.
         self.default_config.Global.cluster_dir = self.cluster_dir
         self.command_line_config.Global.cluster_dir = self.cluster_dir
 
         # Set the search path to the cluster directory
         self.config_file_paths = (self.cluster_dir,)
 
     def find_config_file_name(self):
         """Find the config file name for this application."""
         # For this type of Application it should be set as a class attribute.
         if not hasattr(self, 'config_file_name'):
             self.log.critical("No config filename found")
 
     def find_config_file_paths(self):
         # Set the search path to the cluster directory
         self.config_file_paths = (self.cluster_dir,)
 
     def pre_construct(self):
         # The log and security dirs were set earlier, but here we put them
         # into the config and log them.
         config = self.master_config
         sdir = self.cluster_dir_obj.security_dir
         self.security_dir = config.Global.security_dir = sdir
         ldir = self.cluster_dir_obj.log_dir
         self.log_dir = config.Global.log_dir = ldir
         pdir = self.cluster_dir_obj.pid_dir
         self.pid_dir = config.Global.pid_dir = pdir
         self.log.info("Cluster directory set to: %s" % self.cluster_dir)
         config.Global.work_dir = unicode(genutils.expand_path(config.Global.work_dir))
         # Change to the working directory. We do this just before construct
         # is called so all the components there have the right working dir.
         self.to_work_dir()
 
     def to_work_dir(self):
         wd = self.master_config.Global.work_dir
         if unicode(wd) != unicode(os.getcwd()):
             os.chdir(wd)
             self.log.info("Changing to working dir: %s" % wd)
 
     def start_logging(self):
         # Remove old log files
         if self.master_config.Global.clean_logs:
             log_dir = self.master_config.Global.log_dir
             for f in os.listdir(log_dir):
                 if f.startswith(self.name + u'-') and f.endswith('.log'):
                     os.remove(os.path.join(log_dir, f))
         # Start logging to the new log file
         if self.master_config.Global.log_to_file:
             log_filename = self.name + u'-' + str(os.getpid()) + u'.log'
             logfile = os.path.join(self.log_dir, log_filename)
             open_log_file = open(logfile, 'w')
         else:
             open_log_file = sys.stdout
         log.startLogging(open_log_file)
 
     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.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.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.pid_dir, self.name + u'.pid')
         if os.path.isfile(pid_file):
             with open(pid_file, 'r') as f:
                 pid = int(f.read().strip())
                 return pid
         else:
             raise PIDFileError('pid file not found: %s' % pid_file)
 
 

IPython/kernel/ipclusterapp.py

@@ -1,455 +1,459 b''
 #!/usr/bin/env python
 # encoding: utf-8
 """
 The ipcluster application.
 """
 
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2008-2009  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 signal
 import sys
 
 if os.name=='posix':
     from twisted.scripts._twistd_unix import daemonize
 
 from IPython.core import release
 from IPython.external import argparse
 from IPython.config.loader import ArgParseConfigLoader, NoConfigDefault
 from IPython.utils.importstring import import_item
 
 from IPython.kernel.clusterdir import (
     ApplicationWithClusterDir, ClusterDirError, PIDFileError
 )
 
 from twisted.internet import reactor, defer
 from twisted.python import log, failure
 
 
 #-----------------------------------------------------------------------------
 # The ipcluster application
 #-----------------------------------------------------------------------------
 
 
 # Exit codes for ipcluster
 
 # This will be the exit code if the ipcluster appears to be running because
 # a .pid file exists
 ALREADY_STARTED = 10
 
 # This will be the exit code if ipcluster stop is run, but there is not .pid
 # file to be found.
 ALREADY_STOPPED = 11
 
 
 class IPClusterCLLoader(ArgParseConfigLoader):
 
     def _add_arguments(self):
         # This has all the common options that all subcommands use
         parent_parser1 = argparse.ArgumentParser(add_help=False)
         parent_parser1.add_argument('--ipython-dir', 
             dest='Global.ipython_dir',type=unicode,
             help='Set to override default location of Global.ipython_dir.',
             default=NoConfigDefault,
             metavar='Global.ipython_dir')
         parent_parser1.add_argument('--log-level',
             dest="Global.log_level",type=int,
             help='Set the log level (0,10,20,30,40,50).  Default is 30.',
             default=NoConfigDefault,
             metavar='Global.log_level')
 
         # This has all the common options that other subcommands use
         parent_parser2 = argparse.ArgumentParser(add_help=False)
         parent_parser2.add_argument('-p','--profile',
             dest='Global.profile',type=unicode,
             help='The string name of the profile to be used. This determines '
             'the name of the cluster dir as: cluster_<profile>. The default profile '
             'is named "default".  The cluster directory is resolve this way '
             'if the --cluster-dir option is not used.',
             default=NoConfigDefault,
             metavar='Global.profile')
         parent_parser2.add_argument('--cluster-dir',
             dest='Global.cluster_dir',type=unicode,
             help='Set the cluster dir. This overrides the logic used by the '
             '--profile option.',
             default=NoConfigDefault,
             metavar='Global.cluster_dir'),
         parent_parser2.add_argument('--work-dir',
             dest='Global.work_dir',type=unicode,
             help='Set the working dir for the process.',
             default=NoConfigDefault,
             metavar='Global.work_dir')
         parent_parser2.add_argument('--log-to-file',
             action='store_true', dest='Global.log_to_file', 
             default=NoConfigDefault,
             help='Log to a file in the log directory (default is stdout)'
         )
 
         subparsers = self.parser.add_subparsers(
             dest='Global.subcommand',
             title='ipcluster subcommands',
             description='ipcluster has a variety of subcommands. '
             'The general way of running ipcluster is "ipcluster <cmd> '
             ' [options]""',
             help='For more help, type "ipcluster <cmd> -h"')
 
         parser_list = subparsers.add_parser(
             'list',
             help='List all clusters in cwd and ipython_dir.',
             parents=[parent_parser1]
         )
 
         parser_create = subparsers.add_parser(
             'create',
             help='Create a new cluster directory.',
             parents=[parent_parser1, parent_parser2] 
         )
         parser_create.add_argument(
             '--reset-config',
             dest='Global.reset_config', action='store_true',
             default=NoConfigDefault,
             help='Recopy the default config files to the cluster directory. '
             'You will loose any modifications you have made to these files.'
         )
 
         parser_start = subparsers.add_parser(
             'start',
             help='Start a cluster.',
             parents=[parent_parser1, parent_parser2]
         )
         parser_start.add_argument(
             '-n', '--number',
             type=int, dest='Global.n',
             default=NoConfigDefault,
             help='The number of engines to start.',
             metavar='Global.n'
         )
         parser_start.add_argument('--clean-logs',
             dest='Global.clean_logs', action='store_true',
             help='Delete old log flies before starting.',
             default=NoConfigDefault
         )
         parser_start.add_argument('--no-clean-logs',
             dest='Global.clean_logs', action='store_false',
             help="Don't delete old log flies before starting.",
             default=NoConfigDefault
         )
         parser_start.add_argument('--daemon',
             dest='Global.daemonize', action='store_true',
             help='Daemonize the ipcluster program. This implies --log-to-file',
             default=NoConfigDefault
         )
         parser_start.add_argument('--no-daemon',
             dest='Global.daemonize', action='store_false',
             help="Dont't daemonize the ipcluster program.",
             default=NoConfigDefault
         )
 
         parser_start = subparsers.add_parser(
             'stop',
             help='Stop a cluster.',
             parents=[parent_parser1, parent_parser2]
         )
         parser_start.add_argument('--signal',
             dest='Global.signal', type=int,
             help="The signal number to use in stopping the cluster (default=2).",
             metavar="Global.signal",
             default=NoConfigDefault
         )
 
 
 default_config_file_name = u'ipcluster_config.py'
 
 
 class IPClusterApp(ApplicationWithClusterDir):
 
     name = u'ipcluster'
     description = 'Start an IPython cluster (controller and engines).'
     config_file_name = default_config_file_name
     default_log_level = logging.INFO
     auto_create_cluster_dir = False
 
     def create_default_config(self):
         super(IPClusterApp, self).create_default_config()
         self.default_config.Global.controller_launcher = \
             'IPython.kernel.launcher.LocalControllerLauncher'
         self.default_config.Global.engine_launcher = \
             'IPython.kernel.launcher.LocalEngineSetLauncher'
         self.default_config.Global.n = 2
         self.default_config.Global.reset_config = False
         self.default_config.Global.clean_logs = True
         self.default_config.Global.signal = 2
         self.default_config.Global.daemonize = False
 
     def create_command_line_config(self):
         """Create and return a command line config loader."""
         return IPClusterCLLoader(
             description=self.description, 
             version=release.version
         )
 
     def find_resources(self):
         subcommand = self.command_line_config.Global.subcommand
         if subcommand=='list':
             self.list_cluster_dirs()
             # Exit immediately because there is nothing left to do.
             self.exit()
         elif subcommand=='create':
             self.auto_create_cluster_dir = True
             super(IPClusterApp, self).find_resources()
         elif subcommand=='start' or subcommand=='stop':
-            self.auto_create_cluster_dir = False
+            self.auto_create_cluster_dir = True
             try:
                 super(IPClusterApp, self).find_resources()
             except ClusterDirError:
                 raise ClusterDirError(
                     "Could not find a cluster directory. A cluster dir must "
                     "be created before running 'ipcluster start'.  Do "
                     "'ipcluster create -h' or 'ipcluster list -h' for more "
                     "information about creating and listing cluster dirs."
                 )
 
     def list_cluster_dirs(self):
         # Find the search paths
         cluster_dir_paths = os.environ.get('IPCLUSTER_DIR_PATH','')
         if cluster_dir_paths:
             cluster_dir_paths = cluster_dir_paths.split(':')
         else:
             cluster_dir_paths = []
         try:
             ipython_dir = self.command_line_config.Global.ipython_dir
         except AttributeError:
             ipython_dir = self.default_config.Global.ipython_dir
         paths = [os.getcwd(), ipython_dir] + \
             cluster_dir_paths
         paths = list(set(paths))
 
         self.log.info('Searching for cluster dirs in paths: %r' % paths)
         for path in paths:
             files = os.listdir(path)
             for f in files:
                 full_path = os.path.join(path, f)
                 if os.path.isdir(full_path) and f.startswith('cluster_'):
                     profile = full_path.split('_')[-1]
                     start_cmd = 'ipcluster start -p %s -n 4' % profile
                     print start_cmd + " ==> " + full_path
 
     def pre_construct(self):
         # IPClusterApp.pre_construct() is where we cd to the working directory.
         super(IPClusterApp, self).pre_construct()
         config = self.master_config
         try:
             daemon = config.Global.daemonize
             if daemon:
                 config.Global.log_to_file = True
         except AttributeError:
             pass
 
     def construct(self):
         config = self.master_config
-        if config.Global.subcommand=='list':
-            pass
-        elif config.Global.subcommand=='create':
+        subcmd = config.Global.subcommand
+        reset = config.Global.reset_config
+        if subcmd == 'list':
+            return
+        if subcmd == 'create':
             self.log.info('Copying default config files to cluster directory '
-            '[overwrite=%r]' % (config.Global.reset_config,))
-            self.cluster_dir_obj.copy_all_config_files(overwrite=config.Global.reset_config)
-        elif config.Global.subcommand=='start':
+            '[overwrite=%r]' % (reset,))
+            self.cluster_dir_obj.copy_all_config_files(overwrite=reset)
+        if subcmd =='start':
+            self.cluster_dir_obj.copy_all_config_files(overwrite=False)
             self.start_logging()
             reactor.callWhenRunning(self.start_launchers)
 
     def start_launchers(self):
         config = self.master_config
 
         # Create the launchers. In both bases, we set the work_dir of
         # the launcher to the cluster_dir. This is where the launcher's
         # subprocesses will be launched. It is not where the controller
         # and engine will be launched.
         el_class = import_item(config.Global.engine_launcher)
         self.engine_launcher = el_class(
             work_dir=self.cluster_dir, config=config
         )
         cl_class = import_item(config.Global.controller_launcher)
         self.controller_launcher = cl_class(
             work_dir=self.cluster_dir, config=config
         )
 
         # Setup signals
         signal.signal(signal.SIGINT, self.sigint_handler)
 
         # Setup the observing of stopping. If the controller dies, shut
         # everything down as that will be completely fatal for the engines.
         d1 = self.controller_launcher.observe_stop()
         d1.addCallback(self.stop_launchers)
         # But, we don't monitor the stopping of engines. An engine dying
         # is just fine and in principle a user could start a new engine.
         # Also, if we did monitor engine stopping, it is difficult to
         # know what to do when only some engines die. Currently, the
         # observing of engine stopping is inconsistent. Some launchers
         # might trigger on a single engine stopping, other wait until
         # all stop.  TODO: think more about how to handle this.
         
         # Start the controller and engines
         self._stopping = False  # Make sure stop_launchers is not called 2x.
         d = self.start_controller()
         d.addCallback(self.start_engines)
         d.addCallback(self.startup_message)
         # If the controller or engines fail to start, stop everything
         d.addErrback(self.stop_launchers)
         return d
 
     def startup_message(self, r=None):
         log.msg("IPython cluster: started")
         return r
         
     def start_controller(self, r=None):
         # log.msg("In start_controller")
         config = self.master_config
         d = self.controller_launcher.start(
             cluster_dir=config.Global.cluster_dir
         )
         return d
             
     def start_engines(self, r=None):
         # log.msg("In start_engines")
         config = self.master_config
         d = self.engine_launcher.start(
             config.Global.n,
             cluster_dir=config.Global.cluster_dir
         )
         return d
 
     def stop_controller(self, r=None):
         # log.msg("In stop_controller")
         if self.controller_launcher.running:
             d = self.controller_launcher.stop()
             d.addErrback(self.log_err)
             return d
         else:
             return defer.succeed(None)
 
     def stop_engines(self, r=None):
         # log.msg("In stop_engines")
         if self.engine_launcher.running:
             d = self.engine_launcher.stop()
             d.addErrback(self.log_err)
             return d
         else:
             return defer.succeed(None)
 
     def log_err(self, f):
         log.msg(f.getTraceback())
         return None
         
     def stop_launchers(self, r=None):
         if not self._stopping:
             self._stopping = True
             if isinstance(r, failure.Failure):
                 log.msg('Unexpected error in ipcluster:')
                 log.msg(r.getTraceback())
             log.msg("IPython cluster: stopping")
             d= self.stop_engines()
             d2 = self.stop_controller()
             # Wait a few seconds to let things shut down.
             reactor.callLater(4.0, reactor.stop)
 
     def sigint_handler(self, signum, frame):
         self.stop_launchers()
         
     def start_logging(self):
         # Remove old log files of the controller and engine
         if self.master_config.Global.clean_logs:
             log_dir = self.master_config.Global.log_dir
             for f in os.listdir(log_dir):
                 if f.startswith('ipengine' + '-'):
                     if f.endswith('.log') or f.endswith('.out') or f.endswith('.err'):
                         os.remove(os.path.join(log_dir, f))
                 if f.startswith('ipcontroller' + '-'):
                     if f.endswith('.log') or f.endswith('.out') or f.endswith('.err'):
                         os.remove(os.path.join(log_dir, f))
         # This will remote old log files for ipcluster itself
         super(IPClusterApp, self).start_logging()
 
     def start_app(self):
         """Start the application, depending on what subcommand is used."""
         subcmd = self.master_config.Global.subcommand
         if subcmd=='create' or subcmd=='list':
             return
         elif subcmd=='start':
             self.start_app_start()
         elif subcmd=='stop':
             self.start_app_stop()
 
     def start_app_start(self):
         """Start the app for the start subcommand."""
         config = self.master_config
         # First see if the cluster is already running
         try:
             pid = self.get_pid_from_file()
         except PIDFileError:
             pass
         else:
             self.log.critical(
                 'Cluster is already running with [pid=%s]. '
                 'use "ipcluster stop" to stop the cluster.' % pid
             )
             # Here I exit with a unusual exit status that other processes
             # can watch for to learn how I existed.
             self.exit(ALREADY_STARTED)
 
         # Now log and daemonize
         self.log.info(
             'Starting ipcluster with [daemon=%r]' % config.Global.daemonize
         )
+        # TODO: Get daemonize working on Windows or as a Windows Server.
         if config.Global.daemonize:
             if os.name=='posix':
                 daemonize()
 
         # Now write the new pid file AFTER our new forked pid is active.
         self.write_pid_file()
         reactor.addSystemEventTrigger('during','shutdown', self.remove_pid_file)
         reactor.run()
 
     def start_app_stop(self):
         """Start the app for the stop subcommand."""
         config = self.master_config
         try:
             pid = self.get_pid_from_file()
         except PIDFileError:
             self.log.critical(
                 'Problem reading pid file, cluster is probably not running.'
             )
             # Here I exit with a unusual exit status that other processes
             # can watch for to learn how I existed.
             self.exit(ALREADY_STOPPED)
         else:
             if os.name=='posix':
                 sig = config.Global.signal
                 self.log.info(
                     "Stopping cluster [pid=%r] with [signal=%r]" % (pid, sig)
                 )
                 os.kill(pid, sig)
             elif os.name=='nt':
                 # As of right now, we don't support daemonize on Windows, so
                 # stop will not do anything. Minimally, it should clean up the
                 # old .pid files.
                 self.remove_pid_file()
 
 def launch_new_instance():
     """Create and run the IPython cluster."""
     app = IPClusterApp()
     app.start()
 
 
 if __name__ == '__main__':
     launch_new_instance()
 

IPython/kernel/launcher.py

@@ -1,867 +1,869 b''
 #!/usr/bin/env python
 # encoding: utf-8
 """
 Facilities for launching IPython processes asynchronously.
 """
 
 #-----------------------------------------------------------------------------
 #  Copyright (C) 2008-2009  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 re
 import sys
 
 from IPython.core.component import Component
 from IPython.external import Itpl
 from IPython.utils.traitlets import Str, Int, List, Unicode, Enum
 from IPython.utils.platutils import find_cmd
 from IPython.kernel.twistedutil import gatherBoth, make_deferred, sleep_deferred
 from IPython.kernel.winhpcjob import (
     WinHPCJob, WinHPCTask,
     IPControllerTask, IPEngineTask,
     IPControllerJob, IPEngineSetJob
 )
 
 from twisted.internet import reactor, defer
 from twisted.internet.defer import inlineCallbacks
 from twisted.internet.protocol import ProcessProtocol
 from twisted.internet.utils import getProcessOutput
 from twisted.internet.error import ProcessDone, ProcessTerminated
 from twisted.python import log
 from twisted.python.failure import Failure
 
 #-----------------------------------------------------------------------------
 # Utilities
 #-----------------------------------------------------------------------------
 
 
 def find_controller_cmd():
     """Find the command line ipcontroller program in a cross platform way."""
     if sys.platform == 'win32':
         # This logic is needed because the ipcontroller script doesn't
         # always get installed in the same way or in the same location.
         from IPython.kernel import ipcontrollerapp
         script_location = ipcontrollerapp.__file__.replace('.pyc', '.py')
         # The -u option here turns on unbuffered output, which is required
         # on Win32 to prevent wierd conflict and problems with Twisted.
         # Also, use sys.executable to make sure we are picking up the 
         # right python exe.
         cmd = [sys.executable, '-u', script_location]
     else:
         # ipcontroller has to be on the PATH in this case.
         cmd = ['ipcontroller']
     return cmd
 
 
 def find_engine_cmd():
     """Find the command line ipengine program in a cross platform way."""
     if sys.platform == 'win32':
         # This logic is needed because the ipengine script doesn't
         # always get installed in the same way or in the same location.
         from IPython.kernel import ipengineapp
         script_location = ipengineapp.__file__.replace('.pyc', '.py')
         # The -u option here turns on unbuffered output, which is required
         # on Win32 to prevent wierd conflict and problems with Twisted.
         # Also, use sys.executable to make sure we are picking up the 
         # right python exe.
         cmd = [sys.executable, '-u', script_location]
     else:
         # ipcontroller has to be on the PATH in this case.
         cmd = ['ipengine']
     return cmd
 
 
 #-----------------------------------------------------------------------------
 # Base launchers and errors
 #-----------------------------------------------------------------------------
 
 
 class LauncherError(Exception):
     pass
 
 
 class ProcessStateError(LauncherError):
     pass
 
 
 class UnknownStatus(LauncherError):
     pass
 
 
 class BaseLauncher(Component):
     """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 cluster_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'')
 
     def __init__(self, work_dir, parent=None, name=None, config=None):
         super(BaseLauncher, self).__init__(parent, name, config)
         self.work_dir = work_dir
         self.state = 'before' # can be before, running, after
         self.stop_deferreds = []
         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.
 
         This must return a deferred that fires with information about the
         process starting (like a pid, job id, etc.).
         """
         return defer.fail(
             Failure(NotImplementedError(
                 'start must be implemented in a subclass')
             )
         )
 
     def stop(self):
         """Stop the process and notify observers of stopping.
 
         This must return a deferred that fires with information about the
         processing stopping, like errors that occur while the process is
         attempting to be shut down. This deferred won't fire when the process
         actually stops. To observe the actual process stopping, see
         :func:`observe_stop`.
         """
         return defer.fail(
             Failure(NotImplementedError(
                 'stop must be implemented in a subclass')
             )
         )
 
     def observe_stop(self):
         """Get a deferred that will fire when the process stops.
 
         The deferred will fire with data that contains information about
         the exit status of the process.
         """
         if self.state=='after':
             return defer.succeed(self.stop_data)
         else:
             d = defer.Deferred()
             self.stop_deferreds.append(d)
             return d
 
     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.
         """
 
         log.msg('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 all the deferreds from :func:`observe_stop`."""
 
         log.msg('Process %r stopped: %r' % (self.args[0], data))
         self.stop_data = data
         self.state = 'after'
         for i in range(len(self.stop_deferreds)):
             d = self.stop_deferreds.pop()
             d.callback(data)
         return data
 
     def signal(self, sig):
         """Signal the process.
 
         Return a semi-meaningless deferred after signaling the process.
 
         Parameters
         ----------
         sig : str or int
             'KILL', 'INT', etc., or any signal number
         """
         return defer.fail(
             Failure(NotImplementedError(
                 'signal must be implemented in a subclass')
             )
         )
 
 
 #-----------------------------------------------------------------------------
 # Local process launchers
 #-----------------------------------------------------------------------------
 
 
 class LocalProcessLauncherProtocol(ProcessProtocol):
     """A ProcessProtocol to go with the LocalProcessLauncher."""
 
     def __init__(self, process_launcher):
         self.process_launcher = process_launcher
         self.pid = None
 
     def connectionMade(self):
         self.pid = self.transport.pid
         self.process_launcher.notify_start(self.transport.pid)
 
     def processEnded(self, status):
         value = status.value
         if isinstance(value, ProcessDone):
             self.process_launcher.notify_stop(
                 {'exit_code':0,
                  'signal':None,
                  'status':None,
                  'pid':self.pid
                 }
             )
         elif isinstance(value, ProcessTerminated):
             self.process_launcher.notify_stop(
                 {'exit_code':value.exitCode,
                  'signal':value.signal,
                  'status':value.status,
                  'pid':self.pid
                 }
             )
         else:
             raise UnknownStatus("Unknown exit status, this is probably a "
                                 "bug in Twisted")
 
     def outReceived(self, data):
         log.msg(data)
 
     def errReceived(self, data):
         log.err(data)
 
 
 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([])
 
     def __init__(self, work_dir, parent=None, name=None, config=None):
         super(LocalProcessLauncher, self).__init__(
             work_dir, parent, name, config
         )
         self.process_protocol = None
         self.start_deferred = None
 
     def find_args(self):
         return self.cmd_and_args
 
     def start(self):
         if self.state == 'before':
             self.process_protocol = LocalProcessLauncherProtocol(self)
             self.start_deferred = defer.Deferred()
             self.process_transport = reactor.spawnProcess(
                 self.process_protocol,
                 str(self.args[0]),  # twisted expects these to be str, not unicode
                 [str(a) for a in self.args],  # str expected, not unicode
                 env=os.environ,
                 path=self.work_dir  # start in the work_dir
             )
             return self.start_deferred
         else:
             s = 'The process was already started and has state: %r' % self.state
             return defer.fail(ProcessStateError(s))
 
     def notify_start(self, data):
         super(LocalProcessLauncher, self).notify_start(data)
         self.start_deferred.callback(data)
 
     def stop(self):
         return self.interrupt_then_kill()
 
     @make_deferred
     def signal(self, sig):
         if self.state == 'running':
             self.process_transport.signalProcess(sig)
 
     @inlineCallbacks
     def interrupt_then_kill(self, delay=2.0):
         """Send INT, wait a delay and then send KILL."""
         yield self.signal('INT')
         yield sleep_deferred(delay)
         yield self.signal('KILL')
 
 
 class LocalControllerLauncher(LocalProcessLauncher):
     """Launch a controller as a regular external process."""
 
     controller_cmd = List(find_controller_cmd(), config=True)
     # Command line arguments to ipcontroller.
     controller_args = List(['--log-to-file','--log-level', '40'], config=True)
 
     def find_args(self):
         return self.controller_cmd + self.controller_args
 
     def start(self, cluster_dir):
         """Start the controller by cluster_dir."""
         self.controller_args.extend(['--cluster-dir', cluster_dir])
         self.cluster_dir = unicode(cluster_dir)
         log.msg("Starting LocalControllerLauncher: %r" % self.args)
         return super(LocalControllerLauncher, self).start()
 
 
 class LocalEngineLauncher(LocalProcessLauncher):
     """Launch a single engine as a regular externall process."""
 
     engine_cmd = List(find_engine_cmd(), config=True)
     # Command line arguments for ipengine.
     engine_args = List(
         ['--log-to-file','--log-level', '40'], config=True
     )
 
     def find_args(self):
         return self.engine_cmd + self.engine_args
 
     def start(self, cluster_dir):
         """Start the engine by cluster_dir."""
         self.engine_args.extend(['--cluster-dir', cluster_dir])
         self.cluster_dir = unicode(cluster_dir)
         return super(LocalEngineLauncher, self).start()
 
 
 class LocalEngineSetLauncher(BaseLauncher):
     """Launch a set of engines as regular external processes."""
 
     # Command line arguments for ipengine.
     engine_args = List(
         ['--log-to-file','--log-level', '40'], config=True
     )
 
     def __init__(self, work_dir, parent=None, name=None, config=None):
         super(LocalEngineSetLauncher, self).__init__(
             work_dir, parent, name, config
         )
         self.launchers = []
 
     def start(self, n, cluster_dir):
         """Start n engines by profile or cluster_dir."""
         self.cluster_dir = unicode(cluster_dir)
         dlist = []
         for i in range(n):
             el = LocalEngineLauncher(self.work_dir, self)
             # Copy the engine args over to each engine launcher.
             import copy
             el.engine_args = copy.deepcopy(self.engine_args)
             d = el.start(cluster_dir)
             if i==0:
                 log.msg("Starting LocalEngineSetLauncher: %r" % el.args)
             self.launchers.append(el)
             dlist.append(d)
         # The consumeErrors here could be dangerous
         dfinal = gatherBoth(dlist, consumeErrors=True)
         dfinal.addCallback(self.notify_start)
         return dfinal
 
     def find_args(self):
         return ['engine set']
 
     def signal(self, sig):
         dlist = []
         for el in self.launchers:
             d = el.signal(sig)
             dlist.append(d)
         dfinal = gatherBoth(dlist, consumeErrors=True)
         return dfinal
 
     def interrupt_then_kill(self, delay=1.0):
         dlist = []
         for el in self.launchers:
             d = el.interrupt_then_kill(delay)
             dlist.append(d)
         dfinal = gatherBoth(dlist, consumeErrors=True)
         return dfinal
 
     def stop(self):
         return self.interrupt_then_kill()
 
     def observe_stop(self):
         dlist = [el.observe_stop() for el in self.launchers]
         dfinal = gatherBoth(dlist, consumeErrors=False)
         dfinal.addCallback(self.notify_stop)
         return dfinal
 
 
 #-----------------------------------------------------------------------------
 # MPIExec launchers
 #-----------------------------------------------------------------------------
 
 
 class MPIExecLauncher(LocalProcessLauncher):
     """Launch an external process using mpiexec."""
 
     # The mpiexec command to use in starting the process.
     mpi_cmd = List(['mpiexec'], config=True)
     # The command line arguments to pass to mpiexec.
     mpi_args = List([], config=True)
     # The program to start using mpiexec.
     program = List(['date'], config=True)
     # The command line argument to the program.
     program_args = List([], config=True)
     # The number of instances of the program to start.
     n = Int(1, config=True)
 
     def find_args(self):
         """Build self.args using all the fields."""
         return self.mpi_cmd + ['-n', 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(MPIExecLauncher, self).start()
 
 
 class MPIExecControllerLauncher(MPIExecLauncher):
     """Launch a controller using mpiexec."""
 
     controller_cmd = List(find_controller_cmd(), config=True)
     # Command line arguments to ipcontroller.
     controller_args = List(['--log-to-file','--log-level', '40'], config=True)
     n = Int(1, config=False)
 
     def start(self, cluster_dir):
         """Start the controller by cluster_dir."""
         self.controller_args.extend(['--cluster-dir', cluster_dir])
         self.cluster_dir = unicode(cluster_dir)
         log.msg("Starting MPIExecControllerLauncher: %r" % self.args)
         return super(MPIExecControllerLauncher, self).start(1)
 
     def find_args(self):
         return self.mpi_cmd + ['-n', self.n] + self.mpi_args + \
                self.controller_cmd + self.controller_args
 
 
 class MPIExecEngineSetLauncher(MPIExecLauncher):
 
     engine_cmd = List(find_engine_cmd(), config=True)
     # Command line arguments for ipengine.
     engine_args = List(
         ['--log-to-file','--log-level', '40'], config=True
     )
     n = Int(1, config=True)
 
     def start(self, n, cluster_dir):
         """Start n engines by profile or cluster_dir."""
         self.engine_args.extend(['--cluster-dir', cluster_dir])
         self.cluster_dir = unicode(cluster_dir)
         self.n = n
         log.msg('Starting MPIExecEngineSetLauncher: %r' % self.args)
         return super(MPIExecEngineSetLauncher, self).start(n)
 
     def find_args(self):
         return self.mpi_cmd + ['-n', self.n] + self.mpi_args + \
                self.engine_cmd + self.engine_args
 
 
 #-----------------------------------------------------------------------------
 # SSH launchers
 #-----------------------------------------------------------------------------
 
+# TODO: Get SSH Launcher working again.
 
 class SSHLauncher(BaseLauncher):
     """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)
     ssh_args = List([], config=True)
     program = List(['date'], config=True)
     program_args = List([], config=True)
     hostname = Str('', config=True)
     user = Str('', config=True)
     location = Str('')
 
     def _hostname_changed(self, name, old, new):
         self.location = '%s@%s' % (self.user, new)
 
     def _user_changed(self, name, old, new):
         self.location = '%s@%s' % (new, self.hostname)
 
     def find_args(self):
         return self.ssh_cmd + self.ssh_args + [self.location] + \
                self.program + self.program_args
 
     def start(self, n, hostname=None, user=None):
         if hostname is not None:
             self.hostname = hostname
         if user is not None:
             self.user = user
         return super(SSHLauncher, self).start()
 
 
 class SSHControllerLauncher(SSHLauncher):
     pass
 
 
 class SSHEngineSetLauncher(BaseLauncher):
     pass
 
 
 #-----------------------------------------------------------------------------
 # Windows HPC Server 2008 scheduler launchers
 #-----------------------------------------------------------------------------
 
 
 # This is only used on Windows.
 def find_job_cmd():
     if os.name=='nt':
         return find_cmd('job')
     else:
         return 'job'
 
 
 class WindowsHPCLauncher(BaseLauncher):
 
     # A regular expression used to get the job id from the output of the 
     # submit_command.
-    job_id_regexp = Str('\d+', config=True)
+    job_id_regexp = Str(r'\d+', config=True)
     # The filename of the instantiated job script.
     job_file_name = Unicode(u'ipython_job.xml', config=True)
     # 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'')
     # The hostname of the scheduler to submit the job to
     scheduler = Str('', config=True)
     job_cmd = Str(find_job_cmd(), config=True)
 
     def __init__(self, work_dir, parent=None, name=None, config=None):
         super(WindowsHPCLauncher, self).__init__(
             work_dir, parent, name, config
         )
 
     @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 ['job.exe']
         
     def parse_job_id(self, output):
         """Take the output of the submit command and return the job id."""
         m = re.search(self.job_id_regexp, 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
         log.msg('Job started with job id: %r' % job_id)
         return job_id
 
     @inlineCallbacks
     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
         ]
         log.msg("Starting Win HPC Job: %s" % (self.job_cmd + ' ' + ' '.join(args),))
         # Twisted will raise DeprecationWarnings if we try to pass unicode to this
         output = yield getProcessOutput(str(self.job_cmd),
             [str(a) for a in args],
             env=dict((str(k),str(v)) for k,v in os.environ.items()),
             path=self.work_dir
         )
         job_id = self.parse_job_id(output)
         self.notify_start(job_id)
         defer.returnValue(job_id)
 
     @inlineCallbacks
     def stop(self):
         args = [
             'cancel',
             self.job_id,
             '/scheduler:%s' % self.scheduler
         ]
         log.msg("Stopping Win HPC Job: %s" % (self.job_cmd + ' ' + ' '.join(args),))
         try:
             # Twisted will raise DeprecationWarnings if we try to pass unicode to this
             output = yield getProcessOutput(str(self.job_cmd),
                 [str(a) for a in args],
                 env=dict((str(k),str(v)) for k,v in os.environ.items()),
                 path=self.work_dir
             )
         except:
             output = 'The job already appears to be stoppped: %r' % self.job_id
         self.notify_stop(output)  # Pass the output of the kill cmd
         defer.returnValue(output)
 
 
 class WindowsHPCControllerLauncher(WindowsHPCLauncher):
 
     job_file_name = Unicode(u'ipcontroller_job.xml', config=True)
     extra_args = List([], config=False)
 
     def write_job_file(self, n):
         job = IPControllerJob(self)
 
         t = IPControllerTask(self)
         # 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.cluster_dir
         # Add the --cluster-dir and from self.start().
         t.controller_args.extend(self.extra_args)
         job.add_task(t)
 
         log.msg("Writing job description file: %s" % self.job_file)
         job.write(self.job_file)
 
     @property
     def job_file(self):
         return os.path.join(self.cluster_dir, self.job_file_name)
 
     def start(self, cluster_dir):
         """Start the controller by cluster_dir."""
         self.extra_args = ['--cluster-dir', cluster_dir]
         self.cluster_dir = unicode(cluster_dir)
         return super(WindowsHPCControllerLauncher, self).start(1)
 
 
 class WindowsHPCEngineSetLauncher(WindowsHPCLauncher):
 
     job_file_name = Unicode(u'ipengineset_job.xml', config=True)
     extra_args = List([], config=False)
 
     def write_job_file(self, n):
         job = IPEngineSetJob(self)
 
         for i in range(n):
             t = IPEngineTask(self)
             # 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.cluster_dir
             # Add the --cluster-dir and from self.start().
             t.engine_args.extend(self.extra_args)
             job.add_task(t)
 
         log.msg("Writing job description file: %s" % self.job_file)
         job.write(self.job_file)
 
     @property
     def job_file(self):
         return os.path.join(self.cluster_dir, self.job_file_name)
 
     def start(self, n, cluster_dir):
         """Start the controller by cluster_dir."""
         self.extra_args = ['--cluster-dir', cluster_dir]
         self.cluster_dir = unicode(cluster_dir)
         return super(WindowsHPCEngineSetLauncher, self).start(n)
 
 
 #-----------------------------------------------------------------------------
 # Batch (PBS) system launchers
 #-----------------------------------------------------------------------------
 
+# TODO: Get PBS launcher working again.
 
 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 Itpl. 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
     # The name of the command line program used to submit jobs.
     submit_command = Str('', config=True)
     # The name of the command line program used to delete jobs.
     delete_command = Str('', config=True)
     # A regular expression used to get the job id from the output of the 
     # submit_command.
     job_id_regexp = Str('', config=True)
     # The string that is the batch script template itself.
     batch_template = Str('', config=True)
     # The filename of the instantiated batch script.
     batch_file_name = Unicode(u'batch_script', config=True)
     # The full path to the instantiated batch script.
     batch_file = Unicode(u'')
 
     def __init__(self, work_dir, parent=None, name=None, config=None):
         super(BatchSystemLauncher, self).__init__(
             work_dir, parent, name, config
         )
         self.batch_file = os.path.join(self.work_dir, self.batch_file_name)
         self.context = {}
 
     def parse_job_id(self, output):
         """Take the output of the submit command and return the job id."""
         m = re.match(self.job_id_regexp, 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
         log.msg('Job started 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.context['n'] = n
         script_as_string = Itpl.itplns(self.batch_template, self.context)
         log.msg('Writing instantiated batch script: %s' % self.batch_file)
         f = open(self.batch_file, 'w')
         f.write(script_as_string)
         f.close()
 
     @inlineCallbacks
     def start(self, n):
         """Start n copies of the process using a batch system."""
         self.write_batch_script(n)
         output = yield getProcessOutput(self.submit_command,
             [self.batch_file], env=os.environ)
         job_id = self.parse_job_id(output)
         self.notify_start(job_id)
         defer.returnValue(job_id)
 
     @inlineCallbacks
     def stop(self):
         output = yield getProcessOutput(self.delete_command,
             [self.job_id], env=os.environ
         )
         self.notify_stop(output)  # Pass the output of the kill cmd
         defer.returnValue(output)
 
 
 class PBSLauncher(BatchSystemLauncher):
     """A BatchSystemLauncher subclass for PBS."""
 
     submit_command = Str('qsub', config=True)
     delete_command = Str('qdel', config=True)
-    job_id_regexp = Str('\d+', config=True)
+    job_id_regexp = Str(r'\d+', config=True)
     batch_template = Str('', config=True)
     batch_file_name = Unicode(u'pbs_batch_script', config=True)
     batch_file = Unicode(u'')
 
 
 class PBSControllerLauncher(PBSLauncher):
     """Launch a controller using PBS."""
 
     batch_file_name = Unicode(u'pbs_batch_script_controller', config=True)
 
     def start(self, cluster_dir):
         """Start the controller by profile or cluster_dir."""
         # Here we save profile and cluster_dir in the context so they
         # can be used in the batch script template as ${profile} and
         # ${cluster_dir}
         self.context['cluster_dir'] = cluster_dir
         self.cluster_dir = unicode(cluster_dir)
         log.msg("Starting PBSControllerLauncher: %r" % self.args)
         return super(PBSControllerLauncher, self).start(1)
 
 
 class PBSEngineSetLauncher(PBSLauncher):
 
     batch_file_name = Unicode(u'pbs_batch_script_engines', config=True)
 
     def start(self, n, cluster_dir):
         """Start n engines by profile or cluster_dir."""
         self.program_args.extend(['--cluster-dir', cluster_dir])
         self.cluster_dir = unicode(cluster_dir)
         log.msg('Starting PBSEngineSetLauncher: %r' % self.args)
         return super(PBSEngineSetLauncher, self).start(n)
 
 
 #-----------------------------------------------------------------------------
 # A launcher for ipcluster itself!
 #-----------------------------------------------------------------------------
 
 
 def find_ipcluster_cmd():
     """Find the command line ipcluster program in a cross platform way."""
     if sys.platform == 'win32':
         # This logic is needed because the ipcluster script doesn't
         # always get installed in the same way or in the same location.
         from IPython.kernel import ipclusterapp
         script_location = ipclusterapp.__file__.replace('.pyc', '.py')
         # The -u option here turns on unbuffered output, which is required
         # on Win32 to prevent wierd conflict and problems with Twisted.
         # Also, use sys.executable to make sure we are picking up the 
         # right python exe.
         cmd = [sys.executable, '-u', script_location]
     else:
         # ipcontroller has to be on the PATH in this case.
         cmd = ['ipcluster']
     return cmd
 
 
 class IPClusterLauncher(LocalProcessLauncher):
     """Launch the ipcluster program in an external process."""
 
     ipcluster_cmd = List(find_ipcluster_cmd(), config=True)
     # Command line arguments to pass to ipcluster.
     ipcluster_args = List(
         ['--clean-logs', '--log-to-file', '--log-level', '40'], config=True)
     ipcluster_subcommand = Str('start')
     ipcluster_n = Int(2)
 
     def find_args(self):
         return self.ipcluster_cmd + [self.ipcluster_subcommand] + \
             ['-n', repr(self.ipcluster_n)] + self.ipcluster_args
 
     def start(self):
         log.msg("Starting ipcluster: %r" % self.args)
         return super(IPClusterLauncher, self).start()