pythonw in py3k sets std{in,out,err} to None

The print statement works without error, even though the stdio file objects are instances
of None. Since they are instances of None, however, the encoding attribute will not
exist, so protect blind attribute access of std{in,out,err} by using a new function, get_stream_enc.

IPython/core/splitinput.py

@@ -1,138 +1,139 b''
 # encoding: utf-8
 """
 Simple utility for splitting user input. This is used by both inputsplitter and
 prefilter.
 
 Authors:
 
 * Brian Granger
 * Fernando Perez
 """
 
 #-----------------------------------------------------------------------------
 #  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 re
 import sys
 
 from IPython.utils import py3compat
 
 #-----------------------------------------------------------------------------
 # Main function
 #-----------------------------------------------------------------------------
 
 # RegExp for splitting line contents into pre-char//first word-method//rest.
 # For clarity, each group in on one line.
 
 # WARNING: update the regexp if the escapes in interactiveshell are changed, as
 # they are hardwired in.
 
 # Although it's not solely driven by the regex, note that:
 # ,;/% only trigger if they are the first character on the line
 # ! and !! trigger if they are first char(s) *or* follow an indent
 # ? triggers as first or last char.
 
 line_split = re.compile("""
              ^(\s*)               # any leading space
              ([,;/%]|!!?|\?\??)?  # escape character or characters
              \s*(%?[\w\.\*]*)     # function/method, possibly with leading %
                                   # to correctly treat things like '?%magic'
              (.*?$|$)             # rest of line
              """, re.VERBOSE)
 
 def split_user_input(line, pattern=None):
     """Split user input into initial whitespace, escape character, function part
     and the rest.
     """
     # We need to ensure that the rest of this routine deals only with unicode
-    line = py3compat.cast_unicode(line, sys.stdin.encoding or 'utf-8')
+    encoding = py3compat.get_stream_enc(sys.stdin, 'utf-8')
+    line = py3compat.cast_unicode(line, encoding)
 
     if pattern is None:
         pattern = line_split
     match = pattern.match(line)
     if not match:
         # print "match failed for line '%s'" % line
         try:
             ifun, the_rest = line.split(None,1)
         except ValueError:
             # print "split failed for line '%s'" % line
             ifun, the_rest = line, u''
         pre = re.match('^(\s*)(.*)',line).groups()[0]
         esc = ""
     else:
         pre, esc, ifun, the_rest = match.groups()
 
     #print 'line:<%s>' % line # dbg
     #print 'pre <%s> ifun <%s> rest <%s>' % (pre,ifun.strip(),the_rest) # dbg
     return pre, esc or '', ifun.strip(), the_rest.lstrip()
 
 class LineInfo(object):
     """A single line of input and associated info.
 
     Includes the following as properties:
 
     line
       The original, raw line
 
     continue_prompt
       Is this line a continuation in a sequence of multiline input?
 
     pre
       Any leading whitespace.
 
     esc
       The escape character(s) in pre or the empty string if there isn't one.
       Note that '!!' and '??' are possible values for esc. Otherwise it will
       always be a single character.
 
     ifun
       The 'function part', which is basically the maximal initial sequence
       of valid python identifiers and the '.' character. This is what is
       checked for alias and magic transformations, used for auto-calling,
       etc. In contrast to Python identifiers, it may start with "%" and contain
       "*".
 
     the_rest
       Everything else on the line.
     """
     def __init__(self, line, continue_prompt=False):
         self.line            = line
         self.continue_prompt = continue_prompt
         self.pre, self.esc, self.ifun, self.the_rest = split_user_input(line)
 
         self.pre_char       = self.pre.strip()
         if self.pre_char:
             self.pre_whitespace = '' # No whitespace allowd before esc chars
         else:
             self.pre_whitespace = self.pre
 
         self._oinfo = None
 
     def ofind(self, ip):
         """Do a full, attribute-walking lookup of the ifun in the various
         namespaces for the given IPython InteractiveShell instance.
 
         Return a dict with keys: found,obj,ospace,ismagic
 
         Note: can cause state changes because of calling getattr, but should
         only be run if autocall is on and if the line hasn't matched any
         other, less dangerous handlers.
 
         Does cache the results of the call, so can be called multiple times
         without worrying about *further* damaging state.
         """
         if not self._oinfo:
             # ip.shell._ofind is actually on the Magic class!
             self._oinfo = ip.shell._ofind(self.ifun)
         return self._oinfo
 
     def __str__(self):
         return "LineInfo [%s|%s|%s|%s]" %(self.pre, self.esc, self.ifun, self.the_rest)

IPython/frontend/terminal/interactiveshell.py

@@ -1,668 +1,668 b''
 # -*- coding: utf-8 -*-
 """Subclass of InteractiveShell for terminal based frontends."""
 
 #-----------------------------------------------------------------------------
 #  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
 #-----------------------------------------------------------------------------
 
 import __builtin__
 import bdb
 import os
 import re
 import sys
 import textwrap
 
 try:
     from contextlib import nested
 except:
     from IPython.utils.nested_context import nested
 
 from IPython.core.error import TryNext, UsageError
 from IPython.core.usage import interactive_usage, default_banner
 from IPython.core.interactiveshell import InteractiveShell, InteractiveShellABC
 from IPython.core.pylabtools import pylab_activate
 from IPython.testing.skipdoctest import skip_doctest
 from IPython.utils import py3compat
 from IPython.utils.terminal import toggle_set_term_title, set_term_title
 from IPython.utils.process import abbrev_cwd
 from IPython.utils.warn import warn, error
 from IPython.utils.text import num_ini_spaces, SList
 from IPython.utils.traitlets import Integer, CBool, Unicode
 
 #-----------------------------------------------------------------------------
 # Utilities
 #-----------------------------------------------------------------------------
 
 def get_default_editor():
     try:
         ed = os.environ['EDITOR']
     except KeyError:
         if os.name == 'posix':
             ed = 'vi'  # the only one guaranteed to be there!
         else:
             ed = 'notepad' # same in Windows!
     return ed
 
 
 def get_pasted_lines(sentinel, l_input=py3compat.input):
     """ Yield pasted lines until the user enters the given sentinel value.
     """
     print "Pasting code; enter '%s' alone on the line to stop or use Ctrl-D." \
           % sentinel
     while True:
         try:
             l = l_input(':')
             if l == sentinel:
                 return
             else:
                 yield l
         except EOFError:
             print '<EOF>'
             return
 
 
 def strip_email_quotes(raw_lines):
     """ Strip email quotation marks at the beginning of each line.
 
     We don't do any more input transofrmations here because the main shell's
     prefiltering handles other cases.
     """
     lines = [re.sub(r'^\s*(\s?>)+', '', l) for l in raw_lines]
     return '\n'.join(lines) + '\n'
 
 
 # These two functions are needed by the %paste/%cpaste magics.  In practice
 # they are basically methods (they take the shell as their first argument), but
 # we leave them as standalone functions because eventually the magics
 # themselves will become separate objects altogether.  At that point, the
 # magics will have access to the shell object, and these functions can be made
 # methods of the magic object, but not of the shell.
 
 def store_or_execute(shell, block, name):
     """ Execute a block, or store it in a variable, per the user's request.
     """
     # Dedent and prefilter so what we store matches what is executed by
     # run_cell.
     b = shell.prefilter(textwrap.dedent(block))
 
     if name:
         # If storing it for further editing, run the prefilter on it
         shell.user_ns[name] = SList(b.splitlines())
         print "Block assigned to '%s'" % name
     else:
         shell.user_ns['pasted_block'] = b
         shell.run_cell(b)
 
 
 def rerun_pasted(shell, name='pasted_block'):
     """ Rerun a previously pasted command.
     """
     b = shell.user_ns.get(name)
 
     # Sanity checks
     if b is None:
         raise UsageError('No previous pasted block available')
     if not isinstance(b, basestring):
         raise UsageError(
             "Variable 'pasted_block' is not a string, can't execute")
 
     print "Re-executing '%s...' (%d chars)"% (b.split('\n',1)[0], len(b))
     shell.run_cell(b)
 
 
 #-----------------------------------------------------------------------------
 # Main class
 #-----------------------------------------------------------------------------
 
 class TerminalInteractiveShell(InteractiveShell):
 
     autoedit_syntax = CBool(False, config=True,
         help="auto editing of files with syntax errors.")
     banner = Unicode('')
     banner1 = Unicode(default_banner, config=True,
         help="""The part of the banner to be printed before the profile"""
     )
     banner2 = Unicode('', config=True,
         help="""The part of the banner to be printed after the profile"""
     )
     confirm_exit = CBool(True, config=True,
         help="""
         Set to confirm when you try to exit IPython with an EOF (Control-D
         in Unix, Control-Z/Enter in Windows). By typing 'exit' or 'quit',
         you can force a direct exit without any confirmation.""",
     )
     # This display_banner only controls whether or not self.show_banner()
     # is called when mainloop/interact are called.  The default is False
     # because for the terminal based application, the banner behavior
     # is controlled by Global.display_banner, which IPythonApp looks at
     # to determine if *it* should call show_banner() by hand or not.
     display_banner = CBool(False) # This isn't configurable!
     embedded = CBool(False)
     embedded_active = CBool(False)
     editor = Unicode(get_default_editor(), config=True,
         help="Set the editor used by IPython (default to $EDITOR/vi/notepad)."
     )
     pager = Unicode('less', config=True,
         help="The shell program to be used for paging.")
 
     screen_length = Integer(0, config=True,
         help=
         """Number of lines of your screen, used to control printing of very
         long strings.  Strings longer than this number of lines will be sent
         through a pager instead of directly printed.  The default value for
         this is 0, which means IPython will auto-detect your screen size every
         time it needs to print certain potentially long strings (this doesn't
         change the behavior of the 'print' keyword, it's only triggered
         internally). If for some reason this isn't working well (it needs
         curses support), specify it yourself. Otherwise don't change the
         default.""",
     )
     term_title = CBool(False, config=True,
         help="Enable auto setting the terminal title."
     )
 
     # In the terminal, GUI control is done via PyOS_InputHook
     from IPython.lib.inputhook import enable_gui
     enable_gui = staticmethod(enable_gui)
     
     def __init__(self, config=None, ipython_dir=None, profile_dir=None,
                  user_ns=None, user_module=None, custom_exceptions=((),None),
                  usage=None, banner1=None, banner2=None, display_banner=None):
 
         super(TerminalInteractiveShell, self).__init__(
             config=config, profile_dir=profile_dir, user_ns=user_ns,
             user_module=user_module, custom_exceptions=custom_exceptions
         )
         # use os.system instead of utils.process.system by default,
         # because piped system doesn't make sense in the Terminal:
         self.system = self.system_raw
 
         self.init_term_title()
         self.init_usage(usage)
         self.init_banner(banner1, banner2, display_banner)
 
     #-------------------------------------------------------------------------
     # Things related to the terminal
     #-------------------------------------------------------------------------
 
     @property
     def usable_screen_length(self):
         if self.screen_length == 0:
             return 0
         else:
             num_lines_bot = self.separate_in.count('\n')+1
             return self.screen_length - num_lines_bot
 
     def init_term_title(self):
         # Enable or disable the terminal title.
         if self.term_title:
             toggle_set_term_title(True)
             set_term_title('IPython: ' + abbrev_cwd())
         else:
             toggle_set_term_title(False)
 
     #-------------------------------------------------------------------------
     # Things related to aliases
     #-------------------------------------------------------------------------
 
     def init_alias(self):
         # The parent class defines aliases that can be safely used with any
         # frontend.
         super(TerminalInteractiveShell, self).init_alias()
 
         # Now define aliases that only make sense on the terminal, because they
         # need direct access to the console in a way that we can't emulate in
         # GUI or web frontend
         if os.name == 'posix':
             aliases = [('clear', 'clear'), ('more', 'more'), ('less', 'less'),
                        ('man', 'man')]
         elif os.name == 'nt':
             aliases = [('cls', 'cls')]
 
 
         for name, cmd in aliases:
             self.alias_manager.define_alias(name, cmd)
 
     #-------------------------------------------------------------------------
     # Things related to the banner and usage
     #-------------------------------------------------------------------------
 
     def _banner1_changed(self):
         self.compute_banner()
 
     def _banner2_changed(self):
         self.compute_banner()
 
     def _term_title_changed(self, name, new_value):
         self.init_term_title()
 
     def init_banner(self, banner1, banner2, display_banner):
         if banner1 is not None:
             self.banner1 = banner1
         if banner2 is not None:
             self.banner2 = banner2
         if display_banner is not None:
             self.display_banner = display_banner
         self.compute_banner()
 
     def show_banner(self, banner=None):
         if banner is None:
             banner = self.banner
         self.write(banner)
 
     def compute_banner(self):
         self.banner = self.banner1
         if self.profile and self.profile != 'default':
             self.banner += '\nIPython profile: %s\n' % self.profile
         if self.banner2:
             self.banner += '\n' + self.banner2
 
     def init_usage(self, usage=None):
         if usage is None:
             self.usage = interactive_usage
         else:
             self.usage = usage
 
     #-------------------------------------------------------------------------
     # Mainloop and code execution logic
     #-------------------------------------------------------------------------
 
     def mainloop(self, display_banner=None):
         """Start the mainloop.
 
         If an optional banner argument is given, it will override the
         internally created default banner.
         """
 
         with nested(self.builtin_trap, self.display_trap):
 
             while 1:
                 try:
                     self.interact(display_banner=display_banner)
                     #self.interact_with_readline()
                     # XXX for testing of a readline-decoupled repl loop, call
                     # interact_with_readline above
                     break
                 except KeyboardInterrupt:
                     # this should not be necessary, but KeyboardInterrupt
                     # handling seems rather unpredictable...
                     self.write("\nKeyboardInterrupt in interact()\n")
 
     def _replace_rlhist_multiline(self, source_raw, hlen_before_cell):
         """Store multiple lines as a single entry in history"""
 
         # do nothing without readline or disabled multiline
         if not self.has_readline or not self.multiline_history:
             return hlen_before_cell
 
         # windows rl has no remove_history_item
         if not hasattr(self.readline, "remove_history_item"):
             return hlen_before_cell
 
         # skip empty cells
         if not source_raw.rstrip():
             return hlen_before_cell
 
         # nothing changed do nothing, e.g. when rl removes consecutive dups
         hlen = self.readline.get_current_history_length()
         if hlen == hlen_before_cell:
             return hlen_before_cell
 
         for i in range(hlen - hlen_before_cell):
             self.readline.remove_history_item(hlen - i - 1)
-        stdin_encoding = sys.stdin.encoding or "utf-8"
+        stdin_encoding = py3compat.get_stream_enc(sys.stdin, 'utf-8')
         self.readline.add_history(py3compat.unicode_to_str(source_raw.rstrip(),
                                     stdin_encoding))
         return self.readline.get_current_history_length()
 
     def interact(self, display_banner=None):
         """Closely emulate the interactive Python console."""
 
         # batch run -> do not interact
         if self.exit_now:
             return
 
         if display_banner is None:
             display_banner = self.display_banner
 
         if isinstance(display_banner, basestring):
             self.show_banner(display_banner)
         elif display_banner:
             self.show_banner()
 
         more = False
 
         if self.has_readline:
             self.readline_startup_hook(self.pre_readline)
             hlen_b4_cell = self.readline.get_current_history_length()
         else:
             hlen_b4_cell = 0
         # exit_now is set by a call to %Exit or %Quit, through the
         # ask_exit callback.
 
         while not self.exit_now:
             self.hooks.pre_prompt_hook()
             if more:
                 try:
                     prompt = self.prompt_manager.render('in2')
                 except:
                     self.showtraceback()
                 if self.autoindent:
                     self.rl_do_indent = True
 
             else:
                 try:
                     prompt = self.separate_in + self.prompt_manager.render('in')
                 except:
                     self.showtraceback()
             try:
                 line = self.raw_input(prompt)
                 if self.exit_now:
                     # quick exit on sys.std[in|out] close
                     break
                 if self.autoindent:
                     self.rl_do_indent = False
 
             except KeyboardInterrupt:
                 #double-guard against keyboardinterrupts during kbdint handling
                 try:
                     self.write('\nKeyboardInterrupt\n')
                     source_raw = self.input_splitter.source_raw_reset()[1]
                     hlen_b4_cell = \
                         self._replace_rlhist_multiline(source_raw, hlen_b4_cell)
                     more = False
                 except KeyboardInterrupt:
                     pass
             except EOFError:
                 if self.autoindent:
                     self.rl_do_indent = False
                     if self.has_readline:
                         self.readline_startup_hook(None)
                 self.write('\n')
                 self.exit()
             except bdb.BdbQuit:
                 warn('The Python debugger has exited with a BdbQuit exception.\n'
                      'Because of how pdb handles the stack, it is impossible\n'
                      'for IPython to properly format this particular exception.\n'
                      'IPython will resume normal operation.')
             except:
                 # exceptions here are VERY RARE, but they can be triggered
                 # asynchronously by signal handlers, for example.
                 self.showtraceback()
             else:
                 self.input_splitter.push(line)
                 more = self.input_splitter.push_accepts_more()
                 if (self.SyntaxTB.last_syntax_error and
                     self.autoedit_syntax):
                     self.edit_syntax_error()
                 if not more:
                     source_raw = self.input_splitter.source_raw_reset()[1]
                     self.run_cell(source_raw, store_history=True)
                     hlen_b4_cell = \
                         self._replace_rlhist_multiline(source_raw, hlen_b4_cell)
 
         # Turn off the exit flag, so the mainloop can be restarted if desired
         self.exit_now = False
 
     def raw_input(self, prompt=''):
         """Write a prompt and read a line.
 
         The returned line does not include the trailing newline.
         When the user enters the EOF key sequence, EOFError is raised.
 
         Optional inputs:
 
           - prompt(''): a string to be printed to prompt the user.
 
           - continue_prompt(False): whether this line is the first one or a
           continuation in a sequence of inputs.
         """
         # Code run by the user may have modified the readline completer state.
         # We must ensure that our completer is back in place.
 
         if self.has_readline:
             self.set_readline_completer()
 
         try:
             line = py3compat.str_to_unicode(self.raw_input_original(prompt))
         except ValueError:
             warn("\n********\nYou or a %run:ed script called sys.stdin.close()"
                  " or sys.stdout.close()!\nExiting IPython!")
             self.ask_exit()
             return ""
 
         # Try to be reasonably smart about not re-indenting pasted input more
         # than necessary.  We do this by trimming out the auto-indent initial
         # spaces, if the user's actual input started itself with whitespace.
         if self.autoindent:
             if num_ini_spaces(line) > self.indent_current_nsp:
                 line = line[self.indent_current_nsp:]
                 self.indent_current_nsp = 0
 
         return line
 
     #-------------------------------------------------------------------------
     # Methods to support auto-editing of SyntaxErrors.
     #-------------------------------------------------------------------------
 
     def edit_syntax_error(self):
         """The bottom half of the syntax error handler called in the main loop.
 
         Loop until syntax error is fixed or user cancels.
         """
 
         while self.SyntaxTB.last_syntax_error:
             # copy and clear last_syntax_error
             err = self.SyntaxTB.clear_err_state()
             if not self._should_recompile(err):
                 return
             try:
                 # may set last_syntax_error again if a SyntaxError is raised
                 self.safe_execfile(err.filename,self.user_ns)
             except:
                 self.showtraceback()
             else:
                 try:
                     f = open(err.filename)
                     try:
                         # This should be inside a display_trap block and I
                         # think it is.
                         sys.displayhook(f.read())
                     finally:
                         f.close()
                 except:
                     self.showtraceback()
 
     def _should_recompile(self,e):
         """Utility routine for edit_syntax_error"""
 
         if e.filename in ('<ipython console>','<input>','<string>',
                           '<console>','<BackgroundJob compilation>',
                           None):
 
             return False
         try:
             if (self.autoedit_syntax and
                 not self.ask_yes_no('Return to editor to correct syntax error? '
                               '[Y/n] ','y')):
                 return False
         except EOFError:
             return False
 
         def int0(x):
             try:
                 return int(x)
             except TypeError:
                 return 0
         # always pass integer line and offset values to editor hook
         try:
             self.hooks.fix_error_editor(e.filename,
                 int0(e.lineno),int0(e.offset),e.msg)
         except TryNext:
             warn('Could not open editor')
             return False
         return True
 
     #-------------------------------------------------------------------------
     # Things related to exiting
     #-------------------------------------------------------------------------
 
     def ask_exit(self):
         """ Ask the shell to exit. Can be overiden and used as a callback. """
         self.exit_now = True
 
     def exit(self):
         """Handle interactive exit.
 
         This method calls the ask_exit callback."""
         if self.confirm_exit:
             if self.ask_yes_no('Do you really want to exit ([y]/n)?','y'):
                 self.ask_exit()
         else:
             self.ask_exit()
 
     #------------------------------------------------------------------------
     # 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_autoindent(self, parameter_s = ''):
         """Toggle autoindent on/off (if available)."""
 
         self.shell.set_autoindent()
         print "Automatic indentation is:",['OFF','ON'][self.shell.autoindent]
 
     @skip_doctest
     def magic_cpaste(self, parameter_s=''):
         """Paste & execute a pre-formatted code block from clipboard.
 
         You must terminate the block with '--' (two minus-signs) or Ctrl-D
         alone on the line. You can also provide your own sentinel with '%paste
         -s %%' ('%%' is the new sentinel for this operation)
 
         The block is dedented prior to execution to enable execution of method
         definitions. '>' and '+' characters at the beginning of a line are
         ignored, to allow pasting directly from e-mails, diff files and
         doctests (the '...' continuation prompt is also stripped).  The
         executed block is also assigned to variable named 'pasted_block' for
         later editing with '%edit pasted_block'.
 
         You can also pass a variable name as an argument, e.g. '%cpaste foo'.
         This assigns the pasted block to variable 'foo' as string, without
         dedenting or executing it (preceding >>> and + is still stripped)
 
         '%cpaste -r' re-executes the block previously entered by cpaste.
 
         Do not be alarmed by garbled output on Windows (it's a readline bug).
         Just press enter and type -- (and press enter again) and the block
         will be what was just pasted.
 
         IPython statements (magics, shell escapes) are not supported (yet).
 
         See also
         --------
         paste: automatically pull code from clipboard.
 
         Examples
         --------
         ::
 
           In [8]: %cpaste
           Pasting code; enter '--' alone on the line to stop.
           :>>> a = ["world!", "Hello"]
           :>>> print " ".join(sorted(a))
           :--
           Hello world!
         """
 
         opts, name = self.parse_options(parameter_s, 'rs:', mode='string')
         if 'r' in opts:
             rerun_pasted(self.shell)
             return
 
         sentinel = opts.get('s', '--')
         block = strip_email_quotes(get_pasted_lines(sentinel))
         store_or_execute(self.shell, block, name)
 
     def magic_paste(self, parameter_s=''):
         """Paste & execute a pre-formatted code block from clipboard.
 
         The text is pulled directly from the clipboard without user
         intervention and printed back on the screen before execution (unless
         the -q flag is given to force quiet mode).
 
         The block is dedented prior to execution to enable execution of method
         definitions. '>' and '+' characters at the beginning of a line are
         ignored, to allow pasting directly from e-mails, diff files and
         doctests (the '...' continuation prompt is also stripped).  The
         executed block is also assigned to variable named 'pasted_block' for
         later editing with '%edit pasted_block'.
 
         You can also pass a variable name as an argument, e.g. '%paste foo'.
         This assigns the pasted block to variable 'foo' as string, without
         dedenting or executing it (preceding >>> and + is still stripped)
 
         Options
         -------
 
           -r: re-executes the block previously entered by cpaste.
 
           -q: quiet mode: do not echo the pasted text back to the terminal.
 
         IPython statements (magics, shell escapes) are not supported (yet).
 
         See also
         --------
         cpaste: manually paste code into terminal until you mark its end.
         """
         opts, name = self.parse_options(parameter_s, 'rq', mode='string')
         if 'r' in opts:
             rerun_pasted(self.shell)
             return
         try:
             text = self.shell.hooks.clipboard_get()
             block = strip_email_quotes(text.splitlines())
         except TryNext as clipboard_exc:
             message = getattr(clipboard_exc, 'args')
             if message:
                 error(message[0])
             else:
                 error('Could not get text from the clipboard.')
             return
 
         # By default, echo back to terminal unless quiet mode is requested
         if 'q' not in opts:
             write = self.shell.write
             write(self.shell.pycolorize(block))
             if not block.endswith('\n'):
                 write('\n')
             write("## -- End pasted text --\n")
 
         store_or_execute(self.shell, block, name)
 
     # Class-level: add a '%cls' magic only on Windows
     if sys.platform == 'win32':
         def magic_cls(self, s):
             """Clear screen.
             """
             os.system("cls")
 
     def showindentationerror(self):
         super(TerminalInteractiveShell, self).showindentationerror()
         print("If you want to paste code into IPython, try the "
               "%paste and %cpaste magic functions.")
 
 
 InteractiveShellABC.register(TerminalInteractiveShell)

IPython/utils/io.py

@@ -1,321 +1,321 b''
 # encoding: utf-8
 """
 IO related utilities.
 """
 
 #-----------------------------------------------------------------------------
 #  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.
 #-----------------------------------------------------------------------------
 from __future__ import print_function
 
 #-----------------------------------------------------------------------------
 # Imports
 #-----------------------------------------------------------------------------
 import sys
 import tempfile
 
 #-----------------------------------------------------------------------------
 # Code
 #-----------------------------------------------------------------------------
 
 
 class IOStream:
 
     def __init__(self,stream, fallback=None):
         if not hasattr(stream,'write') or not hasattr(stream,'flush'):
             if fallback is not None:
                 stream = fallback
             else:
                 raise ValueError("fallback required, but not specified")
         self.stream = stream
         self._swrite = stream.write
 
         # clone all methods not overridden:
         def clone(meth):
             return not hasattr(self, meth) and not meth.startswith('_')
         for meth in filter(clone, dir(stream)):
             setattr(self, meth, getattr(stream, meth))
 
     def write(self,data):
         try:
             self._swrite(data)
         except:
             try:
                 # print handles some unicode issues which may trip a plain
                 # write() call.  Emulate write() by using an empty end
                 # argument.
                 print(data, end='', file=self.stream)
             except:
                 # if we get here, something is seriously broken.
                 print('ERROR - failed to write data to stream:', self.stream,
                       file=sys.stderr)
 
     def writelines(self, lines):
         if isinstance(lines, basestring):
             lines = [lines]
         for line in lines:
             self.write(line)
 
     # This class used to have a writeln method, but regular files and streams
     # in Python don't have this method. We need to keep this completely
     # compatible so we removed it.
 
     @property
     def closed(self):
         return self.stream.closed
 
     def close(self):
         pass
 
 
 class IOTerm:
     """ Term holds the file or file-like objects for handling I/O operations.
 
     These are normally just sys.stdin, sys.stdout and sys.stderr but for
     Windows they can can replaced to allow editing the strings before they are
     displayed."""
 
     # In the future, having IPython channel all its I/O operations through
     # this class will make it easier to embed it into other environments which
     # are not a normal terminal (such as a GUI-based shell)
     def __init__(self, stdin=None, stdout=None, stderr=None):
         self.stdin  = IOStream(stdin, sys.stdin)
         self.stdout = IOStream(stdout, sys.stdout)
         self.stderr = IOStream(stderr, sys.stderr)
 
 # setup stdin/stdout/stderr to sys.stdin/sys.stdout/sys.stderr
-stdin = IOStream(sys.stdin)
-stdout = IOStream(sys.stdout)
-stderr = IOStream(sys.stderr)
+stdin = sys.stdin if not sys.stdin else IOStream(sys.stdin)
+stdout = sys.stdout if not sys.stdout else IOStream(sys.stdout)
+stderr = sys.stderr if not sys.stderr else IOStream(sys.stderr)
 
 
 class Tee(object):
     """A class to duplicate an output stream to stdout/err.
 
     This works in a manner very similar to the Unix 'tee' command.
 
     When the object is closed or deleted, it closes the original file given to
     it for duplication.
     """
     # Inspired by:
     # http://mail.python.org/pipermail/python-list/2007-May/442737.html
 
     def __init__(self, file_or_name, mode="w", channel='stdout'):
         """Construct a new Tee object.
 
         Parameters
         ----------
         file_or_name : filename or open filehandle (writable)
           File that will be duplicated
 
         mode : optional, valid mode for open().
           If a filename was give, open with this mode.
 
         channel : str, one of ['stdout', 'stderr']
         """
         if channel not in ['stdout', 'stderr']:
             raise ValueError('Invalid channel spec %s' % channel)
 
         if hasattr(file_or_name, 'write') and hasattr(file_or_name, 'seek'):
             self.file = file_or_name
         else:
             self.file = open(file_or_name, mode)
         self.channel = channel
         self.ostream = getattr(sys, channel)
         setattr(sys, channel, self)
         self._closed = False
 
     def close(self):
         """Close the file and restore the channel."""
         self.flush()
         setattr(sys, self.channel, self.ostream)
         self.file.close()
         self._closed = True
 
     def write(self, data):
         """Write data to both channels."""
         self.file.write(data)
         self.ostream.write(data)
         self.ostream.flush()
 
     def flush(self):
         """Flush both channels."""
         self.file.flush()
         self.ostream.flush()
 
     def __del__(self):
         if not self._closed:
             self.close()
 
 
 def file_read(filename):
     """Read a file and close it.  Returns the file source."""
     fobj = open(filename,'r');
     source = fobj.read();
     fobj.close()
     return source
 
 
 def file_readlines(filename):
     """Read a file and close it.  Returns the file source using readlines()."""
     fobj = open(filename,'r');
     lines = fobj.readlines();
     fobj.close()
     return lines
 
 
 def raw_input_multi(header='', ps1='==> ', ps2='..> ',terminate_str = '.'):
     """Take multiple lines of input.
 
     A list with each line of input as a separate element is returned when a
     termination string is entered (defaults to a single '.'). Input can also
     terminate via EOF (^D in Unix, ^Z-RET in Windows).
 
     Lines of input which end in \\ are joined into single entries (and a
     secondary continuation prompt is issued as long as the user terminates
     lines with \\). This allows entering very long strings which are still
     meant to be treated as single entities.
     """
 
     try:
         if header:
             header += '\n'
         lines = [raw_input(header + ps1)]
     except EOFError:
         return []
     terminate = [terminate_str]
     try:
         while lines[-1:] != terminate:
             new_line = raw_input(ps1)
             while new_line.endswith('\\'):
                 new_line = new_line[:-1] + raw_input(ps2)
             lines.append(new_line)
 
         return lines[:-1]  # don't return the termination command
     except EOFError:
         print()
         return lines
 
 
 def raw_input_ext(prompt='',  ps2='... '):
     """Similar to raw_input(), but accepts extended lines if input ends with \\."""
 
     line = raw_input(prompt)
     while line.endswith('\\'):
         line = line[:-1] + raw_input(ps2)
     return line
 
 
 def ask_yes_no(prompt,default=None):
     """Asks a question and returns a boolean (y/n) answer.
 
     If default is given (one of 'y','n'), it is used if the user input is
     empty. Otherwise the question is repeated until an answer is given.
 
     An EOF is treated as the default answer.  If there is no default, an
     exception is raised to prevent infinite loops.
 
     Valid answers are: y/yes/n/no (match is not case sensitive)."""
 
     answers = {'y':True,'n':False,'yes':True,'no':False}
     ans = None
     while ans not in answers.keys():
         try:
             ans = raw_input(prompt+' ').lower()
             if not ans:  # response was an empty string
                 ans = default
         except KeyboardInterrupt:
             pass
         except EOFError:
             if default in answers.keys():
                 ans = default
                 print()
             else:
                 raise
 
     return answers[ans]
 
 
 class NLprinter:
     """Print an arbitrarily nested list, indicating index numbers.
 
     An instance of this class called nlprint is available and callable as a
     function.
 
     nlprint(list,indent=' ',sep=': ') -> prints indenting each level by 'indent'
     and using 'sep' to separate the index from the value. """
 
     def __init__(self):
         self.depth = 0
 
     def __call__(self,lst,pos='',**kw):
         """Prints the nested list numbering levels."""
         kw.setdefault('indent',' ')
         kw.setdefault('sep',': ')
         kw.setdefault('start',0)
         kw.setdefault('stop',len(lst))
         # we need to remove start and stop from kw so they don't propagate
         # into a recursive call for a nested list.
         start = kw['start']; del kw['start']
         stop = kw['stop']; del kw['stop']
         if self.depth == 0 and 'header' in kw.keys():
             print(kw['header'])
 
         for idx in range(start,stop):
             elem = lst[idx]
             newpos = pos + str(idx)
             if type(elem)==type([]):
                 self.depth += 1
                 self.__call__(elem, newpos+",", **kw)
                 self.depth -= 1
             else:
                 print(kw['indent']*self.depth + newpos + kw["sep"] + repr(elem))
 
 nlprint = NLprinter()
 
 
 def temp_pyfile(src, ext='.py'):
     """Make a temporary python file, return filename and filehandle.
 
     Parameters
     ----------
     src : string or list of strings (no need for ending newlines if list)
       Source code to be written to the file.
 
     ext : optional, string
       Extension for the generated file.
 
     Returns
     -------
     (filename, open filehandle)
       It is the caller's responsibility to close the open file and unlink it.
     """
     fname = tempfile.mkstemp(ext)[1]
     f = open(fname,'w')
     f.write(src)
     f.flush()
     return fname, f
 
 
 def raw_print(*args, **kw):
     """Raw print to sys.__stdout__, otherwise identical interface to print()."""
 
     print(*args, sep=kw.get('sep', ' '), end=kw.get('end', '\n'),
           file=sys.__stdout__)
     sys.__stdout__.flush()
 
 
 def raw_print_err(*args, **kw):
     """Raw print to sys.__stderr__, otherwise identical interface to print()."""
 
     print(*args, sep=kw.get('sep', ' '), end=kw.get('end', '\n'),
           file=sys.__stderr__)
     sys.__stderr__.flush()
 
 
 # Short aliases for quick debugging, do NOT use these in production code.
 rprint = raw_print
 rprinte = raw_print_err

IPython/utils/py3compat.py

@@ -1,175 +1,183 b''
 # coding: utf-8
 """Compatibility tricks for Python 3. Mainly to do with unicode."""
 import __builtin__
 import functools
 import sys
 import re
 import types
 
 orig_open = open
 
 def no_code(x, encoding=None):
     return x
 
+# to deal with the possibility of sys.std* not being a stream at all
+def get_stream_enc(stream, default=None):
+    if not hasattr(stream, 'encoding') or not stream.encoding:
+        return default
+    else:
+        return stream.encoding
+
 def decode(s, encoding=None):
-    encoding = encoding or sys.stdin.encoding or sys.getdefaultencoding()
+    encoding = get_stream_enc(sys.stdin, encoding) or sys.getdefaultencoding()
     return s.decode(encoding, "replace")
 
 def encode(u, encoding=None):
-    encoding = encoding or sys.stdin.encoding or sys.getdefaultencoding()
+    encoding = get_stream_enc(sys.stdin, encoding) or sys.getdefaultencoding()
     return u.encode(encoding, "replace")
-    
+
+
 def cast_unicode(s, encoding=None):
     if isinstance(s, bytes):
         return decode(s, encoding)
     return s
 
 def cast_bytes(s, encoding=None):
     if not isinstance(s, bytes):
         return encode(s, encoding)
     return s
 
 def _modify_str_or_docstring(str_change_func):
     @functools.wraps(str_change_func)
     def wrapper(func_or_str):
         if isinstance(func_or_str, basestring):
             func = None
             doc = func_or_str
         else:
             func = func_or_str
             doc = func.__doc__
         
         doc = str_change_func(doc)
         
         if func:
             func.__doc__ = doc
             return func
         return doc
     return wrapper
 
 if sys.version_info[0] >= 3:
     PY3 = True
     
     input = input
     builtin_mod_name = "builtins"
     
     str_to_unicode = no_code
     unicode_to_str = no_code
     str_to_bytes = encode
     bytes_to_str = decode
     cast_bytes_py2 = no_code
     
     def isidentifier(s, dotted=False):
         if dotted:
             return all(isidentifier(a) for a in s.split("."))
         return s.isidentifier()
     
     open = orig_open
     
     MethodType = types.MethodType
     
     def execfile(fname, glob, loc=None):
         loc = loc if (loc is not None) else glob
         exec compile(open(fname, 'rb').read(), fname, 'exec') in glob, loc
     
     # Refactor print statements in doctests.
     _print_statement_re = re.compile(r"\bprint (?P<expr>.*)$", re.MULTILINE)
     def _print_statement_sub(match):
         expr = match.groups('expr')
         return "print(%s)" % expr
     
     @_modify_str_or_docstring
     def doctest_refactor_print(doc):
         """Refactor 'print x' statements in a doctest to print(x) style. 2to3
         unfortunately doesn't pick up on our doctests.
         
         Can accept a string or a function, so it can be used as a decorator."""
         return _print_statement_re.sub(_print_statement_sub, doc)
     
     # Abstract u'abc' syntax:
     @_modify_str_or_docstring
     def u_format(s):
         """"{u}'abc'" --> "'abc'" (Python 3)
         
         Accepts a string or a function, so it can be used as a decorator."""
         return s.format(u='')
 
 else:
     PY3 = False
     
     input = raw_input
     builtin_mod_name = "__builtin__"
     
     str_to_unicode = decode
     unicode_to_str = encode
     str_to_bytes = no_code
     bytes_to_str = no_code
     cast_bytes_py2 = cast_bytes
     
     import re
     _name_re = re.compile(r"[a-zA-Z_][a-zA-Z0-9_]*$")
     def isidentifier(s, dotted=False):
         if dotted:
             return all(isidentifier(a) for a in s.split("."))
         return bool(_name_re.match(s))
     
     class open(object):
         """Wrapper providing key part of Python 3 open() interface."""
         def __init__(self, fname, mode="r", encoding="utf-8"):
             self.f = orig_open(fname, mode)
             self.enc = encoding
         
         def write(self, s):
             return self.f.write(s.encode(self.enc))
         
         def read(self, size=-1):
             return self.f.read(size).decode(self.enc)
         
         def close(self):
             return self.f.close()
         
         def __enter__(self):
             return self
         
         def __exit__(self, etype, value, traceback):
             self.f.close()
     
     def MethodType(func, instance):
         return types.MethodType(func, instance, type(instance))
     
     # don't override system execfile on 2.x:
     execfile = execfile
     
     def doctest_refactor_print(func_or_str):
         return func_or_str
 
 
     # Abstract u'abc' syntax:
     @_modify_str_or_docstring
     def u_format(s):
         """"{u}'abc'" --> "u'abc'" (Python 2)
         
         Accepts a string or a function, so it can be used as a decorator."""
         return s.format(u='u')
 
     if sys.platform == 'win32':
         def execfile(fname, glob=None, loc=None):
             loc = loc if (loc is not None) else glob
             # The rstrip() is necessary b/c trailing whitespace in files will
             # cause an IndentationError in Python 2.6 (this was fixed in 2.7,
             # but we still support 2.6).  See issue 1027.
             scripttext = __builtin__.open(fname).read().rstrip() + '\n'
             # compile converts unicode filename to str assuming
             # ascii. Let's do the conversion before calling compile
             if isinstance(fname, unicode):
                 filename = unicode_to_str(fname)
             else:
                 filename = fname
             exec compile(scripttext, filename, 'exec') in glob, loc
     else:
         def execfile(fname, *where):
             if isinstance(fname, unicode):
                 filename = fname.encode(sys.getfilesystemencoding())
             else:
                 filename = fname
             __builtin__.execfile(filename, *where)

IPython/utils/text.py

@@ -1,760 +1,760 b''
 # encoding: utf-8
 """
 Utilities for working with strings and text.
 """
 
 #-----------------------------------------------------------------------------
 #  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 __main__
 
 import locale
 import os
 import re
 import shutil
 import sys
 import textwrap
 from string import Formatter
 
 from IPython.external.path import path
 from IPython.testing.skipdoctest import skip_doctest_py3
 from IPython.utils import py3compat
 from IPython.utils.io import nlprint
 from IPython.utils.data import flatten
 
 #-----------------------------------------------------------------------------
 # Code
 #-----------------------------------------------------------------------------
 
 # Less conservative replacement for sys.getdefaultencoding, that will try
 # to match the environment.
 # Defined here as central function, so if we find better choices, we
 # won't need to make changes all over IPython.
 def getdefaultencoding():
     """Return IPython's guess for the default encoding for bytes as text.
 
     Asks for stdin.encoding first, to match the calling Terminal, but that
     is often None for subprocesses.  Fall back on locale.getpreferredencoding()
     which should be a sensible platform default (that respects LANG environment),
     and finally to sys.getdefaultencoding() which is the most conservative option,
     and usually ASCII.
     """
-    enc = sys.stdin.encoding
+    enc = py3compat.get_stream_enc(sys.stdin)
     if not enc or enc=='ascii':
         try:
             # There are reports of getpreferredencoding raising errors
             # in some cases, which may well be fixed, but let's be conservative here.
             enc = locale.getpreferredencoding()
         except Exception:
             pass
     return enc or sys.getdefaultencoding()
 
 def unquote_ends(istr):
     """Remove a single pair of quotes from the endpoints of a string."""
 
     if not istr:
         return istr
     if (istr[0]=="'" and istr[-1]=="'") or \
        (istr[0]=='"' and istr[-1]=='"'):
         return istr[1:-1]
     else:
         return istr
 
 
 class LSString(str):
     """String derivative with a special access attributes.
 
     These are normal strings, but with the special attributes:
 
         .l (or .list) : value as list (split on newlines).
         .n (or .nlstr): original value (the string itself).
         .s (or .spstr): value as whitespace-separated string.
         .p (or .paths): list of path objects
 
     Any values which require transformations are computed only once and
     cached.
 
     Such strings are very useful to efficiently interact with the shell, which
     typically only understands whitespace-separated options for commands."""
 
     def get_list(self):
         try:
             return self.__list
         except AttributeError:
             self.__list = self.split('\n')
             return self.__list
 
     l = list = property(get_list)
 
     def get_spstr(self):
         try:
             return self.__spstr
         except AttributeError:
             self.__spstr = self.replace('\n',' ')
             return self.__spstr
 
     s = spstr = property(get_spstr)
 
     def get_nlstr(self):
         return self
 
     n = nlstr = property(get_nlstr)
 
     def get_paths(self):
         try:
             return self.__paths
         except AttributeError:
             self.__paths = [path(p) for p in self.split('\n') if os.path.exists(p)]
             return self.__paths
 
     p = paths = property(get_paths)
 
 # FIXME: We need to reimplement type specific displayhook and then add this
 # back as a custom printer. This should also be moved outside utils into the
 # core.
 
 # def print_lsstring(arg):
 #     """ Prettier (non-repr-like) and more informative printer for LSString """
 #     print "LSString (.p, .n, .l, .s available). Value:"
 #     print arg
 #
 #
 # print_lsstring = result_display.when_type(LSString)(print_lsstring)
 
 
 class SList(list):
     """List derivative with a special access attributes.
 
     These are normal lists, but with the special attributes:
 
         .l (or .list) : value as list (the list itself).
         .n (or .nlstr): value as a string, joined on newlines.
         .s (or .spstr): value as a string, joined on spaces.
         .p (or .paths): list of path objects
 
     Any values which require transformations are computed only once and
     cached."""
 
     def get_list(self):
         return self
 
     l = list = property(get_list)
 
     def get_spstr(self):
         try:
             return self.__spstr
         except AttributeError:
             self.__spstr = ' '.join(self)
             return self.__spstr
 
     s = spstr = property(get_spstr)
 
     def get_nlstr(self):
         try:
             return self.__nlstr
         except AttributeError:
             self.__nlstr = '\n'.join(self)
             return self.__nlstr
 
     n = nlstr = property(get_nlstr)
 
     def get_paths(self):
         try:
             return self.__paths
         except AttributeError:
             self.__paths = [path(p) for p in self if os.path.exists(p)]
             return self.__paths
 
     p = paths = property(get_paths)
 
     def grep(self, pattern, prune = False, field = None):
         """ Return all strings matching 'pattern' (a regex or callable)
 
         This is case-insensitive. If prune is true, return all items
         NOT matching the pattern.
 
         If field is specified, the match must occur in the specified
         whitespace-separated field.
 
         Examples::
 
             a.grep( lambda x: x.startswith('C') )
             a.grep('Cha.*log', prune=1)
             a.grep('chm', field=-1)
         """
 
         def match_target(s):
             if field is None:
                 return s
             parts = s.split()
             try:
                 tgt = parts[field]
                 return tgt
             except IndexError:
                 return ""
 
         if isinstance(pattern, basestring):
             pred = lambda x : re.search(pattern, x, re.IGNORECASE)
         else:
             pred = pattern
         if not prune:
             return SList([el for el in self if pred(match_target(el))])
         else:
             return SList([el for el in self if not pred(match_target(el))])
 
     def fields(self, *fields):
         """ Collect whitespace-separated fields from string list
 
         Allows quick awk-like usage of string lists.
 
         Example data (in var a, created by 'a = !ls -l')::
             -rwxrwxrwx  1 ville None      18 Dec 14  2006 ChangeLog
             drwxrwxrwx+ 6 ville None       0 Oct 24 18:05 IPython
 
         a.fields(0) is ['-rwxrwxrwx', 'drwxrwxrwx+']
         a.fields(1,0) is ['1 -rwxrwxrwx', '6 drwxrwxrwx+']
         (note the joining by space).
         a.fields(-1) is ['ChangeLog', 'IPython']
 
         IndexErrors are ignored.
 
         Without args, fields() just split()'s the strings.
         """
         if len(fields) == 0:
             return [el.split() for el in self]
 
         res = SList()
         for el in [f.split() for f in self]:
             lineparts = []
 
             for fd in fields:
                 try:
                     lineparts.append(el[fd])
                 except IndexError:
                     pass
             if lineparts:
                 res.append(" ".join(lineparts))
 
         return res
 
     def sort(self,field= None,  nums = False):
         """ sort by specified fields (see fields())
 
         Example::
             a.sort(1, nums = True)
 
         Sorts a by second field, in numerical order (so that 21 > 3)
 
         """
 
         #decorate, sort, undecorate
         if field is not None:
             dsu = [[SList([line]).fields(field),  line] for line in self]
         else:
             dsu = [[line,  line] for line in self]
         if nums:
             for i in range(len(dsu)):
                 numstr = "".join([ch for ch in dsu[i][0] if ch.isdigit()])
                 try:
                     n = int(numstr)
                 except ValueError:
                     n = 0;
                 dsu[i][0] = n
 
 
         dsu.sort()
         return SList([t[1] for t in dsu])
 
 
 # FIXME: We need to reimplement type specific displayhook and then add this
 # back as a custom printer. This should also be moved outside utils into the
 # core.
 
 # def print_slist(arg):
 #     """ Prettier (non-repr-like) and more informative printer for SList """
 #     print "SList (.p, .n, .l, .s, .grep(), .fields(), sort() available):"
 #     if hasattr(arg,  'hideonce') and arg.hideonce:
 #         arg.hideonce = False
 #         return
 #
 #     nlprint(arg)
 #
 # print_slist = result_display.when_type(SList)(print_slist)
 
 
 def esc_quotes(strng):
     """Return the input string with single and double quotes escaped out"""
 
     return strng.replace('"','\\"').replace("'","\\'")
 
 
 def qw(words,flat=0,sep=None,maxsplit=-1):
     """Similar to Perl's qw() operator, but with some more options.
 
     qw(words,flat=0,sep=' ',maxsplit=-1) -> words.split(sep,maxsplit)
 
     words can also be a list itself, and with flat=1, the output will be
     recursively flattened.
 
     Examples:
 
     >>> qw('1 2')
     ['1', '2']
 
     >>> qw(['a b','1 2',['m n','p q']])
     [['a', 'b'], ['1', '2'], [['m', 'n'], ['p', 'q']]]
 
     >>> qw(['a b','1 2',['m n','p q']],flat=1)
     ['a', 'b', '1', '2', 'm', 'n', 'p', 'q']
     """
 
     if isinstance(words, basestring):
         return [word.strip() for word in words.split(sep,maxsplit)
                 if word and not word.isspace() ]
     if flat:
         return flatten(map(qw,words,[1]*len(words)))
     return map(qw,words)
 
 
 def qwflat(words,sep=None,maxsplit=-1):
     """Calls qw(words) in flat mode. It's just a convenient shorthand."""
     return qw(words,1,sep,maxsplit)
 
 
 def qw_lol(indata):
     """qw_lol('a b') -> [['a','b']],
     otherwise it's just a call to qw().
 
     We need this to make sure the modules_some keys *always* end up as a
     list of lists."""
 
     if isinstance(indata, basestring):
         return [qw(indata)]
     else:
         return qw(indata)
 
 
 def grep(pat,list,case=1):
     """Simple minded grep-like function.
     grep(pat,list) returns occurrences of pat in list, None on failure.
 
     It only does simple string matching, with no support for regexps. Use the
     option case=0 for case-insensitive matching."""
 
     # This is pretty crude. At least it should implement copying only references
     # to the original data in case it's big. Now it copies the data for output.
     out=[]
     if case:
         for term in list:
             if term.find(pat)>-1: out.append(term)
     else:
         lpat=pat.lower()
         for term in list:
             if term.lower().find(lpat)>-1: out.append(term)
 
     if len(out): return out
     else: return None
 
 
 def dgrep(pat,*opts):
     """Return grep() on dir()+dir(__builtins__).
 
     A very common use of grep() when working interactively."""
 
     return grep(pat,dir(__main__)+dir(__main__.__builtins__),*opts)
 
 
 def idgrep(pat):
     """Case-insensitive dgrep()"""
 
     return dgrep(pat,0)
 
 
 def igrep(pat,list):
     """Synonym for case-insensitive grep."""
 
     return grep(pat,list,case=0)
 
 
 def indent(instr,nspaces=4, ntabs=0, flatten=False):
     """Indent a string a given number of spaces or tabstops.
 
     indent(str,nspaces=4,ntabs=0) -> indent str by ntabs+nspaces.
 
     Parameters
     ----------
 
     instr : basestring
         The string to be indented.
     nspaces : int (default: 4)
         The number of spaces to be indented.
     ntabs : int (default: 0)
         The number of tabs to be indented.
     flatten : bool (default: False)
         Whether to scrub existing indentation.  If True, all lines will be
         aligned to the same indentation.  If False, existing indentation will
         be strictly increased.
 
     Returns
     -------
 
     str|unicode : string indented by ntabs and nspaces.
 
     """
     if instr is None:
         return
     ind = '\t'*ntabs+' '*nspaces
     if flatten:
         pat = re.compile(r'^\s*', re.MULTILINE)
     else:
         pat = re.compile(r'^', re.MULTILINE)
     outstr = re.sub(pat, ind, instr)
     if outstr.endswith(os.linesep+ind):
         return outstr[:-len(ind)]
     else:
         return outstr
 
 def native_line_ends(filename,backup=1):
     """Convert (in-place) a file to line-ends native to the current OS.
 
     If the optional backup argument is given as false, no backup of the
     original file is left.  """
 
     backup_suffixes = {'posix':'~','dos':'.bak','nt':'.bak','mac':'.bak'}
 
     bak_filename = filename + backup_suffixes[os.name]
 
     original = open(filename).read()
     shutil.copy2(filename,bak_filename)
     try:
         new = open(filename,'wb')
         new.write(os.linesep.join(original.splitlines()))
         new.write(os.linesep) # ALWAYS put an eol at the end of the file
         new.close()
     except:
         os.rename(bak_filename,filename)
     if not backup:
         try:
             os.remove(bak_filename)
         except:
             pass
 
 
 def list_strings(arg):
     """Always return a list of strings, given a string or list of strings
     as input.
 
     :Examples:
 
         In [7]: list_strings('A single string')
         Out[7]: ['A single string']
 
         In [8]: list_strings(['A single string in a list'])
         Out[8]: ['A single string in a list']
 
         In [9]: list_strings(['A','list','of','strings'])
         Out[9]: ['A', 'list', 'of', 'strings']
     """
 
     if isinstance(arg,basestring): return [arg]
     else: return arg
 
 
 def marquee(txt='',width=78,mark='*'):
     """Return the input string centered in a 'marquee'.
 
     :Examples:
 
         In [16]: marquee('A test',40)
         Out[16]: '**************** A test ****************'
 
         In [17]: marquee('A test',40,'-')
         Out[17]: '---------------- A test ----------------'
 
         In [18]: marquee('A test',40,' ')
         Out[18]: '                 A test                 '
 
     """
     if not txt:
         return (mark*width)[:width]
     nmark = (width-len(txt)-2)//len(mark)//2
     if nmark < 0: nmark =0
     marks = mark*nmark
     return '%s %s %s' % (marks,txt,marks)
 
 
 ini_spaces_re = re.compile(r'^(\s+)')
 
 def num_ini_spaces(strng):
     """Return the number of initial spaces in a string"""
 
     ini_spaces = ini_spaces_re.match(strng)
     if ini_spaces:
         return ini_spaces.end()
     else:
         return 0
 
 
 def format_screen(strng):
     """Format a string for screen printing.
 
     This removes some latex-type format codes."""
     # Paragraph continue
     par_re = re.compile(r'\\$',re.MULTILINE)
     strng = par_re.sub('',strng)
     return strng
 
 def dedent(text):
     """Equivalent of textwrap.dedent that ignores unindented first line.
 
     This means it will still dedent strings like:
     '''foo
     is a bar
     '''
 
     For use in wrap_paragraphs.
     """
 
     if text.startswith('\n'):
         # text starts with blank line, don't ignore the first line
         return textwrap.dedent(text)
 
     # split first line
     splits = text.split('\n',1)
     if len(splits) == 1:
         # only one line
         return textwrap.dedent(text)
 
     first, rest = splits
     # dedent everything but the first line
     rest = textwrap.dedent(rest)
     return '\n'.join([first, rest])
 
 def wrap_paragraphs(text, ncols=80):
     """Wrap multiple paragraphs to fit a specified width.
 
     This is equivalent to textwrap.wrap, but with support for multiple
     paragraphs, as separated by empty lines.
 
     Returns
     -------
 
     list of complete paragraphs, wrapped to fill `ncols` columns.
     """
     paragraph_re = re.compile(r'\n(\s*\n)+', re.MULTILINE)
     text = dedent(text).strip()
     paragraphs = paragraph_re.split(text)[::2] # every other entry is space
     out_ps = []
     indent_re = re.compile(r'\n\s+', re.MULTILINE)
     for p in paragraphs:
         # presume indentation that survives dedent is meaningful formatting,
         # so don't fill unless text is flush.
         if indent_re.search(p) is None:
             # wrap paragraph
             p = textwrap.fill(p, ncols)
         out_ps.append(p)
     return out_ps
 
 
 class EvalFormatter(Formatter):
     """A String Formatter that allows evaluation of simple expressions.
     
     Note that this version interprets a : as specifying a format string (as per
     standard string formatting), so if slicing is required, you must explicitly
     create a slice.
     
     This is to be used in templating cases, such as the parallel batch
     script templates, where simple arithmetic on arguments is useful.
 
     Examples
     --------
     
     In  [1]: f = EvalFormatter()
     In  [2]: f.format('{n//4}', n=8)
     Out [2]: '2'
     
     In  [3]: f.format("{greeting[slice(2,4)]}", greeting="Hello")
     Out [3]: 'll'
     """
     def get_field(self, name, args, kwargs):
         v = eval(name, kwargs)
         return v, name
 
 @skip_doctest_py3
 class FullEvalFormatter(Formatter):
     """A String Formatter that allows evaluation of simple expressions.
     
     Any time a format key is not found in the kwargs,
     it will be tried as an expression in the kwargs namespace.
     
     Note that this version allows slicing using [1:2], so you cannot specify
     a format string. Use :class:`EvalFormatter` to permit format strings.
     
     Examples
     --------
     
     In [1]: f = FullEvalFormatter()
     In [2]: f.format('{n//4}', n=8)
     Out[2]: u'2'
     
     In [3]: f.format('{list(range(5))[2:4]}')
     Out[3]: u'[2, 3]'
 
     In [4]: f.format('{3*2}')
     Out[4]: u'6'
     """
     # copied from Formatter._vformat with minor changes to allow eval
     # and replace the format_spec code with slicing
     def _vformat(self, format_string, args, kwargs, used_args, recursion_depth):
         if recursion_depth < 0:
             raise ValueError('Max string recursion exceeded')
         result = []
         for literal_text, field_name, format_spec, conversion in \
                 self.parse(format_string):
 
             # output the literal text
             if literal_text:
                 result.append(literal_text)
 
             # if there's a field, output it
             if field_name is not None:
                 # this is some markup, find the object and do
                 # the formatting
 
                 if format_spec:
                     # override format spec, to allow slicing:
                     field_name = ':'.join([field_name, format_spec])
 
                 # eval the contents of the field for the object
                 # to be formatted
                 obj = eval(field_name, kwargs)
 
                 # do any conversion on the resulting object
                 obj = self.convert_field(obj, conversion)
 
                 # format the object and append to the result
                 result.append(self.format_field(obj, ''))
 
         return u''.join(py3compat.cast_unicode(s) for s in result)
 
 @skip_doctest_py3
 class DollarFormatter(FullEvalFormatter):
     """Formatter allowing Itpl style $foo replacement, for names and attribute
     access only. Standard {foo} replacement also works, and allows full
     evaluation of its arguments.
 
     Examples
     --------
     In [1]: f = DollarFormatter()
     In [2]: f.format('{n//4}', n=8)
     Out[2]: u'2'
     
     In [3]: f.format('23 * 76 is $result', result=23*76)
     Out[3]: u'23 * 76 is 1748'
     
     In [4]: f.format('$a or {b}', a=1, b=2)
     Out[4]: u'1 or 2'
     """
     _dollar_pattern = re.compile("(.*?)\$(\$?[\w\.]+)")
     def parse(self, fmt_string):
         for literal_txt, field_name, format_spec, conversion \
                     in Formatter.parse(self, fmt_string):
             
             # Find $foo patterns in the literal text.
             continue_from = 0
             txt = ""
             for m in self._dollar_pattern.finditer(literal_txt):
                 new_txt, new_field = m.group(1,2)
                 # $$foo --> $foo
                 if new_field.startswith("$"):
                     txt += new_txt + new_field
                 else:
                     yield (txt + new_txt, new_field, "", None)
                     txt = ""
                 continue_from = m.end()
             
             # Re-yield the {foo} style pattern
             yield (txt + literal_txt[continue_from:], field_name, format_spec, conversion)
 
 
 def columnize(items, separator='  ', displaywidth=80):
     """ Transform a list of strings into a single string with columns.
 
     Parameters
     ----------
     items : sequence of strings
         The strings to process.
 
     separator : str, optional [default is two spaces]
         The string that separates columns.
 
     displaywidth : int, optional [default is 80]
         Width of the display in number of characters.
 
     Returns
     -------
     The formatted string.
     """
     # Note: this code is adapted from columnize 0.3.2.
     # See http://code.google.com/p/pycolumnize/
 
     # Some degenerate cases.
     size = len(items)
     if size == 0:
         return '\n'
     elif size == 1:
         return '%s\n' % items[0]
 
     # Special case: if any item is longer than the maximum width, there's no
     # point in triggering the logic below...
     item_len = map(len, items) # save these, we can reuse them below
     longest = max(item_len)
     if longest >= displaywidth:
         return '\n'.join(items+[''])
 
     # Try every row count from 1 upwards
     array_index = lambda nrows, row, col: nrows*col + row
     for nrows in range(1, size):
         ncols = (size + nrows - 1) // nrows
         colwidths = []
         totwidth = -len(separator)
         for col in range(ncols):
             # Get max column width for this column
             colwidth = 0
             for row in range(nrows):
                 i = array_index(nrows, row, col)
                 if i >= size: break
                 x, len_x = items[i], item_len[i]
                 colwidth = max(colwidth, len_x)
             colwidths.append(colwidth)
             totwidth += colwidth + len(separator)
             if totwidth > displaywidth:
                 break
         if totwidth <= displaywidth:
             break
 
     # The smallest number of rows computed and the max widths for each
     # column has been obtained. Now we just have to format each of the rows.
     string = ''
     for row in range(nrows):
         texts = []
         for col in range(ncols):
             i = row + nrows*col
             if i >= size:
                 texts.append('')
             else:
                 texts.append(items[i])
         while texts and not texts[-1]:
             del texts[-1]
         for col in range(len(texts)):
             texts[col] = texts[col].ljust(colwidths[col])
         string += '%s\n' % separator.join(texts)
     return string