##// END OF EJS Templates
Release 5.0.0b4...
Release 5.0.0b4 What's new since b3: Small delay since last beta release, sorry about that, a few reasons though: - We / I slowed down a bit on the beta release pace, there was a few bugs/issues/User interface annoyance that needed a new version of prompt_toolkit to be released, there was no reasons to make extra-beta without this new version released. - We backed-down on some changes on which we had a few disagrements (more below). - I got sick, and had to catch up with backlog. Anyway, Jonathan release prompt_toolkit 1.0.1 and 1.0.2, so even if you don't try this new beta, upgrading prompt_toolkit will fix some of your issues. == Move back `TerminalInteractiveShell` to it's old place. This is a bigger breaking change since 5.0b3, the `TerminalInteractiveshell` moved back from `IPython.terminal.ptshell` to `IPython.terminal.interactiveshell`, if you've updated your project recently to adapt to this change we're sorry, but despite the fact that the version pre 5.0 and post 5.0 are relatively different the cost of conditional import for project depending on us appeared to be too high. So it's now easier to migrate from 4.0 to 5.0 as the class have the same name, and same location == Option name and default changed. `TerminalInteractiveShell.display_completions_in_column` is now gone. It was not present on 4.x so no API breakage there, and is now replaced by `TerminalInteractiveShell.display_completions` and is a enum that gained a 3rd mode for the completer: `readlinelike` for those of you that regret readline. This give us more flexibility for further options. Would appreciate testing of this new layout from vi user. The two other mode now being `column` and `multicolumn`. By popular request, `multicolumn` is not the default value for the previous option. == bug fixed: - quit/exit broken in ipdb - Copy/Past broken on windowm - Unicode broken on windows - function signature garbled when using `object?` - issue with paging text with `?` - completer could get stuck. See the complete git log for more informations.

File last commit:

r15763:5dd3598e
r22560:1b487f7a
Show More
rlineimpl.py
74 lines | 2.7 KiB | text/x-python | PythonLexer
# -*- coding: utf-8 -*-
""" Imports and provides the 'correct' version of readline for the platform.
Readline is used throughout IPython as::
import IPython.utils.rlineimpl as readline
In addition to normal readline stuff, this module provides have_readline
boolean and _outputfile variable used in IPython.utils.
"""
import sys
import warnings
_rlmod_names = ['gnureadline', 'readline']
have_readline = False
for _rlmod_name in _rlmod_names:
try:
# import readline as _rl
_rl = __import__(_rlmod_name)
# from readline import *
globals().update({k:v for k,v in _rl.__dict__.items() if not k.startswith('_')})
except ImportError:
pass
else:
have_readline = True
break
if have_readline and (sys.platform == 'win32' or sys.platform == 'cli'):
try:
_outputfile=_rl.GetOutputFile()
except AttributeError:
warnings.warn("Failed GetOutputFile")
have_readline = False
# Test to see if libedit is being used instead of GNU readline.
# Thanks to Boyd Waters for the original patch.
uses_libedit = False
if have_readline:
# Official Python docs state that 'libedit' is in the docstring for libedit readline:
uses_libedit = _rl.__doc__ and 'libedit' in _rl.__doc__
# Note that many non-System Pythons also do not use proper readline,
# but do not report libedit at all, nor are they linked dynamically against libedit.
# known culprits of this include: EPD, Fink
# There is not much we can do to detect this, until we find a specific failure
# case, rather than relying on the readline module to self-identify as broken.
if uses_libedit and sys.platform == 'darwin':
_rl.parse_and_bind("bind ^I rl_complete")
warnings.warn('\n'.join(['', "*"*78,
"libedit detected - readline will not be well behaved, including but not limited to:",
" * crashes on tab completion",
" * incorrect history navigation",
" * corrupting long-lines",
" * failure to wrap or indent lines properly",
"It is highly recommended that you install gnureadline, which is installable with:",
" pip install gnureadline",
"*"*78]),
RuntimeWarning)
# the clear_history() function was only introduced in Python 2.4 and is
# actually optional in the readline API, so we must explicitly check for its
# existence. Some known platforms actually don't have it. This thread:
# http://mail.python.org/pipermail/python-dev/2003-August/037845.html
# has the original discussion.
if have_readline:
try:
_rl.clear_history
except AttributeError:
def clear_history(): pass
_rl.clear_history = clear_history