##// END OF EJS Templates
First step in reintegrating Jedi...
First step in reintegrating Jedi If Jedi is installed expose a private API use it with prompt toolkit. Jedi does not _yet_ provide all the completion IPython has, so this is still a bit awkward. In order to debug this (and see what is Jedi provided we for now inject a fake Jedi/IPython delimiter in the menu. Jedi completion and this behavior are enabled by default, but could likely be opt-in. Add also a number of debug flags to be able to track why jedi is not working, and/or what completions are found by IPython and not Jedi. That should give us a bit of heads up and feedback to know whether we can remove part of the IPython completer, and more especially if we can drop `python_matches`. Once `python_matches` is dropped and some other of the current matchers are either dropped or converted to the new API, that should simplify the internal quite a bit. That would just be too much for an already BIG pull-request.

File last commit:

r11067:5076556c
r23284:3ff1be2e
Show More
nonascii2.py
4 lines | 155 B | text/x-python | PythonLexer
Thomas Kluyver
Add test for non-ascii characters in docstrings...
r11067 # coding: iso-8859-5
# (Unlikely to be the default encoding for most testers.)
# БЖџрстуфхцчшщъыьэюя <- Cyrillic characters
'Ўт№Ф'