1
0
forked from VimPlug/jedi
Commit Graph

326 Commits

Author SHA1 Message Date
Dave Halter
06004ad2f5 Some minor refactorings. 2017-10-09 20:32:28 +02:00
Dave Halter
bedff46735 Simplify usages. It should also work way better, now. 2017-10-08 20:13:24 +02:00
Dave Halter
4ddf7bf56d Remove the disabling of dynamic flow information
We should be able to handle this anyway also in completions. Don't hide issues here.
2017-10-07 10:52:43 +02:00
Dave Halter
51d2ffb078 Use sys path mostly from project and move some sys path stuff around. 2017-10-05 10:06:28 +02:00
Dave Halter
383f749026 Move the initial sys path generation into a new project class. 2017-10-02 20:19:55 +02:00
Dave Halter
0762c9218c Move arguments to a separate module. 2017-10-01 13:29:28 +02:00
Dave Halter
16011a91af Move iterable to context/iterable. 2017-09-30 17:41:21 +02:00
Dave Halter
3c75f27376 Move the base Context stuff to another module to keep context free for imports. 2017-09-30 16:46:07 +02:00
Dave Halter
8f177eea07 Move the ModuleContext to a separate module. 2017-09-29 13:24:48 +02:00
Dave Halter
d99d4deebf Merge branch 'values' 2017-09-28 16:19:38 +02:00
Dave Halter
d9d3aeb5bc Move more functions to the syntax tree module. 2017-09-28 09:16:43 +02:00
Dave Halter
d6a04b2928 Remove the deprecated attributes from Jedi. 2017-09-20 18:27:29 +02:00
Dave Halter
0c01a3b823 The sys.modules implementation did not work properly with newly created files.
Fixes #886.
2017-09-20 10:06:02 +02:00
Dave Halter
260aef943a Increase Python's recursion limit
Currently there is still the possiblity that Jedi fails with a recursion error,
because the stack is too small. (see #861) By increasing it we improve the
situation.

Probably we should just be switching away from this extreme amount of recursion
and move to queueing which would also allow to use other algorithms such as
breadth-first-search.
2017-09-18 10:26:42 +02:00
Dave Halter
f4ba71f6a3 Move the recursion limit settings to the recursion module. 2017-09-17 14:08:39 +02:00
Dave Halter
5ff7e3dbbe Actually do goto when follow_imports is used
Fixes #945.
2017-09-13 00:28:49 +02:00
Dave Halter
4a544c29ea Fix a follow_imports (goto) issue. 2017-09-11 23:32:10 +02:00
Dave Halter
619acbd2ca Goto didn't work well on imports in __init__.py files.
Fixes #956.
2017-09-11 21:48:37 +02:00
Dave Halter
882f8029ea Use split_lines and python_bytes_to_unicode directly. 2017-09-03 18:38:00 +02:00
Dave Halter
9663e343c2 Almost the last switch to _get_definition. 2017-09-02 22:42:01 +02:00
Dave Halter
fe9be9fe09 source_to_unicode -> python_bytes_to_unicode. 2017-08-15 20:09:48 +02:00
Dave Halter
f9e31dc941 Refactor splitlines -> split_lines. 2017-08-15 19:55:50 +02:00
Dave Halter
e0485b032e Fix some stuff to make parso work again. 2017-06-02 00:00:31 +02:00
Dave Halter
24cd603fcf Some more parso adaptations. 2017-05-26 09:08:34 -04:00
Dave Halter
3f36824a94 Parso changed load_python_grammar to load_grammar. 2017-05-25 12:41:19 -04:00
Dave Halter
ef2e2f343e Fix some warnings. 2017-05-25 12:24:21 -04:00
Dave Halter
b90589b62e Some changes because parso has changed. 2017-05-22 15:42:42 -04:00
Dave Halter
f35f1b9676 Add the cache_path parameter to parso calls. 2017-05-20 10:08:48 -04:00
Dave Halter
50c7137437 splitlines and source_to_unicode are utils of parso. 2017-05-20 09:55:16 -04:00
Dave Halter
9bca3d39f5 Actually use parso now instead of Jedi. 2017-05-19 14:20:14 -04:00
Dave Halter
84d8279089 Import.paths -> Import.get_paths. 2017-05-07 15:47:34 +02:00
Matthias Bussonnier
6f1ee0cfa8 Use stacklevel in warnings or filters don't work.
In particular with the right stacklevel IPython will display the warning
if code is directly entered by the user. Without this info it does not.

Use the opportunity to add in the warning since when things are
deprecated. This leads to one less lookup of information for the user.
2017-04-29 20:13:19 +02:00
Dave Halter
deb028c3fb Move get_statement_of_position out of the parser tree. 2017-04-15 02:23:08 +02:00
Dave Halter
65a6c61dc6 Remove nodes_to_execute in favor of a function in parser_utils. 2017-04-08 12:59:49 +02:00
Dave Halter
a6829ca546 Use the cache variables in more straight forward fashion. 2017-03-31 23:10:39 +02:00
Dave Halter
6a9c2f8795 Start using ContextualizedNode for py__iter__. 2017-03-28 01:34:07 +02:00
Dave Halter
8a35a04439 Remove the module path from the parser tree.
Some static analysis tests are still failing.
2017-03-27 18:13:32 +02:00
Dave Halter
7874026ee5 A lot of work toward a better diff parser API. 2017-03-25 01:51:03 +01:00
Dave Halter
26cce4d078 Add the grammar as an argument to saving the parser.
This makes collisions of different grammars when loading from the cache impossible.
2017-03-22 18:32:49 +01:00
Dave Halter
2cb565561d Replace the diff parser imports with the modified path. 2017-03-21 22:10:01 +01:00
Dave Halter
0f66a3c7a8 Remove the module attribute from the parser. 2017-03-18 03:53:34 +01:00
Dave Halter
d0b6d41e99 Remove the old star import cache, because it's not even used. 2017-03-18 03:30:23 +01:00
Dave Halter
448bfd0992 Move the python parser tree. 2017-03-16 17:20:32 +01:00
Dave Halter
97fc3bc23c Refactored the parser calls. Now it's possible to use jedi.parser.python.parse to quickly parse something. 2017-03-14 00:38:58 +01:00
Dave Halter
425fba5e95 Move the parser.fast module to parser.diff. 2017-01-10 19:15:47 +01:00
Dave Halter
5b9e5f96aa Merge with master. 2017-01-02 13:05:45 +01:00
Dave Halter
57857b6332 Remove the ImportWrapper and replace it with something simpler. 2016-12-17 16:08:37 +01:00
Dave Halter
575352d4b6 Start cleaning up the interpreter module. 2016-12-15 09:49:34 +01:00
Dave Halter
eaf0100446 Some analysis improvements. 2016-12-11 15:03:19 +01:00
Dave Halter
2be5da3f85 Fix some old regression tests. 2016-12-07 01:30:30 +01:00