9f45f18ad1
Added a grammar param to the parser.
2014-11-24 01:10:39 +01:00
c152a1c58b
Actually replace tree with representation (in all the imports).
2014-11-23 19:46:52 +01:00
cc465364d3
Fixes towards better MergedArray and partial functions.
2014-11-20 13:33:05 +01:00
ce5d428d22
At least functions generate docstrings again.
2014-11-20 01:37:18 +01:00
fe7c750c2c
Param is now the parent of its names and not just a helper class.
2014-10-28 02:05:44 +01:00
585e92ac9f
Temporarily disable evaluate/compiled stuff, because it interferes with the current changes of the parser. This will be undone.
2014-10-08 17:45:49 +02:00
d7836c1034
Add a comment and link to issue #483
2014-09-29 11:57:38 +02:00
42596dba15
Merge remote-tracking branch 'upstream/dev' into fix_runtime_error
...
Conflicts:
jedi/evaluate/imports.py
2014-09-29 11:53:35 +02:00
d1ae447362
Simplify code
2014-09-29 09:50:49 +02:00
c4e45916c6
Modules also use a NamePart as a name, now.
2014-09-22 12:52:48 +02:00
fb86388890
Fix RuntimeError: the PyQt5.QtCore and PyQt4.QtCore modules both wrap the QObject class
2014-09-13 12:18:34 +02:00
bb5ffe9343
CompiledObject.name returns a Name now, not a string. This is more consistent with the Jedi design and doesn't lead to bugs while ducktyping.
2014-09-03 19:43:21 +02:00
18204c4c19
By trying to get rid of search_name in usages, we had to fix an issue with imports:
...
If used like 'follow(is_goto)', it could return a ModuleWrapper instead of a Name, which is what we actually want.
2014-09-03 19:30:00 +02:00
9abc8a19e7
By adding a py__class__ method to CompiledObject and Class, we Jedi is able to understand isinstance checks, now.
...
This also includes a CheckAttribute class in evaluatue.compiled, because it's way easier to generalize the AttributeErrors there.
2014-08-20 14:46:18 +02:00
c9542cbc04
Start implementing an is_class function that will determine if an object is a class or not in the future.
2014-08-20 11:43:25 +02:00
09ca47fa93
Introduce a dedicated isinstance function implementation.
2014-08-19 23:57:59 +02:00
425290aa8f
Fix an issue with partial keyword inputs.
2014-08-14 12:25:00 +02:00
ec7b3bf433
refactor py_base to py__bases__, because that's the general naming schema
2014-08-13 14:34:37 +02:00
cd433adf84
Speedup object lookup even further in classes.
2014-08-13 14:17:57 +02:00
8ed89e8245
implement !=, ==, is, is not operators to work in if statements (they also work in in non if contexts and return a bool value.), includes tests.
2014-08-12 17:59:19 +02:00
fb1dba269a
re-enable the interpretation of the None keyword
2014-08-12 14:38:56 +02:00
33e5a3280a
Remove IsScope in favor of an is_scope function.
...
This function was partially implemented anway. Now we've also added a function called 'get_parent_scope', to make it easy to get a scope of a Call, Statement, whatever.
2014-08-12 01:19:19 +02:00
242072976a
use py__mro__ in a classes scope_names_generator
2014-08-11 23:53:45 +02:00
e7e7bd29e8
fix generator tests (multiple yields must be called with an if random.
2014-08-06 12:45:38 +02:00
15ec0a77fe
a first very simple implementation of reachable/unreachable return statements.
2014-08-05 17:02:16 +02:00
6e5d80a6b2
builtins shouldn't be unique if called by compiled.create
2014-08-01 15:51:59 +02:00
59b8c6b015
CompiledObjects should execute everything when reading the return information from docstring (because it's always types, not values)
2014-07-31 17:16:24 +02:00
50fa3a732d
actually start checking if the integration tests are instances on both sides of the comparison. This wasnt necessary for just autocompletion, but it's way more important now.
2014-07-31 14:58:32 +02:00
0851e7667e
A module shouldn't be callable.
2014-07-30 17:07:57 +02:00
7bd76022bf
get rid of the whole is_callable stuff, because now we can just check for hasattr(obj, 'py__call__')
2014-07-30 15:50:47 +02:00
1e6a950aec
further progress in changing to py__call__
2014-07-30 14:40:56 +02:00
852cdad754
Operator fixes. Subclass of Simple, now.
2014-07-22 16:02:34 +02:00
2696d95d70
fix dotted_from_fs_path for Windows
2014-07-19 10:33:08 +09:00
13ada3154b
replace get_defined_names with scope_names_generator in a lot of places (the cleanup still needs to be done, though).
2014-07-02 16:12:49 +02:00
5b7c869323
types also add to completions, also for compiled objects. removed a few lines of code that complicated the process as well.
2014-06-28 12:09:43 +02:00
d444ef9e15
setdefault fix
2014-06-26 15:23:20 +02:00
47205dd7f3
change the implementation of compiled.load_module and always use the sys.modules cache after an import, because it's easier. Doesn't require any logic.
2014-06-25 18:57:07 +02:00
ddca14980e
introduce an is_scope function to make it easier to work with scopes
2014-06-20 17:47:42 +02:00
ea271c8047
replace _sre fake module literals with undefined types.
2014-06-20 11:18:54 +02:00
a373818965
fix function execution mutable list issue
2014-06-16 17:12:27 +02:00
4e596060b9
test for is_nested failure
2014-05-16 13:00:13 +02:00
a2b483b4f5
None issue fix for static analysis
2014-05-13 01:21:32 +02:00
b6ec589997
refactor general array lookup method get_index_types
2014-05-12 15:23:48 +02:00
6bec7ce847
speed up Builtin lookups
2014-04-25 14:12:28 +02:00
c790f88678
speed up precedences.
2014-04-25 13:28:01 +02:00
ff810d9ece
cache compiled.CompiledObject.defined_names attribute, improves test suite performance by 20%, numpy by more than 50%. Awesome!
2014-04-14 02:20:22 +02:00
3fced34544
the last commit featured an incomplete caching for compiled objects. The current one should improve this.
2014-04-13 21:55:29 +02:00
4bc55be103
significant speedup due to compiled caching
2014-04-13 16:31:38 +02:00
27645af6d2
make a slow inspect.getdoc() call lazy, which improves performance for numpy completions by 30%
2014-04-11 11:53:11 +02:00
d2dc39e0c2
also allow unicode, bytes, bytearray, dict -> fixes #297
2014-04-07 16:16:31 +02:00
Dave Halter