Support --shadow-file with --native-parser#21623
Conversation
This comment has been minimized.
This comment has been minimized.
ilevkivskyi
left a comment
There was a problem hiding this comment.
LG, but there are couple problems in tests.
| == Return code: 0 | ||
|
|
||
| [case testShadowFileWithNativeParserParallel] | ||
| # cmd: mypy --native-parser --shadow-file a.py a_shadow.py main.py a.py b.py |
There was a problem hiding this comment.
The test name says Parallel but the command line doesn't specify --num-workers. You should add e.g. --num-workers=4.
| reveal_type(x) | ||
| [out] | ||
| a.py:2: note: Revealed type is "str" | ||
| == Return code: 0 |
There was a problem hiding this comment.
I think it is worth adding an incremental test for --shadow-file in parallel mode as well.
This comment has been minimized.
This comment has been minimized.
| tmp/b.py:2: note: Revealed type is "builtins.int" | ||
|
|
||
| [case testIncrementalNativeParserShadowFileIntroduced] | ||
| # flags: --native-parser --num-workers=2 |
There was a problem hiding this comment.
Oh, btw, all testcheck.py tests are run in parallel mode in the parallel CI jobs. So I am not sure we need to keep both here. (The cmdline.test however is not run by testcheck.py, so keeping a separate parallel test there makes sense.)
|
According to mypy_primer, this change doesn't affect type check results on a corpus of open source code. ✅ |
No description provided.