Fixed Auto Discovery of .test files Fixes #8650 #10108
Conversation
mypy/test/testcheck.py
Outdated
|
|
||
| for check_files in os.listdir(path2loc): | ||
| # Append only if it has the pattern ( check-*-*.test) | ||
| if re.match(r'check-\w+-?\w+?.test', check_files): |
There was a problem hiding this comment.
Thanks for your contribution.
FIlename check-union-or-syntax.test is not matched by this regex.
There was a problem hiding this comment.
@kamilturek thanks for pointing this. I'll update the regex.
| if sys.platform == 'win32': | ||
| path2loc = os.path.abspath(__file__).split('mypy\\test')[0] + 'test-data\\unit' | ||
| else: | ||
| path2loc = os.path.join(os.path.abspath(__file__).split('mypy/test')[0], 'test-data/unit') |
There was a problem hiding this comment.
This is hard to reason about. Can you unify the cases so that the same implementation works on both Windows and other platforms? You can probably use os.path.join and/or os.sep.
| for check_files in os.listdir(path2loc): | ||
| # Append only if it has the pattern ( check-*-*.test) | ||
| if re.match(r'check-\w+-?\w+?-?\w+?.test', check_files): | ||
| typecheck_files.append(check_files) |
There was a problem hiding this comment.
Can you move this and the above logic to a helper module, so that this can be reused in other places easily?
For example, we could have something like this in this file:
typecheck_files = expand_test_files('test-data/unit/check-*.test')
You can use glob.glob to expand patterns instead of re, since regular expressions are more error-prone as they require escaping, etc.
|
Any progress on this PR? @MiHarsh |
|
Superseded by #13143 |
Fixes #8650
After this PR is merged, this while running testcheck.py , it would automatically create the list of .test files. The main benefit is that if in future there is addition of test files, we would not have to add the new file into testcheck.py manually, it would automatically grab them.
Further I have verified the changes on my local.
Please merge this PR. I am new to open source and this is my first PR.
Please review @JukkaL , @ethanhs and @davidzwa Sir