Only produce python3 warning with python3 warnings enabled#24
Only produce python3 warning with python3 warnings enabled#24jaraco merged 1 commit intojaraco:masterfrom asottile:patch-2
Conversation
As it is, this warning is quite difficult to avoid in a number of circumstances that would automatically be fixed in python3.x. As such, this warning is often not useful. Here’s a few examples where I’ve encountered this that are a bit annoying to fix: - flake8: takes commandline arguments (sys.argv is native strings (bytes py2, str py3)) which are passed to configparser - flake8: passes string literals to configparser (string literals are native strings) - entrypoints: passes os.path.join(sys.path entry, …) to configparser (sys.path is native strings) These warnings only appear in python2.x and any sane library is already testing against python3.x where these would be errors. As such, I propose either outright removing this warning, or (as this PR does) guarding it by `sys.py3kwarning` which is only `True` when running with `python2 -3 ...`
| Return list of successfully read files. | ||
| """ | ||
| if PY2 and isinstance(filenames, bytes): | ||
| if PY2 and isinstance(filenames, bytes) and sys.py3kwarning: |
There was a problem hiding this comment.
This additional guard should have been added inside the if, only for the warnings.warn().
By adding it here, filenames = [filenames] is not executed.
This is breaking flake8 (which depends on configparser under Python 2) for me, setup.cfg of my project seems to be ignored with this change, while it works normally if I remove it.
I see that this code has just been removed in the master branch, are you planning to release configparser 3.7.0 soon?
There was a problem hiding this comment.
are you planning to release configparser 3.7.0 soon?
Yes. I actually thought I had. Indeed, I'd tagged it, just not pushed the releases. I'll do that now.
There was a problem hiding this comment.
Thanks! Both 3.5.3 and 3.7.0 fix the issue for me.
|
Apologies on this, I should have been more careful :( Thanks for fixing this! |
|
I too could have been more careful. I put too much trust in the test suite and the age of the (original) pull request. I found a broader lesson in the whole ordeal, though, which you may appreciate. |
4 participants
As it is, this warning is quite difficult to avoid in a number of circumstances that would automatically be fixed in python3.x. As such, this warning is often not useful. Here’s a few examples where I’ve encountered this that are a bit annoying to fix:
These warnings only appear in python2.x and any sane library is already testing against python3.x where these would be errors.
As such, I propose either outright removing this warning, or (as this PR does) guarding it by
sys.py3kwarningwhich is onlyTruewhen running withpython2 -3 ...