-
-
Notifications
You must be signed in to change notification settings - Fork 1.2k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Require python 3.7.2+
#5921
Require python 3.7.2+
#5921
Conversation
Pull Request Test Coverage Report for Build 2147752109
💛 - Coveralls |
With respect to coverage. The first decrease seems "correct", it is no longer covered. Do we want to remove it? The second coverage decline comes from |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We could wait a little for this one to avoid issue with Python 3.6 when backporting fixes on 2.13. What do you think?
I think we should remove dead code no longer accessed. |
Fine with me! Shall we say first week of April? |
Alternatively we could merge this release 2.14.0 without python 3.6 right now, and move everything that is in 2.14 now to 2.15.0 (then we support 2.14 and python 3.7). |
I would suggest to wait a bit longer before dropping 3.6, even if it's only in a dev branch. Once it's committed, there sure will be a lot of cleanup and update PRs which will make bugfixes / backports even more difficult. A few weeks more or even a month doesn't make much of a difference IMO.
Additionally, it would be better not to release new minor versions with so little time after one another. The fast bugfix releases are much appreciated though! |
The idea I had (following #5988 where the test fail for python 3.6) is to not do any bug fixes for python 3.6 / pylint 2.13 and just support pylint 2.14 / python 3.7. |
🤔 Not sure about that. I would still recommend to wait. I agree pylint 2.14.0 should be the one which requires 3.7, but it would make more sense to include the whole cleanup there as well. Until we're ready, bugfixes for 2.13.x still need to be tested on 3.6. After all, the particular issue with 3.6 wasn't a big deal and easy to resolve. It should be possible to support it for a few more weeks now, too. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Not a full review, but from what I've seen this looks about right.
Some notes
- We didn't add CI tests for
pypy-3.8
, right? - This should definitely be done separately, but we should make sure to update
py-version
. Just a warning, that might trigger a lot a pylint typing warning 😄
https://github.com/PyCQA/pylint/blob/96d7e898cc863a6854cd0870ba4069c8f7d7851b/pylintrc#L56-L57
@@ -361,12 +361,9 @@ def _check_broken_noreturn(self, node: Union[nodes.Name, nodes.Attribute]) -> No | |||
if ( | |||
isinstance(inferred, (nodes.FunctionDef, nodes.ClassDef)) | |||
and inferred.qname() in TYPING_NORETURN | |||
# In Python 3.6, NoReturn is alias of '_NoReturn' | |||
# In Python 3.7 - 3.8, NoReturn is alias of '_SpecialForm' |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I would leave the second line.
No we need to exclude line number checking. Something for another PR :)
I tried, and yeah that's also something for a follow-up 😄 See OP. |
There's a checklist in #4683 I assumed we were going to do it later on, but doing the whole checklist here would make sense. |
Not so sure about that. It might be easier to review if they are split up into multiple PRs. |
@Pierre-Sassoulas Shall we merge this Monday 11th? We could do one final hotfix this weekend if needed without the trouble of having the hotfix supporting |
Yes let's do it. I think we can (try to) release 2.14 fairly soon after. This way we don't have to backport to 2.13 and think about python 3.6 anymore |
There is quite a lot still on the |
The plan is still to merge this after a likely 2.13.6? Or are we cancelling that milestone? |
I think so, but I think we can already finish the review of the changes here. I'll fix the merge conflict first though. |
I need to create a test for the namespace changes but I've been quite busy lately and the main blocker for 2.13.6 is #6141 that needs to be reviewed by Marc ideally. We could release 2.13.6 without this issue in the worst case scenario. And the issue with 3.6 would be manageable anyway. I think we can merge this one. |
CONTRIBUTORS.txt
if you are a new contributor.doc/whatsnew/<current release.rst>
.Type of Changes
Description
Two additional follow up PRs would be necessary after this:
min_pyver=3.7
. That would be quite a large diff so I have excluded it from this PR.py-version
inpylintrc
. This causes a number of warnings about usingfrom __future__ import annotations
. So that warrants its own PR.Just a draft to see if I got everything that needs to be changed for an initial PR.
Closes #4310
First step toward #4683