Release 1.8 planning

[wesleywright]

I'm planning to make a 1.8 release sometime around Dec 15.

Please post here any issues you'd like to see fixed in the release, or PRs you'd like to have merged. I'll cut the release branch around early next week.

Tasks

Beta Give feedback

1. Fix crash with type alias to Callable[[Unpack[Tuple[Any, ...]]], Any] #16541
   +0
2. Allow type ignores of PEP 695 constructs #16608
   +0
3. Cherry-pick 16541 once landed
4. Cherry-pick 16608 once landed
5. cherry-pick 16655
6. Don't check plugin-generated functions #16524
   +0

[AlexWaygood]

Sounds great! Here's a few PRs that might be worth considering:

• I've got what I think is a fix for a crash here. The fix might be a bit naive (not sure), but I'd love it if someone could take a look: Fix crash with type alias to Callable[[Unpack[Tuple[Any, ...]]], Any] #16541.
• This one fixes a regression in mypy 1.7: Don't check plugin-generated functions #16524
• This one doesn't really help us in any way, but it's a very easy change for us to make and it makes life easier for some distros, so why not: Use python2 instead of docutils to test cases with no stubs available #16513
• I'd love to see this one get in, if possible. It looks pretty simple, and it exposes some typing errors in CPython's Argument Clinic tool (which I help maintain): Do not intersect types in isinstance checks if at least one is final. #16330

[AlexWaygood]

There's a few more outstanding issues reporting regressions in mypy 1.7 that are worth noting, but I don't think any of them have PRs linked to them yet. I personally wouldn't hold up the 1.8 release for any of them:

• [Regression 1.6.x -> 1.7] Invalid "Returning Any from function declared..." #16499
• Regression with covariant type through built-in function #16476
• Possible regression when narrowing using isinstance #16462
• I think this one was mostly fixed already, but not sure: Various stubgen regressions in mypy 1.7.0 #16486. (I believe Fix stubgen regressions with pybind11 and mypy 1.7 #16504 fixed a lot of the reported problems.)

[hauntsaninja]

I'd like to see #16608 merged

[wesleywright]

I've cut a release-1.8 branch for this release. It's based on the latest master commit (7c33e7c) so it should already contain #16504, #16330, and #16513.

[wesleywright]

cherry-picked #16608 onto release-1.8

[wesleywright]

cherry-picked #16541. hopefully #16524 lands in the next couple of days.

[AlexWaygood]

Looks like our CI is broken on master (https://github.com/python/mypy/actions/runs/7186189373/job/19571043378), apparently because of pypa/hatch#1113. (We simulate various typed packages as part of our tests; they all use hatchling as their backend.)

[hauntsaninja]

Thanks, #16655 to fix that. (We switched to hatchling because setuptools editable installs no longer work well with type checkers)

[hauntsaninja]

I'd also say we shouldn't block 1.8 on #16524. I can make a point release for it whenever its merged.

[wesleywright]

Cherry-picked #16655. I'll likely finish out the release in the next couple of days; I'll cherry-pick #16524 if it lands before then, otherwise I'll just move forward.

[cdce8p]

It might be a bit late but the last typeshed update #16665 includes some useful improvements for asyncio functions with callbacks. Those are among the first uses of TypeVarTuples in typeshed.

• Use PEP-646 in stubs for asyncio typeshed#11015 by @AlexWaygood

[wesleywright]

1.8 release blogpost: https://mypy-lang.blogspot.com/2023/12/mypy-18-released.html

[intgr]

@ymyzk Can you please add mypy 1.8.0 to mypy-play.net?

[ymyzk]

@intgr Thanks for the reminder. Yes, I just added mypy 1.8.0: https://mypy-play.net/?mypy=1.8.0&python=3.12&flags=version

[meshy]

Should this issue be closed now that 1.8 has been released?