Release 1.1 planning

[koogoro]

I'm planning to make release 1.1 by the end of February.

Please post here any issues you'd like to see fixed in the release, or PRs you'd like to have merged into the release branch

[Avasam]

It's small, but I'd really like #14437 for QoL improvement in stubtest.
Another typeshed QoL improvement (and mypy bug with aliased properties) that would be nice is #13975 / #6700
and stubtest name mangling weirdness: #14560

Out of all my issues, these are the 3 I care more about right now and would also help typeshed.

[cdce8p]

Fix for a crash with dataclass_transform

• Fix crash with dataclass_transform cache #14734

Edit:

• [dataclass_transform] detect transform spec changes in incremental mode #14695 resolves the issue as well. Would be great if that could be cherry-picked to release-1.1.

--
Additionally, this one would resolve one of the issues with pydantic dataclasses. The other one needs to be fixed by pydantic itself.

• [dataclass_transform] include __dataclass_fields__ in transformed types #14752

[hauntsaninja]

+1 to getting in the currently open dataclass_transform PRs

[koogoro]

It looks like #14437 was included in the base of the release branch, and #14695 and #14752 have been cherry-picked.

[hauntsaninja]

Could cherry pick #14792

[tmke8]

Support for @override would be nice. The following PR implements it:

• Add support for PEP 698 - override decorator #14609

But there was an open question about the interaction with @overload.

[koogoro]

Cherry-picked #14792. It might be better to wait for the next release for a new feature, since the release is already a bit late at this point.

[AlexWaygood]

Argh, you might have to cherry-pick these two to fix the mypyc build on the 1.1 branch — sorry:

• Support next types-setuptools update #14781
• Unpin the types-setuptools build dependency #14787

[AlexWaygood]

Here's a PR to fix the mypyc build on the 1.1. branch:

• [Release 1.1] Cherry-pick some mypyc build fixes #14820

[Viicos]

By any chance do you think this one could be included as well?

• Allow any type to be returned by __new__ #14471

It has been requested by a couple of people, and unfortunately I'm kind of stuck on this one. If you think it is fixable without too much hassle I'd be grateful to get help on this one!

(I'm posting it here as it might be considered as a feature and not a fix according to the new release versioning scheme).

[koogoro]

@Viicos Let's wait on this change until the next release. These are intended to happen monthly, so it shouldn't be too long until the next one. In general, I'd prefer not to cherry-pick changes onto the release at this point unless it's fixing a bug that would be introduced by this release.

[Viicos]

@Viicos Let's wait on this change until the next release. These are intended to happen monthly, so it shouldn't be too long until the next one. In general, I'd prefer not to cherry-pick changes onto the release at this point unless it's fixing a bug that would be introduced by this release.

Sure thanks for the answer. There's no hurry for this one anyway

[koogoro]

Release made! Here's the blog post: https://mypy-lang.blogspot.com/2023/03/mypy-111-released.html
(The version number is 1.1.1 since 1.1.0 was not building on wheels.)
I'll keep this open for a little bit to track potential bugs introduced in this release that we'd want to create an expedited 1.1.2 release for.

[AlexWaygood]

We've just had a regression reported over at typeshed:

• dataclasses.is_dataclass() assumed always true when passed type typeshed#9854

This regression was caused by python/typeshed#9362. It's already been fixed on the typeshed main branch via python/typeshed#9758, but the fix didn't make it into 1.1.1. (This is my error.)

[koogoro]

#14853 might be a candidate for a point release if the fix is simple enough.

[jchalupka-pan]

#14853 might be a candidate for a point release if the fix is simple enough.

@koogoro The issue has been fixed. Not hurrying up of course, I just don't know if you have been notified somehow :)

[wesleywright]

I've got a couple of dataclass_transform fixes inflight:

• [dataclass_transform] support implicit default for "init" parameter in field specifiers #14870 (not sure how big this will be when it comes out of Draft, but hopefully small?)
• [dataclass_transform] fix frozen behavior for base classes with direct metaclasses #14878 is a simple and straightforward fix

[AlexWaygood]

This fixes a (minor) regression; could be worth backporting:

• Add error code to typing_extensions suggestion #14881