Skip to content

Conversation

sobolevn
Copy link
Member

Let's see how it goes!

@sobolevn
Copy link
Member Author

There are many stubtest failures. What do @JelleZijlstra and @srittau think about adding them to allowlist.txt for now with a special comment?

@srittau
Copy link
Collaborator

srittau commented Sep 12, 2025

As a first step, I think we should add --ignore-disjoint-bases to the stubtest call (with a FIXME comment). This should get rid of some of the failures.

@srittau
Copy link
Collaborator

srittau commented Sep 12, 2025

There are also quite a few entries that were stubtest bugs that were fixed and can be removed, which is good of course. I'm okay with adding the remaining entries (with FIXMEs to look at them more closely).

@brianschubert
Copy link
Member

I'm working on patches for the __new__ and @disjoint_base related errors. I could try adding them to this branch if you'd like to save some allowlist churn, else I'll open separate PRs

@sobolevn
Copy link
Member Author

@brianschubert awesome! Thank you, please feel free to push them here :)

error: grpc.aio.ClientCallDetails.__new__ is inconsistent, stub parameter "cls" differs from runtime parameter "_cls"
For reference:
https://github.com/python-excel/xlrd/blob/3a19d22014d7b3f3041b7188d21a653c18c709bf/xlrd/formula.py#L675

```
/home/runner/work/typeshed/typeshed/stubs/xlrd/xlrd/formula.pyi
  /home/runner/work/typeshed/typeshed/stubs/xlrd/xlrd/formula.pyi:66:9 - error: Mismatch between signature of __new__ and __init__ in class "Ref3D"
    Signature of __init__ is "(atuple: tuple[int, int, int, int, int, int, int, int, int, int, int, int]) -> None"
    Signature of __new__ is "(iterable: Iterable[int] = ..., /) -> Ref3D" (reportInconsistentConstructor)
1 error, 0 warnings, 0 informations
```

```
$ uv run -w xlrd python -c 'import xlrd; print(xlrd.formula.Ref3D.__new__ is tuple.__new__)'
True
```
@brianschubert
Copy link
Member

I pushed some commits which fix the __new__-related errors, remove the unused allowlist entries, and add allowlist entries for the remaining errors. I'll open a separate PR for the @disjoint_base errors to keep this diff smaller.

Copy link
Contributor

According to mypy_primer, this change has no effect on the checked open source code. 🤖🎉

@sobolevn
Copy link
Member Author

Amazing work, thank you!

@srittau srittau merged commit 4b0623f into main Sep 14, 2025
103 checks passed
@srittau srittau deleted the sobolevn-patch-1 branch September 14, 2025 01:33
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants