Implement PEP 765 · Issue #130080 · python/cpython · GitHub

This issue serves as a tracker for implementing PEP-765.

Linked PRs

pythongh-130080: return in finally in subprocess.py

pythongh-130080: implement PEP 765

gh-130080: return in finally in subprocess.py (#130081)

pythongh-130080: implement PEP 765

pythongh-130080: move _Py_EnsureArrayLargeEnough to a separate header…

gh-130080: move _Py_EnsureArrayLargeEnough to a separate header so it…

pythongh-130080: move _Py_EnsureArrayLargeEnough to a separate header…

gh-130080: implement PEP 765 (#130087)

🚀

Final task, mark the the PEP as Final:

Once a PEP has been accepted, the reference implementation must be completed. When the reference implementation is complete and incorporated into the main source code repository, the status will be changed to “Final”.

https://peps.python.org/pep-0001/#pep-review-resolution

There are a bunch of warnings when executing the test suite now:

in Lib/test/test_grammar.py (e.g., test_break_in_finally)
in Lib/test/test_functools.py (e.g., test_double_wrapped_methods)
in Lib/multiprocessing/connection.py (but this one has already been reported in return-in-finally in multiprocessing/connection.py #131233)

I'm re-opening the issue so that we fix the issues in the tests but you can open a separate issue if you want.

pythongh-130080: fix warnings in tests

gh-130080: fix warnings in tests (#131471)

pythongh-130080: implement PEP 765 (python#130087)

pythongh-130080: fix warnings in tests (python#131400)

#130087 seems to have introduce a regression were AST folding is now applied even if it shouldn't be. E.g.

# test.py
match a:
  case -1+2j:
    pass

With Python 3.13 (or 3.14a6)

$ python3.13 -m ast test.py
        ...
            match_case(
               pattern=MatchValue(
                  value=BinOp(
                     left=UnaryOp(
                        op=USub(),
                        operand=Constant(value=1)),
                     op=Add(),
                     right=Constant(value=2j))),
               body=[...

After #130087

$ python -m ast test.py
        ...
            match_case(
               pattern=MatchValue(
                  value=Constant(value=(-1+2j))),
               body=[...

pythongh-130080: do not fold match case constants in unoptimized AST

gh-130080: do not fold match case constants in unoptimized AST (#131577)

#130087 seems to have introduce a regression were AST folding is now applied even if it shouldn't be. E.g.

Thanks, this is now fixed.

@iritkatriel I believe this memset is unnecessary:

cpython/Python/ast_opt.c

Line 943 in 90b82f2

memset(&state, 0, sizeof(_PyASTOptimizeState));

Here is PR: #131745

We do this in other places - set everything to 0 and then override other fields. If someone adds a new field which should be initialised to 0 they will probably not bother to update initialisation.

I see. Closing it then.

I just upgraded to 3.14 (from 3.13) and now I can no longer return a custom value from a function when returning early without copy and pasting my return value everywhere?

def test(fail = False):
    try:
        if fail:
            raise Exception('fail')
        else:
            print('try')
            return # Will never reach Done
    except Exception as e:
        print(e)
    finally:
        print('Finally')
        return (0, 1, 2) # Syntaxerror in 3.14
    print('Done')
    return (0, 1, 2) # Unreachable when returning early

Who tf comes up with this nonsense?!?

v1.8.9 遵循PEP 765规范重构代码。(python/cpython#130080)

Implement PEP 765 · Issue #130080 · python/cpython · GitHub

Repository navigation

Implement PEP 765 #130080

Linked PRs

4 remaining items

Metadata

Assignees

Labels

Projects

Milestone

Relationships

Development

Implement PEP 765 · Issue #130080 · python/cpython · GitHub

Repository navigation

Implement PEP 765 #130080

Description

Linked PRs

Activity

hugovk commented on Mar 17, 2025

picnixz commented on Mar 18, 2025

4 remaining items

cdce8p commented on Mar 22, 2025

iritkatriel commented on Mar 23, 2025

WolframAlph commented on Mar 25, 2025

iritkatriel commented on Mar 25, 2025

WolframAlph commented on Mar 28, 2025

defkev commented on Oct 10, 2025

Metadata

Metadata

Assignees

Labels

Projects

Milestone

Relationships

Development

Issue actions