failed to format a file with a with-statement and a named-expression

[15r10nk]

Describe the bug

The following code can not be parsed/formatted by black:

with {} as name_3, (name_2 := name_4):

    def name_3[**name_1]():
        pass

black reported the following error:

> black -l 100 -C bug.py
error: cannot format bug.py: Cannot parse: 1:26: with {} as name_3, name_2 := name_4:

Oh no! 💥 💔 💥
1 file failed to reformat.

but it can be parsed by cpython:

from ast import parse
parse(
    'with {} as name_3, (name_2 := name_4):\n'
    '\n'
    '    def name_3[**name_1]():\n'
    '        pass\n'
)

result:

returncode: 0

Environment

• Black's version: current main (dbb14ea)
• OS and Python version: Linux/Python 3.12.6 (main, Sep 9 2024, 22:11:19) [Clang 18.1.8 ]

Additional context

The bug was found by pysource-codegen (see #3908)
The issue description above was created by a script. Let me know if you think it can be improved.
One think I noted is that the following code causes also a formatting failure:

with a, (name_2 := name_4):
        pass

But pysource-minimize did not found this minimal example for some reason.

[15r10nk]

The reason why it did not found a simpler example was that I had to pass the `--target-version' parameter.

[tusharsadhwani]

@15r10nk I am able to reproduce the smaller case without the target version parameter as well:

$ cat x.py
with x, (a := b):
    pass

$ black x.py
error: cannot format x.py: Cannot parse for target version Python 3.9: 1:10: with x, a := b:

Oh no! 💥 💔 💥
1 file failed to reformat.

[MeGaGiGaGon]

The issue happens inside remove_with_parens , which eventually ends up removing the () from the (a:=b) at

if isinstance(node.children[1], Node):
    remove_with_parens(node.children[1], node)

My guess at a fix would be that the if needs to get the and not any(leaf.type == token.COLONEQUAL for leaf in node.leaves()) check added to it.

[tusharsadhwani]

I just jumped into visit_atom because that's where the LineGenerator was deviating from expected output, and found a fix there.

[MeGaGiGaGon]

👍 much cleaner that what I found

[15r10nk]

@tusharsadhwani thank you for fixing this so fast. I have already created the next issue #4632.