red-knot: adapt fuzz-parser to work with red-knot

[connorskees]

Adds a --bin argument to the fuzz-parser script, which makes it possible to fuzz either ruff or red_knot.

This fuzzing finds a few panics already in some seeds:

>>> python scripts/fuzz-parser/fuzz.py --bin red_knot 0-9
Running `cargo build --release` since no test executable was specified...
Concurrently running the fuzzer on 10 randomly generated source-code files...
Ran fuzzer successfully on seed 1                            [1/10]
Ran fuzzer successfully on seed 8                            [2/10]
Ran fuzzer successfully on seed 5                            [3/10]
Ran fuzzer successfully on seed 0                            [4/10]
Ran fuzzer successfully on seed 2                            [5/10]
Ran fuzzer successfully on seed 6                            [6/10]
Ran fuzzer successfully on seed 3                            [7/10]
Ran fuzzer on seed 7                                         [8/10]
The following code triggers a bug:

async def name_0[*name_1](name_2: name_0(), /):
    pass

Ran fuzzer on seed 9                                         [9/10]
The following code triggers a bug:

class name_5[**name_1](name_5[name_3]):
    pass

Ran fuzzer on seed 4                                        [10/10]
The following code triggers a bug:

class name_2[name_0](*name_2):
    pass

Bugs found in the following seeds:
4 7 9

This was tested manually by running the above command for --bin ruff and --bin red_knot.

Resolves #14157, though does not add the fuzzing to CI due to the existing failures.

[sharkdp]

This fuzzing finds a few panics already in some seeds:

class name_2[name_0](*name_2):

Probably similar to the panics here related to self-referential class bases:

ruff/crates/red_knot_workspace/tests/check.rs

Lines 269 to 272 in ac23c99

	// related to circular references in class definitions
	("crates/ruff_linter/resources/test/fixtures/pyflakes/F821_26.py", true, false),
	("crates/ruff_linter/resources/test/fixtures/pyflakes/F811_19.py", true, false),
	("crates/ruff_linter/resources/test/fixtures/pyupgrade/UP039.py", true, false),

And/or here: #14333

[github-actions]

ruff-ecosystem results

Linter (stable)

✅ ecosystem check detected no linter changes.

Linter (preview)

✅ ecosystem check detected no linter changes.

Formatter (stable)

✅ ecosystem check detected no format changes.

Formatter (preview)

✅ ecosystem check detected no format changes.

[MichaReiser]

Nice, thank you. Do you think it would be hard to instead test if the exit code isn't 1 or zero (panic should use something else) instead of adding the new flag to red knot?

[connorskees]

Oh I've always assumed panic! exited with 1, but that seems not to be the case. Great suggestion, thank you, that is a lot better

[AlexWaygood]

Thanks @connorskees, this is great! I pushed a few nitpicks, but nothing major.