Fix python linting errors

[tgross35]

These were flagged by ruff, run using the config in #112482

[rustbot]

r? @Mark-Simulacrum

(rustbot has picked a reviewer for you, use r? to override)

[tgross35]

@rustbot label -T-bootstrap -T-compiler -T-libs -T-rustdoc

[rustbot]

The Miri subtree was changed

cc @rust-lang/miri

Some changes occurred in compiler/rustc_codegen_gcc

cc @antoyo

[GuillaumeGomez]

Changes related to rustdoc nd rustc_codegen_gcc look good to me. Could you add the lint check into CI too please?

[tgross35]

Thanks, I'm in talks about doing that now at #112482 🙂

[tgross35]

All of these changes are trivial best practice items: == None -> is None, and unused variables or imports. The
miri runner could be a serious gotcha (as-is, setting env within the test functions would quietly update the environment for all future runs - hope you never set MIRI_SKIP_UI_CHECKS=1)

[saethlin]

Yeah I looked over the Miri changes. That's a classic footgun, thanks for the fix.

[RalfJung]

I don't understand the miri script diff, could you explain?

[saethlin]

Python default arguments are initialized when the function definition itself is evaluated, not when the function is called.

deffunc(key, x={}):
x[key] ="oof"print(x)
func("first")
func("second")

Prints:

{'first': 'oof'}
{'first': 'oof', 'second': 'oof'}

Therefore, it is best practice to always use one of Python's immutable types as the initializer for a optional argument. It's pretty rare to see anything other than None.

[tgross35]

Mutable containers as default args get instantiated once per definition (not once per call), and they get the same scope as the function itself. This can just lead to some surprises:

deftest(skip, env={}):
ifskip:
env[“MIRI_SKIP_EVERYTHING”] = “1”
print(skip, env)
test(skip=True)
test(skip=False)

output:

# expected
True, {‘MIRI_SKIP_EVERYTHING’: ‘1’}
# surprise!
False, {‘MIRI_SKIP_EVERYTHING’: ‘1’}

It’s working fine here since env isn’t modified within the function, but it’s just the kind of thing that’s easy to quietly “whoops” at some point

edit: I see you beat me to it while I was still editing @saethlin 🙂

[RalfJung]

Oh wow wtf, I had no idea.^^ Good catch (and a massive footgun in Python, who would ever want mutable default arguments...).

[Mark-Simulacrum]

@bors r+

[bors]

📌 Commit 22d00dc has been approved by Mark-Simulacrum

It is now in the queue for this repository.