refine pyupgrade's TimeoutErrorAlias lint (UP041) to remove false positives

[BurntSushi]

Previously, this lint had its alias detection logic a little
backwards. That is, for Python 3.11+, it would only detect
asyncio.TimeoutError as an alias, but it should have also detected
socket.timeout as an alias. And in Python <3.11, it would falsely
detect asyncio.TimeoutError as an alias where it should have only
detected socket.timeout as an alias.

We fix it so that both asyncio.TimeoutError and socket.timeout are
detected as aliases in Python 3.11+, and only socket.timeout is
detected as an alias in Python 3.10.

Fixes #8565

Test Plan

I tested this by updating the existing snapshot test which had erroneously
asserted that socket.timeout should not be replaced with TimeoutError in Python
3.11+. I also added a new regression test that targets Python 3.10 and ensures
that the suggestion to replace asyncio.TimeoutError with TimeoutError does not
occur.

[github-actions]

ruff-ecosystem results

Linter (stable)

✅ ecosystem check detected no linter changes.

Linter (preview)

✅ ecosystem check detected no linter changes.

[alex-700]

@BurntSushi hi! Thanks for the PR!

If I understand correctly, you decide to leave linting asyncio.TimeoutError even for py310 (but make the fix unsafe).
IMHO, it looks a bit counterintuitive.
The user, targeted to py310, can do nothing with the triggered lint about asyncio.TimeoutError, but suppressing.
Should we just do not produce a warning about asyncio.TimeoutError in the case of py310?

[charliermarsh]

Ah yeah, I think we may have misdiagnosed in the original issue. We probably shouldn't raise on asyncio.TimeoutError in Python 3.10 at all. The initial PR actually suggests that this was the intention:

But then has this code:

/// Return `true` if an [`Expr`] is an alias of `TimeoutError`.
fn is_alias(expr: &Expr, semantic: &SemanticModel, target_version: PythonVersion) -> bool {
    semantic.resolve_call_path(expr).is_some_and(|call_path| {
        if target_version >= PythonVersion::Py311 {
            matches!(call_path.as_slice(), [""] | ["asyncio", "TimeoutError"])
        } else {
            matches!(
                call_path.as_slice(),
                [""] | ["asyncio", "TimeoutError"] | ["socket", "timeout"]
            )
        }
    })
}

We probably want to return both errors if >= Python 3.11, and only socket.timeout error if Python 3.10?

[alex-700]

@charliermarsh yes.
Something like

/// Return `true` if an [`Expr`] is an alias of `TimeoutError`.
fn is_alias(expr: &Expr, semantic: &SemanticModel, target_version: PythonVersion) -> bool {
    semantic.resolve_call_path(expr).is_some_and(|call_path| {
        if target_version >= PythonVersion::Py311 {
            matches!(
                call_path.as_slice(),
                [""] | ["asyncio", "TimeoutError"] | ["socket", "timeout"]
            )
        } else {
            matches!(call_path.as_slice(), [""] | ["socket", "timeout"])
        }
    })
}

^ to minimize a difference from the current code.

Or maybe something like

/// Return `true` if an [`Expr`] is an alias of `TimeoutError`.
fn is_alias(expr: &Expr, semantic: &SemanticModel, target_version: PythonVersion) -> bool {
    semantic
        .resolve_call_path(expr)
        .is_some_and(|call_path| match call_path.as_slice() {
            [""] => true,
            ["socket", "timeout"] if target_version >= PythonVersion::Py310 => true,
            ["asyncio", "TimeoutError"] if target_version >= PythonVersion::Py311 => true,
            _ => false,
        })
}

or (implicitly used the information that target_version >= py310)

/// Return `true` if an [`Expr`] is an alias of `TimeoutError`.
fn is_alias(expr: &Expr, semantic: &SemanticModel, target_version: PythonVersion) -> bool {
    semantic
        .resolve_call_path(expr)
        .is_some_and(|call_path| {
            matches!(call_path.as_slice(), [""] | ["socket", "timeout"])
            || matches!(call_path.as_slice(), ["asyncio", "TimeoutError"] if target_version >= PythonVersion::Py311)
        })
}

P.S. I am actually does not understand why the check for [""] is here 😄

[BurntSushi]

Right, the unsafe fix on Python 3.10 for asyncio.TimeoutError was what I was uncertain about. I structured this PR such that all I have to do is drop the last commit. I fixed the logic as @charliermarsh indicated above in the first commit. :-)

[BurntSushi]

Okay, I've dropped the last commit and updated the PR description.

[charliermarsh]

Excellent, thanks!