Convert several for loops in dataclasses module to list comprehensions

[keithasaurus]

Feature or enhancement

Proposal:

There are several for loops in the dataclasses module where performance would presumably improve with list comprehensions. Arguably the code becomes a bit clearer as well.

Has this already been discussed elsewhere?

This is a minor feature, which does not need previous discussion elsewhere

Links to previous discussion of this feature:

No response

Linked PRs

• gh-114011: Convert for loops in dataclasses module to comprehensions #114012

[carljm]

It's rarely advisable to presume that something will improve performance without benchmarking it :) In all Python versions prior to 3.12, it's likely that if anything this conversion would have been a performance regression. In 3.12+ I expect it shouldn't be a regression anymore, due to comprehension inlining, but I doubt it will be a meaningful improvement, either. Without benchmarks, we should not consider performance as a factor.

I do personally find the code a bit nicer with the comprehensions, but we don't usually take PRs for such minor and subjective code changes, just due to the inherent risk of any change.

[terryjreedy]

Yesterday I, as main idlelib maintainer, replaced 6 awkward lines in an IDLE test method, including a nested function, with a much clearer one-line dict comprehension. If I were the main dataclass maintainer, I would likely merge the two return changes, maybe with a format tweak, as they improve the parallelism of the if-else returns. But think the current dataclass module experts, @ericvsmith and @carljm should decide this one. I might check whether the affected functions are covered by tests.