Which is a good argument for suggesting the message be on by default, but not, IMHO, a reason to bypass linters entirely and do it ourselves.
I would generally agree, but just don’t really see what problem this pattern is solving. For me that is one of the most important conclusions of Irit’s research: all “valid” patterns are replaced with a trivial other pattern that avoids the pitfalls of the current one.
There’s certainly enough consensus here to go to ruff on behalf of the core team and say that this deserves an on-by-default warning.
At my day job we turn off all linter messages and only turn them on when somebody wants to spend the time to fix all issues, which you can imagine is almost never. I think it would still be good to indicate to ruff that this is a very useful message, but again that would likely not reach all audiences. (In fear of pinging them and involving them in a discussion they don’t want to get involved in, @charliermarsh this might be relevant for you and your team).
I’m not sure you can answer this, but would there be something that could convince you to support this proposal? For me the evidence of misuse and possibility of replacing it with a (in my opinion) clearer pattern is enough, but I’m wondering if there is other evidence that is currently lacking that could convince you.