Add unnecessary_underscores to core/recommended set.

[kallentu]

Describe the rule you'd like to see added and to what rule set

https://dart.dev/tools/linter-rules/unnecessary_underscores

When a user is using multiple underscores (__) and that variable isn't referenced, they should use a wildcard variable _ instead.

Enabling this lint will allow us to avoid that pattern which is also the point of having the wildcard variables feature 😀

This should at least be in the recommended lint set.
I'd love to see it in the core lint set though.

We should be encouraging users to use wildcard variables where they can and avoid multiple underscores all together. Wildcard variables provide static errors when used and would prevent errors like referencing an intentionally unused parameter by accident.

Additional context

See paragraph above.

[natebosch]

I'm in favor. It's mostly a style fix but I'm inclined to put it in the core set.

[lrhn]

I'm not sure it warrantes being in core.
It's not any kind of warning sign that code contains __, it's just old code that still works just as well.
It's unlikely to cause an error, it's likely not in public API. It's a small implementation detail that there is no reason to ding people points for on Pub.dev.

[natebosch]

it's likely not in public API.

For public API this might warrant inclusion in the core set, but you're right that it will be incredibly rare. It's definitely not worth splitting the implementation just to be able to include it in the core set.

no reason to ding people points for on Pub.dev.

There's also no (good) reason for maintainers of published packages who are anyways upgrading their language version to not apply the quick fixes for this lint.

I'd be happy enough to see this in the recommended set, but I do think it is likely to do more good (alert maintainers of a readability improvement they may have overlooked) than harm (annoy maintainers of published packages who prefer ugly code to any amount of churn?) if we put it in core.

[lrhn]

The lint should also only apply to code that is 3.7 or above (that's where we released it, right?), so if you're not touching your old code, you won't lose points.
If you move your code base to 3.7, you'll get the warning from a recommended lint anyway and will likely just fix it.

I don't think it'll matter much for adaption whether the lint is in core or not. I expect authors that edit their code to use the recommended lints.
I do prefer to only put lints in core if they matter, meaning that someone else seeing that the code doesn't satisfy the core lints should feel cautious about the code quality.
I don't think this lint qualifies.