Incosistency between else_if_without_else and needless_else

[c410-f3r]

else_if_without_else is allowed in restriction while needless_else warns in style although both deal with the same pattern. It is then probably a good move to put both in the same category.

[y21]

I personally disagree. They deal with the opposite pattern rather, and one is (in my experience) more idiomatic than the other.
Requiring an else on every single else if even when it's not needed sounds quite restrictive. An empty else OTOH seems much more sane to warn and is likely a mistake, possibly leftover from refactoring.

The same sort of argument can be made for implicit_return (restriction) and needless_return (style). Omitting return; as the last statement of a function is idiomatic Rust so it seems good to have needless_return be warn by default.

[Alexendoo]

Yeah having both of these warn by default would create a cycle where we repeatedly ask the user to add an else {} then on the next run ask them to remove it

[c410-f3r]

Looks like this is a desired undocumented behaviour, or is there any?

For those interested in consistency, semicolon_inside_block and semicolon_inside_block among possibly others are in the same category with the same level of restriction.

[Alexendoo]

There are several pairs of lints that suggest opposites to each other, but the only case where they should be in the same category is if that category is restriction, meaning they're both considered individually opt in

[c410-f3r]

@Alexendoo #11666