fix: semicolon_inside_block don't fire if block is surrounded by parens

[ada4a]

changelog: [semicolon_inside_block]: don't fire if block is surrounded by parens

supersedes #15391, fixes #15380

I changed to approach to ignore blocks surrounded by parens, as suggested in #15391 (comment).

To do that, I took inspiration from https://rust-lang.zulipchat.com/#narrow/channel/257328-clippy/topic/attempted.20fix.20for.20disallowed_macros/near/532461016 and realized that this lint would benefit from being early-pass rather than late-pass, since that surfaces ExprKind::Paren which we now can explicitly ignore. Well, implicitly rather -- the lint now matches only on a StmtKind::Semi that contains a ExprKind::Block, without a ExprKind::Paren layer inbetween.

[rustbot]

r? @Alexendoo

rustbot has assigned @Alexendoo.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[github-actions]

Lintcheck changes for 168f1c5

Lint	Added	Removed	Changed
clippy::semicolon_outside_block	192	0	17

This comment will be updated if you push new changes

[ada4a]

So apparently the lint got smarter as a side effect of this PR.. I don't really understand what exactly caused that, but the suggestions do seem to be correct, so I've added a quick test case based on one of the added lint firings. If/when we figure out what caused the change, I'd of course give the commits and the test case more sensible names, and maybe minimize the test case further

[ada4a]

also no idea why the note: [...] implied by [...] thing disappeared -- probably because of the late-to-early change as well?..