Add lint could_be_unsized

[Jarcho]

fixes: #1368
fixes: #2308

Worklist of things still to do:

• Finish checking types for constraints. e.g. dyn Iterator<Item = T> would require T be Sized
• Actually resolve dependencies. e.g. T: Foo<U> need to check the constraint from Foo to determine whether T and U need to be Sized
• Check impl bodies for types. e.g. struct Foo<T>(..) should consider impl<T> Foo<T> {..} to determine it T could be ?Sized (will be done in a future PR)
• Check impl bodes for traits. e.g. impl<T> Foo<T> for Bar<T> {..} could be T: ?Sized

changelog: Add lint could_be_unsized

[rust-highfive]

r? @giraffate

(rust-highfive has picked a reviewer for you, use r? to override)

[Jarcho]

This should be mostly done now. Just need to take another pass through the code and clean things up. Linting types can wait for another PR, but everything else is linted.

There is a false positive when adding constraints to trait associated types. e.g.

trait Assoc {
    type Type: ?Sized;
}
fn foo<T: Assoc<Type = T>>(x: &T) -> &T::Assoc where T::Assoc: Sized { ... }

This will suggest add ?Sized to T even though that won't compile. I don't think it's a blocker. Unsized associated types are unusual to start with.

I'd like to keep it opt-in right now due to the possibility of ICE's. It runs successfully on all the crates in lintcheck, but I'd like better testing before enabling it by default.

[Jarcho]

I think it's ready now. It can remain in the nursery for a bit.

[bors]

☔ The latest upstream changes (presumably #8576) made this pull request unmergeable. Please resolve the merge conflicts.

[Jarcho]

@rust-lang/clippy ping to whoever wants to take this one.

[bors]

☔ The latest upstream changes (presumably #10415) made this pull request unmergeable. Please resolve the merge conflicts.

[Jarcho]

r? clippy

[Manishearth]

Overall structure looks good but I don't have the time for a thorough review.

r? @blyxyas you might find this one fun

[rustbot]

Could not assign reviewer from: blyxyas.
User(s) blyxyas are either the PR author, already assigned, or on vacation, and there are no other candidates.
Use r? to specify someone else to assign.

[blyxyas]

I'll take a look

r? blyxyas

[bors]

☔ The latest upstream changes (presumably #12259) made this pull request unmergeable. Please resolve the merge conflicts.

[xFrednet]

Hey @blyxyas, this is a ping from triage. Would you mind taking a look at this PR if you have the time? Otherwise, you can reassign it, by commenting r? clippy.

[blyxyas]

Yeah, I had the intention of looking at it because it has been been rolled like 3 times. But with how many reviews I get assigned, + the performance stuff, I simply son't see any time to review this.

r? clippy

[xFrednet]

Hey, this is a ping from triage. @dswij can you give this PR a review? It's totally fine if you don't have the time right now, you can reassign the PR to a random team member using r? clippy.

@rustbot ready

[xFrednet]

Guess I can take this one on.

@Jarcho would you mind rebasing? Then we can also see a better lintcheck sample size :D

[xFrednet]

r? xFrednet

[xFrednet]

This is still on my todo list, but I didn't have the time to take a look at it in the last month. The next week will also be full, but then I'll hopefully have time to review this :)