trait-system: Recover deferred closure calls after errors

[Dnreikronos]

fixes #157951

this already reported the right errors, then typeck kept going and hit the deferred closure call path. that path assumed it could always find a fn trait impl after closure kind inference. with this repro, the earlier errors mean that lookup can fail, so it should recover instead of iceing.

i think keeping this as recovery is the least surprising fix here. the compiler has already told the user what's wrong, so turning the later invariant into another hard failure doesn't buy much.

also drops the weird e0746 help for this closure case. there's no written ret ty to edit, so suggestions like \impl f\ or \�ox<dyn box::new(f)>\ were just noise. added the next-solver ui test for the repro.

[rustbot]

r? @nnethercote

rustbot has assigned @nnethercote.
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

Why was this reviewer chosen?

The reviewer was selected based on:

• Owners of files modified in this PR: compiler
• compiler expanded to 73 candidates
• Random selection from 19 candidates

[nnethercote]

Can you squash the commits together? I don't think they're particularly meaningful in isolation. Once that's done, looks good.

View changes since this review

[nnethercote]

@rustbot author

[rustbot]

Reminder, once the PR becomes ready for a review, use @rustbot ready.