Convert strict_provenance_lints to late lint passes

[hanna-kruppe]

Right now, these lints run during type checking of cast expressions, but they work fine as regular late lint passes, so this PR converts them to passes.

Besides the generic reasons to prefer regular lint passes over "hardwired" lints, the move solves one problem unique to cast type checking: emitting diagnostics directly in CastCheck::do_check causes problems for clippy reusing the cast type checking code.

Tracking issue: #130351

[rustbot]

r? @oli-obk

rustbot has assigned @oli-obk.
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 18 candidates

[hanna-kruppe]

I'm not sure if it makes sense to keep this test. From the information in #137588 it seems like:

1. The lint running and emitting a diagnostic is somehow a load-bearing part of the reproduction, so this test likely won't work as regression test any more with the changes in this PR.
2. Nobody ever pinned down the root cause of the ICE, figured out which change fixed it, or found a reproduction that doesn't trigger a bunch of other compiler errors along the way.

I don't know how to update the test so it still has a chance of being useful as a regression test. And it doesn't seem like there's anyone who understood #137588 well enough to help with this.

View changes since the review