Emit a note that static bounds from HRTBs are a bug

[jackh726]

This note isn't perfect, but opening this to either 1) land as is or 2) get some feedback on how to improve it

Let r? @compiler-errors and cc. @nikomatsakis

[rustbot]

rustc_error_messages was changed

cc @davidtwco, @compiler-errors, @JohnTitor, @estebank, @TaKO8Ki

[jackh726]

@estebank maybe you have some thoughts on how to make this point to the whole clause?

[compiler-errors]

I thought a bit about this since you had mentioned it earlier -- it's probably not generalizable to point to the whole clause because we can have things like where Ty: Trait + OtherTrait -- the choice for the spans in predicates_of to point to just the trait part was probably made for that reason.

[jackh726]

Right, it's not simply enough to say "point to the entire clause. Instead, I'm imagining something like:

for<'a> T: Trait<'a>
^^^^^^^^^^^^^^^^^^^^

for<'a> T: Send + Trait<'a>
^^^^^^^^^^        ^^^^^^^^^

for<'a> T: Trait<'a> + Send
^^^^^^^^^^^^^^^^^^^^

[jackh726]

The problem is that we currently have no way (that I know of) to even get that span info from an InstantiatedPredicate. We lose that when generating the GenericPredicates, and there's no mapping back, AFAICT. I was looking a bit at instead storing like the owner and which predicate it is, but that also doesn't quite work, since lowering is done in a bit more complicated fashion.

[estebank]

Yeah, we do lose too much metadata when dealing with predicates :-/

Ideally the spans should be something along the lines of

for<'a> T: Send + Trait<'a>
^^^^^^^           ^^^^^^^^^

but it's perfectly fine to land as is.

[bors]

☔ The latest upstream changes (presumably #101560) made this pull request unmergeable. Please resolve the merge conflicts.

[compiler-errors]

r=me after rebased past conflict

[compiler-errors]

nit: there's some trivial lift macro i think

[jackh726]

There is not (that I could find)

[compiler-errors]

TrivialTypeTraversalAndLiftImpls is what it's called, which itself uses CloneLiftImpls if you don't want to derive the traversal/folder traits. But up to you to keep this manual impl.

[jackh726]

I'm going to wait for the stabilization PR to merge, since they're going to conflict and this will be easier to rebase.

[jackh726]

@bors r=compiler-errors

[bors]

📌 Commit aae37f8 has been approved by compiler-errors

It is now in the queue for this repository.

[Mark-Simulacrum]

Marking as a perf regression per results from this PR's merge in the rollup (#101805 (comment)). (For clarity: the results are from just this PR in that comment, and match the effects of the rollup as a whole).

It looks like we partially reverted in #101902, but that didn't fully recover the performance here.