Use hir::Path instead of ast::Path in HIR Attribute's.

[eddyb]

The goal here was to avoid using ast::Path in the HIR and having serialization logic for ast::{Ty,Expr,...} (inside generic arguments), despite them not existing for attribute paths.
We can't remove Nonterminals "as easily", but we bug! for those already in some places.

This seemed a trivial change at first, but now I'm not so sure, I've had to make a lot of code work on either AST or HIR Attributes, using generics (over the Path parameter).

Maybe we should just have path: Vec<Ident> in Attribute, instead of {ast,hir}::Path?
EDIT: opened #56492 for that variant.

r? @petrochenkov cc @nikomatsakis

[petrochenkov]

We can't remove Nonterminals "as easily", but we bug! for those already in some places.

That's what I immediately thought - even if AST paths are removed, there still can be pieces of AST in tokens.

Is it possible to bug!() on generic arguments in AST paths as well?
I think that would be the best solution for the short term problem of unblocking #56511.
(If some generic arguments still reach HIR as a consequences of error recovery, they can be filtered away during lowering.)

[petrochenkov]

Actually, looking at #56511 I no longer understand why pieces of AST in HIR is an issue.
#56511 removes encodable/decodable imple from HIR, but AST is still serializable, so what's the problem?

[petrochenkov]

Regarding attributes in HIR, I the long term goal is to have some semantic representation for attributes in HIR (hir::Attribute) based on name resolution results, and not just tokens.
This means no hir::MetaItem, hir::MetaItemKind etc, because no token parsing happens on HIR attributes, all the parsing and validation is done during AST -> HIR lowering.
hir::Attributes would probably still contain paths (hir::Path), but mostly for diagnostics.

In that sense this PR is closer to the long term goal than #56492.

[bors]

☔ The latest upstream changes (presumably #56557) made this pull request unmergeable. Please resolve the merge conflicts.

[XAMPPRocky]

Triage; @eddyb Hello, have you been able to get back to this PR?

[TimNN]

Ping from triage @eddyb: What is the status of this PR?

[Dylan-DPC-zz]

ping from triage @eddyb Unfortunately we haven't heard from you on this in a while, so I'm closing the PR to keep things tidy. Don't worry though, if you'll have time again in the future please reopen this PR, we'll be happy to review it again!