Multiple diagnostic paths missing CoroutineClosure, worse async closure error messages

[SebTardif]

Bug

Multiple diagnostic and lint code paths match ty::Closure and/or ty::Coroutine but miss ty::CoroutineClosure, producing less precise error messages or lint false negatives for async closures.

Affected locations

File	Line	Impact
rustc_borrowck/src/diagnostics/explain_borrow.rs	213	Shows raw type instead of "closure" for async closure destructors
rustc_borrowck/src/diagnostics/region_errors.rs	182	is_closure_fn_mut returns false for async FnMut closures, causing less specific region error
rustc_lint/src/dangling.rs	105	DANGLING_POINTERS_FROM_LOCALS lint silently skips async closures
rustc_hir_analysis/src/hir_ty_lowering/errors.rs	973	Missing "point at closure" span for async closure trait bound failures
rustc_hir_typeck/src/fn_ctxt/checks.rs	1400, 1628	Missing "callable defined here" note and call-site matching for async closures
rustc_hir_typeck/src/method/suggest.rs	1827	Missing "point at closure" span in method resolution errors (same pattern as hir_ty_lowering)
rustc_trait_selection/src/error_reporting/infer/mod.rs	1782, 1850, 2556	Missing "closure has signature" note, missing "does not fulfill lifetime" suggestion, and missing TyCategory for async closures

Pattern

Same CoroutineClosure-missing-from-match pattern found across Rounds 63-64. Each is a one-line addition of ty::CoroutineClosure(..) to an existing match arm.

Severity

All P-low, diagnostic/lint quality only. No ICEs or wrong codegen. The most impactful is the dangling.rs lint false negative, which silently skips async closures returning dangling pointers.