Suppress garbled suggestions from strict provenance lints in macros

[Dnreikronos]

The strict provenance lints (lossy_provenance_casts, fuzzy_provenance_casts) build suggestions using span arithmetic (shrink_to_lo(), shrink_to_hi(), .to()). When the cast sits inside a macro, those span operations produce broken output like $e as ).addr() because the spans don't map cleanly back to source text.

The fix wraps suggestion fields in Option and checks can_be_used_for_suggestions() before constructing them. The lint itself still fires, you just don't get the garbled suggestion. Same approach already used in op.rs and static_mut_refs.rs in this crate.

What changed

• errors.rs: sugg fields in LossyProvenanceInt2Ptr and LossyProvenancePtr2Int are now Option<...>
• cast.rs: suggestion construction in both lossy_provenance_ptr2int_lint and fuzzy_provenance_int2ptr_lint is guarded behind can_be_used_for_suggestions()
• New regression test confirms both lints fire on casts inside macros, with no suggestion block in output

Fixes #156850

[rustbot]

r? @TaKO8Ki

rustbot has assigned @TaKO8Ki.
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

[mu001999]

@rustbot author

r? me

View changes since this review

[rustbot]

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

[Dnreikronos]

@rustbot ready