fix const_item_mutation lint to use needs_drop instead of has_dtor

[Rachit2323]

Description:
The const_item_mutation lint was only checking if the outer type has a direct Drop impl (has_dtor). But if a field inside the type has a Drop impl, that drop logic can also observe the mutation — so the warning should be suppressed in that case too.

This fixes a false positive where the lint would warn on code like:

O.inner.val = 42;

even when Inner has a Drop impl that prints the value — meaning the mutation IS observable and the warning is wrong.

Fix: replace has_dtor check with needs_drop, which checks the whole type including all fields inside it.

[rustbot]

Some changes occurred to MIR optimizations

cc @rust-lang/wg-mir-opt

[rustbot]

r? @jackh726

rustbot has assigned @jackh726.
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 74 candidates
• Random selection from 17 candidates

[rabindra789]

Nice fix! While testing this branch, I noticed one behaviour change. on the base commit this lint warns for traits associated consts whose type mentions Self, but on this branch it doesn't. Is that expected, or should this case still lint?

[Rachit2323]

Thanks for catching this! It was unintended. Fixed it — for types containing type parameters like Self, needs_drop is too conservative so I now skip the suppression and let the lint fire as before.

[rabindra789]

Thanks for checking and fixing it! Glad we caught that regression.

View changes since this review

[rustbot]

This PR was rebased onto a different main commit. Here's a range-diff highlighting what actually changed.

Rebasing is a normal part of keeping PRs up to date, so no action is needed—this note is just to help reviewers.

[jackh726]

If you can squash, r=me.