Generate Deref assist for generic structs

[phyBrackets]

Fixing issue #20574

The fix ensures the assist correctly appears for generic structs without existing Deref implementations while maintaining accuracy for non-generic types.

[A4-Tacks]

It seems that hir::Type::impls_trait behavior has changed recently.
Previously, it always returned false for generic types, but now it always returns true

[ChayimFriedman2]

Hi @phyBrackets, are you still around?

[rustbot]

This PR was rebased onto a different master 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.

[phyBrackets]

Hi @ChayimFriedman2 @A4-Tacks Sorry for the delay, been busy with other stuffs, I've rebased the patch onto master and updated the fix, for generic structs, instead of relying on impls_trait (which now returns true for generic types regardless of actual implementations), the fix routes through has_actual_impl which checks for concrete impl blocks in the module.

Let me know if you have any other feedback.

[ChayimFriedman2]

I commented on the approach, but honestly I think this is entirely redundant. The behavior of impls_trait() is a bug, and we'll fix it (eventually), then there will be no need for this.

View changes since this review

[ChayimFriedman2]

A much simpler and more correct approach is to list the impls by Impl::all_for_type(db, strukt.ty(db)), then filter by Impl::trait_().

[ChayimFriedman2]

There is also no need to differentiate if the struct has generic args.

[rustbot]

☔ The latest upstream changes (possibly #22049) made this pull request unmergeable. Please resolve the merge conflicts.