simplify MoveDataBuilder::move_path_for with new UnDerefer::co_iter_projections

[beepster4096]

• I did not use an LLM to create a change in this PR.
• I used an LLM to create a change in this PR, and I have explained below how it was used.

Split off an old design of #160780

MoveDataBuilder::move_path_for has been kind of messy since Derefer was moved before drop elaboration because of borrowck problems with the UnDerefer::iter_projections iterator. This PR adds co_iter_projections, returning a ProjectionCoroutine which takes a &UnDerefer for each iteration. Now, move_path_for doesn't have an outlined copy of add_move_path anymore which is a lot nicer.

Since my other PR doesn't need this anymore, if its a perf regression I'll probably just close the PR.

Issues:

• Is it a major perf regression from re-accessing the hashmap??
  • At least should be identical for places that don't have derefs in the middle
• I'm not sure if co_iter_projections is the best name

Possible alternatives:

• mem::take on self.data.rev_lookup.un_derefer?

[rustbot]

r? @oli-obk

rustbot has assigned @oli-obk.
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 75 candidates
• Random selection from 19 candidates