[llvm] [MachineCopyPropagation] Recognise and delete no-op moves produced after forwarded uses (PR #129889)

[Martin Storsjö via llvm-commits]

mstorsjo wrote:

> Thanks in either case for the report. Do you mind giving #137348 a run with your original problem case to see if it fixes all the issues you are seeing or if some remain?

Thanks for the very quick fix! This does seem to avoid the issue; a Clang built with a Clang with this patch, no longer crashes for the single reproducer input that I had. I've started a longer full build chain which will try to build a larger number of test projects with the produced Windows/arm64 compiler, I'll know by tomorrow if that fixed things fully or not. But the fix for this issue does look good, thanks! (And it would have taken me ages to isolate, dig in and figure out what's broken and how to fix it!)

https://github.com/llvm/llvm-project/pull/129889