https://gcc.gnu.org/bugzilla/show_bug.cgi?id=97286
--- Comment #15 from Hongtao Liu <liuhongt at gcc dot gnu.org> --- (In reply to Andrew Pinski from comment #14) > SLSR is definitely the wrong place for this. This is more like an out of ssa. > > Also it needs to handle more than just VCEs, it needs to handle "nop" casts > too. Any testcase to show nop casts, here in the PR only VCEs create redundant moves. > > Oh LLM patches are questionable; even though GCC does NOT currently have a > policy; it is close to have one. If we base on what binutils (and the sound > of it, glibc) have it will ban patches which are co-authored by a LLM. But according to [1] Using an LLM to inspire or help create a patch might be OK, and I have manually reviewed and refined the code to make it follow GCC design. Or maybe you mean in that case we should remove Co-Authored-By: Claude Sonnet 4.5 <[email protected]> in the commit message? [1] https://sourceware.org/binutils/wiki/LLM_Generated_Content
