Hello! I have sent an explanation on ICE in try_combine on pr112494.c [1],and an argument that explains why we can safely ignore non-COMPARISON_P mode changes [2].
Can we proceed with the proposed solution? [1] https://gcc.gnu.org/pipermail/gcc-patches/2023-November/638726.html [2] https://gcc.gnu.org/pipermail/gcc-patches/2023-December/639137.html Thanks, Uros.
