https://gcc.gnu.org/bugzilla/show_bug.cgi?id=128867
--- Comment #3 from Haoxin Tu <haoxintu at gmail dot com> --- (In reply to Drea Pinski from comment #1) > Confirmed. I bet this is a dup of another bug which is older. The problem > is not with using the ranger but rather the ranger does not handle fp > correctly for comparisons. Thanks for your quick confirmation! I tried my best not to report duplicate cases. Indeed, there are some relevant but different cases - PR 126535 (ranger op1_range on the UNEQ edge of islessgreater; fails since GCC 13, while this one is trunk-only); - PR 126545 (you just mentioned; same, this only fails on the trunk).
