https://gcc.gnu.org/bugzilla/show_bug.cgi?id=24021
Richard Biener <rguenth at gcc dot gnu.org> changed: What |Removed |Added ---------------------------------------------------------------------------- CC|richard.guenther at gmail dot com |rguenth at gcc dot gnu.org --- Comment #15 from Richard Biener <rguenth at gcc dot gnu.org> --- I think that for match.pd we should rely mostly on global ranges which means extending SSA name ranges for floats. Indeed I think the most important parts will be tracking whether values can be NaNs, Infs or whether they are known to be negative, positive or zero.