https://gcc.gnu.org/bugzilla/show_bug.cgi?id=86975
sandra at gcc dot gnu.org changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |sandra at gcc dot gnu.org
--- Comment #1 from sandra at gcc dot gnu.org ---
Given that this code is originating in the tree reassoc1 pass, wouldn't it be
better to just make that pass prefer to generate a code sequence with an
unsigned constant on all targets? It should work just as well as the current
negative-constant expansion does on targets that support unsigned constants,
right?
