https://gcc.gnu.org/bugzilla/show_bug.cgi?id=84487
Peter Bergner <bergner at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Status|RESOLVED |REOPENED
CC| |bergner at gcc dot gnu.org
Resolution|FIXED |---
--- Comment #31 from Peter Bergner <bergner at gcc dot gnu.org> ---
(In reply to Christophe Lyon from comment #30)
> Since this was backported to gcc-8-branch, I've noticed:
> FAIL: gfortran.dg/pr51434.f90 -O2 execution test
> FAIL: gfortran.dg/pr51434.f90 -Os execution test
> on arm and aarch64
I'm seeing this on powerpc64le-linux as well, also only on FSF 8 branch. I'm
unsure if the patch is at fault or whether it's exposing a latent issue, since
I can get the tests to pass if I add -fno-tree-vrp. Maybe there's some
interaction between this patch and VRP?
I'm reopening the bug until this is resolved.
