http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59745
Jakub Jelinek <jakub at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |jakub at gcc dot gnu.org
--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 31792
--> http://gcc.gnu.org/bugzilla/attachment.cgi?id=31792&action=edit
gcc49-pr59745.patch
Totally untested fix. As this is only reproduceable in bootstrapped build, -O0
built stage1 doesn't fail, this is just a guess what the problem is, will try
to verify later on with bootstrap (but, because of that I think adding a
reduced testcase is not really useful either). If that is the case, it would
be a latent bug before, as determine_offset has been called before for
read/read and could put items into the affine cache.
