https://gcc.gnu.org/bugzilla/show_bug.cgi?id=122627

--- Comment #2 from Jeffrey A. Law <law at gcc dot gnu.org> ---
Yea, definitely another case of an out of bounds access.  This time it was a
write corrupting the heap.

I'm testing a bigger hammer for this class of problems.  After register
elimination (which is where we trigger use of addptr) I put in a call to
potentially resize the equivalency arrays and removed the spot fix I did last
week for pr122321.

With that change pr122321 as well as this one both compile successfully.  I
didn't have a testcase for the other spot change we did back in Sept and the
code paths to the failure point aren't as clear.  So I'm leaving the Sept spot
check alone.

Anyway, testing a fix.

Reply via email to