Richard Guenther writes:
> This fixes PR50082 by completely defering to fold what kind of
> overflow warning to emit from tree-ssa-forwprop.c. It also moves
> overflow handling to the single place that decides whether and
> what we optimize.
>
> Bootstrapped and tested on x86_64-unknown-linux-gn
This fixes PR50082 by completely defering to fold what kind of
overflow warning to emit from tree-ssa-forwprop.c. It also moves
overflow handling to the single place that decides whether and
what we optimize.
Bootstrapped and tested on x86_64-unknown-linux-gnu.
Ian, does this look ok?
Thanks,
This "fixes" PR50082 by passing comparison operands stripped off
useless conversions to the canonicalization routine. This makes
sure to fold the testcase early from a point where emitting
a overflow warning is properly constrained.
Bootstrapped and tested on x86_64-unknown-linux-gnu, applied to
