https://gcc.gnu.org/bugzilla/show_bug.cgi?id=24542
rsandifo at gcc dot gnu.org <rsandifo at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |rsandifo at gcc dot gnu.org
--- Comment #14 from rsandifo at gcc dot gnu.org <rsandifo at gcc dot gnu.org>
---
Heh, was in the process of reopening this too, but Richard beat
me too it.
FWIW, I agree this is worth providing as an option. Another justification
is the different promotion handling between u32 = u16 op u16 and
u64 = u32 op u32.
"auto" (which wasn't a thing when the PR was first filed) might also
increase the chances of accidentally pushing promotions to the root of
a multi-statement calculation.
I don't think the false positive/negative ratio matters too much for
the option itself. If it works then I think it's worth having.
IMO the ratio only becomes important if we're considering enabling
this by default (unlikely), -Wall (unsure) or -Wextra (seems feasible).
