On Thu, 4 Apr 2024 at 17:33, Jonathan Wakely <jwak...@redhat.com> wrote: > > On Thu, 4 Apr 2024 at 17:29, Ulrich Drepper <drepper....@gmail.com> wrote: > > > > On Thu, Apr 4, 2024 at 5:29 PM Jonathan Wakely <jwak...@redhat.com> wrote: > > > I would appreciate more eyes on this to confirm my conclusions about > > > negative int_type values, and the proposed fix, make sense. > > > > The way something like this is handled in glibc's ctype functions is > > that both branches are considered. For isXXX(c) whether c is -v or > > 256-v the same value is returned (except for EOF which is -1). This > > caused the least number of bad surprises. > > > > You could here also perform similar actions. > > Yes, my first attempt to fix PR93672 did exactly that, see > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93672#c1 > > But since it doesn't work for '\xff' (because that's EOF when char is > signed) it only handles 127 of the 128 possible bugs ;-) > I'm also not sure it's conforming, since the standard specifies how > the matching is done, and that won't match negative chars.
I might suggest relaxing the standard to allow it though...
