[PATCH v6 04/11] stdio-common: Add __translated_number_width

[Dmitry V. Levin]

On Mon, Jan 30, 2023 at 12:03:24PM +0100, Florian Weimer wrote:
> * Dmitry V. Levin:
> 
> > On Mon, Jan 30, 2023 at 11:42:21AM +0100, Florian Weimer via Libc-alpha wrote:
> >> * Dmitry V. Levin:
> >> 
> >> > Hi,
> >> >
> >> > On Fri, Dec 16, 2022 at 09:15:22PM +0100, Florian Weimer via Libc-alpha wrote:
> >> >> This function will be used to compute the width of a number
> >> >> after i18n digit translation.
> >> >
> >> > I haven't bisected, but I suppose it was this changeset that introduced
> >> > a regression reported by the strace test suite.
> >> >
> >> > glibc-2.36$ cat <<'EOF' |gcc -xc - && ./a.out
> >> > #include <stdio.h>
> >> > int main() { printf("%03d\n", 1); return 0; }
> >> > EOF
> >> > 001
> >> >
> >> > The same test on master prints 0001 instead of 001.
> >> 
> >> Sorry, I can't reproduce it.  Did you build glibc yourself from upstream
> >> sources, or did you get the build from somewhere else?
> >
> > This was initially reported by the strace test suite running on
> > rawhide-test.fedorainfracloud.org, and I suppose the patch submitted
> > by Andreas today fixes it.
> 
> Andreas' fix is for %#03o, though.  Perhaps the reduction above is
> incorrect?

Oops, I posted the wrong reduction, it was originally %#03o indeed.


-- 
ldv