On Wed, 12 Mar 2025 at 19:55, Jonathan Wakely <jwakely.gcc@xxxxxxxxx> wrote:
>
> On Wed, 12 Mar 2025 at 19:42, Jonathan Wakely <jwakely.gcc@xxxxxxxxx> wrote:
> >
> > On Wed, 12 Mar 2025 at 18:38, Dennis Clarke <dclarke@xxxxxxxxxxxxx> wrote:
> > > OKay, so the timezone Europe/London is absent somewhere?
> > >
> > > t$
> > > t$ ls -l /usr/share/zoneinfo/Europe/London
> > > -rw-r--r-- 1 root root 3664 Nov 23 14:04 /usr/share/zoneinfo/Europe/London
> > > t$
> > >
> > >
> > > Seems to exist on the machine.
> > >
> > >
> > > So that is baffling.
> >
> > As I said, only the /usr/share/zoneinfo/tzdata.zi file matters, not
> > the Europe/London file.
> >
> > On my machine it contains that zone:
> >
> > $ grep Europe/London /usr/share/zoneinfo/tzdata.zi
> > Z Europe/London -0:1:15 - LMT 1847 D
> > L Europe/London GB
> > L Europe/London GB-Eire
> > L Europe/London Europe/Guernsey
> > L Europe/London Europe/Isle_of_Man
> > L Europe/London Europe/Jersey
> > L Europe/London Europe/Belfast
>
> My guess is something in the tzdata.zi file which cannot be parsed,
> causing an error when loading it, so that a UTC-only fallback is used.
>
> If that's true, the program below should print "ersatz"
>
> > Could you please compile and run this program with this installed gcc-14:
> >
> > #include <chrono>
> > #include <iostream>
> >
> > int main()
> > {
> > std::cout << std::chrono::get_tzdb().version << '\n';
> > }
>
> I suspect the problem is https://gcc.gnu.org/PR116657
But maybe the Debian-specific https://gcc.gnu.org/PR109857
