Re: Bug#1121780: Please revert 3.13.9 version bump to help m68k

On Wed, Dec 3, 2025 at 5:08=E2=80=AFAM John Paul Adrian Glaubitz
<[email protected]> wrote:
>
> Hi Adrian,
>
> On Tue, 2025-12-02 at 15:31 +0100, John Paul Adrian Glaubitz wrote:
> > On Tue, 2025-12-02 at 16:25 +0200, Adrian Bunk wrote:
> > > Source: python3-defaults
> > > Version: 3.13.9-1
> > > Severity: normal
> > > X-Debbugs-Cc: [email protected]
> > > User: [email protected]
> > > Usertags: m68k
> > >
> > > I do not think the 3.13.9 version bump is necessary, and going
> > > back to 3.13.7 would allow m68k to participate in the 3.14 transition=
.
> > >
> > > (Based on the timing I suspect the python3.13/14 FTBFS on m68k
> > >  started with GCC 15, but this is just a guess.)
> >
> > Yes, it did. There is a bug in Python upstream for which I provided a p=
atch
> > but while the compilation error is fixed, the Python interpreter itself=
 eventually
> > locks up for whatever I have not been able to figure out.
> >
> > I think it would be best to build Python 3.14 with gcc-14 for the time =
being
> > until the underlying problem has been sorted out. But I have just so mu=
ch capacity
> > I can spend on all of this before burning out.
>
> FWIW, I have built and uploaded Python 3.14 manually now in a chroot wher=
e gcc-defaults
> was downgraded to gcc-14 and that built without any problems. So it's def=
initely related
> to gcc-15.

>From the GCC-15 release notes (<https://gcc.gnu.org/gcc-15/changes.html>):

    C23 by default: GCC 15 changes the default language version for C
    compilation from -std=3Dgnu17 to -std=3Dgnu23. If your code relies on o=
lder
    versions of the C standard, you will need to either add -std=3D to your
    build flags, or port your code; see the porting notes.

The first thing I would try is to see if it is related to C23.

Jeff