glib build fail for armhf

  • Open
  • quality assurance status badge
Details
2 participants
  • Ludovic Courtès
  • Mathieu Othacehe
Owner
unassigned
Submitted by
Mathieu Othacehe
Severity
normal
M
M
Mathieu Othacehe wrote on 26 Nov 2020 10:31
(address . bug-guix@gnu.org)
87blfkihvi.fsf@gnu.org
Hello,

The glib build is broken on armhf due to a timeout in the test suite,
see:

It looks like it's also broken on the 1.2.0 branch.

This prevents the evaluation of the guix-modular specification, see:
system for that specification until this is understood.

Thanks,

Mathieu
L
L
Ludovic Courtès wrote on 27 Nov 2020 12:08
(name . Mathieu Othacehe)(address . othacehe@gnu.org)(address . 44881@debbugs.gnu.org)
87lfent5u8.fsf@gnu.org
Hi Mathieu,

Mathieu Othacehe <othacehe@gnu.org> skribis:

Toggle quote (6 lines)
> The glib build is broken on armhf due to a timeout in the test suite,
> see:
> https://ci.guix.gnu.org/log/m6rd864j30khc6k60gr2wqr1pz40f1di-glib-2.62.6-bin.
>
> It looks like it's also broken on the 1.2.0 branch.

Yeah. We were discussing it with Chris Baines on IRC. My hypothesis is
that those tests time out on low-end ARMv7 devices (like
guix-x15.sjd.se).

Previously, we were building them on the more powerful OverDrives
probably, and indeed, the tests pass there. Berlin now has a substiute
for this one built on an OverDrive.

Toggle quote (4 lines)
> This prevents the evaluation of the guix-modular specification, see:
> https://ci.guix.gnu.org/eval/19035/log/raw. I disabled the armhf-linux
> system for that specification until this is understood.

I think we can re-enable it now?

Thanks,
Ludo’.
M
M
Mathieu Othacehe wrote on 1 Dec 2020 10:49
(name . Ludovic Courtès)(address . ludo@gnu.org)(address . 44881@debbugs.gnu.org)
87o8jd3lfh.fsf@gnu.org
Hey Ludo,

Toggle quote (6 lines)
>> This prevents the evaluation of the guix-modular specification, see:
>> https://ci.guix.gnu.org/eval/19035/log/raw. I disabled the armhf-linux
>> system for that specification until this is understood.
>
> I think we can re-enable it now?

Sure, I enabled it again. Maybe we should consider disabling those tests
on core-updates then.

Thanks,

Mathieu
?