#1026445 mutter: test failure on armhf and sometimes armel: ../../src/xcb_io.c:626: _XAllocID: Assertion `ret != inval_id' failed

Package:
src:mutter
Source:
src:mutter
Submitter:
Simon McVittie
Date:
2026-01-10 11:15:02 UTC
Severity:
normal
Tags:
#1026445#5
Date:
2022-12-20 11:47:10 UTC
From:
To:
Recent uploads of mutter have had a FTBFS on armhf and sometimes armel,
with this test failure in "mutter:core+mutter/wayland / xwayland":

This appears to be an assertion failure in libX11 rather than in mutter
itself, so it might be a bug in libX11 rather than in mutter, or perhaps
mutter is using it wrong.

This seems to happen consistently on armhf. On armel, retrying the build
of 43.2-1 was successful, but then 43.2-3 failed with the same symptom.

The same test failure has not been seen on arm64 or on non-ARM
architectures, for whatever reason (in particular, other slower
architectures like riscv64 and mips*el don't seem to have this
problem). Is there anything known to be different about armel and armhf
specifically?

    smcv
#1026445#10
Date:
2022-12-20 21:23:18 UTC
From:
To:
Control: severity -1 important

In 43.2-4 I've downgraded failures in this test to be non-fatal,
reducing the severity of this issue.

This is probably because d/rules in mutter explicitly skips the
tests on riscv64 and mips*el. I'd prefer to re-enable the tests on
all architectures (even if all failures are ignored on some of them)
now that it's using Meson, in which all tests have a finite timeout,
but that will probably need to happen via experimental in order to avoid
disrupting migration.

One important and possibly relevant difference between 32-bit ARM and
arm64 is that on 32-bit ARM, we explicitly set the default driver in
mutter's fork of cogl to be OpenGL|ES 2 instead of OpenGL 3, using
non-upstreamed patches. I'd like to be able to stop applying those
patches, but that will need input from users of proprietary GPU drivers
on ARM.

    smcv
#1026445#17
Date:
2023-03-18 11:04:28 UTC
From:
To:
Hi Simon,

I assume this is a PEBKAC issue, but I got this error too in a different
condition and as such it might help with this bug. If not, please ignore.

I cloned mutter's Salsa repo and added Salsa's default CI pipeline and it
consistently fails on the 'build i386' job in the same test case
("mutter:core+mutter/wayland / xwayland") with the same assertion failure.
See https://salsa.debian.org/diederik/mutter/-/jobs/4063218

I _think_ Salsa's CI runners run on amd64 and thus fails on an i386 job.
I also _think_ that armhf and armel are (often?) build on an arm64 native
host, but this could be dependent on which specific host it happens to get
build and thus cause the inconsistent error/non-error situation?

HTH and if not, sorry for the noise,
  Diederik
#1026445#22
Date:
2024-07-27 15:54:11 UTC
From:
To:
I'm finally doing that today.

Thank you,
Jeremy Bícha
#1026445#27
Date:
2026-01-10 10:08:33 UTC
From:
To:
Es gibt eine Familienspende in Höhe von 1.850.000,00 USD von Cheng Charlie
Saephan. Bitte antworten Sie für weitere Informationen. Denken Sie daran,
Ihrer Familie und den Bedürftigen in Ihrer Umgebung Gutes zu tun.

Dies ist bereits der zweite Versuch, Sie zu erreichen. Bitte antworten Sie
für weitere Details.
#1026445#32
Date:
2026-01-10 10:08:33 UTC
From:
To:
Es gibt eine Familienspende in Höhe von 1.850.000,00 USD von Cheng Charlie
Saephan. Bitte antworten Sie für weitere Informationen. Denken Sie daran,
Ihrer Familie und den Bedürftigen in Ihrer Umgebung Gutes zu tun.

Dies ist bereits der zweite Versuch, Sie zu erreichen. Bitte antworten Sie
für weitere Details.