Re: glib2.0: FTBFS on s390x: gio/tests/gdbus-connection-slow.c:53 should not be reached
On Thu, 18 Sep 2025 at 19:41:40 +0100, Simon McVittie wrote:
This is asserting that when a subprocess sends a D-Bus signal, calls
g_dbus_connection_flush_sync(), and immediately exits, the parent
process receives the signal within 1 second. Presumably this arbitrary
timeout is too short for a mainframe designed for batch processing, or
something.
I think this is probably just a timeout that is too short and needs to
be longer. The test continues (cancelling the timeout) as soon as the
signal is received, so extending it to, say, 10 seconds would not
actually make the test any slower.
I don't know why 2.86.0-3 fails this when 2.86.0-2 succeeded.
I still don't know why this would have changed. This test hasn't changed
since 2024 and the code it's testing hasn't changed recently either.
On Fri, 19 Sep 2025 at 05:20:16 +0000, Pranav P wrote:
I will look into this and see if I can figure out something.
I was unable to reproduce the failure on zelenka, so if that was the
first thing you were going to try, you can skip that...
Do the s390x buildds do multiple builds in parallel, or do they run
multiple VMs with a large scheduling interval, or some other factor like
that that could result in arbitrary delays?
And has their setup changed recently?
Thanks,
smcv
Reply to: