[Pkg-utopia-maintainers] Bug#1091290: xdg-desktop-portal: FTBFS on armhf: handle_add_notification: 'g_variant_equal (notification, arg_notification)' should be TRUE

[Simon McVittie]

Control: retitle -1 xdg-desktop-portal: FTBFS on armhf: handle_add_notification: 'g_variant_equal (notification, arg_notification)' should be TRUE
Control: tags -1 + moreinfo

On Mon, 23 Dec 2024 at 18:44:04 +0100, Lucas Nussbaum wrote:
> During a rebuild of all packages in sid, this package failed to build
> on armhf.

Is this test failure reproducible? It seems to be building OK on
reproducible-builds.org.

If the problem is not reproducible or is only reproducible on relatively
slow machines, it might be as simple as an arbitrary timeout that is too
short for slower CPUs.

> > (/<<PKGBUILDDIR>>/obj-arm-linux-gnueabihf/tests/../src/xdg-desktop-portal:2182414): xdg-desktop-portal-WARNING **: 12:57:45.805: Failed to create RealtimeKit proxy: Could not connect: No such file or directory

I don't think this warning is actually anything to do with the test
failure. The reason for failure seems to be one or both of these assertion
failures "'res' should be FALSE" and "g_variant_equal (notification,
arg_notification)' should be TRUE" (they're probably related):

> > # Start of portal tests
> > # Start of notification tests
> > ok 1 /portal/notification/exists
> > ok 2 /portal/notification/basic
> > ok 3 /portal/notification/buttons
> > not ok /portal/notification/bad-arg - xdg-desktop-portal:ERROR:../tests/notification.c:137:notification_fail: 'res' should be FALSE
> > Bail out!
...
> > **
> > xdg-desktop-portal:ERROR:../tests/notification.c:137:notification_fail: 'res' should be FALSE
> > error: **
> > xdg-desktop-portal:ERROR:../tests/backend/notification.c:65:handle_add_notification: 'g_variant_equal (notification, arg_notification)' should be TRUE
> > 
> > Bail out! xdg-desktop-portal:ERROR:../tests/backend/notification.c:65:handle_add_notification: 'g_variant_equal (notification, arg_notification)' should be TRUE

Version 1.19.x in experimental has some improvements to this test which
might help it to pass, or might at least give better assertion failure
messages when it fails.

    smcv