glib-networking-2.80.0-4.eln144 failed ELN Mass Rebuild 144: # Start of oscp tests not ok /tls/gnutls/connection/oscp/must-staple - GLib-GIO:ERROR:../gio/gcancellable.c:674:cancellable_source_cancelled: assertion failed: (callback_was_not_called) Bail out! stderr: ** GLib-GIO:ERROR:../gio/gcancellable.c:674:cancellable_source_cancelled: assertion failed: (callback_was_not_called) Build log: https://kojipkgs.fedoraproject.org//work/tasks/5859/126675859/build.log Reproducible: Always
Wow, this is a bug in GCancellable. That's pretty impressive. I will start by reassigning this bug to GLib component. For ELN mass rebuild purposes, I would just trigger a rebuild. It's a race condition and *probably* won't hit the same race twice. I don't know how to trigger an ELN build myself.
(Sorry Paul. Wrong component....)
(In reply to Johnny Hughes from comment #0) > GLib-GIO:ERROR:../gio/gcancellable.c:674:cancellable_source_cancelled: > assertion failed: (callback_was_not_called) Hey Marco, do you think this was fixed by https://gitlab.gnome.org/GNOME/glib/-/merge_requests/2765? Hopefully? If so, I will backport that fix. Otherwise, I should report a new upstream bug.
Actually, looking at discussion in that merge request, there is a regression and we should not backport it. We should just rerun the glib-networking build in ELN and hope the test doesn't fail again. I'm going to close this since I don't think this requires any other action at this time.
(In reply to Michael Catanzaro from comment #4) > Actually, looking at discussion in that merge request, there is a regression > and we should not backport it. Hm, actually, the test failure here might itself be the regression. I see there's another merge request https://gitlab.gnome.org/GNOME/glib/-/merge_requests/4206 to fix it. We'll handle it upstream regardless.
Reopening. I had thought this was fixed by https://src.fedoraproject.org/rpms/glib2/c/5f6b1b57bbe1a656bf9e916a489dc27371066004?branch=rawhide but apparently not.
*** Bug 2340230 has been marked as a duplicate of this bug. ***
That commit didn't add the patch to the specfile, so it was never applied in the build.
Thank you Yaakov... I didn't notice.
OK, really fixed now.