On Tue, 2013-10-22 at 10:07 -0400, Emilio Pozuelo Monfort wrote: > (Note that sometimes test failures are > just race conditions because the tests run on loaded/slow machines).
Over time I've been trying to fix these race conditions, because unlike Debian's package builds, gnome-continuous runs the glib tests continuously: easily 50+ times a day, 365 days a year, leveraging the InstalledTests model: https://live.gnome.org/GnomeGoals/InstalledTests This makes it a lot easier to both find races, and since it operates on the git repositories instead of immutable tarballs, that makes it easier to *fix* them: https://bugzilla.gnome.org/show_bug.cgi?id=705152 https://bugzilla.gnome.org/show_bug.cgi?id=704699 https://bugzilla.gnome.org/700460 https://bugzilla.gnome.org/show_bug.cgi?id=700855 https://git.gnome.org/browse/glib/commit/?h=ff8f37ac05490c67fc3be5e87ff5756f1a275da3 GLib's test suite seems pretty good to me at this point, although I don't really keep historical data about which specific tests failed, so I only tend to notice tests that are closer to 50/50% races instead of 1%/99%. _______________________________________________ gtk-devel-list mailing list gtk-devel-list@gnome.org https://mail.gnome.org/mailman/listinfo/gtk-devel-list
