Hide Forgot
Description of problem: Sometimes when I run upstream tests, one of them (test-async-cancel) fails. Version-Release number of selected component (if applicable): gnome-vfs2-2.24.2-7.el6.x86_64 How reproducible: Cca 35 % Steps to Reproduce: 1. As root, install all dependencies (GConf2-devel libxml2-devel zlib-devel glib2-devel bzip2-devel ORBit2-devel openjade automake libtool intltool autoconf gtk-doc perl-XML-Parser libsmbclient-devel openssl-devel gamin-devel krb5-devel avahi-glib-devel hal-devel dbus-devel dbus-glib-devel libacl-devel libselinux-devel keyutils-libs-devel). 2. Get the gnome-vfs2 src rpm and put it to a regular user's home directory. 3. As regular user: mkdir ~/rpmbuild 4. As regular user: rpm2cpio ~/gnome-vfs2-*.el6.src.rpm | cpio -idmv 5. As regular user: cd ~/rpmbuild/BUILD/gnome-vfs-*/ 6. As regular user: make check Actual results: ... Testing directory loads..** Message: test failed: load directory cancel 2: job never went away delay 1 ... FAIL: test-async-cancel ... =================== 1 of 6 tests failed =================== make[2]: *** [check-TESTS] Error 1 make[2]: Leaving directory `/home/test/rpmbuild/BUILD/gnome-vfs-2.24.2/test' make[1]: *** [check-am] Error 2 make[1]: Leaving directory `/home/test/rpmbuild/BUILD/gnome-vfs-2.24.2/test' make: *** [check-recursive] Error 1 Expected results: ... PASS: test-async-cancel ... ================== All 6 tests passed ================== ... Additional info: Bug is also reported at https://bugzilla.gnome.org/show_bug.cgi?id=618089 It looks strange to me that there are 35 tests available in the test directory but only 6 of them are run by "make check".
I can reproduce similar failures. There are "synchronous" test cases, which call async gnome-vfs methods and waits for callbacks to set some flags. The wait is implemented following way: for (i = 0; i < MAX_THREAD_WAIT; i++) { g_thread_yield (); g_main_context_iteration (NULL, FALSE); if (*wait_for_it) { return TRUE; } } It seems the test cases fail, because the wait is not long enough. Errors can be eliminated if you increase MAX_THREAD_WAIT constant, however the constant needs to be increased a lot, e.g. 10000 to "be sure" that it doesn't fail. Maybe g_main_context_iteration, or g_thread_yield changed its behavioral a bit over the time. I don't know...
OK, you can disable the test.
Tested on x86_64 (VM). Tested version: gnome-vfs2-2.24.2-8.el6.src.rpm Applied scenario: - Get the gnome-vfs2 src rpm and put it to user's home directory. - mkdir -p ~/rpmbuild/SOURCES - cd ~/rpmbuild/SOURCES - rpm2cpio ~/gnome-vfs2-*.el6.src.rpm | cpio -idmv - rpmbuild -bc gnome-vfs2.spec - cd ~/rpmbuild/BUILD/gnome-vfs-*/ - make check Results: PASS: test-acl PASS: test-address PASS: test-escape PASS: test-uri PASS: ./auto-test ================== All 5 tests passed ==================
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://rhn.redhat.com/errata/RHBA-2016-0764.html