Bug 356141 – gnome-base/libgnome-keyring-2.32.0 hangs on tests

• Status: RESOLVED FIXED
• Product: Gentoo Linux

[Description Markos Chandras (RETIRED) 2011-02-22 23:03:25 UTC]

make[1]: Leaving directory `/var/tmp/portage/gnome-base/libgnome-keyring-2.32.0/
Making check in egg
make[1]: Entering directory `/var/tmp/portage/gnome-base/libgnome-keyring-2.32.0
Making check in .
make[2]: Entering directory `/var/tmp/portage/gnome-base/libgnome-keyring-2.32.0
make[2]: Nothing to be done for `check-am'.
make[2]: Leaving directory `/var/tmp/portage/gnome-base/libgnome-keyring-2.32.0/
Making check in tests
make[2]: Entering directory `/var/tmp/portage/gnome-base/libgnome-keyring-2.32.0
make  check-am
make[3]: Entering directory `/var/tmp/portage/gnome-base/libgnome-keyring-2.32.0
TEST_DATA=./test-data gtester -m=slow ./run-auto-test
TEST: ./run-auto-test... (pid=15964)


Been in this place for 4 hours... Is this expected? If so please add a warning like other ebuilds to saying that tests will take too much time. Otherwise please restrict them.

The current stable passes tests successfully

[Comment 1 Pacho Ramos 2011-02-23 08:56:51 UTC]

Unable to reproduce, please attach full build.log. 

We try to run tests skipping dbus tests as told with upstream some time ago, for that, we run tests as follows:

src_test() {
        unset DBUS_SESSION_BUS_ADDRESS
        emake check || die "tests failed"
}

I think that, for some reason, dbus tests are not being skipped for you (without build.log is a bit harder to know that), maybe some other variable should be unset :-/

Anyway, this shouldn't block stabilization -> tests pass ok for me and I am using keyrings every day without issues

[Comment 2 Markos Chandras (RETIRED) 2011-02-23 09:44:40 UTC]

Created attachment 263613 [details]
build.log

[Comment 3 Pacho Ramos 2011-02-23 19:56:51 UTC]

Yes, dbus tests are being run and they shouldn't, could you add a "set > set.out" command in src_test phase and attach set.out to check if some more variables should be unset?

[Comment 4 Markos Chandras (RETIRED) 2011-03-02 18:20:18 UTC]

Created attachment 264393 [details]
set.out

[Comment 5 Pacho Ramos 2011-03-02 18:38:58 UTC]

Looks clean enough, maybe you could try to unset GSETTINGS_BACKEND and XDG_SESSION_COOKIE for tests :-/

[Comment 6 Markos Chandras (RETIRED) 2011-03-05 19:15:58 UTC]

Still hangs

[Comment 7 Pacho Ramos 2011-03-06 11:58:45 UTC]

This shouldn't occur per:
https://bugzilla.gnome.org/show_bug.cgi?id=626455#c3

And as can be seen reading tests/gtest-helpers.c:
[...]

        /* Need to have DBUS running */
        address = g_getenv ("DBUS_SESSION_BUS_ADDRESS");
        if (!address || !address[0]) {
                g_printerr ("\nNo DBUS session available, skipping tests!\n\n");
                return FALSE;
        }
[...]

does it also occur when you emerging the package as follows:
unset DBUS_SESSION_BUS_ADDRESS && emerge -1 libgnome-keyring
?

If also fails, does it fail in the same way (-> dbus tests not being skipped) when manually compiling and running tests after unsetting that variable?

[Comment 8 Pacho Ramos 2011-05-06 18:53:29 UTC]

(In reply to comment #7)
> This shouldn't occur per:
> https://bugzilla.gnome.org/show_bug.cgi?id=626455#c3
> 
> And as can be seen reading tests/gtest-helpers.c:
> [...]
> 
>         /* Need to have DBUS running */
>         address = g_getenv ("DBUS_SESSION_BUS_ADDRESS");
>         if (!address || !address[0]) {
>                 g_printerr ("\nNo DBUS session available, skipping
> tests!\n\n");
>                 return FALSE;
>         }
> [...]
> 
> does it also occur when you emerging the package as follows:
> unset DBUS_SESSION_BUS_ADDRESS && emerge -1 libgnome-keyring
> ?
> 
> If also fails, does it fail in the same way (-> dbus tests not being skipped)
> when manually compiling and running tests after unsetting that variable?

ping

[Comment 9 Pacho Ramos 2011-06-04 10:34:45 UTC]

Any news here?

[Comment 10 Markos Chandras (RETIRED) 2011-06-04 10:57:59 UTC]

It still hangs. unset DBUS_SESSION_BUS_ADDRESS does not help. Running them manually ( make check ) hangs as well

[Comment 11 Pacho Ramos 2011-06-04 14:39:01 UTC]

I would then really appreciate if you could report this problem to upstream directly since I don't know what is different in your system and doesn't really expect to get it solved without involving them.

Thanks

[Comment 12 Pacho Ramos 2011-06-19 12:01:49 UTC]

Did you reported this?

[Comment 13 Markos Chandras (RETIRED) 2011-06-20 11:51:02 UTC]

(In reply to comment #12)
> Did you reported this?

No sorry. I got busy

[Comment 14 Markos Chandras (RETIRED) 2011-06-22 07:06:05 UTC]

I have reported it upstream

[Comment 15 Pacho Ramos 2011-06-22 09:11:34 UTC]

Thanks

[Comment 16 Pacho Ramos 2011-08-27 19:55:20 UTC]

Can you please reply to upstream questions?

Thanks
https://bugzilla.gnome.org/show_bug.cgi?id=653150#c1

[Comment 17 Markos Chandras (RETIRED) 2011-08-29 19:31:36 UTC]

Sorry, this bug slipped under my radar. Please keep this bug open so I wont forget again. I will take care of the upstream bug as well

[Comment 18 Markos Chandras (RETIRED) 2011-08-30 11:23:10 UTC]

Please have a look on the upstream bug. A patch is attached there. However, it wont make it to gnome-2.32 upstream tree since the codebase has changed substantially. I tested the patch in my stable box and I can verify that tests in libgnome-keyring work as expected

[Comment 19 Pacho Ramos 2011-09-12 08:41:46 UTC]

+  12 Sep 2011; Pacho Ramos <[e-mail]> libgnome-keyring-2.32.0.ebuild,
+  +files/libgnome-keyring-2.32.0-drop-test.patch,
+  +files/libgnome-keyring-2.32.0-hang-tests.patch,
+  +files/libgnome-keyring-2.32.0-run-test.patch:
+  Apply multiple fixes for tests, also a one to prevent hangs by Markos
+  Chandras (bug #356141).
+