[Webkit-unassigned] [Bug 115352] [gstreamer] Make sure gstreamer source element is thread-safe

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=115352


Andre Moreira Magalhaes <andrunko at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
 Attachment #202292|0                           |1
        is obsolete|                            |




--- Comment #20 from Andre Moreira Magalhaes <andrunko at gmail.com>  2013-08-13 10:56:03 PST ---
Created an attachment (id=208654)
 --> (https://bugs.webkit.org/attachment.cgi?id=208654&action=review)
Updated patch against upstream master

So, first of all, this is the updated patch against upstream master with conflicts fixed, etc.

I investigated the regressions on #117141 and #117688 and here goes my findings:

1 - Bug #117688 (http/tests/security/video-poster-cross-origin-crash.html crash on GTK bots):

Running this test with both upstream/master and upstream/master with the updated patch applied I get the exactly same results:
$ TEST_RUNNER_TEST_PLUGIN_PATH=./WebKitBuild/Debug/TestNetscapePlugin/.libs/libtestnetscapeplugin.so TEST_RUNNER_INJECTED_BUNDLE_FILENAME=WebKitBuild/Debug/Libraries/.libs/libTestRunnerInjectedBundle.so Tools/jhbuild/jhbuild-wrapper --gtk run ./WebKitBuild/Debug/Programs/WebKitTestRunner LayoutTests/http/tests/security/video-poster-cross-origin-crash.html 
Gtk-Message: Failed to load module "overlay-scrollbar"
Gtk-Message: Failed to load module "canberra-gtk-module"
Gtk-Message: Failed to load module "overlay-scrollbar"
Gtk-Message: Failed to load module "canberra-gtk-module"
Content-Type: text/plain
>>>
#EOF
#EOF
#EOF

** (WebKitTestRunner:24868): CRITICAL **: socket_embed_hook: assertion `spi_global_register != NULL' failed
LEAK: 1 WebContext
LEAK: 8 WebCoreNode

2 - Bug #117141 (fast/js/cyclic-proto.html crash on Qt bots):
This bug seems to be crashing on LocalStorage and I can't see how this patch is related.

That said I believe this bug is not related to the crashes, but if someone can enlighten me here or get me some better backtrace with the crashes I may be able to fix any possible regression. It is good to note that this patch fixes a bunch of thread related issues with the webkit gstreamer source element.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.