[Bug 725809] ghostpad: rare crash because of missing reference count on its target pad

[GStreamer (bugzilla.gnome.org)]

https://bugzilla.gnome.org/show_bug.cgi?id=725809
  GStreamer | gstreamer (core) | git

George Kiagiadakis <kiagiadakis.george> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|1.3.1                       |HEAD

--- Comment #5 from George Kiagiadakis <kiagiadakis.george at gmail.com> 2014-03-07 10:39:14 UTC ---
I have tried, but a unit test is very difficult to write in a way that it
reproduces the problem at least most of the times. To reproduce it, the other
thread needs to fully destroy the target pad while the thread running
gst_ghost_pad_set_target() is between the acquisition of the target pad pointer
and its unlinking from the internal pad. This probably means that the OS
scheduler needs to unschedule the thread running gst_ghost_pad_set_target() for
a big amount of time in between these two events.

The application that this bug was originally found in would usually crash once
every 30 to 40 hours (sometimes even more!) of running a continuous loop that
creates and destroys a full pipeline with rtspsrc every 5 seconds.

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