[Bug 747852] pad: idle probe doesn't block pad from pushing data

[GStreamer (GNOME Bugzilla)]

https://bugzilla.gnome.org/show_bug.cgi?id=747852

--- Comment #8 from Thiago Sousa Santos <thiagossantos at gmail.com> ---
This is not exactly as trivial as it seems. When the idle probe is called after
data has passed the pad, no data is flowing from the pad as it is running the
idle probe code. As the idle probe doesn't do any blocking by itself, it
happens that all data pushed from the idle probe callback is allowed to pass.
The user has full control of what happens in the idle probe.

When the idle probe is called directly from the gst_pad_add_probe the streaming
thread can try to push data and we need to block that but without disabling the
idle probe from doing what it wants. For example, relinking that will cause
reconfigure events to be sent. Or any smarter decision involving caps queries.

One possible solution is to allow data from the idle probe thread to
unconditionally pass.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.