GitLab

This is based on collectpads2 and is synchronizing
all streams based on the running time.

New features compared to old videomixer:
 * Synchronizing frames on the running time
 * Improved and simplified negotiation
 * Full QoS support
 * Variable framerate support

Fixes bug #626048, #624905.

Before we required a complete date in matroskademux but in
id3demux for example only the year or year and month was possible too.

Fixes bug #628454.

MPEG doesn't have a static size per frame, so don't pretend it has one
and fail when capturing because it doesn't match. Instead mark the size
as unknown and let the read frame grabbing method use a reasonable fallback
value (assuming that's only for actual streaming formats)

Fixes bug #628349.

And don't post an error message if downstream returns UNEXPECTED.

And don't post an error message if buffer allocation failed because
of UNEXPECTED, which only means that downstream wants us to EOS now.

And properly handle UNEXPECTED and WRONG_STATE.

And as a result, don't ignore WRONG_STATE and NOT_LINKED.
Both mean that it's a good idea to pass them upstream instead
of pretending that everything is good.

And document why wrong-state doesn't need an error message.

Fixes bug #628020.

This now means, that we have absolutely no handwritten assembly anymore
in videomixer and it's also faster now when using SSE.

videomixer: Add orc implementation for blending

videomixer: Fix example pipelines

imagefreeze: Add test for checking if imagefreeze correctly returns UNEXPECTED after the first buffer

If setting the pipeline to PLAYING before issuing the seek, buffers
are already arriving at the sink before the seek is handled and
will have the wrong timestamps and everything.

Fixes bug #625547.

Due to a seek the srcpad task could be started in rare circumstances although
it shouldn't be started anymore because no upstream buffer is available.

This fixes a subtle race condition, that caused bufferalloc to fail
with wrong-state due to a seek but caused it to be not retried as
it should.

Before there could be rounding errors when calculating the duration,
resulting in timestamp + duration being smaller than the next buffer's
timestamp.

Allows the application to modify the client name used to connect when
connecting to the PulseAudio daemon. Note however that updating the
property after the element reached the READY state will have no
effect until the next NULL->READY transition.

Fixes bug #627174.

Before they contained the URL before the actual failure. The other
way around makes more sense and we do the same in other elements
like filesrc.

Fixes bug #627289.

Fixes bug #627162

Avoid to create a new PA context for each new client by using a hash
table containing the list of ring-buffers and the shared PA context
for each client. Doing this will improve application memory usage in
the cases where multiple pipelines involving multiple pulsesink
elements are used.

Fixes bug #624338.

If the application requests a state-change and pulsesink fails to open
the ring_buffer device the mainloop attribute of the sink should be
cleaned up to avoid future state-change (NULL->READY) failures.

Fixes bug #627341.