GitLab

Make the audioresample test work with CK_FORK=no, and
turn a g_print into a GST_INFO.

The fix for green lines introduced by commit
35fdfcc6 caused invalid memory accesses
for even widths. This patch fixes it.

Fixes bug #567660.

When we're not parsing HTTP, return EPARSE when we get an HTTP
message.

Should really have been READABLE and WRITABLE, but those are hard to
add whilst maintaining backwards compatibility. See #343615.

API: GST_MIXER_TRACK_READONLY
API: GST_MIXER_TRACK_WRITEONLY

The macro is called GST_DISABLE_GST_DEBUG, not GST_DISABLE_DEBUG.

Add a property that makes videorate skip to the first buffer it
receives instead of padding the stream from segment start to the
first real buffer.

Fixes bug #567928.

Add string array with human readable names for format and type to be used in log
statements.

Don't try to print the RTP timestamp as a GstClockTime, it's just a guint32.

Fixes #594757

Fixes bug #571610.

Don't blindly return EPARSE when http mode is disabled.
Restore old http mode after temporarily setting it to TRUE.

Check for pulsesink < 0.10.17 because it includes code that is now included in
baseaudiosink. Disable that code in baseaudiosink to be compatible with the
older version.

A green border could be visible when converting to Y444 or RGB, because
the last chroma samples weren't copied correctly

- Buffer sizes were computed different from ffmpegcolorspace
- Green bar on right size for widths not divisable by 4

videotestsrc rounds chroma down. This causes it to omit the last chroma
value completely for odd widths when the chroma is downsampled.
This patch special cases the last pixel to not be rounded down.

Fixes bug #397419.

Take the time of the clock so that the last_time field is set. This is important
for sinks that restart their internal ringbuffer after a caps change and need to
know the last know position.

We want to fail early when an older libtheora release is used that does
not support Y444 or Y42B formats, so use a getcaps function that does
this.

Fixes bug #594165.

Some functions in libtheora can return an error, but that error cannot
ever happen inside theoraenc. In those cases assert that it doesn't.

Make it easy to experiment with different stop states (NULL and READY)

When going to NULL, we reset the ringbuffer so that it starts beck from 0. We
also make sure that the clock is updated with the elapsed time so that it
alsways increments even when the ringbuffer goes back to 0. When this happened
we need to adjust the sample position for the reset ringbuffer.

Fixes #594136

Add a property to disable rendering of video frames during preroll. This
will only work for videosinks that use the new ::show_frame() vfunc instead
of overriding basesink's preroll and render vfuncs directly.

API: GstVideoSink:show-preroll-frame

Add ::show_frame() vfunc which maps to basesink's ::preroll and ::render
vfuncs and add some gtk-doc chunks.

API: GstVideoSinkClass::show_frame()