GitLab

Store the offset and caps when allocating a buffer during seeking, and then
allocate a new buffer with buffer_alloc before we push it out. This ensures
that in all respects the first buffer decoded during seeking behaves like
all other buffers, including allowing downstream re-negotiation.

Fixes #579808

JPEG markers may be preceded by additional 0xff.  jpegdec should
skip over these, even before the end marker.
See #579808

Also handle input with just one quant table, simply duplicate the quant table.
Handle invalid SOF correctly and some small cleanups.
Fixes #578257

Marco Ballesio authored Apr 29, 2009 and René Stadler committed May 06, 2009

Fixes #580746.

Read the timestamp of the incomming buffer before we push it in the adapter and
flush it out again as the buffer might be unreffed then and we read from invalid
memory.
Fixes #581444.

Free the dynamic pads data in finalize.
Fixes #581432

Don't require width/height on the caps. Use the SOF header to find width/height
and fall back to the caps if there is no SOF. Also use the SOF info to find the
subsampling and quantization tables used. This allows us to set the right type
value in the JPEG rtp header.

Deprecate the quality property, it's unused now and it was used wrongly before.

Always send full quant tables for now until we have some code to detect default
ones.

Fixes #580880

Use the width and the height from the payload headers and set them on the
output caps for added awesomeness.

Fix quant parsing, we need to check the type in the lower 6 bits.

Add first bits of caching quantization tables.

The libjpeg api says that we need to set the colorspace before we call
_set_defaults(). Indeed, if we don't do that we end up with some very freaky
non-standard quant table and huffman table indexes.

And don't check for gtk+ when it's not needed (ie. if examples are disabled)

Fix various version of find_stream_by_* by not trying to convert an int to a
pointer and vice versa, for portability reasons.

Fixes #581333

Fix a typo in the dummy NAT packet sending code.

Fixes #581329

Server eof (e.g. connection closed) is announced as connection closed,
so better record state and act accordingly to prevent (read/write)
errors during subsequent teardown/cleanup sequences.  #Fixes 580851.(c).

timestamps following flush/seek should be consistent between
UDP and TCP interleaved case.  Fixes #580851.(b).

A max range that overflows should not be trusted,
nor should a max range that equals the min range.
Fixes #580851.(a).

We can use the SKIP seek flag to instruct the server to send data faster then
normal but with the same bandwidth.
Fixes #537609

Use the new binary search method for finding the right index entry faster.

Round up the scaled U and V width and height so that we always draw the correct
amount of pixels to fill the complete image.

Fixes #569611

Sprinkle some G_LIKELY around.

Avoid traversing and dumping the tree when debugging is not activated.

Add support for subtitle pictures.
Fixes #568278.

Always signal the waiters in the async callbacks. Especially for the volume
callbacks since this might cause deadlocks.

We need to release the state lock before trying to wait for the task to end
because the task might also take the lock.

Fixes #577671

fix demuxing of m4v streams with ac-3 audio
Fixes #580554

getaddrinfo errors should be reported with gai_strerror instead of errno as
spotted by MikeS.

Shuffle the input bits according to RFC 3551 for G726 payloads.
Add option to force the previous behaviour.

Fixes #567140

Add a debugging category, add some comments and remove _peek_parent().

Bump taglib requirement.