It can happen that negotiation fails during get_buffer(), but then
we don't retry later and never return NOT_NEGOTIATED upstream...
and instead run into assertions.
https://bugzilla.gnome.org/show_bug.cgi?id=708769
As part of the changes to support the "avc3" variant of the ISO-BMFF
(see bug #702004) a new stream-format has been created
(video/x-h264, stream-format="avc3", alignment="au") that requires
changes to gstavcodecmap to exclude this format because avdec_h264
expects the SPS and PPS to be in the codec_data.
Amendment 2 of ISO/IEC 14496-15 (AVC file format) is defining a new
structure for fragmented MP4 called "avc3". The principal difference
between AVC1 and AVC3 is the location of the codec initialisation data
(e.g. SPS, PPS). In AVC1 this data is placed in the initial MOOV box
(moov.trak.mdia.minf.stbl.stsd.avc1) but in AVC3 this data goes in the
first sample of every fragment.
https://bugzilla.gnome.org/show_bug.cgi?id=702004
We need to reload the defaults for the codec after closing it,
otherwise we can't access codec information like the supported
sample rates and can crash.
https://bugzilla.gnome.org/show_bug.cgi?id=707040
It is enabled although we pass --disable-hwaccels apparently and
makes the build fail on Windows because libav looks for the d3d headers
in the wrong places. We can't make use of these features via libav.
https://bugzilla.gnome.org/show_bug.cgi?id=706601
libav can write slightly after the plane end in some SIMD optimized
functions. The extra padding value needs to be at least 16+stride_align
for each plane, so just increase the bottom padding value for the output
frame.
https://bugzilla.gnome.org/show_bug.cgi?id=694299
version 9.8:
- kmvc: Clip pixel position to valid range
- kmvc: Use fixed sized arrays in the context
- indeo: Reject negative array indexes
- indeo: Check for reference when inheriting motion vectors
- indeo: Properly forward the error codes
- mjpeg: Check the unescaped size for overflows
- wmapro: Error out on impossible scale factor offsets
- wmapro: Check the min_samples_per_subframe
- wmapro: Return early on unsupported condition
- wmapro: Check num_vec_coeffs against the actual available buffer
- wmapro: Make sure there is room to store the current packet
- lavc: Move put_bits_left in put_bits.h
- 4xm: Do not overread the source buffer in decode_p_block
- 4xm: Check bitstream_size boundary before using it
version 9.7:
Most of the following fixes resulted from test samples that the Google
Security Team has kindly made available to us:
- 4xm: fix several programming errors to avoid crashes, etc.
- apetag: use int64_t for filesize
- jpegls: Fix invalid writes to memory
- ljpeg: use the correct number of components in YUV
- mjpeg: Validate sampling factors
- mjpegdec: properly report unsupported disabled features
- mjpegdec: validate parameters in mjpeg_decode_scan_progressive_ac
- mpegvideo: allocate sufficiently large scratch buffer for interlaced vid
- pixdesc: mark gray8 as pseudopal
- smacker: fix several programming errors to avoid crashes, etc.
- tiff: do not overread the source buffer
- vmd: drop incomplete chunks and spurious samples
- vmdav: convert to bytestream2 to avoid invalid reads and writes
- wavpack: check packet size early
- wavpack: use bytestream2 in wavpack_decode_block
- wavpack: validate samples size parsed in wavpack_decode_block
If coded_width/_height is supplied, the codec might use it as the
width/height and if it is wrong can lead to segfaults or video
corruption.
This is specially harmful on renegotiation scenarios where the
resolution changed. There seems to be no specific function for reseting
the AV Context in libav, so just set it directly.
https://bugzilla.gnome.org/show_bug.cgi?id=702003
