The patch introduces race conditions.

This allows vp9dsp.h to only include the VP9 types header, and not the
decoder skeleton interface which is for hardware decoders (dxva2/vaapi).

The advantage here is that the internal software decoder interface is
not exposed to the DSP functions or the hardware accelerations.

If ret is NULL, a dummy common holder is created to hold *all* the
parallel function returns, which gets written concurrently. This commit
simplify the whole logic by simply not writing to that holder when not
set.

Signed-off-by: Marton Balint <cus@passwd.hu>

Signed-off-by: Marton Balint <cus@passwd.hu>

Needed for the C+11 atomics. Also change add_cxxflags to check_cxxflags.
Reviewed-by: James Almer <jamrial@gmail.com>
Signed-off-by: Marton Balint <cus@passwd.hu>

Reviewed-by: Matthieu Bouron <matthieu.bouron@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>

Also limits remuxing to audio, video and subtitle streams.

refer to SPEC:
Annex E. The FLV File Format said:
E.3 TheFLVFileBody have a table:
Field            Type    Comment
PreviousTagSize0 UI32    Always 0
Reviewed-by: Bela Bodecs <bodecsb@vivanet.hu>
Reviewed-by: Hendrik Leppkes <h.leppkes@gmail.com>
Signed-off-by: Steven Liu <lq@chinaffmpeg.org>

Fixes memleak.

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

adding demuxer and other logs should be easy
This forces single threaded decoding for simplicity
It also requires pthreads, this could be avoided either with
some lockless tricks or simply by assuming av_log would never be called from
another thread.

Fixes Ticket5521

Previous version reviewed by Stefano
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

They are unofficial extensions to the format for the time being, not
an experimental feature.

Signed-off-by: James Almer <jamrial@gmail.com>

* commit '064f19f3':
  avconv: support parsing bitstream filter options

This commit is a noop, see 5ef19590Merged-by: James Almer <jamrial@gmail.com>

* commit 'ecd2ec69':
  mov: Evaluate the movie display matrix

This commit is a noop, see 7010ebdfMerged-by: James Almer <jamrial@gmail.com>

* commit 'b90c8a3d':
  fate: Add tests for mov display matrix

Adapted to use ffprobe -show_entries
Merged-by: James Almer <jamrial@gmail.com>

* commit '7d308bf8':
  avprobe: Add -show_stream_entry to get a single stream property

This commit is a noop, we have a generic -show_entry option.
Merged-by: Clément Bœsch <u@pkh.me>

* commit '218ed725':
  openssl: Allow newer TLS versions than TLSv1

This commit is a noop, see e8634fb9Merged-by: Clément Bœsch <u@pkh.me>

* commit 'dad7514f':
  xcb: Add all the libraries to the link line explicitly

This commit is a noop. It appears we already link against the xcb shape
library since 54170a33.
Merged-by: Clément Bœsch <u@pkh.me>

* commit 'c541a44e':
  Revert "rtmpproto: Don't include a client version in the unencrypted C1 handshake"
Merged-by: Clément Bœsch <u@pkh.me>

* commit '801ac715':
  qsv: Be informative when reporting that no data has been consumed
Merged-by: Clément Bœsch <u@pkh.me>

* commit '30015305':
  Use avpriv_request_sample() where appropriate

Only the roqvideo chunk is merged because we actually support 24bpp
flic, see 5781c983.
Merged-by: Clément Bœsch <u@pkh.me>

* commit '07cac07c':
  dash: Use correct ISO C scanf conversion specifier

This commit is a noop: the use of SCN (scanf) format is wrong here.
Merged-by: Clément Bœsch <u@pkh.me>

* commit '3ec6f855':
  srt: Adjust signedness of sscanf format strings

This commit is a noop, a different fix is included in the big -Wformat
patch under review
(http://ffmpeg.org/pipermail/ffmpeg-devel/2017-March/209239.html)
Merged-by: Clément Bœsch <u@pkh.me>

* commit '7a2b2b6a':
  dxtory: Drop nonsense ISO C printf conversion specifiers for standard types
Merged-by: Clément Bœsch <u@pkh.me>

* commit 'c454dfcf':
  Use ISO C printf conversion specifiers where appropriate

This commit is a noop, an equivalent patch is currently under review on
the mailing-list: http://ffmpeg.org/pipermail/ffmpeg-devel/2017-March/209239.htmlMerged-by: Clément Bœsch <u@pkh.me>

Ends up identical to Libav.

This reduces diff with Libav.