lavfi/selectivecolor: simplify crazy mid val computations

lavfi/transpose: add missing const options flags

Fixes ffmpeg -h filter=transpose

MAINTAINERS: update

wmavoice: protect against zero-energy in adaptive gain control.

Otherwise the scale factor becomes NaN, resulting in corrupt output.
Fixes #5426.

wmavoice: move overflow handling to common code.

wmavoice: reindent.

wmavoice: disable bitstream checking.

The checked bitstream reader does that already. To allow parsing of
superframes split over a packet boundary, we always decode the last
superframe in each packet at the start of the next packet, even if
theoretically we could have decoded it. The last superframe in the
last packet is decoded using AV_CODEC_CAP_DELAY.

wmavoice: move wmavoice_flush() up.

checkasm/vp9: benchmark all sub-IDCTs (but not WHT or ADST).

avformat/hlsenc: strftime identifiers and segment index
 in filenames

Putting date/time values into segment filenames is very usefull.
But to produce non-conflicting segment filenames with -use_localtime
option with date/time
values in hls_segment_filename option, sometimes is not enough.
Like in cases when multiple segments produced in the same second.
But hlsenc currently does not make possible to use segment index (%d) at
the

same time whe use_localtime is in effect, due to identifier conflict.
This patch makes possible to use strftime identifiers and still put
segment index (%d) at same time in segment filenames by introducing
second_level_segment_index flag. When -use_localtime is active,
identifier %d is for month day index, so %%d is the segment index
placeholder. This enhanced behaviour only exists when new
second_level_segment_index flag is specified.
For instance putting 'segment_%Y%m%d%H%M%S_%%05d.ts' value into
-hls_segment_filename option and specifing -hls_flags
second_level_segment_index and -use_localtime 1, may produce segment
filename as 'segment_20161230235758_00002.ts'

An example:

ffmpeg -loglevel info -y -f lavfi -i color=c=red:size=640x480:r=25 -f
lavfi -i anullsrc=r=44100:cl=stereo -c:v mpeg2video -g 25 -acodec aac
-cutoff 20000 -ac 2 -ar 44100 -ab 192k -f hls -hls_time 3 -hls_list_size
5 -hls_flags delete_segments+second_level_segment_index -use_localtime 1
-hls_segment_filename "segment_%Y%m%d%H%M%S_%%05d.ts" stream.m3u8

will produce segments filenames:

....
segment_20161227005902_00013.ts
segment_20161227005902_00014.ts
segment_20161227005902_00015.ts
segment_20161227005903_00016.ts
segment_20161227005903_00017.ts
segment_20161227005903_00018.ts
segment_20161227005903_00019.ts
segment_20161227005903_00020.ts
....

Signed-off-by: Bela Bodecs <bodecsb@vivanet.hu>

Fix bug when incrementing initial_prog_date_time when removing segments

initial_prog_date_time shouldn't be adjusted when deleting segments
from disk, but rather when segments are removed from the playlist.

Signed-off-by: Jesper Ek <deadbeef84@gmail.com>
Signed-off-by: Steven Liu <lq@chinaffmpeg.org>

avformat/matroskadec: Fix OOM on long streams

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

avdevice/decklink_enc: do not reference this after freeing it

Fixes Coverity CID 1396863.

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

avdevice/decklink_dec: properly initialize no_video variable

Fixes Coverity CID 1396859.

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

ffmpeg: Check avcodec_parameters_to_context() for failure

Fixes CID1396241

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

avcodec/ylc: thread safe initialization is possible with this codec

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avcodec/ylc: add frame threading support

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avcodec/ylc: do not leak memory at uninit

Signed-off-by: Paul B Mahol <onemda@gmail.com>

doc/examples/http_multiclient: Fix resource leak

Fixes CID1396269

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

flv demuxer supports live rtmp inputs but there is no any info about it in the docs.

Signed-off-by: Bela Bodecs <bodecsb@vivanet.hu>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/magicyuv: export colorspace and color_range for YUV

Signed-off-by: Paul B Mahol <onemda@gmail.com>

NVENC: Update check for Lookahead

Reviewed-by: Timo Rothenpieler <timo@rothenpieler.org>

avcodec/lossless_videodsp: fix output of add_hfyu_left_pred_int16_c()

It is now bitexact with the ssse3 and sse4.1 versions of the function.

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: James Almer <jamrial@gmail.com>

avformat/hlsenc: detecting duplicated segment filenames
  ffmpeg-devel

with use_localtime parameter hlsenc may produce identical filenames for
different but still existing segments. It happens when
hls_segment_filename contains
syntacticaly correct but inadequate format parameters. Currently there
is no any log message when such a situaton occurs but these cases should
be avoided in most times. This patch generate warning log messages in
these cases.

ticketID: #6043

Signed-off-by: Bela Bodecs <bodecsb@vivanet.hu>
Signed-off-by: Steven Liu <lingjiujianke@gmail.com>

avcodec/error_resilience: Move variable initialization down, remove unneeded inits

This makes the code faster and easier to read

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

ffplay: add startup volume option

Fixes Ticket 5389.

Reviewed-by: Marton Balint <cus@passwd.hu>
Signed-off-by: Ganesh Ajjanagadde <gajjanag@alum.mit.edu>

avfilter/af_amerge: properly handle unknown input layouts

Reviewed-by: Nicolas George <george@nsup.org>
Signed-off-by: Marton Balint <cus@passwd.hu>

avcodec/error_resilience: Merge surrounding status checks

Simplifies code and is also faster

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

avcodec/error_resilience: Factor block parity check out

This makes the code noticably faster when there are lots of blocks

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

avfilter/vf_w3fdif: add >8 but <16 bit support

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avcodec/pixlet: simplify lowpass_prediction() function

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avutil/random_seed: Reduce the time needed on systems with very low precission clock()

This should fix issues on BSD
CLOCKS_PER_SEC is 128 on BSD while SUSv2 requires it to be a million

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

tools/zmqsend: Do not truncate fgetc() return

Fixes CID1396836

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

lavfi/buffersrc: push the frame deeper if requested.

Reduce peak memory consumption with ffmpeg in certain cases.

lavf/framesync: detect EOF immediately.

Fix an infinite loop in forward_status_change().

avcodec/utvideo: fix mistake using wrong arguments for left and lefttop pixel components

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avcodec/pixlet: make sure scaling factors are not zero

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avcodec/pixlet: avoid some overflows

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avcodec/pixlet: fix undefined behaviour in postprocess_chroma

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avfilter/framequeue: add missing check for ASSERT_LEVEL

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

tests/audio_fifo: fix buffer allocation for non planar formats

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

swscale/swscale: Fix dereference of stride array before null check

Fixes: CID1396263
Fixes: CID1396271
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avformat/avidec: skip odml master index chunks in avi_sync

Fixes pts gaps when reading AVI files > 256GiB generated by FFmpeg.

Signed-off-by: Tobias Rapp <t.rapp@noa-archive.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avutil/tests/random_seed: eliminate goto

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

avfilter/vf_deband: do not use uninitialized value

Fixes coverity report.

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avcodec/utvideodec: add SIMD support for median prediction for planar formats

~10% faster overall.

Signed-off-by: Paul B Mahol <onemda@gmail.com>

lavfi: avfilter_graph_request_oldest: request a frame again before returning.

With min_samples, if a frame arrives but is too small, it clears
frame_wanted_out. In most cases, the destination filter would be
activated again later because of frame_wanted_out on its own
outputs, but not sinks.
avfilter_graph_request_oldest() is doing the work of the sink
itself, and is therefore allowed to use frame_blocked_in.

avcodec/asvdec: return correct value in case of error

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avcodec/pixlet: check out of bounds pfx value

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avcodec/pixlet: remove unnecessary double to float conversion

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

avfilter/af_amerge: detect EOF immediately

Fix an infinite loop in forward_status_change().

Signed-off-by: Nicolas George <george@nsup.org>
Signed-off-by: Marton Balint <cus@passwd.hu>

avutil/tests: Improved code coverage for random_seed

Signed-off-by: Thomas Turner <thomastdt@googlemail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec: add Apple Pixlet decoder

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avfilter/vf_deband: add planes coupling mode

Signed-off-by: Paul B Mahol <onemda@gmail.com>

fate: Add test for ticket 6024, truncated decoding mode

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

ffserver_config: Check for failure to allocate FFServerIPAddressACL

Fixes CID1396537

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

tests/avstring: free the pointer after calls to av_d2str()

Fixes memleaks.

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

doc/general: mention recently added PCM codecs

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avutil/random_seed: Improve get_generic_seed() with higher precission clock()

Tested-by: Thomas Turner <thomastdt@googlemail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

doc/filters: Fix vsbmc option name.

Reported-by: Антон Приходько

lavfi: take_samples: free frames after taking all samples.

avformat/wavdec: add support for decoding 24.0 and 16.8 floating point pcm formats

Fixes #5602 and #5603.

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avcodec: add pcm_f16le and pcm_f24le decoder

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avcodec/wmaprodec: cleanup extradata dumping

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avformat/mpegenc: Check for av_mallocz() failure

Fixes CID1257787

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

avformat/img2dec: Remove dead code from psd_probe()

Fixes CID1397124

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

avfilter/vf_psnr: add gray10 and gray12 support

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avutil/random_seed: Use uint64 instead of uint8 for struct to avoid potential alignment issues

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

avutil: Added selftest for libavutil/audio_fifo.c

Signed-off-by: Thomas Turner <thomastdt@googlemail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

swscale: save ebx register when it is not available

Configure checks if the ebx register can be used for asm and it has to
be saved if and only if this is not the case.
Without this the build fails when configuring with --toolchain=hardened
--disable-pic on i386 using gcc 4.8:
error: PIC register clobbered by '%ebx' in 'asm'

In that case gcc 4.8 reserves the ebx register for the GOT needed for
PIE, so it can't be used in asm directly.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

avutil: Improved test coverage for avstring.c

Signed-off-by: Thomas Turner <thomastdt@googlemail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/magicyuv: add 10 bit support

Signed-off-by: Paul B Mahol <onemda@gmail.com>

lavfi/framequeue: avoid empty structs.

Fix compilation on MSVC.

avformat/mov.: Make the process of uuid-xmp atom faster.

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

lavfi/atempo: avoid false triggering an assertion failure

Steps to reproduce:
./ffmpeg_g -f s16be -i /dev/null -af atempo=0.5 -y /tmp/atempo.wav

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

doc/examples/decoder_targeted: Limit max pixels for fuzzing

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

lavfi: make filter_frame non-recursive.

A lot of changes happen at the same time:

- Add a framequeue fifo to AVFilterLink.

- split AVFilterLink.status into status_in and status_out: requires
  changes to the few filters and programs that use it directly
  (f_interleave, split, filtfmts).

- Add a field ready to AVFilterContext, marking when the filter is ready
  and its activation priority.

- Add flags to mark blocked links.

- Change ff_filter_frame() to enqueue the frame.

- Change all filtering functions to update the ready field and the
  blocked flags.

- Update ff_filter_graph_run_once() to use the ready field.

- buffersrc: always push the frame immediately.

lavfi: add FFFrameQueue API.

avcodec/get_bits: add av_assert2 to get_bits_long()

Signed-off-by: Paul B Mahol <onemda@gmail.com>

opus: remove redundant ff_celt_window2 declaration

Signed-off-by: Rostislav Pehlivanov <atomnuker@gmail.com>

avformat: Auto-detect mjpeg 2000 in mpeg-ts

This makes it possible to decode motion jpeg 2000
encoded in a transport stream without a correct PMT/PAT.

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

avformat/tls: add tls url_get_file_handle

Support url_get_file_handle on TLS streams.

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

lavc: Fix ticket 6024, truncated mode decoding

The assumption that avcodec_send_packet makes regarding decoders
consuming the entire packet is not true if the codec supports
truncated decoding mode and the truncated flag is turned on.

Steps to reproduce:
./ffmpeg_g -flags truncated \
-i "http://samples.ffmpeg.org/MPEG2/test-ebu-422.40000.pakets.ts" \
-c:v ffv1 -c:a copy -y /tmp/truncated.nut

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

avcodec/mips: version 1 of vc1dsp optimizations for loongson mmi

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

avfilter/vf_ssim: add >8 bit depth suppport

Signed-off-by: Paul B Mahol <onemda@gmail.com>

avformat/wavdec: Eliminate goto for clang -O0 DCE

Clang is not able to eliminate the reference to ff_spdif_probe() when
there is a goto target in the same block and optimization is disabled.

This fixes the following build failure on OS X:
  ./configure --disable-everything --disable-doc \
    --enable-decoder=pcm_s16le --enable-demuxer=wav \
    --enable-protocol=file --disable-optimizations --cc=clang
  make
  ...
  Undefined symbols for architecture x86_64:
    "_ff_spdif_probe", referenced from:
        _set_spdif in libavformat.a(wavdec.o)
  ld: symbol(s) not found for architecture x86_64

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

avutil/imgutils: Clarify doxy for av_image_check_size2()

Reviewed-by: Andreas Cadhalpun <andreas.cadhalpun@googlemail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

fate/psd: add tests for uncompress and rle samples

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

avcodec/h263dec: Return the correct error code in explode mode

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/smvjpegdec: return the packet size instead of 0

Most decoders return the amount of data used.
This is more consistent

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/tdsc: return the packet size instead of 0

Most decoders return the amount of data used.
This is more consistent

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/screenpresso: return the packet size instead of 0

Most decoders return the amount of data used.
This is more consistent

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avcodec/rscc: return the packet size instead of 0

Most decoders return the amount of data used.
This is more consistent

Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Michael Niedermayer <michael@niedermayer.cc>

avfilter/vsrc_testsrc: draw_bar: make sure width is not negative

Reported-by: Josh de Kock
Signed-off-by: Paul B Mahol <onemda@gmail.com>

avformat/mp3dec: fix msan warning when verifying mpa header

MPEG Audio frame header must be 4 bytes. If we fail to read
4 bytes bail early to avoid Use-of-uninitialized-value msan error.
Reference https://crbug.com/666874.

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

tiff: fix overflows when calling av_reduce

The arguments of av_reduce are signed, so the cast to uint64_t is misleading.

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

mov: prevent overflow during bit rate calculation

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

cafdec: prevent overflow during bit rate calculation

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

4xm: prevent overflow during bit rate calculation

Reviewed-by: Michael Niedermayer <michael@niedermayer.cc>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

doc/filters: fix channel names in sofalizer example

Signed-off-by: Paul B Mahol <onemda@gmail.com>