mpjpegdec: fix memory leak in probe function

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

Merge commit 'a6f19d6a9f8d1e08653d9d77581e8c823f4955c2'

* commit 'a6f19d6a9f8d1e08653d9d77581e8c823f4955c2':
  configure: Support MSVC 2015

Conflicts:
configure
libavutil/internal.h

Merged-by: Michael Niedermayer <michaelni@gmx.at>

swr: Fix ASSERT_LEVEL warning

Found-by: cehoyos
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/hq_hqa: Fix signness of tag

Fixes Ticket4509

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/apng: Add partial support for blending with PAL8 pixel format

Currently restricted to blending pixels that only contain either
0 or 255 in their alpha components

Signed-off-by: Donny Yang <work@kota.moe>
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/apng: Add support for blending with GRAY8A pixel format

Signed-off-by: Donny Yang <work@kota.moe>
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/apng: Add blending support for non-alpha pixel formats

Signed-off-by: Donny Yang <work@kota.moe>
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/apng: Dispose previous frame properly

The spec specifies the dispose operation as how the current (i.e., currently
being rendered) frame should be disposed when the next frame is blended onto it

This is contrary to ffmpeg's current behaviour of interpreting the dispose
operation as how the previous (i.e., already rendered) frame should be disposed

This patch fixes ffmpeg's behaviour to match those of the spec, which involved
a rewrite of the blending function

Signed-off-by: Donny Yang <work@kota.moe>
Reviewed-by: Paul B Mahol <onemda@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

lavf/tls: let the user specify what name to verify against

This can be useful for debugging, or in scenarios where the user
doesn't want to use the system's DNS settings for whatever reason.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

configure: Support MSVC 2015

The C runtime C99 compatibility had been improved a lot and it now
rejects some of the compatibility defines provided for the older
versions.

Many thanks to Ray for the time spent testing.

Bug-Id: 864
CC: libav-stable@libav.org

avformat/mxfenc: Allow overriding /manual setting of the signal standard

previous patch reviewed-by: tim nicholson <nichot20@yahoo.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

atrac3plus: give the phase_shift flag a better name.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

atrac3plus: add support for GHA phase inversion.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avformat/version: Bump version for single jpeg muxer

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

doc/ffserver: Add entry for missing jpeg variant

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

ffserver: Use singlejpeg muxer for jpeg

Fixes Ticket4218

Based on patch by: Otávio Ribeiro
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avformat: Add single jpeg muxer

Needed to fix Ticket4218

Based on patch by: Otávio Ribeiro
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/aacdec: Do not return a uninitialized value

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit 'bc76c46943272515805d7ac48ca39f14826d1fed'

* commit 'bc76c46943272515805d7ac48ca39f14826d1fed':
  aac: Wait to know the channels before allocating frame

Conflicts:
libavcodec/aacdec.c

See: 676a395ab903cac623c5d6ddd0928c789e08a59e
Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit 'a188108ebf28ebac9d2b8fc7d5b391aef45698b3'

* commit 'a188108ebf28ebac9d2b8fc7d5b391aef45698b3':
  aac: Support channel configurations 11 and 12

Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '677c804aa3a78d61b21e6423165a252846c20f0e'

* commit '677c804aa3a78d61b21e6423165a252846c20f0e':
  aac: correctly map 7.1ch-wide AAC from FDK AAC encoder

Sample: FDK_7.1ch_wide.aac
Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '252d6200c36e7eaa79f8d5205b7d731179e94897'

* commit '252d6200c36e7eaa79f8d5205b7d731179e94897':
  avio: Add avio_put_str16be

Conflicts:
doc/APIchanges
libavformat/avio.h
libavformat/aviobuf.c
libavformat/version.h

The FFmpeg implementation is kept as requested by ubitux
Doxy improvements merged

Merged-by: Michael Niedermayer <michaelni@gmx.at>

Merge commit '9b56ac74b170d12027fbc81f581a451a709f1105'

* commit '9b56ac74b170d12027fbc81f581a451a709f1105':
  mpjpeg: Initial implementation

Conflicts:
Changelog
libavformat/allformats.c
libavformat/version.h

Merged-by: Michael Niedermayer <michaelni@gmx.at>

avformat/hdsenc: Change duration from single to to double precision

This slightly improves precision

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avformat/rmenc: Remove float usage

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/libstagefright: Check for pthread_create() failure

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avfilter: fix a few 5 spaces indent

avdevice/iec61883: Check pthread init for failures

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

ffserver: Check for ffio_set_buf_size() failure

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avformat/aviobuf: Check for ffio_set_buf_size() failure

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avformat/ffmdec: Check ffio_set_buf_size() return value

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

ffserver: Check allocations (likely not all)

Reviewed-by: "Reynaldo H. Verdejo Pinochet" <reynaldo@osg.samsung.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

aac: Wait to know the channels before allocating frame

The channel configuration can be delivered only by the PCE,
try to parse it first and not try to decode until a channel
configuration is set.

CC: libav-stable@libav.org

aac: Support channel configurations 11 and 12

These are defined in ISO/IEC 14496-3:2009/PDAM 4 for 6.1 and 7.1.
It also defines another 7.1 layout with configuration 14, that one
is not added here for now.

11: 3/3.1    FC FL+FR BL+BR BC LFE
12: 3/2/2.1  FC FL+FR SiL+SiR BL+BR LFE

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

aac: correctly map 7.1ch-wide AAC from FDK AAC encoder

FDK AAC encoder outputs SCE(front)+CPE(front)+CPE(back)+CPE(back) on
MODE_7_1_REAR_SURROUND configuration.
Since decoder couldn't properly map 4 back channels, decoding failed
unless -request_channel_layout 0x8000000000000000 has been specified.
Now we treat first CPE(back) as CPE(side) on channel mapping.

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

avio: Add avio_put_str16be

mpjpeg: Initial implementation

Support only streams with Content-Length.

Signed-off-by: Luca Barbato <lu_zero@gentoo.org>

fate: add missing avdevice dependency to closed caption test

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

avcodec/allcodecs: Re-order nvenc encoders below x264/5.

We shouldn't pick nvenc by default.

Signed-off-by: Philip Langdale <philipl@overt.org>

vf_colormatrix: calculate coefficients only once

calc_coefficients is no longer being called every frame

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/mpegvideo: Use av_memdup() for allocating thread_context

Also check for allocation failure

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/mpegvideo: Clear thread_context array before allocating

This is probably redundant but its safer

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/mpegvideo: Merge thread context initialization loops

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/nvenc: Add 'nvenc_h264' as an alternative name for 'nvenc'

This allows us to offer the same codec name that libav uses. We don't have
a special way to do aliases, so it's all a bit more verbose than you'd want
but such is life.

Signed-off-by: Philip Langdale <philipl@overt.org>

avcodec/nvenc: Rename nvenc_h265 to nvenc_hevc

For the sake of compatibility, and because pretty much everything else in the
codebase calls it HEVC.

Signed-off-by: Philip Langdale <philipl@overt.org>

Revert "avcodec: Rename nvenc.c to nvenc_a.c, to avoid conflict with the other implementation"

This reverts commit d0d0913702358c61dfb9c5ac8b141cd203524869.

Signed-off-by: Philip Langdale <philipl@overt.org>

Revert "Merge commit 'b08caa87c35a768ec0abb16b1e99c3a85f1df28e'"

This reverts commit d8bbb99c51af08c76dd0a66588fecc5c0106844d, reversing
changes made to d0d0913702358c61dfb9c5ac8b141cd203524869.

Signed-off-by: Philip Langdale <philipl@overt.org>

Revert "nvenc: remove cuda.h requirement for nvenc_a"

This reverts commit f10497042f092d56f82f536276d5998d7fd1e4c4.

Signed-off-by: Philip Langdale <philipl@overt.org>

Revert "avcodec/Makefile: fix checkheaders for nvenc_b"

This reverts commit 3be811a528461816063059377fcb2ac60eabfa77.

Signed-off-by: Philip Langdale <philipl@overt.org>

Revert "Changelog: Drop duplicate NVENC entry"

This reverts commit 440fa7758b687bbb0007f85497eed8bb9aec96bd.

Signed-off-by: Philip Langdale <philipl@overt.org>

avutil/colorspace: Remove RGB_TO_Y/U/V

They are unused and colorspace.h is not a installed header

Found-by: Timo Rothenpieler <timo@rothenpieler.org>
Reviewed-by: Clément Bœsch <u@pkh.me>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/exr: fix crash caused by merge

Various header informations need to be reset when decoding next frame.
Regression since: 95582b5c

Fixes ticket #4597.
Signed-off-by: Paul B Mahol <onemda@gmail.com>

configure: improve the checks for gmtime_r and localtime_r

They are inlined wrapper functions inside the time.h header on MinGW-w64, so
neither check_func() or check_func_headers() work with them.

Reviewed-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: James Almer <jamrial@gmail.com>

Changelog: Drop duplicate NVENC entry

Found-by: Timo Rothenpieler <timo@rothenpieler.org>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avdevice/x11grab: use av_clip() instead of nested min & max

Note: [xy]_off and screen_[wh] variables are int, as well as
X11GrabContext.{width,height,x_off,y_off}.

avcodec/snowenc: use av_clip() instead of nested min & max

Note: AVCodecContext.refs is an int field.

avcodec/mpegvideo: use av_clip() instead of nested min & max

Note: MpegEncContext.mb_{y,height} are int fields, as well as local off
variable.

ffplay: use av_clip() instead of nested min & max

Note: {wanted,min,max}_samples are int variables.

hevc: fix typo (mpv -> mvp).

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Merge remote-tracking branch 'cigaes/master'

* cigaes/master:
  lavf/http: Add simple autodetection for client HTTP method, based on AVIO_FLAG_READ.
  lavf/http: Indent else-clause.
  lavf/http: Properly process HTTP header on listen.
  lavf/http: Rudimentary error handling for HTTP requests received from clients.
  lavf/http: Process HTTP header before sending response.
  lavf/http: Document method option.

Merged-by: Michael Niedermayer <michaelni@gmx.at>

swresample: fix initilaize/initialize typo

doc/filters: fix inaccuarte/inaccurate typo

lavf/http: Add simple autodetection for client HTTP method, based on AVIO_FLAG_READ.

Signed-off-by: Stephan Holljes <klaxa1337@googlemail.com>

lavf/http: Indent else-clause.

Signed-off-by: Stephan Holljes <klaxa1337@googlemail.com>

lavf/http: Properly process HTTP header on listen.

Signed-off-by: Stephan Holljes <klaxa1337@googlemail.com>

lavf/http: Rudimentary error handling for HTTP requests received from clients.

Signed-off-by: Stephan Holljes <klaxa1337@googlemail.com>

lavf/http: Process HTTP header before sending response.

Signed-off-by: Stephan Holljes <klaxa1337@googlemail.com>

lavf/http: Document method option.

Signed-off-by: Stephan Holljes <klaxa1337@googlemail.com>

avformat/mxfenc: Add mxf muxer avclass

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/s302m: Only set the sample rate when some data is output

This way ffplay chooses the mp2 stream for Ticket3890

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avformat/utils: Do not select audio streams with unknown sample rate in av_find_best_stream()

Together with the next commit this prevents non-PCM S302M from being selected unless either
it can be decoded or the user selects passthrough/copy

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/s302m: Check for non PCM Streams

Allow the user to choose what to do with the non PCM data through AVOptions

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

flvdec: fix lack of duration for some files

Fixes #4579

Signed-off-by: Ganesh Ajjanagadde <gajjanagadde@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avformat/flvdec: Show unknown tags at debug level

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avformat/flvdec: Add TYPE_ONCAPTIONINFO

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avformat/flvdec: increase buffer size for parsing metadata string key

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Revert addition of a few fixed_dsp functions to APIChanges

The header is not installed currently so the changes should not be in APIChanges

Found-by: Ronald S. Bultje and others
This reverts part of commit 19ed3e35a5cd188de69f3635f56d9e51c2f720d5.

avutil/ppc/cpu: add include avassert.h

fix build failure on ppc64

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avutil/softfloat: Fix alternative implementation of av_normalize1_sf()

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avutil/softfloat: Fix alternative implementation of av_normalize_sf()

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

arm: only enable setend on ARMv6

Without this check it causes SIGILL crashes on ARMv5.

Reviewed-by: Michael Niedermayer <michaelni@gmx.at>
Signed-off-by: Andreas Cadhalpun <Andreas.Cadhalpun@googlemail.com>

avutil/softfloat: Move av_sincos_sf() from header to c file

The function is quite big and trigonometric functions should not really
be used in speed critical code

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

libavutil/doc: Changes in documentation due to changes to fixed_dsp

New functions are added to fixed_dsp, so the documentation is changed
accordingly.

Signed-off-by: Nedeljko Babic <nedeljko.babic@imgtec.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avutil/softfloat: Add more tests for the normalization

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avutil/softfloat: Fix thresholds in av_normalize_sf()

Found-by: Nedeljko Babic <Nedeljko.Babic@imgtec.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/rdft: Use more specific error codes

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

lavf/matroskaenc: Write Block Keyframe correctly

Per matroska Block Structure [1], for keyframes 0th bit of the flag
should not be set (unlike SimpleBlocks). For Blocks, keyframes is
inferred by the absence of ReferenceBlock element (as done by
matroskadec). This CL writes the flag correctly and inserts the
ReferenceBlock element for non-keyframes. The timestamp inserted is
that of the immediately preceding frame (which is true for VP8 and VP9
- the only 2 codecs using the matroska block element as of now). It
also considers all non-video frames (audio, subtitles, metadata) to
be keyframes.

[1] http://www.matroska.org/technical/specs/index.html#block_structure

Signed-off-by: Vignesh Venkatasubramanian <vigneshv@google.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/utils: Assert that audio decoders do not report using more data than was input

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/alsdec: Check for overread

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/atrac3plusdec: consume only as many bytes as available

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/adpcm: Check for overreads

See: vlc ticket 14649
Reported-by: carl
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/mips: MSA (MIPS-SIMD-Arch) optimizations for HEVC idct functions

This patch adds MSA (MIPS-SIMD-Arch) optimizations for HEVC idct functions in new file hevc_idct_msa.c
Adds new generic macros (needed for this patch) in libavutil/mips/generic_macros_msa.h

Signed-off-by: Shivraj Patil <shivraj.patil@imgtec.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

libavutil/softfloat: Fix av_normalize1_sf bias.

av_normalize1_sf doesn't properly address border case when mantis is
exactly -0x40000000.

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

lavf/riffenc: Set correct block align for mp2.

Rounding by Michael.

Fixes ticket #4565.

lavfi/tinterlace: Double aspect ratio for modes merge and pad.

swresample/resample: fix typos

Found-by: wm4 <nfxjfg@googlemail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/proresenc_anatoliy: Use more specific error codes

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avcodec/snow: replace unspecific error code

Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

swresample/swresample: Cleanup on init failure.

This avoids leaks if the user doest call swr_close() after a failed init

Found-by: James Almer <jamrial@gmail.com>
Reviewed-by: James Almer <jamrial@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

swresample: Add swr_get_out_samples()

Previous version reviewed-by: Pavel Koshevoy <pkoshevoy@gmail.com>
Previous version reviewed-by: wm4 <nfxjfg@googlemail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>