Rename ff_put_string to avpriv_put_string

This allows using it from libavformat as well. This will be used
by the RTP/JPEG depacketizer.

Signed-off-by: Martin Storsjö <martin@martin.st>

mjpeg: Rename some symbols to avpriv_* instead of ff_*

These symbols will be used from the RTP/JPEG depacketizer.

Signed-off-by: Martin Storsjö <martin@martin.st>

file: Add S_ISFIFO compatability macro

Not all systems have S_ISFIFO.

Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

x86/mpegaudiodec: fix compilation failure on cygwin with undefined imdct36_blocks_avx

this is a regression since today

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

dsputil_template: Reindent after recent change

Forgot to keep it pretty in e62e455f7b4c5b148f5a289ba823dcc71d6ca389.

Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

x86/h264dsp_init: put a HAVE_YASM back

Should fix compilation on open solaris

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

yadif: cosmetics

tiff/doubles2str: check for truncation

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

tiff/doubles2str: factor out component_len

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

gxfenc: assert that the to be written data has not been truncated.

Otherwise out of array reads would happen later

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

gxfenc: dont duplicate buffer size

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

ffprobe: avoid unsafe usage of the snprintf() return code

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

srtdec: fix unsafe snprintf() return usage

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

ass: assert that the timsstamps fitted in the buffers

Failure of the assert would cause various problems later
if we continue.

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

ffmpeg_filter: fix unsafe snprintf() usage

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

ffserver: fix unsafe snprintf() return usage.

Found-by: "Ronald S. Bultje" <rsbultje@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

ffserver: increase buffer size in http_send_too_busy_reply()

also add a assert0 to make sure the data is not truncated (which
should be impossible now)

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

Merge remote-tracking branch 'qatar/master'

* qatar/master:
  swscale: Provide the right alignment for external mmx asm
  x86: Replace checks for CPU extensions and flags by convenience macros
  configure: msvc: fix/simplify setting of flags for hostcc
  x86: mlpdsp: mlp_filter_channel_x86 requires inline asm

Conflicts:
libavcodec/x86/fft_init.c
libavcodec/x86/h264_intrapred_init.c
libavcodec/x86/h264dsp_init.c
libavcodec/x86/mpegaudiodec.c
libavcodec/x86/proresdsp_init.c
libavutil/x86/float_dsp_init.c
libswscale/utils.c
libswscale/x86/swscale.c

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

aacpsy: use exp2(f) instead of pow(f)(2,...)

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

ffprobe: use exp2 instead of pow(2, ...)

imc: use exp2 instead of pow(2,...)

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

lavu/samplefmt: fix/clarify documentation for av_samples_fill_arrays()

In particular, avoid improper use of the term "channels" in place of
"planes", and fix erroneous documentation for the linesize parameter.

lavu/samplefmt: add note regarding buf validity in av_samples_fill_arrays() doxy

lavfi: drop deprecated and pointless avfilter_default_end_frame() function

Fix warning when compiling boxblur.

While this is technically a major API break, practically there will be no
one using that function since the filtering API is mostly private, so
that function alone is not usable.

dsputil/me/pixels: Actually use av_restrict

Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

mpeg4videodec: Fix "warning: dc_pred_dir may be used uninitialized in this function"

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

smush: read subversion for version==1 too.

This avoids a gcc warning, otherwise it should have no effect.
its a bit more consistent too

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

swr: add av_cold to init/free functions

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

swscale: Provide the right alignment for external mmx asm

This reverts parts of e0c6cce4472. There is external mmx asm that
requires this alignment.

This fixes crashes when using swscale in builds with external mmx,
without inline assembly.

Signed-off-by: Martin Storsjö <martin@martin.st>

dct-test: fix order of evaluation bug

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

ffserver: fix order of evaluation bugs

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

swr/auto_matrix: treat a single channel as mono instead of failing

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

swr/auto_matrix: treat stereo downmix like stereo.

Based on code by John Stebbins
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

swr/auto_matrix: add clean layout function

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

swr/auto_matrix: move layouts to local variables

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

x86: Replace checks for CPU extensions and flags by convenience macros

This separates code relying on inline from that relying on external
assembly and fixes instances where the coalesced check was incorrect.

random_seed: Replace a VLA with a normal array

Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

configure: msvc: fix/simplify setting of flags for hostcc

Signed-off-by: Mans Rullgard <mans@mansr.com>

x86: mlpdsp: mlp_filter_channel_x86 requires inline asm

Signed-off-by: Martin Storsjö <martin@martin.st>

yuv4mpeg: fix potential alignment bug

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

Merge remote-tracking branch 'qatar/master'

* qatar/master:
  mov_chan: Only set the channel_layout if setting it to a nonzero value
  mov_chan: Reindent an incorrectly indented line
  mp2 muxer: mark as AVFMT_NOTIMESTAMPS.
  x86: float_dsp: fix ff_vector_fmac_scalar_avx() on Win64
  x86: more specific checks for availability of required assembly capabilities
  x86: avcodec: Drop silly "_mmx" suffix from dsputil template names
  fate: Drop redundant setting of FUZZ to 1
  cavsdsp: set idct permutation independently of dsputil
  x86: allow using add_hfyu_median_prediction_cmov on any cpu with cmov

Conflicts:
libavcodec/x86/dsputil_mmx.c
libavformat/mp3enc.c

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

ffio_read: dont spam the user at EOF with a truncation error

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

flashsv2enc: only encode diff blocks when needed

A flashsv2 block may have a "diff block" to indicate which scan lines of
the block are actually encoded.  However, this diff block need not be
used when the entire block is coded.

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

mpegts: Fix potential memory leak in mpegts_read_packet()

This leak can happen in case of some error paths.

Found-by: Tanami, Ohad
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

avio: replace ETIMEDOUT by EIO

ETIMEDOUT is not available on all platforms

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

mov_chan: Only set the channel_layout if setting it to a nonzero value

If regularly parsing new chan atoms (as in rtpdec_qt), but the
chan atoms don't actually contain any channel layout, don't reset
the value that the caller has filled in (by guessing).

Signed-off-by: Martin Storsjö <martin@martin.st>

mov_chan: Reindent an incorrectly indented line

Signed-off-by: Martin Storsjö <martin@martin.st>

mp2 muxer: mark as AVFMT_NOTIMESTAMPS.

Signed-off-by: Alex Converse <alex.converse@gmail.com>

flashsv2enc: remove experimental flag, the encoder seems working now

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

libswr: remove redundant ARCH_X86, MMX* implicates X86

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

swsresample: Fix unprotected inline asm

Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

vf_mp: Do not build if inline assembly is not available

Rather than modify the mplayer filter sources, just disable
vf_mp if inline assembly is not available.

Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

dshow: Change WINBOOL to BOOL

WINBOOL is MinGW-specific, and since both MSVC and MinGW
have BOOL, use that instead.

Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

flashsv2enc: move blockbuffer realloc to reconfigure_at_keyframe()

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

flashv2enc: reallocate not only on block count changes but on dimension changes.

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

flashsv2enc: Replace a VLA with a heap alloc

Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

flashsv2enc: fix prev-Z-prime encoding

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

x86: float_dsp: fix ff_vector_fmac_scalar_avx() on Win64

The SWAP macro does not work for explicit xmm/ymm usage, so instead just move
the scalar value from xmm2 to xmm0.

vf_idet: Remove emms

There is no MMX code anyway.

Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

proresdec2: avoid VLA and use SliceContext instead.

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

cache: Only include unistd.h if it exists

This follows suite from f3be3597079be7cd7adbb8392c32e408cadd3da2.

Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

file: Only include unistd.h if it exists

It is included for the open/read/write/close functions. On
MSVC, where this header does not exist, the same functions
are provided by io.h, which is already included.

On windows, these functions are provided by io.h. Make sure
io.h is included if it exists, regardless of the setmode
function.

Signed-off-by: Martin Storsjö <martin@martin.st>
Conflicts:

configure

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

random_seed: Only read /dev/*random if we have unistd.h

unistd.h is used for open/read/close, but if this header does not
exist, there's probably no use in trying to open /dev/*random
at all.

Signed-off-by: Martin Storsjö <martin@martin.st>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

rtpdec: Don't explicitly include unistd.h any longer

unistd.h used to be required for gethostname. On windows, gethostname
is provided by winsock2.h. Now network.h includes both unistd.h and
winsock2.h if they exist.

Signed-off-by: Martin Storsjö <martin@martin.st>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

wtvdec: Remove unused strings.h header

Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

vf_colormatrix: Drop unused strings.h header

It already uses av_strcasecmp.

Signed-off-by: Derek Buitenhuis <derek.buitenhuis@gmail.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

x86: more specific checks for availability of required assembly capabilities

Makefile: remove old codec_names.h on distclean

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

gitignore: codec_names.h is no more

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

lavd/sdl: decrease debug info notice log level from AV_LOG_INFO to VERBOSE

Decrease log spam.

lavd/sdl: remove trailing dot in messages

This is consistent with the apparently prevailing convention.

lavf/utils: fix typo in has_codec_parameters

Replace "unspecified sample size" with "unspecified frame size". +10l.

img2dec: Don't leave AVIOContexts open on zero byte files

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

x86: avcodec: Drop silly "_mmx" suffix from dsputil template names

Merge remote-tracking branch 'qatar/master'

* qatar/master:
  mov_chan: Pass a separate AVIOContext for reading
  af_asyncts: check return value from lavr when flushing.
  mss2: simplify loop in decode_rle()
  mss12: avoid unnecessary division in arith*_get_bit()
  mss2: do not try to read too many palette entries
  mpegvideo: set AVFrame fields to NULL after freeing the base memory
  configure: Set the right cc_e flags for msvc

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

lavf: factor codec id forcing out

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

lavfi: add qp_table_size

This avoid recalculating it and in case w/h changed avoids crashes.

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

fate: Drop redundant setting of FUZZ to 1

cavsdsp: set idct permutation independently of dsputil

CAVS uses its own idct so using dsputil to set the permutation
is fragile.

Signed-off-by: Mans Rullgard <mans@mansr.com>

x86: allow using add_hfyu_median_prediction_cmov on any cpu with cmov

For some reason add_hfyu_median_prediction_cmov is only selected
on 3Dnow-capable CPUs, even though it uses no 3Dnow instructions.
This patch allows it to be selected on any cpu with cmov with the
possibility of being overridden by the mmxext version.

Signed-off-by: Mans Rullgard <mans@mansr.com>

faxcompr: fix out of array read

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

lavc/libx264: remap X264_LOG_INFO loglevel from AV_LOG_INFO to VERBOSE

AV_LOG_INFO is more geared towards messages to be read by the user, the
statistics shown by libx264 with X264_LOG_INFO are more useful at the
debugging level.

Help reducing the log spam.

lavd/dshow: use AV_OPT_TYPE_IMAGE_SIZE

Signed-off-by: Stefano Sabatini <stefasab@gmail.com>

lavd/dshow: support video codec and pixel format selection

Signed-off-by: Stefano Sabatini <stefasab@gmail.com>

lavu/opt: reindent after last commit

mov_chan: Pass a separate AVIOContext for reading

This fixes crashes when called from rtpdec_qt, where
AVFormatContext->pb is null, a crash present since 3bab7cd128.

Signed-off-by: Martin Storsjö <martin@martin.st>

af_asyncts: check return value from lavr when flushing.

Fixes an infinite loop on flush when avresample_get_delay() still
reports some samples but avresample_convert() doesn't return any data.

lavfi: 10l fix () placement

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

lavfi: factor copy_video_props() out

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

libavfilter: pass QP table through the filter chain

Any volunteers to port the pp and spp filters from libmpcodec?

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

lavc/vp6: Disable deblock filtering for Simple Profile

In vp6 Advanced Profile, deblock filtering is conditionally enabled in
each frame header.  In Simple Profile it should always be off.  vp6 was
inheriting the wrong default from ff_vp56_init.

Signed-off-by: Ben Jackson <ben@ben.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Add Dolby/DPLII downmix support to libswresample

Based on code by John Stebbins <jstebbins.hb@gmail.com>

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

swr: update rematrix coeffs to match AC-3

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

swr: update copyright years

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

lavc/raw-test: invert the order of key/value when printing the fourcc->pix_fmt map

Print
FOURCC: PIX_FMT
rather than
PIX_FMT: FOURCC

This seems more consistent with the help message:
-l                list the pixel format for each fourcc

Remove possibly redundant/confusing comment.

lavf/flvenc: apply various log fixes/clarifications

Should improve user feedback in case of errors.

lavf/flvenc: return meaningful error codes

mss2: simplify loop in decode_rle()

It calculates the sum of power of two series, which can be done in one step.

Suggested by Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Kostya Shishkov <kostya.shishkov@gmail.com>

mss12: avoid unnecessary division in arith*_get_bit()

That division can be replaced with a comparison:
((c->value - c->low) << 1) + 1 >= range

By expanding 'range' definition and simplifying this inequation we obtain
the final expression.

Suggested by Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Kostya Shishkov <kostya.shishkov@gmail.com>

mss2: do not try to read too many palette entries

Reported by Michael Niedermayer <michaelni@gmx.at>

Signed-off-by: Kostya Shishkov <kostya.shishkov@gmail.com>