Format detection and internal frame initialization is moved to a
separate init_image() function, which is called when all the tags have
been read, and so both BitsPerSample and SamplesPerPixel information
has been collected.

This fixes decoding of the file 11.tiff from roundup issue #1925.

Based on a patch by Kostya Shishkov <kostya.shishkov@gmail.com>.

Signed-off-by: Diego Biurrun <diego@biurrun.de>

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

This is safer and helps debugging.

Signed-off-by: Diego Biurrun <diego@biurrun.de>

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

Fix trac issue #188

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

Fixes private options.

Signed-off-by: Anton Khirnov <anton@khirnov.net>

The _fast integer types provide no realworld benefits, but may introduce
portability issues and are just plain ugly.

int/unsigned is the natural memory access type for CPUs, using sized types
for temporary variables, counters and similar just increases code size and
can possibly cause a slowdown.

Adding _POSIX_C_SOURCE to CPPFLAGS globally produces all sorts of problems
since it causes certain system functions to be hidden on some (BSD) systems.
The solution is to only add the flag on systems that really require it, i.e.
glibc-based ones.

This change makes BSD systems compile out-of-the-box without the need for
adding specific flags manually.  It also allows dropping a number of flags
set manually on a file-per-file basis, but were only present to work around
breakage introduced by the presence of _POSIX_C_SOURCE.

Also add _XOPEN_SOURCE to CPPFLAGS for glibc systems.  We use XSI extensions
in several places already, so it is preferable to define it globally instead
of littering source files with individual #defines only needed for glibc.

Fixes NULL pointer dereference.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

The QuickTime RLE encoder only supports 16, 24, and 32-bit color. This patch adds support
for 8-bit grayscale.

Signed-off-by: John Horigan <john@glyphic.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Also delete some unused deblock asm macros.

Problem description, preliminary review discussion at
http://thread.gmane.org/gmane.comp.video.ffmpeg.devel/127731

Signed-off-by: Anton Khirnov <anton@khirnov.net>

default_val was recently changes from double to a union, current code
wasn't updated for that.

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

On cygwin, math.h needs to be included before float.h because of a bug
in the system headers. Including libavutil/libm.h first works around
this issue.

Longer discussion of the topic:
http://thread.gmane.org/gmane.comp.video.ffmpeg.devel/128582

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

seems git missed them and we temporary lost our improvments in them.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Fix decoding of file b.tif, trac issue #168.

Signed-off-by: Diego Biurrun <diego@biurrun.de>

Tested to fix Haiku H264/10bit fate failures, may also fix others.

This is possibly a temporary solution

Includes AVX versions from x264.

Looks better for some cases, worse for others, overall not much difference.
Its more correct though.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Just some cosmetics & comments, the only functional change
"error_resilience: use s->last_picture for accessing last MVs."
Is so buggy that it needs a full rewrite, guess -1.0 PSNR loss wasnt
enough for ronald to realize there was a problem.

Help debugging, safer.

Add link to official specifications, remove author duplication and
others.

Remove one level of indirection, simplify code.

Remove one level of indirection/improve readability.

Also rename variable "extradata" to "palette" ff_cmap_read_palette()
and extract_header(), more meaningful.

Remove the problematic av_dlog() call

This breaks API and ABI.

Should fix compile on systems missing yasm/nasm.

Fixes issue 2679.