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

* commit 'e5c2794a':
  x86: consistently use unaligned movs in the unaligned bswap
Merged-by: Michael Niedermayer <michaelni@gmx.at>

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

This fixes a warning on openbsd and is more robust
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Except for the vintage preset, the values are defined by Lou Logan based
on the ones found in Adobe Photoshop CS6.
Signed-off-by: Clément Bœsch <ubitux@gmail.com>
Signed-off-by: Lou Logan <lou@lrcd.com>

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

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

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

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

Found-by: durandal_1707
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

It appears gcc forgets to increment the byte stream pointer at -O3
I was not able to spot a reason/excuse for it to do that. It
also disappears if the function isnt inlined.

This should fix the gif fate failures on open & free bsd

See: [FFmpeg-devel] [PATCH] gifdec: workaround old gcc 4.2 mis-compiling code
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

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

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

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

* qatar/master:
  sndio_dec: Add missing includes for av_gettime()

no change, as this has already been fixed in FFmpeg
Merged-by: Michael Niedermayer <michaelni@gmx.at>

Fixes fate errors in asv1, ffvhuff and huffyuv on x86_32.

The round to zero behavior is maintained, possibly this should
be changed to round to nearest.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

This should silence many valgrind warnings about use of uninitialized
data, all the warnings i looked at where false positives having their
uninitialized data not actually used.

The same effect could be achieved by listing all code that touches
the pixel buffer in the valgrind suppression file.
Note, valgrind also fails to trace the origins correctly.
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Note: a lot of muxers already do the flushing explicitly.

Remove the temporary hack for amerge and replace it with a
generic solution.

Also downgrade the error to a warning.

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

Change:
"source and destination color space are identical"

which is a true fact, but doesn't tell nothing about what's wrong, to:
"Source and destination color space must not be identical"

which tells more explicitly what the problem is.