This commit is based on libav's implementation and
makes sure to compare output timestamps together.
It also reduces the differences with avconv.

The changes to the test reference files are caused
by an additional packet at the end, the timestamp
of the frame encoded by this packet is always
strictly below the limit stated by the -t option.

This code is intended for errors in external libraries
when no corresponding error code can be found.
AVERROR_UNKNOWN is too vague for that and looks like AVERROR_BUG.

Some API require a buffer and return a specific error
if it is too small. This error code helps to map codes.

This prevents long loops and OOM in case of damaged timestamps in some
cases
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

sample:

http://samples.mplayerhq.hu/V-codecs/yuv8/MAILTEST.AVIReviewed-by: compn <tempn@twmi.rr.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Update documentation to match code behavior after previous changes, also
reword the examples section.

This should fix debug 2048 amongth other things
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>

* qatar/master:
  cosmetics: Consistently use C-style comments with multiple inclusion guards
  anm: fix a few Doxygen comments
  misc typo and wording fixes
  attributes: add av_noreturn
  attributes: drop pointless define guards
  configure: do not disable av_always_inline with --enable-small
  flvdec: initial stream switch support
  avplay: fix write on freed memory for rawvideo
  snow: remove a VLA used for edge emulation
  x86: lavfi: fix gradfun/yadif build with mmx/sse disabled
  snow: remove the runs[] VLA.
  snow: Check mallocs at init
  flacdec: remove redundant setting of avctx->sample_fmt

Conflicts:
	ffplay.c
	libavcodec/h264.c
	libavcodec/snow.c
	libavcodec/snow.h
	libavcodec/snowdec.c
	libavcodec/snowenc.c
	libavformat/flvdec.c
	libavutil/attributes.h
	tools/patcheck
Merged-by: Michael Niedermayer <michaelni@gmx.at>

The new name is more auto-explicative, as the pb is used for the segment
list file.

Better rather than silently exit.

Also add a comment for clarifying the logic.

Fix trac ticket #1290.

Add doc/mips.txt document with short explanation regarding targeted
platform for MIPS optimizations and regarding MIPS copyright notice.
Signed-off-by: Nedeljko Babic <nbabic@mips.com>
Signed-off-by: Stefano Sabatini <stefasab@gmail.com>

Option help message tell what the option *does*, not what the set field
represents.

Allow to deal with the list file printing only in a single point, thus
simplifying logic, and allow a bit of factorization (no special case
needed when printing the first file name of the list).

Reviewed-by: Jean First <jeanfirst@gmail.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>

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

Also use it in the declaration of the various exit_program
implementations in avtools.

inspired by a clang-scan report.

the av_-prefixed attributes must not be defined outside of this file

Currently, --enable-small turns av_always_inline into plain inline,
which is more or less ignored by the compiler.  While the intent of
this is probably to reduce code size by avoiding some inlining, it
has more far-reaching effects.

We use av_always_inline in two situations:

1. The body of a function is smaller than the call overhead.
   Instances of these are abundant in libavutil, the bswap.h
   functions being good examples.

2. The function is a template relying on constant propagation
   through inlined calls for sane code generation.  These are
   often found in motion compensation code.

Both of these types of functions should be inlined even if targeting
small code size.

Although GCC has heuristics for detecting the first of these types,
it is not always reliable, especially when the function uses inline
assembler, which is often the reason for having those functions in
the first place, so making it explicit is generally a good idea.

The size increase from inlining template-type functions is usually
much smaller than it seems due to different branches being mutually
exclusive between the different invocations.  The dead branches can,
however, only be removed after inlining and constant propagation have
been performed, which means the initial cost estimate for inlining
these is much higher than is actually the case, resulting in GCC
often making bad choices if left to its own devices.

Furthermore, the GCC inliner limits how much it allows a function to
grow due to automatic inlining of calls, and this appears to not take
call overhead into account.  When nested inlining is used, the limit
may be hit before the innermost level is reached.  In some cases, this
has prevented inlining of type 1 functions as defined above, resulting
in significant performance loss.
Signed-off-by: Mans Rullgard <mans@mansr.com>

Codec change midstream gets mapped to a separate stream.

Do not assume avpacket and the decoded frames are independent.

To be absolutely sure and not sprinkle av_free_packet around the code
the call had been placed before getting the frame and on the error path.

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

These functions are defined conditionally so any uses need to have
preprocessor guards.
Signed-off-by: Mans Rullgard <mans@mansr.com>

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

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

It is same as avpkt->size.
Signed-off-by: Paul B Mahol <onemda@gmail.com>

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