Address trac ticket #1642.

This is useful for the M3U8 format incoming patch.

Allow factorization for the incoming patches, and avoid code/logic
duplication.

Decouple segment index wrapping and segment count logic. The counter
should not be reset since it detects the segment increasing timeframe.

This seems the expected behavior.

Address trac ticket #1599.

This allows the iteration callbacks to discover the internal class and
options, and show them when required.

This is required for iterating through the filter class (without
incurring into a loop).

This is due to the design of components iteration through AVClass
child_class_next() callback, which requires that two components cannot
share the same class.

Consistent with --enable-doc, and easier to type.

The previous code would release the buffer and get it
immediately after, assuming the contents is the same.

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

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

Also print a warning if neither quality nor bitrate is specified
and use the libvpx default bitrate in this case.

The idea of using the default bitrate is from Luca Barbato
Reviewed-by: James Zern <jzern@google.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Reviewed-by: James Zern <jzern@google.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Add files created for MP3 optimization that have MIPS copyright
 notice in them to list of files in mips.txt.
Signed-off-by: Nedeljko Babic <nbabic@mips.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

For some compilers ISA older than mips32r2 is set as default.
 Because of them assembler flag "-mips32r2" needs to be added
 for CPUs that support mips32r2 ISA in part of configure script
 where ISA capabilities are checked.
"-mhard-float" assembler flag is also added for CPUs that have FPU.
Signed-off-by: Nedeljko Babic <nbabic@mips.com>
Signed-off-by: Michael Niedermayer <michaelni@gmx.at>

Require AV_PERM_PRESERVE and leave harmless permissions
alone for the next filter.

This is only necessary because of the poll_frame implementation.
Removing it altogether would be another solution.

The branch with the test was already impossible.

Only write needs to be removed, other can be left.