GitLab

Fixes load/play behaviour with liboggplay for test case
Reported by Chris Pearce

"liboggz unaligned access fixes v2" patch by Vladimir Vukicevic (:vlad)

remove the unnecessary call to AC_FUNC_REALLOC, which caused
a build failure on these platforms

oggz_comments_decode calls _oggz_comment_add_byname with "value" as NULL
under certain situations. This can then sometimes cause
_oggz_comment_add_byname to try to run strcmp on NULL, which fails.

ie. when all accumulated pages have granulepos -1 and thus
should all be kept when advancing the accumulator

Accumulate pages even for start=0, or tracks with no granuleshift,
for cases where continued pages occur at the chop time.

Includes 2003 project files, and the old 6.0 stuff still

Squash multiple instances of the same comment name=value,
don't hang on scanning comment list by value comparison.
See http://lists.xiph.org/pipermail/ogg-dev/2008-September/001196.html

For portability, to mingw. Reported by Frank Barchard:
http://lists.xiph.org/pipermail/ogg-dev/2009-June/001483.html

bail out of oggz_seek_end() immediatly if get_prev_start_page() fails.

Reported by Ondrej Certik

and add missing hyphens in SEE ALSO sections

Since theora version 3.2.1, the mapping from granules to time
has been shifted by 1, such that the first frame of a newly
encoded file is numbered 1, not 0. Detect that in auto_theora(),
and also report the Theora version in oggz info.

Adds oggz_bounded_seek_set() public API call.

For a single file, comma-separated output, don't put the
terminating newline on stdout, so that the program output
can be used in an HTML5 <video> codecs attribute.

(merged from hogg man page examples)

add 'codecs' tool, to show a list of codecs present in a bitstream (only the first link in a chain is supported)

throughout (#include with "" not <>)