MJPEG Tools Svn

Use extended family info for newer cpus (4 bit field could only go to 15)

Couple include files to get prototypes for strcmp and usleep

Fixed: mpeg2enc output bit rate exceeding maximum rate due to faulty
algorithm for self-tuning pass-2 reencoding. New algorithm implemented
that won't have this problem
Fixed: mpeg2enc use of uninitialised variable in pass-2 rate controller.
Fixed: mplex incorrect substream id for multiplexing for dvd authoring.

1) Document that $n (frame count) is no longer subsituted for. The author of
the program hard coded 0 for the value of '$n'.
2) DO NOT USE size_t if '-1' is going to be used as a sentinel! Something like
'if (jpegsize < 0) ...' does not work if jpegsize is size_t. Use ssize_t
instead!

Forgot to comment that the informational message about the input and output
rates needs to be printed BEFORE the '-r' value is clobbered. If -c is given
the program's logic copies the source frame rate over top of the -r value
before entering the conversion. Fix was to move the print statement out from
the conversion routine into the main program before the variable was overwritten

jpegsize of 0 should be treated, I think, the same as "end of file". It's
probably a corrupt file anyhow if jpegsize ends up being 0.

Fix the lavpipe manpage - the warning to remove comments was not clear/strong
enough it seems.

Also in pipelist.c be a little more consistent using mjpeg_error_exit1. Also
check for premature end of file better.

Fix typo - it's "maximum" not "maimum" :)

Cleanup.

# automake 1.9 could use AM_PROG_STDC but 1.10 became picky and needs AM_CC_C_0