[mythtv-commits] mythtv branch master updated by beirdo. v0.26-pre-833-g68a6f02

Git Repo Owner noreply at mythtv.org
Thu Jul 12 05:13:42 UTC 2012


The branch, master has been updated on the
mythtv repository by gitolite user beirdo.
       via  68a6f02fadd8a574f2c39c91fad3f7b544915fc9 (commit)
      from  fc980fa364d51aa504bdb1792f465b06e53ad91e (commit)

Those revisions listed above that are new to this repository have
not appeared on any other notification email; so we list those
revisions in full, below.

- Log -----------------------------------------------------------------
commit 68a6f02fadd8a574f2c39c91fad3f7b544915fc9
Author:    Gavin Hurlbut <ghurlbut at mythtv.org> at Wed, 11 Jul 2012 22:05:42 -0700
Committer: Gavin Hurlbut <ghurlbut at mythtv.org> at Wed, 11 Jul 2012 22:05:42 -0700
URL:       http://code.mythtv.org/cgit/mythtv/commit/?id=68a6f02fadd8a574f2c39c91fad3f7b544915fc9

Attempt to plug a memory leak in MPEG2 "lossless" transcode
Seems that when we hit the 2000 frames in use (i.e. queued), we very
ungracefully reject more frames, "No more queue slots!", and then we promptly
leak the packet we were going to put in that frame.  Added a few more
av_free_packet() to hopefully plug the leak.

Also, since it seems the HDHR Prime (and perhaps other sources) record MPEG2
that can sometimes have exceedingly large GOP sizes.  This code was mostly
designed for use with ivtv and others with predicably small GOP sizes.  To
try to get the "No more queue slots!" to disappear, I have increased the
max frames from 2000 to 20000.  This could still use a rework, but hopefully
will fix the short-term issue.

Thanks to rkulagow for reporting this mess. :)



-----------------------------------------------------------------------

Summary of changes:
 mythtv/programs/mythtranscode/mpeg2fix.cpp |    9 ++++++++-
 1 files changed, 8 insertions(+), 1 deletions(-)

-- 



More information about the mythtv-commits mailing list