[MPlayer-dev-eng] Proposed patch for ASF lock up due to consecutive video packets with no valid PTS.
rob.krakora at messagenetsystems.com
Mon Oct 12 21:00:31 CEST 2009
Basically the problem occurs when a frame arrives with no PTS and giant
invalid PTS value is returned by design and on the next frame it is used in
a calculation and the frame time ends up being over 1000000 and it is passed
to a sleep function causing the application to hang for hours on end. The
overall design looks sound except for this one parasitic case which appears
to be valid. I will attempt another patch that allows for longer frames...
On Mon, Oct 12, 2009 at 2:22 PM, Uoti Urpala <uoti.urpala at pp1.inet.fi>wrote:
> On Mon, 2009-10-12 at 14:06 -0400, Robert Krakora wrote:
> > I improved the patch (see below)...I don't recall getting any responses
> > this yet...
> The video.c code does look buggy. Does playback work if you use
> -correct-pts (which avoids using that video.c code at all)? I don't have
> any asf files on hand here to easily test.
> Your patch assumes that frame durations over one second are invalid.
> However there are valid files with frames longer than that.
> MPlayer-dev-eng mailing list
> MPlayer-dev-eng at mplayerhq.hu
Senior Software Engineer
101 East Carmel Dr. Suite 105
Carmel, IN 46032
(317)566-1677 Ext. 206
More information about the MPlayer-dev-eng