[FFmpeg-cvslog] r25380 - trunk/libavformat/utils.c
Aurelien Jacobs
aurel
Thu Oct 7 21:04:48 CEST 2010
On Thu, Oct 07, 2010 at 08:08:00PM +0200, Michael Niedermayer wrote:
> On Thu, Oct 07, 2010 at 03:03:41PM +0200, Aurelien Jacobs wrote:
> > On Thu, Oct 07, 2010 at 12:40:43PM +0200, Michael Niedermayer wrote:
> > > On Thu, Oct 07, 2010 at 11:24:41AM +0200, Aurelien Jacobs wrote:
> > > > On Wed, Oct 06, 2010 at 11:50:16PM +0200, Michael Niedermayer wrote:
> > > > > On Wed, Oct 06, 2010 at 11:46:28PM +0200, Michael Niedermayer wrote:
> > > > > > On Wed, Oct 06, 2010 at 10:52:26PM +0200, aurel wrote:
> > > > > > > Author: aurel
> > > > > > > Date: Wed Oct 6 22:52:26 2010
> > > > > > > New Revision: 25380
> > > > > > >
> > > > > > > Log:
> > > > > > > dynamically use nb_streams instead of static use of MAX_STREAMS
> > > > > >
> > > > > > revert this, its exloitable
> > > > >
> > > > > ret = av_read_packet(ic, pkt);
> > > > > }while(ret == AVERROR(EAGAIN));
> > > > > if (ret != 0)
> > > > > break;
> > > > > read_size += pkt->size;
> > > > > st = ic->streams[pkt->stream_index];
> > > > > if (pkt->pts != AV_NOPTS_VALUE &&
> > > > > start_time[pkt->stream_index] != AV_NOPTS_VALUE) {
> > > > > ^^^^^^^^^^^^^^^
> > > > > only crash it seems but still
> > > > >
> > > > > start_time should not be an array but a value in AVStream
> > > >
> > > > I guess you mean something like this ?
> > >
> > > no, this sets start_time wrongly if its not set.
> > > either the code has to use first_dts if start_time is unavailable or
> > > a new field has to be added to AVstream
> >
> > OK. So the attached patch should be better.
> >
> > Aurel
> > utils.c | 21 +++++++--------------
> > 1 file changed, 7 insertions(+), 14 deletions(-)
> > 6d307a5b6eeb5f6d65c706e2a3514dbbda69de81 start_time2.diff
>
> ok if tested
Applied.
> also make sure you remove things that became unneeded (like local nb_streams
> i think)
Yes sure. I had this in mind, but I didn't wanted to clutter the patch
with unrelated change.
Aurel
More information about the ffmpeg-cvslog
mailing list