Re: [FFmpeg-devel] [PATCH v2] mov: Add option to keep exact packet sequence after seeking

2015-03-18 Thread Derek Buitenhuis
On 3/17/2015 2:19 PM, Derek Buitenhuis wrote: [...] --- libavformat/isom.h| 1 + libavformat/mov.c | 46 +++--- libavformat/version.h | 4 ++-- 3 files changed, 38 insertions(+), 13 deletions(-) Ping. -Derek

Re: [FFmpeg-devel] [PATCH v2] mov: Add option to keep exact packet sequence after seeking

2015-03-18 Thread Michael Niedermayer
On Wed, Mar 18, 2015 at 01:28:05PM +, Derek Buitenhuis wrote: On 3/17/2015 2:19 PM, Derek Buitenhuis wrote: [...] --- libavformat/isom.h| 1 + libavformat/mov.c | 46 +++--- libavformat/version.h | 4 ++-- 3 files changed, 38

Re: [FFmpeg-devel] [PATCH v2] mov: Add option to keep exact packet sequence after seeking

2015-03-18 Thread Michael Niedermayer
On Wed, Mar 18, 2015 at 02:04:29PM +, Derek Buitenhuis wrote: On 3/18/2015 2:01 PM, Michael Niedermayer wrote: LGTM but like the previous version i did not verify that it actually produces the same sequence of packets, i dont think its critical to verify that ... I've verified this

Re: [FFmpeg-devel] [PATCH v2] mov: Add option to keep exact packet sequence after seeking

2015-03-18 Thread Derek Buitenhuis
On 3/18/2015 2:01 PM, Michael Niedermayer wrote: LGTM but like the previous version i did not verify that it actually produces the same sequence of packets, i dont think its critical to verify that ... I've verified this patch on several pathological samples I have, and it works. - Derek

[FFmpeg-devel] [PATCH v2] mov: Add option to keep exact packet sequence after seeking

2015-03-17 Thread Derek Buitenhuis
The current behavior may produce a different sequence of packets after seeking, compared to demuxing linearly from the beginning. This is because the MOV demuxer seeks in each stream individually, based on timestamp, which may set each stream at a slightly different position than if the file would