Michael Niedermayer <[EMAIL PROTECTED]> added the comment:

On Sun, Feb 17, 2008 at 09:24:33AM -0000, wg wrote:
> 
> wg <[EMAIL PROTECTED]> added the comment:
> 
> Bringing duration into this discussion is just ridiculous.
> (Just check how often and where "duration" occurs in ffmpeg.c)

?

> "duration" is clearly unset for subtitle streams and
> cannot be set unless you know the future.

mov and avi for example have an index and thus could easily set
the duration but thats not relevant. What is, is that without
some kind of duration (be it fixed, limited per spec or known
per index or duration field in the container) you cannot detect
discontinuities. So if duration isnt set the bug is that ffmpeg
detects discontinuities where it actually can not.

> 
> Maybe the fix for discontinuity could be cleaner by setting
> next_pts to AV_NOPTS_VALUE in the CODEC_TYPE_SUBTITLE cases,
> but basically it's certainly correct and _required_.

Subtitle streams are NOT special. There is no need for handling
them special.
If ffmpeg.c doesnt use duration properly for discontinuity detection
then thats a bug and its not less of a bug without subtitles.
Fixing that bug would fix subtitles, adding a special case would
not fix that bug.
You can check the bugtracker there are several discontinuity bugs IIRC.
The discontinuity detection is far from perfect, this should be improved
not workarounds be added to skip the buggy code in various special cases.

> 
> Same for the "-t total" case, which msg1551 don't address
> at all.

Its a seperate issue and doesnt belong in here.

> And I'm not going to split this into "one issue per source line
> changed" 

No of course not, its one issue for each seperate thing. And you of course
dont have to do anything.
Also sending proper bugreports instead of patches would be an option. Even
with a patch i would like to have a test case if possible!

Anyway http://ffmpeg.mplayerhq.hu/general.html#SEC25 clearly says
"Also please do not submit a patch which contains several unrelated changes."

> because, uhm, it is the _same_ issue 

Not for us

[...]
-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

There will always be a question for which you do not know the correct awnser.

______________________________________________________
FFmpeg issue tracker <[EMAIL PROTECTED]>
<https://roundup.mplayerhq.hu/roundup/ffmpeg/issue356>
______________________________________________________

Reply via email to