[vlc-devel] Re: vlc: svn commit r18906 (jpsaman)

[Jean-Paul Saman]

Ross,

It is our intention to submit/merge these patches with the Live555 code. 
  We (M2X and VideoLAN team) do not want to fork live555. However this 
is work in progress and needs some more testing before it will be 
offered to you at the live555 e-mail list.

The patch and the live555 demuxer rewrite in VLC seems to keep a better 
sync in RTSP seeking (with DSS and VLC), but at the sametime doesn't fix 
the problem at all. Trying this patch in vlc-0.8.6a made the seeking 
problem even worse. Which is why I hung on to it for a few weeks (and 
that create a fuzz/rumor). In the meantime I expected an update from 
thedj. As you know thedj is working on implementing the last needed bit 
(RTPInfo support).

It was my decision to put it in vlc unstable archive, since it seems to 
fix a few bugs/problems some people are having with RTSP in vlc-trunk. I 
thought it would be wise to give it more testing, so to shake out any 
regressions that might be in there before pushing it upstreams to 
live555 project.

Ross Finlayson wrote:
>> live-starttime.patch the patch needs latest live555 source code.
> 
> Could someone please reassure me about these 'patches'?  It's important 
> that there not be any divergence between the official LIVE555 source 
> code release (available from our web site) and the versions of LIVE555 
> code that get used to build VLC.  In particular, I want people to always 
> be able to build VLC - without any loss of functionality - using the 
> official current LIVE555 code base (available from 
> <http://www.live555.com/public/live555-latest.tar.gz>).

> I notice, however, that people have been referring to patches that seem 
> to modify the LIVE555 code to add new functionality - e.g., the 
> "playStartTime()" function.  At first glance, this new functionality 
> looks good, and I am happy to see it - but at some point the changes 
> should be posted to the "live-devel" list, so that they can be 
> incorporated into the official LIVE555 source release.
> 
> Are these patches just a temporary reflection of 'work in progress' that 
> people have been doing to modify their own copies of the LIVE555 code, 
> so that they can test new functionality in VLC, while ensuring that the 
> VLC nightly builds compile OK?  If so, then that's good (provided that 
> the LIVE555 code changes eventually get fed back into the official 
> source code, as noted above).  If, however, these patches represent a 
> permanent divergence in the LIVE555 code base, then that's something 
> that I definitely want to avoid.  (Because you folks are generally 
> highly clueful, I doubt that's the case, but I just want to be sure :-)

Gts,
Jean-Paul Saman.

-- 
This is the vlc-devel mailing-list, see http://www.videolan.org/vlc/
To unsubscribe, please read http://developers.videolan.org/lists.html