[MPlayer-dev-eng] [PATCHES] VOBsub cleanup

[Clément Bœsch]

On Tue, Jan 18, 2011 at 08:14:40AM +0100, Reimar Döffinger wrote:
> On Mon, Jan 17, 2011 at 11:19:13PM +0100, Clément Bœsch wrote:
> > On Mon, Jan 17, 2011 at 09:04:43PM +0100, Reimar Döffinger wrote:
> > > > 2) load_vob_subtitle, just like the other hack, seems at the moment unable
> > > >    to load more than one subtitle at a time; can anyone confirm that? It
> > > >    seems not very problematic since there is generally a single vobsub
> > > >    file for multiple language but just to know :)
> > > 
> > > In theory you should be able to load one subtitle per vo_vobsub struct.
> > > However having more than one probably isn't tested at all, and changing
> > > vo_vobsubs during playback might be real tricky.
> > > However, it means it would be a good idea if load_vob_subtitle returned
> > > the allocated vobsub struct instead of assigning it to vo_vobsub itself.
> > > 
> > 
> > Why? Doesn't this mean in case we want to load more subtitles we would
> > have to change the prototype? In the current situation (after the patch),
> > we would simply have to change the assignation to an item in an
> > array/linked list instead of vo_vobsub; just like the way load_subtitle is
> > designed.
> > 
> > I'm not aware at all of the vobsub system, so I certainly miss your point
> > here.
> 
> Well, loading != selecting the subtitle.
> Assigning vo_vobsub means it gets selected for display, that's why assigning
> it inside that functions seems wrong.
> Your point is likely right as well though.
> 
> > > > 3) If this patch gets applied, I'll edit a bit my -sub-paths patch to
> > > >    allow vobsub to be loaded in specified directories, but it will only
> > > >    load the first find one (if you confirm (2)); it is fine?
> > > 
> > > It is fine (though as said I'd appreciate it if you try to ensure you make
> > > doing this properly not harder, i.e. keep assinging vo_vobsub as much top-level
> > > as reasonably possible).

Ok, changed by passing the vo_vobsub instance (I can't make the function
return it because of the error handling). I'll commit this modified
version in 3 days if no one object.

-- 
Clément B.