[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]

Re: OpenSceneGraph



On Sun, 2007-07-01 at 09:47 -0600, Rick L Vinyard Jr wrote:
> Here is Ben's (VTP upstream) reply regarding OSG 2.0:
>  > Rick,
>  >
>  > 2.0 changes the OSG API a fair deal, and doesn't have any compelling
>  > features that i've yet seen. Hence, not much hurry on the adoption. We'll
>  > probably move to it after the next VTP release.
> 
> Ralf Corsepius wrote:
> > On Sun, 2007-07-01 at 11:25 +0200, Hans de Goede wrote:
> >   
> >> Ralf Corsepius wrote:
> >>     
> >>> The background behind all this: 
> >>>
> >>> Do people consider it worth to have OSG-1 and OSG-2 packages in
> >>> parallel?
> >>>       
> >> Note I'm not an OSG user, but judging from this thread, yes that sounds like 
> >> the best solution.
> >>     
> > It's an option I've been considering. Unfortunately OSG upstream has
> > committed a couple of decisions which render parallel installation
> > complicated (e.g. they dropped pkgconfig support :( )
> >
> >   
> >>> The run-time environments/packages could rather easily be made
> >>> installable in parallel (e.g. by introducing a set of OSG-2 packages), 
> >>> but implementing this for the devel packages would be non-trivial.
> >>>       
> >> If it really is a lot of pain to make them parallel installable,
> >>     
> > The devel libs conflict, so the only option I see is installing OSG-2's
> > devel libs into a %{_libdir}/osg-2.0 subdir.
> >   
> Personally, I'd rather see the OSG-1 libs in an osg-1.0 subdir,
This would be incompatible to OSG-1 as it has been shipped until now.

>  and 
> basically treat the OSG-1 packages as legacy packages. That way, the 
> OpenSceneGraph package can just move ahead to 2.0 and packages that need 
> OSG-1 will have to explicitly state it.
This is the replacing "OSG-1 by compat-OSG-1 packages" idea. It's only
applicable to the run-time packages, but not to the devel packages. 


> >>  why not make 
> >> the -devel packages conflict each other?
> >>     
> > That's one option.
> >   
> I rather not see that.

Neither do I, but ... it seems inevitable.

Ralf



[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]