Re: The Inevitable Compiz/Xgl/libcm Question...

On Thu, 2006-02-16 at 14:15 -0500, Soeren Sandmann wrote:
> John Stowers wrote:
> > Soeren (and others),
> >  
> > I have been reading the news of Xgl/Compiz with great interest. I have 
> > been browsing the source code of compiz and libcm in the appropriate 
> > cvs servers. I have read the notes of your XDevConf presentation and 
> > also that of Davids at 
> > 
> > <>.
> >  
> > I (and i'm sure many others) are wondering how these things fit 
> > together and what direction development of metacity will take.
> Well, compiz and metacity are similar in that they are both combined 
> window and compositing managers. They are both going to require many of 
> the same infrastructure changes to work well.
> Both are based on GL, so they have basically the same potential for 
> effects. There are no effects in compiz that couldn't also be done in 
> metacity, and there is very little metacity can do that compiz couldn't 
> also do.
> The big benefit I see to extending metacity is that it is an incremental 
> improvement. There has been tons of work put into metacity to make it 
> usable, handle broken legacy application and integrate well with GNOME. 
> Extending metacity makes use of that work.

To me the inevitable question about Compiz/Xgl/libcm is, how does Xgl
fit in there.
Do compiz and libcm+metacity work both on a regular Xorg? I've heard
they do.
What benefits does Xgl give, and does it benefit the metacity
How does AIGLX branch of xorg fit in there, too?

I've heard rumors that accelerated indirect branch of Xorg will give the
acceleration necessary for having the compositor run at an acceptable
speed (it does or doesn't without? Going to test in a bit), and that
AIGLX can be merged into Xorg itself, while Xgl means launching a
separate server on top. Is this somewhat correct?

So, you can see I'm having a hard time understanding what benefits Xgl
gives, that AIGLX doesn't, and if metacity could (optionally) use these
benefits, whatever they be.

With Regards,
Mart Raudsepp

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