Re: Cairo api for standard display API, back-cairo redesign & saving/restoring state to not destroy the current path.

[Banlu Kemiyatorn]

On Mon, 07 Feb 2005 22:30:41 -0700, Aredridel <address@hidden> wrote:

> Any reason not to just depend on Cairo? It's a fast, well thought out
> API from what I can see (no experience yet, but I was looking it over
> today, after talking to pippin.)
> 
> It's a clean dependency, not going to drag in a whole tree -- just
> Cairo, and optionally Glitz.

Yes, there's. On some systems that cairo isn't being ported,
adding dependency on cairo would be a big waste. Or if
we just want to use GDI+ backend, and can't compile it
because we don't have cairo.

I'm thinking if we can implement a pure glitz backend, ie. improving 
tessellation code in the back-xlib to use glitz, would be an interesting
idea too.

In general, I think keeping the recipe of the backend as dynamic
as possible is the best choice.

I agree cairo's api is really well thought out, but a down side is that
it can confuse those who already familair with postscript API. That's
the only reason not to give it too much momentum than other supported
display APIs.

BTW. I never know Cairo is fast (ie. than libart), but I have heard it is
much much faster since last time I tried and I hope I'll have time
to test it soon.