[cairo] scaled_font fini & directfb

[Chris Wilson]

On Sun, 27 Dec 2009 02:21:00 -0800, L A <l.thee.a at gmail.com> wrote:
> Couple of questions, so that I understand this better:
> 
> - Could you eloborate on "simply does not provide an acceleration
> architecture that can be used by Cairo"?
> - What kind of an architecture are we talking about?
> - What is missing in DFB?

Essentially Cairo needs a shaderful acceleration architecture. If the GPU
cannot do tessellation, then we need to feed geometries as the most
compact form of vertices we can supply. So we basically need vertex
attributes coupled with vertex and fragments shaders in order to
accelerate anything more complex than a blt.

At the very least, we require composite! (And just compare the performance
using render-0.10 versus render-0.0 to see why that alone is inadequate.)

> - Why cant Cairo do transformations and scaling through DFB?

Generally it cannot (because as far as I am aware, there is insufficient
support in the dfb api), but if the transformations you want can be mapped
on to the dfb support, depending upon your application that may be
sufficient to achieve reasonable performance. In general, it is not and
the moment you fallback you have lost (many times over).

> - Are there any plans for a gles (1.1 or 2.0) backend?

I'm not planning one. I'm not aware of an open source GLES stack, so that
diversification of the already horrible GL landscape I've ignored. Having
said that, within a couple of extra limitations (which are often required
for the hardware anyway) the algorithms targeting GL should be easy to
reuse with GLES.
-- 
Chris Wilson, Intel Open Source Technology Centre