Re: Cairo with OpenGLES
Luca Barbato <[email protected]>
| Newsgroups | gmane.comp.lib.cairo |
|---|---|
| Message-ID | <[email protected]> |
On 7/22/11 12:39 PM, Chris Wilson wrote: > On Fri, 22 Jul 2011 13:02:53 +0300, Alexandros Frantzis<[email protected]> wrote: >> cairo trunk contains a working gl/gles2 backend (gles 1.1 is not >> supported). There are some things that still need work, like shader >> workarounds for platforms that don't support GL_OES_texture_npot and >> also proper support for reading from/writing to image surfaces (e.g. if >> they have a pixel format not natively supported by GLES2). Another >> important aspect is performance optimizations. > > And the critical item on my list there is a high-level interface for the > GL backend. We're using the tessellation routines designed for xrender all > too frequently, and not taking advantage of simple geometry... > > And Andrea has got me excited about doing cairo_shape_t again. The benefit > here is that it is meant to be a backend cache of the tessellated > geometry. That would require application level support to fully exploit. > -Chris > Might be nice also split egl from gles2 so vg might get more consideration and people with a separate 2d vector unit might enjoy it. Yes, I know that there are bugs, yes, I know that vg implementations aren't as well behaving as gles2, but still having a nicer support would help mesa to bugfix their implementation. lu -- cairo mailing list [email protected] http://lists.cairographics.org/mailman/listinfo/cairo