[Bug 217908] Re: Pixellated Images in Firefox/Opera due to incorrect EXTEND_PAD implementation in several video drivers

[Tom Jaeger]

Chris Wilson wrote:
> I'm sorry I haven't take the time to fully review the cairo patch, but I
> just want you to be aware of the work Eric Anholt did a couple of years
> ago to support the extend repeat modes and server-side gradients (but we
> failed to merge in a timely manner):
> http://cgit.freedesktop.org/~anholt/cairo/log/?h=server-gradients

Thanks, I wasn't aware of this.  The patch we're dealing with here is
basically a (small) subset of David Reveman's patch.  I'll see if I can
find the time to update the gradient part of the patch next weekend and
port it to xcb.

> If you care to implement something along those lines, then I will be
> happy to merge it into upstream post-1.10 (which should give us a while
> to check the behaviour on non-Ubuntu systems). And also provide the
> circumstances where implementing h/w accelerated gradients becomes
> worthwhile. I suspect that even with s/w fallbacks switching to XRender
> gradients should be a performance win due to reduced transport costs and
> potentially improved pixmap allocation.

I would think so, too.  The nice thing here is that we are guaranteed
that gradients are handled in the server by software, so we don't have
to worry about drivers implementing them incorrectly.  And if we ever
decide to enable hardware acceleration for gradients, as Michel Dänzer
suggested in [1], driver maintainers will be instantly aware of the
issue, because most (all?) drivers will currently crash if the source's
pDrawable is NULL.


[1]
http://lists.freedesktop.org/archives/xorg-devel/2009-February/000238.html

-- 
Pixellated Images in Firefox/Opera due to incorrect EXTEND_PAD implementation in several video drivers
https://bugs.launchpad.net/bugs/217908
You received this bug notification because you are a member of Mozilla
Bugs, which is subscribed to firefox in ubuntu.