On Mon, Mar 28, 2016 at 3:06 PM, Patrick Walton <pwal...@mozilla.com> wrote:

> Keep in mind that Servo without WR doesn't do dirty rects and invalidation
> either. (I had patches to do it but they weren't complete and bitrotted.)
>

Ok, so that helps to explain Manish's results above.


> The question in my mind is whether it'd be better to build the CPU fallback
> on top of WebRender or whether it's best to build it on top of Skia. There
> are ways to build invalidation on top of WebRender, so I'm not sure we gain
> much by keeping the old painting code around. Furthermore, WebRender 2 has
> the potential to make invalidation and dirty rects easy. It's not clear
> whether WR2 will work, but signs are positive enough that there's a good
> chance that any invalidation work we did today would be obsoleted by the
> WR2 effort. The upshot to me is that it's likely that maintaining the
> existing painting code will not prove helpful for invalidation work anyhow.
>

This makes sense to me.


>
> Patrick
> On Mar 28, 2016 1:50 PM, "Bobby Holley" <bobbyhol...@gmail.com> wrote:
>
> > In general, does the software-fallback path for WebRender mean "really
> > really slow"? If WebRender avoids dirty-rects on the grounds that
> painting
> > is free, then a GL-based software path is going to be a lot slower than a
> > classic software rendering path.
> >
> > On Mon, Mar 28, 2016 at 10:33 AM, <lbergst...@mozilla.com> wrote:
> >
> > > On Monday, March 21, 2016 at 11:45:06 AM UTC-5, Jack Moffitt wrote:
> > > > I propose the following straw man transition plan:
> > > >
> > > > 1. Keep -c, -g, -w command line options as they are, but switch the
> > > > default setting to WebRender.
> > > > 2. Remove -g.
> > > > 3. Add in an llvmpipe software-only fallback and replace -c with
> that.
> > >
> > > I like this plan, or at least the first two steps of it. Does #1
> require
> > > getting any additional WPT/CSS tests working, or is WebRender basically
> > > already there?
> > >
> > > I think #3 might need some more thought to figure out what scenarios it
> > > will support (printing? X11 remoting? headless mode?) to see if
> llvmpipe
> > > will support all of those scenarios. But it'd be great if it did and
> > would
> > > let us reduce the number of rendering backends.
> > > - Lars
> > > _______________________________________________
> > > dev-servo mailing list
> > > dev-servo@lists.mozilla.org
> > > https://lists.mozilla.org/listinfo/dev-servo
> > >
> > _______________________________________________
> > dev-servo mailing list
> > dev-servo@lists.mozilla.org
> > https://lists.mozilla.org/listinfo/dev-servo
> >
> _______________________________________________
> dev-servo mailing list
> dev-servo@lists.mozilla.org
> https://lists.mozilla.org/listinfo/dev-servo
>
_______________________________________________
dev-servo mailing list
dev-servo@lists.mozilla.org
https://lists.mozilla.org/listinfo/dev-servo

Reply via email to