[PATCH weston] README: introduce libweston

[Bryce Harrington]

On Thu, Jul 16, 2015 at 11:55:03AM +0300, Pekka Paalanen wrote:
> On Wed, 15 Jul 2015 09:52:38 -0700
> Bryce Harrington <bryce at osg.samsung.com> wrote:
> 
> > On Wed, Jul 15, 2015 at 01:03:24PM +0100, Daniel Stone wrote:
> > > Hi,
> > > 
> > > On 14 July 2015 at 11:07, Pekka Paalanen <ppaalanen at gmail.com> wrote:
> > > > From: Pekka Paalanen <pekka.paalanen at collabora.co.uk>
> > > >
> > > > What is libweston and where do we intend to go with it.
> > > 
> > > Sounds good to me, although the bit about always using
> > > --with-cairo=image just makes me lean even more towards removing the
> > > other options.
> > 
> > What exactly does that affect inside Weston?  Is it just some of the
> > example clients, or does it affect stuff that actual users would care
> > about?
> 
> It mainly affects the toytoolkit, which flavour of Cairo it will use
> for everything. It doesn't affect the compositor as such, but as it
> affects all demo clients which includes weston-desktop-shell, it will
> impact all users.
> 
> Until we actually fix that, I recommend people to build with the
> default flavour "image".
> 
> A quick fix would be to have a runtime switch in toytoolkit to ask for
> cairo-egl, and default to cairo-image. This way we'd limit the
> cairo-egl impact to just the demos that actually ask for it.
> 
> A more involved fix would be to get rid of cairo-egl completely, and
> remove the support from toytoolkit. Then demos that depended on it
> would need to get their decorations drawn with cairo-image, copied into
> textures, and blitted in GL to the surface.

Mind if we postpone eliminating this functionality for a few months?
I'd like to do some experimentation with cairo performance under wayland
and I'd like to quantify the benefit/cost of cairo-egl before we chuck
it.

Bryce