Re: Fedora, Gnome, Weyland and sawfish

Wayland forbids clients to access other clients (or at least it is
designed/meant to do it).


On Sat, 15 Apr 2017 20:41:44 +1000
Allan Duncan <[email protected]> wrote:

> On 15/04/17 18:17, Robert 'Bobby' Zenz wrote:
> > Can you provide a link to what you are referring, please?
> > 
> > Well, X is a monolithic thing. But Wayland either drops features or
> > delegates them to someone else, and most of the time they seem to
> > brush off requests for features with "well, each implementation can
> > always implement that (in the way they want)". Which leaves us with
> > the current situation.
> > 
> > Take RedShift for example, previously we had one tool that worked
> > everywhere without questions. Now with Wayland it is either in the
> > compositor or not at all, which means that every compositor has to
> > implement the features RedShift provided (and people are even
> > cheering for it, see Gnome).
> > 
> > As previously said, I'm kinda bitter about that because I see it as
> > a step back, but hey, X is not going anywhere within the next two
> > decades, so I guess its kinda okay for now.
> 
> After I posted my last, I read through the wayland doc material
> written by Kristian Høgsberg of Intel (came with the wayland modules).
> 
> Although it explains the overall changes made, mainly a reduction of
> the control path, it appears that there isn't any fundamental change.
> There is still a series of clients that generate buffers and the
> compositor which displays them, now directly between the kernel and
> the clients, not a side call from the central X block.
> 
> Unexplained was the nitty gritty such as where something like the
> window manager would exist.  Under X is it one of the clients, or is
> it called from the X compositor?  If it would still be a client then
> the level of conversion
> would be presumably less daunting, although a tall order for this
> small group.
> 


-- 
Sawfish ML