XDG-Shell Patches Get Moving For Wayland

[phoronix]

Phoronix: XDG-Shell Patches Get Moving For Wayland

After a lot of mailing list discussions amongst developers that have a stake in Wayland and early patches sent out, the latest xdg-shell patches were formally distributed today on the developers' mailing list. The xdg-shell is a new protocol living outside of the core Wayland protocol...

http://www.phoronix.com/vr.php?view=MTUyODY

[89c51]

May i ask why a protocol for something as important as a desktop was not one of the priorities (if that is the case) and there is a need for replacing the existing protocol??

[antognolli]

Don't forget Enlightenment!

I'm an Enlightenment developer, so you can say that the protocol has support and input from Enlightenment developers too!

[valeriodean]

I'm an Enlightenment developer, so you can say that the protocol has support and input from Enlightenment developers too!

You are right, especially because, I guess, the Qt 5 porting efforts leave the KDE devs without too much time to help in xdg-shell development, isn't it? I seen just Martin to appear in the wayland's mailing list to discuss the CSD/SSD part. Correct me if I forgot something.

[jonnor]

May i ask why a protocol for something as important as a desktop was not one of the priorities (if that is the case) and there is a need for replacing the existing protocol??

Because X.org has served, and continues to serve, the desktop usecases pretty OK, unlike the mobile/embedded usecases. For embedded there is also less need for full-fledged, general solutions (the most complex), one can often accept quite basic features compare to desktop - hence it is an easier starting-point. As a last thing, there is very little money in desktop Linux compared to embedded (and sever but that is less relevant here).

You could interpret this as "desktop is less important".

[mrugiero]

May i ask why a protocol for something as important as a desktop was not one of the priorities (if that is the case) and there is a need for replacing the existing protocol??

Who said it was not one of the priorities? If it needs many iterations to get done, it is probably best outside of the core protocol. Otherwise, you'd probably end up with an X11 situation, where you made a lot of assumptions in the core protocol which became too obsolete to be workable.
Also, if this is actually equivalent to EWMH, this was outside of the core protocol for X, and was developed quite some time later. This kind of hints are really desktop oriented, so adding it to the core would just mean bloat in mobile use cases, so that's another reason to keep it separate. wl_shell should only contain what's common to both use cases, IMO.
Ultimately, they probably waited so the actual interested parties (DE makers) get a say, so it is a better product for them. For this, they need to actually be porting it to know what improvements are needed. Another, extra reason, might be related to the fact some of the most powerful backers are interested in mobile, and everyone scratches their own itches first.