The Future Direction & Purpose Of Wayland's Weston Is Being Revisited

[Delgarde]

Haven't all the DEs around linux implemented their own compositors effectively making Weston just a test thing?

GNOME has their own compositor, and I believe KDE are doing the same... it's more work, but gives them a bit more control over how things work.

But I think some of the smaller projects would be looking at porting to Wayland by re-implementing their shell as a Weston plugin, rather than writing everything from scratch.

[daniels]

But afaik Kristian switched to another project a long ago, maybe other folks you mention either don't work on Wayland either, or only partially.

Well, a quick check of the commit logs will tell you.

[Delgarde]

But afaik Kristian switched to another project a long ago, maybe other folks you mention either don't work on Wayland either, or only partially.

Hardly any open-source developers work on a single project full time. Generally, they'll have one project they spend a lot of time on, but a dozen or so others that they also contribute regularly to. The secondary projects are usually related in some way, e.g a Wayland developer might also have some involvement with graphics drivers, and some involvement with the desktops that are moving to Wayland.

[89c51]

Um, apart from how Collabora sponsors the vast majority of work from Pekka, myself, Jonny, Emilio, Fred, Tomeu, etc; Red Hat similarly for Marek, Peter, Hans, and formerly Jasper (now Endless); Jolla for Giulio; Samsung for Derek and Bryce; Intel for Jason and Kristian; etc etc, ad nauseum.

AFAI understand none of all those people is just maintaning/reviewing patches for wayland (and related projects). Something like linus does for the kernel in a way (gets payed to do just that). Maybe it's not needed but pekka complained in a way in the last paragraph hence my comment/worry.

BTW since you answer questions. Is it true that people are working on a WL remote protocol (not related to RDP or whatever else). Was it just an idea and will never get to production or we will see something when the desktop stuff is there???

[Awesomeness]

Wayland, like BTRFS, is just around the corner.

Wayland is not around the corner. It is already there and is stable.

No they're not.
Having support (1) for Wayland, having full support (2), and having full well tested support (3) are 3 different things. Qt5 is at (1). We the end users mean (3).

SailfishOS uses Qt and Wayland and does so since 2013.

[Almindor]

Wayland is not around the corner. It is already there and is stable.


SailfishOS uses Qt and Wayland and does so since 2013.

This!

I use/develop for the Jolla phone and it works wonderfuly. The UI is stable and smooth, no jerkyness, much more responsive than any android. Qt5 works fine there, my question is why are we not seeing this on the desktop?

[jonnor]

re: Jolla mobile is using Wayland since 2013.

I use/develop for the Jolla phone and it works wonderfuly. The UI is stable and smooth, no jerkyness, much more responsive than any android. Qt5 works fine there, my question is why are we not seeing this on the desktop?

Because:
1) The desktop is a much more general purpose and thus complex user experience. People expect many more interactions to work. Like multiple windows, multiple workspaces, manual+automatic window management, subsurfaces, multiple input devices, even multiple seats.
2) Desktop has a bunch of legacy code to consider. All your existing apps must either be ported to Wayland or use XWayland compat. The window managers/compositors must be ported. And because X11 will be around for a long while still, apps+compositors must add support parallel support for Wayland, and without introducing bugs in the X11 version.
The apps and compositors are generally also bigger in terms of lines-of-code, partly due to 1).
3) There is much less commercial interest in GNU/Linux desktop compared to embedded, and even mobile.

Also, it really did not help that Wayland support is only in Qt5/Gtk3. Many apps have not made that transition yet, which is pre-requisite for native Wayland support...

[Delgarde]

my question is why are we not seeing this on the desktop?

Because code doesn't write itself. Someone needs do the work of porting all the desktop pieces to Wayland, and open-source projects never have enough people to do all the things they want.

[Awesomeness]

1) The desktop is a much more general purpose and thus complex user experience. People expect many more interactions to work. Like multiple windows, multiple workspaces, manual+automatic window management, subsurfaces, multiple input devices, even multiple seats.
2) Desktop has a bunch of legacy code to consider. All your existing apps must either be ported to Wayland or use XWayland compat. The window managers/compositors must be ported. And because X11 will be around for a long while still, apps+compositors must add support parallel support for Wayland, and without introducing bugs in the X11 version.
The apps and compositors are generally also bigger in terms of lines-of-code, partly due to 1).
3) There is much less commercial interest in GNU/Linux desktop compared to embedded, and even mobile.

Also, it really did not help that Wayland support is only in Qt5/Gtk3. Many apps have not made that transition yet, which is pre-requisite for native Wayland support...

And nothing of all that has anything to do with the state of Wayland itself.
Wayland is stable and in shipping products.

[asdfblah]

After all these years of development just imagine how much legacy code and cruft Wayland has taken on.

I think we're due to replace Wayland with something newer.

We should ask Lennart Poettering to do it for us.