Years After Wayland 1.0, Will 2016 Be The Year Of The Wayland Desktop?

[johnc]

OK so people will have to suffer poorer performance and bugs for a while (wayland will not be perfect in the beginning). But they don't gain anything. Niiiiiice! I wonder why Windows is so much more popular....

MS has been suffering from all the same sort of mental defects that have plagued desktop Linux in recent years.

[Scias]

Hopefully not, there are a lot of very common features which could be implemented in a completely WM-agnostic way using X11 calls only which are now impossible on Wayland due to 'security' 'fixes'.

Nothing prevents developers to agree on a common library or protocols to handle security-sensitive features such as buffer sharing (for color pickers, screencast/screenshot software etc) just like they did for libinput.
So give it some time. As far as I know the EWMH (and even the ICCCM) that standardized a lot of stuff between X11 window managers weren't born with Xorg.

Another misconception is thinking X does a lot of stuff while actually it doesn't. Today almost everything is rendered client-side and OpenGL stuff completely bypass X. X doesn't do any rendering or buffer allocation anymore since DRI2.

Basically the only thing X does noadays is broadcasting all your input and buffers to every possible process on the system and require a shitload of costly useless copies and IPCs when used with a compositor.

[abu_shawarib]

Benefits will often be for developers where they no longer have to worry as much about bad design decisions made 20 or 30 years ago.[/url]

Some people still suffer from this

[zanny]

I want to see Wayland enabling some application that's not possible/is very hard to do with X.
Because so far I haven't seen developers raving about how much easier it si to develop for Wayland.

Because application developers do not develop for Wayland. They use GTK or Qt and those toolkits have implemented Wayland backends. But for whats "hard / impossible to do with X" try tearfree input redirection. Its X11 its impossible for xrandr provideroffloadsink or provideroutputsource without tearing or repainting issues.

The very hard part about X11 is having an application that is secure from hostile other binaries from stealing their keystrokes or what they paint to the screen.

[simotek]

I'm surprised anyone is defending X. It's just a mess. Wayland has some drawbacks as a result of its design decisions, but I don't think there can be much question that it is an important step towards making linux a modern and useable desktop OS.

Foolish you for thinking everyone using Linux wants a modern useable desktop :-P
I do but some people just like the excuse to complain the old ways of the old days are being broken

[Delgarde]

Hopefully not, there are a lot of very common features which could be implemented in a completely WM-agnostic way using X11 calls only which are now impossible on Wayland due to 'security' 'fixes'.

I assume you're thinking of stuff like 3rd-party screen capture tools? If so, there's no doubt that that *is* a security fix... unless you like the idea of arbitrary applications (including dodgy web content exploiting browser/plugin bugs) being able to monitor everything that happens on your desktop? Support for such tools is still desirable, but it needs to be done in a much better way than the X11 method provides.

[curaga]

I thought one big advantage which Wayland would bring would be a perfectly V-synced desktop with no tearing whatsoever.

Yes, except as a Radeon user I've had that on X since I moved to Radeon in 2007 or 2008...

[ileonte]

I assume you're thinking of stuff like 3rd-party screen capture tools? If so, there's no doubt that that *is* a security fix... unless you like the idea of arbitrary applications (including dodgy web content exploiting browser/plugin bugs) being able to monitor everything that happens on your desktop? Support for such tools is still desirable, but it needs to be done in a much better way than the X11 method provides.

Depends what you mean when you say '3rd party'. Right now on X11 you mix and match applications from different 'worlds' and they will (mostly) work - for example you can use something like KSnapshot pretty much without problems on GNOME or XFCE. You can't do that anymore on Wayland since the compositor is the only one that has the full picture - unless the app knows how to speak DIRECTLY to whichever compositor you're using it won't work.

Likewise on the global keyboard shortcuts front - X11 has a mechanism that allows an application to register global shortcuts regardless of what desktop environment/window manager you are using. On Wayland this is AGAIN pushed into the compositor which means that each one will have it's own quirky way of doing it (if at all). Is the X11 implementation of this basically a global keylogger ? Yes. Could this be fixed while keeping the functionality ? Absolutely.

One more thing - remember when an app could position its window where it wanted ? Like for example an application that you 'minimize' to the system tray remembering the size and position of its window so that it restores to the state it was in before being closed ? Yeah, you can't do that anymore either because that's a security risk apparently.

[ileonte]

Nothing prevents developers to agree on a common library or protocols to handle security-sensitive features such as buffer sharing (for color pickers, screencast/screenshot software etc) just like they did for libinput.
So give it some time. As far as I know the EWMH (and even the ICCCM) that standardized a lot of stuff between X11 window managers weren't born with Xorg.

Another misconception is thinking X does a lot of stuff while actually it doesn't. Today almost everything is rendered client-side and OpenGL stuff completely bypass X. X doesn't do any rendering or buffer allocation anymore since DRI2.

Basically the only thing X does noadays is broadcasting all your input and buffers to every possible process on the system and require a shitload of costly useless copies and IPCs when used with a compositor.

I wholeheartedly agree with you on all points. All I'm saying is that IF/when this magical gathering of developers takes place it will take more than a year for everything to be synced and working together in a nice way. There's very little hope that the whole of 2016 is enough time to get everyone in the same boat.

[r_a_trip]

OK so people will have to suffer poorer performance and bugs for a while (wayland will not be perfect in the beginning). But they don't gain anything. Niiiiiice! I wonder why Windows is so much more popular....

Wayland has way less cruft and workarounds, has a better security model and "every frame is perfect", so no screen tearing. I'd say that is end user gain enough. The performance will come.

Also, the whole low level graphic stack in Linux (the kernel) has been overhauled to make Wayland possible. X.org has gained from this as well, but the impetus came from Wayland.

Wayland is also ready for the next decades. X.org isn't. If X.org needs to be brought into the 21st century, they need to gut it and make it something not X11. Oh wait... that is called Wayland.