Mainlining XWayland To Be Discussed Next Month

[phoronix]

Phoronix: Mainlining XWayland To Be Discussed Next Month

XWayland, the implementation that allows for X11 applications to run inside Wayland by running a root-less X.Org Server, will be talked about next month for possible upstreaming in the X.Org Server...

http://www.phoronix.com/vr.php?view=MTE1ODE

[mark45]

1) Hopefully they finish what they're currently working on on X.org, then put X.org into maintenance mode and all devs move over to work solely on Wayland to get it done sooner. It's been developed since 2008, that's quite some time already.
2) Hopefully the (Wayland) video presentations from XDC will have good audio, so far the audio has typically been bad (while the video quality was better).

[archibald]

1) Hopefully they finish what they're currently working on on X.org, then put X.org into maintenance mode and all devs move over to work solely on Wayland to get it done sooner. It's been developed since 2008, that's quite some time already.

Compare to X, being developed since 2008 makes it barely out of nappies. I think a cautious approach with something like X would be best: it has a long history (being developed since the 1980's?) and is vital to so many appications and operating systems that the consequences of migrating too soon could be severe.

As far as I can see (please correct me if I'm wrong) the main advantage of Wayland is that the old and almost unused (yet required) code in X can be ditched. Given that, I can't see why switching to Wayland is something that is needed as urgently as so many people think. What am I missing?

[mark45]

As far as I can see (please correct me if I'm wrong) the main advantage of Wayland is that the old and almost unused (yet required) code in X can be ditched. Given that, I can't see why switching to Wayland is something that is needed as urgently as so many people think. What am I missing?

It's actually a fair amount of time, since it's not really rocket science and the end project is relatively small (up to 20K lines) and for the last (2 ?) years more devs have been involved (full time?).
Also, it's not urgent, but it's pretty desirable.

As to Wayland's merits: having like 20 times less code is just one reason, another one is that Wayland does all the stuff the right way (utf8 for drag-n-drop, less processes, GPU-oriented, yada-yada-yada) - all of this combined makes a world of difference for the devs working with the graphics stack of the OS and for those creating GUI toolkits and it's easier for new devs to learn the stack and provide improvements.
But for non-geek users there's much less to be excited about, they only benefit seriously in the mid and long term.

[frantaylor]

1) Hopefully they finish what they're currently working on on X.org, then put X.org into maintenance mode and all devs move over to work solely on Wayland to get it done sooner. It's been developed since 2008, that's quite some time already.
2) Hopefully the (Wayland) video presentations from XDC will have good audio, so far the audio has typically been bad (while the video quality was better).

You're aware that X development started in 1984, and it wasn't really usable by normal humans until about 1992, right? That's 8 years! Window systems don't exactly write themselves.

X11 will undoubtedly remain under continuous development for a long time to come. There is no problem with running both X11 and Wayland applications at the same time on the same display. There's no reason to deprecate or eliminate X11. X11 application development will continue on non-Linux platforms for a long time to come. gtk, qt, etc. will gain wayland functionality, but the X11 versions of these toolkits will remain.

The Linux world often forgets that it's just a small part of the computing community. Just because Linux is not putting X11 front and center any more doesn't mean that everyone else is going to drop it like a hot potato. There are a LOT of internal enterprise apps out there that have been running on X windows for 20+ years, on those old standard war-horse Unixes (whose names I dare not mention for fear of starting a flame war), these apps (and the highly-paid consultants who write and maintain them) are going nowhere. X11 is going NOWHERE while there are still companies paying big $$$ to maintain their programs on X11.

"all devs move over to work solely on Wayland to get it done sooner."

Does ANYONE read "The Mythical Man Month" anymore??? You can't make nine women pop out a baby in a month. You can't speed up a software project by adding people to it. It just doesn't work that way. Adding people slows down the project, because the productive people have to take time to bring the new people up to speed. The new people make no contribution until they get up to speed. More people means more meetings and more management structure and fewer people actually getting work done.

More people is clearly better in the testing phase, because meetings etc. are not necessary. Most software companies know this already, they just add 3 to the version number, pass off the betas as products, and get their customers to do their QA work for them.

Wayland is awesome and it's going to make the Linux desktop a much more pleasant experience, but that doesn't mean X has to depart the scene. It will live in its own code base and will continue to be developed by those who care about it.

[archibald]

As to Wayland's merits: having like 20 times less code is just one reason, another one is that Wayland does all the stuff the right way (utf8 for drag-n-drop, less processes, GPU-oriented, yada-yada-yada) - all of this combined makes a world of difference for the devs working with the graphics stack of the OS and for those creating GUI toolkits and it's easier for new devs to learn the stack and provide improvements.
But for non-geek users there's much less to be excited about, they only benefit seriously in the mid and long term.

Thank you for explaining, I can certainly see the (large) benefits now. I misunderstood people's enthusiasm as wanting an immediate non-geek benefit.

[89c51]

Thank you for explaining, I can certainly see the (large) benefits now. I misunderstood people's enthusiasm as wanting an immediate non-geek benefit.

Isn't it also fixing things like the whole boot blinking>black screen>blinking>Xcursor black screen>DM mess. I think it can be started right after kms starts.

I am not sure if it will also fix the copy/paste/clipboard mess.

[bug!]

I hope they will fix or implement something similar to Alt+Tab so we can exit from games.
Anyone have info about this?

Edit: I know about virtual consoles like Ctrl+Alt+Fn, they are good for now but in the long run it would be better to have something more convenient, specific for this task.

[89c51]

Also could someone mention the benefits of the new input method the devs are working on???




@bug

Wasn't there an article here that the virtual consoles will be replaced with something based on WL with the same functionality??

[mark45]

Isn't it also fixing things like the whole boot blinking>black screen>blinking>Xcursor black screen>DM mess. I think it can be started right after kms starts.

I am not sure if it will also fix the copy/paste/clipboard mess.

Yes, as you mentioned, it also fixes some parts of the input mess, and the boot blinking, the copy/paste has been fixed (it won't hang, and clipboard strings are in utf8 by default now), and it has many other improvements that non-geek user's won't notice (immediately) and it's hard to count them all.