GNOME Will Move Full-Speed With Wayland Support

[moilami]

What if Canonical is bluffing and Mir's only purpose is to speed up Wayland deployment?

ROFL, that would be simply Epic

[F i L]

I can't stand Gnome anyway. Goodbye and good riddance! If KDE pulls the same stunt and Canonical does nothing about it and tries to force me to use Unity than it looks like I'll be on 12.04 for a long time to come. X.org is and has always been my friend. Fuck all these new kids on the block.

And suddenly... you realize why Mir was a bad thing.

[Hermit]

words is all canonical has too. at this point mir barely does anything and is far less functional/complete than wayland. yet canonical claims they can have mir finished, have backends for GTK/QT finished, have unity totally rewritten in QT5/QML and ported to mir by 2014? Don't make me laugh.

The key issue here is future sustainability. I'm pretty sure Canonical can get a usable display server out of Mir, using most of the architectural concepts of Wayland and leveraging XWayland into XMir (at least they admit they are standing on the shoulders of giants in RAOF's latest G+ post). The binary drivers probably won't be a problem. If nVidia/AMD are going to do binary drivers for Wayland/Mir, it makes sense for them to do a unified EGL driver that supports the widest possible audiences. It may increase the QA testing cost but that's a cost the GPU vendors will just have to bear.

However, the patches for Mesa and the toolkits WILL be a problem. Right now they have out-of-tree patches for Mesa EGL and Qt (QMir). They'll probably do the same for Gtk+ (GMir?). Those patches are not going into mainline anytime soon. How long are they prepared to maintain those out-of-tree patches? The cost to maintain those patches will only increase over time. The longer they're standing out on an island by themselves the more distance there will be between them and the rest of the open source world. The only way this can change is if Mir becomes a wild resounding success and gets picked up by other distros. Then they will have more leverage to force the Mesa and toolkit people to merge in their patches.

And what about the other lesser known toolkits, like EFL and SDL (both already have Wayland support)? Didn't Valve just hire the lead developer of SDL Sam Lantinga? You would think SDL would figure pretty prominently in their Linux gaming plans. How would this affect Valve's commitment to target Ubuntu as the main Linux gaming platform?

[jrch2k8]

Regardless of whether you prefer wayland or mir, both are currently lacking a certain amount of what some would call essential functionality in comparison to X. Off the top of my head:
- Multi-GPU support
- Hybrid laptop (PowerXpress, Optimus) support
- modeswitching API
- multi-display API
These are obviously not insurmountable, but still a lot of work. Just something to keep in mind.

fix me if am wrong
- i think multigpu is actually here partially, since KMS/drm recognize my 4850X2 as 2 separeted 4850 that can be enable at the same time[with 2 sxerver], in wayland should be pretty fast to support just matching the visuals to number N gpu and since prime exist at KMS/DRM level too, it should be quite straightforward to share buffers to allow move thing from gpu N to gpu N+1, i agree crossfire/SLI seem huge work for FOSS stack not sure if blobs will do it internally so its transparent for wayland
- hybrid see below
- i think i read in mailinglist or phoronix that it exist and is in revision[wlrandr]
- im not sure here but KMS/DRM seems quite smart handling outputs since i have in 4850x2 4 dvi outputs in which 2 are actual DVI monitors 1 is an hdmi/dvi converter for my hdtv and the last is an old RGB/DVI converted and all my monitors are detected properly and initialized at boot with its proper modesettings, so it shouldnt be quite trivial to assign specific visual an specific output? so compositors just have to ask wayland, hey give me the active outputs and a list of active GPU's!!! and the compositor let me choose later how i want it to work? ofc keeping a safe default like only 1st output is active or all outputs are clones until you specify it otherwise

no expert so feel free to fix any misunderstanding i could have

[johnc]

And now that I think about it, I can see fedora possibly adopting wayland by default even without proprietary driver support, fedora really isn't a distro to give a crap about proprietary drivers.

And hence why Fedora isn't a distro for most end-users.

Not to pick on them alone, because frankly this "who gives a f about proprietary drivers" mentality seems to be widely held by "the community" and it's evident in the end product. And that's precisely why the stuff put out by "the community" is wholly inadequate for your typical user... and thus why we need a distro backed by a company that understands the needs of most users and not just l33t arrogant pricks.

[Hamish Wilson]

Not to pick on them alone, because frankly this "who gives a f about proprietary drivers" mentality seems to be widely held by "the community" and it's evident in the end product. And that's precisely why the stuff put out by "the community" is wholly inadequate for your typical user... and thus why we need a distro backed by a company that understands the needs of most users and not just l33t arrogant pricks.

Yeah, me liking to use a driver that properly integrates with my system makes me a l33t arrogant prick...

[AJSB]

And they are aiming at spring 2014! Does that mean that if Mir wasn't announced there wouldn't be a usable Wayland display server one year from now? I really do understand why Canonnical couldn't keep on waiting for it to ship.

I believe so.

MIR was the best thing ever *announced* for LINUX. PERIOD.

Wayland was crrrrrrraaaaaaawwwwwwwwlllllinnnnnnnng at snail pace.

At that rate, it would have take 2-3 more years to reach a truly usable state.

MIR was a true Earthquake , Wayland devs started to finally move their arses.

[jrch2k8]

And hence why Fedora isn't a distro for most end-users.

Not to pick on them alone, because frankly this "who gives a f about proprietary drivers" mentality seems to be widely held by "the community" and it's evident in the end product. And that's precisely why the stuff put out by "the community" is wholly inadequate for your typical user... and thus why we need a distro backed by a company that understands the needs of most users and not just l33t arrogant pricks.

well to be honest is not a mentality is a very valid technical reason since 99%[to not say all] commercial drivers are huge undebuggable messes that are pretty nasty to work with that include a bazillion dirty hacks and a bazillion security issues that cannot be fixed unless the manufacturer consider it worth it[in most cases they don't give a rat ass since those fixes don't give them more FPS] that forcibly remove your right to use your hardware at your will[remove a driver so you are forced to buy a new one for example].

so this is more a plea saying "dudes the hardware and the inside silicon is yours, so please give out documentation so we can implement proper native secure drivers"[gallium is a very nice technology is just getting slower due to lack of man power].

so the point is to use a huge insecure blackbox driver i'll use windows that is equally insecure and blackboxed instead

P.S nobody[with 2 neurons working and literate in GPU coding] is aking manufacturer to opensource those huge ugly blobs, the community only ask for low level ASM documentation which in no way can affect their silicon secrets[UVD and Pureview are hold for a different reason tho <--MPAA mafia]

[AJSB]

ROFL, that would be simply Epic

+1 !!!

I dunno if it was a bluff but it sure made a Earthquake in (a) Way(land).

[jrch2k8]

I believe so.

MIR was the best thing ever *announced* for LINUX. PERIOD.

Wayland was crrrrrrraaaaaaawwwwwwwwlllllinnnnnnnng at snail pace.

At that rate, it would have take 2-3 more years to reach a truly usable state.

MIR was a true Earthquake , Wayland devs started to finally move their arses.

you are so uninformed that make me feel sad.

wayland is ready as spec from quite some time, we are waiting for toolkits and DE enviroments[Qt 5.1 and GTK3] and ofc Apps after that and Mir will face the same if ever gets ready.As a note GTK3 and Qt had been porting from quite some time to wayland too and mir have nothing to do with it, DE devs just said publicly they will not support Mir upstream[for a while at least] and they will support Wayland instead and in the case of this thread[beyond the flashy title] are announcing they consider is good enough the current state of the ports to start pushing a full port at max speed.

please literate yourself before post in a flashy title article so you avoid make yourself look like a 15 years old idiot kid