User: Password:
|
|
Subscribe / Log in / New account

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 8, 2013 22:44 UTC (Sat) by nix (subscriber, #2304)
Parent article: The Wayland Situation: Facts About X vs. Wayland (Phoronix)

I'm afraid the whole thing strikes me as so one-sided it borders on disingenuous :( I'd like to like Wayland, but it still gets my hackles up every time I think about it.

Eric's argument for why not to call it X12 is bizarre -- because if it were called X12 people who care about X might care about it, as if they won't care if it's called something else! A better reason for its not being called X12 is that it is clearly not an evolution of X11 in the way X11 was an evolution of X10: it's not a network protocol at all...

"Multi-monitors is a client problem". That's code for "a random half of your applications will eventually support your second monitor: the other half won't", and is just as broken as client window decorations for exactly the same reason. How this is better than X, where nothing but SDL (sigh, fixed in SDL 2) needed hacking for multiple monitors on my system, is not at all clear. (I have a good few (non-free) games that insist on rendering on only one monitor, which is reasonable, but every one of them without exception can be forced to render on a specific one and turn the other one off.)

Nobody uses core X, he says. My Emacs disagrees, it's running on a headless server using network transparency, plus client-side fonts, right now and working quite happily with very low volumes of network traffic and effectively instantaneous screen updates.

"A higher-performance version of VNC", the protocol in which scrolling a nearly-completely-black screen with a bit of text on it over a 1Gbps network takes half a second or so and is juddery as hell. Wonderful. (From other comments on LWN, I understand that VNC is not quite so bad anymore, and that this should be fixed 'properly' by having every single toolkit independently reimplement network transparency, only not a single toolkit has actually done that to my knowledge. Doubtless they will all use totally different authentication mechanisms and have their own network-related security holes. If they reimplement it at all, which I doubt.)

"Every frame is perfect". I don't understand why this is such an issue. Total amount of flashing or tearing I have had in X since I got a graphics card newer than an S3: none, not even when doing simultaneous overlapping textured video playback and oolite. I don't know what the hell the Wayland people are doing with X to make it tear all the time, maybe it's compositing. If compositing is making X creak at the seams this much, maybe a rethink is necessary, but personally given the choice between network transparency and wobbly windows I'd jump at network transparency every time.


(Log in to post comments)

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 8, 2013 22:51 UTC (Sat) by josh (subscriber, #17465) [Link]

> Nobody uses core X, he says. My Emacs disagrees, it's running on a headless server using network transparency, plus client-side fonts, right now and working quite happily with very low volumes of network traffic and effectively instantaneous screen updates.

You've just proven the point. Emacs uses client-side drawing and client-side fonts, like everything else these days. The only relevant applications left using core X drawing/font/etc primitives, or anything other than windows and pixmaps, are the X test suites.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 8, 2013 23:43 UTC (Sat) by nix (subscriber, #2304) [Link]

You've just proven the point. Emacs uses client-side drawing and client-side fonts, like everything else these days. The only relevant applications left using core X drawing/font/etc primitives, or anything other than windows and pixmaps, are the X test suites.
Really? Xaw is a client-side drawing library, is it? 'cos my Emacs is using Xaw. It's using it because of a ten-year-old unfixed bug in Gtk whereby disconnecting from the X server, then shutting it down, then reconnecting to the new one, segfaults. Oops! This is... problematic for emacs --daemon usage, so anyone using that (anyone shutting their desktop down overnight and leaving their Emacs running on another machine) is still stuck using Xaw.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 9, 2013 1:02 UTC (Sun) by josh (subscriber, #17465) [Link]

You specifically said "plus client-side fonts" in your description, so I assumed you ran a version of Emacs using client-side fonts.

Do you have a link for the GTK+ bug report about the segfault when connecting to a new X server?

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 9, 2013 5:46 UTC (Sun) by dakas (guest, #88146) [Link]

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 18, 2013 11:05 UTC (Tue) by nix (subscriber, #2304) [Link]

This bug number is somewhat unusual among bugs since it is now wired into the Emacs source code in a warning printed whenever you try to use --daemon on a Gtk Emacs! It has caused a lot of trouble in the past...

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 10, 2013 5:05 UTC (Mon) by Russ.Dill@gmail.com (guest, #52805) [Link]

Wait...why are you running emacs over X? WTF?

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 10, 2013 8:19 UTC (Mon) by dgm (subscriber, #49227) [Link]

A typical novice mistake. Everyone worth it's bits knows that The Right Way(TM) to do it is run X over Emacs.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 10, 2013 15:04 UTC (Mon) by dakas (guest, #88146) [Link]

Everyone worth it's bits knows that The Right Way(TM) to do it is run X over Emacs.
No, the right way is to use a shell connection (not X) into the machine in question and have it transparently shuffle the files into your local copy of Emacs.

For example, if you do C-x C-f /ssh:frodo@barad-dur.mordor.xxx:/mnt/doom RET then you'll get secure access to /mnt/doom as user frodo on host barad-dur.mordor.xxx without ever having to leave your local copy of Emacs.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 10, 2013 16:47 UTC (Mon) by mathstuf (subscriber, #69389) [Link]

It was really foolish to allow anyone else write access to /mnt/doom. He really needed a better sysadmin.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 10, 2013 23:58 UTC (Mon) by ssmith32 (subscriber, #72404) [Link]

lol. Just finished up a re-read of the trilogy. Thanks for the chuckle :)

Take care,
-stu

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 11, 2013 3:06 UTC (Tue) by madscientist (subscriber, #16861) [Link]

TRAMP is great, but not always the best solution. At work I have a monster system. At home I have a mediocre system. I want to do compiling, debugging, etc. on my monster system--from within Emacs obviously. Also, I don't want to have to recreate my entire Emacs session at home: I can use emacsclient to bring up a remote frame in my at-work Emacs session, displayed on my system at home, and have all my buffers etc., even unsaved work, right there.

Distributed file access is important, but it's not the only thing.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 11, 2013 5:49 UTC (Tue) by oldtomas (guest, #72579) [Link]

By default, these days TRAMP tries to start processes on the box the current buffer comes from. See https://www.gnu.org/software/emacs/manual/html_node/tramp.... Invoking M-x shell "from within" a tramp buffer will bring you a shell on the remote box, and invoking M-x compile will try to run the compile command there too (Emacs 23.4.1 as it comes with Debian, so nothing too recent here).

So it'd be worth to give that a try. That said, there are doubtlessly advantages to the remote X setup, which I do appreciate in other occassions.

For your case (keeping the session alive), Emacs server might be an option too.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 18, 2013 11:09 UTC (Tue) by nix (subscriber, #2304) [Link]

Uh... unless you use TTYs exclusively, the Emacs server *relies* on remote X to pop up a frame on the screen in front of you rather than wherever it was started. It's not an alternative to remote X: it *is* remote X, just a rather more unusual variant in which one program can be connected to multiple X "Display"s simultaneously.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 20, 2013 18:28 UTC (Thu) by daglwn (guest, #65432) [Link]

My emacs 24 doesn't pop up frames when using TRAMP to invoke shells.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 25, 2013 20:57 UTC (Tue) by nix (subscriber, #2304) [Link]

Aaah. Sorry, I completely misread it and had everything turned around. You're right, of course. (And that's a nifty feature: how long's it been there, I wonder...)

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 18, 2013 11:04 UTC (Tue) by nix (subscriber, #2304) [Link]

Do you mean "why am I running it remotely"? I'm running it on X to get lots of fonts and colours and keystrokes (I use shift, control, meta, super and hyper in different combinations and regularly wish there were more bucky bits!)

I'm running it remotely because the remote system (actually 'on the other side of my desk' on the other end of an unshared GbE connection, so not that remote) is faster than the local one, has ECCRAM where the local desktop (like most desktops) has normal RAM, has far *more* RAM which Emacs always likes, has the house RAID array local rather than over NFS, and is the machine that I don't shut down every night. Hence emacs --daemon is a lifesaver.

I cannot imagine that this is a particularly rare use case in these days of expensive power in Europe, given that desktops are often more power-hungry than headless servers. (Though it is true that perhaps not all that many people have a headless server at home, it's certainly not rare among developers, and among those developers a hell of a lot of them likely use remote X or would use something like it if it were available on their OS: remote headless servers are a lot less useful if you can't treat them 'mostly like your desktop' for worky stuff. But perhaps the Linux desktop is no longer targetted at developers?!)

As an aside, I do wish TuxOnIce hadn't gone moribund. I used to be able to suspend my desktop rather than shutting it down...

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 8, 2013 23:13 UTC (Sat) by stevenj (guest, #421) [Link]

"Multi-monitors is a client problem". That's code for "a random half of your applications will eventually support your second monitor: the other half won't", and is just as broken as client window decorations for exactly the same reason.
Read "client" as "toolkit". Applications don't roll their own GUI from scratch anymore, so only a few toolkits need to (and probably will) implement the required "client" functionality for the vast majority of applications to benefit.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 8, 2013 23:45 UTC (Sat) by nix (subscriber, #2304) [Link]

Yeah. Hence 'a random half'. The random half using toolkit A, which implemented multi-monitor support, versus toolkit B, which didn't bother. (I've seen just how slowly some toolkits pick up features. I have *no* confidence in them implementing even multi-monitor support properly, let alone the entire network transparency layer that the Wayland boosters appear to believe will be implemented in each one.)

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 9, 2013 5:03 UTC (Sun) by butlerm (guest, #13312) [Link]

The straightforward way to do this would be to implement a common midlayer graphics API one level higher than Wayland with support for rendering across multiple screens and network transparency. Then the toolkit developers could target that interface instead of requiring those things to be implemented on a toolkit by toolkit basis, which sounds crazy to me.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 10, 2013 8:21 UTC (Mon) by dgm (subscriber, #49227) [Link]

Shall we call it X12?

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 10, 2013 17:45 UTC (Mon) by butlerm (guest, #13312) [Link]

Two major differences - X is a protocol, this would be a library interface. X requires IPC where it isn't necessary, this would not. The protocols used for network transparency should be independent of the API. X went horribly wrong by wiring the two together - if you use any other protocol, you get the worst of both worlds - a hard limitation to what the X protocol can support and a bunch of gratuitous IPCs to a slow and undignified protocol adapter in-between.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 11, 2013 7:48 UTC (Tue) by mmarq (guest, #2332) [Link]

Un-wire it.

And BTW take it to kronos, what ever gets to be a kronos standard is the one that will win.

I like the idea for compatibility
http://hack.org/mc/texts/gosling-wsd.pdf

Something very simple, even perhaps CoreX11 is too much (don't know), the remote case i would keep the protocol and download a conformant piece to every machine... its client, everybody downloads VNC etc if they wanted.

Most important of all take it kronos... let me see, about **ALL** the industry is there (fame and jobs).

The reality is that you are getting obsolete gentleman, the arguing is intense, but you seem to have lost the boat. Input(X) is bad ? ... what about streaminput ? ( i know NIH)... composition ? there is already a OpenWF http://www.khronos.org/openwf/ ( yes i know NIH) GL ? ... well they are at OpenGL4.3, the all edifice is being build around EGL i think (Linux DS is based on whatever fix for a problem seems the best by a few heads in a particular time, irrespective of others)

X, Wayland, MIR ??? ... i think this isn't going anywhere.

Perhaps Android DS can became full industry compliant (Kronos), and we can scrap all that s**t, no matter if many think that in point A or B their lady is much better.

If everybody coding for X (apps) could continue that would be awesome, but that is the only reason i see to give them the trouble to adopt another window system and protocol... **standard and lasting**... techs particularities are terceary importance, after all software can change all the time(to where is what might not be the wise thing, even if some tech argument is valid)

(it has do to with philosophy, complex systems above the "animal" don't evolve by competition, natural evolution crap, they destroy each other instead, like animals that eat each other... the natural paradigma of the human relation is cooperation not animal fagomania)

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 14, 2013 14:51 UTC (Fri) by daniels (subscriber, #16193) [Link]

Khronos's window system stuff (OpenWF, StreamInput, etc) has been around for many years now. And still no-one cares.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 9, 2013 16:25 UTC (Sun) by Cyberax (✭ supporter ✭, #52523) [Link]

You're reading it wrong...

Multi-display will work without any changes in applications right now. After all, applications won't care about who's going to be displaying their buffers with pixels and the compositor can send them to whatever displays it wants.

But if you are going to do something more complicated, like switchable graphics, then you're going to have to support it within the applications. It'd be nice to have it on the Mesa (OpenGL) level, though.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 9, 2013 7:16 UTC (Sun) by callegar (guest, #16148) [Link]

Flashing and tearing is an issue. The ability to give presentations with no flickering on opengl effects (and particularly fading) is important. And unless compositing is disabled, flickering and flashing are there on X, even with intel or nvidia graphics.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 10, 2013 15:13 UTC (Mon) by drag (guest, #31333) [Link]

with compositing disabled the graphic ugliness of X is made much worse.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 18, 2013 11:12 UTC (Tue) by nix (subscriber, #2304) [Link]

I have compositing disabled. No flashing. At all. Maybe if I used programs that were really slow to repaint I'd see it -- but I guess none of them are anymore, not with modern CPU speeds. Maybe back in the 90s... heck, I remember when Enlightenment was huge and bloated and slow, while now it's a bit heftier than it used to be and is renowned for being trim and fast!

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 10, 2013 3:53 UTC (Mon) by daniels (subscriber, #16193) [Link]

> Nobody uses core X, he says. My Emacs disagrees [...]

Sure, and so does xeyes.

xeyes

Posted Jun 10, 2013 8:59 UTC (Mon) by oldtomas (guest, #72579) [Link]

Are you serious? My sarcasm-o-meter is giving contradictory readings...

xeyes

Posted Jun 18, 2013 11:27 UTC (Tue) by nix (subscriber, #2304) [Link]

He's being patronizing and sarcastic because, well, as far as I can tell because I've knocked a hole in his argument and he has no constructive response he can give. Depressing.

It would be less depressing if I hadn't been banging on about the same subject in virtually every thread on the same subject here for *years* and if Daniel hadn't been reading those threads as well and hadn't completely ignored them or been just as patronizing in all of those.

I wouldn't make such a fuss, except, again, this workflow is the one in which I have earned every penny I have ever earned. And it's not at all a subtle or unlikely workflow, unlike the workflow of someone making their living from a remote instance of fucking xeyes.

(Why yes, Daniel *has* just made me very angry. Way to piss off your users and make your users suspect that the Wayland developers don't give a damn about how their actual *users* are using the system they're trying to replace, mate. Well done. Very well done. Do I have any confidence in Wayland at all? Not after *that* comment.)

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 18, 2013 11:23 UTC (Tue) by nix (subscriber, #2304) [Link]

Oh, I'm sorry. Obviously the Emacs in which I have done *all my work for the last fifteen years* and have earned *every penny I have ever earned* is only as important as bloody xeyes.

Sheesh. Talk about patronizing.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 19, 2013 0:07 UTC (Wed) by daniels (subscriber, #16193) [Link]

Yes, and I make a living out of vim. What you're saying, effectively, is that X11 is an abomination and inferior to SSH, just because vim under SSH is much more efficient than under an X11 terminal emulator.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 25, 2013 20:34 UTC (Tue) by nix (subscriber, #2304) [Link]

What? I never said any such thing. If you want to invent words and put them in my mouth, you are welcome to, but that still doesn't mean I actually said them.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 25, 2013 20:52 UTC (Tue) by daniels (subscriber, #16193) [Link]

Right, it'd be putting words in your mouth. Just like you saying I think text scrolling is as pointless as xeyes, is putting words in my mouth. EOT I hope.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 10, 2013 12:25 UTC (Mon) by drag (guest, #31333) [Link]

> Eric's argument for why not to call it X12 is bizarre -- because if it were called X12 people who care about X might care about it, as if they won't care if it's called something else! A better reason for its not being called X12 is that it is clearly not an evolution of X11 in the way X11 was an evolution of X10: it's not a network protocol at all...

It probably was a bit tongue in cheek.

As he pointed out in the article X12 has had a sort of 'draft' status for a long time now and has gone no-where. Long before Wayland came along.

If anything the Wayland remoting will be superior because it acknowledges how modern applications actually want to work rather then pretending that server side rendering primitives have any relevance any more.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 10, 2013 13:17 UTC (Mon) by kskatrh (subscriber, #73410) [Link]

Something called X12 has existed for years and is entirely unrelated to X11 (http://www.x12.org/).

X, X11, X Window System were terrible names IMO, and if anyone was seriously developing a follow-on to X11 called X12 I'd say they were severely misguided.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 14, 2013 15:31 UTC (Fri) by Lennie (guest, #49641) [Link]

People use the same or similar names for different things:

http://en.wikipedia.org/wiki/X10_%28industry_standard%29

This is normal.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 10, 2013 20:14 UTC (Mon) by Corkscrew (guest, #65853) [Link]

> It probably was a bit tongue in cheek.

I'm not sure it was tongue in cheek. The impression I get is that the X community has a very entrenched focus on a particular group of power users. (For example all the commenters on here for whom network transparency is more important than tearing.) That makes it hard to effect major changes - you'll always be treading on someone's toes.

The only way to avoid being shouted down by this group is to essentially create a local fork of the *community* and pull resources (e.g. developers) across in a controlled fashion. In the absence of a meatspace version of Git, the easiest way to do this is to start an entirely new project, as this creates a barrier to entry for naysayers. Hence Wayland (and Mir for that matter).

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 10, 2013 21:17 UTC (Mon) by rahvin (subscriber, #16953) [Link]

The "fork" also allows people to keep using X until they are convinced that Wayland is appropriate for their workload. This is what bothers me so much about people that are up in arms about Wayland. X11 isn't going anywhere. It hasn't changed much in a decade and it will likely continue to not change much for the next decade.

If Wayland becomes popular at some point down the road you might need to switch to use some application BUT I'd be willing to bet that at that point whatever issue you have with Wayland will be mostly addressed (though you will never address the underlying assumption that client side drawing is better) or at least band-aided over to be tolerable.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 18, 2013 11:31 UTC (Tue) by nix (subscriber, #2304) [Link]

X11 isn't going anywhere --- but if people start targetting applications for Wayland, every single such application is an application I suddenly cannot run remotely with anything like the transport efficiency of X apps, particularly not if they are using anything for which GlyphSets would have been used (I don't really care if e.g. 3D games start to use Wayland).

And *that* is a loss of functionality. As mmarq has pointed out, Wayland only doesn't reduce functionality if *nobody targets it*. In which case its existence seems rather pointless.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 11, 2013 22:38 UTC (Tue) by marcH (subscriber, #57642) [Link]

> The impression I get is that the X community has a very entrenched focus on a particular group of power users. (For example all the commenters on here for whom network transparency is more important than tearing.)

... as opposed to the majority of "normal" users of the hugely popular Linux Desktop?

I doubt the power users have more than one girlfriend in average.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 12, 2013 13:50 UTC (Wed) by khim (subscriber, #9252) [Link]

... as opposed to the majority of "normal" users of the hugely popular Linux Desktop?

As opposed to the majority of "normal" users who don't use Linux Desktop (and increasingly don't use desktop at all). For these Wayland makes perfect sense... if you forget one tiny yet game-changing detail. Release date. Wayland released in 2003-2005 and then used in the "smartphones revolution" of 2007-2012 will be huge game changer. But it was not released neither in 2005 nor in 2007. Instead it was conceived in 2008 and released in the end of 2012.

Why "normal" users should use Wayland now when they have already made a different choice is enigma to me. I guess the hope at this point it to port Android (or may be BBX?) on top of Wayland - but I'm not seeing any movements in this direction.

The Wayland Situation: Facts About X vs. Wayland (Phoronix)

Posted Jun 11, 2013 7:59 UTC (Tue) by mmarq (guest, #2332) [Link]

Delusions myfriend delusions...

Whatever follows what is being cocked wit OpenMAX infrastructure has a chance, else niche.


Copyright © 2018, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds