Submitted by We do not condone or support Canonical in the course of action they have chosen, and will not carry XMir patches upstream.
Ubuntu has to do virtually all its work on Xmir drivers by itself. No one else supports it.
We do not condone or support Canonical in the course of action they have chosen, and will not carry XMir patches upstream.
Ubuntu has to do virtually all its work on Xmir drivers by itself. No one else supports it.
Is there any technical reason for why these patches were not accepted?
Well, technically Intel is invested heavily in Wayland. Hah, nice use of the word “technically”
Anyway, this only means that Canonical has a bit more work on their hands. It’s basically people telling them something along the lines of “You want Mir instead of Wayland then make it work yourselves”.
We do not condone or support Canonical in the course of action they have chosen
That is to vague, what are they *exactly* not condoning?
Using Mir while there is Wayland which all distros except Unbuntu want to build upon. Intel sees no point in Canonical making a diversion.
Edited 2013-09-09 22:37 UTC
To certain point I agree, why bloat the driver just for one distro?, maybe if Mir gets more adoption intel will reconsider.
Question: as someone who isn’t a Linux guy can you or anybody else tell me what is so bad about Mir? Reading the Wiki it seems to take good parts from Wayland and some of the good parts of Android…what EXACTLY is so bad about that?
And isn’t competition good? Even Thom has written about how X-Server just isn’t built for desktops ( there really isn’t a point in using a client/server model when both are on the same box) so you now have two different choices, Mir and Wayland, and this is bad…why exactly?
I don’t know, maybe I’m missing something, but it seems like having two new replacements for X-Server that can compete and motivate each other which sounds like a great idea to me.
If I had to fathom a guess, I’d be guessing its because everyone else is backing Wayland at the moment, and, Mir is simply going to fragment the community.
Also, I believe that Ubuntu originally backed Wayland, but changed their mind.
This is actually a VERY good move long term for Linux. Yes it will make things bad for Ubuntu. However, as both Mir and Wayland uses different code, supporting both would lead to a significant amount of development hours wasted globally (which could be spent on making Linux more competitive, and bringing wayland up to parity with Xorg).
I also believe that Wayland was announced ages before Mir. If this is true, it is irresponsible for Canonical to steam ahead with their own solution.
Oh yes, they were the most enthusiastic of supporters of Wayland a while back… Shuttleworth declaring that Ubuntu would be running natively on it a couple of releases ago.
Then nothing much happened to deliver on that boast, and eventually it was revealed that they’d started their own competing project that they’d been working on in secret for the past year… supposedly because Wayland had a bunch of imaginary defects that made it unacceptable for them (similar defects, oddly enough, to the ones that Mir got caught with in one of the earlier previews).
Actually, Canonical proposed Wayland changed the way Wayland handled multitouch and Wayland didn’t want to change.
So Mir now exists.
That is what happened (as far as I know/understand it).
Edited 2013-09-10 11:55 UTC
Who is everyone else?
Apparently AMD have stated they won’t be supporting Wayland. This leaves Intel who is, and Nvidia who haven’t said yet. If Nvidia do support Wayland I guess their sales will go up while AMD’s fall, but only in small amounts which don’t matter to them I suppose.
Their Catalyst drivers won’t. But the open source AMD driver will.
Yeah, and we all know how good they are. /sarcasm
Actually, the open source drivers are not that bad now for PC use (not laptops). I have been able to play many games using Wine and natively at 30 to 60 FPS.
This time last year the open source drivers simply were not fast or complete enough. But as of Ubuntu 13.04 and Fedora 18 and 19 they are okay for a PC if you are using an AMD 6870 (I have no idea how other cards perform).
Hopefully when the power management patches are enabled by default Laptops will not spontaneously set on fire
.
Here is info regarding the patches:
http://lists.freedesktop.org/archives/dri-devel/2013-June/040436.ht…
Edited 2013-09-10 19:55 UTC
I should clarify that I’m not talking about their open source drivers, I’m referring to their binary drivers.
The open source drivers have to potential to work very well with the rest of the existing infrastructure (no shims, devs talk to each other)
This will make sense if a users is a GPU driver developer, but for the vast majority of users, buying an AMD GPU, install Linux, install open source drivers only because of an FOSS-only ideology, then playing games with sub-par performance doesn’t make sense.
So the open AMD drivers are still wasteful with electricity on the desktop…
That is correct!
As noted the code is there and hopefully it will be on by default soon. I am sure we can all understand why the power management is nice on PCs, but is vital on Laptops.
I guess he was talking about every vendor who has actually made some kind of move to push Linux display technology forward.
I say this because as far as I know, AMD and NVidia’s current plan is to stick with Xorg as long as they’ll be able to get away with it. Which kind of makes sense, since it’s not as if they would make any money out of the effort of going through a Wayland- or Mir-friendly driver rewrite.
Edited 2013-09-10 19:02 UTC
AFAIK, one of the possible transition plans is for them to simply provide EGL support and leaving all the rest up to wayland. This should theoretically reduce the driver maintenance burden.
ATM Nvidia reimplements large sections of Xorg in order to have a working display manager. Amd uses more of the existing X infrastructure (probably why their driver is of lower quality and more likely to break).
So let me get this strait. If submit patches to your project and you refuse the patches and tell me to go pound sand. So i sit down at my desk and write my own code. Now its irresponsible that I didn’t use your code? Who the hell are you to dictate to anyone what code they should or shouldn’t write?
Well, my guess would be that if it’d be too much work to support just one ‘fork’ over the other dozens on the wayland path, then it’s not worth the effort. If, given some miracle, mir would take off and become popular, then they just might reconsider later.
We’ve traditionally had one display server.
First we had XFree86, but that was forked to become Xorg. Later the Xorg developers decided to write a new display server to replace Xorg since some of the code was so poor that they couldn’t even understand how it worked. They went on to create Wayland.
Canonical initially announced support and that they would immediately begin work towards Unity on Wayland. Later, after a year of secret development they announced Mir along with an article full of lies to sell Mir to the public.
Intel is most likely not condoning Canonical’s attempt to fragment the display server, and also their deceitful marketing practices..
Mir is not being developed in the interests of the community, and lives in a bubble where any contributions only really benefit Canonical. Intel and many other companies and people work to develop and support Wayland in the spirit of the community so it is not surprising at all that Intel would push this code back down to Ubuntu where it belongs.
Mir is a single distribution solution so xMir code doesn’t belong upstream.
I would have done the same thing.
Yep, and Canonical will do just that, resulting in graphics libraries needing to support two new backends instead of one, and neither will be done well. Fucking wonderful!
Everyone, except Canonical I guess, expected Mir to cause a major mess for the global Linux community. Did Canonical care?
Edited 2013-09-10 00:20 UTC
Did you expect them to, given past behaviors? I think everyone be they application developers, library maintainers, graphics driver gurus, etc should just make a concerted effort to stop working on anything related to Canonical. They want to do it themselves? Let’s see them do just that.
Note, I’m not being spiteful here. I would actually love to see this exact result happen not to destroy Canonical, but so that we might eventually get a somewhat integrated ecosystem. On which side such integration will happen is something I cannot predict. The wider community has more resources, but very little direction. Canonical are the reverse: they obviously know what they want and why, but aren’t focusing as much of their resources on it as they should and are being hampered in some ways by the community. I think Canonical should be put through the crucible, so to speak. I think they have it in them but, as long as they’re concerned with the wider community’s development and reactions, they won’t ever focus on themselves enough to make their operating system happen.
Yup, when you live in the FOSS world, this is exactly what happens. If you want the code to be open and allow people to do whatever the hell they want with it, don’t get too butt-hurt when they do just that.
This is what kills Linux on the desktop. Since you don’t have a person or group in charge of the ship, it just kind of wanders all over the place with no real direction. Even if you get 99% of the community going in the same direction, all it takes is the other 1% to screw it up, esp if that 1% happens to have a lot of $$.
Edited 2013-09-10 00:23 UTC
More like there is no ship to begin with and never will be. Just a whole bunch of rafts built from parts of a larger forest.
As for the other guy, no it doesn’t mean graphics libraries will have to support two back ends. Quite the opposite really. If everybody but Ubuntu goes to Wayland, it’s on Ubuntu to handle the support or become increasingly irrelevant.
I agree with the first but I think Canonical is trying to fix just that. Actually build a ship and control it as much as possible instead of being dependent on the community which goes all over the place and not necessarily along Canonical’s vision
yes, but it is it’s own ship. Canonical want to have total control over it, close development with no community involvement and worst of all, the agreement about the ability to relicense the code.
That’s why nobody, except Canonical, is going on that boat.
Good luck then!
Not saying I support it but it isn’t any different than what Google did with Android which is also Linux deep under but not compatible with any distro and yet people like to say Linux now rules because Android does.
Edited 2013-09-10 14:09 UTC
Yep, that’s pretty much Ubuntu’s problem in a nutshell: it’s in that no-man’s land where it’s just successful enough to make Linux fanboys bitter & resentful, but not successful enough that they’ll brag about it or try to assume undue credit for Canonical’s achievements.
The same one as used by the KWin developers – an unwillingness to accept patches that are useful only for a single distribution. Such patches are only a maintenance burden for upstream developers, adding extra code paths that they don’t use, but which need to be kept working when making changes.
And it’s not enough to say “Canonical devs will maintain it”, because even ignoring trust issues, reality doesn’t work that way – there are no clear demarcations between “stuff for upstream devs” and “stuff for Mir devs”. It’s *all* stuff for upstream devs…
This kind of crap is why you will never see Linux get past 1% on your average pc. Too many variables to count on any support from anyone. It’s too bad though, as a replacement for X11 is badly needed for home desktops. Anyone know if Intel is, or is planning to, get behind Wayland?
TFA said that they’re heavily invested in Wayland, and have even hired Wayland’s founder.
Intel is one of the main contributors to Wayland, it is being pushed as the technology for Tizen and Jolla. One of the Jolla developers was the one that coded libhybris in order to support Android drivers on glibc systems with Wayland, and is used by Mir too, so they are late on this race
Edited 2013-09-09 21:48 UTC
Lets not be too dramatic given that Intel is already supporting Wayland and it is only Ubuntu as far as I know who are hell bent on having their own project – for everyone else such as Fedora, SuSE etc. they’ll be using Wayland with Ubuntu eventually having to come up with some compromise for the sake of not being left out in the cold. IMHO the display server is one of the last sore points of Linux that really needs addressing and if done right along with support from GPU vendors there is a greater chance that other parts will fall into place.
If you want to shoot yourself in the foot, the community will be more than happy to let you do it all by yourself.
Yeah. Too bad it’s the users that get hit by the bullets.
Edited 2013-09-10 00:11 UTC
Ricochet is the word you are looking for
To be fair, Wayland was started years before Mir, and the creation of Mir is only possible due to the efforts of X developers building the path towards wayland. And yes, those developers include some from intel.
Canonical are 100% at fault here, and should have just helped wayland instead of creating Mir. Wayland is already usable, and will be available for Fedora 20.
If you want to use desktop linux, you should be running a distro that uses systemd (or is moving towards it), and uses wayland.
We have that common stack for KDE, Gnome, XFCE, Mate, LXDE, etc. etc.
The *only* place it differs is in Canonical’s Unity; making their name even more ironic.
It’s not the FOSS community at fault here. It’s canonical – a company.
Available as a partly-working technical preview, remember. Wayland (and desktops on Wayland) are coming along nicely, but what ships in F20 isn’t going to be ready for end users.
Last I heard, Gnome Shell was now running natively on Wayland, but there are lots of bits not yet working. I don’t know about KDE, but I assume they’re in a similar state…
Sssh, mere technicalities :p
The point is that mir won’t be fully functional first, so they’re causing division for no reason – everything mir will be able to do will be possible in a wayland client.
KWin needs to finish the port to Qt5 and QML2 first, but it is progressing very well.
KF5 and plasma desktop 2 are not that far away.
Don’t deceive others and don’t deceive yourself. The only thing Wayland worked as a preview AFAIK is through rebeccablack. I don’t know if there are some latest developments. The only thing that is clear at the moment is MIR is going to ship with 14.04 LTS. Thanks to Wayland and several other existing libraries that MIR uses if not because of them, MIR is impossible to deliver. And for Wayland, in F20?
The point was that Mir will not be functional and available before wayland, because they’re still using an x-compatibility layer.
If their NIH syndrome were not so bad, they could have just helped with wayland, and it would have been fully functional way before mir will be.
Ubuntu had good reason to think that the dominant commercial Linux community that doesn’t like it (i.e. Red Hat) would act to thwart Ubuntu’s attempt at a unique independent user experience, exactly like they did with GNOME 3.0. Remember that GNOME fought tooth and nail against letting the common user know how to shut down a PC…even when sleep or hibernation would crash the majority of PCs running Linux. Ubuntu at least fought this absolute idiocy.
Remember that Wayland has been around for FIVE YEARS with trivial progress before Mir was announced, and wasn’t inclined to accept Ubuntu’s patches. Given that, I don’t fault Ubuntu, and Intel’s action is understandable given their developer association with Red Hat.
I’m not an Ubuntu fanboy, as my time is 60/20/20 for Windows/Ubuntu/Sabayon. In the end, IMHO, Red Hat is jealous of Ubuntu’s fame with annoyance at Ubuntu’s trivial support for core elements such as the Linux kernel. However, IMO, such support should be fairly supported in proportion to profits, not fame.
You have made several factual inaccuracies.
Intel is the primary backer of Wayland. Red Hat does work on other parts of the graphics stack including the input system but most of the contributors are either from Intel or Collabora
Canonical was the first commercial vendor to announce that they will be supporting Wayland however they never contributed to it and developed Mir for nearly an year in private without ever talking to Wayland developers who work for Intel and not Red Hat.
They also made some incorrect assertions about Wayland to justify creating Mir but had to fix it quickly after the PR disaster.
Also Wayland developers made enormous strides in making the transition from Xorg smoother including working on DRI and XWayland and Mir wouldn’t even have been possible without all that work according to the Mir developers themselves.
That explains it even more, doesn’t it?
So you’re saying they behave exactly like RH does on many of it’s projects?
Why is Intel to be blamed for Canonical’s NIH exactly?
Please, Red Hat has nothing to do with Gnome decision despite their influence. Replace Gnome 3.0 by Unity, you have the same experience until it was fixed. Show how easily to install Unity on other distributions. Gnome Shell fully supports suspend and hibernation without crashing unless you use an obscure driver. Common users as you call have no problem using Gnome Shell. Shutdown feature was always available by pressing Alt to replace “Switch” prior to 3.6. That behaviour can be changed by either dconf or Tweak tools (now part of Gnome 3.10 by default)
Display server is argually the most complicated software to work on meaning the process are non-trivial considering the complexity of the nearly 30 years old X server. Without these five years done on Wayland, Mir would not exist, period. Many Canonical patches were either distro specific, hacks or bad hence reject. Wayland is distribution agnostic backed by several organisations, one of them is Intel hence this topic.
To quote a French expression: “qui s’excuse s’accuse”. How Red Hat be jealous considering its biggest contributions in the Linux ecosystem especially the kernel Canonical heavily depends on?
Because Ubuntu has made inroads in the sector where RH failed spectacularly and never made a dent, namely desktop? (No Fedora doesn’t count and it’s pretty much also a failure in this sector)
Edited 2013-09-10 08:20 UTC
How But Redhat gave up on a providing an desktop years ago (before Ubuntu even came out). I know RHEL has a desktop configuration but it isn’t where they make their money.
Edited 2013-09-10 08:24 UTC
Yeah, exactly.
How could they give up something that they never had or even tried to have?
http://www.techhive.com/article/144749/article.html
Redhat 7.3, 8 and 9 were pretty good desktop Operating systems, and you could pick it up in PC world in the UK in a nice shrink wrap box.
http://www.amazon.com/Red-Hat-Linux-9-0-Personal/dp/B00008QODZ
Most of the reviews are from 2003.
Edited 2013-09-10 17:29 UTC
As you probably know lucas_maximus I’ve followed RHEL for quite a long time, long before RedHat 7 hit the shelves. Those were never intended for consumer desktops, they were geared for and great for workstation class systems and servers.
Thanks for correcting me on that, however it was easy to get the opposite impression.
I remember, RH announce a global linux desktop or something I forgot the exact term. And then eventually gave up the idea.
Fedora succeed enough on my disk sectors.
I like Linux distros that are simply Linux distros. Ubuntu is slowly departing from that ecosystem.
RH charge for their Dekstop, while Ubuntu is free with paid extras. I think this is the cause of popularity.
No, not really. RH desktop is too much focused on enterprise stability (read: hopelessly outdated) to be of interest to consumers.
RH already gave up on the consumer desktop because they couldn’t make a dent.
No, they never tried to enter the consumer market because they didn’t believe it was profitable. Don’t try to rewrite history with your ignorance.
You’re wrong! RH tried that:
http://www.techshout.com/software/2007/10/red-hat-global-desktop-an…
And then eventually gave up:
http://blogs.computerworld.com/red_hat_abandons_desktop_linux_for_c…
Edited 2013-09-12 05:19 UTC
Global desktop wasn’t meant to be a consumer desktop at all. It was a Intel specific program designed for targeted markets.
Is that why your own article states:
“The Global Desktop is not merely a consumer version of the recently released RHEL (Red Hat Enterprise Linux) Desktop 5. The RHEL Desktop 5 is meant for enterprises that need high levels of desktop security and comprehensive network-based management tools.”
Now, put your fingers in your ears and start shouting to drown out the logic and reason.
Ever tried redhat linux 8.0? That’s about 12 years ago. It’s, AFAIK, one of the best linux desktop. Quite easy installation, quite beautiful, and quite useful, for early linux desktop. Its free.
And then, they abandon that ship to focus on RHEL. Redhat linux then succeeded by Fedora, but fedora quality worse than their predecessor.
CMIIW.
I’m sorry to say, but Ubuntu has done nearly NOTHING to advance GNU/Linux on the desktop. GNU/Linux still represents less than 2% of the desktop market (consumer or enterprise). That’s absolutely laughable!
To add insult to injury I’m a person who strongly advocates for open source, so to say that GNU/Linux has made in-roads into the common desktop marketplace is just wishful thinking. Has it gotten better, sure, is it worth being out chest over, no. Your average user still can’t install software without jumping through various hoops to get what they want.
With regards to enterprise stability, would you care to tell me how many enterprise desktops that you support? I support about 2500 machines running various GNU/Linux distributions including CentOS, RHEL, Fedora, Suse and Ubuntu and while CentOS, RHEL and Suse could be considered “hopelessly outdated” they’re rock solid and just work. They have features that people who run in a real enterprise environment *NEED* to manage them. They don’t change much in between releases and are easy targets to test.
Fedora and Ubuntu are not enterprise desktops at all. They’re all the latest and greatest bits thrown together and they break *often*. This is expected in a home user desktop or a development environment where you want to test what will become the next enterprise desktop. I look forward to hearing you complain about how Canonical’s “Enterprise Desktop” is hopelessly outdated in 3-5 years as that seems to be what people are complaining about WRT CentOS/RHEL/Suse enterprise desktop offerings that are now already 3-5 years old.
Guess what! People actually want stability. It’s why Canonical now has a LTS release in the first place. Your average enterprise doesn’t give a flying fig newton about the latest and greatest! They care about stability, manageability and maintainability! It’s an enterprise desktop not a freaking home PC that nobody gives a shit about!
Red Hat decided that the HOME DESKTOP environment was dead. There was NOTHING there and there still isn’t. A GNU/Linux home desktop doesn’t matter! There’s no money there and there won’t be for quite some time. Ask Canonical how many home users that they have with Canonical support contracts. I would hazard to guess not many.
Oh, and I should mention that many of my enterprise desktops have a full suite of development tools available to them including the latest compilers, toolkits, libraries and the like. All with this “hopelessly outdated” enterprise environment.
To be fair with Fedora, that is the first distro I used as a mail server. And it has many unique technologies. The only problem is that it becomes obsolete after 18 months of usage. So I default my desktop to Ubuntu L.T.S. for practical reasons. Ubuntu having MIR is not a problem for users, its a problem for other distributions, since we still be using Ubuntu on business desktops. I will be watching Fedora from time to time.
http://fedoraproject.org/wiki/Releases/19/FeatureList
That’s the main reason why all of the whinging over this “issue” seems so overblown. It’s essentially the same as what happened with Nokia – there’s the self-serving myth where Maemo/Maego/whatever was poised for imminent success, if only big mean Microsoft hadn’t come along and smothered it in the crib. But reality is that Maemo didn’t NEED to be smothered by Microsoft, because it was already stillborn – Nokia had the better part of a decade to build a Linux-based mobile OS, and instead they wasted the time with directionless floundering, political in-fighting, etc.
Nokia doomed their Linux OS long before Windows Phone was even a glimmer in Microsoft’s eye. And what makes Nokia’s missteps truly inexcusable was when Palm came along in 2009, started their own Linux-based mobile OS and basically overtook Nokia from a standing start. Palm went from conception to having a product to market… in less time than it took Nokia to decide on a UI toolkit. It’s like the fable of the tortoise and the hare… except in this case the hare still won even after sleeping through most of the race. ‘Course, for Palm, their “prize” was becoming part of a figurative rabbit stew, but that’s besides the point.
Same thing with Wayland/Mir: the current “problem” would never have existed in the first place if Wayland’s development wasn’t so glacially slow. I guess Canonical was just supposed to sit around with their thumbs up their asses, while the Wayland developers spent another 5 years bike-shedding…
“Same thing with Wayland/Mir: the current “problem” would never have existed in the first place if Wayland’s development wasn’t so glacially slow. I guess Canonical was just supposed to sit around with their thumbs up their asses, while the Wayland developers spent another 5 years bike-shedding..”
Nonsense. Wayland developers (who are really Xorg developers btw) wrote major portions of KMS, DRI etc that made the transition away from Xorg even possible and Mir simple wouldn’t even exist without all that work done from Wayland developers and in the open source world, if you want to speed up a project even further, you submit patches rather than write a new alternative from scratch.
I predict this will be worse than Amiga mythos…
I remember fighting this at gnome-mailing list. And I’ve stopped subscribing to that list because of this. I think, from what I’ve experience from that mailing list, most GNOME developers knew nothing about the needs of a common business desktop PC, or just blinded. Their reasoning is that, they want to force users to use hibernate so that if some problem occurs, the more bugs they received, then they can fix hibernate issues in computers. But the subscribers(I think many were developers) were mostly using LAPTOPS, and I told them that in a business, we need to properly shutdown computers, unlike laptops. And it took them many releases before they gave up to this foolish reasoning and offered users the way how to SHUT DOWN computer. Previously they told users to LOG-OUT so that they can shutdown or press the ALT to reveal the hidden menu.
Woohh…
Condone? There’s a strike force heading towards the Isle of Man isn’t there? One of these days I’m going to end up getting an Ivy Bridge, but stop acting like you’re Darth Vader.
Really? I’m pretty sure Intel’s not the only GPU company out there. Do you have some secret knowledge about what the other manufacturers are planning?
So, Intel, Red Hat, SUSE, are all backing Wayland. Great – I’m looking forward to seeing if it meets my needs (and hoping that if it doesn’t, Xorg or Mir will be a viable option). Canonical wants to implement their own display server? Great – I’m looking forward to seeing if it meets my needs (and hoping that if it doesn’t, Xorg or Wayland will be a viable option).
It has been developed by the NSA to spy on you.
Mine will never support it.
http://www.templeos.org
Nobody needs Intel anyway.
This guy impersonated me.
Yeah, the fact that the comment was coherent & didn’t contain any paragraphs longer than 500 words was a bit of a giveaway.