Intel Reverts Plans, Will Not Support Ubuntu's XMir

[Guest]

How did you determine that? I have read claims that the L4 kernel is the most used kernel in the world. It is said to be present in basically all mobile phones (as the baseband processor kernel) and various other embedded devices.

For what it is worth, I have yet to find another monolithic kernel that is more widely used than Linux, but I do not have hard numbers on actual usage.

You're just dumb or trolling. For 99.99% of people it would be obvious what me and dee. meant. L4 doesn't match.

[LinuxGamer]

I am already aware of that, but I was keeping what I said short. Anyway, Canonical wants to go their own route. Redhat did this with systemd versus upstart. The advantage of this that they avoid the red tape involved in collaboration, which is likely the exact reason Redhat decided to go its own way on systemd. They could have improved upstart, but it would have taken much more time to do that in a way that was mutually agreeable to both Redhat and Canonical. Canonical likely feels the same way about X, Wayland, etcetera.

You do understand that systemd is not a Red Hat project?

[mrugiero]

Redhat did not want to work with Canonical on upstart in the same way Canonical does not want to work with Redhat on Wayland. They are both within their rights to refuse to work together. The fact that they do anything at all is more than I can say for many here.

While that's true, I need to point out that there is a more valid commercial reason for Red Hat to do so, since upstart uses Canonical's CLA, which makes it asymmetrically licensed. Wayland is MIT for everyone, period. upstart, Mir and most Canonical projects, are GPL for everyone except Canonical; for Canonical they are whatever they want them to be. So, commercially, you are giving your competition a tool, but you can't use that tool as much as them. That's a real reason not to contribute to their projects, on the business side.
And before you mention Red Hat's CLA, remember theirs only defaults code without an explicit license to MIT.

You're simplistic way of argumenting is pathetic. I did not claim that in every big company it is different. I said something about free software as well, which you ignored. Intel contributes to free software... hmm.. wtf did I mean?

No, let's not think, let's just say that you're right anyway! That's the spirit!

It's not simplistic, it's pointing out the contradiction on your own post. Intel devs aren't some hippie programmers that work for free, they are getting paid, and when you get paid by your time, the one paying gets to choose what you put your time on; you always have the option to quit if you don't like what they make you work in, and in the case of free software you can quit and do what you want on the code base, with a fork if needed. Might be ugly, but that's how it works. I don't like it very much either, but if they'd kept the patch, they'd get the responsibility of maintaining it. And Intel doesn't want to pay for that, as simple as it gets. And even more, then, you have a single distro support, you expect a developer you don't pay (because Intel clearly doesn't want to pay for it) to maintain it, then it's probably simpler to just make it out of tree, dependent on Canonical, since chances are a Canonical dev will maintain such a support, you are just making it easy for him, saving him the extra reviews on the driver mailing list. And if it turns out it's an independent dev maintaining it, it's still simpler, because bugs in machines using Mir will probably be in Mir's launchpad, so chances are a repo in Mir's launchpad gets you everything you need to maintain it in the same place.

[Teho]

The advantage of this that they avoid the red tape involved in collaboration, which is likely the exact reason Redhat decided to go its own way on systemd.

Then again systemd is more collaboratively developed and widely adopted than Upstart ever was so I find that argument bit hard to believe. I mean the original design was developed by Kay Sievers from Novell (at that time) and Lennart Poettering from Red Hat. To my understanding they started the project on their free time.

They could have improved upstart, but it would have taken much more time to do that in a way that was mutually agreeable to both Redhat and Canonical. Canonical likely feels the same way about X, Wayland, etcetera.

The thing is that systemd and Upstart are fundamentally different whereas Wayland and Mir are essentially the same. To my knowledge the only fundamental difference is between client (Wayland) and server (Mir) allocated buffers and even that could be done on Wayland (and has been done apparently). Red Hat did collaborate with Upstart upstream though. Canonical never did the same for Wayland. They _never_ brought up the issues they had with Wayland to its upstream and clearly brought up their lack of understanding of Wayland when they released the Mir spec sheet (that had quite a bit of misinformation about Wayland).

[blackout23]

I am already aware of that, but I was keeping what I said short. Anyway, Canonical wants to go their own route. Redhat did this with systemd versus upstart. The advantage of this that they avoid the red tape involved in collaboration, which is likely the exact reason Redhat decided to go its own way on systemd. They could have improved upstart, but it would have taken much more time to do that in a way that was mutually agreeable to both Redhat and Canonical. Canonical likely feels the same way about X, Wayland, etcetera.

Ehh you are aware that systemd is not a RedHat project? Even two Arch Linux devs work on it.

[nomadewolf]

I salute Intel for this!

[LinuxGamer]

I salute Intel for this!

we all do

[chrisb]

That would require Canonical to remove the contributor agreement. Obvious that would never happen. Not for Red Hat, not for Debian, not for anyone.

No, they could've just forked it.

[LinuxGamer]

No, they could've just forked it.

you don't know GPLv3 well do you

[mrugiero]

No, they could've just forked it.

And we'd be in the same situation as using systemd, we'd be using a different init system in Ubuntu than in other distributions. And that wouldn't be contributing upstream, which was the point.