Wayland 1.3 Release Candidates Are Now Out

[phoronix]

Phoronix: Wayland 1.3 Release Candidates Are Now Out

Kristian H?gsberg has put out the first test releases of the forthcoming Wayland 1.3 release and reference Weston compositor...

http://www.phoronix.com/vr.php?view=MTQ2NzA

[mark45]

It's a bugfix release and since the features are internal or weston-related, it doesn't deserve a new point release.

[mark45]

The features are highly Gnome related, so it does need a release. Gnome 3.10 is here in a few days and Fedora needs to release Wayland supported packages soon. So sure this is needed when the primary Wayland desktop and primary Wayland distribution ask for it.

Which is what I'm saying, it should be like 1.2.2 instead of 1.3.
1.2.1 included some internal features and lots of bugfixes, like this release.

But it doesn't really matter, I'm just picky I guess.

[dee.]

They need to look like they're making progress. They themselves realized how serious and fast Canonical is about the development of Mir so they have to look like they're keeping up.

Does anyone bite your pathetic trolling attempts anymore?

[TheBlackCat]

Which is what I'm saying, it should be like 1.2.2 instead of 1.3.
1.2.1 included some internal features and lots of bugfixes, like this release.

From the announcements, it is pretty clear why they are using 1.3 vs. 1.2.3 (emphasis added):

1.2.1:

This release is mainly bug fixes and documentation tweaks, but a lotof them. We also have the touch support for toytoolkit and other clients. This is probably more of a feature, but it barely affects existing code paths or change weston internal API, so it's a safe addition.

1.3:

I believe I picked up most recent patches and fixes recently on the list. In particular, I merged Rob and Neils patch to send out events to all listeners for wl_pointer, wl_keyboard and wl_touch from a client. This patch enables clutter-gtk and webkitgtk, but touches core event delivery code paths, so look for regressions there.

So, as with many projects, only non-disruptive changes are allowed in x.x.x releases. If the change has a significant potential for regressions or changes to users, it is kept for x.x releases.

[r_a_trip]

They need to look like they're making progress. They themselves realized how serious and fast Canonical is about the development of Mir so they have to look like they're keeping up.

I'll bite. It doesn't matter how fast Canonical moves with Mir or how slow Wayland progresses. Despite your boundless hope that the world will unite on Mir, this is not going to happen.

The way that Canonical introduced Mir, with a big tada and a lot of technical falsehoods about Wayland, stepped on a lot of toes. That Mir was conceived in total darkness as a secret skunkworks project, kept under wraps for 9 months, also didn't endear any outsider to Mir.

Mir is asymmetrically licensed under the GPLv3 + CLA. This means that no-one but Canonical can license a Mir implementation under less free conditions. For mobile, with the specific requirements of that industry in place, Mir under the GPLv3 is useless to any distributor but Canonical. Wayland in contrast is licensed under MIT and this means everybody is free to use it as free software or to close up the code. This is completely symmetrical. No need to beg ($$) Canonical for a license exception. This alone reduces the chance of third party Mir uptake dramatically.

You may think that Canonical is big enough to bend the entire Linux community to their converged vision and on Canonical's terms, but they are not. They are a relatively small company, that has captured a smidge of the non-technical minded computing market (in comparison to the whole non-technical computing market). A smidge that is not really paying the bills to boot.

What you think is a titanic battle of wills between mighty Canonical and the rest of the puny Linux developers, is in fact a technical rift between tiny Ubuntu and the rest of the massive Linux ecosphere. Ubuntu is breaking off of the larger Linux ice shelf. So you will have a whole family of diverse Operating Systems, powered by Linux, GNU, Wayland and systemd and there will be Ubuntu, powered by Linux, GNU, Mir and upstart. Where do you think the synergies will happen?

Ubuntu is becoming its own thing. If that is bad or good, I'll leave to interested stakeholders. I no longer have any skin in that game. It will probably mean that Ubuntu will keep diverging from what is traditionally understood as a Linux distribution. Since upstream support for Ubuntu is becoming strained, I wouldn't be surprised if Canonical opts to write its own development environment, unique to Ubuntu. This will mean yet another FOSS platform, next to Linux, *BSD, Haiku, ReactOS, etc. but if running that makes you happy, it makes me happy that you have found your thing.

[dh04000]

Mir is asymmetrically licensed under the GPLv3 + CLA. This means that no-one but Canonical can license a Mir implementation under less free conditions.

Wait, wait, wait.... are you telling me that the opensource community is mad that a piece of OSS can't just be forked and released under a less free license? Being able to just show up, take someone else's hard work, fork it and license it under a less free title and develop it outside of the GPL is now a GOOD THING in the open source community? So what Oracle does is a good thing now?

[dee.]

I'll bite. It doesn't matter how fast Canonical moves with Mir or how slow Wayland progresses. Despite your boundless hope that the world will unite on Mir, this is not going to happen.

The way that Canonical introduced Mir, with a big tada and a lot of technical falsehoods about Wayland, stepped on a lot of toes. That Mir was conceived in total darkness as a secret skunkworks project, kept under wraps for 9 months, also didn't endear any outsider to Mir.

Mir is asymmetrically licensed under the GPLv3 + CLA. This means that no-one but Canonical can license a Mir implementation under less free conditions. For mobile, with the specific requirements of that industry in place, Mir under the GPLv3 is useless to any distributor but Canonical. Wayland in contrast is licensed under MIT and this means everybody is free to use it as free software or to close up the code. This is completely symmetrical. No need to beg ($$) Canonical for a license exception. This alone reduces the chance of third party Mir uptake dramatically.

The license is a red herring. It doesn't really matter that much, GPLv3 license is fine, and if Canonical wants to make money with a dual-licensing scheme, that's their right and entirely acceptable.

What matters more is that Mir is developed in such a way that no other DE can even consider it feasible to support it. Mir is being developed for Unity only, it's tailored for Canonical's needs and not anyone else's, Canonical makes no promises to maintain any kind of compatibility with any alternative implementations, or even compatibility within their own releases - they've explicitly stated that they will not maintain a stable server-side API. This is not a software that is suitable for various needs and use cases. Especially, when there is already Wayland, which is made to run everywhere and suit everyone's needs. So it really makes no sense for anyone other than Canonical to support Mir. There's no benefit in it to anyone else.

When Intel starts selling their Tizen-based ultrabooks, they're going to increase the Linux marketshare more than Ubuntu ever did. Intel has enough power and resources to bring Linux to the reach of average consumers, in regular retail stores, right next to win8 laptops. People can compare and contrast, see how much better Tizen runs (with Wayland) when compared to win8.

So Canonical is really making a bad mistake with Mir. They're fighting a battle they have no chance of winning, and it will end badly for them unless they repent. I hope it doesn't come to that.

[dh04000]

He said the asymmetry is unfair, that's it. The right level of freedom or restrictions may vary for different parts of the software stack, but asymmetry is always wrong.

Wayland is symmetric, MIR is asymmetric. MIR is always the wrong choice for every one but Canonical.

I wasn't aware GPL was ever unfair. Everyone is free to use any GPL code as free GPL code. Multi-developer licensed GPL can have issues. Ask VLC how they had to rewrite perfectly good code due to a number of developers not responding or simply refusing change the license on their code when the main project decided too (I seem to remember reading ~30% of VLC code had thier devs refuse to change license and had to have the code rewritten, but I don't remember of that's true or just my bad memory) . Pretty shitty situation that is smart to avoided by CLA, that way one asshole with an axe to grind can't kill/damage an entire project.

link, but the whole story isn't found in this one link: http://www.videolan.org/press/lgpl-libvlc.html

[dee.]

Even if MIR was any good, do you really believe the rest of the industry would sign yet another CLA just to be a part of MIR?

Sure. People sign CLA's to contribute to lots of things. CLA's don't really matter all that much.