The First Benchmarks Of Unity On XMir: There's A Performance Hit

[gQuigs]

* A fully open source license

BS. The GPL isn't open source now?

[timothyja]

BS. The GPL isn't open source now?

I think he is talking about contributors having to agree to pass the copyrights of their code to Canonical.

Edit: In other word its open source today but Canonical have the rights to make it not so. Although the same is true for Wayland with the MIT licence a company can fork it and not release the code so its much the same either way.

[chrisb]

Yes and Unity in 13.10 will run on XMir not Mir. So Unity will also be slower than just running on a regular X Server.

Indeed. That is a shame.

Yeah, which is why it makes no sense to run a whole DE on top of it. When you run a X-based DE on top of XMir, it gives no benefit whatsoever. You can't run Mir apps on it, you can't run Wayland apps on it.

Yes, obviously running Unity on XMir is not the long term plan. I presume this is being done to get Mir some wider usage before Unity/Mir becomes the default.

Prove it.

You want me to prove that Wayland is faster than Wayland+X11? It should be, it doesn't have to do all of the crufty X11 stuff.

[seb24]

I think he is talking about contributors having to agree to pass the copyrights of their code to Canonical.

Edit: In other word its open source today but Canonical have the rights to make it not so. Although the same is true for Wayland with the MIT licence a company can fork it and not release the code so its much the same either way.

Is not exactly that. For Mir Canonical have the right to "sublicence" the work not to change the main licence. It means that the Mir project will stay in GPLv3 but can have some derivative licence.

[Teho]

Although the same is true for Wayland with the MIT licence a company can fork it and not release the code so its much the same either way.

The difference is that in case of Mir Canonical has rights to the code that others don't while everyone has the same rights to Wayland source code. Also the GPLv3 license is bit problematic from various projects; it's for example too restrictive for various embedded setups and possibly phone OEMs. This means that Canonical can still sell Mir under less restrictive or propietary license for OEMs but no one else can. Therefore it's possible that Mir can't be used by anyother company but Canonical without a license from Canonical on phones. Mir is therefore not an option for other GNU/Linux phone OSes like Sailfish and Tizen; both of which btw plan to use Wayland in the future.

Also projects like KWin have no intrests in GPLv3 license and want to remain GPLv2+ so they can't really use code from Mir at all.

For Mir Canonical have the right to "sublicence" the work not to change the main licence. It means that the Mir project will stay in GPLv3 but can have some derivative licence.

To give a simple example, suppose a company launches a new GPL-licensed project and asks contributors to sign a Harmony copyright assignment agreement with the “only OSI-approved licenses” outbound option selected. The company is then entirely free to license out all contributions under, say, the (OSI-approved) 3-clause BSD license, which in turn does nothing to restrict the company from privately licensing the project code, including contributions, under a proprietary, closed-source license. This is not some novel scenario, but what Harmony adds is the illusion of constraint, which I am concerned may mislead contributors who are relatively unfamiliar with open source licensing.

--Source, so yes they can make it proprietary and stop releasing code under GPLv3-license.

[seb24]

--Source, so yes they can make it proprietary and stop releasing code under GPLv3-license.

As I understand no (in the case of the Canonical CLA) :

As a condition on the exercise of this right, We agree to also
license the Contribution under the terms of the license or
licenses which We are using for the Material on the
Submission Date.

I understand that can do some sub-licence but always licence in the original licence too.

Source

[ravas]

I'm personally not surprised by these results at all - Canonical got a head start since they decided to base this off of Android's code, to some degree anyway. As said before, a compatibility layer will always result in poorer performance, HOWEVER, that means worse performance compared to native. So for example, a game compiled to work for Wayland will perform better than the X version in Xwayland. But, Xwayland could still potentially perform better than X11. But considering how young Mir is, I'm not surprised it performed worse, and its performance loss is far from "what a shame, just kill it".

At this point I'm finding it a bit tough to figure out which display server will end up being the best replacement to X:
Pros of Mir over Wayland:
* MUCH faster development
* Supposed to get Android driver support
* A seemingly more devoted team

Pros way Wayland over Mir:
* Targets all DEs in mind
* Seems to be more thought-out in a technical standpoint
* Better multi-seat support
* A fully open source license
* Probably will be more light-weight in the end

Actually, all of the pros listed for Mir belong on the Wayland side.
XMir only exists due to years of Wayland development (it is XWayland renamed).
The Android driver support exists due to Wayland development (libhybris).
Wayland has a lot more involvement from the general ecosystem.

Basically, Canonical came in after years of the Wayland developers doing all the grunt work, put the pieces together with new names and let everybody think they had done something amazingly quickly.

[jrch2k8]

1. first surprise for the ubutrolls wayland protocol is 95% ready and the only thing left is tackling the corner cases and has been ready to go since december 2012.

2. second surprise we are not waiting for wayland at all this days, we are waiting for the toolkit that started to migrate full stem ahead, as far as i can tell SDL2.0/Qt5.1/gtk3.10/EFL git

3. third surprise Video support is here vdpau/vaapi/mplayer/gstreamer/HW overlays [ofc most code is still in git]

4. fourth surprise Gnome git can actually run on wayland except GDM [still have few porting glitches if you don't use gtk+ git]

5. Qt5.1 apps like Qtcreator/qtdesigner run on wayland already and QML/C++ classes QCompositor already works pretty impressively

6. Mir was developed faster? just amuse me and check of the LOC and refactor needed in the entire graphic stack to make something like Mir/wayland even technically feasibly?[hint it ways passes the couple of millons of LOC in Glamor/cairo/mesa/kernel code/DDX/dri2-3/Drivers/etc] and this technical miracle that required to rewrite basically every part of the graphic stack almost from scratch for years has 0 Canonical contributions.

So yes mir got coded faster because wayland + community did all the massive heavy lifting while canonical waited[without 1 freaking commit] until it was good enough for them to start and in some cases take solutions from wayland code[read their bazaar and wayland git and you will see some funny things in there]

So wayland being more community don't have cool demos[we are geeks after all, we don't demo we port] but the next release of every major desktop environment[except unity] will support wayland natively, not demos or layers just native[EFL will come first and then Gnome and last KDE SC 5]

[Teho]

As I understand no (in the case of the Canonical CLA) :

I understand that can do some sub-licence but always licence in the original licence too.

Source

The quote was talking exactly about that clause. In practise it doesn't prevent proprietary only licensing.

[chrisb]

Also the GPLv3 license is bit problematic from various projects; it's for example too restrictive for various embedded setups and possibly phone OEMs. This means that Canonical can still sell Mir under less restrictive or propietary license for OEMs but no one else can.

Changing the license of Mir won't eliminate GPLv3 software from Ubuntu Touch.