Cinnamon 2.0 Desktop Is Readied For Release

[phoronix]

Phoronix: Cinnamon 2.0 Desktop Is Readied For Release

Linux Mint's Cinnamon 2.0 desktop fork of the GNOME Shell has been tagged and is being readied for release...

http://www.phoronix.com/vr.php?view=MTQ4MTQ

[uid313]

Faster?

I hope it is faster!

I tried out Cinnamon because it looked promising, but ended up quickly ditching it.
The problem is that it is too slow!

Bringing up the launcher menu takes several hundred milliseconds.

[Kostas]

Have they ported any of the numerous (not just limited to) performance improvements that Gnome-shell has had ever since they forked or is their codebase the mess it was?

[Yorgos]

Petra is πέτρα, which means rock

[Daktyl198]

Have they ported any of the numerous (not just limited to) performance improvements that Gnome-shell has had ever since they forked or is their codebase the mess it was?

**removed all that I said before**

post from Clem on segfault:

When we talk about Cinnamon being independent, our primary concern isn’t for it to be different, it’s for us to be able to port it to other distros and backport it to supported Mint releases without worrying about which version of GNOME is available. In other words, we don’t want Cinnamon to be its own for marketing reasons, but for technical and compatibility reasons. It has to be down to us whether Cinnamon works here and there, we don’t want to rely on third party projects.

That’s the reason we forked gjs for instance, to stop it from being different underneath us in every single release. In other components we go further than that and take the opportunity to improve things, we’re getting started here with cinnamon-control-center for instance and you’ve seen the work we put on Cinnamon itself and nemo of course.

If you don't get it, they pull in everything upstream and only fork the things that break every other release, or things they want to customize a bit (like the display settings module, where they've added a few design enhancements).

[Dukenukemx]

Petra is πέτρα, which means rock

Don't be such a μαλάκας.

[dee.]

Great news, I just hope they've fixed the memory leaks that were in 1.8.

[Luke]

To do that every release would be piles of work

Have they ported any of the numerous (not just limited to) performance improvements that Gnome-shell has had ever since they forked or is their codebase the mess it was?

If they were going to always follow upstream GNOME, there would be no reason to fork it. GNOME 3 is impossible to maintain extensions for, I know because before I found Cinnamon I used gnome-shell frippery, always had to pin gnome-shell when following alphas until I could port the extensions myself (sometimes possible) or they got updated.

I guess the way to port Cinnamon to new upstream code would be to treat the patches applied to the last used version of shell as though they were extensions, apply Cinnamon changes since then to them, then start the debugging process. If they want to release anywhere near with Ubuntu et all do, that just about requires using a previous distro version of shell. A good time for this sort of thing might be once only, when not only shell but Firefox and major GNOME apps are all Wayland compatable. I suspect it will be a lot of work.

As for the menu (really the main issue) , there are a lot of aftermarket menus for Cinnamon, some of which are much faster. I am using Cinnamon 1.9, which will become Cinnamon 2.0, and lately the menus have gotten a lot faster. Before, even on a Radeon 6750 on "low" power before kernel 3.10, the menu was slow. Now it mostly works fine, but the changes are on top of enabling DPM on that graphics card. Bug was said to be having to reload the icon list with every opening, and that I think was inherited from GNOME. Gnome-shell was actually worse about the default menu when run on my netbook, would hardly open those at all. Frippery's menus worked just fine on ANYTHING, which suggests that just porting the gnome-shell frippery menus to Cinnamon would make a better replacement menu than any now in existance.

Also avoid icon themes filled with .svg images. They make any version of Nautilus, Caja, or Nemo ever written bog down. I just rewrote my "ubuntustudio-legacy" icons using only presized png icons, based on current ubuntustudio icons after a libglib update broke my old ones that dated back to 2008. That made a HUGE difference in file manager responsiveness, as did using light icon themes free of .svg images when testing distros on Pentium III class machines.

[peppercats]

Cinnamon is a nice DE, only problem is that it's slow and a memory hog. Maybe they've fixed it.

[Luke]

I've yet to see any really fast desktop based on the "tablet model" or on Clutter

Cinnamon is a nice DE, only problem is that it's slow and a memory hog. Maybe they've fixed it.

Cinnamon is a fork of gnome-shell, gnome-shell is based on the Clutter toolkit. There are two DE's that tried to base themselves on the the "tablet" model: Unity and Gnome-shell. Both are slow compare to old school stacking window managers or pre-Unity Compiz, both are heavy. When Ubuntu was first working on Unity, both Mutter and the Clutter toolkit got the blame. Shell has a lot of Javascript, but if the use of a scripting language was the culprit I would expect Unity to blow away gnome-shell in a speed comparison, and it does not.

Cinnmon is pretty, but I only use it in my big video editing machines where a fat DE is trival and no DE is slow. It's still the best combination of gnome-shell's good features with those of GNOME 2 I have ever seen. On my netbook I use IceWM to save battery as much as anything else. Each has its place, I like and use both of them in their roles. If you need an end-user desktop that is fast I would recommend MATE over anything else. For a hacker IceWM is faster yet, but you have to know how to set it up.

The only way I ever see anything based on Shell or Unity being fast is if Wayland or Mir radically speed up the compositing/Xserver situation by getting X out of the way, and then all the applications get ported to both Wayland and to Mir (due to the "sectarian divide") to keep Xwayland/Xmir overhead out of the picture. If they are still slow, then Clutter is the reason. At that point, to fix the problem might require more people on Clutter's team, not driving them away with the sort of infighting I see so much of, not just here but elsewhere in life.

I look forward to seeing Cinnamon on Wayland, then we shall see...