Xfce's Xfwm4 Sees Wayland Port With Wlroots

[ThanosApostolou]

I wonder couldn't they use directly sway as a compositor instead of implementing a new one based on wlroots? Creating a new wayland compositor is a huge task which needs many resources even when a library like wlroots is used. As far as I know sway supports floating windows and windows borders instead of just tiling mode. So, wouldn't it be possible so that they make xfce session starting sway with its own config using floating windows by default and port the xfce "window manager settings" to have some simple options for changing this config? They wouldn't have the same flexibility and maybe some options would be missing but I believe they could contribute to sway directly so that they support any customization they want withing the sway config.

[aufkrawall]

Free RAM is there to be used, to test if a DE is capable of performing with lets say 1 GB of RAM you have to test on a machine with 1GB of RAM and not use 8 GB and see how much is left. From experience I can tell that XFCE is totally usable with 1GB, Gnome and KDE not so much.

RAM is to be used by filling it with data in a way that makes sense. It is not the task of a DE to do so. There is the shell etc. and some dependencies like Qt that make other parts of the DE snappy when they are cached, but that's it. This is just few hundreds MB at max (without bloat like Akonadi).

Flushing memory or looking at the startup memory consumption doesn't give any indication on real life usage of a DE.

Again, a DE is not a prefetcher like Superfetch.

[Anux]

Again, a DE is not a prefetcher like Superfetch.

Thats not what I wanted to say. Just that starting up a DE is of no use if you don't do anything useful with it. Therefor my example with the 1 GB machine. As soon as you run into your 1GB limit you will notice a heavy performance drop and that will happen a little later with XFCE and even later with smaller DEs.

[guglovich]

Given Xfce's limited resources, this is simply a project killer. The benefits are minimal, the time commitment is enormous. Xfce used to have the best UX, now KDE and Deepin are taking it over. They could simply switch to someone else's implementation at a later date.

[billyswong]

I'm glad to see xfce adopting Wayland, the last of the 3 "main desktop environments" (maybe they should change to "wfce" now )

But the best part is that they opted for wlroots (which aims to be a more generic compositor than Mutter/Kwin) and I hope other projects like Mate and Lxde can work together with them to improve wlroots into a better compositor, rather than spending their limited resources in a more private solution.

I wonder couldn't they use directly sway as a compositor instead of implementing a new one based on wlroots? Creating a new wayland compositor is a huge task which needs many resources even when a library like wlroots is used. As far as I know sway supports floating windows and windows borders instead of just tiling mode. So, wouldn't it be possible so that they make xfce session starting sway with its own config using floating windows by default and port the xfce "window manager settings" to have some simple options for changing this config? They wouldn't have the same flexibility and maybe some options would be missing but I believe they could contribute to sway directly so that they support any customization they want withing the sway config.

Personally I don't mind if they write their own wlroots-based compositor or use an existing one. As I am currently using Compiz, I am expecting when Wayland support of MATE is ready, I should be able to switch over to Wayfire. (Or do I expect too much?)

[Quackdoc]

There is a large lack of a good DE atop wlroots despite IMO wlroots being the most featureful thing (at least features I care about) to build atop off.

[ubuntuweston]

I thought adlocode, the developer who undertook this task, is not directly working with xfce but just hacking on his own? Meaning that we don't even know if upstream xfce will go the route he's taking?

[Ironmask]

Given Xfce's limited resources, this is simply a project killer. The benefits are minimal, the time commitment is enormous. Xfce used to have the best UX, now KDE and Deepin are taking it over. They could simply switch to someone else's implementation at a later date.

It's kind of a "damned if you do, damned if you don't" deal, since Xfce basically has no developers. It's a dead project. I love it, but like a lot of other people in this thread, I gave up on it, it simply doesn't have the resources to stay alive anymore. It's like the HURD, it's technically an active project, but by "active" they mean "keeping the lights on". I really respect the ongoing work with Xfce of course, I know the developers are trying really hard with it, but there just aren't enough of them. They need to adapt, or else they'll stagnate and die with Xorg, but they don't have the resources to adapt. It's not a project killer if the project was already dead.

[birdie]

If people tipped me I'd make a proper Linux based comparison between DEs RAM consumption in Linux. The guy who did it in FreeBSD, well, I'm not a fan of his methodology.

Speaking of XFCE: it's nearly dead. There's some activity here and there but not much else. This project is external to XFCE, basically run by a single person I guess.

[project_phelius]

Wasn't Xfce experimenting with a Mir-based compositor originally? I thought I remembered reading about that a few years ago.
Whatever happened to that effort and was there any conclusions reached, or was it just abandoned?
Has Wlroots surpassed Mir in the core functionality that matters most, or is it just Wlroots has more people and all the momentum, now.
Just curious.