I've been using Xfce for like last 25 years. I'm immensely thankful to these folks for creating and maintaining a DE that gets the job done, allows ample configuration, is stable, and otherwise stays out of your way. Incessant improvement without dramatic changes is a rare and valuable skill.
Wayland support is huge news. Of course, it's still limited, and Wayland itself is still limited in many ways. But this appears to be the road forward, and Xfce is taking it, showing how it's future-proof. I like how it was done by factoring out things related to the Wayland support into a library, while maintaining most of the rest. This hopefully shows that the code is in a good shape, well-structured.
I really like the XFCE implementation on Kali Linux because the desktop menu(?) includes a search bar. When I install XFCE on other distros like Ubuntu and Arch, this feature is missing out of the box. Is this a standard feature and/or is there a standard method to add it to other Linux distros?
Their own terminal emulator or they poke around the system and use the first one they find. Ideally this happens through the dbus so you can't trace it with strace -f.
Wayland still intrigues me - it was first released 16 years ago but still support like in XFCE is "experimental". I really don't know much about Wayland but its initial goals and the excitement if caused (at first) seemed always very noble.
There's definitely a lesson to be learned from Wayland's rollout. Technically it's a vastly superior approach to system graphics. The major roadblock to its adoption has been that it didn't deal with all the "extras" that X provided to desktop distributions as well. Things like copy/paste, screen sharinf, and others don't have a single standard way of accomplishing through Wayland because those aren't things related to Waylands core goals. As a result, it's been a no man's land of distributions waiting to see if someone else will build a standard for those things they like, or building a very opinionated standard that no one else likes.
Future projects wanting to overturn crusty old incumbents by supercharging the worst part of that incumbents' process should at least provide a side project where all the extra bits that maybe people liked can grow a replacement, even though it's not the focus.
> provide a side project where all the extra bits that maybe people liked can grow a replacement
The whole point of freedesktop.org was to provide such a side project in the first place. The fact that it doesn't seem to be doing this with any effectiveness at present is the real issue.
the decision to roll window management and keyboard management and screen management into one thing called "compositor" looks very cavalier to me. It resulted in a need to duplicate so many things that are modular and pluggable under X.
(Repeating myself, I wish somebody created a Wayland compositor that explicitly delegates most tasks to plugins, providing a stable interface, so that you could use your own WM, your own keyboard switcher, etc. Thankfully, screenshooting is using a similar approach already, but in a limited way. The idea is so obvious that somebody must have attempted this already.)
> The idea is so obvious that somebody must have attempted this already
You'd think that, wouldn't you?
The closest thing I know to a universal Wayland server that's WM and desktop independent is Mir.
I blame the disappearance of the one concerted competitive rival effort to the primarily Red Hat-backed complex of Wayland+GNOME+Flatpak to a single HN thread.
In my experience, HN commenters are often surprised when I say that I find this place to frequently be toxic, negative, with a strong element of groupthink. It's all on display in this thread:
If you ignore things like randomly running into screen sharing problems or issues configuring XYZ thing because Gnome or whoever can't be arsed to expose it in their Wayland implementation. I keep trying Wayland and then swapping back to X11. Less issues and I never notice the things Wayland "solves", just what it breaks.
Indeed. I kept hearing "Wayland solves problems with two monitors at different refresh rates" But when I set up mismatching monitors, X11 didn't skip a beat (1440p native/120Hz, 1440p (being fed to a 4k native monitor)/60Hz
Conversely, the worst issue I had with dual monitors was with Wayland: Wayfire thought it was a brilliant idea to not support "window spanning two monitors" as default policy.
It feels to me like Wayland, GNOME and Systemd all come from the same sort of developer brain infection making them build vast, all-inclusive structures that are explicitly hostile to anyone who doesn't have the exact same underlying assumptions and needs.
Yeah, I have the same experiences with Wayland. In fact, whenever I see it being mentioned and that it may be the default, I tend to just try to avoid the program in which it is default because I've always had issues with it. X11 works. These days are long gone where you have to modify X11 config as well, just run "nvidia-xconfig" or whatever. I do not remember the last time I had to mess around with X11. Wayland less so, as many essential features were not working. X11 has more than one clipboards as well, and I have no clue if Wayland has them, too, and we are talking about something very basic.
The protocol might be good, but my hardware is not supported by wlroots. Whether this is the fault of upstream or the hardware vendor does not matter. I cannot run wlroots, ever. Period.
It does matter, because one post before you felt entitled to put the blame on the protocol. Nuance is pretty important when you take into account the difference in amount of resources available to the two groups.
Maybe I'm misunderstanding you when you argue against desktop environments and wlroots compositors having "solid" wayland implementations due to not working under nvidia cards. If that's the case I might argue that you could present your point in a less terse manner.
...And just the other day I was thinking about maybe trying XFCE out :)
Now, currently I'm using LXQt (which is probably even more bare bones and "get the job done" DE), what can XFCE give me? I mean, I certainly see a point when people switch to XFCE from Gnome or KDE. But what about if coming from the other side of the spectrum?
Yeah, I understand that it sort of sits in between. But what are those "additional" features compared to a more spartan DEs? More included apps / accessories? More access to the machine's properties (like, more GUIs for various options)? Something else?
Good for you. If you're happy with it, that's great.
The thing is this: the controls on a vertical taskbar should be horizontal, and they should rearrange themselves in rows according to its width. So, horizontal buttons with horizontal text for each window, and status icons in columns and rows.
Any and all desktops with a start menu and a taskbar are just variations of Windows 95, and this is how Windows 95 worked. Most of the ripoffs don't do it right, and can't. A few (e.g. KDE) do it but really badly, so you get a HUGE start button and a HUGE clock, for instance.
Most people don't remember Win95 now, but I do. It really annoys me when the knock-offs can't do things the original did 29 years ago.
Can you imagine if someone came along with a new version of Vim, but they'd changed all the keystrokes and replaced all the : commands with English verbs? Instead of :wq or :x you typed SAVE FILE? People would laugh mockingly and go use something else.
That's how I feel about a half-done version of the taskbar-and-start-menu interface, like in MATE or Cinnamon.
Original Windows 95 fit into 32MB of disk space because that was the original maximum size of FAT16 drives. It was a tiny, simple OS. If someone presents me with a UI that copies the Win95 one, and basically every other FOSS desktop except GNOME >=3, Unity and Pantheon does, then the least I expect is an accurate copy of the very simple functionality that Windows 95 provided in 4MB of RAM.
Take that away, and it's ruined. It's Vim with no command mode and only cursor keys: you might as well use Notepad.
Notepad is perfectly usable and it's all a lot of people need. I'm not mocking anyone happy with it. I expect a bit more though.
Wayland support is huge news. Of course, it's still limited, and Wayland itself is still limited in many ways. But this appears to be the road forward, and Xfce is taking it, showing how it's future-proof. I like how it was done by factoring out things related to the Wayland support into a library, while maintaining most of the rest. This hopefully shows that the code is in a good shape, well-structured.
Congratulations with the release!