Also @x0x7 this project sure is interesting but I've been thinking: at this point I might as well just switch to XWayland. There's a good variety of compositors, they actually look good and implement good ideas, and it's more secure since apps are isolated so it doesn't allow for keyloggers. What do you think?
I've already switched to it. I wonder if steam will end up picking up XLibre. SteamOS is already built on Arch which supports XLibre through the AUR. And it's only a quick hope from Arch to Artix considering a hybrid Arch/Artix system is as easy as adding both repos to the pacman.conf.
The reason why XLibre could make a lot of sense is that it will have the broadest compatibility with applications and all of the efficiency gains of Wayland go away when you consider that XLibre can do the same things and many of those direct buffer edits have already been made to XLibre.
Imagine if not only Steam switched to XLibre but Valve contributed to the development.
This goes beyond me but for some reason the developer of KiCad says he needs proper X11 to do everything he's doing. Why, I don't know. But X11 has a lot of parts. There is a lot to cross implement or bridge to. Part of the idea of Wayland is to start fresh without past bloat so if they are really going to implement all parts or bridge to every nook and cranny of it then one could argue why not just stick with X11?
As both try to get full coverage of each others features and performance, and XLibre aims to do that, it almost doesn't matter.
So while it doesn't matter I'm going to stick with XLibre just to go against the grain of IBM trying to kill XOrg. I think both should exist. Open source is benefited by their being options and not throwing out good projects and allowing contribution to everything that can be improved.
That's a theme of programming. Don't do a rewrite when improvements of an existing thing can cover it. To the extent Xorg has stagnated and created demand for Wayland some of that has been engineered.
People aren't changing to a new kernel project every year. Instead we improve Linux. But that doesn't mean we actively try to kill BSD. This isn't how open source is supposed to work. Also the fact that LLVM and clang exist doesn't mean we kill GCC. But IBM takes over management of open source and then tries to kill things. I say fuck that. I run Artix. It's fuck you Linux. Do all the things that controlling corporations don't want you to do with your Linux.
Yeah good point. I want to stay away from big corporations that take over Linux as well. My distro of choice is Slackware which doesn't have systemd. But the advantage of Wayland + XWayland would be to at least have better security in programs that have been ported over. I'm not opposed to all change. I miss Firefox with XUL addons for example but I will admit that the old Firefox simply wasn't robust. Do you remember it freezing all the time? The new Firefox is still better (assuming you're using a privacy-friendly fork). Other examples are Appimages and Flatpaks for cross-distro program packaging (and in the former case portability). I have a similar feeling about Wayland that the reason for its development is valid. Apparently I had already used a version of Gnome that's ported to wayland so I've already tested it out and it works fine, although I admit Gnome seemed to lack a lot of features and now I wonder if that's because of wayland.
See Gnome is a good example of why a rewrite or replacing with something new isn't always good. The temptation as a developer, and its really hard to estimate which one is correct on a case to case, is that the core of the software you wrote isn't that complicated and you are more experienced so if there is a major organization gain for starting over implementing all the features won't be that hard.
The practical reality of rewrites is the years of putting in features into something wasn't entirely waste and when you swap to developing a replacement 90% of these features aren't coming over. This how you get gnome that is always re-inventing and the latest and greatest but feels more basic than even XFCE.
So when tempted to overhaul everything it's often better to do some major overhauling of a few targeted sub-components than it is to throw out almost everything. Going even further back the throwing out of Gnome 2 for Gnome 3 still pisses me off. They killed it because they knew Gnome 3 was shitburgers, and no one would use it while Gnome 2 existed.
I'm moving more toward using the latest and greatest as a consumer but that doesn't mean killing old things. Developers shouldn't be in the business of destroying things. If you have to kill something good and effective to get people to use your new thing it just means your new thing isn't good enough yet. So the opposite side of things would be Python 3 vs Python 2. Everyone should have moved a lot faster to it because Python 3 is just entirely an improvement and 90% similar. They didn't throw out features. And despite that Ubuntu kept people stuck on hybrid python 2 and 3 still to this day.
From what I understand XLibre is taking the Python 3 like approach. Something sucks in X11 and isn't fast... Ok, let's fix that.
Log in
@JasonCarswell
Also @x0x7 this project sure is interesting but I've been thinking: at this point I might as well just switch to XWayland. There's a good variety of compositors, they actually look good and implement good ideas, and it's more secure since apps are isolated so it doesn't allow for keyloggers. What do you think?
I've already switched to it. I wonder if steam will end up picking up XLibre. SteamOS is already built on Arch which supports XLibre through the AUR. And it's only a quick hope from Arch to Artix considering a hybrid Arch/Artix system is as easy as adding both repos to the pacman.conf.
The reason why XLibre could make a lot of sense is that it will have the broadest compatibility with applications and all of the efficiency gains of Wayland go away when you consider that XLibre can do the same things and many of those direct buffer edits have already been made to XLibre.
Imagine if not only Steam switched to XLibre but Valve contributed to the development.
But muh security.
What about XWayland does that not work?
This goes beyond me but for some reason the developer of KiCad says he needs proper X11 to do everything he's doing. Why, I don't know. But X11 has a lot of parts. There is a lot to cross implement or bridge to. Part of the idea of Wayland is to start fresh without past bloat so if they are really going to implement all parts or bridge to every nook and cranny of it then one could argue why not just stick with X11?
As both try to get full coverage of each others features and performance, and XLibre aims to do that, it almost doesn't matter.
So while it doesn't matter I'm going to stick with XLibre just to go against the grain of IBM trying to kill XOrg. I think both should exist. Open source is benefited by their being options and not throwing out good projects and allowing contribution to everything that can be improved.
That's a theme of programming. Don't do a rewrite when improvements of an existing thing can cover it. To the extent Xorg has stagnated and created demand for Wayland some of that has been engineered.
People aren't changing to a new kernel project every year. Instead we improve Linux. But that doesn't mean we actively try to kill BSD. This isn't how open source is supposed to work. Also the fact that LLVM and clang exist doesn't mean we kill GCC. But IBM takes over management of open source and then tries to kill things. I say fuck that. I run Artix. It's fuck you Linux. Do all the things that controlling corporations don't want you to do with your Linux.
Yeah good point. I want to stay away from big corporations that take over Linux as well. My distro of choice is Slackware which doesn't have systemd. But the advantage of Wayland + XWayland would be to at least have better security in programs that have been ported over. I'm not opposed to all change. I miss Firefox with XUL addons for example but I will admit that the old Firefox simply wasn't robust. Do you remember it freezing all the time? The new Firefox is still better (assuming you're using a privacy-friendly fork). Other examples are Appimages and Flatpaks for cross-distro program packaging (and in the former case portability). I have a similar feeling about Wayland that the reason for its development is valid. Apparently I had already used a version of Gnome that's ported to wayland so I've already tested it out and it works fine, although I admit Gnome seemed to lack a lot of features and now I wonder if that's because of wayland.
See Gnome is a good example of why a rewrite or replacing with something new isn't always good. The temptation as a developer, and its really hard to estimate which one is correct on a case to case, is that the core of the software you wrote isn't that complicated and you are more experienced so if there is a major organization gain for starting over implementing all the features won't be that hard.
The practical reality of rewrites is the years of putting in features into something wasn't entirely waste and when you swap to developing a replacement 90% of these features aren't coming over. This how you get gnome that is always re-inventing and the latest and greatest but feels more basic than even XFCE.
So when tempted to overhaul everything it's often better to do some major overhauling of a few targeted sub-components than it is to throw out almost everything. Going even further back the throwing out of Gnome 2 for Gnome 3 still pisses me off. They killed it because they knew Gnome 3 was shitburgers, and no one would use it while Gnome 2 existed.
I'm moving more toward using the latest and greatest as a consumer but that doesn't mean killing old things. Developers shouldn't be in the business of destroying things. If you have to kill something good and effective to get people to use your new thing it just means your new thing isn't good enough yet. So the opposite side of things would be Python 3 vs Python 2. Everyone should have moved a lot faster to it because Python 3 is just entirely an improvement and 90% similar. They didn't throw out features. And despite that Ubuntu kept people stuck on hybrid python 2 and 3 still to this day.
From what I understand XLibre is taking the Python 3 like approach. Something sucks in X11 and isn't fast... Ok, let's fix that.