Announcement

Collapse
No announcement yet.

X.Org Server 21.1 Will Aim To Release In The Coming Weeks

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #21
    Originally posted by Sonadow View Post
    Right, screw this. I've had enough.

    At this rate desktop Linux will never ever fully transition to Wayland and developers will simply continue to flog the zombie that is Xorg and the Xserver. Things like VR, HUDs and even newer exciting stuff will simply get grafted onto Xorg like the shitbag that it is. Let's continue to deal with having a kernel driver, a drm driver and an Xorg ddx driver just to draw the desktop instead of having everything done by the drm driver. And while we're at it, let's also continue to enjoy the clusterfuck of device-specific input drivers compared to have everything punted off to a general Wayland protocol extension.

    Everyone might as well simply roll back all the efforts to create Wayland compositors and software that actually function properly on Wayland because the Xorg maintainers don't even have the will to stick with their plans of permanently abandoning the Xserver once and for all even after 13 years of Wayland development. If things screen sharing and remote desktop access don't work properly on Wayland, don't bother with wasting effort to address them because "switch back to an Xorg session" is going to be the easy way out.
    My workflow depends on Synergy/Barrier to share Mouse/Keyboard and Clipboard between two computers.

    Unfortunately, Synergy & Barrier do not support Wayland yet. [1] & [2]

    There is some exploratory preliminary efforts to support Wayland made by a barrier dev, but according to him, it's a lot of work and "each wayland compositor will need a separate implementation to support it" [3]. It will be necessary to add a handful of APIs and boilerplate to Wayland to make it possible, and it is not a one person Job.

    In the meantime, all Synergy/Barrier users are effectively stuck on X.org
    We still want to play in VR, and the latests desktop features, because why not?

    As is, Wayland is just not an option for me (yet), it is 0% usable.
    And I am very glad that someone is picking up the slack to finally make X releases again until Wayland is able to support my workflow.

    [1] https://github.com/symless/synergy-core/issues/4090
    [2] https://github.com/debauchee/barrier/issues/109
    [3] https://github.com/debauchee/barrier...ment-818691872
    Last edited by sheepdestroyer; 10 August 2021, 07:38 AM.

    Comment


    • #22
      Originally posted by Sonadow View Post

      The reason desktop Linux is full of technologies and features that are perpetually incomplete simply boils down to the fact that nobody has any desire to finish them as long as there is an easy fallsafe option.

      Wayland doesn't work well enough? Fall back to Xorg sessions.
      Flatpaks and Snaps don't work? Fall back to Appimage.

      And heck, you got shit tons of people out there making no shame of boasting that they will always recommend the fallbacks to users, or refusing to use the new stuff in favour of the older stuff. As long as nobody pulls the rug out on these easy fallbacks, nothing ever gets completed because everybody will just insist on using those fallbacks and the teething problems on the their replacements will never get addressed due to lack of interest.

      Who the hell will ever have any incentive to finish what was started under such situations?

      If the developers think they cannot ever drop Xorg and the Xserver, then kill Wayland right now and continue to focus on developing and improving Wayland. Otherwise, sever all development of Xorg and Xserver, and focus strictly on the parts that matter like XWayland.

      Make up your damn mind. Xorg or Wayland?
      The key bit here is not if someone releases a new XOrg version but when key DEs and distributions depreciate Xorg.
      Usually, you'd expect Fedora and/or Gnome to take the lead but I'm not sure if RedHat is ready to take this step.

      For a uniform DE, you'd need to go even one step further and depreciate XWayland or at least announce depreciation, e.g. for some GTK and/or Gnome version. However, I cannot even imagine the amount of cry and tears that will follow on this forum.

      Anyhow this is the main weakness of open-source on the desktop (among many strengths). Open source tends to go a path of least resistance and often this means heavy compromises and limited uniformity - there are tons of examples.

      Regarding readiness, Wayland is largely ready and improving fast. The obvious points to improve are
      - Games/steam etc
      - Remote desktop: this is a weakness of the Linux desktop since forever. Gnome remote desktop is more or less limited to screen sharing as opposed to proper remote login. Also, Gnome is now looking into replacing VNC with RDP (or at least make RDP a first class citizen). While VNC needs to be replaced, adopting a MS controlled protocol is probably problematic long term. Of course, adopting SPICE (the open source alternative) would require even more developer time..

      Comment


      • #23
        Originally posted by mppix View Post
        The key bit here is not if someone releases a new XOrg version but when key DEs and distributions depreciate Xorg.
        Usually, you'd expect Fedora and/or Gnome to take the lead but I'm not sure if RedHat is ready to take this step.
        Deprecating Xorg could never even begin to happen as long as Nvidia stubbornly refused to properly support Wayland, because like it or not, Nvidia still accounts for the majority (or at least a great many) of Linux users out there. Now that apparently Nvidia has finally seen reason, it's only a matter of time.

        For a uniform DE, you'd need to go even one step further and depreciate XWayland or at least announce depreciation, e.g. for some GTK and/or Gnome version. However, I cannot even imagine the amount of cry and tears that will follow on this forum.
        Umm, nope. The whole point of creating Xwayland is for Wayland to maintain backwards compatibility with current X apps, which is what will ultimately (after Wayland itself is ready of course) render bare-metal X unnecessary in the mid/long term and will allow us to retire it for good. Xwayland is a core part of the Wayland stack and is fundamental for the Linux desktop to transition to Wayland; asking to deprecate it as a way to help Wayland is like asking to deprecate Wine/Proton as a way to help Linux gaming.

        Comment


        • #24
          Kudos to that person who voluntary stepped up.

          Comment


          • #25
            Originally posted by JackLilhammers View Post

            That's just plain wrong.
            Adoption should not be forced, especially on a niche platform like Linux!

            The best (and only sane) way to encourage the switch to Wayland is to make it better for the end users.
            The display server, compositor, whatever, must be transparent to the user.

            Once the user won't be able to tell the difference between Xorg and Wayland, everybody will use it
            I don't believe it's even practically possible since Wayland breaks so many things Xorg-server provided out of the box.

            For one, I believe Wayland is a complete failure, a concept which should be buried. The fact that under Wayland each DE must reimplement a display manager is simply outrageous. Wayland will continue to be a fringe graphical environment unless Wayland developers roll out a universal display manager (akin to Xorg-server) on top of which you can code and build lightweight window managers in less than 1K lines of code without reimplementing a ton of basic features like drag-n-drop, window stacking and moving, global shortcuts, clipboard management, systray, desktop recording and remote desktop features, and even graphical environment configuration file, etc. etc. etc. No one in the Wayland community bats an eye that when a Wayland DM crashes, everything crashes. Wayland DM cannot be replaced or restarted on the fly.

            Wayland in its current form SUCKS HARD. Unix has always been about modularity, yet Wayland insists on duplicating a TON of effort.
            Last edited by birdie; 10 August 2021, 07:56 AM.

            Comment


            • #26
              Originally posted by user1 View Post
              But I just hate being forced to use incomplete software.
              This is the exact wrong way to look at this. In the case of something like Wayland, what is "done" supposed to mean? By nature of the fact that this topic is about an X11 release, we could be confused as to why the X Window System still isn't "done" after 34 years but we don't. We could also say "Why isn't Linux done yet?" but we know that some software, by the nature of what it does, is never "done".

              Originally posted by user1 View Post
              Let's face it, even though Wayland has progressed tremendously in the last year or two, it still hasn't reached feature parity to X.
              It's not supposed to. One of the reasons that Wayland was created the way it is is because X.org is bloated with features that are no longer really used but are difficult to strip out.

              One "feature" that people attribute to X.org, screen and window capture, is just a security flaw that's talked about like a feature. Sure, window and screen capture is desirable and isn't part of Wayland but it doesn't need to be. That's been proven by Pipewire.

              Originally posted by user1 View Post
              For example, gaming still sucks on it because of the forced VSync so there is still no true fullscreen unredirection like on X.
              There's a merge request for disabling v-sync but my understanding is that nothing prevents a Wayland compositor from using direct scan out for full screen applications and that wlroots already does this.

              Originally posted by user1 View Post
              Or the fact that the Wayland desktop can't recover from crashes and other rough edges.
              I'm not exactly sure what this is supposed to mean.

              Originally posted by user1 View Post
              I know that Wayland is much better under the hood than X, but at lest to me, currently, whats point of it when X just works much better for me without any issues whatsoever?
              And for me, a person with a dual monitor setup that needs fractional scaling on one of them, Wayland works much better. I remember trying to use fractional scaling on X and it would result in some applications scaling down 50% instead of up which made the application less readable on the higher resolution monitor and unreadable on the lower resolution monitor.

              Comment


              • #27
                Originally posted by Myownfriend View Post
                One "feature" that people attribute to X.org, screen and window capture, is just a security flaw that's talked about like a feature. Sure, window and screen capture is desirable and isn't part of Wayland but it doesn't need to be. That's been proven by Pipewire.
                Oh, poor Wayland fans continue to perpetuate the falsehood. Show me a single person who's been hacked using Xorg "insecurity" "features", I dare you.

                Comment


                • #28
                  Originally posted by Myownfriend View Post
                  There's a merge request for disabling v-sync but my understanding is that nothing prevents a Wayland compositor from using direct scan out for full screen applications and that wlroots already does this.
                  13 years in and we still have features which are yet to be merged or being worked on. Like people above have mentioned you don't replace something with something utterly broken, incomplete and ... crashy.

                  Comment


                  • #29
                    Originally posted by Sonadow View Post
                    Right, screw this. I've had enough.

                    At this rate desktop Linux will never ever fully transition to Wayland and developers will simply continue to flog the zombie that is Xorg and the Xserver. Things like VR, HUDs and even newer exciting stuff will simply get grafted onto Xorg like the shitbag that it is. Let's continue to deal with having a kernel driver, a drm driver and an Xorg ddx driver just to draw the desktop instead of having everything done by the drm driver. And while we're at it, let's also continue to enjoy the clusterfuck of device-specific input drivers compared to have everything punted off to a general Wayland protocol extension.

                    Everyone might as well simply roll back all the efforts to create Wayland compositors and software that actually function properly on Wayland because the Xorg maintainers don't even have the will to stick with their plans of permanently abandoning the Xserver once and for all even after 13 years of Wayland development. If things screen sharing and remote desktop access don't work properly on Wayland, don't bother with wasting effort to address them because "switch back to an Xorg session" is going to be the easy way out.
                    I mean, the funny thing here is that most of the things you talked about are becoming the norm on Xorg as well. libinput processes the majority of X input these days, generic DRI drivers work fine...

                    Not saying X is special or great or anything, but it still has stuff going for it.

                    Comment


                    • #30
                      Originally posted by Sonadow View Post
                      The reason desktop Linux is full of technologies and features that are perpetually incomplete simply boils down to the fact that nobody has any desire to finish them as long as there is an easy fallsafe option.
                      How very nice that you have the time to play guinea pig. I actually have work to get done and Linux with X11 is currently a better option than either Linux with Wayland, Windows 10, or macOS for the way I use my system.

                      I was a die-hard KDE user until all the distros tried that trick with KDE version... 4.2, I think. I switched to LXDE to get a snappy, non-crashy desktop and didn't come back to KDE until somewhere around 5.10.

                      If it helps to calm you down, think of X.org Server 21.1 as Devuan

                      ...and, to be frank, I'll probably stay on X11 until someone comes up with the necessary extensions (and gets them adopted) to allow the compositor to restart without losing my session, the way it does when it's an X11 WM.

                      Hell, if I'm forced to use Wayland for hardware support, I might explore options for running my entire desktop inside something like an Xpra+XWayland-over-localhost stack.

                      Originally posted by mppix View Post
                      For a uniform DE, you'd need to go even one step further and depreciate XWayland or at least announce depreciation, e.g. for some GTK and/or Gnome version. However, I cannot even imagine the amount of cry and tears that will follow on this forum.
                      Well, that's a pretty stupid idea. The #1 reason that Microsoft has held such a monopoly since MS-DOS started to take off is that they know how much people care about compatibility with their old software and pour a lot of work into maintaining that compatibility.

                      Hell, when Canonical started making noise about dropping 32-bit x86 multiarch support, I started researching what distro to switch to in order to retain compatibility with 32-bit Wine and my 32-bit Humble Bundle games without having to choke down all the usability papercuts of the RPM ecosystem.

                      Comment

                      Working...
                      X