Announcement

Collapse
No announcement yet.

Red Hat Developing New xwayland-run & wlheadless-run Utilities

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

  • #41
    Originally posted by bearoso View Post

    Not true at all. The Wayland spec might have come from them, but there is no one Wayland software project. The people implementing Wayland compositors are mostly those who used to do desktop environments, and it was clear from the start that many of them had never been display or graphics programmers before.
    You are just making things up. Major contributors to Wayland are also been heavy contributors to Xorg and much of ecosystem is shared between them. Just to name a few people I am aware of

    * Daniel Stone - XKB maintainer and current Weston developer
    * Simon Ser - long term Mesa contributor and Wlroots developer
    * Peter Hutterer - libinput used by both Xorg and Wayland
    * Olivier Fourdan - Founder of Xfce and current maintainer of XWayland

    Btw, much of this is funded by usage outside of desktop environments since nobody pays for Linux on the desktop. I challenge you to name a single person who doesn't have display or graphics programming experience who are major contributors to Wayland

    Comment


    • #42
      Originally posted by jacob View Post

      That's a gnome problem, not a wayland problem. Any bug in a gnome extension can kill the whole gnome shell. It does it on X11 too. Wayland prevents something like a dock to be a totally separate application with no connection whatsoever to the compositor, like it can be on X11 (there are pros and cons to that), but there is absolutely nothing in wayland that forces compositor extensions to run in the same process and with the same memory space. It's a gnome design flaw (arguably a big one).



      Wayland in no way infringes on user freedom. For the rest, it's a matter of opinion. For example, the gnome's Mutter was the first production-grade compositor and all the other projects could have just used it. Nothing forced everyone to implement their own compositor like previously the various desktops were not each writing their own X11 server. I find this argument especially strange given that so many people always claim (partly wrongly) that Linux is "all about choice". They insist that everyone must reinvent the flat tire with their own window manager, dock and file manager, but they expect the compositor to be standardised (but the init system must not be, but the libc must be - except if it's musl, etc.....)
      stop with these silly semantics debates. when I say wayland problem I mean "linux desktop" problem. If it inhibits linux desktop developers for running their apps in a consistent way across wayland compositoris it's not a wayland problem? That is such a tedious and absurd argument, yet it is made over and over. gnome is the number #1 wayland compositor. if it sucks it reflects on wayland and hurts wayland, hence it's a problem. Gnome opinions and what should and shouldn't be standardized has clearly not served anyone here least of all gnome. If the gnome group wanted to make mutter the standard for compositors they couldn't have done a worse job.

      In the X11 world there was a set of standards that window managers had to implement, but it wasn't entirely limiitted to X11 window manager hints, there's been lsb, xdg standards .desktop files, etc. The linux desktop doesn't have the manpower for all these compositors to go off in their own directions and design an implement standards for everything users have come to expect, especially given the corpus of existing x11 applications. I'm not saying the compositors have to change a single thing, all I'm saying it's been really awful, there's a bunch of missing functionality compared to x11 (again doesn't mean it has to be an wayland protocol, but it has to be implemented somehow), there's a lot of compositor compatability issues, there's a lot of potential for sharing code, and again, whatever the history of these projects, they aren't staffed like they used to be. My advice is that they need to consolidate around libraries and standardize protocols whenever possible and align their roadmaps in the best possible way.

      My argument for "user freedom" is not the way you state it. I'm saying that gnomes decision to create a "consistent desktop" (in opposition to "user freedom") has entirely backfired. First because it prevented the protocolization of certain desktop functions, and second it's reinforced a monolithic desktop, which is prone to crashing the entire desktop when the failure is quite isolated. Gnome shell doesn't have any protocols, while wlroiots, weston, kde have dozens. Are we really saying that what these compositors are doing are so different they don't have the same basic needs? Of course not, we know that a large part of the wayland community fought against any standardization, preventing consensus, compatability, code reuse. and that's "user freedom" I'm actually advocating for, just design things properly, and people will figure it out.

      Comment


      • #43
        Originally posted by fitzie View Post
        stop with these silly semantics debates. when I say wayland problem I mean "linux desktop" problem. If it inhibits linux desktop developers for running their apps in a consistent way across wayland compositoris it's not a wayland problem? That is such a tedious and absurd argument, yet it is made over and over. gnome is the number #1 wayland compositor. if it sucks it reflects on wayland and hurts wayland, hence it's a problem. Gnome opinions and what should and shouldn't be standardized has clearly not served anyone here least of all gnome. If the gnome group wanted to make mutter the standard for compositors they couldn't have done a worse job.

        It's not semantics, if you are pointing out a problem it's important to know where that problem is. There is nothing inherently wrong with mutter, in fact it's the most mature wayland compositor out there and the one with the best user experience by a long shot. Gnome shell on the other hand made some bad design decisions a long while ago and ever since it's being hampered by them, like its badly concieved hooks for extensions. It's been a problem on X11 and it's still a problem on wayland.

        Originally posted by fitzie View Post
        In the X11 world there was a set of standards that window managers had to implement, but it wasn't entirely limiitted to X11 window manager hints, there's been lsb, xdg standards .desktop files, etc. The linux desktop doesn't have the manpower for all these compositors to go off in their own directions and design an implement standards for everything users have come to expect, especially given the corpus of existing x11 applications. I'm not saying the compositors have to change a single thing, all I'm saying it's been really awful, there's a bunch of missing functionality compared to x11 (again doesn't mean it has to be an wayland protocol, but it has to be implemented somehow), there's a lot of compositor compatability issues, there's a lot of potential for sharing code, and again, whatever the history of these projects, they aren't staffed like they used to be. My advice is that they need to consolidate around libraries and standardize protocols whenever possible and align their roadmaps in the best possible way.
        Hence my comment that nothing mandates anyone to write the millionth compositor. To the extent that there need to be several desktops at all, they could very well use one compositor. It's their decision to develop their own, and their decision alone. Analogies between wayland and xorg are very misleading but as far as they go, the compositor is the X server, not the window manager. XFCE or KDE for example could very easily implement their desktop management on top of mutter in the same way gnome shell does. Instead they decided to go it alone. Their call, but if you have a problem with that, then understand that it's their problem. Not gnome's and not wayland's.

        Incidentally, one again, wayland is not aiming to bring over all the existing X applications and replicate all their functionalities. It's not meant to. Consider it akin to the transition from MacOS classic to OSX: the old applications were supported through virtualisation for a while, then deprecated and, in many cases, not replaced.
        Last edited by jacob; 29 November 2023, 10:18 PM.

        Comment


        • #44
          Originally posted by fitzie View Post

          Gnome shell doesn't have any protocols, while wlroiots, weston, kde have dozens. Are we really saying that what these compositors are doing are so different they don't have the same basic needs?
          GNOME Shell "doesn't have any protocols" is a nonsensical statement. GNOME Shell or more specifically Mutter uses many of the same Wayland protocols that all the other Wayland implementations uses. This is a prime example of ignorant users.

          Comment


          • #45
            Originally posted by spicfoo View Post

            Funny that you pretend you know what most Xorg users want or how they handled other transitions. You are only speaking for yourself and nobody else.



            Some users are indeed ignorant and that's ok but they also confidently claim they know more about the tech than free and open source developers who have spend decades working on it and if you do that, I don't believe you.

            I'm actually reading the arguments made by the xorg users, and they all point to deficiencies in wayland. Of course there's cranks out there that don't explain their opinion, but I'm not just basing this on me. It's people that you that consistently strawman this debate over and over again. You want experts here's an expert:

            Wayland. It comes up a lot: “Bug X fixed in the Plasma Wayland session.” “The Plasma Wayland session has now gained support for feature Y.” And it’s in the news quite …



            Wayland has not been without its problems, it’s true. Because it was invented by shell-shocked X developers, in my opinion it went too far in the other direction. Wayland’s minimal core protocols are lacking most of the features that non-trivial apps and desktops actually need to work–such as screen locking, screen sharing, cross-app window activation, non-integer scaling, and so on. Compositors all needed to come up with ways to do these things themselves. And that need for each compositor to implement everything itself fragments development efforts and disadvantages small teams without the expertise of heavy-hitting graphics developers. These are real problems and we shouldn’t sweep them under the rug.

            You want real users, here's fellow forum members
            Phoronix: KDE On Wayland: "The Biggest Thing Needed Now Is Adoption By 3rd Party Apps" Given the recent discussions stemming from Fedora 40 planning to ship KDE Plasma 6 and drop the KDE Plasma X11 session to focus solely on Wayland for the next-gen KDE desktop, prominent KDE developer Nate Graham has written a


            This is basically a rant.
            I really want to be a full on Wayland (KWin) user but the problem is it just doesn't work right as an end user, and I'm starting to lose faith that it ever will. I'm in ops for a small IT company, and use Arch on both my work laptop (Intel) and my home PC (nVidia), but when it comes to GUI I may as well be end user. Here's a bunch of my use cases where Wayland is a real nuisance:
            - Basically anything with my desktop. Blame nVidia all you want, but the simple fact is my browser flickers if I look at it funny on KWin Wayland and it doesn't on the X11 session. Sometimes the gnome terminal just decides to break transparency. Sometimes an entire monitor won't clear old output and it'll look like old crashed Windows XP explorer trails.
            - Any video sharing at all. This is now happening on both nVidia and Intel. It's just unstable. Its output will freeze if any app is maximized or full screen. This already made me look a fool at work (Including calls of "install windows lol") more than once. Unusable.
            - Moonlight sharing. KWin Wayland just spams my logs and the input grabbing in seamless mode does not work right. It will completely flood my journal (annoying) and won't always actually sync clicks and inputs, basically making it unusable. Dumb. Also, the flatpak is unusable with Wayland (but not with X11!).
            - Thunderbird. It will literally not start on Wayland anymore on my desktop. No idea why.
            - Returning from sleep. Sometimes it works, other times I have to kill the session or reboot. Sometimes it breaks the lock screen and I have to jump to a TTY. Dumb.
            - Gaming performance out the box. It's just better on X11.
            - Recovering from OOM crashes. I can almost never get to a TTY when I'm using the Wayland session (this mostly happens on the laptop, so it's not nVidia).
            - Input latency. Often you will have insane judders where the cursor is just stuck for no reason in the Wayland session.
            - The entire DPI debacle with text looking blurry or not scaling right. This one actually struck my colleague harder (GNOME user) to the point he switched back to X11 and just settled on a single scale for both monitors despite being suboptimal.
            - Speaking of which, basically no "legacy" Java application ever "feels" right (missing icons, bad scaling, misplaced menus, menus with window borders that shouldn't have them, etc.) on KWin Wayland, and we use a ton of those.
            For these reasons and way more it most of the time just feels like Wayland is reinventing the wheel for no reason or just to avoid maintaining something that while massively flawed, just worked. Stuff like this, the perpetual technological clout chasing with no regard to compatibility is the reason Linux will never be a mainstream desktop OS. Ever.

            This is now really funny because the anti-xorg crowd have been busy arguing with themselves, and every time someone points this out to them, they come back and ignore it.

            Comment


            • #46
              Originally posted by fitzie View Post

              I'm actually reading the arguments made by the xorg users...
              Are you really? Your ability to selectively copy/paste some links or text still doesn't allow you to pretend that you know what most Xorg users want. Plural of anecdote is not data. Riddle me this, your first blog post is about KDE moving to use Wayland by default. If it was so deficient, why is KDE moving to using it by default? Why is Xorg foundation backing Wayland? Why is nobody willing to maintain Xorg server releases?

              Comment


              • #47
                Originally posted by spicfoo View Post

                GNOME Shell "doesn't have any protocols" is a nonsensical statement. GNOME Shell or more specifically Mutter uses many of the same Wayland protocols that all the other Wayland implementations uses. This is a prime example of ignorant users.
                my argument is that for a significant chunk of capabilities for which there is an established protocol created by kde or wlroots or weston, gnome doesn't use it. Not to say they don't support any protocol. you are playing gotcha while I'm actually trying to explain something to you. The fact is that many things that are done in some standardized IPC way in other compositors are done inprocess because gnome is a monolithic architecture. The only way for you to see certain states of gnome is by writing javascript and running inside gnome-shell itself, not via dbus, standard wayland protocol etc.

                If you are so quick to call me ignorant instead of properly trying to understand my point, then I will quit trying to explain things to you.

                It's so silly because it's terribly obvious that gnome shell architectures a mess, they do some patchwork by moving things into different threads, but it's not built for purpose, and it should be multiprocess with standardize protocols shared by other compositors where appropriate. gnome-shell devs are upset that purism saw this and wrote phosh/phoc, and they will surly be upset when cosmic comes out again, showing them that gnome-shell architecture is a dead end. I've seen their rejection of wayland protocol standards with the statement "we don't want to allow user freedom in this aspect". But they did it by ossifying their shell, and contributing negatively to compositor compatibility issues and lack of capabilities we have today.

                Comment


                • #48
                  Originally posted by fitzie View Post

                  my argument is that for a significant chunk of capabilities for which there is an established protocol created by kde or wlroots or weston, gnome doesn't use it. Not to say they don't support any protocol.
                  What you originally said was "Gnome shell doesn't have any protocols". You don't get to now pretend you didn't say that.

                  Originally posted by fitzie View Post
                  Running inside gnome-shell itself, not via dbus, standard wayland protocol etc.
                  GNOME Shell uses D-Bus extensively and does use most of the Wayland protocols as well. More examples of ignorance.

                  Comment


                  • #49
                    Originally posted by spicfoo View Post

                    Are you really? Your ability to selectively copy/paste some links or text still doesn't allow you to pretend that you know what most Xorg users want. Plural of anecdote is not data. Riddle me this, your first blog post is about KDE moving to use Wayland by default. If it was so deficient, why is KDE moving to using it by default? Why is Xorg foundation backing Wayland? Why is nobody willing to maintain Xorg server releases?

                    i'm saying all the xorg users I've heard are saying their bias towards xorg over wayland isn't some fondness of antiquity or whatever strawman you believe, it's technical issues with wayland.

                    Originally posted by spicfoo View Post

                    What you originally said was "Gnome shell doesn't have any protocols". You don't get to now pretend you didn't say that.



                    GNOME Shell uses D-Bus extensively and does use most of the Wayland protocols as well. More examples of ignorance.
                    I've tried to explain my point, when I said gnome shell doesn't have any protocols, it is was to say unlike wlroots and weston and kde who have all created namedspaced wayland protocols for their use, somehow gnome hasn't created any named spaced protocols like all the other compositors. this wasn't to mean that they don't use any wayland protocols at all. This in of itself isn't proof that gnome-shell is has a flawed architecture, but it is compounding evidence, and the flaws in gnome-shell are not controversial to most people. Gnome has been a somewhat good user of dbus, and is pushing along xdg portals, but it's long overdue, and the fallout has been there. the entire wayland protocol standardization process was a mess, and it took wlroots to start pushing for the gnome team preventing progress there. all this is true.

                    but you want to stick to calling me names so I'm done with this conversation, you win have a great day.

                    Comment


                    • #50
                      Originally posted by fitzie View Post

                      i'm saying all the xorg users I've heard are saying their bias towards xorg over wayland isn't some fondness of antiquity or whatever strawman you believe
                      I don't think all Xorg users are biased towards Xorg in the first place and I also never said anything about what I believe. In other words, that's the real strawman here.

                      Originally posted by fitzie View Post
                      I've tried to explain my point, when I said gnome shell doesn't have any protocols, it is was to say unlike wlroots and weston and kde who have all created namedspaced wayland protocols for their use, somehow gnome hasn't created any named spaced protocols like all the other compositors.
                      Weston hasn't done what you are claiming here and GNOME Shell uses D-Bus interfaces or unstable namespaces instead of its own. It isn't indicative of any architecture flaws like you seem to think. You don't have a point. You are just cobbling together random unrelated things into an incoherent conclusion

                      Comment

                      Working...
                      X