Announcement

Collapse
No announcement yet.

The Ongoing Work For Native Wine Wayland Support

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

  • #11
    Originally posted by avis View Post
    How are they going to tackle absolute window positioning which is required by many applications including Adobe Suite and which is AFAIK not part of the Wayland protocol because reasons?
    You find the following in the PDF and the presentation and the experimental branch.
    Coarse position synchronization (exp.)
    then note about
    Wayland (logical space)
    This is the thing Wayland wine support has been finding that majority of windows applications don't need absolute window positioning to work. Logical space is good enough for majority of windows applications.

    Avis the thing to remember from the wine prototype of Wayland support less than 1 in 10 Windows applications in fact use absolute window positioning in any complex way.

    Not all applications in the Adobe Suite in fact require absolute window positioning used in any complex way that was found in the early wine wayland prototype.

    Remember multi window gimp the windows version does in fact work in the prototype of wine for Wayland.

    One of the reasons why absolute position is not part of the Wayland protocol is failure to find examples where it really required. Example after Example have turned out not to require absolute position even after a person put it forwards as real world example that required it.

    Avis get the problem yet. Wayland developers don't add stuff to protocol without functional use case. For absolute position no one has found one that passes full inspection. The presumed need of absolute position may be a limitation myth we have been using for decades.

    Wayland Window could be floating in a 4D space falling into a black hole for all the application knows about it. Of course since this is the case Wayland applications can kind of write its own rules of it own relative space(also called logical space).

    Avis lot of ways X11/windows compared to Wayland is like newtons physics vs General Relativity. Yes Wayland being the general relativity.

    Wayland is a good chance to re-look at some of these long term beliefs and check out if they are valid or not.

    Comment


    • #12
      Originally posted by cl333r View Post
      They dumped everything into a 9 hour video, genius...
      perfect, fits into one work day

      Comment


      • #13
        If I'm not mistaken, Wayland has that (IMO terrible) design idea where it does not handle the input events (mouse, keyboard; therefore an application must gather those input events from some other source. Or something alike that. Which is completely unlike X11.

        So WinE must now gather input events from libinput, or something similar, to be able to use Wayland?

        Edit: at 5:35:17 "Wayland is somewhat special in that events are not events on surfaces but rather events on a dedicated input object"

        Does that mean that Wayland provides input events, but it does not specify to which window events belong to (an application might open many windows)? That would be nuts.
        Last edited by drastic; 25 October 2023, 05:04 PM.

        Comment


        • #14
          Originally posted by cl333r View Post
          They dumped everything into a 9 hour video, genius...
          Its a livestream vod and they gave you time codes

          Comment


          • #15
            Originally posted by drastic View Post
            If I'm not mistaken, Wayland has that (IMO terrible) design idea where it does not handle the input events (mouse, keyboard; therefore an application must gather those input events from some other source. Or something alike that. Which is completely unlike X11.

            So WinE must now gather input events from libinput, or something similar, to be able to use Wayland?

            Edit: at 5:35:17 "Wayland is somewhat special in that events are not events on surfaces but rather events on a dedicated input object"

            Does that mean that Wayland provides input events, but it does not specify to which window events belong to (an application might open many windows)? That would be nuts.
            Its a side effect of what wine is doing. Notice how wine is using xdg_surface for basically everything. xdg-surface is not in fact an independent window in the way Wine wayland uses it. What wine wayland is doing would be like MDI windows on windows for everything there is a side effect from doing this.

            Comment


            • #16
              Originally posted by avis View Post
              How are they going to tackle absolute window positioning which is required by many applications including Adobe Suite and which is AFAIK not part of the Wayland protocol because reasons?
              Did you watch the video?

              The presenter seemed pretty upfront about the idea that some applications that have specialty needs may need to stick with x11 only for now, but that their current algorithms are generally good enough for most applications.

              In the long run, perhaps a new wayland extension will be added for proper wine support, but it sounds like they are focusing on getting what they can do now working and released first.

              Comment


              • #17
                Originally posted by smitty3268 View Post

                Did you watch the video?

                The presenter seemed pretty upfront about the idea that some applications that have specialty needs may need to stick with x11 only for now, but that their current algorithms are generally good enough for most applications.

                In the long run, perhaps a new wayland extension will be added for proper wine support, but it sounds like they are focusing on getting what they can do now working and released first.
                It's not about only Wine FFS

                NOTE: We are currently working on an alternative solution that preserves (most of) the compatibility with this protocol but does not expose a global coordinate system. Please...




                Kudos to oiaohm for perfect gaslighting. Wayland is indeed perfect and people are just fucking around - they absolutely don't need absolute window positioning.

                Comment


                • #18
                  Originally posted by treba View Post

                  Err, they do - they are even switching over the browser, see Lacros.
                  Err, they don't. They are going to switch, however. News from August 2, 2023: https://arstechnica.com/gadgets/2023...asier-updates/

                  “Previously ChromeOS was using a homemade graphics stack called "Freon," but now with Wayland, it'll be on the new and normal desktop Linux graphics stack. Google's 2016 move to Freon was at a time when it could have moved from X11 (the old, normal desktop Linux graphics stock) directly to Wayland, but it decided to take this custom detour instead.​”

                  Comment


                  • #19
                    Originally posted by Vistaus View Post

                    Err, they don't. They are going to switch, however.
                    They haven't enabled Lacros by default yet, but they already use Wayland in order to run Linux apps.

                    Comment


                    • #20
                      Originally posted by avis View Post
                      It's not about only Wine FFS

                      NOTE: We are currently working on an alternative solution that preserves (most of) the compatibility with this protocol but does not expose a global coordinate system. Please...




                      Kudos to oiaohm for perfect gaslighting. Wayland is indeed perfect and people are just fucking around - they absolutely don't need absolute window positioning.
                      No I did not say wayland was perfect. Do note that 249 "Draft: Add xdg-alignment protocol for window-placement control" is relative positioning not absolute positioning.

                      249 is basically what would happens if extend https://wayland.app/protocols/xdg-shell#xdg_positioner to be able to work in a broader sense.

                      Would have paid for your to read 249.

                      The reality here you may have been gaslighting the complete time that windows/X11 absolute positioning is in fact required in many cases.

                      Please note I am not saying that absolute positioning of some form is not required at all. But there is a very high probability that windows/x11 absolute position is only really need for a very small corner case.

                      Yes Windows and X11 developers in a lot of cases have been doing lots of code to do relative position operations that are covered by xdg_positioner design in X11 and windows form of absolute position.

                      avis the problem here wine does show a few things. Also those who study the problem have found a few things as well.

                      1) Single 0.0 origin point model does this match GPU hardware this is no it does not. Yes each connected monitor on a GPU has it own 0.0 origin point for absolute position on output.
                      2) GPU buffer operations by GPU design are meant to use relative to buffer operations.
                      3) Number of applications in fact using absolute position operations because they need absolute position is insanely low.
                      4) there is a different type of absolute position to what X11 and Windows use. Absolute position based on output device this now would be how to get a xdg_positioner anchor.

                      Avis see problem yet Windows and X11 idea of absolute position is limiting. Relative position+ means to get per output anchors can do a lot more. Of course majority of applications will not need per output anchors.

                      Comment

                      Working...
                      X