Announcement

Collapse
No announcement yet.

NVIDIA 555.42.02 Linux Beta Brings Wayland Explicit Sync, GSP Firmware Used By Default

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

  • #11
    Originally posted by Weasel View Post
    Because it can't do basic shit by design, such as querying window absolute positions, or positioning them, or parenting a process' window inside of another app as a child (YES EVEN IF BOTH APPS PERMIT EACH OTHER!).

    Broken by design.
    They really need to have a protocol for the positioning. A window should be able to request a position and retrieve its own positions. Firefox has been annoying me about that. Compositor could refuse the request, but they should be able to request it.

    Comment


    • #12
      Originally posted by Weasel View Post
      Bparenting a process' window inside of another app as a child
      That is actually very easy, I've done that for a customer demo a couple of months ago.
      Incorporating the main window of two child processes into the host application and even nicely dealing with any sub windows the two children might open.

      The child doesn't even need to know about this, any unmodified Wayland application will do.


      Comment


      • #13
        Working mouse capture for x11 clients.

        Comment


        • #14
          Originally posted by jeisom View Post

          They really need to have a protocol for the positioning. A window should be able to request a position and retrieve its own positions. Firefox has been annoying me about that. Compositor could refuse the request, but they should be able to request it.
          A protocol it's being worked on

          Hello everyone! This is a new attempt to resolve the issues clients designed for stacking window managers are facing when they want to set their own...

          Comment


          • #15
            Originally posted by jeisom View Post
            They really need to have a protocol for the positioning.
            As someone else already mentioned above as I was typing, it is being worked on.

            Wayland does still have shortfalls, that is known and acknowledged. However, Wayland does not have the biggest shortfall of all - the shortfall that X.org/X11 has.

            Wayland has a dedicated maintainer.
            Last edited by ezst036; 21 May 2024, 02:06 PM.

            Comment


            • #16
              Originally posted by xhustler View Post
              The windowing system is not the place to restrict what applications are and are not allowed to do.

              If you want to restrict what applications can do, use a sandbox (or disallow these protocols in your compositor via a configuration setting).
              This will never work. You can't just take a candy from software developers, just look how they are screaming when new API (wayland) doesn't offer them their usual toys. Imagine what will happen when someone will restrict something within existing API. All apps will just refuse to run, giving user a link to howto about adding an app to trusted list. They will check and ask for full control regardless of whether they need it. So, windowing system is exact place where apps ability to mess with user of other apps should be defined. Sandboxes, antiviruses, proactive intrusion detectors - all these are needed because fucked up security within existing API.

              It is funny, when linux adepts brag about linux security, meaning obsolete and useless UGO access control, while actively opposing efforts to make linux desktop secure.

              Comment


              • #17
                Originally posted by ezst036 View Post

                As someone else already mentioned above as I was typing, it is being worked on.

                Wayland does still have shortfalls, that is known and acknowledged. However, Wayland does not have the biggest shortfall of all - the shortfall that X.org/X11 has.

                Wayland has a dedicated maintainer.
                And a whole lot of good that does. It seems for many things you want to do with Wayland there are still two answers: either you need another protocol, which is being considered or worked on or there is a protocol, but you have to implement it yourself anyway.
                Last edited by bug77; 21 May 2024, 02:25 PM.

                Comment


                • #18
                  To be fair, the devs themselves admit this issue is caused specifically by the way that Blender implements some of these things. Afaik, most if not all of that could be solved with a refactor. Which isn't to say that it's not BS and shouldn't be necessary, but more than it's not really a limitation of Wayland so much as an annoyance.

                  Comment


                  • #19
                    Originally posted by bug77 View Post
                    Bah, Arch doesn't pick up betas, I will have to wait a little bit longer.
                    Though tbh the biggest Wayland offender lately hasn't been Nvidia. It's been Chromium.
                    ​​



                    Also the biggest wayland offender isn't nvidia or chromium. It's wayland. Maybe you missed the memo about --ozone-platform-hint no longer being supported, instead you just use --ozone-platform now. Chromium is fine. Nvidia was there before wayland was, the responsibility for poor wayland support doesn't fall only on Nvidia's shoulders, about half of it falls on the shoulders of the wayland devs. But Nvidia has been accomodating the eccentricities of the wayland devs, gradually, maybe even a bit begrudgingly, and certainly slowly, but they have.

                    Originally posted by ezst036 View Post
                    Wayland has a dedicated maintainer.

                    Maybe he'll go down in history as the worst maintainer to ever attempt to maintain anything... for how spectacularly he has failed.

                    Freedesktop, the same organization responsible for wayland is also responsible for X11, did u know that? It's basically their decision to not have a dedicated maintainer for X11. For some reason they really have a massive boner for getting people to use wayland and not use X, so they do just about everything in their power to cripple the development of X to try and make it lose it's massive lead over wayland, in fact they spend so much effort on this that they forget to actually develop wayland to such a point where it could maybe theoretically in some universe compete with X.
                    Last edited by rabcor; 21 May 2024, 03:21 PM.

                    Comment


                    • #20
                      Originally posted by bug77 View Post
                      And a whole lot of good that does. It seems for many things you want to do with Wayland there are still two answers: either you need another protocol, which is being considered or worked on or there is a protocol, but you have to implement it yourself anyway.
                      There is a problem here. The first prototype of wayland did not use the Wayland protocol in fact used dbus protocol the reality is wayland protocol comes about because dbus limitations and performance issues were a problem. From day 1 of wayland protocol it was Wayland protocol for some things and dbus protocol for others.

                      Yet we still have people wanting to shove everything though the Wayland protocol.

                      Also wayland development process is slow mostly as you get into the details as merge request 264 shows lot of thing we have been doing have been causing lots of issues. Ok yes works for me you don't see those issues.

                      Like the one lets allow application to place window were ever. What happens when that runs into start menu locaiton and other items like it. Applications under x11 have not had a clean way to find out where they should not attempt to put windows.

                      Xrandr issue of not put the screen result back after application end under x11 is because you don't have zones/grouping of applications as 264 implements to know when all the applications wanting X different screen res have in fact closed and the screen should return to normal res.

                      bug77 you name is kind of right. When you go digging there is over 770 know X11 issues all linked to window placement. Yes all issues that person could go a decade or more and not hit one of them. But then some other poor sucker hits like 50 of them in a day.

                      Horrible part here is lot of the issues that effect X11 also effect MacOS and Windows so you cannot go and copy someone else work to fix the problem. This creates the long wayland protocol extension development process. Wayland is a chance to fix a lot of long term issues. Problem is fixing those long term issues is lots and lots of research to collect up all the issues effecting a particular area and redoing until everything is covered.

                      Comment

                      Working...
                      X