Announcement

Collapse
No announcement yet.

KDE Plasma 5.25's Exciting Improvements For KWin, Continued Wayland Optimizations

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

  • KDE Plasma 5.25's Exciting Improvements For KWin, Continued Wayland Optimizations

    Phoronix: KDE Plasma 5.25's Exciting Improvements For KWin, Continued Wayland Optimizations

    With Plasma 5.25 now under a soft feature freeze ahead of its official release next month, KDE developers are focusing on bug fixing as well as talking more about all of the changes they managed to land this cycle...

    https://www.phoronix.com/scan.php?pa...asma-5.25-KWin

  • #2
    I find KWinFT based on wlroots a lot more promising: https://subdiff.org/blog/2020/the-kwinft-project/

    Comment


    • #3
      Wayland on KDE 5.24 has been a pleasant experience for me. No more big crashes, most functionality is in place, very nice indeed. Now if only more apps jumped into Wayland...

      Comment


      • #4
        Originally posted by darkdragon-001 View Post
        I find KWinFT based on wlroots a lot more promising: https://subdiff.org/blog/2020/the-kwinft-project/
        Unfortunately in the real world it always worked way worse for me
        ## VGA ##
        AMD: X1950XTX, HD3870, HD5870
        Intel: GMA45, HD3000 (Core i5 2500K)

        Comment


        • #5
          Originally posted by darkbasic View Post
          Unfortunately in the real world it always worked way worse for me
          Indeed not too long ago, not even Firefox was usable. It might be related to shortcomings of the Firefox Wayland implementation, but even if this was true, it doesn't help me as a user when it works with other Wayland compositors.
          I'm also incclined to think that with all the progress being made on upstream KWin, upstream devs had a valid point by keeping things more conservative.
          Anyway, I'll give a new KWinFT version a try regardless. It certainly deserves attention and perhaps some new version will turn out as a pleasant surprise.

          Comment


          • #6
            Does anyone has an explanation/inside view why the kwin-ft approach of using wlroots is not desirable? why looks like kwin current approach is to double the wayland implementation effort instead of converging into wlroots?

            Comment


            • #7
              Originally posted by C8292 View Post
              Does anyone has an explanation/inside view why the kwin-ft approach of using wlroots is not desirable? why looks like kwin current approach is to double the wayland implementation effort instead of converging into wlroots?
              It probably has something to do with wlroots being written in C, whereas KDE is C++. And it also depends on where wlrrots was when KDE's Wayland efforts started.

              Comment


              • #8
                Originally posted by [email protected] View Post
                Wayland on KDE 5.24 has been a pleasant experience for me. No more big crashes, most functionality is in place, very nice indeed. Now if only more apps jumped into Wayland...
                Same here, except for one nasty crash and lag bug in Lokalize, which forces me to run it under XWayland. The bug has been filed on the KDE Bugs website and has been open for a long time now. If only they could fix that, then I would have everything on Wayland (except for Steam and Wine).

                Comment


                • #9
                  Originally posted by darkdragon-001 View Post
                  I find KWinFT based on wlroots a lot more promising: https://subdiff.org/blog/2020/the-kwinft-project/
                  Unfortunately for anyone who wants to try out KwinFT, interest from folks packaging for distributions actually seems to be going backwards. Unless you want to build from source (which many won't want to do for a core piece of their GUI), official packages only exist for Manjaro, with AUR packages for Arch proper. There was mention of the Fedora KDE SIG taking interest and setting up a COPR, but no packages were ever built there. There used to be packages in Zawertun's KDE COPR, but those haven't been built since 5.22 six months ago. I imagine Fedora lost interest when KDE started focusing heavily on pushing Wayland along. If official packages never get built for Debian/Ubuntu, Fedora, OpenSUSE, it doesn't feel like this will ever gain meaningful traction.

                  Comment


                  • #10
                    Originally posted by C8292 View Post
                    Does anyone has an explanation/inside view why the kwin-ft approach of using wlroots is not desirable? why looks like kwin current approach is to double the wayland implementation effort instead of converging into wlroots?
                    KwinFT "complete rewrite everything as big bang" is not desirable: https://www.joelonsoftware.com/2000/...ver-do-part-i/

                    Comment

                    Working...
                    X