Announcement

Collapse
No announcement yet.

KDE Plasma 6.0 Lands More Performance Optimizations, Better Wayland Gaming Experience

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

  • polarathene
    replied
    Originally posted by gfunk View Post
    Its not really clear anywhere what the difference or benefits of KwinFT are , anyways glad to see reducing input lag and gaming improvements are a focus as part of plasma 6
    There were some great blog posts when KwinFT was getting started by Roman, and they highlighted some of the advantages, some that used alternative screen management libs to support KwinFT but could be used with Kwin too IIRC. Since then there's been a bunch of rework upstream by the Plasma devs with Kwin and other packages regarding display management, so I'm not sure how these compare now.

    KwinFT doesn't have the same scheduler that Kwin had been working on improving since which gave us those additional settings for the compositor to favor latency vs smoothness, and the cursor improvement ties into that I think, so probably not applicable to KwinFT?

    In VMware on Wayland at least, Plasma 5.27 kwin cursor is not stuck in the VM guest window but was in releases about a year ago, whereas KwinFT 5.27 last I checked was stuck. For VMs in general though as my linked issue earlier shows KwinFT isn't all that usable, no issues with Kwin and some bugs with effects that used to cause problems there have also been resolved.

    Leave a comment:


  • polarathene
    replied
    Originally posted by romangg View Post
    polarathene Sorry that I haven't yet looked into your issue. After Valve stopped funding my project and instead concentrated fully on KWin, I had to start working in some commercial projects to cover daily costs of living. So I naturally must concentrate on a subset of specific topics in KWinFT to work on in the time that's left.

    Regarding GitLab vs GitHub. I'm currently evaluating a move to GitHub. But the reason is more that the GitLab CI policy is kind of annoying with contributors often having to run their own custom CI runners to being able to contribute.
    Yeah no worries, I work on open-source projects unpaid too I understand

    Prioritize what is important for you with KwinFT. Github for CI is great, I've used Github Actions quite a bit, you may have interest in adopting GoReleaser + CloudSmith if that's viable for packaging KwinFT for a wider audience. I might be able to assist a bit with that if I can find the time.

    As a user since there is still quite a bit going on and releases are infrequent, and now especially with the transition to Qt / Plasma 6.x; it's mostly a project I just keep an eye on and every few months peek through issues and merge activity. I'm glad that you still find time to work on KwinFT. When it isn't broken in weird ways in QEMU / VMware, I can try it there more confidently before trusting it on bare metal

    After the 6.0 release, if you can find the time to put together a blog post, or even the project README to clarify any notable differences between Kwin and KwinFT as they have diverged a fair bit now and that is quite apparent in integration/settings and perhaps features there is probably some pro/con list between the two.

    From a dev perspective I really like the approach KwinFT has taken, but the user perspective might seem a bit muddy (the README stated values and my issue, while a bit niche due to VM environment aren't exactly encouraging). That sort of comparison between Kwin and KwinFT is only going to become more difficult for users to individually assess, it may be important to clarify what tradeoffs are involved or set some expectations.

    No expectation to engage in that issue I raised any time soon. I think it's probably best to wait until Plasma 6.0 is out and KwinFT has a release that can be tested alongside it. Other parts of the ecosystem may have also improved. For all I know, it may be upstream with wlroots as I did not try any other wlroots based compositors.

    Leave a comment:


  • gfunk
    replied
    Its not really clear anywhere what the difference or benefits of KwinFT are , anyways glad to see reducing input lag and gaming improvements are a focus as part of plasma 6

    Leave a comment:


  • Vistaus
    replied
    romangg I will surely try again with Plasma 6, but not on Manjaro (I don't like that distro).

    And I do support the move to GitHub. I know a lot of people are fan of GitLab, but to me it's the most crappy experience ever. GitHub may be run by Microsoft, but it has a much nicer workflow and a much better and cleaner UI. I like working on GitHub 10000000% better and that's no exaggeration!

    Leave a comment:


  • romangg
    replied
    timofonic Thanks for the ping. I'm gonna need to remember the notion "sectarian attitudes", what you said earlier. That describes certain behavioral patterns quite well.

    polarathene Sorry that I haven't yet looked into your issue. After Valve stopped funding my project and instead concentrated fully on KWin, I had to start working in some commercial projects to cover daily costs of living. So I naturally must concentrate on a subset of specific topics in KWinFT to work on in the time that's left.

    Regarding GitLab vs GitHub. I'm currently evaluating a move to GitHub. But the reason is more that the GitLab CI policy is kind of annoying with contributors often having to run their own custom CI runners to being able to contribute.

    Vistaus Thanks for trying from time to time. I hope you'll try again after the next release with Qt/Plasma 6. In general I recommend Manjaro to test KWinFT though, because they provide a package on their own and do some testing.

    smitty3268 I don't think that's what he wants to say. I was able to remove a lot of KWinFT' backend code and use wlroots instead and that didn't mean "replacing" the all of KWinFT. But of course that did cost me a lot of time and effort. In this time I wasn't able to work on features and bug fixes. Something that most users find more attractive than increasing overall code efficiency.

    Leave a comment:


  • smitty3268
    replied
    Originally posted by timofonic View Post
    I think using wlroots and improve that common codebase should be the norm between most DEs/WMs, not the exception.
    It seems like your argument is that Kwin should just be abandoned and replaced. It's certainly an option, but not one that I think is particularly popular among most kwin users.

    Leave a comment:


  • Vistaus
    replied
    Originally posted by Steffo View Post

    Why is it a shame? What's the point of financing two KDE compositors? When I read forum posts about KWinFT, I don't see the promise of stability when users report crashes or black screen.
    I've tried several times in the past 1.5 years to get it running, but I can't get it to even display *something*. On several distros, btw, including Void, Arch and openSUSE Tumbleweed.

    Leave a comment:


  • Termy
    replied
    Originally posted by NotMine999 View Post
    A forum topic mentioning Wayland and nobody stepping up to whine & complain about Wayland not having feature "[fill in the blank]".
    All right, all right, i'll bite the bullet: I really hope they implement gamma control for plasma 6 wayland!
    (ok, not a wayland feature thats missing, but at least i tried )

    Leave a comment:


  • awesz
    replied
    Framebufferland compositors are improving at really fast rates. Maybe in another decade and a half they will finally come close to catch up on X11 WMs?

    Leave a comment:


  • NotMine999
    replied
    Something is wrong here.

    A forum topic mentioning Wayland and nobody stepping up to whine & complain about Wayland not having feature "[fill in the blank]".

    The world must certainly be coming to an end soon.

    Leave a comment:

Working...
X