This year was a very interesting year in the development in KWin. After having been the maintainer for several years, I knew that I would not be able to continue being maintainer for personal reasons. Personally I had tried to lower my contributions for quite some time already and encouraged others to do reviews, keeping out of some reviews completely. In 2018 several of such code contributions landed which I hadn’t looked at all and which work out greatly – an example is the new excellent blur effect which I didn’t look at the code at all.
When I stepped down as maintainer I had to read many negative and fearful comments doubting the future of KWin. Personally I was very positive that me stepping down would not have a negative impact, in fact I even hoped for positive impact as it gives new blood the chance to fill up the gap and bring in new ideas. I had become very conservative over time.
So I just run some git stats [1] over the KWin repository to try to verify the assumption that me stepping down had no negative impact: In 2017 there were 614 non scripty commits. The author with most commits was me with 387. Overall 37 developers contributed to KWin.
In 2018 (01.01.2018-26.12.2018) there were 644 non scripty commits authored by 48 developers. The developer with most contributions is Vlad (241). I am no longer the top contributor in KWin (“only” 115 commits) – the last time this happened was 2008, the year I joined KWin development.
I am very happy about this development. We have new developers like Vlad and Alex working on areas which had been neglected for quite some time. The way the effect system improved thanks to them is really great. We have developers like David and Roman improving the Wayland support and overall KWin core. The knowledge about KWin gets spread and the development work is spread over more shoulders. Having a single developer doing the majority of commits is not always healthy. We can see the positive effects this year: we have more contributors overall and more contributors contributing multiple patches.
[1] git shortlog -sne –since=”01 Jan 2018″ –before=”01 Jan 2019″ and git shortlog -sne –since=”01 Jan 2017″ –before=”01 Jan 2018″
Any chance that Kwin will be built on wlroots in the future! It would be great if many wayland compositors were compatible and shared code.
Wlroots is a great project but doesn’t make sense for KWin. Please see: https://blog.martin-graesslin.com/blog/2015/03/why-dont-you-just/ – it predates wlroots, but all arguments hold as well.
KWin probably has a lot of wlroots functionality built in. It doesn’t seem like there would be much benefit to building directly on wlroots. With that said, it would be great if KWin supported some of the Wayland extensions that wlroots and Sway support. The Layer Shell, screen capture, and temperature shifting protocols are particularly useful for Wayland usability and cross compatibility.
It is indeed the cross compatibility that I was aiming for with a wlroots-based kwin. It would be great if one could run a wayland Plasma with sway just like it now is possible to replace kwin with i3 under x11.
The major thing I am missing in my sway session compared to my Plasma/i3 is krunner as app/file/url launcher.
I’ve noticed something strange with Kwin. If I use OpenGL 2 or OpenGL 3 backends the toolbar and kicker menu remain solid. If I use Xrander or use the llvmpipe driver they become semi-transparent. Is Kicker and the toolbar supposed to be semi-transparent as default?
That’s the blur effect.
With OpenGL I get the blur effect on the menu. I guess it only works on GL so you get transparency effect on xrender.
This is awesome news. You’re still contributing and kwin is getting better and better. Happy days!
In these troubled times stepping down to bring new blood is alike to let project die… or worse. New gen is trying hard to move as far as possible from all the previous generations, and methodes they use are… controversial, to tell it mildly.
I’m afraid KDE will never be the same. Hopefully generation next to new one will fix the mistakes.
when people don’t send patches you guys complain, when people send patches you guys complain. talk is easy, show us the code.
First of all thank you very much for all those years of hard work! You did really great job! I agree that having a single major contributor is not healthy for a project. It’s much better when there are more than few major contributors in case of some changes in personal life which makes you to stop development. I’m very happy that we have Vlad and Alex working on KWin! I hope you will still do some minor contributions. KWin is great, it drives whole KDE Plasma and is super important for KDE ecosystem.
Also I hope to see more improvements for KWin from Wayland side. If KWin in future could use Vulkan instead of OpenGL that would be amazing! 🙂
Best wishes, Martin!
Thanks for all of your hard work on KWin Martin. And thanks for your blog! I’ve always found your posts informative yet concise and easy to understand. It’s a neat look into the plumbing of my favourite software suite.