The last week I spent in the BlueSystems office in Barcelona for this years Plasma sprint. It was a wonderful timeing given that we are just approaching the last stages of preparing our first release based on Qt 5 and KDE Frameworks 5. We had some very good discussions and summaries of those can be read in sebas’s blog as well as on the Plasma mailing list.
My main focus for the sprint was obviously KWin and the Wayland port of the workspace. In that area I mostly worked together with Alex. For example we looked at all of the kde-workspace module to figure out what currently expects X11 in an unconditional way. Such code needs either to be ported or at least to be made runtime platform detectable.
Overall I tried to update the team on the current state of KWin/5 and KWin/Wayland and to highlight the next steps. I have some ideas on how to further change KWin/5 and the sprint was obviously a good opportunity to discuss the ideas in person and to get feedback. Of course I will also announce the changes on my bloge once they are implemented.
Some of the overall changes we discussed will also affect KWin. We are considering to split our monolithic kde-workspace module and that would most likely result in an own repository for KWin. This could allow us in future to experiment with more things. One of the ideas I got during the sprint was to do rapid release cycles for KWin once we did the first release on Qt 5 in order to bring the new Wayland code faster to the users.
Our most important discussions in the Plasma teams might have been a look at our tooling and processes. We identified quite some shortcomings and have good ideas on how to address those. And we did not even keep it by just talking we started to implement the changes right at the sprint. An outcome is that we sent summary mails about the discussions to the mailing list. These summaries are intended as blue prints and not as decisons. So anybody can provide feedback and it’s not required to have been at the sprint. We hope that this makes our community way more inclusive.
From the tooling perspective we also had a look at a FLOSS replacement for Trello boards and are quite interested in one of them. If we implement it, this could become a wonderful tool to get the work better distributed and more inviting for the community. It should make it much easier for developers to just go to the board and pick a task. Personally I’m quite satisfied with the experience I made with the Trello board for KWin so far and am glad that KWin could act as an experimental area for this workflow change. I use this opportunity to point out that the Trello board still has a large number of easy tasks and I’m still adding new tasks whenever I find a new one.
May I ask what is the Trello replacement you’ve been looking at? I’m interested obviously, as I find Trello a very good tool, unfortunately privative.
We had a look at cantas
Congratulations to such a successful event. I think the direction is great, both in terms of technical decisions and community progress. I look forward to your more technical blog posts during the future months.
From the KWin point of view: What benefits will Wayland bring? I guess also much of the old OpenGL code is not compatible, due to the fact that Wayland uses OpenGL ES. What about maintaining the old OpenGL backends? Would it make sense to make a KWin fork without the Xorg-backends?
I think you are badly informed, OpenGL works just fine on Wayland.