Looking back, I think it might be 9 months since my last “real” post. Oh, well. I have to admit my personal/work commitments (which I’m gonna be writing about soon) took me away from blogging.
Anyway, what better occasion than the latest Solid sprint could have brought me back to blogging? Many of the participants already wrote a report, so I’ll just give you an overview on how things went from my point of view.
First of all, this sprint had a peculiarity compared to our last ones: lots of new people and new faces. I can’t stress how awesome it is to have new skilled and committed metalworkers. Secondly, it has been great to see more people seriously committed to powermanagement, which up to last year has been a pretty much solo effort. Oliver, Lukas and Alex have been invaluable in helping fixing bugs, refactoring the code and adding new features. So, a quick list of what we have done:
- Refactored the logic parts of core vs. actions. The core now (finally) acts as a pure policy handler, whereas all the logic needed for each action (e.g.: suspend, hibernate, etc…) is now completely boxed into the actions.
- Refactored brightness handling. Now we have a way for reacting to brightness changes when the hardware does, which means no heuristics in trying to understand when and how the brightness changed. We are also planning to remove the KAuth helper.
- Started working towards having powerdevil’s logic completely asynchronous.
- Refactored KCM’s logic to have actions really capable of advertising their support on the system. This means now the KCM shows only those actions which are really supported.
- Fixed several bugs in keyboard brightness handling, which should be rock solid by 4.11.
- Fixed other bugs here and there, which became trivial to fix after all the work above.
Overall, this means an improved and more stable experience when it comes to powermanagement in KDE 4.11. We’re still working on making this stellar for the final release, but I can definitely say the backbone is there.
On a side note, some of our changes obviously broke backwards compatibility with our internal DBus interfaces. Thankfully Hrvoje noticed the fact that most of workspace was using those for interacting with powermanagement. Needless to say this is plain wrong. If you are a developer, you should instead use Solid::PowerManagement, which has all the needed logic for inhibiting and notifying a resume from suspend. This applies also to workspace developers, given you can actually use the very same class for requesting sleep states. This, though, has been already fixed by the Review Request linked above.
More than that, me and Lukas managed to have a quick chat and a plan for integrating solid-partitioner, which is the final result of last year’s summer of code by Lisa. We’ll come up with a joined plan really soon.
On this topic, me and Martin finally managed to sync up with the plans for Polkit in KDE in 2013. We’re planning an exciting GSoC project, so stay tuned.
I guess that’s pretty much all. We had lots of fun and did lots of work as usual; it’s really great to have a concrete taste of the community every once in a while. Speaking of which: Akademy is coming, and this year I had the honor and privilege to be part of the committee of the event. Needless to say, I hope to see you all there.