Now that I am back home I finally find some time to blog about my impressions of Akademy this year. Overall it was a really great Akademy at an awesome location and wonderfully organized by the local team. Thanks a lot for your work!
For me it was the already the fifth Akademy – I hadn’t realize that before and was quite surprised, I still remember my first Akademy in Mechelen as it were yesterday. This year Akademy was a very special one as I had been a member of the jury for the Akademy Awards. I am very happy with the people and products we awarded this year. Camilla, Lydia, Kevin and Nicolás all do an awesome job in their area and all help shaping the future of KDE.
From a conference point of view my personal highlights were the talks about Qt 5 and Frameworks 5 (thanks to Thiago, Kevin and David) as well as Mirko’s talk and the talk about defensive publications. I’ll probably start to document some of the “innovations” inside KWin soon (personally I do not consider anything inside KWin being an innovation, but I see that we should have it published before someone gets a patent on what we did years ago).
My highlight of the days after the conference is the training on QtQuick 2 by our friends from KDAB. I was afraid that I would not learn anything new, but there was lots I did not know and I realized how I can improve my QML code quite a bit. Some of that (especially based on feedback from Nuno) already made it into the 4.9 release. But maybe most important is that I got an idea how QML based effects could look like which could contain shader elements as will be available in QtQuick 2.
While nothing is implemented yet I thought about how our Fade effect could look like in QML:
WindowEffect { id: myWindow opacity: 0.0 onClosed: opacity = 0.0 Component.onCompleted: myWindow.opacity = 1.0 Behavior on opacity { NumberAnimation { duration: 150 } } }
Well we will see how this idea will evolve.
Another highlight of the days after the conference was a session where Jeroen van Meeuwen was present and explained his plans for our bugtracker. I am very glad that there is someone who cares and wants to help us getting bugs.kde.org in a usable shape again. I am really looking forward to this happening and I hope that our developers will support his efforts and start using the bugtracker. I am convinced that bugzilla is our most important and most neglected piece of architecture and a proper usage will help us producing better products (and with that I don’t mean that we fix bugs).
Of course I hardly did any coding, but I mentored a BoF where I explained a little bit about the internals of KWin. I hope to see some patches and that is more important than me coding. But I fixed a few bugs I found by pure chance. And I must say this makes me rather unhappy. If a functionality is broken for almost a year and it’s me noticing then something is really broken. That I don’t use the functionality, that’s just fine. But that no user noticed that the logout effect has not been working at all in 4.8 is quite shocking.
So please my dear users: if you see some “obvious” bug where you think it must have already been reported, please report it nevertheless. Not each obvious bug might be obvious to the developers (I only suspend my system and over the last year got only logged-out by a freezing BLOB). I rather have a duplicate too many than a broken feature.
That reminds me: we currently have a release candidate out. Get it, test it!
hi…
I’ve got an ati graphic card with the catalyst driver and logout never worked. I thought this is normal because of catalyst. I hope your fix will make it working even with catalyst. Thanks for your great work!!!
no, that has nothing to do with it. Best just switch to the radeon driver.
Behavior is great states are not 😉
Hm, since I usually look, away from the Monitor once the “Shutdown” Button is pressed I never really noticed that it was broken. So I would say that for me personally that said effect is not really important.
YES! Slowly but steadily you do seem to be willing to move the effect system over to QtQuick 2 🙂 Something i suggested a while ago which you where opposed to iirc 😉
I hope it’s going to work out since it does feel like the most powerful way to have effects in a easily sharable manner.
Anyway, for the bug you found, the logout effect. Do you have any plans on how to detect such bugs in the future? The only way i think could be possible is by making unit tests if you want to do it in an automated manner..
well no, it’s not moving the effect system over to QtQuick. It’s just allowing the effects to be written with QtQuick which is quite different from what you proposed. IIRC you suggested to use the QtQuick Scene Graph for rendering.
hehehe, i actually suggested all of it, but only because it seems so powerful to me 🙂
Either way, moving the effects to QtQuick 2 or allowing effects to be made in QtQuick 2 is enough for me to like it very much!
The reasoning behind that is that it becomes very easy to use existing shaders and try things out without much problems like compiler issues or crashing everything if you do something wrong. Not that i can even write shaders, yet.. :p
Thanks again for the top BoF. Delve into the code, the past and future of KWin, well – just really good knowledge.
See what I can do about some patches.