Leaking memory is impolite. It’s messy, it can suggest logic bugs, and thanks to AI grifters RAM is expensive.

Unfortunately C++ makes it rather easy to leak memory. Fortunately we have tools to find such leaks. One such tool is Leak Sanitizer (LSAN) from the Address Sanitizer (ASAN) family. It’s using compiler-based instrumentation for the code to reports any leaks after the program terminates.

KDE’s CI infrastructed has ASAN enabled for a while. However the leak sanitizer part was explicitly disabled, so no leaks were reported as part of the CI build and test run. This is because a number of projects have pre-exisiting memory leaks that would cause the C build to fail. Of course those should be fixed eventually, but in order to do that we need to know where they are. Also, for projects that currently don’t have any leaks we want to enforce keeping it that way.

For many years Plasma comes with its own system online accounts system, known as KAccounts. The idea is simple: In Systemsettings you log into a given online service once, and then several applications can use that login, instead of loggin in inside each application separately.

The number of services available and applications making use of them changed a bit over recent years. As of right now the following services are supported:

A few weeks ago Qt 6.8 was released, delivering many fixes and improvements for our software. Some of them were contributed by yours truly, and in this post I want to highlight some of them.

They relate to graphics tablet/stylus input on Wayland. Before we go into the fixes let’s have a quick overview of the flow of tablet input events on Wayland:

The genesis of input events is in the kernel driver for the particular tablet, which talks to the hardware (via USB, bluetooth etc). The kernel passes events to userspace via the evdev system. On the userspace side a Wayland compositor reads those events. Most compositors leverage libinput for this, a library that transforms raw evdev events into something more usable for the compositor while applying some cleaning and configuration. The compositor then delivers events to native Wayland applications using the tablet-v2 protocol. For legacy XWayland applications the compositor uses this protocol to send events to XWayland, which translates them into things X11 apps understand. The compositor can apply useful transformations to the input, like which area of the screen the tablet is mapped to, changing the pressure curve of the pen, or binding buttons to keyboard shortcuts.

This year’s Akademy was a special one for me in many ways.

First of all, instead of me travelling to Akademy it took place in my hometown of Würzburg, Germany. While I did have a hand in organizing it, most of the credit for it goes to Tobias and David. I had a lot of fun introducing people to my area and the concept of drinking wine on a bridge.

In a previous post I talked about using the QML Language Server for KDE development. Since writing that post a few things happened, so it’s time for an update.

I mentioned that when using Kate qmlls should work out of the box when opening a QML file. That’s mostly true, there is one problem though. Depending on your distribution the binary for qmlls has a different name. Sometimes it’s qmlls, sometimes qmlls6 or qmlls-qt6. You may need to adjust the LSP Server settings in Kate to match the name on your system.

Last week, like many other people, I was in Berlin for the Mega^WGoals sprint. Natually the three goals (Sustainability, Accessibility, and Automatability; the three abilities) attracted a diverse crowd of people that brought also other topics, so it turned into a proper megasprint.

Being interested in all three goals made it a bit callenging to follow all relevant discussions unfortunately, but on the flip side it never got boring.

Most of my goal-related work was towards the automation goal. One thing I was working on is a CI job that checks for spelling mistakes in the code, so that those can be caught when doing a MR. A while ago I create a website that tracks which KDE projects are ported to Qt6. What started out as a joke for a talk turned out to be a useful tool for planning porting work. During the print I fixed the site to actually work correctly again and, most importantly, changed the text from “No” to “Yes” since most projects actually use Qt6 now. For a while the site also has auto-generated reports for other things, like showing which projects don’t yet have clang-format applied to them or which projects don’t enforce passing test in the CI. I used the latter list to enable this for the remaining projects that don’t have failing tests right now and prepared a change to the CI system that enforces passing tests by default. In the same spirit others and I also fixed some currently failing tests. We also discussed the idea of extending the site with more checks and turning it into a proper KDE site that isn’t hosted on my personal infrastructure.

For a while Qt has been offering qmlls, a Language Server Protocol implementation for QML. This allows popular text editors like Kate (and some lesser known ones like Visual Studio Code) to work with QML code without having to write dedicated support for it.

Naturally many people are eager to use it to hack on KDE code. When trying to do that you probably have encountered some frustration with things not working as expected. So what’s going on there, and how can we make it work?

KNotifications is KDE’s framework for creating popup notifications. It supports Linux, Windows, macOS, and Android, making it, to my knowledge, the most complete cross-platform library for this available in C++. This makes it natually interesting to use for non-KDE Qt application developers.

However there is one aspect that makes it less attractive for third-party developers: It’s number of dependencies. As of KNotifications 5.110 it depends on the following other KDE Frameworks:

As you probably have seen from other people’s blog posts there was the 2023 Plasma Sprint last week. It was generously hosted by TUXEDO Computers in their offices in Augsburg, Germany. Many thanks to TUXEDO for that!

Other people have already well summarized what happend there, so let’s have a look at what I have been doing:

Together with Kai Uwe, Volker, and Ismael I looked at notifications. This includes internal simplifications in KNotifications, API design questions, a proposed V2 for the notification portal API, and a new UI for per-event configuration in the notification settings module.

A month has passed since my last monthly post about my work as KDE Software Platform Engineer. What have I been up to since then?

As usual not everything I did ended up as committed code. A lot of my work is reviewing other people’s code, discussing ideas, and generally being useful to the community.

One area I’ve been focussing on is our infrastructure for global shortcuts. These are currently handled by the KGlobalAccel framework. This frameworks contains the runtime component that manages global shortcuts as well as an application-facing library to interact with the runtime. The runtime and interface library being in the same project has caused us some issues in the past. To address these the runtime part is now split out into a separate project and part of the Plasma group. You can read up more details on this here. This change also allowed for some further changes that made launching applications from global shortcuts both simpler code-wise and more robust. I have also worked on another set of changes that will address some of the pain points that are currently present in our global shortcuts system. While doing that I also addressed some issues I found in the underlying KDE Frameworks libraries.