The first Libre Application Summit in Portland just finished this week and I had the chance to go.
This was the first ever conference I’ve been to and it was great to meet everybody I knew from online in person such as Christian Hergert, Garrett Regier, Sriram Ramkrishna, and so many more who were all very welcoming. Portland was also a great setting for the conference with lots of beautiful sites and I hope to be able to visit again. With that out of the way onto more technical topics.
So going into this I already had a head start on Flatpak; I have already packaged most projects I maintain and already host a repo for them. It was great to see how everybody else was equally sold on Flatpak though especially Andrew Walton’s talk about VMware’s current distribution nightmare and how they see Flatpak as the clear solution to that.
For those of you who do not know Meson is a new build system which has made some noise in the GNOME community over the past few years. I have a more positive opinion of Autotools than most so I was a bit skeptical that Meson could replace it entirely. Jussi Pakkanen gave two talks over Meson and speaking with him over the week I am more sold on the idea of migrating to it.
So one major pain in building is Windows; Currently the HexChat Project maintains a custom build script written in PowerShell which is as bad as it sounds. Meson has the concept of Wraps which allow easily building an entire stack from source which is valuable on Windows. I have yet to use it in practice but it is certainly on my todo list.
I have occasionally contributed to GNOME Builder and as an IDE a difficult task is to sanely parse and modify an Autotools projects. Meson unlike Autotools exposes JSON data for compilation flags and targets which can allow tools to get actually usable information. I have already started work integrating this in a branch that is a work in progress but already it can build and run projects. The internal API of Builder is a bit Autotools focused in this area, there are some Python binding problems, and we need to work with Meson to expose more information but it should land this cycle.
I did briefly discuss IRC at a BoF mentioning the current state of initiatives like IRCv3. The end conclusion is just that we need to get in contact with the infrastructure of GNOME to push for more modern server features and possibly expose things like chat history which is already doable at the protocol level.
There was a BoF over how to monitize applications and I was honestly disappointed by this discussion. Technologies like Flatpak and the future website of FlatHub, a hosted service for building and distributing flatpaks, help this but I don’t feel like any new useful strategies were brought up and the issue of getting the community at large to want to donate/pay, especially for existing projects, is a bigger blocker than some realize. This is just something that will need to be discussed further.