Alright, I started writing some unit tests for Feeds and I wanted to integrate them into meson, so that I can just run
meson test or
ninja test during build and have them run automagically.
As an added bonus, they should also run when creating a flatpak package, so that I don’t accidentally push a broken built to users.
Here’s the very simplistic way I accomplished this task:
This doesn’t change too much in terms of development, but it surely makes the app more discoverable. It also provides this very nice new URL for Feeds’ website which is under gnome’s domain name (that’s pretty cool, uh?).
As for Feeds’ development, lately I’ve begun a process of small cleanups here and there, and most importantly design changes.
Feeds version 0.8 is coming. This release have been a little overdue, since I finally decided to change the name of the app from
gnome-feeds to just
gfeeds. This has to do with the use of the GNOME name, and besides even GNOME apps don’t use the GNOME prefix in their names.
Someone criticized the name
gfeeds, because it’s kind of dumb to have GNOME/GTK app names to start with a g, similarly to how dumb it is to have KDE apps start with a k. But here’s the catch: the g in
gfeeds stands for Gabriele, I pulled a sneaky on you.
Anyway, the new name should have no problems whatsoever, and besides I learned how to transition to a new appid on Flathub.
It’s not as seamless as you would think, but it’s easy enough. You’ll have to re-submit your app as if it was a new one, and specify that it’s just an appid change. There’s some infrastructure in place to transition from one appid to another automagically, and it involves a little extra json file in the old repo. Here you can see my PR for the new appid and the whole discussion around it. Hope it helps.
My current setup isn’t perfect but it manages to automate flatpak builds. This is incredibly useful, as when people open issues and I try to fix them, I can just tell them to install the lastest flatpak from the CI/CD and test the fixes.
Also, for power users and curious minds I imagine it could be compelling to be able to install the latest development snapshot without jumping through many hoops.
Anyway, let’s get down to business, here’s the
.gitlab-ci.yml file to let the magic happen:
So I recently got myself a shiny new AMD Radeon RX 5700 XT GPU (that’s a mouthful).
Let’s see how it does on Linux.
First, I use Arch (btw) and that makes things a lot easier, at least for now. As of writing this, Navi support is still not in the mainline kernel, but it should come with Linux 5.3.
The code starts to feel a little more complex, and it would probably use another small refactoring, but that’s not too important right now.
Let’s lay out a new roadmap, with the features I implemented, and some new ones I want to add:
Alright, first I want to thank @Etamuk again. He’s being very active with the German translation, and any time I make a commit, soon after I already see he has opened a merge request with the updated translation. That’s just impressive, and I’m truly greatful for his awesome contributions!
But now let’s get down to business with the new features.
This blog is fairly new, and so I still have to set up everything. Comments are obviously a very important part of a blog, especially if you want any kind of feedback on the stuff you’re doing.
The lazy solution is pretty obvious: Disqus the privacy nightmare.
I like the wild idea of respecting my readers’ privacy, and so I started looking for alternative solutions.
The one solution I chose to adopt is an open source comment platform called Commento. It works just like Disqus, except since it doesn’t make money from violating people’s privacy, you can’t really use it for free.
Hey people. I finally got around at creating a static blog.
npm to install plugins is pretty clever, and besides at the end of the day I am the only one that has to use it.
It has a more complex structure compared to the aforementioned solutions, but if it gets the job done I can’t complain too much.