Recently, my sons are puting me under pressure as they want to play games, and I refuse to reconnect the WII. I refuse because this system is closed beyond repair: I am not even allowed to install free software on the material I've bought. That's not strictly forbiden, but each time I'll update the system (which happens when I run a new game on the console), every program that is not signed by the material constructor will get erased. No, really, my sons won't play with it anymore if possible.

But then, I have to feed them with games on Debian to divert them from the WII. That is why I updated the frogatto package around chrismas, so that they can enjoy the latest version.

Similarly, I updated the minetest package last week, and it just entered Debian unstable last night. This is version 0.4.6, released on April 3. while the previously packaged version was 0.3.1, released at least one year ago. The minetest community is very active, with about one release per month, with many changes each time.

Most of the packaging work was already done by Matthew Bekkema, so I mainly had to sort his patches and complete it. Also, I made the package compatible with git-buildpackage to ease the packaging of future upstream releases. I had to reset the git repo for that, unfortunately, but it will worth it in the long run.

Another difficulty was that upstream does not distribute tarballs, but simply tag the code on github. This is usually very similar, but since they split their code base in several github repo (to keep their size reasonable), I had to repack their source to merge the repos back. I don't like messing with upstream tarballs, but it sounds like I had no other possibility here.

Now, I can updating the package for the next version should be very simple, and in the meanwhile, my sons are very happy with this version. I am considering fixing an internationalization bug, where only the C++ engine is translated while the lua mods cannot be translated yet. There is already an attempt to i18n the mods, but it seems very difficult to put into production. I think that it'd be better to use classical tools to maintain translations as this approach.

Another interesting idea would be to come up with a mod or a modified client where you can specify the code of several robots working for you. I'm dreaming of a blockly interface for that, it would definitly rock!