I had a very productive weekend from a hacking perspective. First, I calculated md5sums for my collection of doom installation media and updated the doom-data page appropriately. I also happened to notice a bug against freedoom which would be solved by doom-data in addition to the existing prboom one.

I learnt about the debian alternatives system, by patching the freedoom package to use alternatives for doom2.wad as a precursor to doom-data being packaged (327500). I was very impressed with how simple (yet effective) it is. I note that redhat has borrowed the alternatives system some time ago. I wish they'd use it more often: Compiling xen stable on FC4 is very difficult, as the 3.x version of gcc is not /usr/bin/cc (and the makefiles redefine CC a few times in different places).

I also learnt a bit about how debconf works. I threw together some template files, taking inspiration from the quake2-data package, and using the suggestions in the debconf-devel(7) manpage to test them out. I was really impressed with debconf. It seems there's scope to do some really powerful things that nobody does yet.

Finally, I've been taking a look at deutex 4.4.902. It's sufficiently different from 4.4.0 for me to have to re-work the debian package quite a lot - however this is a good thing. One of the additions is a new -rate option which allows you to control what happens to sound files which don't have the normal sample rate. We've been using a patched version of deutex to achieve this in freedoom's build process, and now that patch will no longer be necessary.

The result of that is we shall be able to provide a proper source package for freedoom in debian in the near future.