Spent far too long today wrestling with the Extras buildsystem. It looks like there's some sort of incompatibility with the current libaudit stuff and the earlier kernels that were running on the build machines. Upgrading them to the current update kernels seems to have made the problems mostly go away. There's still one odd hang that I haven't quite figured out. Hopefully dcbw will have a better idea.

Did a little bit of prodding of my deadlock on current xen-unstable with full disk vbds, but no luck tracking it down yet. Rusty says that's his usual test configuration, so it at least works for some people some of the time. That makes me feel a little bit better about spending time debugging it.

Also on tap for tomorrow/this week, I need to sit down and think a little about package selection and what it should really look like — the yum backend for anaconda gives me a good chance to rework some of this stuff so it'd be nice to get rid of some of the current warts. Suggestions from the peanut gallery can be left as usual. I have to wonder if the current group paradigm is most of what's wrong… it's pretty similar to the way “Another OS” does things but we really have a far different situation with the greater granularity in packaging. We also have a lot more functionality included, so the same sort of thing starts to break down a little bit.

This evening, sat down to figure out why my iRiver wasn't getting a fstab entry created (since I want to sync some new music to it…); turns out it's due to some oddity in the hal fdi files, so updated the bug. It's like a whole different world to me after years of (bad) kudzu hacks 😉 Even so, I agree with luis that a general solution is much better than every distro doing it their own way.

And now, it's time to continue working my way through all of the South Park episodes released on DVD thus far…