SYA 710 Kernel Compile Lab

It's a week or two late - I've been busy with the Func project, to the exclusion of much else - but after a few attempts the kernel compilation lab is done. I grabbed the freshest 2.6.27 off kernel.org, grabbed the required packages, and off I went to spin up a new kernel. There were some initial problems with one step in the compilation of some particular component ('no rule to make target 'n/n' on a file 'ngen.so''), but a bit of Googling quickly turned up a solution. The solution (which involved simple answering a single question about external firmware loaders differently during 'make menuconfig') was also from a Fedora Core 9 user, so perhaps there's a bug somewhere there that neds to be worked out. One of the later commands in the course notes regarding kernel compilation, the syntax offered for 'mkinitrd' in particular, did not work for me, and I've added a note with the syntax that did to this page of the wiki.

Byebye badware.

Woot. Just finished removing everything related to Mono from my main Ubuntu system. Feels good.

Func Project Report 1

So, after some time of thought, high-level study, and waiting for decisions to be made regarding exactly who's writing exactly what, real work on actually getting things done with Func seems to have begun in earnest.

Chris Tyler cooked up and cloned a small handful of Fedora VMs to use for our development and testing network, and so tonight I've taken to setting them up and trying to get all the pieces in place for a working Func environment - luckily, there don't seem to be too many. Installing Func itself was a breeze, but upon moving to actually setting them up to interact with one another it seems they expect a working DNS system around them - so, my current task is setting up bind. Luckily, that's something I've done a few times, so doing it again should only be a brief distraction from setting up Func itself.

I'm still trying to get a firmer grasp on the Python language itself; while most of the concepts are familiar ones, I'm finding the actual syntax to be somewhat unappealing - to my eye, at least, it just doesn't read very well, and the flow seems somewhat unnatural to me. Nevertheless, while it seems unlikely that Python will become my language of choice, it's nothing a bit more time and practice won't solve.

I've been skimming the actual Func code, the modules in particular, and while I haven't quite attained complete understanding of all the intricacies yet, my grasp is growing better the more I read.

Back to bind.conf for me.

SPR720 Packaging Lab

So I've just completed the SPR720 packaging lab. For my package, I'd chosen Csound, a C-like language for audio synthesis. The .spec file required some modification to be suitable, as Csound uses a Python-based build system called 'scons' and an accompanying 'install.py' script rather than the usual './configure && make' routine, but this part was fairly minor. Some additional modifications were required to compensate for Csound's tendency to, by default, place several of it's executables in '/bin' instead of '/usr/bin'.

All in all, the process wasn't unusually dificult, but it might be nice if the documentation were given a bit more polish - maybe a seperate, more rudimentary tutorial on the procedures with more examples to accompany the main document.

The RPM files produced can be found here here.