One of the things that makes it really difficult to create
shishi parsers at the moment is the fact that you create a node which has to jump to the next node in the chain. For instance, given
ab, you could write a node which matches
a which then goes on to the node which matches
b. Unfortunately, when you create the
a node, you haven't created the
b node yet. So you end up with a dangling pointer and have to backpatch and it's ugly as all hell.
Things have also been pretty stressful since our main firewall failed on Saturday, cutting off all traffic in and out of the network. We've had operators periodically rebooting it ever since, and it became apparent that the “4 hour response” we got from our suppliers was rather useless. Although it took us until Monday morning to ascertain that it really was a problem with the firewall and not our fault, it's taken until late this afternoon to get a working replacement. We even got so desperate as to knock up our own hotspare with an OpenBSD box, two gigabyte ethernet cards,
ipf and lots of Perl magic. But that's probably a success story for another time. Of course, setting up our own replacement made the real PIX quite happy…
I was not too happy with the idea at first, but I became surprised at how well it worked out. For starters, we got a particularly good day - and it looks like today's going to be another great one. Second, Charlbury is an amazingly pretty little village; it's a very traditional Cotswolds village - there were even people playing cricket on the village green as we walked up from the station. It also helped that Charlbury Beer Festival is much more of a family event than most other festivals; there was a band playing, a barbecue, and plenty of people came with their children so it was a good day out all round.
I grabbed the CVS HEAD kernel and started building it; I didn't get particularly far, because after correcting a few simple syntax errors, I found that the build process required the
-fapple-kext option to the compiler. Um, the what option? Time to check out a new version of
cc from the Apple CVS. Only to find that that didn't define the
-fapple-kext option either. Wonderful.
Simon Myer's talk on our very own Request Tracker was particularly funny; Marcus Brinkmann gave a great presentation on GNU Hurd, something that may be worth looking into in a while. Dave Jones and Marcelo Tosatti were there, talking about Linux kernel maintainance; I intend to bug them further about potential use of RT for kernel patch tracking.