Hack, hack, hack.
I must say that ircd code is horrible. And I like to think that I know something about horrible code… I mean, there was fixing nxpr, fixing xtracs, OpenAFS, random bits of other code written by MIT or CMU. So, horrible code I've seen. But dancer-ircd manages to take this to a new level. Most of it isn't the fault of the OPN guys, I mean, lots of these defines had to have been in the base hybrid distribution or they wouldn't be there at all. Suffice it to say that it's a mess. Going through and peppering the things with fixmes and then going through and trying to fix them is probably wise though, as there are some things there that are bound to cause problems eventually otherwise. Not today, not next week… but in six months. Most of it should be trivially fixable though. Good “spare time” project for the next day or two.