while that would likely break login for many servers, one could argue that this is simply a matter of reading the server feedback (typically a notification of illegal characters in the gecos) and then adjusting the value of it - I think though that because the login information is essentially program wide rather than per-server, it might not be a good idea to stick UTF8 in places where it is guaranteed to cause problems most of the time by virtue of almost no servers currently natively supporting unicode (at least for the time being).
now that every OS has unicode enabled major clients (and it should be fairly obvious that xchat for windows and trilliant etc constitute a windows minority), let's see if the server world decides to move to unicode enabled IRCd's first. This will probably not happen until mirc 6.17 has been debugged and a stable confirmation of its unicode capabilities in the form of 6.18 is released I think.
