[IRCServices Coding] Odd problems with Services 5.0a33
Mark Hetherington
mark at ctcp.net
Sun May 19 10:15:24 PDT 2002
As other people have posted, I seem to also have the seemingly random
segfault followed by corrupted nickname database problem (Services is
linked to Unreal). AFAICT nothing has happened nor any messages issued to
services to trigger this so I assume it is during one of Services timed
updates:
[ircservices.log] Services terminating: Segmentation fault
[ircservices.log] IRC Services 5.0a33 starting up
[ircservices.log] FATAL: database/version4: Invalid format in nick.db
Another seemingly random problem that I have seen once or twice in the past
but almost every time I start services with this latest alpha is:
[ircservices.log] IRC Services 5.0a33 starting up
[ircservices.log] sockets: [v]sockprintf() with NULL socket!
[ircservices.log] sockets: [v]sockprintf() with NULL socket!
[ircservices.log] sockets: [v]sockprintf() with NULL socket!
[ircservices.log] sockets: [v]sockprintf() with NULL socket!
[ircservices.log] sockets: [v]sockprintf() with NULL socket!
[ircservices.log] sockets: [v]sockprintf() with NULL socket!
[ircservices.log] sockets: [v]sockprintf() with NULL socket!
The number of socket errors varies from none to several.
While going through the logs since the upgrade and looking for possible
causes, I came across the following during one startup:
[ircservices.log] nickserv/main: Expiring nickname jcat-buffy
[ircservices.log] database/version4: Extension data found for nonexisting
nick `jcat-buffy'
I guess this means that expiring of nicknames is not correctly cleaning up
extension data. I am not sure what services will do with this now redundant
extension data so maybe the message could be changed to include what has
been done about it. E.g. "database/version4: Deleting extension data found
for nonexisting nick `jcat-buffy'".
I will be playing around with various things to try and get more info on
the problems, but wanted to report them ASAP in case there is something
obvious that could be causing them.
--
Mark.