Added: 4a5
** tentative fix in tab-tree r653
Changed: 19c20
** in tab-tree
** in tab-tree - r532
Changed: 23c24
** lock(userkill{userid}) ???
** tentative fix in tab-tree, rev 655
Critical
Important
- Should not try to kill an infinite number of times or lock up when an agent causes a nick collision
- tentative fix in tab-tree r653
Normal
- Parse bug in numeric 340 (userip reply)
- Channel topics should be restored properly on initial sync.
- We should determine if a channel is empty by timestamp rather than number of users.
- Only possible, i think, with SJOIN. and I believe only on the next join! so we know if it WAS empty, but not IF empty. - tabris
- AKICK gets confused when a user is identified to two nicks, one of which has status and the other being akicked.
- Confused? i believe that get_best_acc would return the highest. which is what would occur if we used a negative acclevel for akick, as I believe we should. - tabris
- AKICK should remove any +e that applies to the user.
- Race condition when the last user in a channel quits and another user joins immediately after.
- Changing root/dropping nicks too fast should not cause races
- waiting on nickreg.nickid
- NickServ should not send a nearexpire notice on nicks that have been released from hold after they are old enough to expire.
- Every line that begins with a ! should not be counted as flooding.
- in tab-tree - r532
Trivial
- Flooding services should not trigger more than one KILL.
- tentative fix in tab-tree, rev 655
- unban help contains numerous errors.
- in tab-tree
Fixed
- Deadlock, probably in chanserv::handle_sjoin.
- Should not change ops on old users when starting up.
- ChanServ INFO should not return $err_deny
- !k should be an alias for !kick
- Bots sometimes don't rejoin a channel after a netsplit.
