> I was wondering what the exact meaning of a HACK(2) notice is supposed to > be.
The meaning of HACK(2) changed between u2.10.10-series servers and u2.10.11 servers. Previously, if a user (not a server) issued a mode command that the server believed they should not have issued, the mode would be bounced, but no notice was sent--this is a HACK(1). With u2.10.11, since all modes from servers are accepted, I chose to make this a HACK(2). Though it's possible this was an incorrect judgement, it doesn't really matter. > Basically the service user is sending a CREATE for a channel it wants to > join (because it hasn't read the burst from the hub ircd yet and so doesn't > know the channel already exists). AFAICT this could perfectly well happen > on normal servers (create a channel while servers are bursting which then > gets bursted from the far side). From the service's PoV it knows that the > user "C" is no longer opped once it has parsed the burst message because it > saw the earlier timestamp. > > However, the ircd says: > > HACK(2): C MODE #twilightzone +o C [989986810] > > and sends a mode back the other way deopping the user. Is this expected? > Could the ircd not just convert the CREATE to a JOIN when propogating it and > not send the HACK notice? (it knows that the service hasn't yet acknowledged > the burst)? This is indeed expected behavior. The server is transmuting the CREATE into a JOIN, it's just not being silent about this anymore. -- Kevin L. Mitchell <[EMAIL PROTECTED]>
