e.g. irc.mozilla.org [14/10/2014 11:58:21] ":fripp.mozilla.org NOTICE aleth-build :*** You cannot list within the first 60 seconds of connecting. Please try again later." [14/10/2014 11:58:21] ":fripp.mozilla.org 321 aleth-build Channel :Users Name" [14/10/2014 11:58:21] ":fripp.mozilla.org 323 aleth-build :End of channel list." So we get a correct 321,323 pair, but that means we won't reattempt LIST for another hour or day or whatever the constant is.

This is indicated by ISUPPORT SECURELIST ("This InspIRCd specific token indicates that secure listing is enabled, and that your initial request for a channel list on connect may be denied until you have been connected for a certain amount of time.") https://github.com/inspircd/inspircd/blob/master/src/modules/m_securelist.cpp

Comment on attachment 8513432 [details] [diff] [review] securelist.diff Review of attachment 8513432 [details] [diff] [review]: ----------------------------------------------------------------- How hard would this be to add a test for? Let me know about the parseInt comment... ::: chat/protocols/irc/ircNonStandard.jsm @@ +56,5 @@ > + // pair, but no list data. > + // We fake the last LIST time so that we will retry LIST the next time > + // the user requires it after the interval specified. > + const kMinute = 60000; > + let waittime = (aMessage.params[1].split(" ")[7] * 1000) || kMinute; Doesn't this need a parseInt or something around it? Nit: waitTime @@ +57,5 @@ > + // We fake the last LIST time so that we will retry LIST the next time > + // the user requires it after the interval specified. > + const kMinute = 60000; > + let waittime = (aMessage.params[1].split(" ")[7] * 1000) || kMinute; > + this._lastListTime = Date.now() - kListRefreshInterval + waittime; This is me just being pedantic, but can you do now + waitTime - refreshInterval? It makes more sense in my head. :)

(In reply to Patrick Cloke [:clokep] from comment #3) > How hard would this be to add a test for? The test we would want here is to detect whether inspircd have decided to change the string they send us, and we can't test for that. > > + let waittime = (aMessage.params[1].split(" ")[7] * 1000) || kMinute; > > Doesn't this need a parseInt or something around it? Nit: waitTime It's not needed as the multiplication enforces type conversion.