View Issue Details
|ID||Project||Category||View Status||Date Submitted||Last Update|
|0006059||unreal||ircd||public||2022-01-18 23:31||2022-01-27 16:41|
|Summary||0006059: Option to block connections that doesn't have a VERSION reply.|
|Description||Although unusual, sometimes some spambots do not have a VERSION reply.|
Currently we can deny connections based on a string(s), but when comes to an empty reply we don't take any actions.
Would be nice, probably, to have a way to also deny connections to clients that do not reply to our VERSION request.
This is because of this on IRC, and it's not the first time I see someone asking for a ways to do so:
22:22:45 <jFm> Hi 22:24:42 <jFm> lets say there are some bots that dont reply ctcp version 22:25:18 <jFm> How can a solution be found with spamfilter? 23:26:58 <jFm> Spamfilter add u zline +24h bot ^(.+![\w-]*[-_][\w-]*@.+:.+)$ 23:27:18 <jFm> What am i missing here?
|Tags||No tags attached.|
|3rd party modules|
I've experimented with that, and there's one major question: how long would the server have to wait for a reply? It is known for clients (for example some instances of irssi) to reply after 80 seconds or more. What should happen before that time? Block the client in the registration phase? Allow them to act normally and kill after the timeout?
Also, some spambot operators fairly quickly learned to send a reply, and later to send various replies gathered from real world when the initial one was blocked.
This has been suggested several times in the past 20 years. I have always been against this - and I think i still am - because IMO it is perfectly OK for a client not to reply to CTCP VERSION. Lots of people configure their client like that (including myself) for flood or privacy reasons. So you will often ban innocent people. And, yes, in addition to that also comes the question on how long to wait for a reply and what to do in the meantime, k4be mentions that clearly.