Hmm... Ok. First -- thank you everyone who has replied. Isomer/stoney`/Richard/Kev -- and whoever else replies after this of course.
Some response to Kev inline. On Tue, 2003-08-19 at 20:28, Kev wrote: > X is able to see through this mask, always, because of the way it's > connected to the network. However, we have to take special care with > X to prevent it from revealing the user's true mask. > Agreed on the special care. Hadn't realized X is able to see through the mask -- but I did test Richard's recommendations and do see that it works now. Thank you. > Unfortunately, you're quite right--this is a *huge* mess. You not only > have to do it on the original /mode, but also any subsequent net junction > mode changes and any usage of /mode b. There's simply no way to represent > this information in the protocol, and there would be a *lot* of overhead, > in terms of memory usage, to store that data. I'm afraid we can't do > this. Ewwww.... Ok. Perhaps my mind oversimplifies things a bit too much. Since X always sees through /mode +x it seems that X could set a ban on the masked host but carry through with that ban when the individual is unmasked. So internally it adds a ban to the full usermask but doesn't display it in any of the ban listings -- until the user returns without masking their host. Of course this all runs back to the same problems that make it so when you ban a named address and the person comes back with their numeric ip instead -- or visa versa. Bah. > > If it's any consolation, if the user is logged in, even if they're not > +x, bans on [EMAIL PROTECTED] work. Of course, that can be > circumvented by not logging in, but that's why we provided the +r channel > mode, which prevents not-logged-in users from joining the channel unless > they're /invite'd. It's not an ideal solution in some situations, > perhaps, but I'm afraid it's the best we can do at the moment... > Thanks.. hadn't realized you'd added +r. Unfortunately most of our users do not log in. > Also note, by the way, that bans on the user's real host mask are also > effective even when the user is +x. I see that now. Thanks, Jay