refuse to honour RSFNC for a RESV nick

[jesopo]

No description provided.

[dwfreed]

This breaks keeping a RESV for Guest* so that the only way a person can have a guest nick is via services. OFTC currently has such a RESV, though I suspect part of our reasoning is that our guest nicks are sequential and we do not have SAVE, so a determined attacker could kill a guesting client by colliding with their nick change.

[jesopo]

we could solve that by having 2 kinds of RSFNC. one that is expected to sometimes fail and one that should never fail, and you'd use the former for any user-supplied nickname

[dwfreed]

Atheme already allows differentiating between a regain and a force change via the fnc_sts handler; FNC_REGAIN is the type value for regain, and FNC_FORCE is the type value for force changes. If we pursue that route, I think RSFNC should remain the FNC_FORCE handler, and we could add something like ENCAP REGAIN for the FNC_REGAIN handler, which would require compliance with all applicable nick rules, just like NICK

[jesopo]

atheme work done here atheme/atheme#864 so now we can think about how this should look over the wire