hunt_server() with wildcard hits wrong server

[df7cb]

/quote shedding quantum* 60 0 :Server maintenance
!oxygen User shedding ENABLED by Myon (~myon@myon.chair); interval: 60 seconds, opers: no, reason: "Server maintenance" [helix]
/quote shedding quantum* off
!oxygen User shedding DISABLED by Myon (~myon@myon.chair) [helix]
/quote shedding quantum.oftc.net 60 0 :Server maintenance
!quantum User shedding ENABLED by Myon (~myon@myon.chair); interval: 60 seconds, opers: no, reason: "Server maintenance" [helix]

[df7cb]

More weirdness:

/motd kin* -> oxygen
/motd kinetic* -> helix
/motd kinetic.o* -> kinetic

[df7cb]

So thanks to rnowak and me continuing to wtf over some code bits, we found the explanation: There is a user QuantumBot connected to oxygen, and a user kineticrootkit[m] connected to helix which made hunt_server find the user first before applying the wildcard to server names.

We should fix that to look at server names first, or remove the user lookups alltogether in the places where it doesn't make much sense (shedding, motd, ?).

[cryogen]

That's just kind of the way it works. It's always better to use fully qualified names instead of wild cards when doing remote commands

[Adam-]

FWIW seems upstream already has this https://github.com/ircd-hybrid/ircd-hybrid/blob/master/src/server.c#L186-L217