That seems to be the right semantics to me. If a user doesn't have
permission to make a call, we wouldn't want to call any pre-hook. However...
I'm looking at RSRpcServices to compare and it looks to me that having
the authz check outside of the AccessController CP is weird. For the
methods
Hi,
I was looking at RSGroupAdminServer#removeRSGroup :
if (master.getMasterCoprocessorHost() != null) {
master.getMasterCoprocessorHost().preRemoveRSGroup(name);
}
However, RSGroupAdminServer#removeRSGroup is called by RSGroupAdminEndpoint
:
