> > Yes, but it complicates an existing very simple plugin. > By a few lines. I find that much easier to accept, in terms of complexity, > than subclassing the plugin and introducing a new config file.
It's not the number of lines as much as the new logic. (I'm not saying the logic is complicated, but it takes a very simple plugin and makes it less very simple.) > > A compromise is still two files - with an additional argument to the > > plugins file for DENY or DENYHARD. > > Or instantiate the same plugin twice and make the failure code and > bad-address file configurable, yes. I think thats what I meant. -R