2008/7/4 Kinkie <[EMAIL PROTECTED]>: > I agree. Maybe even more structure and data can be shown than what's > in SNMP, and that's exactly the kind of things I'd like to think > about.
Well, you should think about it as layers. There should be a way to pull data out of Squid and export it somewhere. That can be one "layer". A hierarchical structured type (like SNMP) is probably good enough for representing anything we want to. There then should be a transport layer for getting it -out- of Squid. This should be HTTP (avoiding XMLRPC, pretty please!) and SNMP. Finally, you can do whatever you want for presentation layer. If the SNMP code was made _Much, Much Easier_ to understand - including the tree definition/iteration/registration code - then it would be a lot easier to add new SNMP variables. You could then just split the SNMP transport from the underlying data gathering method and actually write a simple HTTP layer for accessing it. > How would everyone feel about changing the cachemgr (possibly breaking > backwards compatibility) so that it can render better an high-level > view of the dataspace, while leaving detailed views (possibly even > more detailed than now) to snmp, xmlrpc and/or structured (but > user-unfriendly) html / plaintext displays? I reckon you could just drop modifying cachemgr as it stands and look at ways of sorting out a proper way of registering reporting modules in the above described SNMP-like way, and then sort out a way of separating out the transport layer. You'd then suddenly get HTTP access for all of the SNMP exported data. Then all of the existing cachemgr information could be represented in this fashion and your HTTP frontends can use that. Adrian
