|
From: Rod J. <rod...@in...> - 2003-12-10 20:13:21
|
I've just been writing a chapter on performance and thinking that a generic caching interceptor would be very useful. Ideally this would allow for pluggable implementations, even using a fancy distributed cache if necessary. But a simple caching interceptor would be a good start. Any thoughts on this? Sounds useful? I think treating caching as an aspect, and having an efficient generic implementation totally distinct from application code could be very useful. We could use pointcuts to target caching, but might also need additional information on a per-method level. E.g. how long the return value could be kept for a particular method. Source-level attributes could help a lot here, but unfortunately we're not there yet. Has anyone implemented something like this? Regards, Rod |