i agree with jukka that unless there are significant performance test results i would not add caching this early. i know that the path-to-item resolution is a little bit slow but maybe we should profile a "performance critical" request in detail to see where the actual problems are.
IMO the "relative-resource" resolution could probably be optimized if it would not need the resolver again. eg: rode.getNode(relPath) compared to: resource.getResolver().getResource(resource, relPath) or is this already handled nicely behind the scenes ? regards, toby On 2/17/08, Jukka Zitting <[EMAIL PROTECTED]> wrote: > Hi, > > On Feb 17, 2008 2:47 PM, Felix Meschberger <[EMAIL PROTECTED]> wrote: > > What do you think if we implement a simple, stupid path to Resource > > cache in the ResourceResolver. [...] > > The cache is solely present to be able to increase performance resolving > > resources. > > Do you have a performance benchmark that points out resource > resolution as a bottleneck? > > IMHO the added complexity of a cache can only be justified if it gives > a clear and measurable performance benefit and if there are no > architecturally cleaner approaches to solving the bottleneck. > > BR, > > Jukka Zitting > -- -----------------------------------------< [EMAIL PROTECTED] >--- Tobias Bocanegra, Day Management AG, Barfuesserplatz 6, CH - 4001 Basel T +41 61 226 98 98, F +41 61 226 98 97 -----------------------------------------------< http://www.day.com >---
