DO NOT REPLY TO THIS EMAIL, BUT PLEASE POST YOUR BUG RELATED COMMENTS THROUGH THE WEB INTERFACE AVAILABLE AT <http://nagoya.apache.org/bugzilla/show_bug.cgi?id=17623>. ANY REPLY MADE TO THIS MESSAGE WILL NOT BE COLLECTED AND INSERTED IN THE BUG DATABASE.
http://nagoya.apache.org/bugzilla/show_bug.cgi?id=17623 Incorrect caching behaviour [EMAIL PROTECTED] changed: What |Removed |Added ---------------------------------------------------------------------------- Severity|Critical |Blocker Status|NEW |ASSIGNED Version|Current CVS 2.0 |Current CVS 2.1 ------- Additional Comments From [EMAIL PROTECTED] 2003-05-30 06:48 ------- Ok, I think I can confirm now, that this is a blocker :( The caching algorithm is based on key/validity pairs. The problem now is, when you use an internal pipeline call, the key is always the same e.g. "cocoon:/a.portlet", regardless of what the pipeline does. The validity for this pipeline is based on the validities of the sitemap components used in this pipeline, so in this test scenary it's the timestamp of the xml files. As they both have the same time stamp, there is currently no way to tell from the outsite that something different is happening. Your suggestion using getKey() instead of getURI() is the right track, but it violates SoC. A Source object should not care about caching too much. One solution is to code extra information into the validity object. This is possible and not too complicated I think - so you won't get wrong content. But in your case this means that always only one language is cached (the last one invoked), so you have to rewrite your pipelines. Doesn't sound got to me. I'm thinking about it...
