Paul Libbrecht wrote: > > Le 18 janv. 08 à 00:27, Sergiu Dumitriu a écrit : > >> Jerome Velociter wrote: >>> AFAIK, xwiki .js files aren't parsed by the velocity renderer. You >>> might find JS mixed with velocity only in templates inside <script> >>> tags. > > That's how much I could see with less and grep in albatross and would be > already very relevant. > >> Not exactly, some of the JS files are parsed (those that get through >> the SkinAction). > > These are the ones I would like to know about. > >> But I don't think that is an issue, they can be cached. > > It is an issue. > In mod_jk, it is classical to configure the "static" resources to be > fully ignored by the servlet container... no caching, just normal apache > httpd download work (with file modifs headers, only-if-changed-since > responses etc). > > Caching is an alternative for performance management, and it relies on > the proper usage of Vary headers... report on experience there is > interesting as well. > >> AFAIR, there is no (outer) script that really depends on the current >> request. > > Then for performance that script should be kind of isolated. > In an ideal solution, all that would be under xwiki/static and it'd be > easy to take part into httpd (which may well be on another host). Right > now, I think that xwiki/skins would already be a good first step. > > If Have an answer to question 1 (or a way to track the answer on > individual files, I think I could let experiments be made... >
How about: - scripts that are taken from /xwiki/skins/albatross/*js are static = scripts that are taken from /xwiki/bin/skin/.../*js are parsed Sergiu _______________________________________________ users mailing list users@xwiki.org http://lists.xwiki.org/mailman/listinfo/users