The JCR model is overly complicated and doesn't really provide value - for example, JCR provides versioning, but I had to write my own versioning anyway because I couldn't implement existing JSPWiki functionality with it (I don't remember specifics, it was many years ago). JCR versioning is meant for something else, obviously.
Even stuff like pre-importing sample pages is hard, 'cos the on-disk format isn't standardized. You need to write your own importer. /Janne On 13 Nov 2012, at 08:42, Florian Holeczek wrote: > Hi Janne, > >> I wouldn't recommend using JCR anymore for JSPWiki. It is, as you point out, >> overly complicated and many of its features (like versioning) are ill suited >> for JSPWiki purposes. Also I'm no longer developing Priha, so someone would >> need to pick that up too. >> >> An redesigned version of the current enhanced with metadata and >> subpage/subwiki support should be sufficient, something like switching a >> WikiName class to everywhere that Strings are being used to denote a page, >> then having it support a simple [subwiki:page1/page2] syntax... > > overly complicated is true unfortunately. However, that's mostly the server > part's problem, isn't it? Did you face any bigger problems on the client side? > If there was a fast and lean server already existing, ready to be packaged > with JSPWiki, it would be quite comfortable to use JCR. That's why I'm really > curious about Jackrabbit OAK. > On the other hand, looking at the amount of resources available, it may be > the better decision to enhance the 2.x API regarding your points above. > > Regards > Florian
