Extending this thread, while have anyone to opine it. I know that only the content (properties and binary under nt:resource) are automatically indexed by jackrabbit for searches, and of course these contents must have be in the repository for the method get works... I would to know if its possible to extend the jackrabbit (core maybe?), to do an index for contents that in fact are represented by a link to. Maybe the implementation get the real content of the file (represented by link) locally for reasons to run the indexer and not store this content into the repository. In case of a query search, in the results will be the value under jcr:data with the link to the file. Like google, you know...
Tks Helio >Let me share my dilema with you. >We have an CVS repository, where we put the project files and documentation in. >I have to make a repository that will have the contents the applications will >need, but for reason to do not duplicate the files already in CVS and that >have got to be available in the repository (visible by the application like >the files were in the repository, but they are in CVS). >If the problem was only this, ok, only just creating a propertie with the link >to file in CVS will at first solve it. But these files that is visible by >application repository and in fact are in CVS, must be fulltext searcheable.... >Any thoughts for this case? Can I do a index for these files, as like they >were in repository, and by a query in jackrabbit I would find the files that >has the content I am looking for? > >If the case was not clear for you I can rebuild it. I am not soo good for >explain, you know... > >Thanks >Helio. >
