> Jukka Zitting wrote:
> >> Writing basic-search isn't to hard if I could translate it
> directly
> >> into lucene queries, but...this of course it not a very nice thing
> >> todo, because jackrabbit has this query grammar sets, but
> I really do
> >> not know how they work (not sure if i want to learn about
> javacc) :-)
> >
> > How about exposing a "Lucene" query language at the Jackrabbit
> > QueryManager level? This way a Jackrabbit-specific client
> could bypass
> > the XPath or SQL parsing and issue queries directly against the
> > underlying Lucene index.
The problem with this approach is that a client shouldn't know about the
lucene index structure. For example, if JCR-1080 is implemented, we
don't want broken clients.
>
> I don't think that's a good idea. besides you don't have to
> know javacc to write a new query syntax for jackrabbit. your
> syntax implementation just need to transform a string
> statement into a org.apache.jackrabbit.core.query.QueryNode tree.
That is something I thought of, but not sure wether it is possible to
translate everything into a QueryNode
>
> AFAIK Julian did something similar in his spi implementation.
I might take a look at that later on, thanks for the pointer,
-Ard
>
> regards
> marcel
>
|