jackrabbit-users mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Darren Hartford" <dhartf...@ghsinc.com>
Subject <basicsearch> [was: Any commercial real world implementations out there]
Date Fri, 18 Aug 2006 14:04:56 GMT
I had made a request to the mailing list several months ago about
mapping a <basicsearch> query to an Xpath/Xquery, but there were no
responses (probably because it is two different domain knowledge areas
that don't normally cross-pollinate).

This does bring up back to the original question - how to enable
Jackrabbit to do DASL <basicsearch>.  As mentioned by others in the
past, two possible approaches could be:

*Map a DASL <basicsearch> to an equivalent JCR/Xquery-Xpath query and
re-map the results back. This could be shared with the JCR expert group
if this works out, or share deficiencies related to JCR-implementations
with WebDAV interfaces.

*Look at a writing a new QueryEngine/Handler
(http://jackrabbit.apache.org/doc/arch/operate/query.html) specifically
for DASL <basicsearch>.

Based on Julian's feedback, it may be that option 1, mapping to an
equivalent Xpath, is not be a good approach after all.


> -----Original Message-----
> From: Julian Reschke [mailto:julian.reschke@gmx.de] 
> Sent: Saturday, August 12, 2006 7:13 AM
> To: users@jackrabbit.apache.org
> Cc: Darren Hartford
> Subject: Re: Any commercial real world implementations out there
> 
> David Nuescheler schrieb:
> > Hi Darren,
> > 
> > thanks for the details.
> > 
> > As far as I understand you can find DASL support for XPath 
> and SQL in 
> > the JCR Server (not the simple server) and it would be a matter of 
> > using the respective libs in the simple server and provide 
> for example 
> > a "basicsearch to jcr xpath"
> > conversion. Is that correct?
> > 
> > Shouldn't be too hard, should it?
> > If this is stalling your development maybe this could be a good 
> > contribution ;)? Anyone interested ??
> 
> I've been looking at translating from Jackrabbit's internal 
> query model to DASL/basicsearch, and this is non-trivial, and 
> definitively will not be fully compliant.
> 
> When I'm done, I may be available to provide a summary. One 
> major problem I already encountered is that DASL basicsearch 
> basically uses scopes (with depth 0, 1, infinity) and 
> conditions on query expressions. 
>   Thus, a query such as
> 
> 	/*/foo
> 
> already can't be translated, because I can't tell DASL to 
> restict the path to resources whose last path segment is 
> "foo". Another problem with the basicsearch scopes is that a 
> search at depth 1 includes things at depth 0, thus query 
> results will need to be filtered.
> 
> Best regards, Julian
> 

Mime
View raw message