cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stefano Mazzocchi <stef...@apache.org>
Subject Re: Repository support for property queries (DASL?)
Date Mon, 12 Apr 2004 19:34:45 GMT
Gianugo Rabellino wrote:

> Rolf Kulemann wrote:
> 
>> Hello,
>>
>> I set up a small cocoon app which uses the Repository interface.
>> I'm now up to a point where would use a kind of property query in order
>> to receive all resources/paths matching the query.
>>
>> Currently the repository does not support a property "search" method.
>>
>> IMHO this is an important feature and should be a concern of the
>> RepositoryPropertyHelper.
>>
>> Another question is, what kind of queroes should be supported? DASL
>> queries???
>>
> 
> Oh, queries... well, I really appreciate your enthousiasm, but my advice 
> is not even dare to think that querying is going to be solved just by an 
> email thread. JSR170 people have been fighting over it quite a while, 
> and overall searching is _really_ a PITA.
> 
> Besides, DASL isn't really a query language but rather a query wrapper: 
> you're free to insert whatever query language you want (DAV:basicsearch 
> being the most prominent one) inside DASL (which boils also to an HTTP 
> method + a bunch of headers).
> 
> Consider also that it's not just the query language, but even how 
> results are reported.
> 
> MHO? I really feel uncomfortable with the repository abstraction as a 
> whole since, after all, it has to be somehow modeled into webdav. I'm 
> fine with it as long as it's lightweight enough, but for more serious 
> needs I'd much rather wait for JSR170 to come out.

FYI, ETA for the reference implementation of JSR170 (aka JCR) in 
slide/proposals/jcrri is 2/4 weeks.

The current state of affairs is that JCR will support two query 
languages: a SQL-like called JCRQL and a sub/superset of XPath. Both are 
optional, and a container has to include at least one of them to be 
compliant.

Now, don't shoot the pianist.

This is what the group came out with, not what I personally like.

The idea is to fix problems with the spec in the reference 
implementation ;-)

-- 
Stefano.


Mime
View raw message