[
https://issues.apache.org/jira/browse/JCR-1052?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12539799
]
Filippo Fratoni commented on JCR-1052:
--------------------------------------
Sorry for very late answer.
I don't consider this request as Bug but Improvement request.
Let me explain better the scenario.
Expectation:
When working with JCR implementation I need to work with paths for every
kind of functional implementation, and paths I can use with
hierarchy manager ar JCR Specification compliant. So when I'm going to
query JCR my expectation is to deal with paths I'm using with
hierarchy manager.
Real case:
When I'm going to query along path of the form "Alevel/Blevel/3Level" this
paths breaks XML Specifications an XPath parser can't build the tree.
JCR Specifications says:
XPath query "must" be run against xml document representation of jcr
structure. They don't assert anything about the owner of the conversion from
jcr names to xml names.
Suggestion:
I'm using JackRabbit to develope complex taxonomies and resource hierarchies
with strong and complex query need. Conversion between JCR end XML become a
big Issue. In this cases an Xpath parser that transparently accept JCR path
can be a great plus. Xpath localNames may contains numeric starting
character to allow tree creation and than may apply ISO9075 conversion.
thanks for paying attention to my request
Filippo
> JCR valid names not allowed in Xpath xml sintax
> ------------------------------------------------
>
> Key: JCR-1052
> URL: https://issues.apache.org/jira/browse/JCR-1052
> Project: Jackrabbit
> Issue Type: Improvement
> Components: xpath
> Affects Versions: 1.3
> Environment: WIndows 2003/xp server/client
> Tomcat 6.0.x
> Reporter: Filippo Fratoni
> Attachments: Xpath_basic_encoding.txt
>
>
> ref: JCR-579
>
> I'm working with Magnolia CMS that relies on Jackrabbit jcr Impl 1.3.
> I can't query valid JCR path using Jackrabbit SearchManager in xpath mode due
> to different spec. between JCR names and XML names.
> The ISO9075 approach in jackrabbit commons library is name based and not path
> based. The result is the need to parse again full path in an Xpath compliant
> way to recursively encode non XML names.
> I think that the owner of this enhancement should be SearchManager class or
> one of its collaborators(the parser for example) to ensure the consistence of
> JackRabbit jcr impl.
> For my project needs I'm going to token path string with "/" separator and
> then apply ISO9075 encode. It will be ok for my actual project.
> Thanks in advance
> Filippo
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
