[ https://issues.apache.org/jira/browse/JCR-1712?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=12624622#action_12624622 ] 

Julian Reschke commented on JCR-1712:
-------------------------------------

> isn't it inconsistent, if an extended jcr-name containing an unknown URI passes the parsing, while parsing a name with unknown prefix fails (the namespace resolver throws NamespaceException)? 

Not really, as the extended name is fully defined without additional session- or repository-related information.

Speaking of which, we need to clarify whether the URI in the extended name *needs* to be registered at all.

For instance, if I just need it as identifier, for instance for permissions, why would I need to repository-register the namespace?

> initially i had the parser behaving like that. but i got rid of that limitation due to the fact, that Jackrabbit currently allows any string to be a namespace URI except for the reserved ones. 

I disagree it's a limitation. What's a limitation is that distinguishing an extended name from a JCR name requires knowledge of the namespace registry. I think *that's* a very undesirable limitation.

People who mint invalid namespace URIs are only affected when they try to use the new notation. I think that's totally acceptable.


> JSR 283: JCR Names
> ------------------
>
>                 Key: JCR-1712
>                 URL: https://issues.apache.org/jira/browse/JCR-1712
>             Project: Jackrabbit
>          Issue Type: Sub-task
>            Reporter: angela
>         Attachments: JCR-1712_1.diff
>
>


-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.