[ 
https://issues.apache.org/jira/browse/JCR-1664?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Nicholas Stuart updated JCR-1664:
---------------------------------

    Attachment: out.patch

path to swap StringRefAddr to match glassfish Reference object.

> JNDI Referencable Issues
> ------------------------
>
>                 Key: JCR-1664
>                 URL: https://issues.apache.org/jira/browse/JCR-1664
>             Project: Jackrabbit
>          Issue Type: Bug
>          Components: jackrabbit-core
>    Affects Versions: 1.5
>         Environment: linux glassfish jndi spring
>            Reporter: Nicholas Stuart
>         Attachments: out.patch
>
>
> I'm questioning the use of Referencable in the BindableResource and 
> BindableResourceFactory classes for the JNDI lookup process. Reason for this 
> is because Referencable needs the Addrs to be in the EXACT order in order for 
> it to be considered the same. (see 
> http://java.sun.com/j2se/1.4.2/docs/api/javax/naming/Reference.html#equals(java.lang.Object)
>  )
> In order for me to get the JNDI reference to be found correctly I had to 
> change the BindableResource.getReference method to swap the order the 
> StringReferences were added to match up what was being passed in by 
> glassfish. This seems EXTREMELY fragile to me as I don't know what order, say 
> JBoss, would pass the StringRefences in in the Reference object for the 
> Factory method.
> Also, another problem is that getReference is binding the class name to 
> BindableRepository class implementation and not javax.jcr.Repository. This 
> again causes them not to match if you follow the example on the wiki on 
> setting up the JNDI reference and use javax.jcr.Repository as the type. This 
> can either be fixed by changing the JNDI reference to use the 
> BindableRepository class or the change the BindableRepository class to set 
> that to the Repository interface. Not sure which would be considered 'better'
> I have a patch that fixes the first issue (at least for glassfish), but not 
> the second. Again, this seems like a really 'breakable' setup right now and 
> not sure what would be better to make sure this is avoided.

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

Reply via email to