[
https://issues.apache.org/jira/browse/JCR-1113?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12525006
]
Felix Meschberger commented on JCR-1113:
----------------------------------------
+1
This would probably also benefit Sling, as I have my own implementation of the
Mapper interface there because I read the descriptors using kxml, which is more
light weight and has far less dependencies than Commons Digester. So I "stole"
much of the code of the DigesterMapperImpl regarding descriptor verification ...
Also, removing the node type stuff from the DigesterMapper would remove the
dependency on the Jackrabbit-Core package, which may or may not be available in
certain use cases. Another reason, why I do not use the DigesterMapperImpl in
Sling.
> Refactor the Mapper & DescriptotReader classes
> ----------------------------------------------
>
> Key: JCR-1113
> URL: https://issues.apache.org/jira/browse/JCR-1113
> Project: Jackrabbit
> Issue Type: Improvement
> Components: jcr-mapping
> Reporter: Christophe Lombart
> Assignee: Christophe Lombart
> Fix For: 1.4
>
>
> I would like to refactor the mappers and the descriptor readers in order to
> :
> * Create an abstract mapper impl because both Mapper classes have a lot of
> code in common (AnnotedObjectMapper & DigesterMapperImpl). Only the readers
> are different. The Mappers can make exactly the same process.
> * The Mapper classes should not have the responsibility to create the jcr
> node types. This can be done outside the mapper and it should be an optional
> operation. There are certainly some use cases where node type creation is not
> necessary. Right now, the annotated object mapper creates jcr node types.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
