[
https://issues.apache.org/jira/browse/JCR-3528?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13592339#comment-13592339
]
Jukka Zitting commented on JCR-3528:
------------------------------------
In general I agree with Angela's comment in JCR-3524 about preferring
explicitly specified test node types over trying to automatically find them. A
repository implementation could rightly have some custom rules for example
about specific node types only being allowed in specific subtrees (like
nt:version only under version storage in jcr:system), which is something that
the TCK couldn't possibly auto-detect.
That said, if the patch fixes the issue at hand and introduces no other
problems, then it's obviously an improvement. +1
> Node type selection for
> CanAddChildNodeCallWithNodeTypeTest#testResidualAndLegalType is not optimal
> ---------------------------------------------------------------------------------------------------
>
> Key: JCR-3528
> URL: https://issues.apache.org/jira/browse/JCR-3528
> Project: Jackrabbit Content Repository
> Issue Type: Improvement
> Components: jackrabbit-jcr-tests
> Reporter: Alex Parvulescu
> Assignee: Alex Parvulescu
> Priority: Minor
> Attachments: JCR-3528.patch
>
>
> Similar to JCR-3524, the node type selection for this test is influenced by
> the underlying storage.
> It can choose the fist node type even if that is invalid for the purpose of
> the test and fail the test.
> For example running against an Oak backend the test can choose _nt:folder_ as
> a parent and _nt:hierarchyNode_ as a child, but because _nt:hierarchyNode_ is
> abstract it cannot be used, so the test fails for the wrong reason.
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
