[
https://issues.apache.org/jira/browse/CMIS-386?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13047302#comment-13047302
]
jay brown commented on CMIS-386:
--------------------------------
I completely agree that based on the schema it is valid for the property to be
completely absent. The point is that having the element present-but-empty is
an equally valid (and accepted) method for expressing 'not set'. So what we
are suggesting is that the TCK not fail for either of the valid responses. As
opposed to what it does today which is fail if the element is present but
not-set (which is also correct).
> OpenCMIS TCK should consider empty cmis:parentId element as a "not set"
> attribute for base type definition
> ----------------------------------------------------------------------------------------------------------
>
> Key: CMIS-386
> URL: https://issues.apache.org/jira/browse/CMIS-386
> Project: Chemistry
> Issue Type: Bug
> Components: opencmis-tck
> Affects Versions: OpenCMIS 0.4.0
> Reporter: David Sanders
> Priority: Minor
>
> The OpenTCK expects a base type's cmis:parentId attribute to be null. An
> empty element should also be accepted assuming that meets the spec's
> definition of "not set". That is, both TypeDefinition.getParentTypeId() ==
> null
> and TypeDefiniton.getParentTypeId().length() == 0 should be accepted.
> OpenCMIS TCK fails with
> FAILURE: Document type spec compliance. (AbstractCmisTest.java:114)
> FAILURE: Base type has parent type! (AbstractSessionTest.java:1357)
> The IBM FileNet CMIS provider returns the following for cmis:document and
> cmis:folder:
> <cmis:parentId></cmis:parentId>
--
This message is automatically generated by JIRA.
For more information on JIRA, see: http://www.atlassian.com/software/jira
