[
https://issues.apache.org/jira/browse/CMIS-915?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14559103#comment-14559103
]
Florian Müller commented on CMIS-915:
-------------------------------------
There is a bit of confusion because there are two exact flags. There is one
returned by getACL, which is optional as you pointed out, and it is treated as
optional by the TCK. There is another exact flag attached to CMIS objects. The
TCK complains about this one here. Unfortunately, the spec isn't clear whether
this is optional or not. (Might be a candidate for the CMIS 1.1 errata.)
Changing the failure to a warning works for me. To be on the safe side, a
server should always set this flag, though.
> Discrepancy between specification and TCK tests
> -----------------------------------------------
>
> Key: CMIS-915
> URL: https://issues.apache.org/jira/browse/CMIS-915
> Project: Chemistry
> Issue Type: Bug
> Components: opencmis-tck
> Affects Versions: OpenCMIS 0.13.0
> Reporter: Gavin Cornwell
>
> In 0.13.0 TCK tests started failing around the isExact flag, for example:
> {noformat}
> FAILURE: ACL (AbstractSessionTest.java:1222) [stacktrace]
> FAILURE: ACL is provided but the isExact flag is not set!
> (AbstractSessionTest.java:1332) [stacktrace]
> INFO: Object is null!
> {noformat}
> The spec
> [implies|http://docs.oasis-open.org/cmis/CMIS/v1.1/os/CMIS-v1.1-os.html#x1-36600010]
> that the isExact parameter is optional but the TCK now declares it must be
> set.
> Reducing the failure to a warning was hinted at on the mailing list, can this
> be done?
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
