[
https://issues.apache.org/jira/browse/OAK-583?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13579172#comment-13579172
]
Alex Parvulescu commented on OAK-583:
-------------------------------------
I removed the fixme from the code, as the mentioned test failures are no longer
there (rev r1446564)
bq. it's just about having it consistent.
+0 I see what you are saying, but the benefit of forcing a type on this
property is not so great in the end.
Using an invalid property name would just render you index definition useless,
but that's also the case with a typo in the definition and there's really no
way to prevent that.
Also, I see the fact that you can use both type string and type name for the
definition as a feature :)
> Inconsistencies in property index definitions
> ---------------------------------------------
>
> Key: OAK-583
> URL: https://issues.apache.org/jira/browse/OAK-583
> Project: Jackrabbit Oak
> Issue Type: Improvement
> Components: core
> Reporter: angela
> Assignee: Alex Parvulescu
>
> while trying to simplify the property-index definitions and adding
> one for the access control content, i detected the following inconsistencies
> with the property "propertyNames":
> - the property is sometimes single valued (e.g. "uuid") and sometimes
> multivalued ("nodetypes")
> IMO the property name implies that it is multivalued. however, i get the
> impression that this doesn't work for unique property index definitions.
> (test failure)
> - the property is defined using Type.STRING (implicit) or Type.STRINGS.
> IMO however, Type.NAMES would be more appropriate.
> - while moving out the user related index, i used type name without noticing.
> i will change to STRING(S) such that we have it consistent (see above).
--
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
