[
https://issues.apache.org/jira/browse/OAK-1181?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13821900#comment-13821900
]
Tobias Bocanegra commented on OAK-1181:
---------------------------------------
do you really need to version the oak:index definition? if not, I would IGNORE
the entire node, by enforcing a {{mix:Indexable}}:
{noformat}
[oak:Indexable] mixin
+ oak:index (oak:QueryIndexDefinition) IGNORE
{noformat}
then the oak:QueryIndexDefinition could indeed extend from nt:unstructured, and
allow to be versioned if oak:index is versioned explicitly.
> Review node type definition for oak:queryIndexDefinition
> --------------------------------------------------------
>
> Key: OAK-1181
> URL: https://issues.apache.org/jira/browse/OAK-1181
> Project: Jackrabbit Oak
> Issue Type: Sub-task
> Components: core
> Reporter: angela
> Fix For: 0.14
>
>
> current definition:
> {noformat}
> [oak:queryIndexDefinition] > nt:unstructured
> - type (STRING) mandatory
> - async (STRING)
> - reindex (BOOLEAN) mandatory IGNORE
> {noformat}
> things to review:
> - name not following naming convention
> - extending from nt:unstructured
> - OPV flags (behavior upon versioning of the parent node)
> - incomplete list of properties compared to those mentioned in query.md
> - 'async' property is string with a predefined value 'async': why not boolean?
> - index content stored underneath the definition has not dedicated node type.
> things to consider:
> - is it expected that an index definition has custom properties?
> - is it expected that such custom properties have other property types that
> STRING/NAME/BOOLEAN? e.g. a binary?
> - is it expected that an index definition has other child nodes that the
> index content itself?
--
This message was sent by Atlassian JIRA
(v6.1#6144)
