[
https://issues.apache.org/jira/browse/JCRVLT-544?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17372515#comment-17372515
]
Dominik Süß commented on JCRVLT-544:
------------------------------------
[~kwin] I can try but I'll be out for roughly the next 4 weeks, therefore I
suggest for now to revert this change in the meantime as it creates a
regression while the previous behavior did work for 10+ years. I agree that it
would be nice to better follow the descriptive nature of CNDs where possible
but the practical and theoretical cases make the current version of vault not
adoptable into production. History tells that if we find a case that we could
control we always find further cases rooted in customer code that we cannot
control and have to stay backwards compatible as well.
> Creation of folders based on parent's default nodetype instead of nt:folder
> causing regression
> ----------------------------------------------------------------------------------------------
>
> Key: JCRVLT-544
> URL: https://issues.apache.org/jira/browse/JCRVLT-544
> Project: Jackrabbit FileVault
> Issue Type: Bug
> Components: vlt
> Affects Versions: 3.4.4
> Reporter: Dominik Süß
> Assignee: Konrad Windszus
> Priority: Major
> Fix For: 3.5.2
>
>
> The change of JCRVLT-417 makes folder creation fall back to the nodetypes
> declared in the nodetype constraint of the primary type. This can cause
> regressions with existing packages.
> In the given case the nodetype cq:Component allows a child node definition
> with primaryType nt:base (which nt:folder derives from) - yet as nt:base has
> no childNodeDefinitions declared any substructures fail to install. The
> nodeType definition of cq:Component is correct as it "allows" anything
> deriving from nt:base, whereas only the derived nodeTypes open up further
> options like substructures.
> //cc [~kwin] introducing the behavioral change.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
