[ 
https://issues.apache.org/jira/browse/OAK-11018?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17873929#comment-17873929
 ] 

Julian Reschke commented on OAK-11018:
--------------------------------------

bq. I'm not sure what you mean with "evil" here.

Setting jcr:uuid "manually" on nodes that are not mix:referenceable (yet).

bq. I would keep this. Or maybe we can clarify it? I don't know what is unclear 
actually.

It doesn't reflect reality, see PR for  
https://issues.apache.org/jira/browse/OAK-11000.

bq. Someone might except duplicate UUIDs to work, and a "uniqueness constraint 
violation" would be unexpected for him.

But that's what happens right now: 
https://github.com/apache/jackrabbit-oak/pull/1622/commits/fbc3bf1fe1ef0b0bb980b804f344d9cc57521e08.

bq. I'm not sure what is the motivation for this issue. Do you plan to add such 
properties?

If FileVault could do it, it would avoid the so-called "stashing" (moving all 
children of the node twice), which is extremely slow in DocumentNodeStore, and 
which has caused that operation to time out for users (leaving the repository 
in a somewhat flaky state, with some nodes being moved "away" to a temp folder, 
and some not).


> doc: clarify warning about setting jcr:uuid on non-referenceable nodes
> ----------------------------------------------------------------------
>
>                 Key: OAK-11018
>                 URL: https://issues.apache.org/jira/browse/OAK-11018
>             Project: Jackrabbit Oak
>          Issue Type: Documentation
>          Components: doc
>            Reporter: Julian Reschke
>            Assignee: Julian Reschke
>            Priority: Minor
>
> [https://jackrabbit.apache.org/oak/docs/differences.html#Identifiers] says 
> (as per change in OAK-2164):
> {quote}Manually adding a property with the name jcr:uuid to a non 
> referenceable node might have unexpected effects as Oak maintains an unique 
> index on jcr:uuid properties. As the namespace jcr is reserved, doing so is 
> strongly discouraged.
> {quote}
> But the tests for OAK-11000 show that this just works as "expected in Oak" 
> (throwing an exception even though no mix:referenceable is present) - as the 
> UUID index is maintained even for nodes that do not have mix:referenceable.
> Should we remove or rephrase that warning?



--
This message was sent by Atlassian Jira
(v8.20.10#820010)

Reply via email to