isn't NodeState related to a transient step? The difference you see would mean
the transient # of children is different from the persisted one (which wouldn't
be choking)?
Le 10 mai 2012 à 14:07, Julian Reschke a écrit :
> On 2012-05-10 13:39, Peter Fry wrote:
>> I am trying to debug why a node can't be added to a parent. I'm getting
>> a ItemExistsException at line 1281 of NodeImpl i.e. in
>>
>>
>>
>> NodeState thisState = data.getNodeState();
>>
>> ChildNodeEntry cne = thisState.getChildNodeEntry(nodeName, 1);
>>
>> if (cne != null) {
>>
>> // there's already a child node entry with that name;
>>
>> // check same-name sibling setting of new node
>>
>> if (!def.allowsSameNameSiblings()) {
>>
>> throw new ItemExistsException(
>>
>> "This node already exists: "
>>
>> + itemMgr.safeGetJCRPath(nodePath));
>>
>> }
>>
>> // check same-name sibling setting of existing node
>>
>> NodeImpl existing = itemMgr.getNode(cne.getId(),
>> getNodeId());
>>
>> if (!existing.getDefinition().allowsSameNameSiblings()) {
>>
>> throw new ItemExistsException(
>>
>> "Same-name siblings not allowed for " +
>> existing);
>>
>> }
>>
>> }
>>
>>
>>
>> But when I query the respository with JCR 2 using:
>>
>>
>>
>> NodeIterator nodes = node.getNodes();
>>
>> while (nodes.hasNext()) {
>>
>> Node node1 = nodes.nextNode();
>>
>> }
>>
>>
>>
>> I don't get the node listed that is causing the collision. How come some
>> nodes show up in the addNode code
>>
>> That don't show up in the node.getNodes() call?
>> ...
>
> Wild guess: permissions?
