[
https://issues.apache.org/jira/browse/JCR-4015?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15467583#comment-15467583
]
Carsten Ziegeler commented on JCR-4015:
---------------------------------------
[~reschke] Proposed patch
{code}
Index: src/main/java/org/apache/jackrabbit/commons/JcrUtils.java
===================================================================
--- src/main/java/org/apache/jackrabbit/commons/JcrUtils.java (Revision
1759423)
+++ src/main/java/org/apache/jackrabbit/commons/JcrUtils.java (Arbeitskopie)
@@ -1449,8 +1449,22 @@
// path denotes root node
return session.getRootNode();
}
+ // find deepest existing parent node
+ String path = absolutePath;
+ String existingPath = null;
+ int currentIndex = path.lastIndexOf('/');
+ while (currentIndex > 0) {
+ path = path.substring(0, currentIndex);
+ // break when first existing parent is found
+ if (session.nodeExists(path)) {
+ existingPath = path;
+ break;
+ }
+ currentIndex = path.lastIndexOf('/');
+ }
// create path relative to the root node
- return getOrCreateByPath(session.getRootNode(),
absolutePath.substring(1),
+ return getOrCreateByPath(existingPath == null ? session.getRootNode()
: session.getNode(existingPath),
+ absolutePath.substring(currentIndex + 1),
createUniqueLeaf, intermediateNodeType, nodeType, autoSave);
}
{code}
> CLONE - jackrabbit-jcr-commons JcrUtils.getOrCreateByPath fails if session is
> not allowed to read root
> ------------------------------------------------------------------------------------------------------
>
> Key: JCR-4015
> URL: https://issues.apache.org/jira/browse/JCR-4015
> Project: Jackrabbit Content Repository
> Issue Type: Bug
> Components: jackrabbit-jcr-commons
> Affects Versions: 2.4.5, 2.6.5, 2.8.1, 2.10.3, 2.12.2
> Reporter: Carsten Ziegeler
>
> JcrUtils.getOrCreateByPath starts to try to find the node from the root down
> to the specified path, therefore if the current session has not access to
> "/", it fails. JcrUtils.getOrCreateByPath must try the longest path first and
> then reduce it until it finds an accessible node. So for example if the input
> is "/a/b/c" instead of testing for "/", "/a", "/a/b" .. it should test
> /a/b/c, /a/b, /a and then /
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
