[ 
https://issues.apache.org/jira/browse/JXPATH-133?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
 ]

Vladimir Orlov updated JXPATH-133:
----------------------------------

    Description: 
During the implementation of JCR nodes support in JXPath (extending 
NodePonter's, NodeIterator's, NodePointerFactory ) I faced with the following 
situation: I was running a xpath query against the context (JXPathContext 
instance) like:

context.iteratePointers("/child::* | @ * ")

After that I got an exception JXPathException: "Cannot compare pointers that do 
not belong to the same tree: '/home' and '/home' " thrown from 
NodePointer.compareNodePointers(...) method. After debugging this case I've 
found that for the node pointers extracted via the "/child::* | @ * " xpath 
expression NodePointer.compareTo(Object) was invoked (and consequently 
NodePointer.compareNodePointers(...) ). In fact all node pointers matching the 
"/child::* | @ * " xpath expression had the same node pointer (I checked it 
against my implementations of NodePointerFactory, NodeIterator's etc.). But as 
I found for these node pointers they were initialized with different clones of 
the same parent node pointer. And the result was that in 
NodePointer.compareTo(Object) method the following "==" comparison evaluated in 
false for the different clones of the same parent node pointer:

        if (parent == pointer.parent) {
            return parent == null ? 0 : parent.compareChildNodePointers(this, 
pointer);
        }

As the workaround for this case I overrode the clone method for my NodePointer 
descendants to return 'this'. I think there is some misbalance made between the 
 usage of clone () method of NodePointer's and "==" comparison of NodePointers.


  was:
During the implementation of JCR nodes support in JXPath (extending 
NodePonter's, NodeIterator's, NodePointerFactory ) I faced with the following 
situation: I was running a xpath query against the context (JXPathContext 
instance) like:

context.iteratePointers("/child::* | @ * ")

After that I got an exception JXPathException: "Cannot compare pointers that do 
not belong to the same tree: '/home' and '/home' " thrown from 
NodePointer.compareNodePointers(...) method. After debugging this case I've 
found that for the node pointers extracted via the "/child::*|@*" xpath 
expression NodePointer.compareTo(Object) was invoked (and consequently 
NodePointer.compareNodePointers(...) ). In fact all node pointers matching the 
"/child::*|@*" xpath expression had the same node pointer (I checked it against 
my implementations of NodePointerFactory, NodeIterator's etc.). But as I found 
for these node pointers they were initialized with different clones of the same 
parent node pointer. And the result was that in NodePointer.compareTo(Object) 
method the following "==" comparison evaluated in false for the different 
clones of the same parent node pointer:

        if (parent == pointer.parent) {
            return parent == null ? 0 : parent.compareChildNodePointers(this, 
pointer);
        }

As the workaround for this case I overrode the clone method for my NodePointer 
descendants to return 'this'. I think there is some misbalance made between the 
 usage of clone () method of NodePointer's and "==" comparison of NodePointers.



> Cloning the node pointers and '==' node pointers comparison
> -----------------------------------------------------------
>
>                 Key: JXPATH-133
>                 URL: https://issues.apache.org/jira/browse/JXPATH-133
>             Project: Commons JXPath
>          Issue Type: Bug
>    Affects Versions: 1.3
>         Environment: JDK 1.5.0_12, JBoss 4.0.5
>            Reporter: Vladimir Orlov
>
> During the implementation of JCR nodes support in JXPath (extending 
> NodePonter's, NodeIterator's, NodePointerFactory ) I faced with the following 
> situation: I was running a xpath query against the context (JXPathContext 
> instance) like:
> context.iteratePointers("/child::* | @ * ")
> After that I got an exception JXPathException: "Cannot compare pointers that 
> do not belong to the same tree: '/home' and '/home' " thrown from 
> NodePointer.compareNodePointers(...) method. After debugging this case I've 
> found that for the node pointers extracted via the "/child::* | @ * " xpath 
> expression NodePointer.compareTo(Object) was invoked (and consequently 
> NodePointer.compareNodePointers(...) ). In fact all node pointers matching 
> the "/child::* | @ * " xpath expression had the same node pointer (I checked 
> it against my implementations of NodePointerFactory, NodeIterator's etc.). 
> But as I found for these node pointers they were initialized with different 
> clones of the same parent node pointer. And the result was that in 
> NodePointer.compareTo(Object) method the following "==" comparison evaluated 
> in false for the different clones of the same parent node pointer:
>         if (parent == pointer.parent) {
>             return parent == null ? 0 : parent.compareChildNodePointers(this, 
> pointer);
>         }
> As the workaround for this case I overrode the clone method for my 
> NodePointer descendants to return 'this'. I think there is some misbalance 
> made between the  usage of clone () method of NodePointer's and "==" 
> comparison of NodePointers.

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.

Reply via email to