[
https://issues.apache.org/jira/browse/XERCESJ-832?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
George Bina updated XERCESJ-832:
--------------------------------
Attachment: XPathMatcher.java
I attached the new XPathMatcher class.
This solves all the 4 open issues related with identity constraints:
XERCESJ-832, XERCESJ-846, XERCESJ-1138 and XERCESJ-1396.
The implementation creates a matcher for each location path and then it just
delegates to these matchers. Each matcher stores a set of indices pointing to
the steps to be matched as the current state. It uses a stack to save these
indices when moving to a child and to restore them when returning from a child
element. We need a set of indices because the descendant step can generate more
steps as current, for example, imagine an XPath like
.//a/a/a/a
When we have a document fragment like
<a>
<a>
<a>
....
after this 3rd a we can expect the next element at all locations marked with ^
below
.//^a/^a/^a/^a
So, in this case we will have 4 indices representing the match state of this
location path.
I reused the IntStack to store these sets, putting the values first, followed
by the length of the set.
If you have any questions just let me know.
> element order matters on keyref checking?
> -----------------------------------------
>
> Key: XERCESJ-832
> URL: https://issues.apache.org/jira/browse/XERCESJ-832
> Project: Xerces2-J
> Issue Type: Bug
> Components: XML Schema 1.0 Structures
> Affects Versions: 2.5.0
> Environment: Operating System: Other
> Platform: Other
> Reporter: Jon Schewe
> Attachments: 24724.patch, dome.xsd, test.xml, test.xml,
> XPathMatcher.java
>
>
> By reversing the order of the Entity tags RoleKeyRef will fail. However to
> get
> to this point in the validation the patch for bug 24692 needs to be applied
> first, otherwise the keys won't propertly be recognized.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
