[
https://issues.apache.org/jira/browse/TUSCANY-3689?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Simon Nash resolved TUSCANY-3689.
---------------------------------
Resolution: Fixed
Fixed under revision r1001528 in the 1.6.1 branch and revision r1027406 in the
1.x trunk.
> databinding-saxon corrupts XML documents when transforming from
> XMLStreamReader to NodeInfo
> -------------------------------------------------------------------------------------------
>
> Key: TUSCANY-3689
> URL: https://issues.apache.org/jira/browse/TUSCANY-3689
> Project: Tuscany
> Issue Type: Bug
> Components: SCA Java Runtime
> Affects Versions: Java-SCA-1.6
> Reporter: Simon Nash
> Assignee: Simon Nash
> Fix For: Java-SCA-1.6.1
>
>
> The xml-bigbank demo fails on Tuscany SCA 1.5, 1.5.1, and 1.6 because of a
> problem with the databinding-saxon transformation from XMLStreamReader to
> NodeInfo. It works on Tuscany SCA 1.4.
> This problem isn't related to the frequent unavailability of the live
> exchange rate RSS feed and the live stock quote web service used by this
> demo. For test purposes I have proved this by stubbing out these remote calls
> and using known good data instead.
> The problem is caused by the two-step transformation from XMLStreamReader to
> NodeInfo. The first step transforms the XMLStreamReader to a DOM using
> org.apache.tuscany.sca.databinding.xml.XMLStreamReader2Node, and the second
> step transforms the DOM to a NodeInfo using
> org.apache.tuscany.sca.databinding.saxon.Node2NodeInfoTransformer. The first
> step is corrupting the name of the top-level node in the XML document by
> changing it to "arg1". The second step works OK.
> The following code in XMLStreamReader2Node is causing the name of the
> top-level node to be corrupted:
> if (node instanceof Document) {
> Document doc = (Document)node;
> return DOMHelper.adjustElementName(context,
> doc.getDocumentElement());
> } else {
> return node;
> }
> The node name is corrupted by the call to DOMHelper.adjustElementName(). This
> name adjustment is necessary when transforming an XML instance document for
> use as a SOAP parameter, but it shouldn't happen when transforming an XML
> document for passing to a Saxon XQuery script..
> I've looked in the obvious places for differences in the code between 1.4 and
> 1.6 that would explain why this corruption didn't occur on 1.4, but I haven't
> been able to find out exactly what changed and triggered the problem.
> The simplest and safest way to fix this without risking disturbing other
> transformation code paths is to write a new dedicated databinding transformer
> for the specific case that isn't working.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
