Merge lp:~wallyworld/bzr4j/less-strict-xml-parsing into lp:bzr4j

[Ian Booth]

Make bzr xmlstatus parsing less strict to handle "bad" bzr xml output.

When the plugin gathers info on changes, it runs bzr xmlstatus
If the resulting xml (as output by bzr) has issues, the xerces parser throws an exception and the plugin operation aborts. eg

2010-09-17 17:55:29,197 [ 148235] WARN - e.xmloutput.XmlAbstractHandler - sax error
org.xml.sax.SAXParseException: The value of attribute "fid" associated with an element type "file" must not contain the '<' character.
        at org.apache.xerces.util.ErrorHandlerWrapper.createSAXParseException(Unknown Source)
        at org.apache.xerces.util.ErrorHandlerWrapper.fatalError(Unknown Source)

An example of some bad xml is:

<file fid="x_Mark_Shuttleworth_<email address hidden>_Fri_Jul_22_22:53:21_2005_21772.2">lib/lp/code/interfaces/branch.py</file>

This is actual bzr output. The problem is that the fid attribute contains a '<' character.

By telling the xml parser to be less strict, the above issue is ignored and the plugin works correctly. Without the fix, changed files are not highlighted and change indicators in the gutter margins do not stick after the file is saved.

=== modified file 'modules/core/src/main/java/org/emergent/bzr4j/core/cli/BzrXmlResult.java'
--- modules/core/src/main/java/org/emergent/bzr4j/core/cli/BzrXmlResult.java	2010-03-02 20:44:47 +0000
+++ modules/core/src/main/java/org/emergent/bzr4j/core/cli/BzrXmlResult.java	2010-09-17 09:35:54 +0000
@@ -39,6 +39,7 @@
   static {
     try {
       sm_saxParserFactory.setFeature("http://apache.org/xml/features/allow-java-encodings",true);
+      sm_saxParserFactory.setFeature("http://apache.org/xml/features/continue-after-fatal-error", true);
     } catch (ParserConfigurationException e) {
       e.printStackTrace();
     } catch (SAXNotRecognizedException e) {