[Bug 1425] New - XPath fails to catch document Element

[bu...@apache.org]

http://nagoya.apache.org/bugzilla/show_bug.cgi?id=1425

*** shadow/1425	Fri Apr 20 02:25:02 2001
--- shadow/1425.tmp.6547	Fri Apr 20 02:25:02 2001
***************
*** 0 ****
--- 1,99 ----
+ +============================================================================+
+ | XPath fails to catch document Element                                      |
+ +----------------------------------------------------------------------------+
+ |        Bug #: 1425                        Product: XalanJ2                 |
+ |       Status: NEW                         Version: 2.0.1                   |
+ |   Resolution:                            Platform: PC                      |
+ |     Severity: Normal                   OS/Version:                         |
+ |     Priority:                           Component: org.apache.xpath        |
+ +----------------------------------------------------------------------------+
+ |  Assigned To: xalan-dev@xml.apache.org                                     |
+ |  Reported By: maillist@nue.et-inf.uni-siegen.de                            |
+ |      CC list: Cc:                                                          |
+ +----------------------------------------------------------------------------+
+ |          URL:                                                              |
+ +============================================================================+
+ |                              DESCRIPTION                                   |
+ Dear Xalan Team
+ 
+ I get unexpected results if I apply the XPath to the attached sample documents. 
+ This affects Xalan 2.0.x 
+ 
+ "(//.)" 
+ 
+ to the XML documents shown below. The problem is that the XPath does not catch 
+ the document Element if comments and/or PIs follow the document root. If I 
+ modify the XPath to explicitly catch the document Element (everything below 
+ document except of comments and PIs):
+ 
+ "(//. | /*[not(self::processing-instruction() or self::comment())])"
+ 
+ then it catches the documentElement. To illustrate this behaviour, I modified 
+ \xml-xalan\java\samples\ApplyXPath\ApplyXPath.java to be a bit more noisy. The 
+ inner part then looks like:
+ 
+ #### Modified SOURCE #########################
+ 
+ NodeIterator nl = XPathAPI.selectNodeIterator(doc, xpath);
+ // Serialize the found nodes to System.out.
+ System.out.println("<output>");
+ String[] nodeTypeString = new String[]{ 
+   "", "ELEMENT",  "ATTRIBUTE", "TEXT_NODE",
+   "CDATA_SECTION", "ENTITY_REFERENCE",
+   "ENTITY", "PROCESSING_INSTRUCTION",
+   "COMMENT", "DOCUMENT", "DOCUMENT_TYPE",
+   "DOCUMENT_FRAGMENT", "NOTATION" };
+ int i = 0;
+ Node n;
+ while ((n = nl.nextNode()) != null) {
+    System.out.println("<node" + ++i + " nodeType=\""
+                       + nodeTypeString[n.getNodeType()] + "\">");
+    serializer.transform(new DOMSource(n),
+                         new StreamResult(System.out));
+    System.out.println("</node" + i + ">");
+    System.out.println();
+ }
+ System.out.println("</output>");
+ 
+ #### INPUT DATA #########################
+ This input data means 8 simgle documents, each document in a sinle line, a 
+ <doc/> element and additional comments/PIs. The first 3 documents miss the 
+ documentElement. 
+ 
+ Bad
+ <doc><a /> </doc><!-- -->         # mixed content and following comment
+ <doc><a /> </doc><?pi?>           # mixed content and following PI
+ <doc><a /><b /></doc><!-- -->     # mixed content and following comment
+ 
+ Good
+ <doc><a /></doc><!-- -->          # 'clean' content and following comment
+ <doc><a /> </doc>                 # mixed content and nothing follows
+ <!-- --><doc><a /> </doc>         # mixed content and preceding comment
+ <?pi?><doc><a /> </doc>           # mixed content and preceding PI
+ <doc><a /><b /></doc>             # mixed ElemContent 
+ 
+ #### SAMPLE OUTPUT #########################
+ 
+ Input was <doc><a /><b /></doc><!-- -->
+ XPath was (//.)
+ Software was Xerces 1.3.1 and Xalan 2.0.1 on JDK1.3.0 on Windows 2000
+ It catches the <a/> Element, the <b/> Element, the folowing comment and the 
+ _DOCUMENT_ node, but not the _ELEMENT_ containing the documentElement. 
+ 
+ OUTPUT:
+ Loading classes, parsing my_input.xml, and setting up serializer
+ Querying DOM using (//.)
+ <output>
+ <node1 nodeType="ELEMENT">
+ <a/></node1>
+ 
+ <node2 nodeType="ELEMENT">
+ <b/></node2>
+ 
+ <node3 nodeType="DOCUMENT">
+ <doc><a/><b/></doc></node3>
+ 
+ <node4 nodeType="COMMENT">
+ </node4>
+ 
+ </output>