add locator support and merge to many contribution

dom-parser.js

@@ -1,11 +1,15 @@
-function DOMParser(){
+function DOMParser(pos){
+	this.recordPositions = pos;
 
 }
 DOMParser.prototype.parseFromString = function(source,mimeType){
 	var sax =  new XMLReader();
 	var handler = new DOMHandler();
 	var defaultNSMap = {};
 	var entityMap = {'lt':'<','gt':'>','amp':'&','quot':'"','apos':"'"}
+	if(this.recordPositions){
+		handler.setDocumentLocator({})
+	}
 	sax.contentHandler = handler;
 	sax.lexicalHandler = handler;
 	sax.errorHandler = handler;
@@ -30,7 +34,10 @@ function DOMHandler() {
     this.errors = [];
     this.cdata = false;
 }
-
+function position(locator,node){
+	node.lineNumber = locator.lineNumber;
+	node.columnNumber = locator.columnNumber;
+}
 /**
  * @see org.xml.sax.ContentHandler#startDocument
  * @link http://www.saxproject.org/apidoc/org/xml/sax/ContentHandler.html
@@ -39,7 +46,7 @@ DOMHandler.prototype = {
 	startDocument : function() {
     	this.document = new DOMImplementation().createDocument(null, null, null);
     	if (this.locator) {
-        	this.document.documentURI = this.locator.getSystemId();
+        	this.document.documentURI = this.locator.systemId;
     	}
 	},
 	startElement:function(namespaceURI, localName, qName, attrs) {
@@ -48,6 +55,8 @@ DOMHandler.prototype = {
 	    var len = attrs.length;
 	    appendElement(this, el);
 	    this.currentElement = el;
+
+		this.locator && position(this.locator,el)
 	    for (var i = 0 ; i < len; i++) {
 	        var namespaceURI = attrs.getURI(i);
 	        var value = attrs.getValue(i);
@@ -69,6 +78,7 @@ DOMHandler.prototype = {
 	},
 	processingInstruction:function(target, data) {
 	    var ins = this.document.createProcessingInstruction(target, data);
+	    this.locator && position(this.locator,ins)
 	    appendElement(this, ins);
 	},
 	ignorableWhitespace:function(ch, start, length) {
@@ -77,12 +87,13 @@ DOMHandler.prototype = {
 		chars = _toString.apply(this,arguments)
 		if(this.currentElement && chars){
 			if (this.cdata) {
-				var cdataNode = this.document.createCDATASection(chars);
-				this.currentElement.appendChild(cdataNode);
+				var charNode = this.document.createCDATASection(chars);
+				this.currentElement.appendChild(charNode);
 			} else {
-				var textNode = this.document.createTextNode(chars);
-				this.currentElement.appendChild(textNode);
+				var charNode = this.document.createTextNode(chars);
+				this.currentElement.appendChild(charNode);
 			}
+			this.locator && position(this.locator,charNode)
 		}
 	},
 	skippedEntity:function(name) {
@@ -91,13 +102,16 @@ DOMHandler.prototype = {
 		this.document.normalize();
 	},
 	setDocumentLocator:function (locator) {
-	    this.locator = locator;
+	    if(this.locator = locator){
+	    	locator.lineNumber = 0;
+	    }
 	},
 	//LexicalHandler
 	comment:function(chars, start, length) {
 		chars = _toString.apply(this,arguments)
-	    var comment = this.document.createComment(chars);
-	    appendElement(this, comment);
+	    var comm = this.document.createComment(chars);
+	    this.locator && position(this.locator,comm)
+	    appendElement(this, comm);
 	},
 
 	startCDATA:function() {
@@ -112,6 +126,7 @@ DOMHandler.prototype = {
 		var impl = this.document.implementation;
 	    if (impl && impl.createDocumentType) {
 	        var dt = impl.createDocumentType(name, publicId, systemId);
+	        this.locator && position(this.locator,dt)
 	        appendElement(this, dt);
 	    }
 	},

dom.js

@@ -946,9 +946,9 @@ function serializeToString(node,buf){
 		}
 		return;
 	case ATTRIBUTE_NODE:
-		return buf.push(' ',node.name,'="',node.value.replace(/[<>&"]/g,_xmlEncoder),'"');
+		return buf.push(' ',node.name,'="',node.value.replace(/[<&"]/g,_xmlEncoder),'"');
 	case TEXT_NODE:
-		return buf.push(node.data.replace(/[<>&]/g,_xmlEncoder));
+		return buf.push(node.data.replace(/[<&]/g,_xmlEncoder));
 	case CDATA_SECTION_NODE:
 		return buf.push( '<![CDATA[',node.data,']]>');
 	case COMMENT_NODE:
@@ -1091,9 +1091,9 @@ try{
 					while(this.firstChild){
 						this.removeChild(this.firstChild);
 					}
-          if(String(data)!=''){
-  					this.appendChild(this.ownerDocument.createTextNode(data));
-          }
+					if(data || String(data)){
+						this.appendChild(this.ownerDocument.createTextNode(data));
+					}
 					break;
 				default:
 					//TODO:

sax.js

@@ -34,9 +34,24 @@ function parse(source,defaultNSMapCopy,entityMap,contentHandler,lexHandler,error
 	}
 	function appendText(end){
 		var xt = source.substring(start,end).replace(/&#?\w+;/g,entityReplacer);
+		locator&&position(start);
 		contentHandler.characters(xt,0,end-start);
 		start = end
 	}
+	function position(start,m){
+		while(start>=endPos && (m = linePattern.exec(source))){
+			startPos = m.index;
+			endPos = startPos + m[0].length;
+			locator.lineNumber++;
+			//console.log('line++:',locator,startPos,endPos)
+		}
+		locator.columnNumber = start-startPos+1;
+	}
+	var locator = contentHandler.locator;
+	var linePattern = /.+(?:\r\n?|\n)|.*$/g
+	var startPos = 0;
+	var endPos = 0;
+
 	var elStack = [{currentNSMap:defaultNSMapCopy}]
 	var closeMap = {};
 	var start = 0;
@@ -61,9 +76,11 @@ function parse(source,defaultNSMapCopy,entityMap,contentHandler,lexHandler,error
 			break;
 			// end elment
 		case '?':// <?...?>
+			locator&&position(i);
 			end = parseInstruction(source,i,lexHandler);
 			break;
 		case '!':// <!doctype,<![CDATA,<!--
+			locator&&position(i);
 			end = parseDCC(source,i,contentHandler,lexHandler);
 			break;
 		default:
@@ -74,6 +91,7 @@ function parse(source,defaultNSMapCopy,entityMap,contentHandler,lexHandler,error
 				return;
 			}else{
 				try{
+					locator&&position(i);
 					var end = parseElementAttribute(source,i,entityReplacer,contentHandler,lexHandler,closeMap,elStack);
 				}catch(e){
 					end = -1;

test/dom/serializer.js

@@ -5,7 +5,7 @@ wows.describe('XML Serializer').addBatch({
   'text node containing "]]>"': function() {
     var doc = new DOMParser().parseFromString('<test/>', 'text/xml');
     doc.documentElement.appendChild(doc.createTextNode('hello ]]> there'));
-    console.assert(doc.documentElement.firstChild.toString() == 'hello ]]&gt; there');
+    console.assert(doc.documentElement.firstChild.toString() == 'hello ]]> there',doc.documentElement.firstChild.toString());
   },
   '<script> element with no children': function() {
     var doc = new DOMParser().parseFromString('<html><script></script></html>', 'text/html');

test/index.js

@@ -42,10 +42,14 @@ DOMParser.prototype.parseFromString = function(data,mimeType){
 	}
 	return doc;
 }
-
+console.log('test dom:')
 require('./dom');
+console.log('test parse-element:')
 require('./parse-element');
+console.log('test node:')
 require('./node');
+console.log('test namespace:')
 require('./namespace');
+console.log('test normalize:')
 require('./html/normalize');
 //require('./big-file-performance');

test/locator.js

@@ -0,0 +1,24 @@
+var wows = require('vows');
+var DOMParser = require('xmldom').DOMParser;
+
+function assertPosition(n, line, col) {
+  console.assert(n.lineNumber == line,'lineNumber:'+n.lineNumber+'/'+line);
+  console.assert(n.columnNumber == col,'columnNumber:'+n.columnNumber+'/'+col);
+}
+
+wows.describe('DOMLocator').addBatch({
+  'node positions': function() {
+    var parser = new DOMParser(true);
+    var doc = parser.parseFromString('<?xml version="1.0"?><!-- aaa -->\n<test>\n  <a attr="value"><![CDATA[1]]>something\n</a>x</test>', 'text/xml');
+    var test = doc.documentElement;
+    var a = test.firstChild.nextSibling;
+    assertPosition(doc.firstChild, 1, 1);
+    assertPosition(doc.firstChild.nextSibling, 1, 1+'<?xml version="1.0"?>'.length);
+    assertPosition(test, 2, 1);
+    //assertPosition(test.firstChild, 1, 7);
+    assertPosition(a, 3, 3);
+    assertPosition(a.firstChild, 3, 19);
+    assertPosition(a.firstChild.nextSibling, 3, 19+'<![CDATA[1]]>'.length);
+    assertPosition(test.lastChild, 4, 5);
+  },
+}).run();