jgexml - npm Package Compare versions

Comparing version 0.1.0 to 0.1.1

jgeXml.js

 /*
 
-stacking xml parser
+The Just-Good-Enough XML Parser
 
@@ -92,4 +92,6 @@ */
 
-		if ((c == '\t') || (c == '\r') || (c == '\n')) { //other unicode spaces are not treated as whitespace
-			c = ' ';
+		if (context.state != sContent) {
+			if ((c == '\t') || (c == '\r') || (c == '\n')) { //other unicode spaces are not treated as whitespace
+				c = ' ';
+			}
 		}
@@ -99,5 +101,8 @@
 
-			context.token = context.token.trim(); // nonstandard space handling
+			if ((context.state != sValue) && (context.state != sComment)) { // && (context.state != sContent)
+				context.token = context.token.trim();
+			}
+
 			context.keepToken = false;
-			if (((context.state & 1) == 1) && (context.token != '')) {
+			if (((context.state & 1) == 1) && ((context.token.trim() != '') || context.state == sValue)) {
 				context.token = context.token.replaceAll('&','&');
@@ -118,3 +123,3 @@ context.token = context.token.replaceAll('"','"');
 				}
-				
+
 				if (context.state == sComment) {
@@ -121,0 +126,0 @@ context.token = context.token.replaceAll('-- ','');

package.json

 {
   "name": "jgexml",
-  "version": "0.1.0",
+  "version": "0.1.1",
   "description": "The Just-Good-Enough XML Parser",
@@ -5,0 +5,0 @@ "main": "jgeXml.js",

pullparser.js

@@ -13,3 +13,3 @@ 'use strict';
 var context = {};
-var prefixLen = 0;
+var depth = 0;
 
@@ -20,9 +20,10 @@ while (!context.state || context.state != jgeXml.sEndDocument) {
 		if (context.state == jgeXml.sElement) {
-			prefixLen += 2;
+			depth++;
 		}
 		else if (context.state == jgeXml.sEndElement) {
-			prefixLen -= 2;
+			depth--;
 		}
-		console.log(jgeXml.getStateName(context.state)+' '+context.position+' '+prefixLen+' '+context.token);
+		console.log(jgeXml.getStateName(context.state)+' '+context.position+' '+depth+' '+context.token);
 	}
+	if (depth != 0) process.exitCode = 1;
 }

pushparser.js

@@ -12,12 +12,13 @@ 'use strict';
 
-var prefixLen = 0;
+var depth = 0;
 
 jgeXml.parse(xml,function(state,token){
 	if (state == jgeXml.sElement) {
-		prefixLen += 2;
+		depth++;
 	}
 	else if (state == jgeXml.sEndElement) {
-		prefixLen -= 2;
+		depth--;
 	}
-	console.log(jgeXml.getStateName(state)+' '+prefixLen+' '+token);
+	console.log(jgeXml.getStateName(state)+' '+depth+' '+token);
+	if (depth != 0) process.exitCode = 1;
 });

xml2json.js

@@ -16,95 +16,141 @@ 'use strict';
 
-function parseString(xml,attributePrefix) {
+function encode(token) {
+	token = token.replaceAll('\\','\\\\');
+	token = token.replaceAll('\r','\\r');
+	token = token.replaceAll('\n','\\n');
+	token = token.replaceAll('\t','\\t');
+	token = token.replaceAll('"','\\"');
+	return token;
+}
 
-	var s = '{';
-	var hasContent = false;
-	var hasAttribute = false;
+var s;
+
+function getString() {
+	return s;
+}
+
+function newContext() {
+	var context = {};
+	context.position = s.length-1; //mutable
+	context.anchor = context.position; //immutable
+	context.hasContent = false;
+	context.hasAttribute = false;
+	return context;
+}
+
+function dump(stack) {
+	for (var i=0;i<stack.length;i++) {
+		console.log('Entry '+i+' points to '+stack[i].anchor+' ->'+s.charAt(stack[i].anchor)+'<');
+	}
+	console.log('--');
+}
+
+function parseString(xml,attributePrefix,valueProperty) {
+
+	var stack = [];
+	var depth = 0; //depth tracks the depth of XML element nesting, not the output JSON
 	var lastElement = '';
-	var stack = [];
-	var finished = 0;
-	stack.push(1);
 
+	s = '{';
+	stack.push(newContext());
+
 	jgeXml.parse(xml,function(state,token){
 
 		if (state == jgeXml.sContent) {
-			if (token != '') { // maybe move this in to only omit hasContent = true ??
-				if (s.charAt(s.length-1) == '{') {
-					s = s.substr(0,s.length-1);
+			if (token != '') {
+				var closeObject = false;
+				// content should be following a property name not the beginning of an object
+				// so remove assumption it was a container
+				if ((!valueProperty) && (s.charAt(s.length-1) == '{')) {
+					s = s.replaceAt(s.length-1,' ');
 				}
+
+				// if we're following another property, separate with a comma
 				if (s.charAt(s.length-1) == '"') {
 					s += ',';
 				}
-				if (hasAttribute) {
-					s += ' "_" : ';
+				// if we have had attributes, this is definitely a container not a primitive
+				// treat this value as an anonymous property
+				if ((stack[stack.length-1].hasAttribute) && (!valueProperty)) {
+					s += ' "' + attributePrefix + '": ';
 				}
 				else {
-					hasContent = true;
+					if (stack[stack.length-1].hasContent) {
+						//create array for mixed content/elements
+						if (s.charAt(stack[stack.length-1].position) != '[') {
+							s = s.insert(stack[stack.length-1].position,'[');
+						}
+						if (valueProperty) {
+							s += ',{';
+							closeObject = true;
+						}
+					}
+					stack[stack.length-1].hasContent = true;
 				}
-				s += '"' + token.replaceAll('"','\\"') + '"';
+				if (valueProperty) {
+					s += '"value": ';
+				}
+				s += '"' + encode(token) + '"';
+				if (closeObject) {
+					s += '}';
+				}
+
 			}
 		}
 		else if (state == jgeXml.sEndElement) {
-			// drop hanging comma
-			if (s.charAt(s.length-1) == ',') {
-				s = s.substr(0,s.length-1);
-			}
-			// if we're in an array, close it
-			if (s.charAt(stack[stack.length-1]) == '[') {
+			//dump(stack);
+
+			if (s.charAt(stack[stack.length-1].position) == '[') {
+				// if we're in an array, close it
 				s += ']';
 			}
-			//// if we're in an object, close it
-			//if (s.charAt(stack[stack.length-2]) == '{') {
-			//	s += '}';
-			//}
-			// close an empty element
-			if (!hasContent) {
+			if (s.charAt(stack[stack.length-1].anchor) == '{') {
+				// if we're in an object, close it
 				s += '}';
 			}
-			hasContent = false;
-			hasAttribute = false;
-			lastElement = token;
-			finished = stack.pop();
+
+			stack.pop();
+			depth--;
+			lastElement = token+'<'+depth+'>';
 		}
 		else if (state == jgeXml.sAttribute) {
+			// if not the first attribute, separate the properties with a comma
 			if (s.charAt(s.length-1) !== '{') {
 				s += ',';
 			}
-			s += '"' + attributePrefix + token + '" : ';
-			hasAttribute = true;
+			s += '"' + attributePrefix + token + '": ';
+			stack[stack.length-1].hasAttribute = true;
 		}
 		else if (state == jgeXml.sValue) {
-			s += '"' + token.replaceAll('"','\\"') + '"';
+			s += '"' + encode(token) + '"';
 		}
 		else if (state == jgeXml.sElement) {
-			hasAttribute = false;
-			if (s.charAt(s.length-1) !== '{') {
+			// if this is not the first property, separate with a comma
+			if (s.charAt(s.length-1) != '{') {
 				s += ',';
 			}
-			if (token != lastElement) {
-				if (s.charAt(stack[stack.length-1]) == '[') {
+
+			if (token+'<'+depth+'>' != lastElement) {
+				// element has changed, if we're in an array, terminate it just before the comma we (maybe) added
+				if (s.charAt(stack[stack.length-1].position) == '[') {
 					s = s.insert(s.length-1,']');
 				}
 				s += '"' + token + '": ';
-				stack[stack.length-1] = s.length;
-				stack.push(s.length);
-				s += '{';
+				stack[stack.length-1].position = s.length; //update position of previous element (array insertion point)
 			}
 			else {
-				// array detected, new element matches previous endElement
+				// array detected, new element matches previous endElement and at same depth
 				// only need to insert array opening bracket once
-				if (s.charAt(stack[stack.length-1]) != '[') {
-					s = s.insert(stack[stack.length-1],'[');
-					//s = s.replaceAt(stack[stack.length-1],'[');
+				if (s.charAt(stack[stack.length-1].position) != '[') {
+					s = s.insert(stack[stack.length-1].position,'[');
 				}
-				stack.push(s.length);
-				//if (s.charAt(s.length-1) == '{') {
-				//	//anonymous object?
-				//	//s = s.substr(0,s.length-1); // drop previous opening brace
-				//	s += '"anon": ';
-				//}
-				s += '{';
 			}
+			lastElement = token+'<'+depth+'>';
+			s += '{'; // here we assume all elements are containers not values, this supports attributes
+			stack.push(newContext());
+			depth++;
 		}
 	});
 
+	// remove final trailing comma, if any
 	if (s.charAt(s.length-1) == ',') {
@@ -121,5 +167,6 @@ s = s.substr(0,s.length-1);
 module.exports = {
-	xml2json : function(xml,attributePrefix) {
-		return parseString(xml,attributePrefix);
-	}
+	xml2json : function(xml,attributePrefix,valueProperty) {
+		return parseString(xml,attributePrefix,valueProperty);
+	},
+	getString : getString
 }

New alerts

License Policy Violation

License

This package is not allowed per your license policy. Review the package's license to ensure compliance.

Found 1 instance in 1 package

Fixed alerts

License Policy Violation

License

This package is not allowed per your license policy. Review the package's license to ensure compliance.

Found 1 instance in 1 package

Improved metrics

Total package byte prevSize

109236

increased by572.51%

Number of package files

30

increased by172.73%

Lines of code

538

increased by10.7%

No dependency changes