parse5 - npm Package Compare versions

Comparing version 1.3.2 to 1.4.0

CHANGELOG.md

@@ -0,1 +1,7 @@
+## 1.4.0
+ * Add: Parser [decodeHtmlEntities](https://github.com/inikulin/parse5#optionsdecodehtmlentities) option.
+ * Add: SimpleApiParser [decodeHtmlEntities](https://github.com/inikulin/parse5#optionsdecodehtmlentities-1) option.
+ * Add: Parser [locationInfo](https://github.com/inikulin/parse5#optionslocationinfo) option.
+ * Add: SimpleApiParser [locationInfo](https://github.com/inikulin/parse5#optionslocationinfo-1) option.
+
 ## 1.3.2
@@ -2,0 +8,0 @@ * Fix: `<form>` processing in `<template>` (GH [#40](https://github.com/inikulin/parse5/issues/40))

lib/common/utils.js

@@ -0,0 +0,0 @@ 'use strict';

lib/serialization/serializer.js

@@ -12,3 +12,5 @@ 'use strict';
 //Default serializer options
-var DEFAULT_OPTIONS = {encodeHtmlEntities: true};
+var DEFAULT_OPTIONS = {
+    encodeHtmlEntities: true
+};
 
@@ -159,3 +161,3 @@ //Escaping regexes
 
-    if(parent && this.treeAdapter.isElementNode(parent))
+    if (parent && this.treeAdapter.isElementNode(parent))
         parentTn = this.treeAdapter.getTagName(parent);
@@ -162,0 +164,0 @@

lib/simple_api/simple_api_parser.js

 'use strict';
 
 var Tokenizer = require('../tokenization/tokenizer'),
-    TokenizerProxy = require('./tokenizer_proxy');
+    TokenizerProxy = require('./tokenizer_proxy'),
+    Utils = require('../common/utils');
 
+//Default options
+var DEFAULT_OPTIONS = {
+    decodeHtmlEntities: true,
+    locationInfo: false
+};
+
 //Skipping handler
@@ -12,12 +19,29 @@ function skip() {
 //SimpleApiParser
-var SimpleApiParser = module.exports = function (handlers) {
+var SimpleApiParser = module.exports = function (handlers, options) {
+    this.options = Utils.mergeOptions(DEFAULT_OPTIONS, options);
     this.handlers = {
-        doctype: handlers.doctype || skip,
-        startTag: handlers.startTag || skip,
-        endTag: handlers.endTag || skip,
-        text: handlers.text || skip,
-        comment: handlers.comment || skip
+        doctype: this._wrapHandler(handlers.doctype),
+        startTag: this._wrapHandler(handlers.startTag),
+        endTag: this._wrapHandler(handlers.endTag),
+        text: this._wrapHandler(handlers.text),
+        comment: this._wrapHandler(handlers.comment)
     };
 };
 
+SimpleApiParser.prototype._wrapHandler = function (handler) {
+    var parser = this;
+
+    handler = handler || skip;
+
+    if (this.options.locationInfo) {
+        return function () {
+            var args = Array.prototype.slice.call(arguments);
+            args.push(parser.currentTokenLocation);
+            handler.apply(handler, args);
+        };
+    }
+
+    return handler;
+};
+
 //API
@@ -35,2 +59,11 @@ SimpleApiParser.prototype.parse = function (html) {
             token.type === Tokenizer.NULL_CHARACTER_TOKEN) {
+
+            if (this.options.locationInfo) {
+                if (this.pendingText === null)
+                    this.currentTokenLocation = token.location;
+
+                else
+                    this.currentTokenLocation.end = token.location.end;
+            }
+
             this.pendingText = (this.pendingText || '') + token.chars;
@@ -43,3 +76,3 @@ }
         }
-    } while (token.type !== Tokenizer.EOF_TOKEN)
+    } while (token.type !== Tokenizer.EOF_TOKEN);
 };
@@ -49,2 +82,5 @@
 SimpleApiParser.prototype._handleToken = function (token) {
+    if (this.options.locationInfo)
+        this.currentTokenLocation = token.location;
+
     if (token.type === Tokenizer.START_TAG_TOKEN)
@@ -65,4 +101,5 @@ this.handlers.startTag(token.tagName, token.attrs, token.selfClosing);
 SimpleApiParser.prototype._reset = function (html) {
-    this.tokenizerProxy = new TokenizerProxy(html);
+    this.tokenizerProxy = new TokenizerProxy(html, this.options);
     this.pendingText = null;
+    this.currentTokenLocation = null;
 };
@@ -69,0 +106,0 @@

lib/simple_api/tokenizer_proxy.js

@@ -15,4 +15,4 @@ 'use strict';
 //NOTE: this proxy simulates adjustment of the Tokenizer which performed by standard parser during tree construction.
-var TokenizerProxy = module.exports = function (html) {
-    this.tokenizer = new Tokenizer(html);
+var TokenizerProxy = module.exports = function (html, options) {
+    this.tokenizer = new Tokenizer(html, options);
 
@@ -63,15 +63,15 @@ this.namespaceStack = [];
 //Token handlers
-TokenizerProxy.prototype._ensureTokenizerState = function (tn) {
+TokenizerProxy.prototype._ensureTokenizerMode = function (tn) {
     if (tn === $.TEXTAREA || tn === $.TITLE)
-        this.tokenizer.state = Tokenizer.RCDATA_STATE;
+        this.tokenizer.state = Tokenizer.MODE.RCDATA;
 
     else if (tn === $.PLAINTEXT)
-        this.tokenizer.state = Tokenizer.PLAINTEXT_STATE;
+        this.tokenizer.state = Tokenizer.MODE.PLAINTEXT;
 
     else if (tn === $.SCRIPT)
-        this.tokenizer.state = Tokenizer.SCRIPT_DATA_STATE;
+        this.tokenizer.state = Tokenizer.MODE.SCRIPT_DATA;
 
     else if (tn === $.STYLE || tn === $.IFRAME || tn === $.XMP ||
              tn === $.NOEMBED || tn === $.NOFRAMES || tn === $.NOSCRIPT) {
-        this.tokenizer.state = Tokenizer.RAWTEXT_STATE;
+        this.tokenizer.state = Tokenizer.MODE.RAWTEXT;
     }
@@ -101,3 +101,3 @@ };
         else
-            this._ensureTokenizerState(tn);
+            this._ensureTokenizerMode(tn);
     }
@@ -119,7 +119,8 @@ };
         else if (tn === $.SCRIPT)
-            this.tokenizer.state = Tokenizer.DATA_STATE;
+            this.tokenizer.state = Tokenizer.MODE.DATA;
     }
 
-    else if ((tn === $.SVG && this.currentNamespace === NS.SVG) || (tn === $.MATH && this.currentNamespace === NS.MATHML))
+    else if ((tn === $.SVG && this.currentNamespace === NS.SVG) ||
+             (tn === $.MATH && this.currentNamespace === NS.MATHML))
         this._leaveCurrentNamespace();
 };

package.json

 {
   "name": "parse5",
   "description": "WHATWG HTML5 specification-compliant, fast and ready for production HTML parsing/serialization toolset for Node.",
-  "version": "1.3.2",
+  "version": "1.4.0",
   "author": "Ivan Nikulin <[e-mail]> (https://github.com/inikulin)",
@@ -6,0 +6,0 @@ "contributors": [

README.md

@@ -71,5 +71,11 @@ <p align="center">
 
-#### &bull; Parser.ctor([treeAdapter])
+#### &bull; Parser.ctor([treeAdapter, options])
 Creates new reusable instance of the `Parser`. Optional `treeAdapter` argument specifies resulting tree format. If `treeAdapter` argument is not specified, `default` tree adapter will be used.
 
+`options` object provides the parsing algorithm modifications:
+#####  options.decodeHtmlEntities
+Decode HTML-entities like `&amp;`, `&nbsp;`, etc.  Default: `true`. **Warning:** disabling this option may cause output which is not conform HTML5 specification.
+#####  options.locationInfo
+Enables source code location information for the nodes. Default: `false`. When enabled, each node (except root node) has `__location` property, which contains `start` and `end` indices of the node in the source code.
+
 *Example:*
@@ -114,5 +120,12 @@ ```js
 
-#### &bull; SimpleApiParser.ctor(handlers)
+#### &bull; SimpleApiParser.ctor(handlers, [options])
 Creates new reusable instance of the `SimpleApiParser`. `handlers` argument specifies object that contains parser's event handlers. Possible events and their signatures are shown in the example.
 
+`options` object provides the parsing algorithm modifications:
+##### options.decodeHtmlEntities
+Decode HTML-entities like `&amp;`, `&nbsp;`, etc.  Default: `true`. **Warning:** disabling this option may cause output which is not conform HTML5 specification.
+##### options.locationInfo
+Enables source code location information for the tokens. Default: `false`. When enabled, each node handler receives `location` object as it's last argument. `location` object contains `start` and `end` indices of the token in the source code.
+
+
 *Example:*
@@ -123,19 +136,19 @@ ```js
 var parser = new parse5.SimpleApiParser({
-    doctype: function(name, publicId, systemId) {
+    doctype: function(name, publicId, systemId /*, [location] */) {
         //Handle doctype here
     },
 
-    startTag: function(tagName, attrs, selfClosing) {
+    startTag: function(tagName, attrs, selfClosing /*, [location] */) {
         //Handle start tags here
     },
 
-    endTag: function(tagName) {
+    endTag: function(tagName /*, [location] */) {
         //Handle end tags here
     },
 
-    text: function(text) {
+    text: function(text /*, [location] */) {
         //Handle texts here
     },
 
-    comment: function(text) {
+    comment: function(text /*, [location] */) {
         //Handle comments here
@@ -171,7 +184,7 @@ }
 
-`options` object provides the serialization algorithm modifications (**Warning:** switching default options causes HTML5 specification violation. However, it may be useful in some cases, e.g. markup instrumentation. Use it on your own risk.)
+`options` object provides the serialization algorithm modifications:
+##### options.encodeHtmlEntities
+HTML-encode characters like `<`, `>`, `&`, etc.  Default: `true`.  **Warning:** disabling this option may cause output which is not conform HTML5 specification.
 
-*  **options.encodeHtmlEntities** - HTML-encode characters like `<`, `>`, `&`, etc.  Default: `true`.
 
-
 *Example:*
@@ -178,0 +191,0 @@ ```js

No alert changes

Improved metrics

Total package byte prevSize

396916

increased by3.11%

Number of package files

25

increased by8.7%

Lines of code

6616

increased by3.86%

Number of lines in readme file

247

increased by5.56%

No dependency changes