Added support for ambiguous ampersands

Author: anthonyjb

build/html-string.js

@@ -0,0 +1,112 @@
+0 info it worked if it ends with ok
+1 verbose cli [ '/usr/local/bin/node', '/usr/local/bin/npm', 'publish' ]
+2 info using [email protected]
+3 info using [email protected]
+4 verbose publish [ '.' ]
+5 verbose cache add [ '.', null ]
+6 verbose cache add name=undefined spec="." args=[".",null]
+7 verbose parsed url { protocol: null,
+7 verbose parsed url   slashes: null,
+7 verbose parsed url   auth: null,
+7 verbose parsed url   host: null,
+7 verbose parsed url   port: null,
+7 verbose parsed url   hostname: null,
+7 verbose parsed url   hash: null,
+7 verbose parsed url   search: null,
+7 verbose parsed url   query: null,
+7 verbose parsed url   pathname: '.',
+7 verbose parsed url   path: '.',
+7 verbose parsed url   href: '.' }
+8 silly lockFile 3a52ce78- .
+9 verbose lock . /home/anthony/.npm/3a52ce78-.lock
+10 verbose tar pack [ '/home/anthony/.npm/HTMLString/1.0.6/package.tgz', '.' ]
+11 verbose tarball /home/anthony/.npm/HTMLString/1.0.6/package.tgz
+12 verbose folder .
+13 info prepublish [email protected]
+14 silly lockFile 1f1177db-tar tar://.
+15 verbose lock tar://. /home/anthony/.npm/1f1177db-tar.lock
+16 silly lockFile 423837ce-npm-HTMLString-1-0-6-package-tgz tar:///home/anthony/.npm/HTMLString/1.0.6/package.tgz
+17 verbose lock tar:///home/anthony/.npm/HTMLString/1.0.6/package.tgz /home/anthony/.npm/423837ce-npm-HTMLString-1-0-6-package-tgz.lock
+18 silly lockFile 1f1177db-tar tar://.
+19 silly lockFile 1f1177db-tar tar://.
+20 silly lockFile 423837ce-npm-HTMLString-1-0-6-package-tgz tar:///home/anthony/.npm/HTMLString/1.0.6/package.tgz
+21 silly lockFile 423837ce-npm-HTMLString-1-0-6-package-tgz tar:///home/anthony/.npm/HTMLString/1.0.6/package.tgz
+22 silly lockFile b99c44c8-ony-npm-HTMLString-1-0-6-package /home/anthony/.npm/HTMLString/1.0.6/package
+23 verbose lock /home/anthony/.npm/HTMLString/1.0.6/package /home/anthony/.npm/b99c44c8-ony-npm-HTMLString-1-0-6-package.lock
+24 silly lockFile b99c44c8-ony-npm-HTMLString-1-0-6-package /home/anthony/.npm/HTMLString/1.0.6/package
+25 silly lockFile b99c44c8-ony-npm-HTMLString-1-0-6-package /home/anthony/.npm/HTMLString/1.0.6/package
+26 silly lockFile 3a52ce78- .
+27 silly lockFile 3a52ce78- .
+28 silly publish { name: 'HTMLString',
+28 silly publish   description: 'An HTML parser written in JavaScript that\'s probably not what you\'re looking for.',
+28 silly publish   version: '1.0.6',
+28 silly publish   keywords: [ 'html', 'parser' ],
+28 silly publish   author:
+28 silly publish    { name: 'Anthony Blackshaw',
+28 silly publish      email: '[email protected]',
+28 silly publish      url: 'https://github.com/anthonyjb' },
+28 silly publish   main: 'build/html-string.js',
+28 silly publish   devDependencies:
+28 silly publish    { grunt: '~0.4.5',
+28 silly publish      'grunt-contrib-clean': '^0.6.0',
+28 silly publish      'grunt-contrib-coffee': '^0.11.1',
+28 silly publish      'grunt-contrib-concat': '^0.5.0',
+28 silly publish      'grunt-contrib-jasmine': '^0.9.2',
+28 silly publish      'grunt-contrib-uglify': '^0.5.1',
+28 silly publish      'grunt-contrib-watch': '^0.6.1' },
+28 silly publish   scripts: { test: 'grunt jasmine --verbose' },
+28 silly publish   repository:
+28 silly publish    { type: 'git',
+28 silly publish      url: 'https://github.com/GetmeUK/HTMLString.git' },
+28 silly publish   license: 'MIT',
+28 silly publish   readme: '# HTMLString\n\n[![Build Status](https://travis-ci.org/GetmeUK/HTMLString.svg?branch=master)](https://travis-ci.org/GetmeUK/HTMLString)\n\n> An HTML parser written in JavaScript that\'s probably not what you\'re looking for.\n\n## Install\n\n**Using bower**\n\n```\nbower install --save HTMLString\n```\n\n**Using npm**\n\n```\nnpm install --save HTMLString\n```\n\n## Building\nTo build the library you\'ll need to use Grunt. First install the required node modules ([grunt-cli](http://gruntjs.com/getting-started) must be installed):\n```\ngit clone https://github.com/GetmeUK/HTMLString.git\ncd HTMLString\nnpm install\n```\n\nThen run `grunt build` to build the project.\n\n## Testing\nTo test the library you\'ll need to use Jasmine. First install Jasmine:\n```\ngit clone https://github.com/pivotal/jasmine.git\nmkdir HTMLString/jasmine\nmv jasmine/dist/jasmine-standalone-2.0.3.zip HTMLString/jasmine\ncd HTMLString/jasmine\nunzip jasmine-standalone-2.0.3.zip\n```\n\nThen open `HTMLString/SpecRunner.html` in a browser to run the tests.\n\nAlternatively you can use `grunt jasmine` to run the tests from the command line.\n\n## Documentation\nFull documentation is available at http://getcontenttools.com/api/html-string\n\n## Browser support\n- Chrome\n- Firefox\n- IE9+\n',
+28 silly publish   readmeFilename: 'README.md',
+28 silly publish   gitHead: '8ca1028b8f00c0ae20e25887e57999c7eb6fb03b',
+28 silly publish   bugs: { url: 'https://github.com/GetmeUK/HTMLString/issues' },
+28 silly publish   homepage: 'https://github.com/GetmeUK/HTMLString',
+28 silly publish   _id: '[email protected]',
+28 silly publish   _shasum: '0dbd2b90ad9ff5634d21f366ecbf9c4954477425',
+28 silly publish   _from: '.' }
+29 verbose request where is /HTMLString
+30 verbose request registry https://registry.npmjs.org/
+31 verbose request id 9915c72d0522b64e
+32 verbose url raw /HTMLString
+33 verbose url resolving [ 'https://registry.npmjs.org/', './HTMLString' ]
+34 verbose url resolved https://registry.npmjs.org/HTMLString
+35 verbose request where is https://registry.npmjs.org/HTMLString
+36 info trying registry request attempt 1 at 21:57:55
+37 http PUT https://registry.npmjs.org/HTMLString
+38 http 401 https://registry.npmjs.org/HTMLString
+39 verbose headers { 'content-type': 'application/json',
+39 verbose headers   'cache-control': 'max-age=300',
+39 verbose headers   'content-length': '42',
+39 verbose headers   'accept-ranges': 'bytes',
+39 verbose headers   date: 'Sun, 23 Oct 2016 20:57:49 GMT',
+39 verbose headers   via: '1.1 varnish',
+39 verbose headers   connection: 'keep-alive',
+39 verbose headers   'x-served-by': 'cache-lcy1130-LCY',
+39 verbose headers   'x-cache': 'MISS',
+39 verbose headers   'x-cache-hits': '0',
+39 verbose headers   'x-timer': 'S1477256268.696056,VS0,VE1166',
+39 verbose headers   vary: 'Accept-Encoding' }
+40 error publish Failed PUT 401
+41 error Error: Could not authenticate getmeuk : HTMLString
+41 error     at RegClient.<anonymous> (/usr/local/lib/node_modules/npm/node_modules/npm-registry-client/lib/request.js:308:14)
+41 error     at Request._callback (/usr/local/lib/node_modules/npm/node_modules/npm-registry-client/lib/request.js:246:65)
+41 error     at Request.self.callback (/usr/local/lib/node_modules/npm/node_modules/request/request.js:236:22)
+41 error     at Request.emit (events.js:98:17)
+41 error     at Request.<anonymous> (/usr/local/lib/node_modules/npm/node_modules/request/request.js:1142:14)
+41 error     at Request.emit (events.js:117:20)
+41 error     at IncomingMessage.<anonymous> (/usr/local/lib/node_modules/npm/node_modules/request/request.js:1096:12)
+41 error     at IncomingMessage.emit (events.js:117:20)
+41 error     at _stream_readable.js:943:16
+41 error     at process._tickCallback (node.js:419:13)
+42 error If you need help, you may report this *entire* log,
+42 error including the npm and node versions, at:
+42 error     <http://github.com/npm/npm/issues>
+43 error System Linux 3.13.0-100-generic
+44 error command "/usr/local/bin/node" "/usr/local/bin/npm" "publish"
+45 error cwd /home/anthony/Desktop/Work/Public/CoffeeScript/github/HTMLString
+46 error node -v v0.10.33
+47 error npm -v 1.4.28
+48 verbose exit [ 1, true ]

build/html-string.min.js

@@ -1,7 +1,7 @@
 {
   "name": "HTMLString",
   "description": "An HTML parser written in JavaScript that's probably not what you're looking for.",
-  "version": "1.0.6",
+  "version": "1.0.7",
   "keywords": [
     "html",
     "parser"

npm-debug.log

@@ -8,7 +8,8 @@
     Turing: '<q id="turings-quote">Machines take me by <br> <span class="suprised">surprise</span> with <i>great&nbsp;frequency.</i></q>',
     Wozniak: 'all the best people in life seem to like LINUX.',
     WozniakNamespaced: 'all the best people in life seem to like <ns:tag ns:attr="foo">LINUX</ns:tag>.',
-    WozniakWhitespace: 'all    the best people in life seem to like LINUX.'
+    WozniakWhitespace: 'all    the best people in life seem to like LINUX.',
+    AmbiguousAmpersand: '&amp; &<a href="/foo?bar=1&zee=2&amp;omm=3&end">amp</a> &foo && &&amp; &end'
   };
 
   describe('HTMLString.String()', function() {
@@ -22,12 +23,17 @@
       string = new HTMLString.String(quotes.WozniakWhitespace, true);
       return expect(string.text()).toBe(quotes.WozniakWhitespace);
     });
-    return it('should parse and render a string (HTML)', function() {
+    it('should parse and render a string (HTML)', function() {
       var string;
       string = new HTMLString.String(quotes.Turing);
       expect(string.html()).toBe(quotes.Turing);
       string = new HTMLString.String(quotes.WozniakNamespaced);
-      expect(string.html()).toBe(quotes.WozniakNamespaced);
+      return expect(string.html()).toBe(quotes.WozniakNamespaced);
+    });
+    return it('should parse and render a string (HTML with ambiguous ampersands)', function() {
+      var string;
+      string = new HTMLString.String(quotes.AmbiguousAmpersand);
+      expect(string.html()).toBe(quotes.AmbiguousAmpersand);
       return console.log(string.html());
     });
   });

package.json

@@ -21,6 +21,9 @@ all the best people in life seem to like <ns:tag ns:attr="foo">LINUX</ns:tag>.
     WozniakWhitespace: '''
 all    the best people in life seem to like LINUX.
     '''
+    AmbiguousAmpersand: '''
+&amp; &<a href="/foo?bar=1&zee=2&amp;omm=3&end">amp</a> &foo && &&amp; &end
+    '''
 }
 
 describe 'HTMLString.String()', () ->
@@ -41,6 +44,10 @@ describe 'HTMLString.String()', () ->
         string = new HTMLString.String(quotes.WozniakNamespaced)
         expect(string.html()).toBe quotes.WozniakNamespaced
 
+    it 'should parse and render a string (HTML with ambiguous ampersands)', () ->
+        string = new HTMLString.String(quotes.AmbiguousAmpersand)
+        expect(string.html()).toBe quotes.AmbiguousAmpersand
+
         console.log string.html()
 
 

spec/html-string-spec.js

@@ -659,6 +659,7 @@ class _Parser
 
         @fsm.addTransition '<', CHAR_OR_ENTITY_OR_TAG, OPENNING_OR_CLOSING_TAG
         @fsm.addTransition '&', CHAR_OR_ENTITY_OR_TAG, ENTITY
+        @fsm.addTransition 'END', CHAR_OR_ENTITY_OR_TAG, null
 
         # Entity
         @fsm.addTransitions ENTITY_CHARS, ENTITY, null, (c) ->
@@ -668,6 +669,19 @@ class _Parser
             @_pushChar("&#{ @entity };")
             @entity = ''
 
+        @fsm.addTransitionAny ENTITY, CHAR_OR_ENTITY_OR_TAG, (c) ->
+            @_pushChar('&')
+            for c in @entity.split('')
+                @_pushChar(c)
+            @entity = ''
+            @_back()
+
+        @fsm.addTransition 'END', ENTITY, null, () ->
+            @_pushChar('&')
+            for c in @entity.split('')
+                @_pushChar(c)
+            @entity = ''
+
         # Opening or closing Tag
         @fsm.addTransitions [' ', '\n'], OPENNING_OR_CLOSING_TAG
         @fsm.addTransitions ALPHA_CHARS, OPENNING_OR_CLOSING_TAG, OPENING_TAG, () ->
@@ -779,7 +793,7 @@ class _Parser
         @fsm.addTransitions ENTITY_CHARS, ATTR_ENTITY_NO_DELIM, null, (c) ->
             @entity += c
 
-        @fsm.addTransitions ENTITY_CHARS, ATTR_ENTITY_SINGLE_DELIM, (c) ->
+        @fsm.addTransitions ENTITY_CHARS, ATTR_ENTITY_SINGLE_DELIM, null, (c) ->
             @entity += c
 
         @fsm.addTransitions ENTITY_CHARS, ATTR_ENTITY_DOUBLE_DELIM, null, (c) ->
@@ -797,6 +811,21 @@ class _Parser
             @attributeValue += "&#{ @entity };"
             @entity = ''
 
+        @fsm.addTransitionAny ATTR_ENTITY_NO_DELIM, ATTR_VALUE_NO_DELIM, (c) ->
+             @attributeValue += '&' + @entity
+             @entity = ''
+             @_back()
+
+        @fsm.addTransitionAny ATTR_ENTITY_SINGLE_DELIM, ATTR_VALUE_SINGLE_DELIM, (c) ->
+             @attributeValue += '&' + @entity
+             @entity = ''
+             @_back()
+
+        @fsm.addTransitionAny ATTR_ENTITY_DOUBLE_DELIM, ATTR_VALUE_DOUBLE_DELIM, (c) ->
+             @attributeValue += '&' + @entity
+             @entity = ''
+             @_back()
+
     # Parsing methods
 
     _back: () ->
@@ -896,6 +925,8 @@ class _Parser
 
             @head++
 
+        @fsm.process('END')
+
         return @string
 
     preprocess: (html) ->