From 50159974c9dff73fcfa1015e5eaf3a9b14e4e65c Mon Sep 17 00:00:00 2001
From: Ella Iseulde Van Dorpe <*@iseulde.com>
Date: Tue, 21 Jul 2015 15:24:24 +0000
Subject: [PATCH] Editor: word count: exclude HTML comments and entities

Also make sure `type` is one of the three allowed types and remove unnecessary regex flags.

See #30966.

Built from https://develop.svn.wordpress.org/trunk@33344


git-svn-id: http://core.svn.wordpress.org/trunk@33316 1a063a9b-81f0-0310-95a4-ce76da25c4cd
---
 wp-admin/js/word-count.js     | 17 +++++++++++++----
 wp-admin/js/word-count.min.js |  2 +-
 wp-includes/version.php       |  2 +-
 3 files changed, 15 insertions(+), 6 deletions(-)

diff --git a/wp-admin/js/word-count.js b/wp-admin/js/word-count.js
index 025412e346..c8cfb2f52d 100644
--- a/wp-admin/js/word-count.js
+++ b/wp-admin/js/word-count.js
@@ -14,14 +14,16 @@
 		shortcodes = this.settings.l10n.shortcodes;
 
 		if ( shortcodes && shortcodes.length ) {
-			this.settings.shortcodesRegExp = new RegExp( '\\[\\/?(?:' + shortcodes.join( '|' ) + ')[^\\]]*?\\]', 'gi' );
+			this.settings.shortcodesRegExp = new RegExp( '\\[\\/?(?:' + shortcodes.join( '|' ) + ')[^\\]]*?\\]', 'g' );
 		}
 	}
 
 	WordCounter.prototype.settings = {
 		HTMLRegExp: /<\/?[a-z][^>]*?>/gi,
+		HTMLcommentRegExp: /<!--[\s\S]*?-->/g,
 		spaceRegExp: /&nbsp;|&#160;/gi,
-		connectorRegExp: /--|\u2014/gi,
+		HTMLEntityRegExp: /&\S+?;/g,
+		connectorRegExp: /--|\u2014/g,
 		removeRegExp: new RegExp( [
 			'[',
 				// Basic Latin (extract)
@@ -60,19 +62,24 @@
 		astralRegExp: /[\uD800-\uDBFF][\uDC00-\uDFFF]/g,
 		wordsRegExp: /\S\s+/g,
 		charactersRegExp: /\S/g,
-		allRegExp: /[^\f\n\r\t\v\u00ad\u2028\u2029]/g,
+		allRegExp: /[^\f\n\r\t\v\u00AD\u2028\u2029]/g,
 		l10n: window.wordCountL10n || {}
 	};
 
 	WordCounter.prototype.count = function( text, type ) {
 		var count = 0;
 
-		type = type || this.settings.l10n.type || 'words';
+		type = type || this.settings.l10n.type;
+
+		if ( type !== 'characters' && type !== 'all' ) {
+			type = 'words';
+		}
 
 		if ( text ) {
 			text = text + '\n';
 
 			text = text.replace( this.settings.HTMLRegExp, '\n' );
+			text = text.replace( this.settings.HTMLcommentRegExp, '' );
 
 			if ( this.settings.shortcodesRegExp ) {
 				text = text.replace( this.settings.shortcodesRegExp, '\n' );
@@ -81,9 +88,11 @@
 			text = text.replace( this.settings.spaceRegExp, ' ' );
 
 			if ( type === 'words' ) {
+				text = text.replace( this.settings.HTMLEntityRegExp, '' );
 				text = text.replace( this.settings.connectorRegExp, ' ' );
 				text = text.replace( this.settings.removeRegExp, '' );
 			} else {
+				text = text.replace( this.settings.HTMLEntityRegExp, 'a' );
 				text = text.replace( this.settings.astralRegExp, 'a' );
 			}
 
diff --git a/wp-admin/js/word-count.min.js b/wp-admin/js/word-count.min.js
index 2bca545f40..0b16f0d8b7 100644
--- a/wp-admin/js/word-count.min.js
+++ b/wp-admin/js/word-count.min.js
@@ -1 +1 @@
-!function(){function a(a){var b,c;if(a)for(b in a)a.hasOwnProperty(b)&&(this.settings[b]=a[b]);c=this.settings.l10n.shortcodes,c&&c.length&&(this.settings.shortcodesRegExp=new RegExp("\\[\\/?(?:"+c.join("|")+")[^\\]]*?\\]","gi"))}a.prototype.settings={HTMLRegExp:/<\/?[a-z][^>]*?>/gi,spaceRegExp:/&nbsp;|&#160;/gi,connectorRegExp:/--|\u2014/gi,removeRegExp:new RegExp(["[","!-@[-`{-~","\x80-\xbf\xd7\xf7","\u2000-\u2bff","\u2e00-\u2e7f","]"].join(""),"g"),astralRegExp:/[\uD800-\uDBFF][\uDC00-\uDFFF]/g,wordsRegExp:/\S\s+/g,charactersRegExp:/\S/g,allRegExp:/[^\f\n\r\t\v\u00ad\u2028\u2029]/g,l10n:window.wordCountL10n||{}},a.prototype.count=function(a,b){var c=0;return b=b||this.settings.l10n.type||"words",a&&(a+="\n",a=a.replace(this.settings.HTMLRegExp,"\n"),this.settings.shortcodesRegExp&&(a=a.replace(this.settings.shortcodesRegExp,"\n")),a=a.replace(this.settings.spaceRegExp," "),"words"===b?(a=a.replace(this.settings.connectorRegExp," "),a=a.replace(this.settings.removeRegExp,"")):a=a.replace(this.settings.astralRegExp,"a"),a=a.match(this.settings[b+"RegExp"]),a&&(c=a.length)),c},window.wp=window.wp||{},window.wp.utils=window.wp.utils||{},window.wp.utils.WordCounter=a}();
\ No newline at end of file
+!function(){function a(a){var b,c;if(a)for(b in a)a.hasOwnProperty(b)&&(this.settings[b]=a[b]);c=this.settings.l10n.shortcodes,c&&c.length&&(this.settings.shortcodesRegExp=new RegExp("\\[\\/?(?:"+c.join("|")+")[^\\]]*?\\]","g"))}a.prototype.settings={HTMLRegExp:/<\/?[a-z][^>]*?>/gi,HTMLcommentRegExp:/<!--[\s\S]*?-->/g,spaceRegExp:/&nbsp;|&#160;/gi,HTMLEntityRegExp:/&\S+?;/g,connectorRegExp:/--|\u2014/g,removeRegExp:new RegExp(["[","!-@[-`{-~","\x80-\xbf\xd7\xf7","\u2000-\u2bff","\u2e00-\u2e7f","]"].join(""),"g"),astralRegExp:/[\uD800-\uDBFF][\uDC00-\uDFFF]/g,wordsRegExp:/\S\s+/g,charactersRegExp:/\S/g,allRegExp:/[^\f\n\r\t\v\u00AD\u2028\u2029]/g,l10n:window.wordCountL10n||{}},a.prototype.count=function(a,b){var c=0;return b=b||this.settings.l10n.type,"characters"!==b&&"all"!==b&&(b="words"),a&&(a+="\n",a=a.replace(this.settings.HTMLRegExp,"\n"),a=a.replace(this.settings.HTMLcommentRegExp,""),this.settings.shortcodesRegExp&&(a=a.replace(this.settings.shortcodesRegExp,"\n")),a=a.replace(this.settings.spaceRegExp," "),"words"===b?(a=a.replace(this.settings.HTMLEntityRegExp,""),a=a.replace(this.settings.connectorRegExp," "),a=a.replace(this.settings.removeRegExp,"")):(a=a.replace(this.settings.HTMLEntityRegExp,"a"),a=a.replace(this.settings.astralRegExp,"a")),a=a.match(this.settings[b+"RegExp"]),a&&(c=a.length)),c},window.wp=window.wp||{},window.wp.utils=window.wp.utils||{},window.wp.utils.WordCounter=a}();
\ No newline at end of file
diff --git a/wp-includes/version.php b/wp-includes/version.php
index d32beec285..2e3367ac43 100644
--- a/wp-includes/version.php
+++ b/wp-includes/version.php
@@ -4,7 +4,7 @@
  *
  * @global string $wp_version
  */
-$wp_version = '4.3-beta3-33343';
+$wp_version = '4.3-beta3-33344';
 
 /**
  * Holds the WordPress DB revision, increments when changes are made to the WordPress DB schema.