From 8103dc5ade56e1fbb90d1b32fa22d06e62ba54cf Mon Sep 17 00:00:00 2001
From: Scott Taylor <wonderboymusic@git.wordpress.org>
Date: Thu, 26 Jun 2014 17:14:47 +0000
Subject: [PATCH] Use less greedy regex in `wptexturize()`. Adds unit tests.

Props miqrogroove.
See #28564.


git-svn-id: https://develop.svn.wordpress.org/trunk@28852 602fd350-edb4-49c9-b593-d223f7449a82
---
 src/wp-includes/formatting.php                 |  8 ++++----
 tests/phpunit/tests/formatting/WPTexturize.php | 16 ++++++++++++++++
 2 files changed, 20 insertions(+), 4 deletions(-)

diff --git a/src/wp-includes/formatting.php b/src/wp-includes/formatting.php
index 5af533932f..77ecbcda72 100644
--- a/src/wp-includes/formatting.php
+++ b/src/wp-includes/formatting.php
@@ -198,7 +198,7 @@ function wptexturize($text, $reset = false) {
 		.	'(?(?=!--)'	// Is this a comment?
 		.		'.+?--\s*>'	// Find end of comment
 		.	'|'
-		.		'.+?>'		// Find end of element
+		.		'[^>]+>'	// Find end of element
 		.	')'
 		. '|'
 		.	'\['		// Find start of shortcode.
@@ -206,7 +206,7 @@ function wptexturize($text, $reset = false) {
 		.	'(?:'
 		.		'[^\[\]<>]'	// Shortcodes do not contain other shortcodes.
 		.	'|'
-		.		'<.+?>' 	// HTML elements permitted. Prevents matching ] before >.
+		.		'<[^>]+>' 	// HTML elements permitted. Prevents matching ] before >.
 		.	')+'
 		.	'\]'		// Find end of shortcode.
 		.	'\]?'		// Shortcodes may end with ]]
@@ -224,12 +224,12 @@ function wptexturize($text, $reset = false) {
 				_wptexturize_pushpop_element( $curl, $no_texturize_tags_stack, $no_texturize_tags );
 			}
 
-		} elseif ( '[' === $first && 1 === preg_match( '/^\[(?:[^\[\]<>]|<.+?>)+\]$/', $curl ) ) {
+		} elseif ( '[' === $first && 1 === preg_match( '/^\[(?:[^\[\]<>]|<[^>]+>)+\]$/', $curl ) ) {
 			// This is a shortcode delimeter.
 
 			_wptexturize_pushpop_element( $curl, $no_texturize_shortcodes_stack, $no_texturize_shortcodes );
 
-		} elseif ( '[' === $first && 1 === preg_match( '/^\[\[?(?:[^\[\]<>]|<.+?>)+\]\]?$/', $curl ) ) {
+		} elseif ( '[' === $first && 1 === preg_match( '/^\[\[?(?:[^\[\]<>]|<[^>]+>)+\]\]?$/', $curl ) ) {
 			// This is an escaped shortcode delimeter.
 
 			// Do not texturize.
diff --git a/tests/phpunit/tests/formatting/WPTexturize.php b/tests/phpunit/tests/formatting/WPTexturize.php
index ea22b62500..7ba7346251 100644
--- a/tests/phpunit/tests/formatting/WPTexturize.php
+++ b/tests/phpunit/tests/formatting/WPTexturize.php
@@ -1305,6 +1305,22 @@ class Tests_Formatting_WPTexturize extends WP_UnitTestCase {
 				'word <img src="http://example.com/wp-content/uploads/2014/06/image-300x216.gif" /> word', // Ensure we are not corrupting image URLs.
 				'word <img src="http://example.com/wp-content/uploads/2014/06/image-300x216.gif" /> word',
 			),
+			array(
+				'[ do texturize "[quote]" here ]',
+				'[ do texturize &#8220;[quote]&#8221; here ]',
+			),
+			array(
+				'[ regex catches this <a href="[quote]">here</a> ]',
+				'[ regex catches this <a href="[quote]">here</a> ]',
+			),
+			array(
+				'[ but also catches the <b>styled "[quote]" here</b> ]',
+				'[ but also catches the <b>styled &#8220;[quote]&#8221; here</b> ]',
+			),
+			array(
+				'[Let\'s get crazy<input>[plugin code="<a href=\'?a[]=100\'>hello</a>"]</input>world]',
+				'[Let&#8217;s get crazy<input>[plugin code="<a href=\'?a[]=100\'>hello</a>"]</input>world]',
+			),
 		);
 	}