From a9072c98beb8944b7f08cd66100a58e3c20769bc Mon Sep 17 00:00:00 2001
From: csteipp <csteipp@wikimedia.org>
Date: Tue, 12 Nov 2013 13:35:31 -0800
Subject: [PATCH] SECURITY: Improve css javascript detection

* Forbid vertical tabs
* Convert Fullwidth, sup/sub script, IPA, and repetition unicode to
ascii, for ie6

Bug: 55332
Change-Id: I41a71b5b8fbecadd0f958cf57cc90d4c2fd9366e
---
 includes/Sanitizer.php       | 46 ++++++++++++++++++++++++++++++-
 tests/parser/parserTests.txt | 64 ++++++++++++++++++++++++++++++++++++++++++++
 2 files changed, 109 insertions(+), 1 deletion(-)

diff --git a/includes/Sanitizer.php b/includes/Sanitizer.php
index 196abd9..50ac84d 100644
--- a/includes/Sanitizer.php
+++ b/includes/Sanitizer.php
@@ -882,6 +882,21 @@ class Sanitizer {
 		$value = preg_replace_callback( $decodeRegex,
 			array( __CLASS__, 'cssDecodeCallback' ), $value );
 
+		// Normalize Halfwidth and Fullwidth Unicode block that IE6 might treat as ascii
+		$value = preg_replace_callback(
+			'/[！-ｚ]/u', // U+FF01 to U+FF5A
+			array( __CLASS__, 'cssNormalizeUnicodeWidth' ),
+			$value
+		);
+
+		// Convert more characters IE6 might treat as ascii
+		// U+0280, U+0274, U+207F, U+029F, U+026A, U+207D, U+208D
+		$value = str_replace(
+			array( 'ʀ', 'ɴ', 'ⁿ', 'ʟ', 'ɪ', '⁽', '₍' ),
+			array( 'r', 'n', 'n', 'l', 'i', '(', '(' ),
+			$value
+		);
+
 		// Remove any comments; IE gets token splitting wrong
 		// This must be done AFTER decoding character references and
 		// escape sequences, because those steps can introduce comments
@@ -897,8 +912,24 @@ class Sanitizer {
 			$value = substr( $value, 0, $commentPos );
 		}
 
+		// S followed by repeat, iteration, or prolonged sound marks,
+		// which IE will treat as "ss"
+		$value = preg_replace(
+			'/s(?:
+				\xE3\x80\xB1 | # U+3031
+				\xE3\x82\x9D | # U+309D
+				\xE3\x83\xBC | # U+30FC
+				\xE3\x83\xBD | # U+30FD
+				\xEF\xB9\xBC | # U+FE7C
+				\xEF\xB9\xBD | # U+FE7D
+				\xEF\xBD\xB0   # U+FF70
+			)/ix',
+			'ss',
+			$value
+		);
+
 		// Reject problematic keywords and control characters
-		if ( preg_match( '/[\000-\010\016-\037\177]/', $value ) ) {
+		if ( preg_match( '/[\000-\010\013\016-\037\177]/', $value ) ) {
 			return '/* invalid control char */';
 		} elseif ( preg_match( '! expression | filter\s*: | accelerator\s*: | url\s*\( !ix', $value ) ) {
 			return '/* insecure input */';
@@ -907,6 +938,19 @@ class Sanitizer {
 	}
 
 	/**
+	 * Normalize Unicode U+FF01 to U+FF5A
+	 * @param character $char
+	 * @return character in ASCII range \x21-\x7A
+	 */
+	static function cssNormalizeUnicodeWidth( $matches ) {
+		$cp = utf8ToCodepoint( $matches[0] );
+		if ( $cp === false ) {
+			return '';
+		}
+		return chr( $cp - 65248 ); // ASCII range \x21-\x7A
+	}
+
+	/**
 	 * @param $matches array
 	 * @return String
 	 */
diff --git a/tests/parser/parserTests.txt b/tests/parser/parserTests.txt
index 11a5516..527c10c 100644
--- a/tests/parser/parserTests.txt
+++ b/tests/parser/parserTests.txt
@@ -5059,6 +5059,70 @@ MSIE CSS safety test: comment in expression
 
 !! end
 
+!! test
+CSS safety test: vertical tab
+!! input
+<p style="font-size: 100px; background-image:url\b(https://www.google.com/images/srpr/logo6w.png)">A</p>
+!! result
+<p style="/* invalid control char */">A</p>
+
+!! end
+
+!! test
+MSIE CSS safety test: Fullwidth
+!! input
+<p style="font-size: 100px; color: ｅｘｐｒｅｓｓｉｏｎ((title='XSSed'),'red')">A</p>
+<div style="top:ＥＸＰＲＥＳＳＩＯＮ(alert())">B</div>
+!! result
+<p style="/* insecure input */">A</p>
+<div style="/* insecure input */">B</div>
+
+!! end
+
+!! test
+MSIE CSS safety test: IPA extensions
+!! input
+<div style="background-image:uʀʟ(javascript:alert())">A</div>
+<p style="font-size: 100px; color: expʀessɪoɴ((title='XSSed'),'red')">B</p>
+!! result
+<div style="/* insecure input */">A</div>
+<p style="/* insecure input */">B</p>
+
+!! end
+
+!! test
+MSIE CSS safety test: sup/sub script
+!! input
+<div style="background-image:url⁽javascript:alert())">A</div>
+<div style="background-image:url₍javascript:alert())">B</div>
+<p style="font-size: 100px; color: expressioⁿ((title='XSSed'),'red')">C</p>
+!! result
+<div style="/* insecure input */">A</div>
+<div style="/* insecure input */">B</div>
+<p style="/* insecure input */">C</p>
+
+!! end
+
+!! test
+MSIE CSS safety test: Repetition markers
+!! input
+<p style="font-size: 100px; color: expres〱ion((title='XSSed'),'red')">A</p>
+<p style="font-size: 100px; color: expresゝion((title='XSSed'),'red')">B</p>
+<p style="font-size: 100px; color: expresーion((title='XSSed'),'red')">C</p>
+<p style="font-size: 100px; color: expresヽion((title='XSSed'),'red')">D</p>
+<p style="font-size: 100px; color: expresﹽion((title='XSSed'),'red')">E</p>
+<p style="font-size: 100px; color: expresﹼion((title='XSSed'),'red')">F</p>
+<p style="font-size: 100px; color: expresｰion((title='XSSed'),'red')">G</p>
+!! result
+<p style="/* insecure input */">A</p>
+<p style="/* insecure input */">B</p>
+<p style="/* insecure input */">C</p>
+<p style="/* insecure input */">D</p>
+<p style="/* insecure input */">E</p>
+<p style="/* insecure input */">F</p>
+<p style="/* insecure input */">G</p>
+
+!! end
 
 !! test
 Table attribute legitimate extension
-- 
1.8.1.4