From d37c017994d2771b41209f571874b096c77e785d Mon Sep 17 00:00:00 2001
From: Fomafix <fomafix@googlemail.com>
Date: Wed, 11 Apr 2018 06:23:50 +0200
Subject: [PATCH] CSSMin::serializeStringValue: Update implementation to new
 specification

The current version of https://www.w3.org/TR/cssom/ is
https://www.w3.org/TR/2016/WD-cssom-1-20160317/

The new specification for CSS string serialization
https://www.w3.org/TR/2016/WD-cssom-1-20160317/#serialize-a-string
has some changes compared to the old specification
https://www.w3.org/TR/2013/WD-cssom-20131205/#serialize-a-string
* U+0000 get replaced by the REPLACEMENT CHARACTER (U+FFFD) instead of
  throwing an exception.
* U+0080 to U+009F are not escaped.

The old implementation has a bug because it selects the byte range
[\x7f-\x9f] and not unicode codepoint range [\u007f-\u009f]. This
breaks the encoding because CSS is in UTF-8 not in ISO 8859-x.

Also add tests to cover CSSMin::serializeStringValue.

Bug: T192048
Change-Id: I894824c216b95dbba461308488fba33121ffea54
---
 includes/libs/CSSMin.php                   | 10 ++----
 tests/phpunit/includes/libs/CSSMinTest.php | 37 ++++++++++++++++++++++
 2 files changed, 40 insertions(+), 7 deletions(-)

diff --git a/includes/libs/CSSMin.php b/includes/libs/CSSMin.php
index 1cbcbde1e0..73825e82e3 100644
--- a/includes/libs/CSSMin.php
+++ b/includes/libs/CSSMin.php
@@ -173,18 +173,14 @@ class CSSMin {
 
 	/**
 	 * Serialize a string (escape and quote) for use as a CSS string value.
-	 * http://www.w3.org/TR/2013/WD-cssom-20131205/#serialize-a-string
+	 * https://www.w3.org/TR/2016/WD-cssom-1-20160317/#serialize-a-string
 	 *
 	 * @param string $value
 	 * @return string
-	 * @throws Exception
 	 */
 	public static function serializeStringValue( $value ) {
-		if ( strstr( $value, "\0" ) ) {
-			throw new Exception( "Invalid character in CSS string" );
-		}
-		$value = strtr( $value, [ '\\' => '\\\\', '"' => '\\"' ] );
-		$value = preg_replace_callback( '/[\x01-\x1f\x7f-\x9f]/', function ( $match ) {
+		$value = strtr( $value, [ "\0" => "\\fffd ", '\\' => '\\\\', '"' => '\\"' ] );
+		$value = preg_replace_callback( '/[\x01-\x1f\x7f]/', function ( $match ) {
 			return '\\' . base_convert( ord( $match[0] ), 10, 16 ) . ' ';
 		}, $value );
 		return '"' . $value . '"';
diff --git a/tests/phpunit/includes/libs/CSSMinTest.php b/tests/phpunit/includes/libs/CSSMinTest.php
index f2d5ef379b..59a1c7c9b7 100644
--- a/tests/phpunit/includes/libs/CSSMinTest.php
+++ b/tests/phpunit/includes/libs/CSSMinTest.php
@@ -22,6 +22,43 @@ class CSSMinTest extends MediaWikiTestCase {
 		] );
 	}
 
+	/**
+	 * @dataProvider serializeStringValueProvider
+	 * @covers CSSMin::serializeStringValue
+	 */
+	public function testSerializeStringValue( $input, $expected ) {
+		$output = CSSMin::serializeStringValue( $input );
+		$this->assertEquals(
+			$expected,
+			$output,
+			'Serialized output must be in the expected form.'
+		);
+	}
+
+	public function serializeStringValueProvider() {
+		return [
+			[ 'Hello World!', '"Hello World!"' ],
+			[ "Null\0Null", "\"Null\\fffd Null\"" ],
+			[ '"', '"\\""' ],
+			[ "'", '"\'"' ],
+			[ "\\", '"\\\\"' ],
+			[ "Tab\tTab", '"Tab\\9 Tab"' ],
+			[ "Space  tab \t space", '"Space  tab \\9  space"' ],
+			[ "Line\nfeed", '"Line\\a feed"' ],
+			[ "Return\rreturn", '"Return\\d return"' ],
+			[ "Next\xc2\x85line", "\"Next\xc2\x85line\"" ],
+			[ "Del\x7fDel", '"Del\\7f Del"' ],
+			[ "nb\xc2\xa0sp", "\"nb\xc2\xa0sp\"" ],
+			[ "AMP&amp;AMP", "\"AMP&amp;AMP\"" ],
+			[ '!"#$%&\'()*+,-./0123456789:;<=>?', '"!\\"#$%&\'()*+,-./0123456789:;<=>?"' ],
+			[ '@[\\]^_`{|}~', '"@[\\\\]^_`{|}~"' ],
+			[ 'ä', '"ä"' ],
+			[ 'Ä', '"Ä"' ],
+			[ '€', '"€"' ],
+			[ '𝒞', '"𝒞"' ], // U+1D49E 'MATHEMATICAL SCRIPT CAPITAL C'
+		];
+	}
+
 	/**
 	 * @dataProvider mimeTypeProvider
 	 * @covers CSSMin::getMimeType
-- 
2.20.1