Sanitizer: Allow attribute names to use any Unicode "Letter" or "Number"

[lhc/web/wiklou.git] / tests / phpunit / includes / SanitizerTest.php

diff --git a/tests/phpunit/includes/SanitizerTest.php b/tests/phpunit/includes/SanitizerTest.php
index 12db1a1..abcf1d4 100644 (file)
--- a/tests/phpunit/includes/SanitizerTest.php
+++ b/tests/phpunit/includes/SanitizerTest.php
@@ -178,6 +178,10 @@ class SanitizerTest extends MediaWikiTestCase {
        public static function provideTagAttributesToDecode() {
                return [
                        [ [ 'foo' => 'bar' ], 'foo=bar', 'Unquoted attribute' ],
+                       [ [ 'עברית' => 'bar' ], 'עברית=bar', 'Non-Latin attribute' ],
+                       [ [ '६' => 'bar' ], '६=bar', 'Devanagari number' ],
+                       [ [ '搭𨋢' => 'bar' ], '搭𨋢=bar', 'Non-BMP character' ],
+                       [ [], 'ńgh=bar', 'Combining accent is not allowed' ],
                        [ [ 'foo' => 'bar' ], '    foo   =   bar    ', 'Spaced attribute' ],
                        [ [ 'foo' => 'bar' ], 'foo="bar"', 'Double-quoted attribute' ],
                        [ [ 'foo' => 'bar' ], 'foo=\'bar\'', 'Single-quoted attribute' ],
@@ -211,7 +215,7 @@ class SanitizerTest extends MediaWikiTestCase {
                        [ [ 'foo6' => 'baz' ], 'foo6=baz', 'Numbers are allowed' ],
 
                        # This bit is more relaxed than XML rules, but some extensions use
-                       # it, like ProofreadPage (see bug 27539)
+                       # it, like ProofreadPage (see T29539)
                        [ [ '1foo' => 'baz' ], '1foo=baz', 'Leading numbers are allowed' ],
                        [ [], 'foo$=baz', 'Symbols are not allowed' ],
                        [ [], 'foo@=baz', 'Symbols are not allowed' ],
@@ -286,7 +290,7 @@ class SanitizerTest extends MediaWikiTestCase {
                        [ ' ', '/* /* */' ],
                        [ 'display: block;', "display:/* foo */block;" ],
                        [ 'display: block;', "display:\\2f\\2a foo \\2a\\2f block;",
-                               'Backslash-escaped comments must be stripped (bug 28450)' ],
+                               'Backslash-escaped comments must be stripped (T30450)' ],
                        [ '', '/* unfinished comment structure',
                                'Remove anything after a comment-start token' ],
                        [ '', "\\2f\\2a unifinished comment'",
@@ -362,4 +366,25 @@ class SanitizerTest extends MediaWikiTestCase {
                        [ '+1 +2', '+1', '+2' ],
                ];
        }
+
+       /**
+        * @dataProvider provideIsReservedDataAttribute
+        */
+       public function testIsReservedDataAttribute( $attr, $expected ) {
+               $this->assertSame( $expected, Sanitizer::isReservedDataAttribute( $attr ) );
+       }
+
+       public static function provideIsReservedDataAttribute() {
+               return [
+                       [ 'foo', false ],
+                       [ 'data', false ],
+                       [ 'data-foo', false ],
+                       [ 'data-mw', true ],
+                       [ 'data-ooui', true ],
+                       [ 'data-parsoid', true ],
+                       [ 'data-mw-foo', true ],
+                       [ 'data-ooui-foo', true ],
+                       [ 'data-mwfoo', true ], // could be false but this is how it's implemented currently
+               ];
+       }
 }