tests/phpunit/includes/SanitizerTest.php

   1 <?php
   2
   3 /**
   4  * @todo Tests covering decodeCharReferences can be refactored into a single
   5  * method and dataprovider.
   6  */
   7 class SanitizerTest extends MediaWikiTestCase {
   8
   9         protected function setUp() {
  10                 parent::setUp();
  11
  12                 AutoLoader::loadClass( 'Sanitizer' );
  13         }
  14
  15         /**
  16          * @covers Sanitizer::decodeCharReferences
  17          */
  18         public function testDecodeNamedEntities() {
  19                 $this->assertEquals(
  20                         "\xc3\xa9cole",
  21                         Sanitizer::decodeCharReferences( '&eacute;cole' ),
  22                         'decode named entities'
  23                 );
  24         }
  25
  26         /**
  27          * @covers Sanitizer::decodeCharReferences
  28          */
  29         public function testDecodeNumericEntities() {
  30                 $this->assertEquals(
  31                         "\xc4\x88io bonas dans l'\xc3\xa9cole!",
  32                         Sanitizer::decodeCharReferences( "&#x108;io bonas dans l'&#233;cole!" ),
  33                         'decode numeric entities'
  34                 );
  35         }
  36
  37         /**
  38          * @covers Sanitizer::decodeCharReferences
  39          */
  40         public function testDecodeMixedEntities() {
  41                 $this->assertEquals(
  42                         "\xc4\x88io bonas dans l'\xc3\xa9cole!",
  43                         Sanitizer::decodeCharReferences( "&#x108;io bonas dans l'&eacute;cole!" ),
  44                         'decode mixed numeric/named entities'
  45                 );
  46         }
  47
  48         /**
  49          * @covers Sanitizer::decodeCharReferences
  50          */
  51         public function testDecodeMixedComplexEntities() {
  52                 $this->assertEquals(
  53                         "\xc4\x88io bonas dans l'\xc3\xa9cole! (mais pas &#x108;io dans l'&eacute;cole)",
  54                         Sanitizer::decodeCharReferences(
  55                                 "&#x108;io bonas dans l'&eacute;cole! (mais pas &amp;#x108;io dans l'&#38;eacute;cole)"
  56                         ),
  57                         'decode mixed complex entities'
  58                 );
  59         }
  60
  61         /**
  62          * @covers Sanitizer::decodeCharReferences
  63          */
  64         public function testInvalidAmpersand() {
  65                 $this->assertEquals(
  66                         'a & b',
  67                         Sanitizer::decodeCharReferences( 'a & b' ),
  68                         'Invalid ampersand'
  69                 );
  70         }
  71
  72         /**
  73          * @covers Sanitizer::decodeCharReferences
  74          */
  75         public function testInvalidEntities() {
  76                 $this->assertEquals(
  77                         '&foo;',
  78                         Sanitizer::decodeCharReferences( '&foo;' ),
  79                         'Invalid named entity'
  80                 );
  81         }
  82
  83         /**
  84          * @covers Sanitizer::decodeCharReferences
  85          */
  86         public function testInvalidNumberedEntities() {
  87                 $this->assertEquals( UTF8_REPLACEMENT, Sanitizer::decodeCharReferences( "&#88888888888888;" ), 'Invalid numbered entity' );
  88         }
  89
  90         /**
  91          * @covers Sanitizer::removeHTMLtags
  92          * @dataProvider provideHtml5Tags
  93          *
  94          * @param String $tag Name of an HTML5 element (ie: 'video')
  95          * @param Boolean $escaped Wheter sanitizer let the tag in or escape it (ie: '&lt;video&gt;')
  96          */
  97         public function testRemovehtmltagsOnHtml5Tags( $tag, $escaped ) {
  98                 $this->setMwGlobals( array(
  99                         'wgUseTidy' => false
 100                 ) );
 101
 102                 if ( $escaped ) {
 103                         $this->assertEquals( "&lt;$tag&gt;",
 104                                 Sanitizer::removeHTMLtags( "<$tag>" )
 105                         );
 106                 } else {
 107                         $this->assertEquals( "<$tag></$tag>\n",
 108                                 Sanitizer::removeHTMLtags( "<$tag>" )
 109                         );
 110                 }
 111         }
 112
 113         /**
 114          * Provide HTML5 tags
 115          */
 116         public static function provideHtml5Tags() {
 117                 $ESCAPED = true; # We want tag to be escaped
 118                 $VERBATIM = false; # We want to keep the tag
 119                 return array(
 120                         array( 'data', $VERBATIM ),
 121                         array( 'mark', $VERBATIM ),
 122                         array( 'time', $VERBATIM ),
 123                         array( 'video', $ESCAPED ),
 124                 );
 125         }
 126
 127         function dataRemoveHTMLtags() {
 128                 return array(
 129                         // former testSelfClosingTag
 130                         array(
 131                                 '<div>Hello world</div />',
 132                                 '<div>Hello world</div>',
 133                                 'Self-closing closing div'
 134                         ),
 135                         // Make sure special nested HTML5 semantics are not broken
 136                         // http://www.whatwg.org/html/text-level-semantics.html#the-kbd-element
 137                         array(
 138                                 '<kbd><kbd>Shift</kbd>+<kbd>F3</kbd></kbd>',
 139                                 '<kbd><kbd>Shift</kbd>+<kbd>F3</kbd></kbd>',
 140                                 'Nested <kbd>.'
 141                         ),
 142                         // http://www.whatwg.org/html/text-level-semantics.html#the-sub-and-sup-elements
 143                         array(
 144                                 '<var>x<sub><var>i</var></sub></var>, <var>y<sub><var>i</var></sub></var>',
 145                                 '<var>x<sub><var>i</var></sub></var>, <var>y<sub><var>i</var></sub></var>',
 146                                 'Nested <var>.'
 147                         ),
 148                         // http://www.whatwg.org/html/text-level-semantics.html#the-dfn-element
 149                         array(
 150                                 '<dfn><abbr title="Garage Door Opener">GDO</abbr></dfn>',
 151                                 '<dfn><abbr title="Garage Door Opener">GDO</abbr></dfn>',
 152                                 '<abbr> inside <dfn>',
 153                         ),
 154                 );
 155         }
 156
 157         /**
 158          * @dataProvider dataRemoveHTMLtags
 159          * @covers Sanitizer::removeHTMLtags
 160          */
 161         public function testRemoveHTMLtags( $input, $output, $msg = null ) {
 162                 $GLOBALS['wgUseTidy'] = false;
 163                 $this->assertEquals( $output, Sanitizer::removeHTMLtags( $input ), $msg );
 164         }
 165
 166         /**
 167          * @dataProvider provideTagAttributesToDecode
 168          * @covers Sanitizer::decodeTagAttributes
 169          */
 170         public function testDecodeTagAttributes( $expected, $attributes, $message = '' ) {
 171                 $this->assertEquals( $expected,
 172                         Sanitizer::decodeTagAttributes( $attributes ),
 173                         $message
 174                 );
 175         }
 176
 177         public static function provideTagAttributesToDecode() {
 178                 return array(
 179                         array( array( 'foo' => 'bar' ), 'foo=bar', 'Unquoted attribute' ),
 180                         array( array( 'foo' => 'bar' ), '    foo   =   bar    ', 'Spaced attribute' ),
 181                         array( array( 'foo' => 'bar' ), 'foo="bar"', 'Double-quoted attribute' ),
 182                         array( array( 'foo' => 'bar' ), 'foo=\'bar\'', 'Single-quoted attribute' ),
 183                         array( array( 'foo' => 'bar', 'baz' => 'foo' ), 'foo=\'bar\'   baz="foo"', 'Several attributes' ),
 184                         array( array( 'foo' => 'bar', 'baz' => 'foo' ), 'foo=\'bar\'   baz="foo"', 'Several attributes' ),
 185                         array( array( 'foo' => 'bar', 'baz' => 'foo' ), 'foo=\'bar\'   baz="foo"', 'Several attributes' ),
 186                         array( array( ':foo' => 'bar' ), ':foo=\'bar\'', 'Leading :' ),
 187                         array( array( '_foo' => 'bar' ), '_foo=\'bar\'', 'Leading _' ),
 188                         array( array( 'foo' => 'bar' ), 'Foo=\'bar\'', 'Leading capital' ),
 189                         array( array( 'foo' => 'BAR' ), 'FOO=BAR', 'Attribute keys are normalized to lowercase' ),
 190
 191                         # Invalid beginning
 192                         array( array(), '-foo=bar', 'Leading - is forbidden' ),
 193                         array( array(), '.foo=bar', 'Leading . is forbidden' ),
 194                         array( array( 'foo-bar' => 'bar' ), 'foo-bar=bar', 'A - is allowed inside the attribute' ),
 195                         array( array( 'foo-' => 'bar' ), 'foo-=bar', 'A - is allowed inside the attribute' ),
 196                         array( array( 'foo.bar' => 'baz' ), 'foo.bar=baz', 'A . is allowed inside the attribute' ),
 197                         array( array( 'foo.' => 'baz' ), 'foo.=baz', 'A . is allowed as last character' ),
 198                         array( array( 'foo6' => 'baz' ), 'foo6=baz', 'Numbers are allowed' ),
 199
 200                         # This bit is more relaxed than XML rules, but some extensions use
 201                         # it, like ProofreadPage (see bug 27539)
 202                         array( array( '1foo' => 'baz' ), '1foo=baz', 'Leading numbers are allowed' ),
 203                         array( array(), 'foo$=baz', 'Symbols are not allowed' ),
 204                         array( array(), 'foo@=baz', 'Symbols are not allowed' ),
 205                         array( array(), 'foo~=baz', 'Symbols are not allowed' ),
 206                         array( array( 'foo' => '1[#^`*%w/(' ), 'foo=1[#^`*%w/(', 'All kind of characters are allowed as values' ),
 207                         array( array( 'foo' => '1[#^`*%\'w/(' ), 'foo="1[#^`*%\'w/("', 'Double quotes are allowed if quoted by single quotes' ),
 208                         array( array( 'foo' => '1[#^`*%"w/(' ), 'foo=\'1[#^`*%"w/(\'', 'Single quotes are allowed if quoted by double quotes' ),
 209                         array( array( 'foo' => '&"' ), 'foo=&amp;&quot;', 'Special chars can be provided as entities' ),
 210                         array( array( 'foo' => '&foobar;' ), 'foo=&foobar;', 'Entity-like items are accepted' ),
 211                 );
 212         }
 213
 214         /**
 215          * @dataProvider provideDeprecatedAttributes
 216          * @covers Sanitizer::fixTagAttributes
 217          */
 218         public function testDeprecatedAttributesUnaltered( $inputAttr, $inputEl, $message = '' ) {
 219                 $this->assertEquals( " $inputAttr",
 220                         Sanitizer::fixTagAttributes( $inputAttr, $inputEl ),
 221                         $message
 222                 );
 223         }
 224
 225         public static function provideDeprecatedAttributes() {
 226                 /** array( <attribute>, <element>, [message] ) */
 227                 return array(
 228                         array( 'clear="left"', 'br' ),
 229                         array( 'clear="all"', 'br' ),
 230                         array( 'width="100"', 'td' ),
 231                         array( 'nowrap="true"', 'td' ),
 232                         array( 'nowrap=""', 'td' ),
 233                         array( 'align="right"', 'td' ),
 234                         array( 'align="center"', 'table' ),
 235                         array( 'align="left"', 'tr' ),
 236                         array( 'align="center"', 'div' ),
 237                         array( 'align="left"', 'h1' ),
 238                         array( 'align="left"', 'p' ),
 239                 );
 240         }
 241
 242         /**
 243          * @dataProvider provideCssCommentsFixtures
 244          * @covers Sanitizer::checkCss
 245          */
 246         public function testCssCommentsChecking( $expected, $css, $message = '' ) {
 247                 $this->assertEquals( $expected,
 248                         Sanitizer::checkCss( $css ),
 249                         $message
 250                 );
 251         }
 252
 253         public static function provideCssCommentsFixtures() {
 254                 /** array( <expected>, <css>, [message] ) */
 255                 return array(
 256                         // Valid comments spanning entire input
 257                         array( '/**/', '/**/' ),
 258                         array( '/* comment */', '/* comment */' ),
 259                         // Weird stuff
 260                         array( ' ', '/****/' ),
 261                         array( ' ', '/* /* */' ),
 262                         array( 'display: block;', "display:/* foo */block;" ),
 263                         array( 'display: block;', "display:\\2f\\2a foo \\2a\\2f block;",
 264                                 'Backslash-escaped comments must be stripped (bug 28450)' ),
 265                         array( '', '/* unfinished comment structure',
 266                                 'Remove anything after a comment-start token' ),
 267                         array( '', "\\2f\\2a unifinished comment'",
 268                                 'Remove anything after a backslash-escaped comment-start token' ),
 269                         array( '/* insecure input */', 'filter: progid:DXImageTransform.Microsoft.AlphaImageLoader(src=\'asdf.png\',sizingMethod=\'scale\');' ),
 270                         array( '/* insecure input */', '-ms-filter: "progid:DXImageTransform.Microsoft.AlphaImageLoader(src=\'asdf.png\',sizingMethod=\'scale\')";' ),
 271                         array( '/* insecure input */', 'width: expression(1+1);' ),
 272                         array( '/* insecure input */', 'background-image: image(asdf.png);' ),
 273                         array( '/* insecure input */', 'background-image: -webkit-image(asdf.png);' ),
 274                         array( '/* insecure input */', 'background-image: -moz-image(asdf.png);' ),
 275                         array( '/* insecure input */', 'background-image: image-set("asdf.png" 1x, "asdf.png" 2x);' ),
 276                         array( '/* insecure input */', 'background-image: -webkit-image-set("asdf.png" 1x, "asdf.png" 2x);' ),
 277                         array( '/* insecure input */', 'background-image: -moz-image-set("asdf.png" 1x, "asdf.png" 2x);' ),
 278                 );
 279         }
 280
 281         /**
 282          * Test for support or lack of support for specific attributes in the attribute whitelist.
 283          */
 284         public static function provideAttributeSupport() {
 285                 /** array( <attributes>, <expected>, <message> ) */
 286                 return array(
 287                         array( 'div', ' role="presentation"', ' role="presentation"', 'Support for WAI-ARIA\'s role="presentation".' ),
 288                         array( 'div', ' role="main"', '', "Other WAI-ARIA roles are currently not supported." ),
 289                 );
 290         }
 291
 292         /**
 293          * @dataProvider provideAttributeSupport
 294          * @covers Sanitizer::fixTagAttributes
 295          */
 296         public function testAttributeSupport( $tag, $attributes, $expected, $message ) {
 297                 $this->assertEquals( $expected,
 298                         Sanitizer::fixTagAttributes( $attributes, $tag ),
 299                         $message
 300                 );
 301         }
 302 }