tests/phpunit/includes/SanitizerTest.php

   1 <?php
   2
   3 class SanitizerTest extends MediaWikiTestCase {
   4
   5         protected function setUp() {
   6                 parent::setUp();
   7
   8                 AutoLoader::loadClass( 'Sanitizer' );
   9         }
  10
  11         function testDecodeNamedEntities() {
  12                 $this->assertEquals(
  13                         "\xc3\xa9cole",
  14                         Sanitizer::decodeCharReferences( '&eacute;cole' ),
  15                         'decode named entities'
  16                 );
  17         }
  18
  19         function testDecodeNumericEntities() {
  20                 $this->assertEquals(
  21                         "\xc4\x88io bonas dans l'\xc3\xa9cole!",
  22                         Sanitizer::decodeCharReferences( "&#x108;io bonas dans l'&#233;cole!" ),
  23                         'decode numeric entities'
  24                 );
  25         }
  26
  27         function testDecodeMixedEntities() {
  28                 $this->assertEquals(
  29                         "\xc4\x88io bonas dans l'\xc3\xa9cole!",
  30                         Sanitizer::decodeCharReferences( "&#x108;io bonas dans l'&eacute;cole!" ),
  31                         'decode mixed numeric/named entities'
  32                 );
  33         }
  34
  35         function testDecodeMixedComplexEntities() {
  36                 $this->assertEquals(
  37                         "\xc4\x88io bonas dans l'\xc3\xa9cole! (mais pas &#x108;io dans l'&eacute;cole)",
  38                         Sanitizer::decodeCharReferences(
  39                                 "&#x108;io bonas dans l'&eacute;cole! (mais pas &amp;#x108;io dans l'&#38;eacute;cole)"
  40                         ),
  41                         'decode mixed complex entities'
  42                 );
  43         }
  44
  45         function testInvalidAmpersand() {
  46                 $this->assertEquals(
  47                         'a & b',
  48                         Sanitizer::decodeCharReferences( 'a & b' ),
  49                         'Invalid ampersand'
  50                 );
  51         }
  52
  53         function testInvalidEntities() {
  54                 $this->assertEquals(
  55                         '&foo;',
  56                         Sanitizer::decodeCharReferences( '&foo;' ),
  57                         'Invalid named entity'
  58                 );
  59         }
  60
  61         function testInvalidNumberedEntities() {
  62                 $this->assertEquals( UTF8_REPLACEMENT, Sanitizer::decodeCharReferences( "&#88888888888888;" ), 'Invalid numbered entity' );
  63         }
  64
  65         /**
  66          * @covers Sanitizer::removeHTMLtags
  67          * @dataProvider provideHtml5Tags
  68          *
  69          * @param String $tag Name of an HTML5 element (ie: 'video')
  70          * @param Boolean $escaped Wheter sanitizer let the tag in or escape it (ie: '&lt;video&gt;')
  71          */
  72         function testRemovehtmltagsOnHtml5Tags( $tag, $escaped ) {
  73                 $this->setMwGlobals( array(
  74                         # Enable HTML5 mode
  75                         'wgHtml5' => true,
  76                         'wgUseTidy' => false
  77                 ) );
  78
  79                 if ( $escaped ) {
  80                         $this->assertEquals( "&lt;$tag&gt;",
  81                                 Sanitizer::removeHTMLtags( "<$tag>" )
  82                         );
  83                 } else {
  84                         $this->assertEquals( "<$tag></$tag>\n",
  85                                 Sanitizer::removeHTMLtags( "<$tag>" )
  86                         );
  87                 }
  88         }
  89
  90         /**
  91          * Provide HTML5 tags
  92          */
  93         public static function provideHtml5Tags() {
  94                 $ESCAPED = true; # We want tag to be escaped
  95                 $VERBATIM = false; # We want to keep the tag
  96                 return array(
  97                         array( 'data', $VERBATIM ),
  98                         array( 'mark', $VERBATIM ),
  99                         array( 'time', $VERBATIM ),
 100                         array( 'video', $ESCAPED ),
 101                 );
 102         }
 103
 104         function dataRemoveHTMLtags() {
 105                 return array(
 106                         // former testSelfClosingTag
 107                         array(
 108                                 '<div>Hello world</div />',
 109                                 '<div>Hello world</div>',
 110                                 'Self-closing closing div'
 111                         ),
 112                         // Make sure special nested HTML5 semantics are not broken
 113                         // http://www.whatwg.org/html/text-level-semantics.html#the-kbd-element
 114                         array(
 115                                 '<kbd><kbd>Shift</kbd>+<kbd>F3</kbd></kbd>',
 116                                 '<kbd><kbd>Shift</kbd>+<kbd>F3</kbd></kbd>',
 117                                 'Nested <kbd>.'
 118                         ),
 119                         // http://www.whatwg.org/html/text-level-semantics.html#the-sub-and-sup-elements
 120                         array(
 121                                 '<var>x<sub><var>i</var></sub></var>, <var>y<sub><var>i</var></sub></var>',
 122                                 '<var>x<sub><var>i</var></sub></var>, <var>y<sub><var>i</var></sub></var>',
 123                                 'Nested <var>.'
 124                         ),
 125                         // http://www.whatwg.org/html/text-level-semantics.html#the-dfn-element
 126                         array(
 127                                 '<dfn><abbr title="Garage Door Opener">GDO</abbr></dfn>',
 128                                 '<dfn><abbr title="Garage Door Opener">GDO</abbr></dfn>',
 129                                 '<abbr> inside <dfn>',
 130                         ),
 131                 );
 132         }
 133
 134         /**
 135          * @dataProvider dataRemoveHTMLtags
 136          */
 137         function testRemoveHTMLtags( $input, $output, $msg = null ) {
 138                 $GLOBALS['wgUseTidy'] = false;
 139                 $this->assertEquals( $output, Sanitizer::removeHTMLtags( $input ), $msg );
 140         }
 141
 142         /**
 143          * @dataProvider provideTagAttributesToDecode
 144          * @covers Sanitizer::decodeTagAttributes
 145          */
 146         function testDecodeTagAttributes( $expected, $attributes, $message = '' ) {
 147                 $this->assertEquals( $expected,
 148                         Sanitizer::decodeTagAttributes( $attributes ),
 149                         $message
 150                 );
 151         }
 152
 153         public static function provideTagAttributesToDecode() {
 154                 return array(
 155                         array( array( 'foo' => 'bar' ), 'foo=bar', 'Unquoted attribute' ),
 156                         array( array( 'foo' => 'bar' ), '    foo   =   bar    ', 'Spaced attribute' ),
 157                         array( array( 'foo' => 'bar' ), 'foo="bar"', 'Double-quoted attribute' ),
 158                         array( array( 'foo' => 'bar' ), 'foo=\'bar\'', 'Single-quoted attribute' ),
 159                         array( array( 'foo' => 'bar', 'baz' => 'foo' ), 'foo=\'bar\'   baz="foo"', 'Several attributes' ),
 160                         array( array( 'foo' => 'bar', 'baz' => 'foo' ), 'foo=\'bar\'   baz="foo"', 'Several attributes' ),
 161                         array( array( 'foo' => 'bar', 'baz' => 'foo' ), 'foo=\'bar\'   baz="foo"', 'Several attributes' ),
 162                         array( array( ':foo' => 'bar' ), ':foo=\'bar\'', 'Leading :' ),
 163                         array( array( '_foo' => 'bar' ), '_foo=\'bar\'', 'Leading _' ),
 164                         array( array( 'foo' => 'bar' ), 'Foo=\'bar\'', 'Leading capital' ),
 165                         array( array( 'foo' => 'BAR' ), 'FOO=BAR', 'Attribute keys are normalized to lowercase' ),
 166
 167                         # Invalid beginning
 168                         array( array(), '-foo=bar', 'Leading - is forbidden' ),
 169                         array( array(), '.foo=bar', 'Leading . is forbidden' ),
 170                         array( array( 'foo-bar' => 'bar' ), 'foo-bar=bar', 'A - is allowed inside the attribute' ),
 171                         array( array( 'foo-' => 'bar' ), 'foo-=bar', 'A - is allowed inside the attribute' ),
 172                         array( array( 'foo.bar' => 'baz' ), 'foo.bar=baz', 'A . is allowed inside the attribute' ),
 173                         array( array( 'foo.' => 'baz' ), 'foo.=baz', 'A . is allowed as last character' ),
 174                         array( array( 'foo6' => 'baz' ), 'foo6=baz', 'Numbers are allowed' ),
 175
 176                         # This bit is more relaxed than XML rules, but some extensions use
 177                         # it, like ProofreadPage (see bug 27539)
 178                         array( array( '1foo' => 'baz' ), '1foo=baz', 'Leading numbers are allowed' ),
 179                         array( array(), 'foo$=baz', 'Symbols are not allowed' ),
 180                         array( array(), 'foo@=baz', 'Symbols are not allowed' ),
 181                         array( array(), 'foo~=baz', 'Symbols are not allowed' ),
 182                         array( array( 'foo' => '1[#^`*%w/(' ), 'foo=1[#^`*%w/(', 'All kind of characters are allowed as values' ),
 183                         array( array( 'foo' => '1[#^`*%\'w/(' ), 'foo="1[#^`*%\'w/("', 'Double quotes are allowed if quoted by single quotes' ),
 184                         array( array( 'foo' => '1[#^`*%"w/(' ), 'foo=\'1[#^`*%"w/(\'', 'Single quotes are allowed if quoted by double quotes' ),
 185                         array( array( 'foo' => '&"' ), 'foo=&amp;&quot;', 'Special chars can be provided as entities' ),
 186                         array( array( 'foo' => '&foobar;' ), 'foo=&foobar;', 'Entity-like items are accepted' ),
 187                 );
 188         }
 189
 190         /**
 191          * @dataProvider provideDeprecatedAttributes
 192          * @covers Sanitizer::fixTagAttributes
 193          */
 194         function testDeprecatedAttributesUnaltered( $inputAttr, $inputEl, $message = '' ) {
 195                 $this->assertEquals( " $inputAttr",
 196                         Sanitizer::fixTagAttributes( $inputAttr, $inputEl ),
 197                         $message
 198                 );
 199         }
 200
 201         public static function provideDeprecatedAttributes() {
 202                 /** array( <attribute>, <element>, [message] ) */
 203                 return array(
 204                         array( 'clear="left"', 'br' ),
 205                         array( 'clear="all"', 'br' ),
 206                         array( 'width="100"', 'td' ),
 207                         array( 'nowrap="true"', 'td' ),
 208                         array( 'nowrap=""', 'td' ),
 209                         array( 'align="right"', 'td' ),
 210                         array( 'align="center"', 'table' ),
 211                         array( 'align="left"', 'tr' ),
 212                         array( 'align="center"', 'div' ),
 213                         array( 'align="left"', 'h1' ),
 214                         array( 'align="left"', 'span' ),
 215                 );
 216         }
 217
 218         /**
 219          * @dataProvider provideCssCommentsFixtures
 220          * @covers Sanitizer::checkCss
 221          */
 222         function testCssCommentsChecking( $expected, $css, $message = '' ) {
 223                 $this->assertEquals( $expected,
 224                         Sanitizer::checkCss( $css ),
 225                         $message
 226                 );
 227         }
 228
 229         public static function provideCssCommentsFixtures() {
 230                 /** array( <expected>, <css>, [message] ) */
 231                 return array(
 232                         array( ' ', '/**/' ),
 233                         array( ' ', '/****/' ),
 234                         array( ' ', '/* comment */' ),
 235                         array( ' ', "\\2f\\2a foo \\2a\\2f",
 236                                 'Backslash-escaped comments must be stripped (bug 28450)' ),
 237                         array( '', '/* unfinished comment structure',
 238                                 'Remove anything after a comment-start token' ),
 239                         array( '', "\\2f\\2a unifinished comment'",
 240                                 'Remove anything after a backslash-escaped comment-start token' ),
 241                         array( '/* insecure input */', 'filter: progid:DXImageTransform.Microsoft.AlphaImageLoader(src=\'asdf.png\',sizingMethod=\'scale\');' ),
 242                         array( '/* insecure input */', '-ms-filter: "progid:DXImageTransform.Microsoft.AlphaImageLoader(src=\'asdf.png\',sizingMethod=\'scale\')";' ),
 243                         array( '/* insecure input */', 'width: expression(1+1);' ),
 244                         array( '/* insecure input */', 'background-image: image(asdf.png);' ),
 245                         array( '/* insecure input */', 'background-image: -webkit-image(asdf.png);' ),
 246                         array( '/* insecure input */', 'background-image: -moz-image(asdf.png);' ),
 247                         array( '/* insecure input */', 'background-image: image-set("asdf.png" 1x, "asdf.png" 2x);' ),
 248                         array( '/* insecure input */', 'background-image: -webkit-image-set("asdf.png" 1x, "asdf.png" 2x);' ),
 249                         array( '/* insecure input */', 'background-image: -moz-image-set("asdf.png" 1x, "asdf.png" 2x);' ),
 250                 );
 251         }
 252
 253         /**
 254          * Test for support or lack of support for specific attributes in the attribute whitelist.
 255          */
 256         public static function provideAttributeSupport() {
 257                 /** array( <attributes>, <expected>, <message> ) */
 258                 return array(
 259                         array( 'div', ' role="presentation"', ' role="presentation"', 'Support for WAI-ARIA\'s role="presentation".' ),
 260                         array( 'div', ' role="main"', '', "Other WAI-ARIA roles are currently not supported." ),
 261                 );
 262         }
 263
 264         /**
 265          * @dataProvider provideAttributeSupport
 266          */
 267         function testAttributeSupport( $tag, $attributes, $expected, $message ) {
 268                 $this->assertEquals( $expected,
 269                         Sanitizer::fixTagAttributes( $attributes, $tag ),
 270                         $message
 271                 );
 272         }
 273 }