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                         'wgUseTidy' => false
  75                 ) );
  76
  77                 if ( $escaped ) {
  78                         $this->assertEquals( "&lt;$tag&gt;",
  79                                 Sanitizer::removeHTMLtags( "<$tag>" )
  80                         );
  81                 } else {
  82                         $this->assertEquals( "<$tag></$tag>\n",
  83                                 Sanitizer::removeHTMLtags( "<$tag>" )
  84                         );
  85                 }
  86         }
  87
  88         /**
  89          * Provide HTML5 tags
  90          */
  91         public static function provideHtml5Tags() {
  92                 $ESCAPED = true; # We want tag to be escaped
  93                 $VERBATIM = false; # We want to keep the tag
  94                 return array(
  95                         array( 'data', $VERBATIM ),
  96                         array( 'mark', $VERBATIM ),
  97                         array( 'time', $VERBATIM ),
  98                         array( 'video', $ESCAPED ),
  99                 );
 100         }
 101
 102         function dataRemoveHTMLtags() {
 103                 return array(
 104                         // former testSelfClosingTag
 105                         array(
 106                                 '<div>Hello world</div />',
 107                                 '<div>Hello world</div>',
 108                                 'Self-closing closing div'
 109                         ),
 110                         // Make sure special nested HTML5 semantics are not broken
 111                         // http://www.whatwg.org/html/text-level-semantics.html#the-kbd-element
 112                         array(
 113                                 '<kbd><kbd>Shift</kbd>+<kbd>F3</kbd></kbd>',
 114                                 '<kbd><kbd>Shift</kbd>+<kbd>F3</kbd></kbd>',
 115                                 'Nested <kbd>.'
 116                         ),
 117                         // http://www.whatwg.org/html/text-level-semantics.html#the-sub-and-sup-elements
 118                         array(
 119                                 '<var>x<sub><var>i</var></sub></var>, <var>y<sub><var>i</var></sub></var>',
 120                                 '<var>x<sub><var>i</var></sub></var>, <var>y<sub><var>i</var></sub></var>',
 121                                 'Nested <var>.'
 122                         ),
 123                         // http://www.whatwg.org/html/text-level-semantics.html#the-dfn-element
 124                         array(
 125                                 '<dfn><abbr title="Garage Door Opener">GDO</abbr></dfn>',
 126                                 '<dfn><abbr title="Garage Door Opener">GDO</abbr></dfn>',
 127                                 '<abbr> inside <dfn>',
 128                         ),
 129                 );
 130         }
 131
 132         /**
 133          * @dataProvider dataRemoveHTMLtags
 134          */
 135         function testRemoveHTMLtags( $input, $output, $msg = null ) {
 136                 $GLOBALS['wgUseTidy'] = false;
 137                 $this->assertEquals( $output, Sanitizer::removeHTMLtags( $input ), $msg );
 138         }
 139
 140         /**
 141          * @dataProvider provideTagAttributesToDecode
 142          * @covers Sanitizer::decodeTagAttributes
 143          */
 144         function testDecodeTagAttributes( $expected, $attributes, $message = '' ) {
 145                 $this->assertEquals( $expected,
 146                         Sanitizer::decodeTagAttributes( $attributes ),
 147                         $message
 148                 );
 149         }
 150
 151         public static function provideTagAttributesToDecode() {
 152                 return array(
 153                         array( array( 'foo' => 'bar' ), 'foo=bar', 'Unquoted attribute' ),
 154                         array( array( 'foo' => 'bar' ), '    foo   =   bar    ', 'Spaced attribute' ),
 155                         array( array( 'foo' => 'bar' ), 'foo="bar"', 'Double-quoted attribute' ),
 156                         array( array( 'foo' => 'bar' ), 'foo=\'bar\'', 'Single-quoted attribute' ),
 157                         array( array( 'foo' => 'bar', 'baz' => 'foo' ), 'foo=\'bar\'   baz="foo"', 'Several attributes' ),
 158                         array( array( 'foo' => 'bar', 'baz' => 'foo' ), 'foo=\'bar\'   baz="foo"', 'Several attributes' ),
 159                         array( array( 'foo' => 'bar', 'baz' => 'foo' ), 'foo=\'bar\'   baz="foo"', 'Several attributes' ),
 160                         array( array( ':foo' => 'bar' ), ':foo=\'bar\'', 'Leading :' ),
 161                         array( array( '_foo' => 'bar' ), '_foo=\'bar\'', 'Leading _' ),
 162                         array( array( 'foo' => 'bar' ), 'Foo=\'bar\'', 'Leading capital' ),
 163                         array( array( 'foo' => 'BAR' ), 'FOO=BAR', 'Attribute keys are normalized to lowercase' ),
 164
 165                         # Invalid beginning
 166                         array( array(), '-foo=bar', 'Leading - is forbidden' ),
 167                         array( array(), '.foo=bar', 'Leading . is forbidden' ),
 168                         array( array( 'foo-bar' => 'bar' ), 'foo-bar=bar', 'A - is allowed inside the attribute' ),
 169                         array( array( 'foo-' => 'bar' ), 'foo-=bar', 'A - is allowed inside the attribute' ),
 170                         array( array( 'foo.bar' => 'baz' ), 'foo.bar=baz', 'A . is allowed inside the attribute' ),
 171                         array( array( 'foo.' => 'baz' ), 'foo.=baz', 'A . is allowed as last character' ),
 172                         array( array( 'foo6' => 'baz' ), 'foo6=baz', 'Numbers are allowed' ),
 173
 174                         # This bit is more relaxed than XML rules, but some extensions use
 175                         # it, like ProofreadPage (see bug 27539)
 176                         array( array( '1foo' => 'baz' ), '1foo=baz', 'Leading numbers are allowed' ),
 177                         array( array(), 'foo$=baz', 'Symbols are not allowed' ),
 178                         array( array(), 'foo@=baz', 'Symbols are not allowed' ),
 179                         array( array(), 'foo~=baz', 'Symbols are not allowed' ),
 180                         array( array( 'foo' => '1[#^`*%w/(' ), 'foo=1[#^`*%w/(', 'All kind of characters are allowed as values' ),
 181                         array( array( 'foo' => '1[#^`*%\'w/(' ), 'foo="1[#^`*%\'w/("', 'Double quotes are allowed if quoted by single quotes' ),
 182                         array( array( 'foo' => '1[#^`*%"w/(' ), 'foo=\'1[#^`*%"w/(\'', 'Single quotes are allowed if quoted by double quotes' ),
 183                         array( array( 'foo' => '&"' ), 'foo=&amp;&quot;', 'Special chars can be provided as entities' ),
 184                         array( array( 'foo' => '&foobar;' ), 'foo=&foobar;', 'Entity-like items are accepted' ),
 185                 );
 186         }
 187
 188         /**
 189          * @dataProvider provideDeprecatedAttributes
 190          * @covers Sanitizer::fixTagAttributes
 191          */
 192         function testDeprecatedAttributesUnaltered( $inputAttr, $inputEl, $message = '' ) {
 193                 $this->assertEquals( " $inputAttr",
 194                         Sanitizer::fixTagAttributes( $inputAttr, $inputEl ),
 195                         $message
 196                 );
 197         }
 198
 199         public static function provideDeprecatedAttributes() {
 200                 /** array( <attribute>, <element>, [message] ) */
 201                 return array(
 202                         array( 'clear="left"', 'br' ),
 203                         array( 'clear="all"', 'br' ),
 204                         array( 'width="100"', 'td' ),
 205                         array( 'nowrap="true"', 'td' ),
 206                         array( 'nowrap=""', 'td' ),
 207                         array( 'align="right"', 'td' ),
 208                         array( 'align="center"', 'table' ),
 209                         array( 'align="left"', 'tr' ),
 210                         array( 'align="center"', 'div' ),
 211                         array( 'align="left"', 'h1' ),
 212                         array( 'align="left"', 'span' ),
 213                 );
 214         }
 215
 216         /**
 217          * @dataProvider provideCssCommentsFixtures
 218          * @covers Sanitizer::checkCss
 219          */
 220         function testCssCommentsChecking( $expected, $css, $message = '' ) {
 221                 $this->assertEquals( $expected,
 222                         Sanitizer::checkCss( $css ),
 223                         $message
 224                 );
 225         }
 226
 227         public static function provideCssCommentsFixtures() {
 228                 /** array( <expected>, <css>, [message] ) */
 229                 return array(
 230                         array( ' ', '/**/' ),
 231                         array( ' ', '/****/' ),
 232                         array( ' ', '/* comment */' ),
 233                         array( ' ', "\\2f\\2a foo \\2a\\2f",
 234                                 'Backslash-escaped comments must be stripped (bug 28450)' ),
 235                         array( '', '/* unfinished comment structure',
 236                                 'Remove anything after a comment-start token' ),
 237                         array( '', "\\2f\\2a unifinished comment'",
 238                                 'Remove anything after a backslash-escaped comment-start token' ),
 239                         array( '/* insecure input */', 'filter: progid:DXImageTransform.Microsoft.AlphaImageLoader(src=\'asdf.png\',sizingMethod=\'scale\');' ),
 240                         array( '/* insecure input */', '-ms-filter: "progid:DXImageTransform.Microsoft.AlphaImageLoader(src=\'asdf.png\',sizingMethod=\'scale\')";' ),
 241                         array( '/* insecure input */', 'width: expression(1+1);' ),
 242                         array( '/* insecure input */', 'background-image: image(asdf.png);' ),
 243                         array( '/* insecure input */', 'background-image: -webkit-image(asdf.png);' ),
 244                         array( '/* insecure input */', 'background-image: -moz-image(asdf.png);' ),
 245                         array( '/* insecure input */', 'background-image: image-set("asdf.png" 1x, "asdf.png" 2x);' ),
 246                         array( '/* insecure input */', 'background-image: -webkit-image-set("asdf.png" 1x, "asdf.png" 2x);' ),
 247                         array( '/* insecure input */', 'background-image: -moz-image-set("asdf.png" 1x, "asdf.png" 2x);' ),
 248                 );
 249         }
 250
 251         /**
 252          * Test for support or lack of support for specific attributes in the attribute whitelist.
 253          */
 254         public static function provideAttributeSupport() {
 255                 /** array( <attributes>, <expected>, <message> ) */
 256                 return array(
 257                         array( 'div', ' role="presentation"', ' role="presentation"', 'Support for WAI-ARIA\'s role="presentation".' ),
 258                         array( 'div', ' role="main"', '', "Other WAI-ARIA roles are currently not supported." ),
 259                 );
 260         }
 261
 262         /**
 263          * @dataProvider provideAttributeSupport
 264          */
 265         function testAttributeSupport( $tag, $attributes, $expected, $message ) {
 266                 $this->assertEquals( $expected,
 267                         Sanitizer::fixTagAttributes( $attributes, $tag ),
 268                         $message
 269                 );
 270         }
 271 }