SanitizerTest.php

Go to the documentation of this file.

<?php
 
class SanitizerTest extends MediaWikiTestCase {
 
    protected function tearDown() {
        MWTidy::destroySingleton();
        parent::tearDown();
    }
 
    public function testDecodeNamedEntities() {
        $this->assertEquals(
            "\xc3\xa9cole",
            Sanitizer::decodeCharReferences( '&eacute;cole' ),
            'decode named entities'
        );
    }
 
    public function testDecodeNumericEntities() {
        $this->assertEquals(
            "\xc4\x88io bonas dans l'\xc3\xa9cole!",
            Sanitizer::decodeCharReferences( "&#x108;io bonas dans l'&#233;cole!" ),
            'decode numeric entities'
        );
    }
 
    public function testDecodeMixedEntities() {
        $this->assertEquals(
            "\xc4\x88io bonas dans l'\xc3\xa9cole!",
            Sanitizer::decodeCharReferences( "&#x108;io bonas dans l'&eacute;cole!" ),
            'decode mixed numeric/named entities'
        );
    }
 
    public function testDecodeMixedComplexEntities() {
        $this->assertEquals(
            "\xc4\x88io bonas dans l'\xc3\xa9cole! (mais pas &#x108;io dans l'&eacute;cole)",
            Sanitizer::decodeCharReferences(
                "&#x108;io bonas dans l'&eacute;cole! (mais pas &amp;#x108;io dans l'&#38;eacute;cole)"
            ),
            'decode mixed complex entities'
        );
    }
 
    public function testInvalidAmpersand() {
        $this->assertEquals(
            'a & b',
            Sanitizer::decodeCharReferences( 'a & b' ),
            'Invalid ampersand'
        );
    }
 
    public function testInvalidEntities() {
        $this->assertEquals(
            '&foo;',
            Sanitizer::decodeCharReferences( '&foo;' ),
            'Invalid named entity'
        );
    }
 
    public function testInvalidNumberedEntities() {
        $this->assertEquals(
            UtfNormal\Constants::UTF8_REPLACEMENT,
            Sanitizer::decodeCharReferences( "&#88888888888888;" ),
            'Invalid numbered entity'
        );
    }
 
    public function testRemovehtmltagsOnHtml5Tags( $tag, $escaped ) {
        MWTidy::setInstance( false );
 
        if ( $escaped ) {
            $this->assertEquals( "&lt;$tag&gt;",
                Sanitizer::removeHTMLtags( "<$tag>" )
            );
        } else {
            $this->assertEquals( "<$tag></$tag>\n",
                Sanitizer::removeHTMLtags( "<$tag>" )
            );
        }
    }
 
    public static function provideHtml5Tags() {
        $ESCAPED = true; # We want tag to be escaped
        $VERBATIM = false; # We want to keep the tag
        return [
            [ 'data', $VERBATIM ],
            [ 'mark', $VERBATIM ],
            [ 'time', $VERBATIM ],
            [ 'video', $ESCAPED ],
        ];
    }
 
    function dataRemoveHTMLtags() {
        return [
            // former testSelfClosingTag
            [
                '<div>Hello world</div />',
                '<div>Hello world</div>',
                'Self-closing closing div'
            ],
            // Make sure special nested HTML5 semantics are not broken
            // https://html.spec.whatwg.org/multipage/semantics.html#the-kbd-element
            [
                '<kbd><kbd>Shift</kbd>+<kbd>F3</kbd></kbd>',
                '<kbd><kbd>Shift</kbd>+<kbd>F3</kbd></kbd>',
                'Nested <kbd>.'
            ],
            // https://html.spec.whatwg.org/multipage/semantics.html#the-sub-and-sup-elements
            [
                '<var>x<sub><var>i</var></sub></var>, <var>y<sub><var>i</var></sub></var>',
                '<var>x<sub><var>i</var></sub></var>, <var>y<sub><var>i</var></sub></var>',
                'Nested <var>.'
            ],
            // https://html.spec.whatwg.org/multipage/semantics.html#the-dfn-element
            [
                '<dfn><abbr title="Garage Door Opener">GDO</abbr></dfn>',
                '<dfn><abbr title="Garage Door Opener">GDO</abbr></dfn>',
                '<abbr> inside <dfn>',
            ],
        ];
    }
 
    public function testRemoveHTMLtags( $input, $output, $msg = null ) {
        MWTidy::setInstance( false );
        $this->assertEquals( $output, Sanitizer::removeHTMLtags( $input ), $msg );
    }
 
    public function testDecodeTagAttributes( $expected, $attributes, $message = '' ) {
        $this->assertEquals( $expected,
            Sanitizer::decodeTagAttributes( $attributes ),
            $message
        );
    }
 
    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' ],
            [
                [ 'foo' => 'bar', 'baz' => 'foo' ],
                'foo=\'bar\'   baz="foo"',
                'Several attributes'
            ],
            [
                [ 'foo' => 'bar', 'baz' => 'foo' ],
                'foo=\'bar\'   baz="foo"',
                'Several attributes'
            ],
            [
                [ 'foo' => 'bar', 'baz' => 'foo' ],
                'foo=\'bar\'   baz="foo"',
                'Several attributes'
            ],
            [ [ ':foo' => 'bar' ], ':foo=\'bar\'', 'Leading :' ],
            [ [ '_foo' => 'bar' ], '_foo=\'bar\'', 'Leading _' ],
            [ [ 'foo' => 'bar' ], 'Foo=\'bar\'', 'Leading capital' ],
            [ [ 'foo' => 'BAR' ], 'FOO=BAR', 'Attribute keys are normalized to lowercase' ],
 
            # Invalid beginning
            [ [], '-foo=bar', 'Leading - is forbidden' ],
            [ [], '.foo=bar', 'Leading . is forbidden' ],
            [ [ 'foo-bar' => 'bar' ], 'foo-bar=bar', 'A - is allowed inside the attribute' ],
            [ [ 'foo-' => 'bar' ], 'foo-=bar', 'A - is allowed inside the attribute' ],
            [ [ 'foo.bar' => 'baz' ], 'foo.bar=baz', 'A . is allowed inside the attribute' ],
            [ [ 'foo.' => 'baz' ], 'foo.=baz', 'A . is allowed as last character' ],
            [ [ 'foo6' => 'baz' ], 'foo6=baz', 'Numbers are allowed' ],
 
            # This bit is more relaxed than XML rules, but some extensions use
            # 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' ],
            [ [], 'foo~=baz', 'Symbols are not allowed' ],
            [
                [ 'foo' => '1[#^`*%w/(' ],
                'foo=1[#^`*%w/(',
                'All kind of characters are allowed as values'
            ],
            [
                [ 'foo' => '1[#^`*%\'w/(' ],
                'foo="1[#^`*%\'w/("',
                'Double quotes are allowed if quoted by single quotes'
            ],
            [
                [ 'foo' => '1[#^`*%"w/(' ],
                'foo=\'1[#^`*%"w/(\'',
                'Single quotes are allowed if quoted by double quotes'
            ],
            [ [ 'foo' => '&"' ], 'foo=&amp;&quot;', 'Special chars can be provided as entities' ],
            [ [ 'foo' => '&foobar;' ], 'foo=&foobar;', 'Entity-like items are accepted' ],
        ];
    }
 
    public function testDeprecatedAttributesUnaltered( $inputAttr, $inputEl, $message = '' ) {
        $this->assertEquals( " $inputAttr",
            Sanitizer::fixTagAttributes( $inputAttr, $inputEl ),
            $message
        );
    }
 
    public static function provideDeprecatedAttributes() {
        return [
            [ 'clear="left"', 'br' ],
            [ 'clear="all"', 'br' ],
            [ 'width="100"', 'td' ],
            [ 'nowrap="true"', 'td' ],
            [ 'nowrap=""', 'td' ],
            [ 'align="right"', 'td' ],
            [ 'align="center"', 'table' ],
            [ 'align="left"', 'tr' ],
            [ 'align="center"', 'div' ],
            [ 'align="left"', 'h1' ],
            [ 'align="left"', 'p' ],
        ];
    }
 
    public function testCssCommentsChecking( $expected, $css, $message = '' ) {
        $this->assertEquals( $expected,
            Sanitizer::checkCss( $css ),
            $message
        );
    }
 
    public static function provideCssCommentsFixtures() {
        return [
            // Valid comments spanning entire input
            [ '/**/', '/**/' ],
            [ '/* comment */', '/* comment */' ],
            // Weird stuff
            [ ' ', '/****/' ],
            [ ' ', '/* /* */' ],
            [ 'display: block;', "display:/* foo */block;" ],
            [ 'display: block;', "display:\\2f\\2a foo \\2a\\2f block;",
                'Backslash-escaped comments must be stripped (T30450)' ],
            [ '', '/* unfinished comment structure',
                'Remove anything after a comment-start token' ],
            [ '', "\\2f\\2a unifinished comment'",
                'Remove anything after a backslash-escaped comment-start token' ],
            [
                '/* insecure input */',
                'filter: progid:DXImageTransform.Microsoft.AlphaImageLoader'
                    . '(src=\'asdf.png\',sizingMethod=\'scale\');'
            ],
            [
                '/* insecure input */',
                '-ms-filter: "progid:DXImageTransform.Microsoft.AlphaImageLoader'
                    . '(src=\'asdf.png\',sizingMethod=\'scale\')";'
            ],
            [ '/* insecure input */', 'width: expression(1+1);' ],
            [ '/* insecure input */', 'background-image: image(asdf.png);' ],
            [ '/* insecure input */', 'background-image: -webkit-image(asdf.png);' ],
            [ '/* insecure input */', 'background-image: -moz-image(asdf.png);' ],
            [ '/* insecure input */', 'background-image: image-set("asdf.png" 1x, "asdf.png" 2x);' ],
            [
                '/* insecure input */',
                'background-image: -webkit-image-set("asdf.png" 1x, "asdf.png" 2x);'
            ],
            [
                '/* insecure input */',
                'background-image: -moz-image-set("asdf.png" 1x, "asdf.png" 2x);'
            ],
            [ '/* insecure input */', 'foo: attr( title, url );' ],
            [ '/* insecure input */', 'foo: attr( title url );' ],
            [ '/* insecure input */', 'foo: var(--evil-attribute)' ],
        ];
    }
 
    public function testEscapeHtmlAllowEntities( $expected, $html ) {
        $this->assertEquals(
            $expected,
            Sanitizer::escapeHtmlAllowEntities( $html )
        );
    }
 
    public static function provideEscapeHtmlAllowEntities() {
        return [
            [ 'foo', 'foo' ],
            [ 'a¡b', 'a&#161;b' ],
            [ 'foo&#039;bar', "foo'bar" ],
            [ '&lt;script&gt;foo&lt;/script&gt;', '<script>foo</script>' ],
        ];
    }
 
    public function testEscapeId( $input, $output ) {
        $this->assertEquals(
            $output,
            Sanitizer::escapeId( $input, [ 'noninitial', 'legacy' ] )
        );
    }
 
    public static function provideEscapeId() {
        return [
            [ '+', '.2B' ],
            [ '&', '.26' ],
            [ '=', '.3D' ],
            [ ':', ':' ],
            [ ';', '.3B' ],
            [ '@', '.40' ],
            [ '$', '.24' ],
            [ '-_.', '-_.' ],
            [ '!', '.21' ],
            [ '*', '.2A' ],
            [ '/', '.2F' ],
            [ '[]', '.5B.5D' ],
            [ '<>', '.3C.3E' ],
            [ '\'', '.27' ],
            [ '§', '.C2.A7' ],
            [ 'Test:A & B/Here', 'Test:A_.26_B.2FHere' ],
            [ 'A&B&amp;C&amp;amp;D&amp;amp;amp;E', 'A.26B.26amp.3BC.26amp.3Bamp.3BD.26amp.3Bamp.3Bamp.3BE' ],
        ];
    }
 
    public function testEscapeIdReferenceList( $referenceList, $id1, $id2 ) {
        $this->assertEquals(
            Sanitizer::escapeIdReferenceList( $referenceList, 'noninitial' ),
            Sanitizer::escapeIdForAttribute( $id1 )
                . ' '
                . Sanitizer::escapeIdForAttribute( $id2 )
        );
    }
 
    public static function provideEscapeIdReferenceList() {
        return [
            [ 'foo bar', 'foo', 'bar' ],
            [ '#1 #2', '#1', '#2' ],
            [ '+1 +2', '+1', '+2' ],
        ];
    }
 
    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
        ];
    }
 
    public function testEscapeIdForStuff( $stuff, array $config, $id, $expected, $mode = null ) {
        $func = "Sanitizer::escapeIdFor{$stuff}";
        $iwFlavor = array_pop( $config );
        $this->setMwGlobals( [
            'wgFragmentMode' => $config,
            'wgExternalInterwikiFragmentMode' => $iwFlavor,
        ] );
        $escaped = call_user_func( $func, $id, $mode );
        self::assertEquals( $expected, $escaped );
    }
 
    public function provideEscapeIdForStuff() {
        // Test inputs and outputs
        $text = 'foo тест_#%!\'()[]:<>&&amp;&amp;amp;';
        $legacyEncoded = 'foo_.D1.82.D0.B5.D1.81.D1.82_.23.25.21.27.28.29.5B.5D:.3C.3E' .
            '.26.26amp.3B.26amp.3Bamp.3B';
        $html5Encoded = 'foo_тест_#%!\'()[]:<>&&amp;&amp;amp;';
        $html5Experimental = 'foo_тест_!_()[]:<>_amp;_amp;amp;';
 
        // Settings: last element is $wgExternalInterwikiFragmentMode, the rest is $wgFragmentMode
        $legacy = [ 'legacy', 'legacy' ];
        $legacyNew = [ 'legacy', 'html5', 'legacy' ];
        $newLegacy = [ 'html5', 'legacy', 'legacy' ];
        $new = [ 'html5', 'legacy' ];
        $allNew = [ 'html5', 'html5' ];
        $experimentalLegacy = [ 'html5-legacy', 'legacy', 'legacy' ];
        $newExperimental = [ 'html5', 'html5-legacy', 'legacy' ];
 
        return [
            // Pure legacy: how MW worked before 2017
            [ 'Attribute', $legacy, $text, $legacyEncoded, Sanitizer::ID_PRIMARY ],
            [ 'Attribute', $legacy, $text, false, Sanitizer::ID_FALLBACK ],
            [ 'Link', $legacy, $text, $legacyEncoded ],
            [ 'ExternalInterwiki', $legacy, $text, $legacyEncoded ],
 
            // Transition to a new world: legacy links with HTML5 fallback
            [ 'Attribute', $legacyNew, $text, $legacyEncoded, Sanitizer::ID_PRIMARY ],
            [ 'Attribute', $legacyNew, $text, $html5Encoded, Sanitizer::ID_FALLBACK ],
            [ 'Link', $legacyNew, $text, $legacyEncoded ],
            [ 'ExternalInterwiki', $legacyNew, $text, $legacyEncoded ],
 
            // New world: HTML5 links, legacy fallbacks
            [ 'Attribute', $newLegacy, $text, $html5Encoded, Sanitizer::ID_PRIMARY ],
            [ 'Attribute', $newLegacy, $text, $legacyEncoded, Sanitizer::ID_FALLBACK ],
            [ 'Link', $newLegacy, $text, $html5Encoded ],
            [ 'ExternalInterwiki', $newLegacy, $text, $legacyEncoded ],
 
            // Distant future: no legacy fallbacks, but still linking to leagacy wikis
            [ 'Attribute', $new, $text, $html5Encoded, Sanitizer::ID_PRIMARY ],
            [ 'Attribute', $new, $text, false, Sanitizer::ID_FALLBACK ],
            [ 'Link', $new, $text, $html5Encoded ],
            [ 'ExternalInterwiki', $new, $text, $legacyEncoded ],
 
            // Just before the heat death of universe: external interwikis are also HTML5 \m/
            [ 'Attribute', $allNew, $text, $html5Encoded, Sanitizer::ID_PRIMARY ],
            [ 'Attribute', $allNew, $text, false, Sanitizer::ID_FALLBACK ],
            [ 'Link', $allNew, $text, $html5Encoded ],
            [ 'ExternalInterwiki', $allNew, $text, $html5Encoded ],
 
            // Someone flipped $wgExperimentalHtmlIds on
            [ 'Attribute', $experimentalLegacy, $text, $html5Experimental, Sanitizer::ID_PRIMARY ],
            [ 'Attribute', $experimentalLegacy, $text, $legacyEncoded, Sanitizer::ID_FALLBACK ],
            [ 'Link', $experimentalLegacy, $text, $html5Experimental ],
            [ 'ExternalInterwiki', $experimentalLegacy, $text, $legacyEncoded ],
 
            // Migration from $wgExperimentalHtmlIds to modern HTML5
            [ 'Attribute', $newExperimental, $text, $html5Encoded, Sanitizer::ID_PRIMARY ],
            [ 'Attribute', $newExperimental, $text, $html5Experimental, Sanitizer::ID_FALLBACK ],
            [ 'Link', $newExperimental, $text, $html5Encoded ],
            [ 'ExternalInterwiki', $newExperimental, $text, $legacyEncoded ],
        ];
    }
 
    public function testInvalidFragmentThrows() {
        $this->setMwGlobals( 'wgFragmentMode', [ 'boom!' ] );
        Sanitizer::escapeIdForAttribute( 'This should throw' );
    }
 
    public function testNoPrimaryFragmentModeThrows() {
        $this->setMwGlobals( 'wgFragmentMode', [ 666 => 'html5' ] );
        Sanitizer::escapeIdForAttribute( 'This should throw' );
    }
 
    public function testNoPrimaryFragmentModeThrows2() {
        $this->setMwGlobals( 'wgFragmentMode', [ 666 => 'html5' ] );
        Sanitizer::escapeIdForLink( 'This should throw' );
    }
}