diff options
| author | Asmir Mustafic <[email protected]> | 2014-12-17 21:31:54 +0100 |
|---|---|---|
| committer | Asmir Mustafic <[email protected]> | 2014-12-17 21:31:54 +0100 |
| commit | 398ebb68c9395a67858d230d0610aa3676bebdc7 (patch) | |
| tree | b0757c652a8991895c33c035d2012e841579f67d | |
| parent | de3a3af4aa683d0dae5bd09b62fe74127974c3ea (diff) | |
| parent | 4df8acdb61a2eb27512cd9d2b2e682399c9bc2e5 (diff) | |
Merge pull request #64 from goetas/i63
Case insensitive tags
| -rw-r--r-- | src/HTML5/Parser/Tokenizer.php | 10 | ||||
| -rw-r--r-- | test/HTML5/Parser/DOMTreeBuilderTest.php | 22 |
2 files changed, 28 insertions, 4 deletions
diff --git a/src/HTML5/Parser/Tokenizer.php b/src/HTML5/Parser/Tokenizer.php index f7b450f..958ade8 100644 --- a/src/HTML5/Parser/Tokenizer.php +++ b/src/HTML5/Parser/Tokenizer.php @@ -203,7 +203,9 @@ class Tokenizer $sequence = '</' . $this->untilTag . '>'; $txt = ''; $tok = $this->scanner->current(); - while ($tok !== false && ! ($tok == '<' && ($this->sequenceMatches($sequence) || $this->sequenceMatches(strtoupper($sequence))))) { + + $caseSensitive = !Elements::isHtml5Element($this->untilTag); + while ($tok !== false && ! ($tok == '<' && ($this->sequenceMatches($sequence, $caseSensitive)))) { if ($tok == '&') { $txt .= $this->decodeCharacterReference(); $tok = $this->scanner->current(); @@ -891,7 +893,7 @@ class Tokenizer $buffer .= $this->scanner->charsUntil($first); // Stop as soon as we hit the stopping condition. - if ($this->sequenceMatches($sequence) || $this->sequenceMatches(strtoupper($sequence))) { + if ($this->sequenceMatches($sequence, false)) { return $buffer; } $buffer .= $this->scanner->current(); @@ -916,7 +918,7 @@ class Tokenizer * see if the input stream is at the start of a * '</script>' string. */ - protected function sequenceMatches($sequence) + protected function sequenceMatches($sequence, $caseSensitive = true) { $len = strlen($sequence); $buffer = ''; @@ -932,7 +934,7 @@ class Tokenizer } $this->scanner->unconsume($len); - return $buffer == $sequence; + return $caseSensitive ? $buffer == $sequence : strcasecmp($buffer, $sequence) === 0; } /** diff --git a/test/HTML5/Parser/DOMTreeBuilderTest.php b/test/HTML5/Parser/DOMTreeBuilderTest.php index 9a7d5d8..5bba7cc 100644 --- a/test/HTML5/Parser/DOMTreeBuilderTest.php +++ b/test/HTML5/Parser/DOMTreeBuilderTest.php @@ -56,6 +56,28 @@ class DOMTreeBuilderTest extends \Masterminds\HTML5\Tests\TestCase $this->assertInstanceOf('\DOMDocument', $doc); $this->assertEquals('html', $doc->documentElement->tagName); } + + public function testStrangeCapitalization() + { + $html = "<!doctype html> + <html> + <head> + <Title>Hello, world!</TitlE> + </head> + <body>TheBody<script>foo</script></body> + </html>"; + $doc = $this->parse($html); + + $this->assertInstanceOf('\DOMDocument', $doc); + $this->assertEquals('html', $doc->documentElement->tagName); + + $xpath = new \DOMXPath( $doc ); + $xpath->registerNamespace( "x", "http://www.w3.org/1999/xhtml" ); + + $this->assertEquals("Hello, world!", $xpath->query( "//x:title" )->item( 0 )->nodeValue); + $this->assertEquals("foo", $xpath->query( "//x:script" )->item( 0 )->nodeValue); + } + public function testDocumentFakeAttrAbsence() { $html = "<!DOCTYPE html><html xmlns=\"http://www.w3.org/1999/xhtml\"><body>foo</body></html>"; |