diff options
| -rw-r--r-- | README.md | 2 | ||||
| -rw-r--r-- | src/HTML5.php | 4 | ||||
| -rw-r--r-- | src/HTML5/Serializer/Serializer.php (renamed from src/HTML5/Serializer.php) | 2 | ||||
| -rw-r--r-- | src/HTML5/Serializer/Traverser.php (renamed from src/HTML5/Traverser.php) | 59 | ||||
| -rw-r--r-- | test/HTML5/Serializer/SerializerTest.php (renamed from test/HTML5/SerializerTest.php) | 12 | ||||
| -rw-r--r-- | test/HTML5/Serializer/TraverserTest.php | 78 |
6 files changed, 105 insertions, 52 deletions
@@ -45,7 +45,7 @@ $dom = HTML5::parse($html); print HTML5::saveHTML($dom); // Or save it to a file: -HTML5::save('out.html'); +HTML5::save($dom, 'out.html'); ?> ``` diff --git a/src/HTML5.php b/src/HTML5.php index c28e4c1..31db307 100644 --- a/src/HTML5.php +++ b/src/HTML5.php @@ -14,7 +14,7 @@ class HTML5 extends \HTML5\Parser { * Save a DOM into a given file as HTML5. */ public static function save($dom, $file) { - $serializer = new \HTML5\Serializer($dom); + $serializer = new \HTML5\Serializer\Serializer($dom); return $serializer->save($file); } @@ -22,7 +22,7 @@ class HTML5 extends \HTML5\Parser { * Convert a DOM into an HTML5 string. */ public static function saveHTML($dom) { - $serializer = new \HTML5\Serializer($dom); + $serializer = new \HTML5\Serializer\Serializer($dom); return $serializer->saveHTML(); } } diff --git a/src/HTML5/Serializer.php b/src/HTML5/Serializer/Serializer.php index 2bed0e8..f16bbe2 100644 --- a/src/HTML5/Serializer.php +++ b/src/HTML5/Serializer/Serializer.php @@ -2,7 +2,7 @@ /** * A simple serializer that walks the DOM tree and outputs HTML5. */ -namespace HTML5; +namespace HTML5\Serializer; /** * Transform a DOM into an HTML5 document. diff --git a/src/HTML5/Traverser.php b/src/HTML5/Serializer/Traverser.php index 25375dd..6d5e805 100644 --- a/src/HTML5/Traverser.php +++ b/src/HTML5/Serializer/Traverser.php @@ -1,5 +1,5 @@ <?php -namespace HTML5; +namespace HTML5\Serializer; /** * Traverser for walking a DOM tree. @@ -12,32 +12,7 @@ namespace HTML5; */ class Traverser { - // TODO: Refactor this into an element mask. - static $block_elements = array( - 'html' => 1, - 'body' => 1, - 'head' => 1, - 'p' => 1, - 'div' => 1, - 'h1' => 1, - 'h2' => 1, - 'h3' => 1, - 'h4' => 1, - 'h5' => 1, - 'h6' => 1, - 'title' => 1, - 'script' => 1, - 'link' => 1, - 'meta' => 1, - 'section' => 1, - 'article' => 1, - 'table' => 1, - 'tbody' => 1, - 'tr' => 1, - 'th' => 1, - 'td' => 1, - //'form' => 1, - ); + static $block_elements = 'html|body|head|p|div|h[1-6]|title|script|link|meta|section|article|table|tbody|tr|th|td'; // TODO: Refactor this into an element mask. static $literal_elements = array( @@ -58,19 +33,7 @@ class Traverser { * input, keygen, link, meta, param, source, track or wbr element, then * continue on to the next child node at this point. */ - static $unary_elements = array( - 'area' => 1, - 'base' => 1, - 'basefont' => 1, - 'bgsound' => 1, - 'br' => 1, - 'col' => 1, - 'command' => 1, - 'embed' => 1, - 'frame' => 1, - 'hr' => 1, - 'img' => 1, - ); + static $unary_elements = 'area|base|basefont|bgsound|br|col|command|embed|frame|hr|img'; /** Namespaces that should be treated as "local" to HTML5. */ static $local_ns = array( @@ -293,16 +256,28 @@ class Traverser { /** * Is an unary tag. + * + * @param string $name + * The name of the element to test. + * + * @return bool + * True if Unary and false otherwise. */ protected function isUnary($name) { - return isset(self::$unary_elements[$name]); + return (bool)preg_match('/^(' . self::$unary_elements . ')$/i', $name); } /** * Is block element. + * + * @param string $name + * The name of the element to test if a block level element + * + * @return bool + * If the element is block level or not. */ protected function isBlock($name) { - return isset(self::$block_elements[$name]); + return (bool)preg_match('/^(' . self::$block_elements . ')$/i', $name); } protected function isLiteral($element) { diff --git a/test/HTML5/SerializerTest.php b/test/HTML5/Serializer/SerializerTest.php index ee8804e..a06fd50 100644 --- a/test/HTML5/SerializerTest.php +++ b/test/HTML5/Serializer/SerializerTest.php @@ -3,9 +3,9 @@ namespace HTML5\Tests; -use \HTML5\Serializer; +use \HTML5\Serializer\Serializer; -require_once 'TestCase.php'; +require_once __DIR__ . '/../TestCase.php'; /** * Test the Serializer. @@ -13,14 +13,14 @@ require_once 'TestCase.php'; * These tests are all dependent upon the parser. So if the parser * fails, the results of the serializer tests may not be conclusive. */ -class SerializerTest extends TestCase { +class SerializerTest extends \HTML5\Tests\TestCase { /** * Parse and serialize a string. */ protected function cycle($html) { $dom = \HTML5::parse($html); - $ser = new \HTML5\Serializer($dom, FALSE); + $ser = new Serializer($dom, FALSE); $out = $ser->saveHTML(); return $out; @@ -32,7 +32,7 @@ class SerializerTest extends TestCase { $dom = \HTML5::parse($html); $this->assertTrue($dom instanceof \DOMDocument, "Canary"); - $ser = new \HTML5\Serializer($dom, FALSE); + $ser = new Serializer($dom, FALSE); $out = $ser->saveHTML(); $this->assertTrue(count($out) >= count($html), 'Byte counts'); @@ -47,7 +47,7 @@ class SerializerTest extends TestCase { $dom = \HTML5::parse($html); $this->assertTrue($dom instanceof \DOMDocument, "Canary"); - $ser = new \HTML5\Serializer($dom, FALSE); + $ser = new Serializer($dom, FALSE); $out = fopen("php://temp", "w"); $ser->save($out); diff --git a/test/HTML5/Serializer/TraverserTest.php b/test/HTML5/Serializer/TraverserTest.php new file mode 100644 index 0000000..0717a75 --- /dev/null +++ b/test/HTML5/Serializer/TraverserTest.php @@ -0,0 +1,78 @@ +<?php +namespace HTML5\Tests; + +use \HTML5\Serializer\Traverser; + +require_once __DIR__ . '/../TestCase.php'; + +class TraverserTest extends \HTML5\Tests\TestCase { + + /** + * Using reflection we make a protected method accessible for testing. + * + * @param string $name + * The name of the method on the Traverser class to test. + * + * @return \ReflectionMethod + * \ReflectionMethod for the specified method + */ + function getProtectedMethod($name) { + $class = new \ReflectionClass('\HTML5\Serializer\Traverser'); + $method = $class->getMethod($name); + $method->setAccessible(true); + return $method; + } + + public function testIsBlock() { + $blocks = array('html', 'body', 'head', 'p', 'div', 'h1', 'h2', 'h3', 'h4', + 'h5', 'h6', 'title', 'script', 'link', 'meta', 'section', 'article', + 'table', 'tbody', 'tr', 'th', 'td', + //'form', + ); + + // Mocking the required input because there is no checking. + $t = new Traverser('', ''); + $method = $this->getProtectedMethod('isBlock'); + + foreach ($blocks as $block) { + $this->assertTrue($method->invoke($t, $block), 'Block test failed on: ' . $block); + + // Also test the uppercase version. + $this->assertTrue($method->invoke($t, strtoupper($block)), 'Block test failed on: ' . strtoupper($block)); + } + + $nonblocks = array('span', 'a', 'img'); + foreach ($nonblocks as $tag) { + $this->assertFalse($method->invoke($t, $tag), 'Block test failed on: ' . $tag); + + // Also test the uppercase version. + $this->assertFalse($method->invoke($t, strtoupper($tag)), 'Block test failed on: ' . strtoupper($tag)); + } + } + + public function testIsUnary() { + $elements = array( 'area', 'base', 'basefont', 'bgsound', 'br', 'col', + 'command', 'embed', 'frame', 'hr', 'img', + ); + + // Mocking the required input because there is no checking. + $t = new Traverser('', ''); + $method = $this->getProtectedMethod('isUnary'); + + foreach ($elements as $element) { + $this->assertTrue($method->invoke($t, $element), 'Unary test failed on: ' . $element); + + // Also test the uppercase version. + $this->assertTrue($method->invoke($t, strtoupper($element)), 'Unary test failed on: ' . strtoupper($element)); + } + + $nonblocks = array('span', 'a', 'div'); + foreach ($nonblocks as $tag) { + $this->assertFalse($method->invoke($t, $tag), 'Unary test failed on: ' . $tag); + + // Also test the uppercase version. + $this->assertFalse($method->invoke($t, strtoupper($tag)), 'Unary test failed on: ' . strtoupper($tag)); + } + } + +}
\ No newline at end of file |