Merge pull request #32 from j0k3r/html5

Add missing HTML5 class
CS
2026-09-27 14:36:23 +00:00 · 2017-06-30 16:50:09 +02:00 · 2017-06-30 16:42:29 +02:00 · 2017-06-30 16:32:37 +02:00 · 2017-05-19 16:16:20 +02:00 · 2017-05-19 16:02:36 +02:00
4 changed files with 64 additions and 53 deletions
@@ -26,7 +26,8 @@
    "require": {
        "php": ">=5.3.3",
        "ext-mbstring": "*",
-        "psr/log": "^1.0"
+        "psr/log": "^1.0",
+        "electrolinux/php-html5lib": "^0.1.0"
    },
    "require-dev": {
        "satooshi/php-coveralls": "~0.6",
@@ -45,19 +45,31 @@ class JSLikeHTMLElement extends \DOMElement
     */
    public function __set($name, $value)
    {
-        if ($name === 'innerHTML') {
+        if ($name !== 'innerHTML') {
+            $trace = debug_backtrace();
+            trigger_error('Undefined property via __set(): ' . $name . ' in ' . $trace[0]['file'] . ' on line ' . $trace[0]['line'], E_USER_NOTICE);
+
+            return;
+        }
+
        // first, empty the element
        for ($x = $this->childNodes->length - 1; $x >= 0; --$x) {
            $this->removeChild($this->childNodes->item($x));
        }

        // $value holds our new inner HTML
-            if ($value !== '') {
+        $value = trim($value);
+        if (empty($value)) {
+            return;
+        }
+
+        // ensure bad entity won't generate warning
+        $previousError = libxml_use_internal_errors(true);
+
        $f = $this->ownerDocument->createDocumentFragment();

        // appendXML() expects well-formed markup (XHTML)
-                // @ to suppress PHP warnings
-                $result = @$f->appendXML($value);
+        $result = $f->appendXML($value);
        if ($result) {
            if ($f->hasChildNodes()) {
                $this->appendChild($f);
@@ -72,7 +84,7 @@ class JSLikeHTMLElement extends \DOMElement
            // We use it (and suppress the warning) because an HTML fragment will
            // be wrapped around <html><body> tags which we don't really want to keep.
            // Note: despite the warning, if loadHTML succeeds it will return true.
-                    $result = @$f->loadHTML('<htmlfragment>' . $value . '</htmlfragment>');
+            $result = $f->loadHTML('<htmlfragment>' . $value . '</htmlfragment>');

            if ($result) {
                $import = $f->getElementsByTagName('htmlfragment')->item(0);
@@ -81,16 +93,11 @@ class JSLikeHTMLElement extends \DOMElement
                    $importedNode = $this->ownerDocument->importNode($child, true);
                    $this->appendChild($importedNode);
                }
-                    } else {
-                        // oh well, we tried, we really did. :(
-                        // this element is now empty
            }
        }
-            }
-        } else {
-            $trace = debug_backtrace();
-            trigger_error('Undefined property via __set(): ' . $name . ' in ' . $trace[0]['file'] . ' on line ' . $trace[0]['line'], E_USER_NOTICE);
-        }
+
+        libxml_clear_errors();
+        libxml_use_internal_errors($previousError);
    }

    /**
@@ -2,6 +2,7 @@

 namespace Readability;

+use HTML5Lib\Parser;
 use Psr\Log\LoggerAwareInterface;
 use Psr\Log\LoggerInterface;
 use Psr\Log\NullLogger;
@@ -284,7 +285,7 @@ class Readability implements LoggerAwareInterface

        $this->html = mb_convert_encoding($this->html, 'HTML-ENTITIES', 'UTF-8');

-        if (!($this->parser === 'html5lib' && ($this->dom = \HTML5_Parser::parse($this->html)))) {
+        if (!($this->parser === 'html5lib' && ($this->dom = Parser::parse($this->html)))) {
            libxml_use_internal_errors(true);

            $this->dom = new \DOMDocument();
@@ -33,6 +33,15 @@ class ReadabilityTest extends \PHPUnit_Framework_TestCase
        $this->assertInstanceOf('DomDocument', $readability->dom);
    }

+    public function testConstructHtml5Parser()
+    {
+        $readability = $this->getReadability('<html/>', 'http://0.0.0.0', 'html5lib');
+
+        $this->assertEquals('http://0.0.0.0', $readability->url);
+        $this->assertInstanceOf('DomDocument', $readability->dom);
+        $this->assertEquals('<html/>', $readability->original_html);
+    }
+
    /**
     * @requires extension tidy
     */
@@ -326,13 +335,6 @@ class ReadabilityTest extends \PHPUnit_Framework_TestCase
        $this->assertNotContains('This text is also an awesome text and you should know that', $readability->getContent()->innerHTML);
    }

-    // public function testConstructParser()
-    // {
-    //     $readability = $this->getReadability('<html/>', 'http://0.0.0.0', 'html5lib');
-
-    //     $this->assertEquals('http://0.0.0.0', $readability->url);
-    // }
-
    // dummy function to be used to the next test
    public function error2Exception($code, $string, $file, $line, $context)
    {
@@ -443,7 +445,7 @@ class ReadabilityTest extends \PHPUnit_Framework_TestCase

    public function testPostFilters()
    {
-        $readability = $this->getReadability('<div>' . str_repeat('<p>This <b>is</b> the awesome content :)</p>', 7) . '</div>', 'http://0.0.0.0');
+        $readability = $this->getReadability('<div>' . str_repeat('<p>This <b>is</b> the awesome content :)</p>', 10) . '</div>', 'http://0.0.0.0');
        $readability->addPostFilter('!<strong[^>]*>(.*?)</strong>!is', '');

        $res = $readability->init();
Author	SHA1	Message	Date
Jérémy BenoistandGitHub	6427dd7371	Merge pull request #32 from j0k3r/html5 Add missing HTML5 class	2017-06-30 16:50:09 +02:00
Jeremy Benoist	613a63c062	CS	2017-06-30 16:42:29 +02:00
Jeremy Benoist	05089bbd03	Add missing HTML5 class	2017-06-30 16:32:37 +02:00
Jérémy BenoistandGitHub	5ac82f3c87	Merge pull request #30 from j0k3r/avoid-php-warning Avoid php warning	2017-05-19 16:16:20 +02:00
Jeremy Benoist	85fb92a042	Fix tests	2017-05-19 16:02:36 +02:00
Jeremy Benoist	f2a43b476c	Avoid PHP Warning This isn't the best solution but the previous one using `@` wasn't really better. Appending a string into a fragment might generate some warning if the string contains bad entity. For example `+`.	2017-05-19 15:37:46 +02:00