Add support for PHP 8.3.

ivopetkov · Nov 2, 2023 · 2d0e3b6 · 2d0e3b6
1 parent f2b5cc9
commit 2d0e3b6
Show file tree

Hide file tree

Showing 4 changed files with 37 additions and 18 deletions.
diff --git a/composer.json b/composer.json
@@ -10,7 +10,7 @@
         }
     ],
     "require": {
-        "php": "7.0.*|7.1.*|7.2.*|7.3.*|7.4.*|8.0.*|8.1.*|8.2.*",
+        "php": "7.0.*|7.1.*|7.2.*|7.3.*|7.4.*|8.0.*|8.1.*|8.2.*|8.3.*",
         "ext-dom": "*"
     },
     "require-dev": {

diff --git a/phpunit.xml b/phpunit.xml
@@ -1,6 +1,6 @@
 <phpunit bootstrap="vendor/autoload.php">
     <testsuites>
-        <testsuite name="">
+        <testsuite name="Tests">
             <directory>./tests/</directory>
         </testsuite>
     </testsuites>

diff --git a/src/HTML5DOMDocument.php b/src/HTML5DOMDocument.php
@@ -87,9 +87,10 @@ public function __construct(string $version = '1.0', string $encoding = '')
      * Load HTML from a string.
      *
      * @param string $source The HTML code.
-     * @param int $options Additional Libxml parameters.
+     * @param integer $options Additional Libxml parameters.
      * @return boolean TRUE on success or FALSE on failure.
      */
+    #[\ReturnTypeWillChange] // While supporting PHP 7
     public function loadHTML($source, $options = 0)
     {
         // Enables libxml errors handling
@@ -131,12 +132,7 @@ public function loadHTML($source, $options = 0)
 
         // Add body tag if missing
         if ($autoAddHtmlAndBodyTags && $source !== '' && preg_match('/\<!DOCTYPE.*?\>/', $source) === 0 && preg_match('/\<html.*?\>/', $source) === 0 && preg_match('/\<body.*?\>/', $source) === 0 && preg_match('/\<head.*?\>/', $source) === 0) {
-            $source = '<body>' . $source . '</body>';
-        }
-
-        // Add DOCTYPE if missing
-        if ($autoAddDoctype && strtoupper(substr($source, 0, 9)) !== '<!DOCTYPE') {
-            $source = "<!DOCTYPE html>\n" . $source;
+            $source = '<html><body>' . $source . '</body></html>';
         }
 
         // Adds temporary head tag
@@ -154,7 +150,7 @@ public function loadHTML($source, $options = 0)
             if (isset($matches[0])) { // has html tag
                 $source = str_replace($matches[0], $matches[0] . '<head>' . $charsetTag . '</head>', $source);
             } else {
-                $source = '<head>' . $charsetTag . '</head>' . $source;
+                $source = '<html><head>' . $charsetTag . '</head></html>' . $source;
                 $removeHtmlTag = true;
             }
             $removeHeadTag = true;
@@ -164,6 +160,11 @@ public function loadHTML($source, $options = 0)
         $source = preg_replace('/&([a-zA-Z]*);/', 'html5-dom-document-internal-entity1-$1-end', $source);
         $source = preg_replace('/&#([0-9]*);/', 'html5-dom-document-internal-entity2-$1-end', $source);
 
+        // Add DOCTYPE if missing
+        if ($autoAddDoctype && strtoupper(substr($source, 0, 9)) !== '<!DOCTYPE') {
+            $source = "<!DOCTYPE html>\n" . $source;
+        }
+
         $result = parent::loadHTML('<?xml encoding="utf-8" ?>' . $source, $options);
         if ($internalErrorsOptionValue === false) {
             libxml_use_internal_errors(false);
@@ -226,10 +227,12 @@ public function loadHTML($source, $options = 0)
 
     /**
      * Load HTML from a file.
-     *
+     * 
      * @param string $filename The path to the HTML file.
-     * @param int $options Additional Libxml parameters.
+     * @param integer $options Additional Libxml parameters.
+     * @return boolean
      */
+    #[\ReturnTypeWillChange] // While supporting PHP 7
     public function loadHTMLFile($filename, $options = 0)
     {
         return $this->loadHTML(file_get_contents($filename), $options);
@@ -342,6 +345,7 @@ public function saveHTML(\DOMNode $node = null): string
             }
             $html = trim($html);
         } else {
+            //$this->modify(self::OPTIMIZE_HEAD);
             $removeHtmlElement = false;
             $removeHeadElement = false;
             $headElement = $this->getElementsByTagName('head')->item(0);
@@ -611,7 +615,7 @@ public function insertHTMLMulti(array $sources)
     /**
      * Applies the modifications specified to the DOM document.
      * 
-     * @param int $modifications The modifications to apply. Available values:
+     * @param integer $modifications The modifications to apply. Available values:
      *  - HTML5DOMDocument::FIX_MULTIPLE_TITLES - removes all but the last title elements.
      *  - HTML5DOMDocument::FIX_DUPLICATE_METATAGS - removes all but the last metatags with matching name or property attributes.
      *  - HTML5DOMDocument::FIX_MULTIPLE_HEADS - merges multiple head elements.
@@ -709,6 +713,9 @@ public function modify($modifications = 0)
                     $stylesToRemove = [];
                     $list = [];
                     foreach ($styles as $style) {
+                        if ($style->parentNode !== $headElement) {
+                            continue;
+                        }
                         $innerHTML = trim($style->innerHTML);
                         if (array_search($innerHTML, $list) === false) {
                             $list[] = $innerHTML;
@@ -753,6 +760,19 @@ public function modify($modifications = 0)
                         $headElement->insertBefore($charsetMetaTag, $headElement->firstChild);
                     }
                 }
+                $scriptTags = $headElement->getElementsByTagName('script');
+                if ($scriptTags->length > 0) {
+                    $nodesToMove = [];
+                    foreach ($scriptTags as $scriptTag) {
+                        if ($scriptTag->parentNode !== $headElement) {
+                            continue;
+                        }
+                        $nodesToMove[] = $scriptTag;
+                    }
+                    foreach ($nodesToMove as $nodeToMove) {
+                        $headElement->appendChild($nodeToMove);
+                    }
+                }
             }
         }
 

diff --git a/tests/Test.php b/tests/Test.php
@@ -125,7 +125,7 @@ public function testHtmlEntities()
     /**
      *
      */
-    public function testInserHTML()
+    public function testInsertHTML()
     {
         // insert beforeBodyEnd
         $source = '<!DOCTYPE html><html><body>'
@@ -275,7 +275,7 @@ public function testInserHTML()
         // Empty content
         $dom = new HTML5DOMDocument();
         $dom->insertHTML('');
-        $expectedSource = '<!DOCTYPE html><html></html>';
+        $expectedSource = '<!DOCTYPE html>';
         $this->assertEquals($expectedSource, $this->removeNewLines($dom->saveHTML()));
 
         // Html tag with attribute
@@ -329,7 +329,7 @@ public function testEmpty()
         $source = '<!DOCTYPE html>';
         $testSource($source, $source);
 
-        $testSource('', '<!DOCTYPE html><html></html>');
+        $testSource('', '<!DOCTYPE html>');
     }
 
     /**
@@ -1277,7 +1277,6 @@ public function testLIBXML_HTML_NODEFDTD()
      */
     public function testLIBXML_HTML_NOIMPLIED()
     {
-
         $content = '<div>hello</div>';
         $dom = new HTML5DOMDocument();
         $dom->loadHTML($content, LIBXML_HTML_NOIMPLIED);
@@ -1446,7 +1445,7 @@ public function testScriptsCDATA()
      * 
      * @return array
      */
-    public function propertyGetterTestDataProvider()
+    static public function propertyGetterTestDataProvider()
     {
         return [
             [