Merge pull request #2675 from mficzel/bugfix/augmenterHandlesUnicodeC…

…haractersCorrectly BUGFIX: HTML augmenter preserves multibyte characters in attributes
neos · Sep 3, 2019 · b06c25a · b06c25a
2 parents 57977a2 + fa57e7e
commit b06c25a
Show file tree

Hide file tree

Showing 2 changed files with 8 additions and 1 deletion.
diff --git a/Classes/Service/HtmlAugmenter.php b/Classes/Service/HtmlAugmenter.php
@@ -69,7 +69,7 @@ protected function getHtmlRootElement($html)
         $domDocument = new \DOMDocument('1.0', 'UTF-8');
         // ignore parsing errors
         $useInternalErrorsBackup = libxml_use_internal_errors(true);
-        $domDocument->loadHTML($html);
+        $domDocument->loadHTML((substr($html, 0, 5) === '<?xml') ? $html : '<?xml encoding="UTF-8"?>' . $html);
         $xPath = new \DOMXPath($domDocument);
         $rootElement = $xPath->query('//html/body/*');
         if ($useInternalErrorsBackup !== true) {

diff --git a/Tests/Unit/Service/HtmlAugmenterTest.php b/Tests/Unit/Service/HtmlAugmenterTest.php
@@ -190,6 +190,13 @@ public function __toString() {
                 'exclusiveAttributes' => null,
                 'expectedResult' => '<p data-bar="öäüß" data-foo="öäüß">valid characters are decoded</p>',
             ),
+            [
+                'html' => '<p data-foo="öäüß🦆">valid characters are untouched</p>',
+                'attributes' => ['data-bar' => 'öäüß🦆'],
+                'fallbackTagName' => null,
+                'exclusiveAttributes' => null,
+                'expectedResult' => '<p data-bar="öäüß🦆" data-foo="öäüß🦆">valid characters are untouched</p>',
+            ],
 
             // exclusive attributes
             array(