[+]: remove content before "<!doctype.*>", otherwise DOMDocument cann…

…ot handle the input -> fix for issue #55
voku · Aug 11, 2020 · 4f70058 · 4f70058
1 parent fa7e00e
commit 4f70058
Show file tree

Hide file tree

Showing 4 changed files with 43 additions and 2 deletions.
diff --git a/CHANGELOG.md b/CHANGELOG.md
@@ -1,3 +1,8 @@
+# Changelog 4.4.8 (2020-08-11) 
+
+- remove content before "<!doctype.*>", otherwise DOMDocument cannot handle the input
+- use a new version of "voku/simple_html_dom" (4.7.22)
+
 # Changelog 4.4.7 (2020-08-11) 
 
 - use a new version of "voku/simple_html_dom" (4.7.21)

diff --git a/composer.json b/composer.json
@@ -19,7 +19,7 @@
   ],
   "require": {
     "php": ">=7.0.0",
-    "voku/simple_html_dom": "~4.7.21",
+    "voku/simple_html_dom": "~4.7.23",
     "ext-dom": "*"
   },
   "require-dev": {

diff --git a/src/voku/helper/HtmlMin.php b/src/voku/helper/HtmlMin.php
@@ -1570,11 +1570,23 @@ private function minifyHtmlDom($html, $multiDecodeNewHtmlEntity): string
         $dom->getDocument()->preserveWhiteSpace = false; // remove redundant white space
         $dom->getDocument()->formatOutput = false; // do not formats output with indentation
 
+        // Remove content before <!DOCTYPE.*> because otherwise the DOMDocument can not handle the input.
+        if (\stripos($html, '<!DOCTYPE') !== false) {
+            /** @noinspection NestedPositiveIfStatementsInspection */
+            if (
+                \preg_match('/(^.*?)<!(?:DOCTYPE)(?: [^>]*)?>/sui', $html, $matches_before_doctype)
+                &&
+                \trim($matches_before_doctype[1])
+            ) {
+                $html = \str_replace($matches_before_doctype[1], '', $html);
+            }
+        }
+
         // load dom
         /** @noinspection UnusedFunctionResultInspection */
         $dom->loadHtml($html);
 
-        $this->withDocType = (\stripos(\ltrim($html), '<!DOCTYPE') === 0);
+        $this->withDocType = (\stripos($html, '<!DOCTYPE') === 0);
 
         $doctypeStr = $this->getDoctype($dom->getDocument());
 

diff --git a/tests/HtmlMinTest.php b/tests/HtmlMinTest.php
@@ -868,6 +868,30 @@ public function testBrokenHtmlExample()
         static::assertSame($expected, $htmlMin->minify($html));
     }
 
+    public function testContentBeforeDoctypeExample()
+    {
+        // init
+        $htmlMin = new HtmlMin();
+        $htmlMin->useKeepBrokenHtml(true);
+
+        $html = '<!-- === BEGIN TOP === --><!DOCTYPE html>
+        <!--[if IE 8]> <html lang="en" class="ie8"> <![endif]-->
+        <!--[if IE 9]> <html lang="en" class="ie9"> <![endif]-->
+        <!--[if !IE]><!-->
+        <html prefix="og: http://ogp.me/ns#" lang="ru">
+        <!--<![endif]-->
+        <head>
+        <!-- Title -->
+        <title>test</title>
+        </head>
+        <body>lall</body></html>
+        ';
+
+        $expected = '<!DOCTYPE html><!--[if IE 8]> <html lang="en" class="ie8"> <![endif]--><!--[if IE 9]> <html lang="en" class="ie9"> <![endif]--><!--[if !IE]><!--><html prefix="og: http://ogp.me/ns#" lang=ru> <!--<![endif]--> <head><title>test</title> <body>lall';
+
+        static::assertSame($expected, $htmlMin->minify($html));
+    }
+
     public function testDoNotCompressTag()
     {
         $minifier = new HtmlMin();