[DomCrawler] Fixed charset detection in html5 meta charset tag

uwej711 · May 26, 2014 · 172e752 · 172e752
1 parent 29341fa
commit 172e752
Show file tree

Hide file tree

Showing 2 changed files with 7 additions and 1 deletion.
diff --git a/src/Symfony/Component/DomCrawler/Crawler.php b/src/Symfony/Component/DomCrawler/Crawler.php
@@ -108,8 +108,10 @@ public function addContent($content, $type = null)
             }
         }
 
+        // http://www.w3.org/TR/encoding/#encodings
+        // http://www.w3.org/TR/REC-xml/#NT-EncName
         if (null === $charset &&
-            preg_match('/\<meta[^\>]+charset *= *["\']?([a-zA-Z\-0-9]+)/i', $content, $matches)) {
+            preg_match('/\<meta[^\>]+charset *= *["\']?([a-zA-Z\-0-9_:.]+)/i', $content, $matches)) {
             $charset = $matches[1];
         }
 

diff --git a/src/Symfony/Component/DomCrawler/Tests/CrawlerTest.php b/src/Symfony/Component/DomCrawler/Tests/CrawlerTest.php
@@ -232,6 +232,10 @@ public function testAddContent()
         $crawler = new Crawler();
         $crawler->addContent('<html><meta http-equiv="Content-Type" content="text/html; charset=utf-8" /><span>中文</span></html>');
         $this->assertEquals('中文', $crawler->filterXPath('//span')->text(), '->addContent() guess wrong charset');
+
+        $crawler = new Crawler();
+        $crawler->addContent(mb_convert_encoding('<html><head><meta charset="Shift_JIS"></head><body>日本語</body></html>', 'SJIS', 'UTF-8'));
+        $this->assertEquals('日本語', $crawler->filterXPath('//body')->text(), '->addContent() can recognize "Shift_JIS" in html5 meta charset tag');
     }
 
     /**