Transcode mixed encodings to UTF-8

Author: clue

composer.json

@@ -17,7 +17,8 @@
         "php": ">=5.3",
         "react/event-loop": "~0.4.0|~0.3.0",
         "clue/buzz-react": "~0.2.0",
-        "ext-simplexml": "*"
+        "ext-simplexml": "*",
+        "neitanod/forceutf8": "~1.4"
     },
     "require-dev": {
         "clue/block-react": "~0.1.0"

src/Io/Loader.php

@@ -26,6 +26,9 @@ public function loadXmlFile($path)
 
     public function loadXmlString($html)
     {
+        // log output often uses garbled ISO-8859-1 and UTF-8 encodings
+        $html = \ForceUTF8\Encoding::toUTF8($html);
+
         // fix invalid markup of outdated ViewVC versions
         // - help link in footer not terminated
         // - selected branch/tag in CVS "sticky tag" dropdown has not attribute value

tests/Io/LoaderTest.php

@@ -39,6 +39,15 @@ public function testHtmlEntities()
         $this->assertEquals('ä… ©', (string)$xml);
     }
 
+    public function testMixedEncodings()
+    {
+        // mixed UTF-8 and ISO-8859-1
+        $str = "<p>ä and \xFC</p>";
+        $xml = $this->loader->loadXmlString($str);
+
+        $this->assertEquals('ä and ü', (string)$xml);
+    }
+
     public function testLoadInvalidMarkupInputNotClosed()
     {
         $str = '<input type="hidden">';