Generalize charset converting

tedious · Aug 28, 2015 · 7cb5de0 · 7cb5de0
1 parent 3fbee96
commit 7cb5de0
Show file tree

Hide file tree

Showing 3 changed files with 58 additions and 22 deletions.
diff --git a/src/Fetch/MIME.php b/src/Fetch/MIME.php
@@ -37,7 +37,7 @@ public static function decode($text, $targetCharset = 'utf-8')
         foreach (imap_mime_header_decode($text) as $word) {
             $ch = 'default' === $word->charset ? 'ascii' : $word->charset;
 
-            $result .= iconv($ch, $targetCharset, $word->text);
+            $result .= Message::charsetConvert($word->text, $ch, $targetCharset) ?: $text;
         }
 
         return $result;

diff --git a/src/Fetch/Message.php b/src/Fetch/Message.php
@@ -520,27 +520,8 @@ protected function processStructure($structure, $partIdentifier = null)
 
             $messageBody = self::decode($messageBody, $structure->encoding);
 
-            if (!empty($parameters['charset']) && $parameters['charset'] !== self::$charset) {
-                $mb_converted = false;
-                if (function_exists('mb_convert_encoding')) {
-                    if (!in_array($parameters['charset'], mb_list_encodings())) {
-                        if ($structure->encoding === 0) {
-                            $parameters['charset'] = 'US-ASCII';
-                        } else {
-                            $parameters['charset'] = 'UTF-8';
-                        }
-                    }
-
-                    $messageBody = @mb_convert_encoding($messageBody, self::$charset, $parameters['charset']);
-                    $mb_converted = true;
-                }
-                if (!$mb_converted) {
-                    $messageBodyConv = @iconv($parameters['charset'], self::$charset . self::$charsetFlag, $messageBody);
-
-                    if ($messageBodyConv !== false) {
-                        $messageBody = $messageBodyConv;
-                    }
-                }
+            if (!empty($parameters['charset'])) {
+                $messageBody = self::charsetConvert($messageBody, $parameters['charset'], self::$charset) ?: $messageBody;
             }
 
             if (strtolower($structure->subtype) === 'plain' || ($structure->type == 1 && strtolower($structure->subtype) !== 'alternative')) {
@@ -575,6 +556,46 @@ protected function processStructure($structure, $partIdentifier = null)
         }
     }
 
+    /**
+     * @param string $text
+     * @param string $from
+     * @param string $to
+     *
+     * @return string|null
+     */
+    public static function charsetConvert($text, $from, $to = null)
+    {
+        if (!$text) {
+            return '';
+        }
+
+        if (null === $to) {
+            $to = self::$charset;
+        }
+
+        $from = strtolower($from);
+        $to   = strtolower($to);
+
+        if ($from === $to) {
+            return $text;
+        }
+
+        $converted = null;
+        if (!$converted && function_exists('mb_convert_encoding') && @mb_check_encoding($text, $from)) {
+            $converted = @mb_convert_encoding($text, $to, $from);
+        }
+
+        if (!$converted && function_exists('iconv')) {
+            $converted = @iconv($from, $to . self::$charsetFlag, $text);
+        }
+
+        if ($converted) {
+            return $converted;
+        }
+
+        return null;
+    }
+
     /**
      * This function takes in the message data and encoding type and returns the decoded data.
      *

diff --git a/tests/Fetch/Test/MessageTest.php b/tests/Fetch/Test/MessageTest.php
@@ -240,6 +240,21 @@ public function testMoveToMailbox()
         $this->assertEquals($sentFolderNumStart + 1, $server->numMessages(), 'Message moved into Sent Folder.');
     }
 
+    public function testCharsetConvert()
+    {
+        $this->assertSame('Привет', Message::charsetConvert(
+            implode(array_map('chr', array(0xF0, 0xD2, 0xC9, 0xD7, 0xC5, 0xD4))),
+            'koi8-r',
+            'utf-8'
+        ));
+
+        $this->assertSame('test', Message::charsetConvert('test', 'unk1', 'unk1'), 'Same charsets not try converting');
+        $this->assertSame('', Message::charsetConvert('', 'unk1', 'unk1'), 'Empty text not try converting');
+
+        $this->assertSame(null, Message::charsetConvert('test', 'unk1', 'utf-8'), 'Null when source charset is unknown');
+        $this->assertSame(null, Message::charsetConvert('test', 'utf-8', 'unk1'), 'Null when destination charset is unknown');
+    }
+
     public function testDecode()
     {
         $quotedPrintableDecoded = "Now's the time for all folk to come to the aid of their country.";