Merge branch 'fix_encode' of https://github.com/shirosaki/roundcubemail into shirosaki-fix_encode

Refactor the new code

program/include/rcmail_sendmail.php

@@ -680,6 +680,15 @@ class rcmail_sendmail
      */
     public function email_input_format($mailto, $count = false, $check = true)
     {
+        // convert to UTF-8 to preserve \x2c(,) and \x3b(;) used in ISO-2022-JP;
+        $charset = $this->options['charset'];
+        if ($charset != RCUBE_CHARSET) {
+            $mailto = rcube_charset::convert($mailto, $charset, RCUBE_CHARSET);
+        }
+        if (preg_match('/ISO-2022/i', $charset)) {
+            $use_base64 = true;
+        }
+
         // simplified email regexp, supporting quoted local part
         $email_regexp = '(\S+|("[^"]+"))@\S+';
 
@@ -711,7 +720,16 @@ class rcmail_sendmail
                 if ($name[0] == '"' && $name[strlen($name)-1] == '"') {
                     $name = substr($name, 1, -1);
                 }
+
+                // encode "name" field
+                if (!empty($use_base64)) {
+                    $name = rcube_charset::convert($name, RCUBE_CHARSET, $charset);
+                    $name = Mail_mimePart::encodeMB($name, $charset, 'base64');
+                }
+                else {
                     $name = stripcslashes($name);
+                }
+
                 $address  = rcube_utils::idn_to_ascii(trim($address, '<>'));
                 $result[] = format_email_recipient($address, $name);
                 $item     = $address;

program/lib/Roundcube/rcube_charset.php

@@ -63,6 +63,8 @@ class rcube_charset
         '238'           => 'WINDOWS-1250',
         'MS950'         => 'CP950',
         'WINDOWS949'    => 'UHC',
+        'WINDOWS1257'   => 'ISO-8859-13',
+        'ISO2022JP'     => 'ISO-2022-JP-MS',
     );
 
     /**
@@ -270,7 +272,7 @@ class rcube_charset
      */
     public static function convert($str, $from, $to = null)
     {
-        $to   = empty($to) ? RCUBE_CHARSET : strtoupper($to);
+        $to   = empty($to) ? RCUBE_CHARSET : self::parse_charset($to);
         $from = self::parse_charset($from);
 
         // It is a common case when UTF-16 charset is used with US-ASCII content (#1488654)
@@ -283,15 +285,6 @@ class rcube_charset
             return $str;
         }
 
-        $aliases = array(
-            'WINDOWS-1257' => 'ISO-8859-13',
-            'US-ASCII'     => 'ASCII',
-            'ISO-2022-JP'  => 'ISO-2022-JP-MS',
-        );
-
-        $mb_from = $aliases[$from] ?: $from;
-        $mb_to   = $aliases[$to] ?: $to;
-
         // Ignore invalid characters
         $mbstring_sc = mb_substitute_character();
         mb_substitute_character('none');
@@ -300,7 +293,7 @@ class rcube_charset
         // using mb_check_encoding() is much slower
         set_error_handler(array('rcube_charset', 'error_handler'), E_WARNING);
         try {
-            $out = mb_convert_encoding($str, $mb_to, $mb_from);
+            $out = mb_convert_encoding($str, $to, $from);
         }
         catch (ErrorException $e) {
             $out = false;

program/lib/Roundcube/rcube_mime.php

@@ -592,6 +592,7 @@ class rcube_mime
         //       Iconv's substr/strlen are 100x slower (#1489113)
 
         if ($charset && $charset != RCUBE_CHARSET) {
+            $charset = rcube_charset::parse_charset($charset);
             mb_internal_encoding($charset);
         }
 

tests/Framework/Charset.php

@@ -65,11 +65,12 @@ class Framework_Charset extends PHPUnit\Framework\TestCase
     {
         return array(
             array('ö', 'ö', 'UTF-8', 'UTF-8'),
-            array('ö', '', 'UTF-8', 'US-ASCII'),
+            array('ö', '', 'UTF-8', 'ASCII'),
             array('aż', 'a', 'UTF-8', 'US-ASCII'),
             array('&BCAEMARBBEEESwQ7BDoEOA-', 'Рассылки', 'UTF7-IMAP', 'UTF-8'),
             array('Рассылки', '&BCAEMARBBEEESwQ7BDoEOA-', 'UTF-8', 'UTF7-IMAP'),
             array(base64_decode('GyRCLWo7M3l1OSk2SBsoQg=='), '㈱山﨑工業', 'ISO-2022-JP', 'UTF-8'),
+            array('㈱山﨑工業', base64_decode('GyRCLWo7M3l1OSk2SBsoQg=='), 'UTF-8', 'ISO-2022-JP'),
         );
     }
 

tests/Framework/Mime.php

@@ -238,6 +238,10 @@ class Framework_Mime extends PHPUnit\Framework\TestCase
                 array("this-is-just-some-blabla-to-make-this-more-than-seventy-five-characters-in-a-row -- this line should be wrapped", 20, "\n"),
                 "this-is-just-some-blabla-to-make-this-more-than-seventy-five-characters-in-a-row\n-- this line should\nbe wrapped",
             ),
+            array(
+                array(rcube_charset::convert("㈱山﨑工業", 'UTF-8', 'ISO-2022-JP'), 1, "\n", true, 'ISO-2022-JP'),
+                rcube_charset::convert("㈱\n山\n﨑\n工\n業", 'UTF-8', 'ISO-2022-JP'),
+            ),
         );
 
         foreach ($samples as $sample) {

tests/Rcmail/Sendmail.php

@@ -7,6 +7,62 @@
  */
 class Rcmail_RcmailSendmail extends PHPUnit\Framework\TestCase
 {
+
+    /**
+     * Data for test_convert()
+     */
+    function data_email_input_format()
+    {
+        return array(
+            array(
+                'name <t@domain.jp>',
+                'name <t@domain.jp>',
+                'UTF-8'
+            ),
+            array(
+                '"first last" <t@domain.jp>',
+                'first last <t@domain.jp>',
+                'UTF-8'
+            ),
+            array(
+                '"first last" <t@domain.jp>, test2@domain.tld,',
+                'first last <t@domain.jp>, test2@domain.tld',
+                'UTF-8'
+            ),
+            array(
+                '<test@domain.tld>',
+                'test@domain.tld',
+                'UTF-8'
+            ),
+            array(
+                'test@domain.tld',
+                'test@domain.tld',
+                'UTF-8'
+            ),
+            array(
+                'ö <t@test.com>',
+                'ö <t@test.com>',
+                'UTF-8'
+            ),
+            array(
+                base64_decode('GyRCLWo7M3l1OSk2SBsoQg==') . ' <t@domain.jp>',
+                '=?ISO-2022-JP?B?GyRCLWo7M3l1OSk2SBsoQg==?= <t@domain.jp>',
+                'ISO-2022-JP'
+            ),
+        );
+    }
+
+    /**
+     * @dataProvider data_email_input_format
+     */
+    function test_email_input_format($input, $output, $charset)
+    {
+        $sendmail = new rcmail_sendmail();
+        $sendmail->options['charset'] = $charset;
+
+        $this->assertEquals($output, $sendmail->email_input_format($input));
+    }
+
     /**
      * Test rcmail_sendmail::identity_select()
      */