Improve line wrapping behavior where message charset is changed by plugins (including html2plaintext conversion)

5 files changed, 34 insertions, 17 deletions

diff --git a/program/include/rcube_bc.inc b/program/include/rcube_bc.inc
index 1894873e6..4130d4dc5 100644
--- a/program/include/rcube_bc.inc
+++ b/program/include/rcube_bc.inc
@@ -338,9 +338,9 @@ function show_bytes($bytes)
     return rcmail::get_instance()->show_bytes($bytes);
 }
 
-function rc_wordwrap($string, $width=75, $break="\n", $cut=false)
+function rc_wordwrap($string, $width=75, $break="\n", $cut=false, $charset=null)
 {
-    return rcube_mime::wordwrap($string, $width, $break, $cut);
+    return rcube_mime::wordwrap($string, $width, $break, $cut, $charset);
 }
 
 function rc_request_header($name)
diff --git a/program/include/rcube_mime.php b/program/include/rcube_mime.php
index 2dbeaf67a..ac3ef9e8a 100644
--- a/program/include/rcube_mime.php
+++ b/program/include/rcube_mime.php
@@ -529,10 +529,11 @@ class rcube_mime
      *
      * @param string $text Text to wrap
      * @param int $length Length
+     * @param string $charset Character encoding of $text
      *
      * @return string Wrapped text
      */
-    public static function format_flowed($text, $length = 72)
+    public static function format_flowed($text, $length = 72, $charset=null)
     {
         $text = preg_split('/\r?\n/', $text);
 
@@ -542,10 +543,10 @@ class rcube_mime
                     $prefix = $regs[0];
                     $level = strlen($prefix);
                     $line  = rtrim(substr($line, $level));
-                    $line  = $prefix . self::wordwrap($line, $length - $level - 2, " \r\n$prefix ");
+                    $line  = $prefix . self::wordwrap($line, $length - $level - 2, " \r\n$prefix ", false, $charset);
                 }
                 else if ($line) {
-                    $line = self::wordwrap(rtrim($line), $length - 2, " \r\n");
+                    $line = self::wordwrap(rtrim($line), $length - 2, " \r\n", false, $charset);
                     // space-stuffing
                     $line = preg_replace('/(^|\r\n)(From| |>)/', '\\1 \\2', $line);
                 }
@@ -565,12 +566,16 @@ class rcube_mime
      * @param int    $width   Line width
      * @param string $break   Line separator
      * @param bool   $cut     Enable to cut word
+     * @param string $charset Charset of $string
      *
      * @return string Text
      */
-    public static function wordwrap($string, $width=75, $break="\n", $cut=false)
+    public static function wordwrap($string, $width=75, $break="\n", $cut=false, $charset=null)
     {
-        $para   = explode($break, $string);
+        if ($charset)
+            mb_internal_encoding($charset);
+
+        $para   = preg_split('/\r?\n/', $string);
         $string = '';
 
         while (count($para)) {
@@ -624,6 +629,9 @@ class rcube_mime
             }
         }
 
+        if ($charset)
+            mb_internal_encoding(RCMAIL_CHARSET);
+
         return $string;
     }
 
diff --git a/program/lib/html2text.php b/program/lib/html2text.php
index dd413e0d6..34c719302 100644
--- a/program/lib/html2text.php
+++ b/program/lib/html2text.php
@@ -135,6 +135,14 @@ class html2text
     var $width = 70;
 
     /**
+     *  Target character encoding for output text
+     *
+     *  @var string $charset
+     *  @access public
+     */
+    var $charset = 'UTF-8';
+
+    /**
      *  List of preg* regular expression patterns to search for,
      *  used in conjunction with $replace.
      *
@@ -347,7 +355,7 @@ class html2text
      *  @access public
      *  @return void
      */
-    function html2text( $source = '', $from_file = false, $do_links = true, $width = 75 )
+    function html2text( $source = '', $from_file = false, $do_links = true, $width = 75, $charset = 'UTF-8' )
     {
         if ( !empty($source) ) {
             $this->set_html($source, $from_file);
@@ -356,6 +364,7 @@ class html2text
         $this->set_base_url();
         $this->_do_links = $do_links;
         $this->width = $width;
+        $this->charset = $charset;
     }
 
     /**
@@ -517,7 +526,7 @@ class html2text
         $text = preg_replace($this->ent_search, $this->ent_replace, $text);
 
         // Replace known html entities
-        $text = html_entity_decode($text, ENT_QUOTES, 'UTF-8');
+        $text = html_entity_decode($text, ENT_QUOTES, $this->charset);
 
         // Remove unknown/unhandled entities (this cannot be done in search-and-replace block)
         $text = preg_replace('/&([a-zA-Z0-9]{2,6}|#[0-9]{2,4});/', '', $text);
@@ -732,14 +741,14 @@ class html2text
      */
     private function _strtoupper($str)
     {
-        $str = html_entity_decode($str, ENT_COMPAT, RCMAIL_CHARSET);
+        $str = html_entity_decode($str, ENT_COMPAT, $this->charset);
 
         if (function_exists('mb_strtoupper'))
             $str = mb_strtoupper($str);
         else
             $str = strtoupper($str);
 
-        $str = htmlspecialchars($str, ENT_COMPAT, RCMAIL_CHARSET);
+        $str = htmlspecialchars($str, ENT_COMPAT, $this->charset);
 
         return $str;
     }
diff --git a/program/steps/mail/sendmail.inc b/program/steps/mail/sendmail.inc
index 348ac9b96..9f5080d0f 100644
--- a/program/steps/mail/sendmail.inc
+++ b/program/steps/mail/sendmail.inc
@@ -559,8 +559,8 @@ if ($isHtml) {
   $plugin['body'] = rcmail_replace_emoticons($plugin['body']);
 
   // add a plain text version of the e-mail as an alternative part.
-  $h2t = new html2text($plugin['body'], false, true, 0);
-  $plainTextPart = rc_wordwrap($h2t->get_text(), $LINE_LENGTH, "\r\n");
+  $h2t = new html2text($plugin['body'], false, true, 0, $message_charset);
+  $plainTextPart = rc_wordwrap($h2t->get_text(), $LINE_LENGTH, "\r\n", false, $message_charset);
   $plainTextPart = wordwrap($plainTextPart, 998, "\r\n", true);
 
   // make sure all line endings are CRLF (#1486712)
@@ -586,9 +586,9 @@ else {
 
   // compose format=flowed content if enabled
   if ($flowed = $RCMAIL->config->get('send_format_flowed', true))
-    $message_body = rcube_mime::format_flowed($message_body, min($LINE_LENGTH+2, 79));
+    $message_body = rcube_mime::format_flowed($message_body, min($LINE_LENGTH+2, 79), $message_charset);
   else
-    $message_body = rc_wordwrap($message_body, $LINE_LENGTH, "\r\n");
+    $message_body = rc_wordwrap($message_body, $LINE_LENGTH, "\r\n", false, $message_charset);
 
   $message_body = wordwrap($message_body, 998, "\r\n", true);
 
diff --git a/program/steps/utils/html2text.inc b/program/steps/utils/html2text.inc
index 67e76c1a7..e17665fec 100644
--- a/program/steps/utils/html2text.inc
+++ b/program/steps/utils/html2text.inc
@@ -5,7 +5,7 @@
  | program/steps/utils/html2text.inc                                     |
  |                                                                       |
  | This file is part of the Roundcube Webmail client                     |
- | Copyright (C) 2005-2010, The Roundcube Dev Team                       |
+ | Copyright (C) 2005-2012, The Roundcube Dev Team                       |
  |                                                                       |
  | Licensed under the GNU General Public License version 3 or            |
  | any later version with exceptions for skins & plugins.                |
@@ -24,7 +24,7 @@ $html = $HTTP_RAW_POST_DATA;
 // Replace emoticon images with its text representation
 $html = rcmail_replace_emoticons($html);
 
-$converter = new html2text($html);
+$converter = new html2text($html, false, true, 0);
 
 header('Content-Type: text/plain; charset=UTF-8');
 print rtrim($converter->get_text());