functions/mime.php

   1 <?php
   2    /** mime.php
   3     **
   4     ** This contains the functions necessary to detect and decode MIME
   5     ** messages.
   6     **
   7     ** $Id$
   8     **/
   9
  10    if (defined('mime_php'))
  11       return;
  12    define('mime_php', true);
  13
  14    require_once('../functions/imap.php');
  15
  16    /** Setting up the objects that have the structure for the message **/
  17
  18    class msg_header {
  19       /** msg_header contains generic variables for values that **/
  20       /** could be in a header.                                 **/
  21
  22       var $type0 = '', $type1 = '', $boundary = '', $charset = '';
  23       var $encoding = '', $size = 0, $to = array(), $from = '', $date = '';
  24       var $cc = array(), $bcc = array(), $reply_to = '', $subject = '';
  25       var $id = 0, $mailbox = '', $description = '', $filename = '';
  26       var $entity_id = 0, $message_id = 0, $name = '';
  27    }
  28
  29    class message {
  30       /** message is the object that contains messages.  It is a recursive
  31           object in that through the $entities variable, it can contain
  32           more objects of type message.  See documentation in mime.txt for
  33           a better description of how this works.
  34        **/
  35       var $header = '';
  36       var $entities = array();
  37
  38       function addEntity ($msg) {
  39          $this->entities[] = $msg;
  40       }
  41    }
  42
  43    /* --------------------------------------------------------------------------------- */
  44    /* MIME DECODING                                                                     */
  45    /* --------------------------------------------------------------------------------- */
  46
  47    // This function gets the structure of a message and stores it in the "message" class.
  48    // It will return this object for use with all relevant header information and
  49    // fully parsed into the standard "message" object format.
  50    function mime_structure ($imap_stream, $header) {
  51
  52       sqimap_messages_flag ($imap_stream, $header->id, $header->id, 'Seen');
  53       $ssid = sqimap_session_id();
  54       $lsid = strlen( $ssid );
  55       $id = $header->id;
  56       fputs ($imap_stream, "$ssid FETCH $id BODYSTRUCTURE\r\n");
  57       //
  58       // This should use sqimap_read_data instead of reading it itself
  59       //
  60       $read = fgets ($imap_stream, 10000);
  61       $bodystructure = '';
  62       while( substr($read, 0, $lsid) <> $ssid &&
  63              !feof( $imap_stream ) ) {
  64          $bodystructure .= $read;
  65          $read = fgets ($imap_stream, 10000);
  66       }
  67       $read = $bodystructure;
  68
  69       // isolate the body structure and remove beginning and end parenthesis
  70       $read = trim(substr ($read, strpos(strtolower($read), 'bodystructure') + 13));
  71       $read = trim(substr ($read, 0, -1));
  72       $end = mime_match_parenthesis(0, $read);
  73       while ($end == strlen($read)-1) {
  74          $read = trim(substr ($read, 0, -1));
  75          $read = trim(substr ($read, 1));
  76          $end = mime_match_parenthesis(0, $read);
  77       }
  78
  79       $msg = mime_parse_structure ($read, 0);
  80       $msg->header = $header;
  81       return $msg;
  82    }
  83
  84    // this starts the parsing of a particular structure.  It is called recursively,
  85    // so it can be passed different structures.  It returns an object of type
  86    // $message.
  87    // First, it checks to see if it is a multipart message.  If it is, then it
  88    // handles that as it sees is necessary.  If it is just a regular entity,
  89    // then it parses it and adds the necessary header information (by calling out
  90    // to mime_get_elements()
  91    function mime_parse_structure ($structure, $ent_id) {
  92
  93       $msg = new message();
  94       if ($structure{0} == '(') {
  95          $ent_id = mime_new_element_level($ent_id);
  96          $start = $end = -1;
  97          do {
  98             $start = $end+1;
  99             $end = mime_match_parenthesis ($start, $structure);
 100
 101             $element = substr($structure, $start+1, ($end - $start)-1);
 102             $ent_id = mime_increment_id ($ent_id);
 103             $newmsg = mime_parse_structure ($element, $ent_id);
 104             $msg->addEntity ($newmsg);
 105          } while ($structure{$end+1} == '(');
 106       } else {
 107          // parse the elements
 108          $msg = mime_get_element ($structure, $msg, $ent_id);
 109       }
 110       return $msg;
 111    }
 112
 113    // Increments the element ID.  An element id can look like any of
 114    // the following:  1, 1.2, 4.3.2.4.1, etc.  This function increments
 115    // the last number of the element id, changing 1.2 to 1.3.
 116    function mime_increment_id ($id) {
 117
 118       if (strpos($id, ".")) {
 119          $first = substr($id, 0, strrpos($id, "."));
 120          $last = substr($id, strrpos($id, ".")+1);
 121          $last++;
 122          $new = $first . "." .$last;
 123       } else {
 124          $new = $id + 1;
 125       }
 126
 127       return $new;
 128    }
 129
 130    // See comment for mime_increment_id().
 131    // This adds another level on to the entity_id changing 1.3 to 1.3.0
 132    // NOTE:  1.3.0 is not a valid element ID.  It MUST be incremented
 133    //        before it can be used.  I left it this way so as not to have
 134    //        to make a special case if it is the first entity_id.  It
 135    //        always increments it, and that works fine.
 136    function mime_new_element_level ($id) {
 137
 138       if (!$id) {
 139           $id = 0;
 140       } else {
 141           $id = $id . '.0';
 142       }
 143
 144       return( $id );
 145    }
 146
 147    function mime_get_element (&$structure, $msg, $ent_id) {
 148
 149       $elem_num = 1;
 150       $msg->header = new msg_header();
 151       $msg->header->entity_id = $ent_id;
 152       $properties = array();
 153
 154       while (strlen($structure) > 0) {
 155          $structure = trim($structure);
 156          $char = $structure{0};
 157
 158          if (strtolower(substr($structure, 0, 3)) == 'nil') {
 159             $text = '';
 160             $structure = substr($structure, 3);
 161          } else if ($char == '"') {
 162             // loop through until we find the matching quote, and return that as a string
 163             $pos = 1;
 164             $text = '';
 165             while ( ($char = $structure{$pos} ) <> '"' && $pos < strlen($structure)) {
 166                $text .= $char;
 167                $pos++;
 168             }
 169             $structure = substr($structure, strlen($text) + 2);
 170          } else if ($char == '(') {
 171             // comment me
 172             $end = mime_match_parenthesis (0, $structure);
 173             $sub = substr($structure, 1, $end-1);
 174             $properties = mime_get_props($properties, $sub);
 175             $structure = substr($structure, strlen($sub) + 2);
 176          } else {
 177             // loop through until we find a space or an end parenthesis
 178             $pos = 0;
 179             $char = $structure{$pos};
 180             $text = '';
 181             while ($char != ' ' && $char != ')' && $pos < strlen($structure)) {
 182                $text .= $char;
 183                $pos++;
 184                $char = $structure{$pos};
 185             }
 186             $structure = substr($structure, strlen($text));
 187          }
 188
 189          // This is where all the text parts get put into the header
 190          switch ($elem_num) {
 191             case 1:
 192                $msg->header->type0 = strtolower($text);
 193                break;
 194             case 2:
 195                $msg->header->type1 = strtolower($text);
 196                break;
 197             case 4: // Id
 198                // Invisimail enclose images with <>
 199                $msg->header->id = str_replace( '<', '', str_replace( '>', '', $text ) );
 200                break;
 201             case 5:
 202                $msg->header->description = $text;
 203                break;
 204             case 6:
 205                $msg->header->encoding = strtolower($text);
 206                break;
 207             case 7:
 208                $msg->header->size = $text;
 209                break;
 210             default:
 211                if ($msg->header->type0 == 'text' && $elem_num == 8) {
 212                   // This is a plain text message, so lets get the number of lines
 213                   // that it contains.
 214                   $msg->header->num_lines = $text;
 215
 216                } else if ($msg->header->type0 == 'message' && $msg->header->type1 == 'rfc822' && $elem_num == 8) {
 217                   // This is an encapsulated message, so lets start all over again and
 218                   // parse this message adding it on to the existing one.
 219                   $structure = trim($structure);
 220                   if ( $structure{0} == '(' ) {
 221                      $e = mime_match_parenthesis (0, $structure);
 222                      $structure = substr($structure, 0, $e);
 223                      $structure = substr($structure, 1);
 224                      $m = mime_parse_structure($structure, $msg->header->entity_id);
 225
 226                      // the following conditional is there to correct a bug that wasn't
 227                      // incrementing the entity IDs correctly because of the special case
 228                      // that message/rfc822 is.  This fixes it fine.
 229                      if (substr($structure, 1, 1) != '(')
 230                         $m->header->entity_id = mime_increment_id(mime_new_element_level($ent_id));
 231
 232                      // Now we'll go through and reformat the results.
 233                      if ($m->entities) {
 234                         for ($i=0; $i < count($m->entities); $i++) {
 235                            $msg->addEntity($m->entities[$i]);
 236                         }
 237                      } else {
 238                         $msg->addEntity($m);
 239                      }
 240                      $structure = "";
 241                   }
 242                }
 243                break;
 244          }
 245          $elem_num++;
 246          $text = "";
 247       }
 248       // loop through the additional properties and put those in the various headers
 249       if ($msg->header->type0 != 'message') {
 250          for ($i=0; $i < count($properties); $i++) {
 251             $msg->header->{$properties[$i]['name']} = $properties[$i]['value'];
 252          }
 253       }
 254
 255       return $msg;
 256    }
 257
 258    // I did most of the MIME stuff yesterday (June 20, 2000), but I couldn't
 259    // figure out how to do this part, so I decided to go to bed.  I woke up
 260    // in the morning and had a flash of insight.  I went to the white-board
 261    // and scribbled it out, then spent a bit programming it, and this is the
 262    // result.  Nothing complicated, but I think my brain was fried yesterday.
 263    // Funny how that happens some times.
 264    //
 265    // This gets properties in a nested parenthesisized list.  For example,
 266    // this would get passed something like:  ("attachment" ("filename" "luke.tar.gz"))
 267    // This returns an array called $props with all paired up properties.
 268    // It ignores the "attachment" for now, maybe that should change later
 269    // down the road.  In this case, what is returned is:
 270    //    $props[0]["name"] = "filename";
 271    //    $props[0]["value"] = "luke.tar.gz";
 272    function mime_get_props ($props, $structure) {
 273
 274       while (strlen($structure) > 0) {
 275          $structure = trim($structure);
 276          $char = $structure{0};
 277
 278          if ($char == '"') {
 279             $pos = 1;
 280             $tmp = '';
 281             while ( ( $char = $structure{$pos} ) != '"' &&
 282                     $pos < strlen($structure)) {
 283                $tmp .= $char;
 284                $pos++;
 285             }
 286             $structure = trim(substr($structure, strlen($tmp) + 2));
 287             $char = $structure{0};
 288
 289             if ($char == '"') {
 290                $pos = 1;
 291                $value = '';
 292                while ( ( $char = $structure{$pos} ) != '"' &&
 293                        $pos < strlen($structure) ) {
 294                   $value .= $char;
 295                   $pos++;
 296                }
 297                $structure = trim(substr($structure, strlen($tmp) + 2));
 298
 299                $k = count($props);
 300                $props[$k]['name'] = strtolower($tmp);
 301                $props[$k]['value'] = $value;
 302             } else if ($char == '(') {
 303                $end = mime_match_parenthesis (0, $structure);
 304                $sub = substr($structure, 1, $end-1);
 305            if (! isset($props))
 306                $props = array();
 307                $props = mime_get_props($props, $sub);
 308                $structure = substr($structure, strlen($sub) + 2);
 309             }
 310             return $props;
 311          } else if ($char == '(') {
 312             $end = mime_match_parenthesis (0, $structure);
 313             $sub = substr($structure, 1, $end-1);
 314             $props = mime_get_props($props, $sub);
 315             $structure = substr($structure, strlen($sub) + 2);
 316             return $props;
 317          } else {
 318             return $props;
 319          }
 320       }
 321    }
 322
 323    //  Matches parenthesis.  It will return the position of the matching
 324    //  parenthesis in $structure.  For instance, if $structure was:
 325    //     ("text" "plain" ("val1name", "1") nil ... )
 326    //     x                                         x
 327    //  then this would return 42 to match up those two.
 328    function mime_match_parenthesis ($pos, $structure) {
 329
 330       $j = strlen( $structure );
 331
 332       // ignore all extra characters
 333       // If inside of a string, skip string -- Boundary IDs and other
 334       // things can have ) in them.
 335       if( $structure{$pos} != '(' )
 336          return( $j );
 337
 338       while( $pos < $j ) {
 339          $pos++;
 340          if ($structure{$pos} == ')') {
 341             return $pos;
 342          } elseif ($structure{$pos} == '"') {
 343             $pos++;
 344             while( $structure{$pos} != '"' &&
 345                $pos < $j ) {
 346                if (substr($structure, $pos, 2) == '\\"')
 347                   $pos++;
 348                elseif (substr($structure, $pos, 2) == '\\\\')
 349                   $pos++;
 350                $pos++;
 351             }
 352          } elseif ( $structure{$pos} == '(' ) {
 353             $pos = mime_match_parenthesis ($pos, $structure);
 354          }
 355       }
 356       echo "Error decoding mime structure.  Report this as a bug!<br>\n";
 357       return( $pos );
 358    }
 359
 360     function mime_fetch_body ($imap_stream, $id, $ent_id ) {
 361         // do a bit of error correction.  If we couldn't find the entity id, just guess
 362         // that it is the first one.  That is usually the case anyway.
 363         if (!$ent_id)
 364             $ent_id = 1;
 365         $sid = sqimap_session_id();
 366         fputs ($imap_stream, "$sid FETCH $id BODY[$ent_id]\r\n");
 367         $data = sqimap_read_data ($imap_stream, $sid, true, $response, $message);
 368         $topline = array_shift($data);
 369         while (! ereg('\\* [0-9]+ FETCH ', $topline) && $data)
 370             $topline = array_shift($data);
 371         $wholemessage = implode('', $data);
 372         if (ereg('\\{([^\\}]*)\\}', $topline, $regs)) {
 373             $ret = substr( $wholemessage, 0, $regs[1] );
 374             /*
 375                 There is some information in the content info header that could be important
 376                 in order to parse html messages. Let's get them here.
 377             */
 378             if( $ret{0} == '<' ) {
 379                 fputs ($imap_stream, "$sid FETCH $id BODY[$ent_id.MIME]\r\n");
 380                 $data = sqimap_read_data ($imap_stream, $sid, true, $response, $message);
 381                 $base = '';
 382                 $k = 10;
 383                 foreach( $data as $d ) {
 384                     if( substr( $d, 0, 13 ) == 'Content-Base:' ) {
 385                         $j = strlen( $d );
 386                         $i = 13;
 387                         $base = '';
 388                         while( $i < $j &&
 389                                ( !isNoSep( $d{$i} ) || $d{$i} == '"' )  )
 390                             $i++;
 391                         while( $i < $j ) {
 392                             if( isNoSep( $d{$i} ) )
 393                                 $base .= $d{$i};
 394                             $i++;
 395                         }
 396                         $k = 0;
 397                     } elseif( $k == 1 && !isnosep( $d{0} ) ) {
 398                         $base .= substr( $d, 1 );
 399                     }
 400                     $k++;
 401                 }
 402                 if( $base <> '' )
 403                     $ret = "<base href=\"$base\">" . $ret;
 404             }
 405         } else if (ereg('"([^"]*)"', $topline, $regs)) {
 406             $ret = $regs[1];
 407         } else {
 408             global $where, $what, $mailbox, $passed_id, $startMessage;
 409             $par = "mailbox=".urlencode($mailbox)."&passed_id=$passed_id";
 410             if (isset($where) && isset($what)) {
 411                 $par .= "&where=".urlencode($where)."&what=".urlencode($what);
 412             } else {
 413                 $par .= "&startMessage=$startMessage&show_more=0";
 414             }
 415             $par .= '&response='.urlencode($response).'&message='.urlencode($message).
 416                     '&topline='.urlencode($topline);
 417
 418             echo   '<b><font color=$color[2]>Body retrieval error. The reason for this is most probably that<BR> ' .
 419                    'the message is malformed. Please help us making future versions<BR> ' .
 420                    "better by submitting this message to the developers knowledgebase!<BR>\n" .
 421                    "<A HREF=\"../src/retrievalerror.php?$par\">Submit message</A><BR>" .
 422
 423                    "<tt>Response:  $response<BR>" .
 424                    "Message:  $message<BR>" .
 425                    "FETCH line:  $topline<BR></tt></font></b>";
 426
 427             fputs ($imap_stream, "$sid FETCH $passed_id BODY[]\r\n");
 428             $data = sqimap_read_data ($imap_stream, $sid, true, $response, $message);
 429             array_shift($data);
 430             $wholemessage = implode('', $data);
 431
 432             $ret = "---------------\n$wholemessage";
 433
 434         }
 435         return( $ret );
 436     }
 437
 438    function mime_print_body_lines ($imap_stream, $id, $ent_id, $encoding) {
 439       // do a bit of error correction.  If we couldn't find the entity id, just guess
 440       // that it is the first one.  That is usually the case anyway.
 441       if (!$ent_id) $ent_id = 1;
 442       $sid = sqimap_session_id();
 443       // Don't kill the connection if the browser is over a dialup
 444       // and it would take over 30 seconds to download it.
 445       set_time_limit(0);
 446
 447       fputs ($imap_stream, "$sid FETCH $id BODY[$ent_id]\r\n");
 448       $cnt = 0;
 449       $continue = true;
 450       $read = fgets ($imap_stream,4096);
 451       // This could be bad -- if the section has sqimap_session_id() . ' OK'
 452       // or similar, it will kill the download.
 453       while (!ereg("^".$sid." (OK|BAD|NO)(.*)$", $read, $regs)) {
 454           if (trim($read) == ')==') {
 455               $read1 = $read;
 456               $read = fgets ($imap_stream,4096);
 457               if (ereg("^".$sid." (OK|BAD|NO)(.*)$", $read, $regs)) {
 458                   return;
 459               } else {
 460                   echo decodeBody($read1, $encoding) .
 461                        decodeBody($read, $encoding);
 462               }
 463           } else if ($cnt) {
 464               echo decodeBody($read, $encoding);
 465           }
 466           $read = fgets ($imap_stream,4096);
 467           $cnt++;
 468       }
 469    }
 470
 471    /* -[ END MIME DECODING ]----------------------------------------------------------- */
 472
 473
 474
 475    /** This is the first function called.  It decides if this is a multipart
 476        message or if it should be handled as a single entity
 477     **/
 478    function decodeMime ($imap_stream, &$header) {
 479       global $username, $key, $imapServerAddress, $imapPort;
 480       return mime_structure ($imap_stream, $header);
 481    }
 482
 483    // This is here for debugging purposese.  It will print out a list
 484    // of all the entity IDs that are in the $message object.
 485    /*
 486    function listEntities ($message) {
 487       if ($message) {
 488          if ($message->header->entity_id)
 489          echo "<tt>" . $message->header->entity_id . ' : ' . $message->header->type0 . '/' . $message->header->type1 . '<br>';
 490          for ($i = 0; $message->entities[$i]; $i++) {
 491             $msg = listEntities($message->entities[$i], $ent_id);
 492             if ($msg)
 493                return $msg;
 494          }
 495       }
 496    }
 497    */
 498
 499    // returns a $message object for a particular entity id
 500    function getEntity ($message, $ent_id) {
 501       if ($message) {
 502          if ($message->header->entity_id == $ent_id && strlen($ent_id) == strlen($message->header->entity_id)) {
 503             return $message;
 504          } else {
 505             for ($i = 0; isset($message->entities[$i]); $i++) {
 506                $msg = getEntity ($message->entities[$i], $ent_id);
 507                if ($msg)
 508                   return $msg;
 509             }
 510          }
 511       }
 512    }
 513
 514     // figures out what entity to display and returns the $message object
 515     // for that entity.
 516     function findDisplayEntity ($message, $textOnly = 1)   {
 517         global $show_html_default;
 518
 519         $entity = 0;
 520
 521         if ($message) {
 522             if ( $message->header->type0 == 'multipart' &&
 523                  ( $message->header->type1 == 'alternative' ||
 524                    $message->header->type1 == 'related' ) &&
 525                  $show_html_default && ! $textOnly ) {
 526                 $entity = findDisplayEntityHTML($message);
 527             }
 528
 529             // Show text/plain or text/html -- the first one we find.
 530             if ( $entity == 0 &&
 531                  $message->header->type0 == 'text' &&
 532                  ( $message->header->type1 == 'plain' ||
 533                    $message->header->type1 == 'html' ) &&
 534                  isset($message->header->entity_id) ) {
 535                 $entity = $message->header->entity_id;
 536             }
 537
 538             $i = 0;
 539             while ($entity == 0 && isset($message->entities[$i]) ) {
 540                 $entity = findDisplayEntity($message->entities[$i], $textOnly);
 541                 $i++;
 542             }
 543         }
 544
 545         return( $entity );
 546     }
 547
 548    // Shows the HTML version
 549    function findDisplayEntityHTML ($message) {
 550       if ($message->header->type0 == 'text' &&
 551           $message->header->type1 == 'html' &&
 552       isset($message->header->entity_id))
 553      return $message->header->entity_id;
 554       for ($i = 0; isset($message->entities[$i]); $i ++) {
 555          $entity = findDisplayEntityHTML($message->entities[$i]);
 556      if ($entity != 0)
 557         return $entity;
 558       }
 559       return 0;
 560    }
 561
 562    /** This returns a parsed string called $body. That string can then
 563        be displayed as the actual message in the HTML. It contains
 564        everything needed, including HTML Tags, Attachments at the
 565        bottom, etc.
 566     **/
 567    function formatBody($imap_stream, $message, $color, $wrap_at) {
 568       // this if statement checks for the entity to show as the
 569       // primary message. To add more of them, just put them in the
 570       // order that is their priority.
 571       global $startMessage, $username, $key, $imapServerAddress, $imapPort,
 572           $show_html_default;
 573
 574       $id = $message->header->id;
 575       $urlmailbox = urlencode($message->header->mailbox);
 576
 577       // Get the right entity and redefine message to be this entity
 578       // Pass the 0 to mean that we want the 'best' viewable one
 579       $ent_num = findDisplayEntity ($message, 0);
 580       $body_message = getEntity($message, $ent_num);
 581       if (($body_message->header->type0 == 'text') ||
 582           ($body_message->header->type0 == 'rfc822')) {
 583
 584          $body = mime_fetch_body ($imap_stream, $id, $ent_num);
 585          $body = decodeBody($body, $body_message->header->encoding);
 586          $hookResults = do_hook("message_body", $body);
 587          $body = $hookResults[1];
 588
 589          // If there are other types that shouldn't be formatted, add
 590          // them here
 591          if ($body_message->header->type1 == 'html') {
 592             if( $show_html_default <> 1 ) {
 593                 $body = strip_tags( $body );
 594                 translateText($body, $wrap_at, $body_message->header->charset);
 595             } else {
 596                 $body = MagicHTML( $body, $id );
 597             }
 598          } else {
 599             translateText($body, $wrap_at, $body_message->header->charset);
 600          }
 601
 602          $body .= "<SMALL><CENTER><A HREF=\"../src/download.php?absolute_dl=true&passed_id=$id&passed_ent_id=$ent_num&mailbox=$urlmailbox&showHeaders=1\">". _("Download this as a file") ."</A></CENTER><BR></SMALL>";
 603
 604          /** Display the ATTACHMENTS: message if there's more than one part **/
 605          $body .= "</TD></TR></TABLE>";
 606          if (isset($message->entities[0])) {
 607             $body .= formatAttachments ($message, $ent_num, $message->header->mailbox, $id);
 608          }
 609          $body .= "</TD></TR></TABLE>";
 610       } else {
 611          $body = formatAttachments ($message, -1, $message->header->mailbox, $id);
 612       }
 613       return( $body );
 614    }
 615
 616    // A recursive function that returns a list of attachments with links
 617    // to where to download these attachments
 618    function formatAttachments ($message, $ent_id, $mailbox, $id) {
 619       global $where, $what;
 620       global $startMessage, $color;
 621       static $ShownHTML = 0;
 622
 623       $body = "";
 624       if ($ShownHTML == 0) {
 625             $ShownHTML = 1;
 626
 627             $body .= "<TABLE WIDTH=100% CELLSPACING=0 CELLPADDING=2 BORDER=0 BGCOLOR=\"$color[0]\"><TR>\n" .
 628                      "<TH ALIGN=\"left\" BGCOLOR=\"$color[9]\"><B>\n" .
 629                      _("Attachments") . ':' .
 630                      "</B></TH></TR><TR><TD>\n" .
 631                      "<TABLE CELLSPACING=0 CELLPADDING=1 BORDER=0>\n" .
 632                      formatAttachments ($message, $ent_id, $mailbox, $id) .
 633                      "</TABLE></TD></TR></TABLE>";
 634
 635             return( $body );
 636       }
 637
 638       if ($message) {
 639          if (!$message->entities) {
 640             $type0 = strtolower($message->header->type0);
 641             $type1 = strtolower($message->header->type1);
 642             $name = decodeHeader($message->header->name);
 643
 644             if ($message->header->entity_id != $ent_id) {
 645                $filename = decodeHeader($message->header->filename);
 646                if (trim($filename) == '') {
 647                   if (trim($name) == '') {
 648                      if( trim( $message->header->id ) == '' )
 649                         $display_filename = 'untitled-[' . $message->header->entity_id . ']' ;
 650                      else
 651                         $display_filename = 'cid: ' . $message->header->id;
 652                      // $display_filename = 'untitled-[' . $message->header->entity_id . ']' ;
 653                   } else {
 654                      $display_filename = $name;
 655                      $filename = $name;
 656                   }
 657                } else {
 658                   $display_filename = $filename;
 659                }
 660
 661                $urlMailbox = urlencode($mailbox);
 662                $ent = urlencode($message->header->entity_id);
 663
 664                $DefaultLink =
 665                   "../src/download.php?startMessage=$startMessage&passed_id=$id&mailbox=$urlMailbox&passed_ent_id=$ent";
 666                if ($where && $what)
 667                   $DefaultLink .= '&where=' . urlencode($where) . '&what=' . urlencode($what);
 668                $Links['download link']['text'] = _("download");
 669                $Links['download link']['href'] =
 670                    "../src/download.php?absolute_dl=true&passed_id=$id&mailbox=$urlMailbox&passed_ent_id=$ent";
 671                $ImageURL = '';
 672
 673                $HookResults = do_hook("attachment $type0/$type1", $Links,
 674                    $startMessage, $id, $urlMailbox, $ent, $DefaultLink,
 675                    $display_filename, $where, $what);
 676
 677                $Links = $HookResults[1];
 678                $DefaultLink = $HookResults[6];
 679
 680                $body .= '<TR><TD>&nbsp;&nbsp;</TD><TD>' .
 681                         "<A HREF=\"$DefaultLink\">$display_filename</A>&nbsp;</TD>" .
 682                         '<TD><SMALL><b>' . show_readable_size($message->header->size) .
 683                         '</b>&nbsp;&nbsp;</small></TD>' .
 684                         "<TD><SMALL>[ $type0/$type1 ]&nbsp;</SMALL></TD>" .
 685                         '<TD><SMALL>';
 686                if ($message->header->description)
 687                   $body .= '<b>' . htmlspecialchars($message->header->description) . '</b>';
 688                $body .= '</SMALL></TD><TD><SMALL>&nbsp;';
 689
 690
 691                $SkipSpaces = 1;
 692                foreach ($Links as $Val) {
 693                   if ($SkipSpaces) {
 694                      $SkipSpaces = 0;
 695                   } else {
 696                      $body .= '&nbsp;&nbsp;|&nbsp;&nbsp;';
 697                   }
 698                   $body .= '<a href="' . $Val['href'] . '">' .  $Val['text'] . '</a>';
 699                }
 700
 701                unset($Links);
 702
 703                $body .= "</SMALL></TD></TR>\n";
 704             }
 705          } else {
 706             for ($i = 0; $i < count($message->entities); $i++) {
 707                $body .= formatAttachments ($message->entities[$i], $ent_id, $mailbox, $id);
 708             }
 709          }
 710          return( $body );
 711       }
 712    }
 713
 714
 715    /** this function decodes the body depending on the encoding type. **/
 716    function decodeBody($body, $encoding) {
 717       $body = str_replace("\r\n", "\n", $body);
 718       $encoding = strtolower($encoding);
 719
 720       global $show_html_default;
 721
 722       if ($encoding == 'quoted-printable') {
 723          $body = quoted_printable_decode($body);
 724
 725
 726          /*
 727             Following code has been comented as I see no reason for it.
 728             If there is any please tell me a mingo@rotedic.com
 729
 730          while (ereg("=\n", $body))
 731             $body = ereg_replace ("=\n", "", $body);
 732         */
 733       } else if ($encoding == 'base64') {
 734          $body = base64_decode($body);
 735       }
 736
 737       // All other encodings are returned raw.
 738       return $body;
 739    }
 740
 741
 742    // This functions decode strings that is encoded according to
 743    // RFC1522 (MIME Part Two: Message Header Extensions for Non-ASCII Text).
 744    function decodeHeader ($string) {
 745       if (eregi('=\\?([^?]+)\\?(q|b)\\?([^?]+)\\?=',
 746                 $string, $res)) {
 747          if (ucfirst($res[2]) == "B") {
 748             $replace = base64_decode($res[3]);
 749          } else {
 750             $replace = ereg_replace("_", " ", $res[3]);
 751         // Convert lowercase Quoted Printable to uppercase for
 752         // quoted_printable_decode to understand it.
 753         while (ereg("(=(([0-9][abcdef])|([abcdef][0-9])|([abcdef][abcdef])))", $replace, $res)) {
 754            $replace = str_replace($res[1], strtoupper($res[1]), $replace);
 755         }
 756             $replace = quoted_printable_decode($replace);
 757          }
 758
 759          $replace = charset_decode ($res[1], $replace);
 760
 761          // Remove the name of the character set.
 762          $string = eregi_replace ('=\\?([^?]+)\\?(q|b)\\?([^?]+)\\?=',
 763              $replace, $string);
 764
 765          // In case there should be more encoding in the string: recurse
 766          return (decodeHeader($string));
 767       } else
 768          return ($string);
 769    }
 770
 771    // Encode a string according to RFC 1522 for use in headers if it
 772    // contains 8-bit characters or anything that looks like it should
 773    // be encoded.
 774    function encodeHeader ($string) {
 775       global $default_charset;
 776
 777      // Encode only if the string contains 8-bit characters or =?
 778      $j = strlen( $string  );
 779      $l = FALSE;                             // Must be encoded ?
 780      $ret = '';
 781      for( $i=0; $i < $j; ++$i) {
 782         switch( $string{$i} ) {
 783            case '=':
 784           $ret .= '=3D';
 785           break;
 786        case '?':
 787           $l = TRUE;
 788           $ret .= '=3F';
 789           break;
 790        case '_':
 791           $ret .= '=5F';
 792           break;
 793        case ' ':
 794           $ret .= '_';
 795           break;
 796       default:
 797           $k = ord( $string{$i} );
 798           if( $k > 126 ) {
 799              $ret .= sprintf("=%02X", $k);
 800              $l = TRUE;
 801           } else
 802              $ret .= $string{$i};
 803         }
 804      }
 805
 806      if( $l )
 807         $string = "=?$default_charset?Q?$ret?=";
 808
 809      return( $string );
 810  }
 811
 812    /*
 813     Strips dangerous tags from html messages.
 814    */
 815
 816    function MagicHTML( $body, $id ) {
 817
 818         global $message, $PHP_SELF, $HTTP_SERVER_VARS;
 819
 820         $j = strlen( $body );   // Legnth of the HTML
 821         $ret = '';              // Returned string
 822         $bgcolor = '#ffffff';   // Background style color (defaults to white)
 823         $leftmargin = '';       // Left margin style
 824         $title = '';            // HTML title if any
 825
 826         $i = 0;
 827         while( $i < $j ) {
 828             if( $body{$i} == '<' ) {
 829                 $tag = $body{$i+1}.$body{$i+2}.$body{$i+3}.$body{$i+4};
 830                 switch( strtoupper( $tag ) ) {
 831                     // Strips the entire tag and contents
 832                     case 'APPL':
 833                     case 'EMBB':
 834                     case 'FRAM':
 835                     case 'SCRI':
 836                     case 'OBJE':
 837                         $etg = '/' . $tag;
 838                         while( $body{$i+1}.$body{$i+2}.$body{$i+3}.$body{$i+4}.$body{$i+5} <> $etg  &&
 839                                $i < $j  ) $i++;
 840                         while( $i < $j && $body{++$i} <> '>' );
 841                         // $ret .= "<!-- $tag removed -->";
 842                         break;
 843                     // Substitute Title
 844                     case 'TITL':
 845                         $i += 5;
 846                         while( $body{$i} <> '>' &&  // </title>
 847                                $i < $j )
 848                                 $i++;
 849                         $i++;
 850                         $title = '';
 851                         while( $body{$i} <> '<' &&  // </title>
 852                                $i < $j ) {
 853                             $title .= $body{$i};
 854                             $i++;
 855                         }
 856                         $i += 7;
 857                         break;
 858                     // Destroy these tags
 859                     case 'HTML':
 860                     case 'HEAD':
 861                     case '/HTM':
 862                     case '/HEA':
 863                     case '!DOC':
 864                     case 'META':
 865                     case 'DIV ':
 866                     case '/DIV':
 867                     case '!-- ':
 868                         $i += 4;
 869                         while( $body{$i}  <> '>' &&
 870                                $i < $j )
 871                             $i++;
 872                         // $i++;
 873                         break;
 874                     case 'STYL':
 875                         $i += 5;
 876                         while( $body{$i} <> '>' &&  // </title>
 877                                $i < $j )
 878                                 $i++;
 879                         $i++;
 880                         // We parse the style to look for interesting stuff
 881                         $styleblk = '';
 882                         while( $body{$i} <> '>' &&
 883                                $i < $j ) {
 884                             // First we get the name of the style
 885                             $style = '';
 886                             while( $body{$i} <> '>' &&
 887                                    $body{$i} <> '<' &&
 888                                    $body{$i} <> '{' &&
 889                                    $i < $j ) {
 890                                if( isnoSep( $body{$i} ) )
 891                                    $style .= $body{$i};
 892                                $i++;
 893                             }
 894                             stripComments( $i, $j, $body );
 895                             $style = strtoupper( trim( $style ) );
 896                             if( $style == 'BODY' ) {
 897                                 // Next we look into the definitions of the body style
 898                                 while( $body{$i} <> '>' &&
 899                                        $body{$i} <> '}' &&
 900                                        $i < $j ) {
 901                                     // We look for the background color if any.
 902                                     if( substr( $body, $i, 17 ) == 'BACKGROUND-COLOR:' ) {
 903                                         $i += 17;
 904                                         $bgcolor = getStyleData( $i, $j, $body );
 905                                     } elseif ( substr( $body, $i, 12 ) == 'MARGIN-LEFT:' ) {
 906                                         $i += 12;
 907                                         $leftmargin = getStyleData( $i, $j, $body );
 908                                     }
 909                                     $i++;
 910                                 }
 911                             } else {
 912                                 // Other style are mantained
 913                                 $styleblk .= "$style ";
 914                                 while( $body{$i} <> '>' &&
 915                                        $body{$i} <> '<' &&
 916                                        $body{$i} <> '}' &&
 917                                        $i < $j ) {
 918                                     $styleblk .= $body{$i};
 919                                     $i++;
 920                                 }
 921                                 $styleblk .= $body{$i};
 922                             }
 923                             stripComments( $i, $j, $body );
 924                             if( $body{$i} <> '>' )
 925                                 $i++;
 926                         }
 927                         if( $styleblk <> '' )
 928                             $ret .= "<style>$styleblk";
 929                         break;
 930                     case 'BODY':
 931                         if( $title <> '' )
 932                             $ret .= '<b>' . _("Title:") . " </b>$title<br>\n";
 933                         $ret .= "<TABLE";
 934                         $i += 5;
 935                         if (! isset($base))
 936                            $base = '';
 937                         $ret .= stripEvent( $i, $j, $body, $id, $base );
 938                         //if( $bgcolor <> '' )
 939                             $ret .= " bgcolor=$bgcolor";
 940                         $ret .= ' width=100%><tr>';
 941                         if( $leftmargin <> '' )
 942                             $ret .= "<td width=$leftmargin>&nbsp;</td>";
 943                         $ret .= '<td>';
 944                         break;
 945                     case 'BASE':
 946                         $i += 5;
 947                         $base = '';
 948                         while( !isNoSep( $body{$i} ) &&
 949                                $i < $j )
 950                                 $i++;
 951                         if( strcasecmp( substr( $base, 0, 4 ), 'href'  ) ) {
 952                                 $i += 5;
 953                                 while( !isNoSep( $body{$i} ) &&
 954                                        $i < $j )
 955                                         $i++;
 956                                 while( $body{$i} <> '>' &&
 957                                        $i < $j ) {
 958                                     if( $body{$i} <> '"' )
 959                                         $base .= $body{$i};
 960                                         $i++;
 961                                 }
 962                                 // Debuging $ret .= "<!-- base == $base -->";
 963                                 if( strcasecmp( substr( $base, 0, 4 ), 'file' ) <> 0 )
 964                                         $ret .= "\n<BASE HREF=\"$base\">\n";
 965                         }
 966                         break;
 967                     case '/BOD':
 968                         $ret .= '</td></tr></TABLE>';
 969                         $i += 6;
 970                         break;
 971                     default:
 972                         // Following tags can contain some event handler, lets search it
 973                         stripComments( $i, $j, $body );
 974                         if (! isset($base))
 975                            $base = '';
 976                         $ret .= stripEvent( $i, $j, $body, $id, $base ) . '>';
 977                         // $ret .= "<!-- $tag detected -->";
 978                 }
 979             } else {
 980                 $ret .= $body{$i};
 981             }
 982             $i++;
 983         }
 984
 985         return( "\n\n<!-- HTML Output ahead -->\n" .
 986                 $ret .
 987                 "\n<!-- END of HTML Output --><base href=\"".
 988                 $HTTP_SERVER_VARS["SERVER_NAME"] . substr( $PHP_SELF, 0, strlen( $PHP_SELF ) - 13 ) .
 989                 "\">\n\n" );
 990    }
 991
 992    function isNoSep( $char ) {
 993
 994         switch( $char ) {
 995             case ' ':
 996             case "\n":
 997             case "\t":
 998             case "\r":
 999             case '>':
1000             case '"':
1001                 return( FALSE );
1002                 break;
1003             default:
1004                 return( TRUE );
1005         }
1006
1007    }
1008
1009    /*
1010       The following function is usefull to remove extra data that can cause
1011       html not to display properly. Especialy with MS stuff.
1012    */
1013
1014    function stripComments( &$i, $j, &$body ) {
1015
1016         while( $body{$i}.$body{$i+1}.$body{$i+2}.$body{$i+3} == '<!--' &&
1017                $i < $j ) {
1018             $i += 5;
1019             while( $body{$i-2}.$body{$i-1}.$body{$i} <> '-->' &&
1020                    $i < $j )
1021                 $i++;
1022             $i++;
1023         }
1024
1025         return;
1026
1027    }
1028
1029    /* Gets the style data of a specific style */
1030
1031    function getStyleData( &$i, $j, &$body ) {
1032
1033         // We skip spaces
1034         while( $body{$i} <> '>' && !isNoSep( $body{$i} ) &&
1035                $i < $j ) {
1036             $i++;
1037         }
1038         // And get the color
1039         $ret = '';
1040         while( isNoSep( $body{$i} ) &&
1041                $i < $j ) {
1042             $ret .= $body{$i};
1043             $i++;
1044         }
1045
1046         return( $ret );
1047    }
1048
1049    /*
1050    Private function for strip_dangerous_tag. Look for event based coded and "remove" it
1051    change on with no (onload -> noload)
1052    */
1053
1054    function stripEvent( &$i, $j, &$body, $id, $base ) {
1055
1056         global $message;
1057
1058         $ret = '';
1059
1060         while( $body{$i} <> '>' &&
1061                $i < $j ) {
1062             $etg = strtolower($body{$i}.$body{$i+1}.$body{$i+2});
1063             switch( $etg ) {
1064                 case '../':
1065                         // Retrolinks are not allowed without a base because they mess with SM security
1066                         if( $base == '' ) {
1067                                 $i += 2;
1068                         } else {
1069                                 $ret .= '.';
1070                         }
1071                         break;
1072                 case 'cid':
1073                     // Internal link
1074                     $k = $i-1;
1075                     if( $body{$i+3} == ':') {
1076                         $i +=4;
1077                         $name = '';
1078                         while( isNoSep( $body{$i} ) &&
1079                                $i < $j  )
1080                             $name .= $body{$i++};
1081                         if( $name <> '' ) {
1082                             $ret .= "../src/download.php?absolute_dl=true&passed_id=$id&mailbox=" .
1083                                         urlencode( $message->header->mailbox ) .
1084                                         "&passed_ent_id=" . find_ent_id( $name, $message );
1085                             if( $body{$k} == '"' )
1086                                 $ret .= '" ';
1087                             else
1088                                 $ret .= ' ';
1089                         }
1090                         if( $body{$i} == '>' )
1091                             $i -= 1;
1092                     }
1093                     break;
1094                 case ' on':
1095                 case "\non":
1096                 case "\ron":
1097                 case "\ton":
1098                     $ret .= ' no';
1099                     $i += 2;
1100                     break;
1101                 case 'pt:':
1102                     if( strcasecmp( $body{$i-4}.$body{$i-3}.$body{$i-2}.$body{$i-1}.$body{$i}.$body{$i+1}.$body{$i+2}, 'script:') == 0 ) {
1103                         $ret .= '_no/';
1104                     } else {
1105                         $ret .= $etg;
1106                     }
1107                     $i += 2;
1108                     break;
1109                 default:
1110                     $ret .= $body{$i};
1111             }
1112             $i++;
1113         }
1114         return( $ret );
1115     }
1116
1117
1118     /* This function trys to locate the entity_id of a specific mime element */
1119
1120     function find_ent_id( $id, $message ) {
1121
1122         $ret = '';
1123         for ($i=0; $ret == '' && $i < count($message->entities); $i++) {
1124
1125             if( $message->entities[$i]->header->entity_id == '' ) {
1126                 $ret = find_ent_id( $id, $message->entities[$i] );
1127             } else {
1128                 if( strcasecmp( $message->entities[$i]->header->id, $id ) == 0 )
1129                     $ret = $message->entities[$i]->header->entity_id;
1130             }
1131
1132         }
1133
1134         return( $ret );
1135
1136     }
1137 ?>