functions/mime.php

   1 <?php
   2
   3 /**
   4  * mime.php
   5  *
   6  * Copyright (c) 1999-2002 The SquirrelMail Project Team
   7  * Licensed under the GNU GPL. For full terms see the file COPYING.
   8  *
   9  * This contains the functions necessary to detect and decode MIME
  10  * messages.
  11  *
  12  * $Id$
  13  */
  14
  15 require_once('../functions/imap.php');
  16 require_once('../functions/attachment_common.php');
  17
  18 /** Setting up the objects that have the structure for the message **/
  19 class msg_header {
  20     /** msg_header contains generic variables for values that **/
  21     /** could be in a header.                                 **/
  22
  23     var $type0 = '', $type1 = '', $boundary = '', $charset = '',
  24         $encoding = '', $size = 0, $to = array(), $from = '', $date = '',
  25         $cc = array(), $bcc = array(), $reply_to = '', $subject = '',
  26         $id = 0, $mailbox = '', $description = '', $filename = '',
  27         $entity_id = 0, $message_id = 0, $name = '', $priority = 3;
  28 }
  29
  30 class message {
  31     /** message is the object that contains messages.  It is a recursive
  32       object in that through the $entities variable, it can contain
  33       more objects of type message.  See documentation in mime.txt for
  34       a better description of how this works.
  35     **/
  36     var $header = '';
  37     var $entities = array();
  38
  39     function addEntity ($msg) {
  40         $this->entities[] = $msg;
  41     }
  42 }
  43
  44 /* --------------------------------------------------------------------------------- */
  45 /* MIME DECODING                                                                     */
  46 /* --------------------------------------------------------------------------------- */
  47
  48 /* This function gets the structure of a message and stores it in the "message" class.
  49  * It will return this object for use with all relevant header information and
  50  * fully parsed into the standard "message" object format.
  51  */
  52 function mime_structure ($imap_stream, $header) {
  53
  54     sqimap_messages_flag ($imap_stream, $header->id, $header->id, 'Seen');
  55     $ssid = sqimap_session_id();
  56     $lsid = strlen( $ssid );
  57     $id = $header->id;
  58     fputs ($imap_stream, "$ssid FETCH $id BODYSTRUCTURE\r\n");
  59     //
  60     // This should use sqimap_read_data instead of reading it itself
  61     //
  62     $read = fgets ($imap_stream, 10000);
  63     $bodystructure = '';
  64     while ( substr($read, 0, $lsid) <> $ssid &&
  65          !feof( $imap_stream ) ) {
  66         $bodystructure .= $read;
  67         $read = fgets ($imap_stream, 10000);
  68     }
  69     $read = $bodystructure;
  70
  71     // isolate the body structure and remove beginning and end parenthesis
  72     $read = trim(substr ($read, strpos(strtolower($read), 'bodystructure') + 13));
  73     $read = trim(substr ($read, 0, -1));
  74     $end = mime_match_parenthesis(0, $read);
  75     while ($end == strlen($read)-1) {
  76         $read = trim(substr ($read, 0, -1));
  77         $read = trim(substr ($read, 1));
  78         $end = mime_match_parenthesis(0, $read);
  79     }
  80
  81     $msg = mime_parse_structure ($read, 0);
  82     $msg->header = $header;
  83
  84     return( $msg );
  85 }
  86
  87 /* this starts the parsing of a particular structure.  It is called recursively,
  88  * so it can be passed different structures.  It returns an object of type
  89  * $message.
  90  * First, it checks to see if it is a multipart message.  If it is, then it
  91  * handles that as it sees is necessary.  If it is just a regular entity,
  92  * then it parses it and adds the necessary header information (by calling out
  93  * to mime_get_elements()
  94  */
  95 function mime_parse_structure ($structure, $ent_id) {
  96
  97   $msg = new message();
  98   if ($structure{0} == '(') {
  99      $ent_id = mime_new_element_level($ent_id);
 100      $start = $end = -1;
 101      do {
 102         $start = $end+1;
 103         $end = mime_match_parenthesis ($start, $structure);
 104
 105         $element = substr($structure, $start+1, ($end - $start)-1);
 106         $ent_id = mime_increment_id ($ent_id);
 107         $newmsg = mime_parse_structure ($element, $ent_id);
 108         $msg->addEntity ($newmsg);
 109      } while ($structure{$end+1} == '(');
 110   } else {
 111      // parse the elements
 112      $msg = mime_get_element ($structure, $msg, $ent_id);
 113   }
 114   return $msg;
 115 }
 116
 117 /* Increments the element ID.  An element id can look like any of
 118  * the following:  1, 1.2, 4.3.2.4.1, etc.  This function increments
 119  * the last number of the element id, changing 1.2 to 1.3.
 120  */
 121 function mime_increment_id ($id) {
 122
 123     if (strpos($id, ".")) {
 124         $first = substr($id, 0, strrpos($id, "."));
 125         $last = substr($id, strrpos($id, ".")+1);
 126         $last++;
 127         $new = $first . "." .$last;
 128     } else {
 129         $new = $id + 1;
 130     }
 131
 132     return $new;
 133 }
 134
 135 /*
 136  * See comment for mime_increment_id().
 137  * This adds another level on to the entity_id changing 1.3 to 1.3.0
 138  * NOTE:  1.3.0 is not a valid element ID.  It MUST be incremented
 139  *        before it can be used.  I left it this way so as not to have
 140  *        to make a special case if it is the first entity_id.  It
 141  *        always increments it, and that works fine.
 142  */
 143 function mime_new_element_level ($id) {
 144
 145   if (!$id) {
 146       $id = 0;
 147   } else {
 148       $id = $id . '.0';
 149   }
 150
 151   return( $id );
 152 }
 153
 154 function mime_get_element (&$structure, $msg, $ent_id) {
 155
 156   $elem_num = 1;
 157   $msg->header = new msg_header();
 158   $msg->header->entity_id = $ent_id;
 159   $properties = array();
 160
 161   while (strlen($structure) > 0) {
 162      $structure = trim($structure);
 163      $char = $structure{0};
 164
 165      if (strtolower(substr($structure, 0, 3)) == 'nil') {
 166         $text = '';
 167         $structure = substr($structure, 3);
 168      } else if ($char == '"') {
 169         // loop through until we find the matching quote, and return that as a string
 170         $pos = 1;
 171         $text = '';
 172         while ( ($char = $structure{$pos} ) <> '"' && $pos < strlen($structure)) {
 173            $text .= $char;
 174            $pos++;
 175         }
 176         $structure = substr($structure, strlen($text) + 2);
 177      } else if ($char == '(') {
 178         // comment me
 179         $end = mime_match_parenthesis (0, $structure);
 180         $sub = substr($structure, 1, $end-1);
 181         $properties = mime_get_props($properties, $sub);
 182         $structure = substr($structure, strlen($sub) + 2);
 183      } else {
 184         // loop through until we find a space or an end parenthesis
 185         $pos = 0;
 186         $char = $structure{$pos};
 187         $text = '';
 188         while ($char != ' ' && $char != ')' && $pos < strlen($structure)) {
 189            $text .= $char;
 190            $pos++;
 191            $char = $structure{$pos};
 192         }
 193         $structure = substr($structure, strlen($text));
 194      }
 195
 196      // This is where all the text parts get put into the header
 197      switch ($elem_num) {
 198         case 1:
 199            $msg->header->type0 = strtolower($text);
 200            break;
 201         case 2:
 202            $msg->header->type1 = strtolower($text);
 203            break;
 204         case 4: // Id
 205            // Invisimail enclose images with <>
 206            $msg->header->id = str_replace( '<', '', str_replace( '>', '', $text ) );
 207            break;
 208         case 5:
 209            $msg->header->description = $text;
 210            break;
 211         case 6:
 212            $msg->header->encoding = strtolower($text);
 213            break;
 214         case 7:
 215            $msg->header->size = $text;
 216            break;
 217         default:
 218            if ($msg->header->type0 == 'text' && $elem_num == 8) {
 219               // This is a plain text message, so lets get the number of lines
 220               // that it contains.
 221               $msg->header->num_lines = $text;
 222
 223            } else if ($msg->header->type0 == 'message' && $msg->header->type1 == 'rfc822' && $elem_num == 8) {
 224               // This is an encapsulated message, so lets start all over again and
 225               // parse this message adding it on to the existing one.
 226               $structure = trim($structure);
 227               if ( $structure{0} == '(' ) {
 228                  $e = mime_match_parenthesis (0, $structure);
 229                  $structure = substr($structure, 0, $e);
 230                  $structure = substr($structure, 1);
 231                  $m = mime_parse_structure($structure, $msg->header->entity_id);
 232
 233                  // the following conditional is there to correct a bug that wasn't
 234                  // incrementing the entity IDs correctly because of the special case
 235                  // that message/rfc822 is.  This fixes it fine.
 236                  if (substr($structure, 1, 1) != '(')
 237                     $m->header->entity_id = mime_increment_id(mime_new_element_level($ent_id));
 238
 239                  // Now we'll go through and reformat the results.
 240                  if ($m->entities) {
 241                     for ($i=0; $i < count($m->entities); $i++) {
 242                        $msg->addEntity($m->entities[$i]);
 243                     }
 244                  } else {
 245                     $msg->addEntity($m);
 246                  }
 247                  $structure = "";
 248               }
 249            }
 250            break;
 251      }
 252      $elem_num++;
 253      $text = "";
 254   }
 255   // loop through the additional properties and put those in the various headers
 256   if ($msg->header->type0 != 'message') {
 257      for ($i=0; $i < count($properties); $i++) {
 258         $msg->header->{$properties[$i]['name']} = $properties[$i]['value'];
 259      }
 260   }
 261
 262   return $msg;
 263 }
 264
 265 /*
 266  * I did most of the MIME stuff yesterday (June 20, 2000), but I couldn't
 267  * figure out how to do this part, so I decided to go to bed.  I woke up
 268  * in the morning and had a flash of insight.  I went to the white-board
 269  * and scribbled it out, then spent a bit programming it, and this is the
 270  * result.  Nothing complicated, but I think my brain was fried yesterday.
 271  * Funny how that happens some times.
 272  *
 273  * This gets properties in a nested parenthesisized list.  For example,
 274  * this would get passed something like:  ("attachment" ("filename" "luke.tar.gz"))
 275  * This returns an array called $props with all paired up properties.
 276  * It ignores the "attachment" for now, maybe that should change later
 277  * down the road.  In this case, what is returned is:
 278  *    $props[0]["name"] = "filename";
 279  *    $props[0]["value"] = "luke.tar.gz";
 280  */
 281 function mime_get_props ($props, $structure) {
 282
 283   while (strlen($structure) > 0) {
 284      $structure = trim($structure);
 285      $char = $structure{0};
 286
 287      if ($char == '"') {
 288         $pos = 1;
 289         $tmp = '';
 290         while ( ( $char = $structure{$pos} ) != '"' &&
 291                 $pos < strlen($structure)) {
 292            $tmp .= $char;
 293            $pos++;
 294         }
 295         $structure = trim(substr($structure, strlen($tmp) + 2));
 296         $char = $structure{0};
 297
 298         if ($char == '"') {
 299            $pos = 1;
 300            $value = '';
 301            while ( ( $char = $structure{$pos} ) != '"' &&
 302                    $pos < strlen($structure) ) {
 303               $value .= $char;
 304               $pos++;
 305            }
 306            $structure = trim(substr($structure, strlen($tmp) + 2));
 307
 308            $k = count($props);
 309            $props[$k]['name'] = strtolower($tmp);
 310            $props[$k]['value'] = $value;
 311         } else if ($char == '(') {
 312            $end = mime_match_parenthesis (0, $structure);
 313            $sub = substr($structure, 1, $end-1);
 314        if (! isset($props))
 315            $props = array();
 316            $props = mime_get_props($props, $sub);
 317            $structure = substr($structure, strlen($sub) + 2);
 318         }
 319         return $props;
 320      } else if ($char == '(') {
 321         $end = mime_match_parenthesis (0, $structure);
 322         $sub = substr($structure, 1, $end-1);
 323         $props = mime_get_props($props, $sub);
 324         $structure = substr($structure, strlen($sub) + 2);
 325         return $props;
 326      } else {
 327         return $props;
 328      }
 329   }
 330 }
 331
 332 /*
 333  *  Matches parenthesis.  It will return the position of the matching
 334  *  parenthesis in $structure.  For instance, if $structure was:
 335  *     ("text" "plain" ("val1name", "1") nil ... )
 336  *     x                                         x
 337  *  then this would return 42 to match up those two.
 338  */
 339 function mime_match_parenthesis ($pos, $structure) {
 340
 341     $j = strlen( $structure );
 342
 343     // ignore all extra characters
 344     // If inside of a string, skip string -- Boundary IDs and other
 345     // things can have ) in them.
 346     if ( $structure{$pos} != '(' ) {
 347         return( $j );
 348     }
 349
 350     while ( $pos < $j ) {
 351         $pos++;
 352         if ($structure{$pos} == ')') {
 353             return $pos;
 354         } elseif ($structure{$pos} == '"') {
 355             $pos++;
 356             while ( $structure{$pos} != '"' &&
 357                     $pos < $j ) {
 358                if (substr($structure, $pos, 2) == '\\"') {
 359                   $pos++;
 360                } elseif (substr($structure, $pos, 2) == '\\\\') {
 361                   $pos++;
 362                }
 363                $pos++;
 364             }
 365         } elseif ( $structure{$pos} == '(' ) {
 366             $pos = mime_match_parenthesis ($pos, $structure);
 367         }
 368     }
 369     echo _("Error decoding mime structure.  Report this as a bug!") . '<br>';
 370     return( $pos );
 371 }
 372
 373 function mime_fetch_body ($imap_stream, $id, $ent_id ) {
 374     // do a bit of error correction.  If we couldn't find the entity id, just guess
 375     // that it is the first one.  That is usually the case anyway.
 376     if (!$ent_id)
 377         $ent_id = 1;
 378     $data = sqimap_run_command ($imap_stream, "FETCH $id BODY[$ent_id]", true, $response, $message);
 379     $topline = array_shift($data);
 380     while (! ereg('\\* [0-9]+ FETCH ', $topline) && $data)
 381         $topline = array_shift($data);
 382     $wholemessage = implode('', $data);
 383     if (ereg('\\{([^\\}]*)\\}', $topline, $regs)) {
 384         $ret = substr( $wholemessage, 0, $regs[1] );
 385         /*
 386             There is some information in the content info header that could be important
 387             in order to parse html messages. Let's get them here.
 388         */
 389         if ( $ret{0} == '<' ) {
 390             $data = sqimap_run_command ($imap_stream, "FETCH $id BODY[$ent_id.MIME]", true, $response, $message);
 391             $base = '';
 392             $k = 10;
 393             foreach( $data as $d ) {
 394                 if ( substr( $d, 0, 13 ) == 'Content-Base:' ) {
 395                     $j = strlen( $d );
 396                     $i = 13;
 397                     $base = '';
 398                     while ( $i < $j &&
 399                            ( !isNoSep( $d{$i} ) || $d{$i} == '"' )  )
 400                         $i++;
 401                     while ( $i < $j ) {
 402                         if ( isNoSep( $d{$i} ) )
 403                             $base .= $d{$i};
 404                         $i++;
 405                     }
 406                     $k = 0;
 407                 } elseif ( $k == 1 && !isnosep( $d{0} ) ) {
 408                     $base .= substr( $d, 1 );
 409                 }
 410                 $k++;
 411             }
 412             if ( $base <> '' )
 413
 414                 $ret = "<base href=\"$base\">" . $ret;
 415         }
 416     } else if (ereg('"([^"]*)"', $topline, $regs)) {
 417         $ret = $regs[1];
 418     } else {
 419         global $where, $what, $mailbox, $passed_id, $startMessage;
 420         $par = "mailbox=".urlencode($mailbox)."&passed_id=$passed_id";
 421         if (isset($where) && isset($what)) {
 422             $par .= "&where=".urlencode($where)."&what=".urlencode($what);
 423         } else {
 424             $par .= "&startMessage=$startMessage&show_more=0";
 425         }
 426         $par .= '&response='.urlencode($response).'&message='.urlencode($message).
 427                 '&topline='.urlencode($topline);
 428
 429         echo   '<b><font color=$color[2]>' .
 430                _("Body retrieval error. The reason for this is most probably that the message is malformed. Please help us making future versions better by submitting this message to the developers knowledgebase!") .
 431                "<A HREF=\"../src/retrievalerror.php?$par\">Submit message</A><BR>" .
 432                '<tt>' . _("Response:") . "$response<BR>" .
 433                _("Message:") . " $message<BR>" .
 434                _("FETCH line:") . " $topline<BR></tt></font></b>";
 435
 436         $data = sqimap_run_command ($imap_stream, "FETCH $passed_id BODY[]", true, $response, $message);
 437         array_shift($data);
 438         $wholemessage = implode('', $data);
 439
 440         $ret = "---------------\n$wholemessage";
 441
 442     }
 443     return( $ret );
 444 }
 445
 446 function mime_print_body_lines ($imap_stream, $id, $ent_id, $encoding) {
 447     // do a bit of error correction.  If we couldn't find the entity id, just guess
 448     // that it is the first one.  That is usually the case anyway.
 449     if (!$ent_id) {
 450         $ent_id = 1;
 451     }
 452     $sid = sqimap_session_id();
 453     // Don't kill the connection if the browser is over a dialup
 454     // and it would take over 30 seconds to download it.
 455
 456     // don´t call set_time_limit in safe mode.
 457     if (!ini_get("safe_mode")) {
 458         set_time_limit(0);
 459     }
 460
 461     fputs ($imap_stream, "$sid FETCH $id BODY[$ent_id]\r\n");
 462     $cnt = 0;
 463     $continue = true;
 464     $read = fgets ($imap_stream,4096);
 465     // This could be bad -- if the section has sqimap_session_id() . ' OK'
 466     // or similar, it will kill the download.
 467     while (!ereg("^".$sid." (OK|BAD|NO)(.*)$", $read, $regs)) {
 468       if (trim($read) == ')==') {
 469           $read1 = $read;
 470           $read = fgets ($imap_stream,4096);
 471           if (ereg("^".$sid." (OK|BAD|NO)(.*)$", $read, $regs)) {
 472               return;
 473           } else {
 474               echo decodeBody($read1, $encoding) .
 475                    decodeBody($read, $encoding);
 476           }
 477       } else if ($cnt) {
 478           echo decodeBody($read, $encoding);
 479       }
 480       $read = fgets ($imap_stream,4096);
 481       $cnt++;
 482     }
 483 }
 484
 485 /* -[ END MIME DECODING ]----------------------------------------------------------- */
 486
 487
 488
 489 /* This is the first function called.  It decides if this is a multipart
 490    message or if it should be handled as a single entity
 491  */
 492 function decodeMime ($imap_stream, &$header) {
 493     global $username, $key, $imapServerAddress, $imapPort;
 494     return mime_structure ($imap_stream, $header);
 495 }
 496
 497 // This is here for debugging purposese.  It will print out a list
 498 // of all the entity IDs that are in the $message object.
 499 /*
 500 function listEntities ($message) {
 501 if ($message) {
 502  if ($message->header->entity_id)
 503  echo "<tt>" . $message->header->entity_id . ' : ' . $message->header->type0 . '/' . $message->header->type1 . '<br>';
 504  for ($i = 0; $message->entities[$i]; $i++) {
 505     $msg = listEntities($message->entities[$i], $ent_id);
 506     if ($msg)
 507        return $msg;
 508  }
 509 }
 510 }
 511 */
 512
 513 /* returns a $message object for a particular entity id */
 514 function getEntity ($message, $ent_id) {
 515     if ($message) {
 516         if ($message->header->entity_id == $ent_id && strlen($ent_id) == strlen($message->header->entity_id)) {
 517             return $message;
 518         } else {
 519             for ($i = 0; isset($message->entities[$i]); $i++) {
 520                 $msg = getEntity ($message->entities[$i], $ent_id);
 521                 if ($msg) {
 522                     return $msg;
 523                 }
 524             }
 525         }
 526     }
 527 }
 528
 529 /*
 530  * figures out what entity to display and returns the $message object
 531  * for that entity.
 532  */
 533 function findDisplayEntity ($message, $textOnly = 1)   {
 534     global $show_html_default;
 535
 536     $entity = 0;
 537
 538     if ($message) {
 539         if ( $message->header->type0 == 'multipart' &&
 540              ( $message->header->type1 == 'alternative' ||
 541                $message->header->type1 == 'related' ) &&
 542              $show_html_default && ! $textOnly ) {
 543             $entity = findDisplayEntityHTML($message);
 544         }
 545
 546         // Show text/plain or text/html -- the first one we find.
 547         if ( $entity == 0 &&
 548              $message->header->type0 == 'text' &&
 549              ( $message->header->type1 == 'plain' ||
 550                $message->header->type1 == 'html' ) &&
 551              isset($message->header->entity_id) ) {
 552             $entity = $message->header->entity_id;
 553         }
 554
 555         $i = 0;
 556         while ($entity == 0 && isset($message->entities[$i]) ) {
 557             $entity = findDisplayEntity($message->entities[$i], $textOnly);
 558             $i++;
 559         }
 560     }
 561
 562     return( $entity );
 563 }
 564
 565 /* Shows the HTML version */
 566 function findDisplayEntityHTML ($message) {
 567
 568     if ( $message->header->type0 == 'text' &&
 569          $message->header->type1 == 'html' &&
 570          isset($message->header->entity_id)) {
 571         return $message->header->entity_id;
 572     }
 573     for ($i = 0; isset($message->entities[$i]); $i ++) {
 574         $entity = findDisplayEntityHTML($message->entities[$i]);
 575         if ($entity != 0) {
 576             return $entity;
 577         }
 578     }
 579
 580     return 0;
 581 }
 582
 583 /* This returns a parsed string called $body. That string can then
 584 be displayed as the actual message in the HTML. It contains
 585 everything needed, including HTML Tags, Attachments at the
 586 bottom, etc.
 587 */
 588 function formatBody($imap_stream, $message, $color, $wrap_at) {
 589     // this if statement checks for the entity to show as the
 590     // primary message. To add more of them, just put them in the
 591     // order that is their priority.
 592     global $startMessage, $username, $key, $imapServerAddress, $imapPort,
 593            $show_html_default;
 594
 595     $id = $message->header->id;
 596     $urlmailbox = urlencode($message->header->mailbox);
 597
 598     // Get the right entity and redefine message to be this entity
 599     // Pass the 0 to mean that we want the 'best' viewable one
 600     $ent_num = findDisplayEntity ($message, 0);
 601     $body_message = getEntity($message, $ent_num);
 602     if (($body_message->header->type0 == 'text') ||
 603         ($body_message->header->type0 == 'rfc822')) {
 604
 605         $body = mime_fetch_body ($imap_stream, $id, $ent_num);
 606         $body = decodeBody($body, $body_message->header->encoding);
 607         $hookResults = do_hook("message_body", $body);
 608         $body = $hookResults[1];
 609
 610         // If there are other types that shouldn't be formatted, add
 611         // them here
 612         if ($body_message->header->type1 == 'html') {
 613             if ( $show_html_default <> 1 ) {
 614                 $body = strip_tags( $body );
 615                 translateText($body, $wrap_at, $body_message->header->charset);
 616             } else {
 617                 $body = MagicHTML( $body, $id );
 618             }
 619         } else {
 620             translateText($body, $wrap_at, $body_message->header->charset);
 621         }
 622
 623         $body .= "<SMALL><CENTER><A HREF=\"../src/download.php?absolute_dl=true&passed_id=$id&passed_ent_id=$ent_num&mailbox=$urlmailbox&showHeaders=1\">". _("Download this as a file") ."</A></CENTER><BR></SMALL>";
 624
 625         /** Display the ATTACHMENTS: message if there's more than one part **/
 626         $body .= "</TD></TR></TABLE>";
 627         if (isset($message->entities[0])) {
 628             $body .= formatAttachments ($message, $ent_num, $message->header->mailbox, $id);
 629         }
 630         $body .= "</TD></TR></TABLE>";
 631     } else {
 632         $body = formatAttachments ($message, -1, $message->header->mailbox, $id);
 633     }
 634     return ($body);
 635 }
 636
 637 /*
 638  * A recursive function that returns a list of attachments with links
 639  * to where to download these attachments
 640  */
 641 function formatAttachments ($message, $ent_id, $mailbox, $id) {
 642   global $where, $what;
 643   global $startMessage, $color;
 644   static $ShownHTML = 0;
 645
 646   $body = "";
 647   if ($ShownHTML == 0) {
 648         $ShownHTML = 1;
 649
 650         $body .= "<TABLE WIDTH=\"100%\" CELLSPACING=0 CELLPADDING=2 BORDER=0 BGCOLOR=\"$color[0]\"><TR>\n" .
 651                  "<TH ALIGN=\"left\" BGCOLOR=\"$color[9]\"><B>\n" .
 652                  _("Attachments") . ':' .
 653                  "</B></TH></TR><TR><TD>\n" .
 654                  "<TABLE CELLSPACING=0 CELLPADDING=1 BORDER=0>\n" .
 655                  formatAttachments ($message, $ent_id, $mailbox, $id) .
 656                  "</TABLE></TD></TR></TABLE>";
 657
 658         return( $body );
 659   }
 660
 661   if ($message) {
 662      if (!$message->entities) {
 663         $type0 = strtolower($message->header->type0);
 664         $type1 = strtolower($message->header->type1);
 665         $name = decodeHeader($message->header->name);
 666
 667         if ($message->header->entity_id != $ent_id) {
 668            $filename = decodeHeader($message->header->filename);
 669            if (trim($filename) == '') {
 670               if (trim($name) == '') {
 671                  if ( trim( $message->header->id ) == '' )
 672                     $display_filename = 'untitled-[' . $message->header->entity_id . ']' ;
 673                  else
 674                     $display_filename = 'cid: ' . $message->header->id;
 675                  // $display_filename = 'untitled-[' . $message->header->entity_id . ']' ;
 676               } else {
 677                  $display_filename = $name;
 678                  $filename = $name;
 679               }
 680            } else {
 681               $display_filename = $filename;
 682            }
 683
 684            $urlMailbox = urlencode($mailbox);
 685            $ent = urlencode($message->header->entity_id);
 686
 687            $DefaultLink =
 688               "../src/download.php?startMessage=$startMessage&passed_id=$id&mailbox=$urlMailbox&passed_ent_id=$ent";
 689            if ($where && $what)
 690               $DefaultLink .= '&where=' . urlencode($where) . '&what=' . urlencode($what);
 691            $Links['download link']['text'] = _("download");
 692            $Links['download link']['href'] =
 693                "../src/download.php?absolute_dl=true&passed_id=$id&mailbox=$urlMailbox&passed_ent_id=$ent";
 694            $ImageURL = '';
 695
 696            /* this executes the attachment hook with a specific MIME-type.
 697             * if that doens't have results, it tries if there's a rule
 698             * for a more generic type. */
 699            $HookResults = do_hook("attachment $type0/$type1", $Links,
 700                $startMessage, $id, $urlMailbox, $ent, $DefaultLink,
 701                $display_filename, $where, $what);
 702            if(count($HookResults[1]) <= 1) {
 703                $HookResults = do_hook("attachment $type0/*", $Links,
 704                $startMessage, $id, $urlMailbox, $ent, $DefaultLink,
 705                $display_filename, $where, $what);
 706            }
 707
 708            $Links = $HookResults[1];
 709            $DefaultLink = $HookResults[6];
 710
 711            $body .= '<TR><TD>&nbsp;&nbsp;</TD><TD>' .
 712                     "<A HREF=\"$DefaultLink\">$display_filename</A>&nbsp;</TD>" .
 713                     '<TD><SMALL><b>' . show_readable_size($message->header->size) .
 714                     '</b>&nbsp;&nbsp;</small></TD>' .
 715                     "<TD><SMALL>[ $type0/$type1 ]&nbsp;</SMALL></TD>" .
 716                     '<TD><SMALL>';
 717            if ($message->header->description)
 718               $body .= '<b>' . htmlspecialchars($message->header->description) . '</b>';
 719            $body .= '</SMALL></TD><TD><SMALL>&nbsp;';
 720
 721
 722            $SkipSpaces = 1;
 723            foreach ($Links as $Val) {
 724               if ($SkipSpaces) {
 725                  $SkipSpaces = 0;
 726               } else {
 727                  $body .= '&nbsp;&nbsp;|&nbsp;&nbsp;';
 728               }
 729               $body .= '<a href="' . $Val['href'] . '">' .  $Val['text'] . '</a>';
 730            }
 731
 732            unset($Links);
 733
 734            $body .= "</SMALL></TD></TR>\n";
 735         }
 736      } else {
 737         for ($i = 0; $i < count($message->entities); $i++) {
 738            $body .= formatAttachments ($message->entities[$i], $ent_id, $mailbox, $id);
 739         }
 740      }
 741      return( $body );
 742   }
 743 }
 744
 745
 746 /** this function decodes the body depending on the encoding type. **/
 747 function decodeBody($body, $encoding) {
 748   $body = str_replace("\r\n", "\n", $body);
 749   $encoding = strtolower($encoding);
 750
 751   global $show_html_default;
 752
 753   if ($encoding == 'quoted-printable') {
 754      $body = quoted_printable_decode($body);
 755
 756
 757      while (ereg("=\n", $body))
 758         $body = ereg_replace ("=\n", "", $body);
 759
 760   } else if ($encoding == 'base64') {
 761      $body = base64_decode($body);
 762   }
 763
 764   // All other encodings are returned raw.
 765   return $body;
 766 }
 767
 768 /*
 769  * This functions decode strings that is encoded according to
 770  * RFC1522 (MIME Part Two: Message Header Extensions for Non-ASCII Text).
 771  */
 772 function decodeHeader ($string, $utfencode=true) {
 773   if (eregi('=\\?([^?]+)\\?(q|b)\\?([^?]+)\\?=',
 774             $string, $res)) {
 775      if (ucfirst($res[2]) == 'B') {
 776         $replace = base64_decode($res[3]);
 777      } else {
 778         $replace = str_replace('_', ' ', $res[3]);
 779     // Convert lowercase Quoted Printable to uppercase for
 780     // quoted_printable_decode to understand it.
 781     while (ereg("(=(([0-9][abcdef])|([abcdef][0-9])|([abcdef][abcdef])))", $replace, $res)) {
 782        $replace = str_replace($res[1], strtoupper($res[1]), $replace);
 783     }
 784         $replace = quoted_printable_decode($replace);
 785      }
 786      /* Only encode into entities by default. Some places
 787         don't need the encoding, like the compose form. */
 788      if ($utfencode){
 789          $replace = charset_decode ($res[1], $replace);
 790      }
 791
 792      // Remove the name of the character set.
 793      $string = eregi_replace ('=\\?([^?]+)\\?(q|b)\\?([^?]+)\\?=',
 794          $replace, $string);
 795
 796      // In case there should be more encoding in the string: recurse
 797      return (decodeHeader($string));
 798   } else
 799      return ($string);
 800 }
 801
 802 /*
 803  * Encode a string according to RFC 1522 for use in headers if it
 804  * contains 8-bit characters or anything that looks like it should
 805  * be encoded.
 806  */
 807 function encodeHeader ($string) {
 808     global $default_charset;
 809
 810     // Encode only if the string contains 8-bit characters or =?
 811     $j = strlen( $string  );
 812     $l = strstr($string, '=?');         // Must be encoded ?
 813     $ret = '';
 814     for( $i=0; $i < $j; ++$i) {
 815         switch( $string{$i} ) {
 816            case '=':
 817           $ret .= '=3D';
 818           break;
 819         case '?':
 820           $ret .= '=3F';
 821           break;
 822         case '_':
 823           $ret .= '=5F';
 824           break;
 825         case ' ':
 826           $ret .= '_';
 827           break;
 828         default:
 829           $k = ord( $string{$i} );
 830           if ( $k > 126 ) {
 831              $ret .= sprintf("=%02X", $k);
 832              $l = TRUE;
 833           } else
 834              $ret .= $string{$i};
 835         }
 836     }
 837
 838     if ( $l ) {
 839         $string = "=?$default_charset?Q?$ret?=";
 840     }
 841
 842     return( $string );
 843 }
 844
 845 /*
 846     Strips dangerous tags from html messages.
 847 */
 848 function MagicHTML( $body, $id ) {
 849
 850     global $message, $HTTP_SERVER_VARS,
 851            $attachment_common_show_images;
 852
 853     $attachment_common_show_images =
 854                      FALSE; // Don't display attached images in HTML mode
 855     $j = strlen( $body );   // Legnth of the HTML
 856     $ret = '';              // Returned string
 857     $bgcolor = '#ffffff';   // Background style color (defaults to white)
 858     $textcolor = '#000000'; // Foreground style color (defaults to black)
 859     $leftmargin = '';       // Left margin style
 860     $title = '';            // HTML title if any
 861
 862     $i = 0;
 863     while ( $i < $j ) {
 864         if ( $body{$i} == '<' ) {
 865             $pos = $i + 1;
 866             $tag = '';
 867             while ($body{$pos} == ' ' || $body{$pos} == "\t" ||
 868                    $body{$pos} == "\n") {
 869                 $pos ++;
 870             }
 871             while (strlen($tag) < 4 && $body{$pos} != ' ' &&
 872                    $body{$pos} != "\t" && $body{$pos} != "\n") {
 873                 $tag .= $body{$pos};
 874                 $pos ++;
 875             }
 876             switch( strtoupper( $tag ) ) {
 877             // Strips the entire tag and contents
 878             case 'APPL':
 879             case 'EMBB':
 880             case 'FRAM':
 881             case 'SCRI':
 882             case 'OBJE':
 883                 $etg = '/' . $tag;
 884                 while ( $body{$i+1}.$body{$i+2}.$body{$i+3}.$body{$i+4}.$body{$i+5} <> $etg  &&
 885                        $i < $j  ) $i++;
 886                 while ( $i < $j && $body{++$i} <> '>' );
 887                 // $ret .= "<!-- $tag removed -->";
 888                 break;
 889             // Substitute Title
 890             case 'TITL':
 891                 $i += 5;
 892                 while ( $body{$i} <> '>' &&  // </title>
 893                        $i < $j )
 894                         $i++;
 895                 $i++;
 896                 $title = '';
 897                 while ( $body{$i} <> '<' &&  // </title>
 898                        $i < $j ) {
 899                     $title .= $body{$i};
 900                     $i++;
 901                 }
 902                 $i += 7;
 903                 break;
 904             // Destroy these tags
 905             case 'HTML':
 906             case 'HEAD':
 907             case '/HTM':
 908             case '/HEA':
 909             case '!DOC':
 910             case 'META':
 911             case 'DIV ':
 912             case '/DIV':
 913             case '!-- ':
 914                 $i += 4;
 915                 while ( $body{$i}  <> '>' &&
 916                        $i < $j )
 917                     $i++;
 918                 // $i++;
 919                 break;
 920             case 'STYL':
 921                 $i += 5;
 922                 while ( $body{$i} <> '>' &&  // </title>
 923                        $i < $j )
 924                         $i++;
 925                 $i++;
 926                 // We parse the style to look for interesting stuff
 927                 $styleblk = '';
 928                 while ( $body{$i} <> '>' &&
 929                        $i < $j ) {
 930                     // First we get the name of the style
 931                     $style = '';
 932                     while ( $body{$i} <> '>' &&
 933                            $body{$i} <> '<' &&
 934                            $body{$i} <> '{' &&
 935                            $i < $j ) {
 936                        if ( isnoSep( $body{$i} ) )
 937                            $style .= $body{$i};
 938                        $i++;
 939                     }
 940                     stripComments( $i, $j, $body );
 941                     $style = strtoupper( trim( $style ) );
 942                     if ( $style == 'BODY' ) {
 943                         // Next we look into the definitions of the body style
 944                         while ( $body{$i} <> '>' &&
 945                                $body{$i} <> '}' &&
 946                                $i < $j ) {
 947                             // We look for the background color if any.
 948                             if ( substr( $body, $i, 17 ) == 'BACKGROUND-COLOR:' ) {
 949                                 $i += 17;
 950                                 $bgcolor = getStyleData( $i, $j, $body );
 951                             } elseif ( substr( $body, $i, 12 ) == 'MARGIN-LEFT:' ) {
 952                                 $i += 12;
 953                                 $leftmargin = getStyleData( $i, $j, $body );
 954                             }
 955                             $i++;
 956                         }
 957                     } else {
 958                         // Other style are mantained
 959                         $styleblk .= "$style ";
 960                         while ( $body{$i} <> '>' &&
 961                                $body{$i} <> '<' &&
 962                                $body{$i} <> '}' &&
 963                                $i < $j ) {
 964                             $styleblk .= $body{$i};
 965                             $i++;
 966                         }
 967                         $styleblk .= $body{$i};
 968                     }
 969                     stripComments( $i, $j, $body );
 970                     if ( $body{$i} <> '>' )
 971                         $i++;
 972                 }
 973                 if ( $styleblk <> '' )
 974                     $ret .= "<style>$styleblk";
 975                 break;
 976             case 'BODY':
 977                 if ( $title <> '' )
 978                     $ret .= '<b>' . _("Title:") . " </b>$title<br>\n";
 979                 $ret .= "<TABLE";
 980                 $i += 5;
 981                 if (! isset($base)) {
 982                     $base = '';
 983                 }
 984                 $ret .= stripEvent( $i, $j, $body, $id, $base );
 985                 $ret .= " bgcolor=$bgcolor width=\"100%\"><tr>";
 986                 if ( $leftmargin <> '' )
 987                     $ret .= "<td width=$leftmargin>&nbsp;</td>";
 988                 $ret .= '<td>';
 989                 if (strtolower($bgcolor) == 'ffffff' ||
 990                     strtolower($bgcolor) == '#ffffff')
 991                     $ret .= '<font color=#000000>';
 992                 break;
 993             case 'BASE':
 994                 $i += 5;
 995                 $base = '';
 996                 while ( !isNoSep( $body{$i} ) &&
 997                        $i < $j ) {
 998                         $i++;
 999                 }
1000                 if ( strcasecmp( substr( $base, 0, 4 ), 'href'  ) ) {
1001                         $i += 5;
1002                         while ( !isNoSep( $body{$i} ) &&
1003                                $i < $j ) {
1004                                 $i++;
1005                         }
1006                         while ( $body{$i} <> '>' &&
1007                                $i < $j ) {
1008                             if ( $body{$i} <> '"' ) {
1009                                 $base .= $body{$i};
1010                             }
1011                             $i++;
1012                         }
1013                         // Debuging $ret .= "<!-- base == $base -->";
1014                         if ( strcasecmp( substr( $base, 0, 4 ), 'file' ) <> 0 ) {
1015                             $ret .= "\n<BASE HREF=\"$base\">\n";
1016                         }
1017                 }
1018                 break;
1019             case '/BOD':
1020                 $ret .= '</font></td></tr></TABLE>';
1021                 $i += 6;
1022                 break;
1023             default:
1024                 // Following tags can contain some event handler, lets search it
1025                 stripComments( $i, $j, $body );
1026                 if (! isset($base)) {
1027                    $base = '';
1028                 }
1029                 $ret .= stripEvent( $i, $j, $body, $id, $base ) . '>';
1030                         // $ret .= "<!-- $tag detected -->";
1031             }
1032         } else {
1033             $ret .= $body{$i};
1034         }
1035     $i++;
1036     }
1037
1038 return( "\n\n<!-- HTML Output ahead -->\n" .
1039         $ret .
1040         "\n<!-- END of HTML Output --><base href=\"".
1041         get_location() . '/'.
1042         "\">\n\n" );
1043 }
1044
1045 function isNoSep( $char ) {
1046
1047     switch( $char ) {
1048     case ' ':
1049     case "\n":
1050     case "\t":
1051     case "\r":
1052     case '>':
1053     case '"':
1054         return( FALSE );
1055         break;
1056     default:
1057         return( TRUE );
1058     }
1059
1060 }
1061
1062 /*
1063   The following function is usefull to remove extra data that can cause
1064   html not to display properly. Especialy with MS stuff.
1065 */
1066
1067 function stripComments( &$i, $j, &$body ) {
1068
1069     while ( $body{$i}.$body{$i+1}.$body{$i+2}.$body{$i+3} == '<!--' &&
1070            $i < $j ) {
1071         $i += 5;
1072         while ( $body{$i-2}.$body{$i-1}.$body{$i} <> '-->' &&
1073                $i < $j )
1074             $i++;
1075         $i++;
1076     }
1077
1078     return;
1079
1080 }
1081
1082 /* Gets the style data of a specific style */
1083
1084 function getStyleData( &$i, $j, &$body ) {
1085
1086     // We skip spaces
1087     while ( $body{$i} <> '>' && !isNoSep( $body{$i} ) &&
1088            $i < $j ) {
1089         $i++;
1090     }
1091     // And get the color
1092     $ret = '';
1093     while ( isNoSep( $body{$i} ) &&
1094            $i < $j ) {
1095         $ret .= $body{$i};
1096         $i++;
1097     }
1098
1099     return( $ret );
1100 }
1101
1102 /*
1103 Private function for strip_dangerous_tag. Look for event based coded and "remove" it
1104 change on with no (onload -> noload)
1105 */
1106
1107 function stripEvent( &$i, $j, &$body, $id, $base ) {
1108
1109     global $message, $base_uri;
1110
1111     $ret = '';
1112
1113     while ( $body{$i} <> '>' &&
1114            $i < $j ) {
1115         $etg = strtolower($body{$i}.$body{$i+1}.$body{$i+2});
1116         switch( $etg ) {
1117         case 'src':
1118             // This is probably a src specification
1119             $k = $i + 3;
1120             while( !isNoSep( $body{$k} )) {
1121                 $k++;
1122             }
1123             if ( $body{$k} == '=' ) {
1124                 /* It is indeed */
1125                 $k++;
1126                 while( !isNoSep( $body{$k} ) &&
1127                        $k < $j ) {
1128                     $k++;
1129                 }
1130                 $src = '';
1131                 while ( $body{$k} <> '>' && isNoSep( $body{$k} ) &&
1132                        $k < $j ) {
1133                     $src .= $body{$k};
1134                     $k++;
1135                 }
1136                 while( !isNoSep( $body{$k} ) &&
1137                        $k < $j ) {
1138                     $k++;
1139                 }
1140                 if ( strtolower( substr( $src, 0, 4 ) ) == 'cid:' ) {
1141                     $src = substr( $src, 4 );
1142                     $src = "../src/download.php?absolute_dl=true&passed_id=$id&mailbox=" .
1143                            urlencode( $message->header->mailbox ) .
1144                            "&passed_ent_id=" . find_ent_id( $src, $message );
1145                 } else if ( strtolower( substr( $src, 0, 4 ) ) <> 'http' ||
1146                             stristr( $src, $base_uri ) ) {
1147                     /* Javascript and local urls goes out */
1148                     $src = '../images/' . _("sec_remove_eng.png");
1149                 }
1150                 $ret .= 'src="' . $src . '" ';
1151                 $i = $k - 3;
1152             } else {
1153                 $ret .= 'src';
1154                 $i = $i + 3;
1155             }
1156
1157             break;
1158         case '../':
1159             // Retrolinks are not allowed without a base because they mess with SM security
1160             if ( $base == '' ) {
1161                     $i += 2;
1162             } else {
1163                     $ret .= '.';
1164             }
1165             break;
1166         case 'cid':
1167             // Internal link
1168             $k = $i-1;
1169             if ( $body{$i+3} == ':') {
1170                 $i +=4;
1171                 $name = '';
1172                 while ( isNoSep( $body{$i} ) &&
1173                        $i < $j  ) {
1174                     $name .= $body{$i++};
1175                 }
1176                 if ( $name <> '' ) {
1177                     $ret .= "../src/download.php?absolute_dl=true&passed_id=$id&mailbox=" .
1178                                 urlencode( $message->header->mailbox ) .
1179                                 "&passed_ent_id=" . find_ent_id( $name, $message );
1180                     if ( $body{$k} == '"' )
1181                         $ret .= '" ';
1182                     else
1183                         $ret .= ' ';
1184                 }
1185                 if ( $body{$i} == '>' )
1186                     $i -= 1;
1187             }
1188             break;
1189         case ' on':
1190         case "\non":
1191         case "\ron":
1192         case "\ton":
1193             $ret .= ' no';
1194             $i += 2;
1195             break;
1196         case 'pt:':
1197             if ( strcasecmp( $body{$i-4}.$body{$i-3}.$body{$i-2}.$body{$i-1}.$body{$i}.$body{$i+1}.$body{$i+2}, 'script:') == 0 ) {
1198                 $ret .= '_no/';
1199             } else {
1200                 $ret .= $etg;
1201             }
1202             $i += 2;
1203             break;
1204         default:
1205             $ret .= $body{$i};
1206         }
1207         $i++;
1208     }
1209     return( $ret );
1210 }
1211
1212
1213 /* This function trys to locate the entity_id of a specific mime element */
1214
1215 function find_ent_id( $id, $message ) {
1216
1217     $ret = '';
1218     for ($i=0; $ret == '' && $i < count($message->entities); $i++) {
1219
1220         if ( $message->entities[$i]->header->entity_id == '' ) {
1221             $ret = find_ent_id( $id, $message->entities[$i] );
1222         } else {
1223             if ( strcasecmp( $message->entities[$i]->header->id, $id ) == 0 )
1224                 $ret = $message->entities[$i]->header->entity_id;
1225         }
1226
1227     }
1228
1229     return( $ret );
1230
1231 }
1232 ?>