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