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, $ent_id ) {
 374
 375     /*
 376      * do a bit of error correction.  If we couldn't find the entity id, just guess
 377      * that it is the first one.  That is usually the case anyway.
 378      */
 379     if (!$ent_id) {
 380         $ent_id = 1;
 381     }
 382
 383     $cmd = "FETCH $id BODY[$ent_id]";
 384     $data = sqimap_run_command ($imap_stream, $cmd, true, $response, $message);
 385     $topline = array_shift($data);
 386     while (! ereg('\\* [0-9]+ FETCH ', $topline) && $data) {
 387         $topline = array_shift($data);
 388     }
 389     $wholemessage = implode('', $data);
 390     if (ereg('\\{([^\\}]*)\\}', $topline, $regs)) {
 391         $ret = substr( $wholemessage, 0, $regs[1] );
 392         /*
 393             There is some information in the content info header that could be important
 394             in order to parse html messages. Let's get them here.
 395         */
 396         if ( $ret{0} == '<' ) {
 397             $data = sqimap_run_command ($imap_stream, "FETCH $id BODY[$ent_id.MIME]", true, $response, $message);
 398             $base = '';
 399             $k = 10;
 400             foreach( $data as $d ) {
 401                 if ( substr( $d, 0, 13 ) == 'Content-Base:' ) {
 402                     $j = strlen( $d );
 403                     $i = 13;
 404                     $base = '';
 405                     while ( $i < $j &&
 406                            ( !isNoSep( $d{$i} ) || $d{$i} == '"' )  )
 407                         $i++;
 408                     while ( $i < $j ) {
 409                         if ( isNoSep( $d{$i} ) )
 410                             $base .= $d{$i};
 411                         $i++;
 412                     }
 413                     $k = 0;
 414                 } elseif ( $k == 1 && !isnosep( $d{0} ) ) {
 415                     $base .= substr( $d, 1 );
 416                 }
 417                 $k++;
 418             }
 419             if ( $base <> '' ) {
 420                 $ret = "<base href=\"$base\">" . $ret;
 421             }
 422         }
 423     } else if (ereg('"([^"]*)"', $topline, $regs)) {
 424         $ret = $regs[1];
 425     } else {
 426         global $where, $what, $mailbox, $passed_id, $startMessage;
 427         $par = 'mailbox=' . urlencode($mailbox) . "&passed_id=$passed_id";
 428         if (isset($where) && isset($what)) {
 429             $par .= '&where='. urlencode($where) . "&what=" . urlencode($what);
 430         } else {
 431             $par .= "&startMessage=$startMessage&show_more=0";
 432         }
 433         $par .= '&response=' . urlencode($response) .
 434                 '&message=' . urlencode($message).
 435                 '&topline=' . urlencode($topline);
 436
 437         echo   '<tt><br>' .
 438                '<table width="80%"><tr>' .
 439                '<tr><td colspan=2>' .
 440                _("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!") .
 441                " <A HREF=\"../src/retrievalerror.php?$par\"><br>" .
 442                _("Submit message") . '</A><BR>&nbsp;' .
 443                '</td></tr>' .
 444                '<td><b>' . _("Command:") . "</td><td>$cmd</td></tr>" .
 445                '<td><b>' . _("Response:") . "</td><td>$response</td></tr>" .
 446                '<td><b>' . _("Message:") . "</td><td>$message</td></tr>" .
 447                '<td><b>' . _("FETCH line:") . "</td><td>$topline</td></tr>" .
 448                "</table><BR></tt></font><hr>";
 449
 450         $data = sqimap_run_command ($imap_stream, "FETCH $passed_id BODY[]", true, $response, $message);
 451         array_shift($data);
 452         $wholemessage = implode('', $data);
 453
 454         $ret = $wholemessage;
 455     }
 456     return( $ret );
 457 }
 458
 459 function mime_print_body_lines ($imap_stream, $id, $ent_id, $encoding) {
 460     // do a bit of error correction.  If we couldn't find the entity id, just guess
 461     // that it is the first one.  That is usually the case anyway.
 462     if (!$ent_id) {
 463         $ent_id = 1;
 464     }
 465     $sid = sqimap_session_id();
 466     // Don't kill the connection if the browser is over a dialup
 467     // and it would take over 30 seconds to download it.
 468
 469     // don´t call set_time_limit in safe mode.
 470     if (!ini_get("safe_mode")) {
 471         set_time_limit(0);
 472     }
 473
 474     fputs ($imap_stream, "$sid FETCH $id BODY[$ent_id]\r\n");
 475     $cnt = 0;
 476     $continue = true;
 477     $read = fgets ($imap_stream,4096);
 478     // This could be bad -- if the section has sqimap_session_id() . ' OK'
 479     // or similar, it will kill the download.
 480     while (!ereg("^".$sid." (OK|BAD|NO)(.*)$", $read, $regs)) {
 481       if (trim($read) == ')==') {
 482           $read1 = $read;
 483           $read = fgets ($imap_stream,4096);
 484           if (ereg("^".$sid." (OK|BAD|NO)(.*)$", $read, $regs)) {
 485               return;
 486           } else {
 487               echo decodeBody($read1, $encoding) .
 488                    decodeBody($read, $encoding);
 489           }
 490       } else if ($cnt) {
 491           echo decodeBody($read, $encoding);
 492       }
 493       $read = fgets ($imap_stream,4096);
 494       $cnt++;
 495     }
 496 }
 497
 498 /* -[ END MIME DECODING ]----------------------------------------------------------- */
 499
 500
 501
 502 /* This is the first function called.  It decides if this is a multipart
 503    message or if it should be handled as a single entity
 504  */
 505 function decodeMime ($imap_stream, &$header) {
 506     global $username, $key, $imapServerAddress, $imapPort;
 507     return mime_structure ($imap_stream, $header);
 508 }
 509
 510 // This is here for debugging purposese.  It will print out a list
 511 // of all the entity IDs that are in the $message object.
 512 /*
 513 function listEntities ($message) {
 514 if ($message) {
 515  if ($message->header->entity_id)
 516  echo "<tt>" . $message->header->entity_id . ' : ' . $message->header->type0 . '/' . $message->header->type1 . '<br>';
 517  for ($i = 0; $message->entities[$i]; $i++) {
 518     $msg = listEntities($message->entities[$i], $ent_id);
 519     if ($msg)
 520        return $msg;
 521  }
 522 }
 523 }
 524 */
 525
 526 /* returns a $message object for a particular entity id */
 527 function getEntity ($message, $ent_id) {
 528     if ($message) {
 529         if ($message->header->entity_id == $ent_id && strlen($ent_id) == strlen($message->header->entity_id)) {
 530             return $message;
 531         } else {
 532             for ($i = 0; isset($message->entities[$i]); $i++) {
 533                 $msg = getEntity ($message->entities[$i], $ent_id);
 534                 if ($msg) {
 535                     return $msg;
 536                 }
 537             }
 538         }
 539     }
 540 }
 541
 542 /*
 543  * figures out what entity to display and returns the $message object
 544  * for that entity.
 545  */
 546 function findDisplayEntity ($message, $textOnly = 1)   {
 547     global $show_html_default;
 548
 549     $entity = 0;
 550
 551     if ($message) {
 552         if ( $message->header->type0 == 'multipart' &&
 553              ( $message->header->type1 == 'alternative' ||
 554                $message->header->type1 == 'related' ) &&
 555              $show_html_default && ! $textOnly ) {
 556             $entity = findDisplayEntityHTML($message);
 557         }
 558
 559         // Show text/plain or text/html -- the first one we find.
 560         if ( $entity == 0 &&
 561              $message->header->type0 == 'text' &&
 562              ( $message->header->type1 == 'plain' ||
 563                $message->header->type1 == 'html' ) &&
 564              isset($message->header->entity_id) ) {
 565             $entity = $message->header->entity_id;
 566         }
 567
 568         $i = 0;
 569         while ($entity == 0 && isset($message->entities[$i]) ) {
 570             $entity = findDisplayEntity($message->entities[$i], $textOnly);
 571             $i++;
 572         }
 573     }
 574
 575     return( $entity );
 576 }
 577
 578 /* Shows the HTML version */
 579 function findDisplayEntityHTML ($message) {
 580
 581     if ( $message->header->type0 == 'text' &&
 582          $message->header->type1 == 'html' &&
 583          isset($message->header->entity_id)) {
 584         return $message->header->entity_id;
 585     }
 586     for ($i = 0; isset($message->entities[$i]); $i ++) {
 587         $entity = findDisplayEntityHTML($message->entities[$i]);
 588         if ($entity != 0) {
 589             return $entity;
 590         }
 591     }
 592
 593     return 0;
 594 }
 595
 596 /* This returns a parsed string called $body. That string can then
 597 be displayed as the actual message in the HTML. It contains
 598 everything needed, including HTML Tags, Attachments at the
 599 bottom, etc.
 600 */
 601 function formatBody($imap_stream, $message, $color, $wrap_at) {
 602     // this if statement checks for the entity to show as the
 603     // primary message. To add more of them, just put them in the
 604     // order that is their priority.
 605     global $startMessage, $username, $key, $imapServerAddress, $imapPort,
 606            $show_html_default;
 607
 608     $id = $message->header->id;
 609     $urlmailbox = urlencode($message->header->mailbox);
 610
 611     // Get the right entity and redefine message to be this entity
 612     // Pass the 0 to mean that we want the 'best' viewable one
 613     $ent_num = findDisplayEntity ($message, 0);
 614     $body_message = getEntity($message, $ent_num);
 615     if (($body_message->header->type0 == 'text') ||
 616         ($body_message->header->type0 == 'rfc822')) {
 617
 618         $body = mime_fetch_body ($imap_stream, $id, $ent_num);
 619         $body = decodeBody($body, $body_message->header->encoding);
 620         $hookResults = do_hook("message_body", $body);
 621         $body = $hookResults[1];
 622
 623         // If there are other types that shouldn't be formatted, add
 624         // them here
 625         if ($body_message->header->type1 == 'html') {
 626             if ( $show_html_default <> 1 ) {
 627                 $body = strip_tags( $body );
 628                 translateText($body, $wrap_at, $body_message->header->charset);
 629             } else {
 630                 $body = MagicHTML( $body, $id );
 631             }
 632         } else {
 633             translateText($body, $wrap_at, $body_message->header->charset);
 634         }
 635
 636         $body .= "<SMALL><CENTER><A HREF=\"../src/download.php?absolute_dl=true&passed_id=$id&passed_ent_id=$ent_num&mailbox=$urlmailbox&showHeaders=1\">". _("Download this as a file") ."</A></CENTER><BR></SMALL>";
 637
 638         /** Display the ATTACHMENTS: message if there's more than one part **/
 639         $body .= "</TD></TR></TABLE>";
 640         if (isset($message->entities[0])) {
 641             $body .= formatAttachments ($message, $ent_num, $message->header->mailbox, $id);
 642         }
 643         $body .= "</TD></TR></TABLE>";
 644     } else {
 645         $body = formatAttachments ($message, -1, $message->header->mailbox, $id);
 646     }
 647     return ($body);
 648 }
 649
 650 /*
 651  * A recursive function that returns a list of attachments with links
 652  * to where to download these attachments
 653  */
 654 function formatAttachments ($message, $ent_id, $mailbox, $id) {
 655   global $where, $what;
 656   global $startMessage, $color;
 657   static $ShownHTML = 0;
 658
 659   $body = "";
 660   if ($ShownHTML == 0) {
 661         $ShownHTML = 1;
 662
 663         $body .= "<TABLE WIDTH=\"100%\" CELLSPACING=0 CELLPADDING=2 BORDER=0 BGCOLOR=\"$color[0]\"><TR>\n" .
 664                  "<TH ALIGN=\"left\" BGCOLOR=\"$color[9]\"><B>\n" .
 665                  _("Attachments") . ':' .
 666                  "</B></TH></TR><TR><TD>\n" .
 667                  "<TABLE CELLSPACING=0 CELLPADDING=1 BORDER=0>\n" .
 668                  formatAttachments ($message, $ent_id, $mailbox, $id) .
 669                  "</TABLE></TD></TR></TABLE>";
 670
 671         return( $body );
 672   }
 673
 674   if ($message) {
 675      if (!$message->entities) {
 676         $type0 = strtolower($message->header->type0);
 677         $type1 = strtolower($message->header->type1);
 678         $name = decodeHeader($message->header->name);
 679
 680         if ($message->header->entity_id != $ent_id) {
 681            $filename = decodeHeader($message->header->filename);
 682            if (trim($filename) == '') {
 683               if (trim($name) == '') {
 684                  if ( trim( $message->header->id ) == '' )
 685                     $display_filename = 'untitled-[' . $message->header->entity_id . ']' ;
 686                  else
 687                     $display_filename = 'cid: ' . $message->header->id;
 688                  // $display_filename = 'untitled-[' . $message->header->entity_id . ']' ;
 689               } else {
 690                  $display_filename = $name;
 691                  $filename = $name;
 692               }
 693            } else {
 694               $display_filename = $filename;
 695            }
 696
 697            $urlMailbox = urlencode($mailbox);
 698            $ent = urlencode($message->header->entity_id);
 699
 700            $DefaultLink =
 701               "../src/download.php?startMessage=$startMessage&passed_id=$id&mailbox=$urlMailbox&passed_ent_id=$ent";
 702            if ($where && $what)
 703               $DefaultLink .= '&where=' . urlencode($where) . '&what=' . urlencode($what);
 704            $Links['download link']['text'] = _("download");
 705            $Links['download link']['href'] =
 706                "../src/download.php?absolute_dl=true&passed_id=$id&mailbox=$urlMailbox&passed_ent_id=$ent";
 707            $ImageURL = '';
 708
 709            /* this executes the attachment hook with a specific MIME-type.
 710             * if that doens't have results, it tries if there's a rule
 711             * for a more generic type. */
 712            $HookResults = do_hook("attachment $type0/$type1", $Links,
 713                $startMessage, $id, $urlMailbox, $ent, $DefaultLink,
 714                $display_filename, $where, $what);
 715            if(count($HookResults[1]) <= 1) {
 716                $HookResults = do_hook("attachment $type0/*", $Links,
 717                $startMessage, $id, $urlMailbox, $ent, $DefaultLink,
 718                $display_filename, $where, $what);
 719            }
 720
 721            $Links = $HookResults[1];
 722            $DefaultLink = $HookResults[6];
 723
 724            $body .= '<TR><TD>&nbsp;&nbsp;</TD><TD>' .
 725                     "<A HREF=\"$DefaultLink\">$display_filename</A>&nbsp;</TD>" .
 726                     '<TD><SMALL><b>' . show_readable_size($message->header->size) .
 727                     '</b>&nbsp;&nbsp;</small></TD>' .
 728                     "<TD><SMALL>[ $type0/$type1 ]&nbsp;</SMALL></TD>" .
 729                     '<TD><SMALL>';
 730            if ($message->header->description)
 731               $body .= '<b>' . htmlspecialchars($message->header->description) . '</b>';
 732            $body .= '</SMALL></TD><TD><SMALL>&nbsp;';
 733
 734
 735            $SkipSpaces = 1;
 736            foreach ($Links as $Val) {
 737               if ($SkipSpaces) {
 738                  $SkipSpaces = 0;
 739               } else {
 740                  $body .= '&nbsp;&nbsp;|&nbsp;&nbsp;';
 741               }
 742               $body .= '<a href="' . $Val['href'] . '">' .  $Val['text'] . '</a>';
 743            }
 744
 745            unset($Links);
 746
 747            $body .= "</SMALL></TD></TR>\n";
 748         }
 749      } else {
 750         for ($i = 0; $i < count($message->entities); $i++) {
 751            $body .= formatAttachments ($message->entities[$i], $ent_id, $mailbox, $id);
 752         }
 753      }
 754      return( $body );
 755   }
 756 }
 757
 758
 759 /** this function decodes the body depending on the encoding type. **/
 760 function decodeBody($body, $encoding) {
 761   $body = str_replace("\r\n", "\n", $body);
 762   $encoding = strtolower($encoding);
 763
 764   global $show_html_default;
 765
 766   if ($encoding == 'quoted-printable') {
 767      $body = quoted_printable_decode($body);
 768
 769
 770      while (ereg("=\n", $body))
 771         $body = ereg_replace ("=\n", "", $body);
 772
 773   } else if ($encoding == 'base64') {
 774      $body = base64_decode($body);
 775   }
 776
 777   // All other encodings are returned raw.
 778   return $body;
 779 }
 780
 781 /*
 782  * This functions decode strings that is encoded according to
 783  * RFC1522 (MIME Part Two: Message Header Extensions for Non-ASCII Text).
 784  */
 785 function decodeHeader ($string, $utfencode=true) {
 786
 787 if ( is_array( $string ) ) {
 788     $string = implode("\n", $string );
 789 }
 790
 791 if (eregi('=\\?([^?]+)\\?(q|b)\\?([^?]+)\\?=',
 792             $string, $res)) {
 793     if (ucfirst($res[2]) == 'B') {
 794         $replace = base64_decode($res[3]);
 795     } else {
 796         $replace = str_replace('_', ' ', $res[3]);
 797         // Convert lowercase Quoted Printable to uppercase for
 798         // quoted_printable_decode to understand it.
 799         while (ereg("(=(([0-9][abcdef])|([abcdef][0-9])|([abcdef][abcdef])))",
 800                $replace, $res)) {
 801             $replace = str_replace($res[1], strtoupper($res[1]), $replace);
 802         }
 803         $replace = quoted_printable_decode($replace);
 804     }
 805     /* Only encode into entities by default. Some places
 806         don't need the encoding, like the compose form. */
 807     if ($utfencode){
 808         $replace = charset_decode ($res[1], $replace);
 809     }
 810
 811     // Remove the name of the character set.
 812     $string = eregi_replace ('=\\?([^?]+)\\?(q|b)\\?([^?]+)\\?=',
 813               $replace, $string);
 814
 815     // In case there should be more encoding in the string: recurse
 816     $string = decodeHeader($string);
 817 }
 818
 819 return ($string);
 820 }
 821
 822 /*
 823  * Encode a string according to RFC 1522 for use in headers if it
 824  * contains 8-bit characters or anything that looks like it should
 825  * be encoded.
 826  */
 827 function encodeHeader ($string) {
 828     global $default_charset;
 829
 830     // Encode only if the string contains 8-bit characters or =?
 831     $j = strlen( $string  );
 832     $l = strstr($string, '=?');         // Must be encoded ?
 833     $ret = '';
 834     for( $i=0; $i < $j; ++$i) {
 835         switch( $string{$i} ) {
 836            case '=':
 837           $ret .= '=3D';
 838           break;
 839         case '?':
 840           $ret .= '=3F';
 841           break;
 842         case '_':
 843           $ret .= '=5F';
 844           break;
 845         case ' ':
 846           $ret .= '_';
 847           break;
 848         default:
 849           $k = ord( $string{$i} );
 850           if ( $k > 126 ) {
 851              $ret .= sprintf("=%02X", $k);
 852              $l = TRUE;
 853           } else
 854              $ret .= $string{$i};
 855         }
 856     }
 857
 858     if ( $l ) {
 859         $string = "=?$default_charset?Q?$ret?=";
 860     }
 861
 862     return( $string );
 863 }
 864
 865 /*
 866     Strips dangerous tags from html messages.
 867 */
 868 function MagicHTML( $body, $id ) {
 869
 870     global $message, $HTTP_SERVER_VARS,
 871            $attachment_common_show_images;
 872
 873     $attachment_common_show_images =
 874                      FALSE; // Don't display attached images in HTML mode
 875     $j = strlen( $body );   // Legnth of the HTML
 876     $ret = '';              // Returned string
 877     $bgcolor = '#ffffff';   // Background style color (defaults to white)
 878     $textcolor = '#000000'; // Foreground style color (defaults to black)
 879     $leftmargin = '';       // Left margin style
 880     $title = '';            // HTML title if any
 881
 882     $i = 0;
 883     while ( $i < $j ) {
 884         if ( $body{$i} == '<' ) {
 885             $pos = $i + 1;
 886             $tag = '';
 887             while ($body{$pos} == ' ' || $body{$pos} == "\t" ||
 888                    $body{$pos} == "\n") {
 889                 $pos ++;
 890             }
 891             while (strlen($tag) < 4 && $body{$pos} != ' ' &&
 892                    $body{$pos} != "\t" && $body{$pos} != "\n") {
 893                 $tag .= $body{$pos};
 894                 $pos ++;
 895             }
 896             switch( strtoupper( $tag ) ) {
 897             // Strips the entire tag and contents
 898             case 'APPL':
 899             case 'EMBB':
 900             case 'FRAM':
 901             case 'SCRI':
 902             case 'OBJE':
 903                 $etg = '/' . $tag;
 904                 while ( $body{$i+1}.$body{$i+2}.$body{$i+3}.$body{$i+4}.$body{$i+5} <> $etg  &&
 905                        $i < $j  ) $i++;
 906                 while ( $i < $j && $body{++$i} <> '>' );
 907                 // $ret .= "<!-- $tag removed -->";
 908                 break;
 909             // Substitute Title
 910             case 'TITL':
 911                 $i += 5;
 912                 while ( $body{$i} <> '>' &&  // </title>
 913                        $i < $j )
 914                         $i++;
 915                 $i++;
 916                 $title = '';
 917                 while ( $body{$i} <> '<' &&  // </title>
 918                        $i < $j ) {
 919                     $title .= $body{$i};
 920                     $i++;
 921                 }
 922                 $i += 7;
 923                 break;
 924             // Destroy these tags
 925             case 'HTML':
 926             case 'HEAD':
 927             case '/HTM':
 928             case '/HEA':
 929             case '!DOC':
 930             case 'META':
 931             //case 'DIV ':
 932             //case '/DIV':
 933             case '!-- ':
 934                 $i += 4;
 935                 while ( $body{$i}  <> '>' &&
 936                        $i < $j )
 937                     $i++;
 938                 // $i++;
 939                 break;
 940             case 'STYL':
 941                 $i += 5;
 942                 while ( $body{$i} <> '>' &&  // </title>
 943                        $i < $j )
 944                         $i++;
 945                 $i++;
 946                 // We parse the style to look for interesting stuff
 947                 $styleblk = '';
 948                 while ( $body{$i} <> '>' &&
 949                        $i < $j ) {
 950                     // First we get the name of the style
 951                     $style = '';
 952                     while ( $body{$i} <> '>' &&
 953                            $body{$i} <> '<' &&
 954                            $body{$i} <> '{' &&
 955                            $i < $j ) {
 956                        if ( isnoSep( $body{$i} ) )
 957                            $style .= $body{$i};
 958                        $i++;
 959                     }
 960                     stripComments( $i, $j, $body );
 961                     $style = strtoupper( trim( $style ) );
 962                     if ( $style == 'BODY' ) {
 963                         // Next we look into the definitions of the body style
 964                         while ( $body{$i} <> '>' &&
 965                                $body{$i} <> '}' &&
 966                                $i < $j ) {
 967                             // We look for the background color if any.
 968                             if ( substr( $body, $i, 17 ) == 'BACKGROUND-COLOR:' ) {
 969                                 $i += 17;
 970                                 $bgcolor = getStyleData( $i, $j, $body );
 971                             } elseif ( substr( $body, $i, 12 ) == 'MARGIN-LEFT:' ) {
 972                                 $i += 12;
 973                                 $leftmargin = getStyleData( $i, $j, $body );
 974                             }
 975                             $i++;
 976                         }
 977                     } else {
 978                         // Other style are mantained
 979                         $styleblk .= "$style ";
 980                         while ( $body{$i} <> '>' &&
 981                                $body{$i} <> '<' &&
 982                                $body{$i} <> '}' &&
 983                                $i < $j ) {
 984                             $styleblk .= $body{$i};
 985                             $i++;
 986                         }
 987                         $styleblk .= $body{$i};
 988                     }
 989                     stripComments( $i, $j, $body );
 990                     if ( $body{$i} <> '>' )
 991                         $i++;
 992                 }
 993                 if ( $styleblk <> '' )
 994                     $ret .= "<style>$styleblk";
 995                 break;
 996             case 'BODY':
 997                 if ( $title <> '' )
 998                     $ret .= '<b>' . _("Title:") . " </b>$title<br>\n";
 999                 $ret .= "<TABLE";
1000                 $i += 5;
1001                 if (! isset($base)) {
1002                     $base = '';
1003                 }
1004                 $ret .= stripEvent( $i, $j, $body, $id, $base );
1005                 $ret .= " bgcolor=$bgcolor width=\"100%\"><tr>";
1006                 if ( $leftmargin <> '' )
1007                     $ret .= "<td width=$leftmargin>&nbsp;</td>";
1008                 $ret .= '<td>';
1009                 if (strtolower($bgcolor) == 'ffffff' ||
1010                     strtolower($bgcolor) == '#ffffff')
1011                     $ret .= '<font color=#000000>';
1012                 break;
1013             case 'BASE':
1014                 $i += 5;
1015                 $base = '';
1016                 while ( !isNoSep( $body{$i} ) &&
1017                        $i < $j ) {
1018                         $i++;
1019                 }
1020                 if ( strcasecmp( substr( $base, 0, 4 ), 'href'  ) ) {
1021                         $i += 5;
1022                         while ( !isNoSep( $body{$i} ) &&
1023                                $i < $j ) {
1024                                 $i++;
1025                         }
1026                         while ( $body{$i} <> '>' &&
1027                                $i < $j ) {
1028                             if ( $body{$i} <> '"' ) {
1029                                 $base .= $body{$i};
1030                             }
1031                             $i++;
1032                         }
1033                         // Debuging $ret .= "<!-- base == $base -->";
1034                         if ( strcasecmp( substr( $base, 0, 4 ), 'file' ) <> 0 ) {
1035                             $ret .= "\n<BASE HREF=\"$base\">\n";
1036                         }
1037                 }
1038                 break;
1039             case '/BOD':
1040                 $ret .= '</font></td></tr></TABLE>';
1041                 $i += 6;
1042                 break;
1043             default:
1044                 // Following tags can contain some event handler, lets search it
1045                 stripComments( $i, $j, $body );
1046                 if (! isset($base)) {
1047                    $base = '';
1048                 }
1049                 $ret .= stripEvent( $i, $j, $body, $id, $base ) . '>';
1050                         // $ret .= "<!-- $tag detected -->";
1051             }
1052         } else {
1053             $ret .= $body{$i};
1054         }
1055     $i++;
1056     }
1057
1058 return( "\n\n<!-- HTML Output ahead -->\n" .
1059         $ret .
1060         "\n<!-- END of HTML Output --><base href=\"".
1061         get_location() . '/'.
1062         "\">\n\n" );
1063 }
1064
1065 function isNoSep( $char ) {
1066
1067     switch( $char ) {
1068     case ' ':
1069     case "\n":
1070     case "\t":
1071     case "\r":
1072     case '>':
1073     case '"':
1074         return( FALSE );
1075         break;
1076     default:
1077         return( TRUE );
1078     }
1079
1080 }
1081
1082 /*
1083   The following function is usefull to remove extra data that can cause
1084   html not to display properly. Especialy with MS stuff.
1085 */
1086
1087 function stripComments( &$i, $j, &$body ) {
1088
1089     while ( $body{$i}.$body{$i+1}.$body{$i+2}.$body{$i+3} == '<!--' &&
1090            $i < $j ) {
1091         $i += 5;
1092         while ( $body{$i-2}.$body{$i-1}.$body{$i} <> '-->' &&
1093                $i < $j )
1094             $i++;
1095         $i++;
1096     }
1097
1098     return;
1099
1100 }
1101
1102 /* Gets the style data of a specific style */
1103
1104 function getStyleData( &$i, $j, &$body ) {
1105
1106     // We skip spaces
1107     while ( $body{$i} <> '>' && !isNoSep( $body{$i} ) &&
1108            $i < $j ) {
1109         $i++;
1110     }
1111     // And get the color
1112     $ret = '';
1113     while ( isNoSep( $body{$i} ) &&
1114            $i < $j ) {
1115         $ret .= $body{$i};
1116         $i++;
1117     }
1118
1119     return( $ret );
1120 }
1121
1122 /*
1123 Private function for strip_dangerous_tag. Look for event based coded and "remove" it
1124 change on with no (onload -> noload)
1125 */
1126
1127 function stripEvent( &$i, $j, &$body, $id, $base ) {
1128
1129     global $message, $base_uri;
1130
1131     $ret = '';
1132
1133     while ( $body{$i} <> '>' &&
1134            $i < $j ) {
1135         $etg = strtolower($body{$i}.$body{$i+1}.$body{$i+2});
1136         switch( $etg ) {
1137         case 'src':
1138             // This is probably a src specification
1139             $k = $i + 3;
1140             while( !isNoSep( $body{$k} )) {
1141                 $k++;
1142             }
1143             if ( $body{$k} == '=' ) {
1144                 /* It is indeed */
1145                 $k++;
1146                 while( !isNoSep( $body{$k} ) &&
1147                        $k < $j ) {
1148                     $k++;
1149                 }
1150                 $src = '';
1151                 while ( $body{$k} <> '>' && isNoSep( $body{$k} ) &&
1152                        $k < $j ) {
1153                     $src .= $body{$k};
1154                     $k++;
1155                 }
1156                 while( !isNoSep( $body{$k} ) &&
1157                        $k < $j ) {
1158                     $k++;
1159                 }
1160                 if ( strtolower( substr( $src, 0, 4 ) ) == 'cid:' ) {
1161                     $src = substr( $src, 4 );
1162                     $src = "../src/download.php?absolute_dl=true&passed_id=$id&mailbox=" .
1163                            urlencode( $message->header->mailbox ) .
1164                            "&passed_ent_id=" . find_ent_id( $src, $message );
1165                 } else if ( strtolower( substr( $src, 0, 4 ) ) <> 'http' ||
1166                             stristr( $src, $base_uri ) ) {
1167                     /* Javascript and local urls goes out */
1168                     $src = '../images/' . _("sec_remove_eng.png");
1169                 }
1170                 $ret .= 'src="' . $src . '" ';
1171                 $i = $k - 2;
1172             } else {
1173                 $ret .= 'src';
1174                 $i = $i + 3;
1175             }
1176
1177             break;
1178         case '../':
1179             // Retrolinks are not allowed without a base because they mess with SM security
1180             if ( $base == '' ) {
1181                     $i += 2;
1182             } else {
1183                     $ret .= '.';
1184             }
1185             break;
1186         case 'cid':
1187             // Internal link
1188             $k = $i-1;
1189             if ( $body{$i+3} == ':') {
1190                 $i +=4;
1191                 $name = '';
1192                 while ( isNoSep( $body{$i} ) &&
1193                        $i < $j  ) {
1194                     $name .= $body{$i++};
1195                 }
1196                 if ( $name <> '' ) {
1197                     $ret .= "../src/download.php?absolute_dl=true&passed_id=$id&mailbox=" .
1198                                 urlencode( $message->header->mailbox ) .
1199                                 "&passed_ent_id=" . find_ent_id( $name, $message );
1200                     if ( $body{$k} == '"' )
1201                         $ret .= '" ';
1202                     else
1203                         $ret .= ' ';
1204                 }
1205                 if ( $body{$i} == '>' )
1206                     $i -= 1;
1207             }
1208             break;
1209         case ' on':
1210         case "\non":
1211         case "\ron":
1212         case "\ton":
1213             $ret .= ' no';
1214             $i += 2;
1215             break;
1216         case 'pt:':
1217             if ( strcasecmp( $body{$i-4}.$body{$i-3}.$body{$i-2}.$body{$i-1}.$body{$i}.$body{$i+1}.$body{$i+2}, 'script:') == 0 ) {
1218                 $ret .= '_no/';
1219             } else {
1220                 $ret .= $etg;
1221             }
1222             $i += 2;
1223             break;
1224         default:
1225             $ret .= $body{$i};
1226         }
1227         $i++;
1228     }
1229     return( $ret );
1230 }
1231
1232
1233 /* This function trys to locate the entity_id of a specific mime element */
1234
1235 function find_ent_id( $id, $message ) {
1236
1237     $ret = '';
1238     for ($i=0; $ret == '' && $i < count($message->entities); $i++) {
1239
1240         if ( $message->entities[$i]->header->entity_id == '' ) {
1241             $ret = find_ent_id( $id, $message->entities[$i] );
1242         } else {
1243             if ( strcasecmp( $message->entities[$i]->header->id, $id ) == 0 )
1244                 $ret = $message->entities[$i]->header->entity_id;
1245         }
1246
1247     }
1248
1249     return( $ret );
1250
1251 }
1252 ?>