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