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, $type = '';
  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 = '', $entities = array();
  37
  38     function addEntity ($msg) {
  39         $this->entities[] = $msg;
  40     }
  41 }
  42
  43 /* --------------------------------------------------------------------------------- */
  44 /* MIME DECODING                                                                     */
  45 /* --------------------------------------------------------------------------------- */
  46
  47 /* This function gets the structure of a message and stores it in the "message" class.
  48  * It will return this object for use with all relevant header information and
  49  * fully parsed into the standard "message" object format.
  50  */
  51 function mime_structure ($imap_stream, $header) {
  52
  53     $ssid = sqimap_session_id();
  54     $lsid = strlen( $ssid );
  55     $id = $header->id;
  56     fputs ($imap_stream, "$ssid FETCH $id BODYSTRUCTURE\r\n");
  57     //
  58     // This should use sqimap_read_data instead of reading it itself
  59     //
  60     $read = fgets ($imap_stream, 9216);
  61     $bodystructure = '';
  62     while ( substr($read, 0, $lsid) <> $ssid &&
  63          !feof( $imap_stream ) ) {
  64         $bodystructure .= $read;
  65         $read = fgets ($imap_stream, 9216);
  66     }
  67     $read = $bodystructure;
  68
  69     // isolate the body structure and remove beginning and end parenthesis
  70     $read = trim(substr ($read, strpos(strtolower($read), 'bodystructure') + 13));
  71
  72     $read = trim(substr ($read, 0, -1));
  73     $end = mime_match_parenthesis(0, $read);
  74     while ($end == strlen($read)-1) {
  75         $read = trim(substr ($read, 0, -1));
  76         $read = trim(substr ($read, 1));
  77         $end = mime_match_parenthesis(0, $read);
  78     }
  79
  80     $msg = mime_parse_structure ($read, 0);
  81     $msg->header = $header;
  82
  83     return( $msg );
  84 }
  85
  86 /* this starts the parsing of a particular structure.  It is called recursively,
  87  * so it can be passed different structures.  It returns an object of type
  88  * $message.
  89  * First, it checks to see if it is a multipart message.  If it is, then it
  90  * handles that as it sees is necessary.  If it is just a regular entity,
  91  * then it parses it and adds the necessary header information (by calling out
  92  * to mime_get_elements()
  93  */
  94 function mime_parse_structure ($structure, $ent_id) {
  95   global $mailbox;
  96   $properties = array();
  97   $msg = new message();
  98   if ($structure{0} == '(') {
  99      $old_ent_id = $ent_id;
 100      $ent_id = mime_new_element_level($ent_id);
 101      $start = $end = -1;
 102      do {
 103         $start = $end+1;
 104         $end = mime_match_parenthesis ($start, $structure);
 105
 106         /* check if we are dealing with a new entity-level */
 107         $i = strrpos($ent_id,'.');
 108         if ($i>0) {
 109             $ent = substr($ent_id, $i+1);
 110         } else {
 111             $ent = '';
 112         }
 113         /* add "forgotten"  parent entities (alternative and relative) */
 114         if ($ent == '0') {
 115             /* new entity levels have information about the type (type1) and
 116             *  the properties. This information is situated at the end of the
 117             *  structure string like for example (example between the brackets)
 118             *  [ "RELATED" ("BOUNDARY" "myboundary" "TYPE" "plain/html") ]
 119             */
 120
 121             /* get the involved properties for parsing to mime_get_properties */
 122             $startprop = strrpos($structure,'(');
 123             $properties_str = substr($structure,$startprop);
 124             $endprop = mime_match_parenthesis ($startprop, $structure);
 125             $propstr = substr($structure, $startprop + 1, ($endprop - $startprop)-1);
 126             /* cut off the used properties */
 127             if ($startprop) {
 128                 $structure_end = substr($structure, $endprop+2);
 129                 $structure = trim(substr($structure,0,$startprop));
 130             }
 131
 132             /* get type1 */
 133             $pos = strrpos($structure,' ');
 134             $type1 = strtolower(substr($structure, $pos+2, (count($structure)-2)));
 135
 136             /* cut off  type1 */
 137             if ($pos && $startprop) {
 138                 $structure = trim(substr($structure, 0, $pos));
 139             }
 140
 141             /* process the found information */
 142             $properties = mime_get_props($properties, $properties_str);
 143             if (count($properties)>0) {
 144                 $msg->header->entity_id = $old_ent_id;
 145                 $msg->header->type0 = 'multipart';
 146                 $msg->header->type1 = $type1;
 147                 for ($i=0; $i < count($properties); $i++) {
 148                     $msg->header->{$properties[$i]['name']} = $properties[$i]['value'];
 149                 }
 150             }
 151             $structure = $structure . ' ' . $structure_end;
 152         }
 153         $element = substr($structure, $start+1, ($end - $start)-1);
 154         $ent_id = mime_increment_id ($ent_id);
 155         $newmsg = mime_parse_structure ($element, $ent_id);
 156         /* set mailbox in case of message/rfc822 entities */
 157         if (isset($newmsg->header->type0) && isset($newmsg->header->type1)) {
 158             if ($newmsg->header->type0 == 'message' && $newmsg->header->type1 == 'rfc822') {
 159                 $newmsg->header->mailbox=$mailbox;
 160             }
 161         }
 162         $msg->addEntity ($newmsg);
 163
 164      } while ($structure{$end+1} == '(');
 165   } else {
 166      // parse the elements
 167     $msg = mime_get_element ($structure, $msg, $ent_id);
 168   }
 169   return $msg;
 170 }
 171
 172
 173 /* Increments the element ID.  An element id can look like any of
 174  * the following:  1, 1.2, 4.3.2.4.1, etc.  This function increments
 175  * the last number of the element id, changing 1.2 to 1.3.
 176  */
 177 function mime_increment_id ($id) {
 178
 179     if (strpos($id, '.')) {
 180         $first = substr($id, 0, strrpos($id, '.'));
 181         $last = substr($id, strrpos($id, '.')+1);
 182         $last++;
 183         $new = $first . '.' .$last;
 184     } else {
 185         $new = $id + 1;
 186     }
 187
 188     return $new;
 189 }
 190
 191 /*
 192  * See comment for mime_increment_id().
 193  * This adds another level on to the entity_id changing 1.3 to 1.3.0
 194  * NOTE:  1.3.0 is not a valid element ID.  It MUST be incremented
 195  *        before it can be used.  I left it this way so as not to have
 196  *        to make a special case if it is the first entity_id.  It
 197  *        always increments it, and that works fine.
 198  */
 199 function mime_new_element_level ($id) {
 200
 201     if (!$id) {
 202         $id = 0;
 203     } else {
 204         $id = $id . '.0';
 205     }
 206
 207     return( $id );
 208 }
 209
 210 function mime_get_element (&$structure, $msg, $ent_id) {
 211
 212   $elem_num = 1;
 213   $msg->header = new msg_header();
 214   $msg->header->entity_id = $ent_id;
 215   $properties = array();
 216   while (strlen($structure) > 0) {
 217      $structure = trim($structure);
 218      $char = $structure{0};
 219
 220      if (strtolower(substr($structure, 0, 3)) == 'nil') {
 221         $text = '';
 222         $structure = substr($structure, 3);
 223      } else if ($char == '"') {
 224         // loop through until we find the matching quote, and return that as a string
 225         $pos = 1;
 226         $text = '';
 227         while ( ($char = $structure{$pos} ) <> '"' && $pos < strlen($structure)) {
 228            $text .= $char;
 229            $pos++;
 230         }
 231         $structure = substr($structure, strlen($text) + 2);
 232      } else if ($char == '{') {
 233          /**
 234           * loop through until we find the matching quote,
 235           * and return that as a string
 236           */
 237          $pos = 1;
 238          $len = '';
 239          while (($char = $structure{$pos}) != '}'
 240                 && $pos < strlen($structure)) {
 241              $len .= $char;
 242              $pos++;
 243          }
 244          $structure = substr($structure, strlen($len) + 4);
 245          $text = substr($structure, 0, $len);
 246          $structure = substr($structure, $len + 1);
 247      } else if ($char == '(') {
 248         // comment me
 249         $end = mime_match_parenthesis (0, $structure);
 250         $sub = substr($structure, 1, $end-1);
 251         $properties = mime_get_props($properties, $sub);
 252         $structure = substr($structure, strlen($sub) + 2);
 253      } else {
 254         // loop through until we find a space or an end parenthesis
 255         $pos = 0;
 256         $char = $structure{$pos};
 257         $text = '';
 258         while ($char != ' ' && $char != ')' && $pos < strlen($structure)) {
 259            $text .= $char;
 260            $pos++;
 261            $char = $structure{$pos};
 262         }
 263         $structure = substr($structure, strlen($text));
 264      }
 265
 266      // This is where all the text parts get put into the header
 267      switch ($elem_num) {
 268         case 1:
 269            $msg->header->type0 = strtolower($text);
 270            break;
 271         case 2:
 272            $msg->header->type1 = strtolower($text);
 273            break;
 274         case 4: // Id
 275            // Invisimail enclose images with <>
 276            $msg->header->id = str_replace( '<', '', str_replace( '>', '', $text ) );
 277            break;
 278         case 5:
 279            $msg->header->description = $text;
 280            break;
 281         case 6:
 282            $msg->header->encoding = strtolower($text);
 283            break;
 284         case 7:
 285            $msg->header->size = $text;
 286            break;
 287         default:
 288            if ($msg->header->type0 == 'text' && $elem_num == 8) {
 289               // This is a plain text message, so lets get the number of lines
 290               // that it contains.
 291               $msg->header->num_lines = $text;
 292
 293            } else if ($msg->header->type0 == 'message' && $msg->header->type1 == 'rfc822' && $elem_num == 8) {
 294               // This is an encapsulated message, so lets start all over again and
 295               // parse this message adding it on to the existing one.
 296               $structure = trim($structure);
 297               if ( $structure{0} == '(' ) {
 298                  $e = mime_match_parenthesis (0, $structure);
 299                  $structure = substr($structure, 0, $e);
 300                  $structure = substr($structure, 1);
 301                  $m = mime_parse_structure($structure, $msg->header->entity_id);
 302
 303                  // the following conditional is there to correct a bug that wasn't
 304                  // incrementing the entity IDs correctly because of the special case
 305                  // that message/rfc822 is.  This fixes it fine.
 306                  if (substr($structure, 1, 1) != '(')
 307                     $m->header->entity_id = mime_increment_id(mime_new_element_level($ent_id));
 308
 309                  // Now we'll go through and reformat the results.
 310                  if ($m->entities) {
 311                     for ($i=0; $i < count($m->entities); $i++) {
 312                        $msg->addEntity($m->entities[$i]);
 313                     }
 314                  } else {
 315                     $msg->addEntity($m);
 316                  }
 317                  $structure = "";
 318               }
 319            }
 320            break;
 321      }
 322      $elem_num++;
 323      $text = "";
 324   }
 325   // loop through the additional properties and put those in the various headers
 326   for ($i=0; $i < count($properties); $i++) {
 327      $msg->header->{$properties[$i]['name']} = $properties[$i]['value'];
 328   }
 329
 330   return $msg;
 331 }
 332
 333 /*
 334  * I did most of the MIME stuff yesterday (June 20, 2000), but I couldn't
 335  * figure out how to do this part, so I decided to go to bed.  I woke up
 336  * in the morning and had a flash of insight.  I went to the white-board
 337  * and scribbled it out, then spent a bit programming it, and this is the
 338  * result.  Nothing complicated, but I think my brain was fried yesterday.
 339  * Funny how that happens some times.
 340  *
 341  * This gets properties in a nested parenthesisized list.  For example,
 342  * this would get passed something like:  ("attachment" ("filename" "luke.tar.gz"))
 343  * This returns an array called $props with all paired up properties.
 344  * It ignores the "attachment" for now, maybe that should change later
 345  * down the road.  In this case, what is returned is:
 346  *    $props[0]["name"] = "filename";
 347  *    $props[0]["value"] = "luke.tar.gz";
 348  */
 349 function mime_get_props ($props, $structure) {
 350
 351   while (strlen($structure) > 0) {
 352      $structure = trim($structure);
 353      $char = $structure{0};
 354      if ($char == '"') {
 355         $pos = 1;
 356         $tmp = '';
 357         while ( ( $char = $structure{$pos} ) != '"' &&
 358                 $pos < strlen($structure)) {
 359            $tmp .= $char;
 360            $pos++;
 361         }
 362         $structure = trim(substr($structure, strlen($tmp) + 2));
 363         $char = $structure{0};
 364
 365         if ($char == '"') {
 366            $pos = 1;
 367            $value = '';
 368            while ( ( $char = $structure{$pos} ) != '"' &&
 369                    $pos < strlen($structure) ) {
 370               $value .= $char;
 371               $pos++;
 372            }
 373            $structure = trim(substr($structure, strlen($value) + 2));
 374            $k = count($props);
 375            $props[$k]['name'] = strtolower($tmp);
 376            $props[$k]['value'] = $value;
 377            if ($structure != '') {
 378                 mime_get_props($props, $structure);
 379            } else {
 380              return $props;
 381            }
 382         } else if ($char == '(') {
 383            $end = mime_match_parenthesis (0, $structure);
 384            $sub = substr($structure, 1, $end-1);
 385            if (! isset($props))
 386               $props = array();
 387               $props = mime_get_props($props, $sub);
 388               $structure = substr($structure, strlen($sub) + 2);
 389            return $props;
 390         }
 391      } else if ($char == '(') {
 392         $end = mime_match_parenthesis (0, $structure);
 393         $sub = substr($structure, 1, $end-1);
 394         $props = mime_get_props($props, $sub);
 395         $structure = substr($structure, strlen($sub) + 2);
 396         return $props;
 397      } else {
 398         return $props;
 399      }
 400   }
 401 }
 402
 403 /*
 404  *  Matches parenthesis.  It will return the position of the matching
 405  *  parenthesis in $structure.  For instance, if $structure was:
 406  *     ("text" "plain" ("val1name", "1") nil ... )
 407  *     x                                         x
 408  *  then this would return 42 to match up those two.
 409  */
 410 function mime_match_parenthesis ($pos, $structure) {
 411
 412     $j = strlen( $structure );
 413
 414     // ignore all extra characters
 415     // If inside of a string, skip string -- Boundary IDs and other
 416     // things can have ) in them.
 417     if ( $structure{$pos} != '(' ) {
 418         return( $j );
 419     }
 420
 421     while ( $pos < $j ) {
 422         $pos++;
 423         if ($structure{$pos} == ')') {
 424             return $pos;
 425         } elseif ($structure{$pos} == '"') {
 426             $pos++;
 427             while ( $structure{$pos} != '"' &&
 428                     $pos < $j ) {
 429                if (substr($structure, $pos, 2) == '\\"') {
 430                   $pos++;
 431                } elseif (substr($structure, $pos, 2) == '\\\\') {
 432                   $pos++;
 433                }
 434                $pos++;
 435             }
 436         } elseif ( $structure{$pos} == '(' ) {
 437             $pos = mime_match_parenthesis ($pos, $structure);
 438         }
 439     }
 440     echo _("Error decoding mime structure.  Report this as a bug!") . '<br>';
 441     return( $pos );
 442 }
 443
 444 function mime_fetch_body($imap_stream, $id, $ent_id ) {
 445
 446     /*
 447      * do a bit of error correction.  If we couldn't find the entity id, just guess
 448      * that it is the first one.  That is usually the case anyway.
 449      */
 450     if (!$ent_id) {
 451         $ent_id = 1;
 452     }
 453
 454     $cmd = "FETCH $id BODY[$ent_id]";
 455     $data = sqimap_run_command ($imap_stream, $cmd, true, $response, $message);
 456
 457     do {
 458         $topline = trim(array_shift( $data ));
 459     } while( $topline && $topline[0] == '*' && !preg_match( '/\* [0-9]+ FETCH.*/i', $topline )) ;
 460     $wholemessage = implode('', $data);
 461     if (ereg('\\{([^\\}]*)\\}', $topline, $regs)) {
 462
 463         $ret = substr( $wholemessage, 0, $regs[1] );
 464         /*
 465             There is some information in the content info header that could be important
 466             in order to parse html messages. Let's get them here.
 467         */
 468         if ( $ret{0} == '<' ) {
 469             $data = sqimap_run_command ($imap_stream, "FETCH $id BODY[$ent_id.MIME]", true, $response, $message);
 470             /* BASE within HTML documents is illegal (see w3 spec)
 471 *            $base = '';
 472 *            $k = 10;
 473 *            foreach( $data as $d ) {
 474 *                if ( substr( $d, 0, 13 ) == 'Content-Base:' ) {
 475 *                    $j = strlen( $d );
 476 *                    $i = 13;
 477 *                    $base = '';
 478 *                    while ( $i < $j &&
 479 *                           ( !isNoSep( $d{$i} ) || $d{$i} == '"' )  )
 480 *                        $i++;
 481 *                    while ( $i < $j ) {
 482 *                        if ( isNoSep( $d{$i} ) )
 483 *                            $base .= $d{$i};
 484 *                        $i++;
 485 *                    }
 486 *                    $k = 0;
 487 *                } elseif ( $k == 1 && !isnosep( $d{0} ) ) {
 488 *                    $base .= substr( $d, 1 );
 489 *                }
 490 *                $k++;
 491 *            }
 492 *            if ( $base <> '' ) {
 493 *                $ret = "<base href=\"$base\">" . $ret;
 494 *            }
 495 *           */
 496         }
 497     } else if (ereg('"([^"]*)"', $topline, $regs)) {
 498         $ret = $regs[1];
 499     } else {
 500         global $where, $what, $mailbox, $passed_id, $startMessage;
 501         $par = 'mailbox=' . urlencode($mailbox) . "&amp;passed_id=$passed_id";
 502         if (isset($where) && isset($what)) {
 503             $par .= '&amp;where='. urlencode($where) . "&amp;what=" . urlencode($what);
 504         } else {
 505             $par .= "&amp;startMessage=$startMessage&amp;show_more=0";
 506         }
 507         $par .= '&amp;response=' . urlencode($response) .
 508                 '&amp;message=' . urlencode($message).
 509                 '&amp;topline=' . urlencode($topline);
 510
 511         echo   '<tt><br>' .
 512                '<table width="80%"><tr>' .
 513                '<tr><td colspan=2>' .
 514                _("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!") .
 515                " <A HREF=\"../src/retrievalerror.php?$par\"><br>" .
 516                _("Submit message") . '</A><BR>&nbsp;' .
 517                '</td></tr>' .
 518                '<td><b>' . _("Command:") . "</td><td>$cmd</td></tr>" .
 519                '<td><b>' . _("Response:") . "</td><td>$response</td></tr>" .
 520                '<td><b>' . _("Message:") . "</td><td>$message</td></tr>" .
 521                '<td><b>' . _("FETCH line:") . "</td><td>$topline</td></tr>" .
 522                "</table><BR></tt></font><hr>";
 523
 524         $data = sqimap_run_command ($imap_stream, "FETCH $passed_id BODY[]", true, $response, $message);
 525         array_shift($data);
 526         $wholemessage = implode('', $data);
 527
 528         $ret = $wholemessage;
 529     }
 530     return( $ret );
 531 }
 532
 533 function mime_print_body_lines ($imap_stream, $id, $ent_id, $encoding) {
 534     // do a bit of error correction.  If we couldn't find the entity id, just guess
 535     // that it is the first one.  That is usually the case anyway.
 536     if (!$ent_id) {
 537         $ent_id = 1;
 538     }
 539     $sid = sqimap_session_id();
 540     // Don't kill the connection if the browser is over a dialup
 541     // and it would take over 30 seconds to download it.
 542
 543     // don´t call set_time_limit in safe mode.
 544     if (!ini_get("safe_mode")) {
 545         set_time_limit(0);
 546     }
 547
 548     fputs ($imap_stream, "$sid FETCH $id BODY[$ent_id]\r\n");
 549     $cnt = 0;
 550     $continue = true;
 551     $read = fgets ($imap_stream,4096);
 552     // This could be bad -- if the section has sqimap_session_id() . ' OK'
 553     // or similar, it will kill the download.
 554     while (!ereg("^".$sid." (OK|BAD|NO)(.*)$", $read, $regs)) {
 555       if (trim($read) == ')==') {
 556           $read1 = $read;
 557           $read = fgets ($imap_stream,4096);
 558           if (ereg("^".$sid." (OK|BAD|NO)(.*)$", $read, $regs)) {
 559               return;
 560           } else {
 561               echo decodeBody($read1, $encoding) .
 562                    decodeBody($read, $encoding);
 563           }
 564       } else if ($cnt) {
 565           echo decodeBody($read, $encoding);
 566       }
 567       $read = fgets ($imap_stream,4096);
 568       $cnt++;
 569     }
 570 }
 571
 572 /* -[ END MIME DECODING ]----------------------------------------------------------- */
 573
 574
 575
 576 /* This is the first function called.  It decides if this is a multipart
 577    message or if it should be handled as a single entity
 578  */
 579 function decodeMime ($imap_stream, &$header) {
 580     global $username, $key, $imapServerAddress, $imapPort;
 581     return mime_structure ($imap_stream, $header);
 582 }
 583
 584 // This is here for debugging purposese.  It will print out a list
 585 // of all the entity IDs that are in the $message object.
 586
 587 function listEntities ($message) {
 588 if ($message) {
 589  if ($message->header->entity_id)
 590  echo "<tt>" . $message->header->entity_id . ' : ' . $message->header->type0 . '/' . $message->header->type1 . '<br>';
 591  for ($i = 0; $message->entities[$i]; $i++) {
 592     $msg = listEntities($message->entities[$i], $ent_id);
 593     if ($msg)
 594        return $msg;
 595  }
 596 }
 597 }
 598
 599
 600 /* returns a $message object for a particular entity id */
 601 function getEntity ($message, $ent_id) {
 602     if ($message) {
 603         if ($message->header->entity_id == $ent_id && strlen($ent_id) == strlen($message->header->entity_id))
 604         {
 605             return $message;
 606         } else {
 607             for ($i = 0; isset($message->entities[$i]); $i++) {
 608                 $msg = getEntity ($message->entities[$i], $ent_id);
 609                 if ($msg) {
 610                     return $msg;
 611                 }
 612             }
 613         }
 614     }
 615 }
 616
 617 /*
 618  * figures out what entity to display and returns the $message object
 619  * for that entity.
 620  */
 621 function findDisplayEntity ($msg, $textOnly = 1)   {
 622     global $show_html_default;
 623
 624     $entity = 0;
 625
 626     if ($msg) {
 627         if ( $msg->header->type0 == 'multipart' &&
 628              ( $msg->header->type1 == 'alternative' ||
 629                $msg->header->type1 == 'mixed' ||
 630                $msg->header->type1 == 'related' ) &&
 631              $show_html_default && ! $textOnly ) {
 632             $entity = findDisplayEntityHTML($msg);
 633         }
 634
 635         // Show text/plain or text/html -- the first one we find.
 636         if ( $entity == 0 &&
 637              $msg->header->type0 == 'text' &&
 638              ( $msg->header->type1 == 'plain' ||
 639                $msg->header->type1 == 'html' ) &&
 640              isset($msg->header->entity_id) ) {
 641             $entity = $msg->header->entity_id;
 642         }
 643
 644         $i = 0;
 645         while ($entity == 0 && isset($msg->entities[$i]) ) {
 646             $entity = findDisplayEntity($msg->entities[$i], $textOnly);
 647             $i++;
 648         }
 649     }
 650
 651     return( $entity );
 652 }
 653
 654 /* Shows the HTML version */
 655 function findDisplayEntityHTML ($message) {
 656
 657     if ( $message->header->type0 == 'text' &&
 658          $message->header->type1 == 'html' &&
 659          isset($message->header->entity_id)) {
 660         return $message->header->entity_id;
 661     }
 662     for ($i = 0; isset($message->entities[$i]); $i ++) {
 663         if ( $message->header->type0 == 'message' &&
 664             $message->header->type1 == 'rfc822' &&
 665             isset($message->header->entity_id)) {
 666             return 0;
 667         }
 668         $entity = findDisplayEntityHTML($message->entities[$i]);
 669         if ($entity != 0) {
 670             return $entity;
 671         }
 672     }
 673
 674     return 0;
 675 }
 676
 677 /*
 678  * translateText
 679  * Extracted from strings.php 23/03/2002
 680  */
 681
 682 function translateText(&$body, $wrap_at, $charset) {
 683     global $where, $what; /* from searching */
 684     global $color; /* color theme */
 685
 686     require_once('../functions/url_parser.php');
 687
 688     $body_ary = explode("\n", $body);
 689     $PriorQuotes = 0;
 690     for ($i=0; $i < count($body_ary); $i++) {
 691         $line = $body_ary[$i];
 692         if (strlen($line) - 2 >= $wrap_at) {
 693             sqWordWrap($line, $wrap_at);
 694         }
 695         $line = charset_decode($charset, $line);
 696         $line = str_replace("\t", '        ', $line);
 697
 698         parseUrl ($line);
 699
 700         $Quotes = 0;
 701         $pos = 0;
 702         $j = strlen( $line );
 703
 704         while ( $pos < $j ) {
 705             if ($line[$pos] == ' ') {
 706                 $pos ++;
 707             } else if (strpos($line, '&gt;', $pos) === $pos) {
 708                 $pos += 4;
 709                 $Quotes ++;
 710             } else {
 711                 break;
 712             }
 713         }
 714
 715         if ($Quotes > 1) {
 716             if (! isset($color[14])) {
 717                 $color[14] = '#FF0000';
 718             }
 719             $line = '<FONT COLOR="' . $color[14] . '">' . $line . '</FONT>';
 720         } elseif ($Quotes) {
 721             if (! isset($color[13])) {
 722                 $color[13] = '#800000';
 723             }
 724             $line = '<FONT COLOR="' . $color[13] . '">' . $line . '</FONT>';
 725         }
 726
 727         $body_ary[$i] = $line;
 728     }
 729     $body = '<pre>' . implode("\n", $body_ary) . '</pre>';
 730 }
 731
 732 /* debugfunction for looping through entities and displaying correct entities */
 733 function listMyEntities ($message) {
 734
 735 if ($message) {
 736     if ($message->header->entity_id) {
 737         echo "<tt>" . $message->header->entity_id . ' : ' . $message->header->type0 . '/' . $message->header->type1 . '<br>';
 738     }
 739     if (!($message->header->type0 == 'message' &&  $message->header->type1 == 'rfc822')) {
 740         if (isset($message->header->boundary) ) {
 741             $ent_id = $message->header->entity_id;
 742             $var = $message->header->boundary;
 743             if ($var !='')
 744             echo "<b>$ent_id boundary = $var</b><br>";
 745         }
 746         if (isset($message->header->type) ) {
 747             $var = $message->header->type;
 748             if ($var !='')
 749             echo "<b>$ent_id type = $var</b><br>";
 750         }
 751         for ($i = 0; $message->entities[$i]; $i++) {
 752             $msg = listMyEntities($message->entities[$i]);
 753         }
 754
 755         if ($msg )  return $msg;
 756     }
 757 }
 758
 759 }
 760
 761
 762
 763 /* This returns a parsed string called $body. That string can then
 764 be displayed as the actual message in the HTML. It contains
 765 everything needed, including HTML Tags, Attachments at the
 766 bottom, etc.
 767 */
 768 function formatBody($imap_stream, $message, $color, $wrap_at) {
 769     // this if statement checks for the entity to show as the
 770     // primary message. To add more of them, just put them in the
 771     // order that is their priority.
 772     global $startMessage, $username, $key, $imapServerAddress, $imapPort, $body,
 773            $show_html_default, $has_unsafe_images, $view_unsafe_images, $sort;
 774
 775     $has_unsafe_images = 0;
 776
 777     $id = $message->header->id;
 778
 779     $urlmailbox = urlencode($message->header->mailbox);
 780     // Get the right entity and redefine message to be this entity
 781     // Pass the 0 to mean that we want the 'best' viewable one
 782     $ent_num = findDisplayEntity ($message, 0);
 783     $body_message = getEntity($message, $ent_num);
 784     if (($body_message->header->type0 == 'text') ||
 785         ($body_message->header->type0 == 'rfc822')) {
 786         $body = mime_fetch_body ($imap_stream, $id, $ent_num);
 787         $body = decodeBody($body, $body_message->header->encoding);
 788         $hookResults = do_hook("message_body", $body);
 789         $body = $hookResults[1];
 790         // If there are other types that shouldn't be formatted, add
 791         // them here
 792         if ($body_message->header->type1 == 'html') {
 793             if ( $show_html_default <> 1 ) {
 794                 $body = strip_tags( $body );
 795                 translateText($body, $wrap_at, $body_message->header->charset);
 796             } else {
 797                 $body = MagicHTML( $body, $id, $message );
 798             }
 799         } else {
 800             translateText($body, $wrap_at, $body_message->header->charset);
 801         }
 802
 803         $body .= "<CENTER><SMALL><A HREF=\"../src/download.php?absolute_dl=true&amp;passed_id=$id&amp;passed_ent_id=$ent_num&amp;mailbox=$urlmailbox&amp;showHeaders=1\">". _("Download this as a file") ."</A></SMALL></CENTER><BR>";
 804         if ($has_unsafe_images) {
 805             if ($view_unsafe_images) {
 806                 $body .= "<CENTER><SMALL><A HREF=\"read_body.php?passed_id=$id&amp;mailbox=$urlmailbox&amp;sort=$sort&amp;startMessage=$startMessage&amp;show_more=0\">". _("Hide Unsafe Images") ."</A></SMALL></CENTER><BR>\n";
 807             } else {
 808                 $body .= "<CENTER><SMALL><A HREF=\"read_body.php?passed_id=$id&amp;mailbox=$urlmailbox&amp;sort=$sort&amp;startMessage=$startMessage&amp;show_more=0&amp;view_unsafe_images=1\">". _("View Unsafe Images") ."</A></SMALL></CENTER><BR>\n";
 809             }
 810         }
 811
 812         /** Display the ATTACHMENTS: message if there's more than one part **/
 813         if (isset($message->entities[1])) {
 814             /* Header-type alternative means we choose the best one to display
 815                so don't show the alternatives as attachment. Header-type related
 816                means that the attachments are already part of the related message.
 817             */
 818             if ($message->header->type1 !='related' && $message->header->type1 !='alternative') {
 819                 $body .= formatAttachments ($message, $ent_num, $message->header->mailbox, $id);
 820             }
 821         }
 822     } else {
 823         $body = formatAttachments ($message, -1, $message->header->mailbox, $id);
 824     }
 825     return ($body);
 826 }
 827
 828 /*
 829  * A recursive function that returns a list of attachments with links
 830  * to where to download these attachments
 831  */
 832 function formatAttachments($message, $ent_id, $mailbox, $id) {
 833     global $where, $what;
 834     global $startMessage, $color;
 835     static $ShownHTML = 0;
 836
 837     $body = '';
 838     if ($ShownHTML == 0) {
 839
 840         $ShownHTML = 1;
 841         $body .= "<TABLE WIDTH=\"100%\" CELLSPACING=0 CELLPADDING=2 BORDER=0 BGCOLOR=\"$color[0]\"><TR>\n" .
 842                 "<TH ALIGN=\"left\" BGCOLOR=\"$color[9]\"><B>\n" .
 843                 _("Attachments") . ':' .
 844                 "</B></TH></TR><TR><TD>\n" .
 845                 "<TABLE CELLSPACING=0 CELLPADDING=1 BORDER=0>\n" .
 846                 formatAttachments($message, $ent_id, $mailbox, $id) .
 847                 "</TABLE></TD></TR></TABLE>";
 848
 849     } else if ($message) {
 850         $header = $message->header;
 851         $type0 = strtolower($header->type0);
 852         $type1 = strtolower($header->type1);
 853         $name = '';
 854         if (isset($header->name)) {
 855             $name = decodeHeader($header->name);
 856         }
 857         if ($type0 =='message' && $type1 == 'rfc822') {
 858
 859             $filename = decodeHeader($message->header->filename);
 860             if (trim($filename) == '') {
 861                 if (trim($name) == '') {
 862                     $display_filename = 'untitled-[' . $message->header->entity_id . ']' ;
 863                 } else {
 864                     $display_filename = $name;
 865                     $filename = $name;
 866                 }
 867             } else {
 868                 $display_filename = $filename;
 869             }
 870
 871             $urlMailbox = urlencode($mailbox);
 872             $ent = urlencode($message->header->entity_id);
 873
 874             $DefaultLink =
 875                 "../src/download.php?startMessage=$startMessage&amp;passed_id=$id&amp;mailbox=$urlMailbox&amp;passed_ent_id=$ent";
 876             if ($where && $what) {
 877                 $DefaultLink .= '&amp;where=' . urlencode($where) . '&amp;what=' . urlencode($what);
 878             }
 879             $Links['download link']['text'] = _("download");
 880             $Links['download link']['href'] =
 881                 "../src/download.php?absolute_dl=true&amp;passed_id=$id&amp;mailbox=$urlMailbox&amp;passed_ent_id=$ent";
 882             $ImageURL = '';
 883
 884             /* this executes the attachment hook with a specific MIME-type.
 885                 * if that doens't have results, it tries if there's a rule
 886                 * for a more generic type. */
 887             $HookResults = do_hook("attachment $type0/$type1", $Links,
 888                 $startMessage, $id, $urlMailbox, $ent, $DefaultLink, $display_filename, $where, $what);
 889             if(count($HookResults[1]) <= 1) {
 890                 $HookResults = do_hook("attachment $type0/*", $Links,
 891                 $startMessage, $id, $urlMailbox, $ent, $DefaultLink,
 892                 $display_filename, $where, $what);
 893             }
 894
 895             $Links = $HookResults[1];
 896             $DefaultLink = $HookResults[6];
 897
 898             $body .= '<TR><TD>&nbsp;&nbsp;</TD><TD>' .
 899                         "<A HREF=\"$DefaultLink\">$display_filename</A>&nbsp;</TD>" .
 900                         '<TD><SMALL><b>' . show_readable_size($message->header->size) .
 901                         '</b>&nbsp;&nbsp;</small></TD>' .
 902                         "<TD><SMALL>[ $type0/$type1 ]&nbsp;</SMALL></TD>" .
 903                         '<TD><SMALL>';
 904             if ($message->header->description) {
 905                 $body .= '<b>' . htmlspecialchars(_($message->header->description)) . '</b>';
 906             }
 907             $body .= '</SMALL></TD><TD><SMALL>&nbsp;';
 908
 909
 910             $SkipSpaces = 1;
 911             foreach ($Links as $Val) {
 912                 if ($SkipSpaces) {
 913                     $SkipSpaces = 0;
 914                 } else {
 915                     $body .= '&nbsp;&nbsp;|&nbsp;&nbsp;';
 916                 }
 917                 $body .= '<a href="' . $Val['href'] . '">' .  $Val['text'] . '</a>';
 918             }
 919
 920             unset($Links);
 921
 922             $body .= "</SMALL></TD></TR>\n";
 923
 924             return( $body );
 925
 926         } elseif (!$message->entities) {
 927
 928             $type0 = strtolower($message->header->type0);
 929             $type1 = strtolower($message->header->type1);
 930             $name = decodeHeader($message->header->name);
 931
 932             if ($message->header->entity_id != $ent_id) {
 933             $filename = decodeHeader($message->header->filename);
 934             if (trim($filename) == '') {
 935                 if (trim($name) == '') {
 936                     if ( trim( $message->header->id ) == '' )
 937                         $display_filename = 'untitled-[' . $message->header->entity_id . ']' ;
 938                     else
 939                         $display_filename = 'cid: ' . $message->header->id;
 940                     // $display_filename = 'untitled-[' . $message->header->entity_id . ']' ;
 941                 } else {
 942                     $display_filename = $name;
 943                     $filename = $name;
 944                 }
 945             } else {
 946                 $display_filename = $filename;
 947             }
 948
 949             $urlMailbox = urlencode($mailbox);
 950             $ent = urlencode($message->header->entity_id);
 951
 952             $DefaultLink =
 953                 "../src/download.php?startMessage=$startMessage&amp;passed_id=$id&amp;mailbox=$urlMailbox&amp;passed_ent_id=$ent";
 954             if ($where && $what) {
 955                $DefaultLink = '&amp;where='. urlencode($where).'&amp;what='.urlencode($what);
 956             }
 957             $Links['download link']['text'] = _("download");
 958             $Links['download link']['href'] =
 959                 "../src/download.php?absolute_dl=true&amp;passed_id=$id&amp;mailbox=$urlMailbox&amp;passed_ent_id=$ent";
 960             $ImageURL = '';
 961
 962             /* this executes the attachment hook with a specific MIME-type.
 963                 * if that doens't have results, it tries if there's a rule
 964                 * for a more generic type. */
 965             $HookResults = do_hook("attachment $type0/$type1", $Links,
 966                 $startMessage, $id, $urlMailbox, $ent, $DefaultLink,
 967                 $display_filename, $where, $what);
 968             if(count($HookResults[1]) <= 1) {
 969                 $HookResults = do_hook("attachment $type0/*", $Links,
 970                 $startMessage, $id, $urlMailbox, $ent, $DefaultLink,
 971                 $display_filename, $where, $what);
 972             }
 973
 974             $Links = $HookResults[1];
 975             $DefaultLink = $HookResults[6];
 976
 977             $body .= '<TR><TD>&nbsp;&nbsp;</TD><TD>' .
 978                         "<A HREF=\"$DefaultLink\">$display_filename</A>&nbsp;</TD>" .
 979                         '<TD><SMALL><b>' . show_readable_size($message->header->size) .
 980                         '</b>&nbsp;&nbsp;</small></TD>' .
 981                         "<TD><SMALL>[ $type0/$type1 ]&nbsp;</SMALL></TD>" .
 982                         '<TD><SMALL>';
 983             if ($message->header->description) {
 984                 $body .= '<b>' . htmlspecialchars(_($message->header->description)) . '</b>';
 985             }
 986             $body .= '</SMALL></TD><TD><SMALL>&nbsp;';
 987
 988
 989             $SkipSpaces = 1;
 990             foreach ($Links as $Val) {
 991                 if ($SkipSpaces) {
 992                     $SkipSpaces = 0;
 993                 } else {
 994                     $body .= '&nbsp;&nbsp;|&nbsp;&nbsp;';
 995                 }
 996                 $body .= '<a href="' . $Val['href'] . '">' .  $Val['text'] . '</a>';
 997             }
 998
 999             unset($Links);
1000
1001             $body .= "</SMALL></TD></TR>\n";
1002             }
1003         } else {
1004             for ($i = 0; $i < count($message->entities); $i++) {
1005                 $body .= formatAttachments($message->entities[$i], $ent_id, $mailbox, $id);
1006             }
1007         }
1008     }
1009     return( $body );
1010 }
1011
1012
1013 /** this function decodes the body depending on the encoding type. **/
1014 function decodeBody($body, $encoding) {
1015   $body = str_replace("\r\n", "\n", $body);
1016   $encoding = strtolower($encoding);
1017
1018   global $show_html_default;
1019
1020   if ($encoding == 'quoted-printable' ||
1021       $encoding == 'quoted_printable') {
1022      $body = quoted_printable_decode($body);
1023
1024
1025      while (ereg("=\n", $body))
1026         $body = ereg_replace ("=\n", "", $body);
1027
1028   } else if ($encoding == 'base64') {
1029      $body = base64_decode($body);
1030   }
1031
1032   // All other encodings are returned raw.
1033   return $body;
1034 }
1035
1036 /*
1037  * This functions decode strings that is encoded according to
1038  * RFC1522 (MIME Part Two: Message Header Extensions for Non-ASCII Text).
1039  * Patched by Christian Schmidt <christian@ostenfeld.dk>  23/03/2002
1040  */
1041 function decodeHeader ($string, $utfencode=true) {
1042     if (is_array($string)) {
1043         $string = implode("\n", $string);
1044     }
1045     $i = 0;
1046     while (preg_match('/^(.{' . $i . '})(.*)=\?([^?]*)\?(Q|B)\?([^?]*)\?=/Ui',
1047                       $string, $res)) {
1048         $prefix = $res[1];
1049         // Ignore white-space between consecutive encoded-words
1050         if (strspn($res[2], " \t") != strlen($res[2])) {
1051             $prefix .= $res[2];
1052         }
1053
1054         if (ucfirst($res[4]) == 'B') {
1055             $replace = base64_decode($res[5]);
1056         } else {
1057             $replace = str_replace('_', ' ', $res[5]);
1058             $replace = preg_replace('/=([0-9a-f]{2})/ie', 'chr(hexdec("\1"))',
1059                                     $replace);
1060             /* Only encode into entities by default. Some places
1061                don't need the encoding, like the compose form. */
1062             if ($utfencode) {
1063                 $replace = charset_decode($res[3], $replace);
1064             }
1065         }
1066         $string = $prefix . $replace . substr($string, strlen($res[0]));
1067         $i = strlen($prefix) + strlen($replace);
1068     }
1069     return( $string );
1070 }
1071
1072 /*
1073  * Encode a string according to RFC 1522 for use in headers if it
1074  * contains 8-bit characters or anything that looks like it should
1075  * be encoded.
1076  */
1077 function encodeHeader ($string) {
1078     global $default_charset;
1079
1080     // Encode only if the string contains 8-bit characters or =?
1081     $j = strlen( $string  );
1082     $l = strstr($string, '=?');         // Must be encoded ?
1083     $ret = '';
1084     for( $i=0; $i < $j; ++$i) {
1085         switch( $string{$i} ) {
1086            case '=':
1087           $ret .= '=3D';
1088           break;
1089         case '?':
1090           $ret .= '=3F';
1091           break;
1092         case '_':
1093           $ret .= '=5F';
1094           break;
1095         case ' ':
1096           $ret .= '_';
1097           break;
1098         default:
1099           $k = ord( $string{$i} );
1100           if ( $k > 126 ) {
1101              $ret .= sprintf("=%02X", $k);
1102              $l = TRUE;
1103           } else
1104              $ret .= $string{$i};
1105         }
1106     }
1107
1108     if ( $l ) {
1109         $string = "=?$default_charset?Q?$ret?=";
1110     }
1111
1112     return( $string );
1113 }
1114
1115 /* This function trys to locate the entity_id of a specific mime element */
1116
1117 function find_ent_id( $id, $message ) {
1118     $ret = '';
1119     for ($i=0; $ret == '' && $i < count($message->entities); $i++) {
1120         if (( $message->entities[$i]->header->type1 == 'alternative') ||
1121             ( $message->entities[$i]->header->type1 == 'related') ||
1122             ( $message->entities[$i]->header->type1 == 'mixed')) {
1123             $ret = find_ent_id( $id, $message->entities[$i] );
1124         } else {
1125             if ( strcasecmp( $message->entities[$i]->header->id, $id ) == 0 )
1126                 $ret = $message->entities[$i]->header->entity_id;
1127         }
1128
1129     }
1130     return( $ret );
1131 }
1132
1133 /**
1134  ** HTMLFILTER ROUTINES
1135  */
1136
1137 /**
1138  * This function returns the final tag out of the tag name, an array
1139  * of attributes, and the type of the tag. This function is called by
1140  * sq_sanitize internally.
1141  *
1142  * @param  $tagname  the name of the tag.
1143  * @param  $attary   the array of attributes and their values
1144  * @param  $tagtype  The type of the tag (see in comments).
1145  * @return           a string with the final tag representation.
1146  */
1147 function sq_tagprint($tagname, $attary, $tagtype){
1148     $me = "sq_tagprint";
1149     if ($tagtype == 2){
1150         $fulltag = '</' . $tagname . '>';
1151     } else {
1152         $fulltag = '<' . $tagname;
1153         if (is_array($attary) && sizeof($attary)){
1154             $atts = Array();
1155             while (list($attname, $attvalue) = each($attary)){
1156                 array_push($atts, "$attname=$attvalue");
1157             }
1158             $fulltag .= ' ' . join(" ", $atts);
1159         }
1160         if ($tagtype == 3){
1161             $fulltag .= " /";
1162         }
1163         $fulltag .= ">";
1164     }
1165     return $fulltag;
1166 }
1167
1168 /**
1169  * A small helper function to use with array_walk. Modifies a by-ref
1170  * value and makes it lowercase.
1171  *
1172  * @param  $val a value passed by-ref.
1173  * @return      void since it modifies a by-ref value.
1174  */
1175 function sq_casenormalize(&$val){
1176     $val = strtolower($val);
1177 }
1178
1179 /**
1180  * This function skips any whitespace from the current position within
1181  * a string and to the next non-whitespace value.
1182  *
1183  * @param  $body   the string
1184  * @param  $offset the offset within the string where we should start
1185  *                 looking for the next non-whitespace character.
1186  * @return         the location within the $body where the next
1187  *                 non-whitespace char is located.
1188  */
1189 function sq_skipspace($body, $offset){
1190     $me = "sq_skipspace";
1191     preg_match("/^(\s*)/s", substr($body, $offset), $matches);
1192     if (sizeof($matches{1})){
1193         $count = strlen($matches{1});
1194         $offset += $count;
1195     }
1196     return $offset;
1197 }
1198
1199 /**
1200  * This function looks for the next character within a string.  It's
1201  * really just a glorified "strpos", except it catches if failures
1202  * nicely.
1203  *
1204  * @param  $body   The string to look for needle in.
1205  * @param  $offset Start looking from this position.
1206  * @param  $needle The character/string to look for.
1207  * @return         location of the next occurance of the needle, or
1208  *                 strlen($body) if needle wasn't found.
1209  */
1210 function sq_findnxstr($body, $offset, $needle){
1211     $me = "sq_findnxstr";
1212     $pos = strpos($body, $needle, $offset);
1213     if ($pos === FALSE){
1214         $pos = strlen($body);
1215     }
1216     return $pos;
1217 }
1218
1219 /**
1220  * This function takes a PCRE-style regexp and tries to match it
1221  * within the string.
1222  *
1223  * @param  $body   The string to look for needle in.
1224  * @param  $offset Start looking from here.
1225  * @param  $reg    A PCRE-style regex to match.
1226  * @return         Returns a false if no matches found, or an array
1227  *                 with the following members:
1228  *                 - integer with the location of the match within $body
1229  *                 - string with whatever content between offset and the match
1230  *                 - string with whatever it is we matched
1231  */
1232 function sq_findnxreg($body, $offset, $reg){
1233     $me = "sq_findnxreg";
1234     $matches = Array();
1235     $retarr = Array();
1236     preg_match("%^(.*?)($reg)%s", substr($body, $offset), $matches);
1237     if (!$matches{0}){
1238         $retarr = false;
1239     } else {
1240         $retarr{0} = $offset + strlen($matches{1});
1241         $retarr{1} = $matches{1};
1242         $retarr{2} = $matches{2};
1243     }
1244     return $retarr;
1245 }
1246
1247 /**
1248  * This function looks for the next tag.
1249  *
1250  * @param  $body   String where to look for the next tag.
1251  * @param  $offset Start looking from here.
1252  * @return         false if no more tags exist in the body, or
1253  *                 an array with the following members:
1254  *                 - string with the name of the tag
1255  *                 - array with attributes and their values
1256  *                 - integer with tag type (1, 2, or 3)
1257  *                 - integer where the tag starts (starting "<")
1258  *                 - integer where the tag ends (ending ">")
1259  *                 first three members will be false, if the tag is invalid.
1260  */
1261 function sq_getnxtag($body, $offset){
1262     $me = "sq_getnxtag";
1263     if ($offset > strlen($body)){
1264         return false;
1265     }
1266     $lt = sq_findnxstr($body, $offset, "<");
1267     if ($lt == strlen($body)){
1268         return false;
1269     }
1270     /**
1271      * We are here:
1272      * blah blah <tag attribute="value">
1273      * \---------^
1274      */
1275     $pos = sq_skipspace($body, $lt+1);
1276     if ($pos >= strlen($body)){
1277         return Array(false, false, false, $lt, strlen($body));
1278     }
1279     /**
1280      * There are 3 kinds of tags:
1281      * 1. Opening tag, e.g.:
1282      *    <a href="blah">
1283      * 2. Closing tag, e.g.:
1284      *    </a>
1285      * 3. XHTML-style content-less tag, e.g.:
1286      *    <img src="blah"/>
1287      */
1288     $tagtype = false;
1289     switch (substr($body, $pos, 1)){
1290     case "/":
1291         $tagtype = 2;
1292         $pos++;
1293         break;
1294     case "!":
1295         /**
1296          * A comment or an SGML declaration.
1297          */
1298         if (substr($body, $pos+1, 2) == "--"){
1299             $gt = strpos($body, "-->", $pos)+2;
1300             if ($gt === false){
1301                 $gt = strlen($body);
1302             }
1303             return Array(false, false, false, $lt, $gt);
1304         } else {
1305             $gt = sq_findnxstr($body, $pos, ">");
1306             return Array(false, false, false, $lt, $gt);
1307         }
1308         break;
1309     default:
1310         /**
1311          * Assume tagtype 1 for now. If it's type 3, we'll switch values
1312          * later.
1313          */
1314         $tagtype = 1;
1315         break;
1316     }
1317
1318     $tag_start = $pos;
1319     $tagname = '';
1320     /**
1321      * Look for next [\W-_], which will indicate the end of the tag name.
1322      */
1323     $regary = sq_findnxreg($body, $pos, "[^\w\-_]");
1324     if ($regary == false){
1325         return Array(false, false, false, $lt, strlen($body));
1326     }
1327     list($pos, $tagname, $match) = $regary;
1328     $tagname = strtolower($tagname);
1329
1330     /**
1331      * $match can be either of these:
1332      * '>'  indicating the end of the tag entirely.
1333      * '\s' indicating the end of the tag name.
1334      * '/'  indicating that this is type-3 xhtml tag.
1335      *
1336      * Whatever else we find there indicates an invalid tag.
1337      */
1338     switch ($match){
1339     case "/":
1340         /**
1341          * This is an xhtml-style tag with a closing / at the
1342          * end, like so: <img src="blah"/>. Check if it's followed
1343          * by the closing bracket. If not, then this tag is invalid
1344          */
1345         if (substr($body, $pos, 2) == "/>"){
1346             $pos++;
1347             $tagtype = 3;
1348         } else {
1349             $gt = sq_findnxstr($body, $pos, ">");
1350             $retary = Array(false, false, false, $lt, $gt);
1351             return $retary;
1352         }
1353     case ">":
1354         return Array($tagname, false, $tagtype, $lt, $pos);
1355         break;
1356     default:
1357         /**
1358          * Check if it's whitespace
1359          */
1360         if (preg_match("/\s/", $match)){
1361         } else {
1362             /**
1363              * This is an invalid tag! Look for the next closing ">".
1364              */
1365             $gt = sq_findnxstr($body, $offset, ">");
1366             return Array(false, false, false, $lt, $gt);
1367         }
1368     }
1369
1370     /**
1371      * At this point we're here:
1372      * <tagname  attribute='blah'>
1373      * \-------^
1374      *
1375      * At this point we loop in order to find all attributes.
1376      */
1377     $attname = '';
1378     $atttype = false;
1379     $attary = Array();
1380
1381     while ($pos <= strlen($body)){
1382         $pos = sq_skipspace($body, $pos);
1383         if ($pos == strlen($body)){
1384             /**
1385              * Non-closed tag.
1386              */
1387             return Array(false, false, false, $lt, $pos);
1388         }
1389         /**
1390          * See if we arrived at a ">" or "/>", which means that we reached
1391          * the end of the tag.
1392          */
1393         $matches = Array();
1394         if (preg_match("%^(\s*)(>|/>)%s", substr($body, $pos), $matches)) {
1395             /**
1396              * Yep. So we did.
1397              */
1398             $pos += strlen($matches{1});
1399             if ($matches{2} == "/>"){
1400                 $tagtype = 3;
1401                 $pos++;
1402             }
1403             return Array($tagname, $attary, $tagtype, $lt, $pos);
1404         }
1405
1406         /**
1407          * There are several types of attributes, with optional
1408          * [:space:] between members.
1409          * Type 1:
1410          *   attrname[:space:]=[:space:]'CDATA'
1411          * Type 2:
1412          *   attrname[:space:]=[:space:]"CDATA"
1413          * Type 3:
1414          *   attr[:space:]=[:space:]CDATA
1415          * Type 4:
1416          *   attrname
1417          *
1418          * We leave types 1 and 2 the same, type 3 we check for
1419          * '"' and convert to "&quot" if needed, then wrap in
1420          * double quotes. Type 4 we convert into:
1421          * attrname="yes".
1422          */
1423         $regary = sq_findnxreg($body, $pos, "[^\w\-_]");
1424         if ($regary == false){
1425             /**
1426              * Looks like body ended before the end of tag.
1427              */
1428             return Array(false, false, false, $lt, strlen($body));
1429         }
1430         list($pos, $attname, $match) = $regary;
1431         $attname = strtolower($attname);
1432         /**
1433          * We arrived at the end of attribute name. Several things possible
1434          * here:
1435          * '>'  means the end of the tag and this is attribute type 4
1436          * '/'  if followed by '>' means the same thing as above
1437          * '\s' means a lot of things -- look what it's followed by.
1438          *      anything else means the attribute is invalid.
1439          */
1440         switch($match){
1441         case "/":
1442             /**
1443              * This is an xhtml-style tag with a closing / at the
1444              * end, like so: <img src="blah"/>. Check if it's followed
1445              * by the closing bracket. If not, then this tag is invalid
1446              */
1447             if (substr($body, $pos, 2) == "/>"){
1448                 $pos++;
1449                 $tagtype = 3;
1450             } else {
1451                 $gt = sq_findnxstr($body, $pos, ">");
1452                 $retary = Array(false, false, false, $lt, $gt);
1453                 return $retary;
1454             }
1455         case ">":
1456             $attary{$attname} = '"yes"';
1457             return Array($tagname, $attary, $tagtype, $lt, $pos);
1458             break;
1459         default:
1460             /**
1461              * Skip whitespace and see what we arrive at.
1462              */
1463             $pos = sq_skipspace($body, $pos);
1464             $char = substr($body, $pos, 1);
1465             /**
1466              * Two things are valid here:
1467              * '=' means this is attribute type 1 2 or 3.
1468              * \w means this was attribute type 4.
1469              * anything else we ignore and re-loop. End of tag and
1470              * invalid stuff will be caught by our checks at the beginning
1471              * of the loop.
1472              */
1473             if ($char == "="){
1474                 $pos++;
1475                 $pos = sq_skipspace($body, $pos);
1476                 /**
1477                  * Here are 3 possibilities:
1478                  * "'"  attribute type 1
1479                  * '"'  attribute type 2
1480                  * everything else is the content of tag type 3
1481                  */
1482                 $quot = substr($body, $pos, 1);
1483                 if ($quot == "'"){
1484                     $regary = sq_findnxreg($body, $pos+1, "\'");
1485                     if ($regary == false){
1486                         return Array(false, false, false, $lt, strlen($body));
1487                     }
1488                     list($pos, $attval, $match) = $regary;
1489                     $pos++;
1490                     $attary{$attname} = "'" . $attval . "'";
1491                 } else if ($quot == '"'){
1492                     $regary = sq_findnxreg($body, $pos+1, '\"');
1493                     if ($regary == false){
1494                         return Array(false, false, false, $lt, strlen($body));
1495                     }
1496                     list($pos, $attval, $match) = $regary;
1497                     $pos++;
1498                     $attary{$attname} = '"' . $attval . '"';
1499                 } else {
1500                     /**
1501                      * These are hateful. Look for \s, or >.
1502                      */
1503                     $regary = sq_findnxreg($body, $pos, "[\s>]");
1504                     if ($regary == false){
1505                         return Array(false, false, false, $lt, strlen($body));
1506                     }
1507                     list($pos, $attval, $match) = $regary;
1508                     /**
1509                      * If it's ">" it will be caught at the top.
1510                      */
1511                     $attval = preg_replace("/\"/s", "&quot;", $attval);
1512                     $attary{$attname} = '"' . $attval . '"';
1513                 }
1514             } else if (preg_match("|[\w/>]|", $char)) {
1515                 /**
1516                  * That was attribute type 4.
1517                  */
1518                 $attary{$attname} = '"yes"';
1519             } else {
1520                 /**
1521                  * An illegal character. Find next '>' and return.
1522                  */
1523                 $gt = sq_findnxstr($body, $pos, ">");
1524                 return Array(false, false, false, $lt, $gt);
1525             }
1526         }
1527     }
1528     /**
1529      * The fact that we got here indicates that the tag end was never
1530      * found. Return invalid tag indication so it gets stripped.
1531      */
1532     return Array(false, false, false, $lt, strlen($body));
1533 }
1534
1535 /**
1536  * This function checks attribute values for entity-encoded values
1537  * and returns them translated into 8-bit strings so we can run
1538  * checks on them.
1539  *
1540  * @param  $attvalue A string to run entity check against.
1541  * @return           Translated value.
1542  */
1543 function sq_deent($attvalue){
1544     $me="sq_deent";
1545     /**
1546      * See if we have to run the checks first. All entities must start
1547      * with "&".
1548      */
1549     if (strpos($attvalue, "&") === false){
1550         return $attvalue;
1551     }
1552     /**
1553      * Check named entities first.
1554      */
1555     $trans = get_html_translation_table(HTML_ENTITIES);
1556     /**
1557      * Leave &quot; in, as it can mess us up.
1558      */
1559     $trans = array_flip($trans);
1560     unset($trans{"&quot;"});
1561     while (list($ent, $val) = each($trans)){
1562         $attvalue = preg_replace("/$ent*(\W)/si", "$val\\1", $attvalue);
1563     }
1564     /**
1565      * Now translate numbered entities from 1 to 255 if needed.
1566      */
1567     if (strpos($attvalue, "#") !== false){
1568         $omit = Array(34, 39);
1569         for ($asc=1; $asc<256; $asc++){
1570             if (!in_array($asc, $omit)){
1571                 $chr = chr($asc);
1572                 $attvalue = preg_replace("/\&#0*$asc;*(\D)/si", "$chr\\1",
1573                                          $attvalue);
1574                 $attvalue = preg_replace("/\&#x0*".dechex($asc).";*(\W)/si",
1575                                          "$chr\\1", $attvalue);
1576             }
1577         }
1578     }
1579     return $attvalue;
1580 }
1581
1582 /**
1583  * This function runs various checks against the attributes.
1584  *
1585  * @param  $tagname         String with the name of the tag.
1586  * @param  $attary          Array with all tag attributes.
1587  * @param  $rm_attnames     See description for sq_sanitize
1588  * @param  $bad_attvals     See description for sq_sanitize
1589  * @param  $add_attr_to_tag See description for sq_sanitize
1590  * @param  $message         message object
1591  * @param  $id              message id
1592  * @return                  Array with modified attributes.
1593  */
1594 function sq_fixatts($tagname,
1595                     $attary,
1596                     $rm_attnames,
1597                     $bad_attvals,
1598                     $add_attr_to_tag,
1599                     $message,
1600                     $id
1601                     ){
1602     $me = "sq_fixatts";
1603     while (list($attname, $attvalue) = each($attary)){
1604         /**
1605          * See if this attribute should be removed.
1606          */
1607         foreach ($rm_attnames as $matchtag=>$matchattrs){
1608             if (preg_match($matchtag, $tagname)){
1609                 foreach ($matchattrs as $matchattr){
1610                     if (preg_match($matchattr, $attname)){
1611                         unset($attary{$attname});
1612                         continue;
1613                     }
1614                 }
1615             }
1616         }
1617         /**
1618          * Remove any entities.
1619          */
1620         $attvalue = sq_deent($attvalue);
1621
1622         /**
1623          * Now let's run checks on the attvalues.
1624          * I don't expect anyone to comprehend this. If you do,
1625          * get in touch with me so I can drive to where you live and
1626          * shake your hand personally. :)
1627          */
1628         foreach ($bad_attvals as $matchtag=>$matchattrs){
1629             if (preg_match($matchtag, $tagname)){
1630                 foreach ($matchattrs as $matchattr=>$valary){
1631                     if (preg_match($matchattr, $attname)){
1632                         /**
1633                          * There are two arrays in valary.
1634                          * First is matches.
1635                          * Second one is replacements
1636                          */
1637                         list($valmatch, $valrepl) = $valary;
1638                         $newvalue =
1639                             preg_replace($valmatch, $valrepl, $attvalue);
1640                         if ($newvalue != $attvalue){
1641                             $attary{$attname} = $newvalue;
1642                         }
1643                     }
1644                 }
1645             }
1646         }
1647         /**
1648          * Turn cid: urls into http-friendly ones.
1649          */
1650         if (preg_match("/^[\'\"]\s*cid:/si", $attvalue)){
1651             $attary{$attname} = sq_cid2http($message, $id, $attvalue);
1652         }
1653     }
1654     /**
1655      * See if we need to append any attributes to this tag.
1656      */
1657     foreach ($add_attr_to_tag as $matchtag=>$addattary){
1658         if (preg_match($matchtag, $tagname)){
1659             $attary = array_merge($attary, $addattary);
1660         }
1661     }
1662     return $attary;
1663 }
1664
1665 /**
1666  * This function edits the style definition to make them friendly and
1667  * usable in squirrelmail.
1668  *
1669  * @param  $message  the message object
1670  * @param  $id       the message id
1671  * @param  $content  a string with whatever is between <style> and </style>
1672  * @return           a string with edited content.
1673  */
1674 function sq_fixstyle($message, $id, $content){
1675     global $view_unsafe_images;
1676     $me = "sq_fixstyle";
1677     /**
1678      * First look for general BODY style declaration, which would be
1679      * like so:
1680      * body {background: blah-blah}
1681      * and change it to .bodyclass so we can just assign it to a <div>
1682      */
1683     $content = preg_replace("|body(\s*\{.*?\})|si", ".bodyclass\\1", $content);
1684     $secremoveimg = "../images/" . _("sec_remove_eng.png");
1685     /**
1686      * Fix url('blah') declarations.
1687      */
1688     $content = preg_replace("|url\(([\'\"])\s*\S+script\s*:.*?([\'\"])\)|si",
1689                             "url(\\1$secremoveimg\\2)", $content);
1690     /**
1691      * Fix url('https*://.*) declarations but only if $view_unsafe_images
1692      * is false.
1693      */
1694     if (!$view_unsafe_images){
1695         $content = preg_replace("|url\(([\'\"])\s*https*:.*?([\'\"])\)|si",
1696                                 "url(\\1$secremoveimg\\2)", $content);
1697     }
1698
1699     /**
1700      * Fix urls that refer to cid:
1701      */
1702     while (preg_match("|url\(([\'\"]\s*cid:.*?[\'\"])\)|si", $content,
1703                       $matches)){
1704         $cidurl = $matches{1};
1705         $httpurl = sq_cid2http($message, $id, $cidurl);
1706         $content = preg_replace("|url\($cidurl\)|si",
1707                                 "url($httpurl)", $content);
1708     }
1709
1710     /**
1711      * Fix stupid expression: declarations which lead to vulnerabilities
1712      * in IE.
1713      */
1714     $content = preg_replace("/expression\s*:/si", "idiocy:", $content);
1715     return $content;
1716 }
1717
1718 /**
1719  * This function converts cid: url's into the ones that can be viewed in
1720  * the browser.
1721  *
1722  * @param  $message  the message object
1723  * @param  $id       the message id
1724  * @param  $cidurl   the cid: url.
1725  * @return           a string with a http-friendly url
1726  */
1727 function sq_cid2http($message, $id, $cidurl){
1728     /**
1729      * Get rid of quotes.
1730      */
1731     $quotchar = substr($cidurl, 0, 1);
1732     $cidurl = str_replace($quotchar, "", $cidurl);
1733     $cidurl = substr(trim($cidurl), 4);
1734     $httpurl = $quotchar . "../src/download.php?absolute_dl=true&amp;" .
1735         "passed_id=$id&amp;mailbox=" . urlencode($message->header->mailbox) .
1736         "&amp;passed_ent_id=" . find_ent_id($cidurl, $message) . $quotchar;
1737     return $httpurl;
1738 }
1739
1740 /**
1741  * This function changes the <body> tag into a <div> tag since we
1742  * can't really have a body-within-body.
1743  *
1744  * @param  $attary  an array of attributes and values of <body>
1745  * @return          a modified array of attributes to be set for <div>
1746  */
1747 function sq_body2div($attary){
1748     $me = "sq_body2div";
1749     $divattary = Array("class"=>"'bodyclass'");
1750     $bgcolor="#ffffff";
1751     $text="#000000";
1752     $styledef="";
1753     if (is_array($attary) && sizeof($attary) > 0){
1754         foreach ($attary as $attname=>$attvalue){
1755             $quotchar = substr($attvalue, 0, 1);
1756             $attvalue = str_replace($quotchar, "", $attvalue);
1757             switch ($attname){
1758             case "background":
1759                 $styledef .= "background-image: url('$attvalue'); ";
1760                 break;
1761             case "bgcolor":
1762                 $styledef .= "background-color: $attvalue; ";
1763                 break;
1764             case "text":
1765                 $styledef .= "color: $attvalue; ";
1766             }
1767         }
1768         if (strlen($styledef) > 0){
1769             $divattary{"style"} = "\"$styledef\"";
1770         }
1771     }
1772     return $divattary;
1773 }
1774
1775 /**
1776  * This is the main function and the one you should actually be calling.
1777  * There are several variables you should be aware of an which need
1778  * special description.
1779  *
1780  * Since the description is quite lengthy, see it here:
1781  * http://www.mricon.com/html/phpfilter.html
1782  *
1783  * @param $body                 the string with HTML you wish to filter
1784  * @param $tag_list             see description above
1785  * @param $rm_tags_with_content see description above
1786  * @param $self_closing_tags    see description above
1787  * @param $force_tag_closing    see description above
1788  * @param $rm_attnames          see description above
1789  * @param $bad_attvals          see description above
1790  * @param $add_attr_to_tag      see description above
1791  * @param $message              message object
1792  * @param $id                   message id
1793  * @return                      sanitized html safe to show on your pages.
1794  */
1795 function sq_sanitize($body,
1796                      $tag_list,
1797                      $rm_tags_with_content,
1798                      $self_closing_tags,
1799                      $force_tag_closing,
1800                      $rm_attnames,
1801                      $bad_attvals,
1802                      $add_attr_to_tag,
1803                      $message,
1804                      $id
1805                      ){
1806     $me = "sq_sanitize";
1807     /**
1808      * Normalize rm_tags and rm_tags_with_content.
1809      */
1810     @array_walk($rm_tags, 'sq_casenormalize');
1811     @array_walk($rm_tags_with_content, 'sq_casenormalize');
1812     @array_walk($self_closing_tags, 'sq_casenormalize');
1813     /**
1814      * See if tag_list is of tags to remove or tags to allow.
1815      * false  means remove these tags
1816      * true   means allow these tags
1817      */
1818     $rm_tags = array_shift($tag_list);
1819     $curpos = 0;
1820     $open_tags = Array();
1821     $trusted = "<!-- begin sanitized html -->\n";
1822     $skip_content = false;
1823
1824     while (($curtag=sq_getnxtag($body, $curpos)) != FALSE){
1825         list($tagname, $attary, $tagtype, $lt, $gt) = $curtag;
1826         $free_content = substr($body, $curpos, $lt-$curpos);
1827         /**
1828          * Take care of <style>
1829          */
1830         if ($tagname == "style" && $tagtype == 2){
1831             /**
1832              * This is a closing </style>. Edit the
1833              * content before we apply it.
1834              */
1835             $free_content = sq_fixstyle($message, $id, $free_content);
1836         }
1837         if ($skip_content == false){
1838             $trusted .= $free_content;
1839         } else {
1840         }
1841         if ($tagname != FALSE){
1842             if ($tagtype == 2){
1843                 if ($skip_content == $tagname){
1844                     /**
1845                      * Got to the end of tag we needed to remove.
1846                      */
1847                     $tagname = false;
1848                     $skip_content = false;
1849                 } else {
1850                     if ($skip_content == false){
1851                         if ($tagname == "body"){
1852                             $tagname = "div";
1853                         } else {
1854                             if (isset($open_tags{$tagname}) &&
1855                                 $open_tags{$tagname} > 0){
1856                                 $open_tags{$tagname}--;
1857                             } else {
1858                                 $tagname = false;
1859                             }
1860                         }
1861                     } else {
1862                     }
1863                 }
1864             } else {
1865                 /**
1866                  * $rm_tags_with_content
1867                  */
1868                 if ($skip_content == false){
1869                     /**
1870                      * See if this is a self-closing type and change
1871                      * tagtype appropriately.
1872                      */
1873                     if ($tagtype == 1
1874                         && in_array($tagname, $self_closing_tags)){
1875                         $tagtype=3;
1876                     }
1877                     /**
1878                      * See if we should skip this tag and any content
1879                      * inside it.
1880                      */
1881                     if ($tagtype == 1 &&
1882                         in_array($tagname, $rm_tags_with_content)){
1883                         $skip_content = $tagname;
1884                     } else {
1885                         if (($rm_tags == false
1886                              && in_array($tagname, $tag_list)) ||
1887                             ($rm_tags == true &&
1888                              !in_array($tagname, $tag_list))){
1889                             $tagname = false;
1890                         } else {
1891                             if ($tagtype == 1){
1892                                 if (isset($open_tags{$tagname})){
1893                                     $open_tags{$tagname}++;
1894                                 } else {
1895                                     $open_tags{$tagname}=1;
1896                                 }
1897                             }
1898                             /**
1899                              * This is where we run other checks.
1900                              */
1901                             if (is_array($attary) && sizeof($attary) > 0){
1902                                 $attary = sq_fixatts($tagname,
1903                                                      $attary,
1904                                                      $rm_attnames,
1905                                                      $bad_attvals,
1906                                                      $add_attr_to_tag,
1907                                                      $message,
1908                                                      $id
1909                                                      );
1910                             }
1911                             /**
1912                              * Convert body into div.
1913                              */
1914                             if ($tagname == "body"){
1915                                 $tagname = "div";
1916                                 $attary = sq_body2div($attary, $message, $id);
1917                             }
1918                         }
1919                     }
1920                 } else {
1921                 }
1922             }
1923             if ($tagname != false && $skip_content == false){
1924                 $trusted .= sq_tagprint($tagname, $attary, $tagtype);
1925             }
1926         } else {
1927         }
1928         $curpos = $gt+1;
1929     }
1930     $trusted .= substr($body, $curpos, strlen($body)-$curpos);
1931     if ($force_tag_closing == true){
1932         foreach ($open_tags as $tagname=>$opentimes){
1933             while ($opentimes > 0){
1934                 $trusted .= '</' . $tagname . '>';
1935                 $opentimes--;
1936             }
1937         }
1938         $trusted .= "\n";
1939     }
1940     $trusted .= "<!-- end sanitized html -->\n";
1941     return $trusted;
1942 }
1943
1944 /**
1945  * This is a wrapper function to call html sanitizing routines.
1946  *
1947  * @param  $body  the body of the message
1948  * @param  $id    the id of the message
1949  * @return        a string with html safe to display in the browser.
1950  */
1951 function magicHTML($body, $id, $message){
1952     global $attachment_common_show_images, $view_unsafe_images,
1953         $has_unsafe_images;
1954     /**
1955      * Don't display attached images in HTML mode.
1956      */
1957     $attachment_common_show_images = false;
1958     $tag_list = Array(
1959                       false,
1960                       "object",
1961                       "meta",
1962                       "html",
1963                       "head",
1964                       "base"
1965                       );
1966
1967     $rm_tags_with_content = Array(
1968                                   "script",
1969                                   "applet",
1970                                   "embed",
1971                                   "title"
1972                                   );
1973
1974     $self_closing_tags =  Array(
1975                                 "img",
1976                                 "br",
1977                                 "hr",
1978                                 "input"
1979                                 );
1980
1981     $force_tag_closing = false;
1982
1983     $rm_attnames = Array(
1984                          "/.*/" =>
1985                          Array(
1986                                "/target/si",
1987                                "/^on.*/si"
1988                                )
1989                          );
1990
1991     $secremoveimg = "../images/" . _("sec_remove_eng.png");
1992     $bad_attvals = Array(
1993         "/.*/" =>
1994             Array(
1995                 "/^src|background|href|action/i" =>
1996                     Array(
1997                           Array(
1998                                 "|^([\'\"])\s*\.\./.*([\'\"])|si",
1999                                 "/^([\'\"])\s*\S+script\s*:.*([\'\"])/si"
2000                                 ),
2001                           Array(
2002                                 "\\1$secremoveimg\\2",
2003                                 "\\1$secremoveimg\\2"
2004                                 )
2005                         ),
2006                 "/^style/si" =>
2007                     Array(
2008                           Array(
2009                                 "/expression\s*:/si",
2010                                 "|url\(([\'\"])\s*\.\./.*([\'\"])\)|si",
2011                                 "/url\(([\'\"])\s*\S+script:.*([\'\"])\)/si"
2012                                ),
2013                           Array(
2014                                 "idiocy:",
2015                                 "url(\\1$secremoveimg\\2)",
2016                                 "url(\\1$secremoveimg\\2)"
2017                                )
2018                           )
2019                 )
2020         );
2021     if (!$view_unsafe_images){
2022         /**
2023          * Remove any references to http/https if view_unsafe_images set
2024          * to false.
2025          */
2026         $addendum = Array(
2027           "/.*/" =>
2028             Array(
2029                 "/^src|background/i" =>
2030                     Array(
2031                           Array(
2032                                 "/^([\'\"])\s*https*:.*([\'\"])/si"
2033                                 ),
2034                           Array(
2035                                 "\\1$secremoveimg\\2"
2036                                 )
2037                         ),
2038                 "/^style/si" =>
2039                     Array(
2040                           Array(
2041                                 "/url\(([\'\"])\s*https*:.*([\'\"])\)/si"
2042                                ),
2043                           Array(
2044                                 "url(\\1$secremoveimg\\2)"
2045                                )
2046                           )
2047                 )
2048           );
2049         $bad_attvals = array_merge($bad_attvals, $addendum);
2050     }
2051
2052     $add_attr_to_tag = Array(
2053                              "/^a$/si" => Array('target'=>'"_new"')
2054                              );
2055     $trusted = sq_sanitize($body,
2056                            $tag_list,
2057                            $rm_tags_with_content,
2058                            $self_closing_tags,
2059                            $force_tag_closing,
2060                            $rm_attnames,
2061                            $bad_attvals,
2062                            $add_attr_to_tag,
2063                            $message,
2064                            $id
2065                            );
2066     if (preg_match("|$secremoveimg|si", $trusted)){
2067         $has_unsafe_images = true;
2068     }
2069     return $trusted;
2070 }
2071 ?>