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 /* --------------------------------------------------------------------------------- */
  19 /* MIME DECODING                                                                     */
  20 /* --------------------------------------------------------------------------------- */
  21
  22 /* This function gets the structure of a message and stores it in the "message" class.
  23  * It will return this object for use with all relevant header information and
  24  * fully parsed into the standard "message" object format.
  25  */
  26
  27 function mime_structure ($bodystructure, $flags=array()) {
  28
  29     // isolate the body structure and remove beginning and end parenthesis
  30     $read = trim(substr ($bodystructure, strpos(strtolower($bodystructure), 'bodystructure') + 13));
  31     $msg = &new message();
  32     $read = trim(substr ($read, 0, -1));
  33     $res = $msg->parseStructure($read);
  34     $msg = $res[0];
  35     $msg->setEnt('0');
  36     if (count($flags)) {
  37        foreach ($flags as $flag) {
  38           $char = strtoupper($flag{1});
  39           switch ($char) {
  40              case 'S':
  41                if (strtolower($flag) == '\\seen') {
  42                   $msg->is_seen = true;
  43                }
  44                break;
  45              case 'A':
  46                if (strtolower($flag) == '\\answered') {
  47                   $msg->is_answered = true;
  48                }
  49                break;
  50              case 'D':
  51                if (strtolower($flag) == '\\deleted') {
  52                   $msg->is_deleted = true;
  53                }
  54                break;
  55              case 'F':
  56                if (strtolower($flag) == '\\flagged') {
  57                   $msg->is_flagged = true;
  58                }
  59                break;
  60              case 'M':
  61                if (strtolower($flag) == '$mdnsent') {
  62                   $msg->is_mdnsent = true;
  63                }
  64                break;
  65              default:
  66                break;
  67           }
  68        }
  69     }
  70 //    listEntities($msg);
  71     return( $msg );
  72 }
  73
  74 /* this starts the parsing of a particular structure.  It is called recursively,
  75  * so it can be passed different structures.  It returns an object of type
  76  * $message.
  77  * First, it checks to see if it is a multipart message.  If it is, then it
  78  * handles that as it sees is necessary.  If it is just a regular entity,
  79  * then it parses it and adds the necessary header information (by calling out
  80  * to mime_get_elements()
  81  */
  82
  83 function mime_fetch_body($imap_stream, $id, $ent_id) {
  84     global $uid_support;
  85     /*
  86      * do a bit of error correction.  If we couldn't find the entity id, just guess
  87      * that it is the first one.  That is usually the case anyway.
  88      */
  89     if (!$ent_id) {
  90         $ent_id = 1;
  91     }
  92     $cmd = "FETCH $id BODY[$ent_id]";
  93
  94     $data = sqimap_run_command ($imap_stream, $cmd, true, $response, $message, $uid_support);
  95     do {
  96         $topline = trim(array_shift( $data ));
  97     } while( $topline && $topline[0] == '*' && !preg_match( '/\* [0-9]+ FETCH.*/i', $topline )) ;
  98
  99     $wholemessage = implode('', $data);
 100     if (ereg('\\{([^\\}]*)\\}', $topline, $regs)) {
 101
 102         $ret = substr( $wholemessage, 0, $regs[1] );
 103         /*
 104             There is some information in the content info header that could be important
 105             in order to parse html messages. Let's get them here.
 106         */
 107         if ( $ret{0} == '<' ) {
 108             $data = sqimap_run_command ($imap_stream, "FETCH $id BODY[$ent_id.MIME]", true, $response, $message, $uid_support);
 109         }
 110     } else if (ereg('"([^"]*)"', $topline, $regs)) {
 111         $ret = $regs[1];
 112     } else {
 113         global $where, $what, $mailbox, $passed_id, $startMessage;
 114         $par = 'mailbox=' . urlencode($mailbox) . "&amp;passed_id=$passed_id";
 115         if (isset($where) && isset($what)) {
 116             $par .= '&amp;where='. urlencode($where) . "&amp;what=" . urlencode($what);
 117         } else {
 118             $par .= "&amp;startMessage=$startMessage&amp;show_more=0";
 119         }
 120         $par .= '&amp;response=' . urlencode($response) .
 121                 '&amp;message=' . urlencode($message).
 122                 '&amp;topline=' . urlencode($topline);
 123
 124         echo   '<tt><br>' .
 125                '<table width="80%"><tr>' .
 126                '<tr><td colspan=2>' .
 127                _("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!") .
 128                " <A HREF=\"../src/retrievalerror.php?$par\"><br>" .
 129                _("Submit message") . '</A><BR>&nbsp;' .
 130                '</td></tr>' .
 131                '<td><b>' . _("Command:") . "</td><td>$cmd</td></tr>" .
 132                '<td><b>' . _("Response:") . "</td><td>$response</td></tr>" .
 133                '<td><b>' . _("Message:") . "</td><td>$message</td></tr>" .
 134                '<td><b>' . _("FETCH line:") . "</td><td>$topline</td></tr>" .
 135                "</table><BR></tt></font><hr>";
 136
 137         $data = sqimap_run_command ($imap_stream, "FETCH $passed_id BODY[]", true, $response, $message, $uid_support);
 138         array_shift($data);
 139         $wholemessage = implode('', $data);
 140
 141         $ret = $wholemessage;
 142     }
 143     return( $ret );
 144 }
 145
 146 function mime_print_body_lines ($imap_stream, $id, $ent_id, $encoding) {
 147     global $uid_support;
 148     // do a bit of error correction.  If we couldn't find the entity id, just guess
 149     // that it is the first one.  That is usually the case anyway.
 150     if (!$ent_id) {
 151         $ent_id = 1;
 152     }
 153     $sid = sqimap_session_id($uid_support);
 154     // Don't kill the connection if the browser is over a dialup
 155     // and it would take over 30 seconds to download it.
 156
 157     // don\8e´t call set_time_limit in safe mode.
 158     if (!ini_get("safe_mode")) {
 159         set_time_limit(0);
 160     }
 161     if ($uid_support) {
 162        $sid_s = substr($sid,0,strpos($sid, ' '));
 163     } else {
 164        $sid_s = $sid;
 165     }
 166
 167     $body = mime_fetch_body ($imap_stream, $id, $ent_id);
 168     echo decodeBody($body, $encoding);
 169     return;
 170 /*
 171     fputs ($imap_stream, "$sid FETCH $id BODY[$ent_id]\r\n");
 172     $cnt = 0;
 173     $continue = true;
 174     $read = fgets ($imap_stream,8192);
 175
 176
 177     // This could be bad -- if the section has sqimap_session_id() . ' OK'
 178     // or similar, it will kill the download.
 179     while (!ereg("^".$sid_s." (OK|BAD|NO)(.*)$", $read, $regs)) {
 180         if (trim($read) == ')==') {
 181             $read1 = $read;
 182             $read = fgets ($imap_stream,4096);
 183             if (ereg("^".$sid." (OK|BAD|NO)(.*)$", $read, $regs)) {
 184                 return;
 185             } else {
 186                 echo decodeBody($read1, $encoding) .
 187                      decodeBody($read, $encoding);
 188             }
 189         } else if ($cnt) {
 190             echo decodeBody($read, $encoding);
 191         }
 192         $read = fgets ($imap_stream,4096);
 193         $cnt++;
 194 //      break;
 195     }
 196 */
 197 }
 198
 199 /* -[ END MIME DECODING ]----------------------------------------------------------- */
 200
 201 /* findDisplayEntity
 202  * Checks to see if $message contains content of type $type0/$type1
 203  * returns the first entity number it finds of that type, or NULL if
 204  * none is found.  Takes optional argument $start to allow the caller
 205  * to continue where they left off
 206  */
 207 function findDisplayEntity($message, $type0, $type1, $start=0) {
 208     if ($message) {
 209         for ($i = $start;isset($message->entities[$i]); $i++) {
 210             $entity = $message->entities[$i];
 211             if ($entity->type0 == $type0 && $entity->type1 == $type1) {
 212                 return $i;
 213             }
 214         }
 215     }
 216     return NULL;
 217 }
 218
 219 // This is here for debugging purposese.  It will print out a list
 220 // of all the entity IDs that are in the $message object.
 221
 222 function listEntities ($message) {
 223   if ($message) {
 224     echo "<tt>" . $message->entity_id . ' : ' . $message->type0 . '/' . $message->type1 . ' parent = '. $message->parent->entity_id. '<br>';
 225     for ($i = 0;isset($message->entities[$i]); $i++) {
 226        echo "$i : ";
 227        $msg = listEntities($message->entities[$i]);
 228
 229        if ($msg) {
 230           echo "return: ";
 231           return $msg;
 232        }
 233     }
 234   }
 235 }
 236
 237 function getPriorityStr($priority) {
 238    $priority_level = substr($priority,0,1);
 239
 240    switch($priority_level) {
 241      /* check for a higher then normal priority. */
 242      case '1':
 243      case '2':
 244         $priority_string = _("High");
 245         break;
 246
 247       /* check for a lower then normal priority. */
 248      case '4':
 249      case '5':
 250         $priority_string = _("Low");
 251         break;
 252
 253      /* check for a normal priority. */
 254      case '3':
 255      default:
 256         $priority_level = '3';
 257         $priority_string = _("Normal");
 258         break;
 259
 260    }
 261    return $priority_string;
 262 }
 263
 264 /* returns a $message object for a particular entity id */
 265 function getEntity ($message, $ent_id) {
 266     return $message->getEntity($ent_id);
 267 }
 268
 269 /*
 270  * translateText
 271  * Extracted from strings.php 23/03/2002
 272  */
 273
 274 function translateText(&$body, $wrap_at, $charset) {
 275     global $where, $what; /* from searching */
 276     global $color; /* color theme */
 277
 278     require_once('../functions/url_parser.php');
 279
 280     $body_ary = explode("\n", $body);
 281     $PriorQuotes = 0;
 282     for ($i=0; $i < count($body_ary); $i++) {
 283         $line = $body_ary[$i];
 284         if (strlen($line) - 2 >= $wrap_at) {
 285             sqWordWrap($line, $wrap_at);
 286         }
 287         $line = charset_decode($charset, $line);
 288         $line = str_replace("\t", '        ', $line);
 289
 290         parseUrl ($line);
 291
 292         $Quotes = 0;
 293         $pos = 0;
 294         $j = strlen( $line );
 295
 296         while ( $pos < $j ) {
 297             if ($line[$pos] == ' ') {
 298                 $pos ++;
 299             } else if (strpos($line, '&gt;', $pos) === $pos) {
 300                 $pos += 4;
 301                 $Quotes ++;
 302             } else {
 303                 break;
 304             }
 305         }
 306
 307         if ($Quotes > 1) {
 308             if (! isset($color[14])) {
 309                 $color[14] = '#FF0000';
 310             }
 311             $line = '<FONT COLOR="' . $color[14] . '">' . $line . '</FONT>';
 312         } elseif ($Quotes) {
 313             if (! isset($color[13])) {
 314                 $color[13] = '#800000';
 315             }
 316             $line = '<FONT COLOR="' . $color[13] . '">' . $line . '</FONT>';
 317         }
 318
 319         $body_ary[$i] = $line;
 320     }
 321     $body = '<pre>' . implode("\n", $body_ary) . '</pre>';
 322 }
 323
 324
 325 /* This returns a parsed string called $body. That string can then
 326 be displayed as the actual message in the HTML. It contains
 327 everything needed, including HTML Tags, Attachments at the
 328 bottom, etc.
 329 */
 330 function formatBody($imap_stream, $message, $color, $wrap_at, $ent_num, $id, $mailbox='INBOX') {
 331     // this if statement checks for the entity to show as the
 332     // primary message. To add more of them, just put them in the
 333     // order that is their priority.
 334     global $startMessage, $username, $key, $imapServerAddress, $imapPort,
 335            $show_html_default, $has_unsafe_images, $view_unsafe_images, $sort;
 336
 337     $has_unsafe_images= 0;
 338     $body = '';
 339     $urlmailbox = urlencode($mailbox);
 340     $body_message = getEntity($message, $ent_num);
 341     if (($body_message->header->type0 == 'text') ||
 342         ($body_message->header->type0 == 'rfc822')) {
 343         $body = mime_fetch_body ($imap_stream, $id, $ent_num);
 344         $body = decodeBody($body, $body_message->header->encoding);
 345         $hookResults = do_hook("message_body", $body);
 346         $body = $hookResults[1];
 347
 348         // If there are other types that shouldn't be formatted, add
 349         // them here
 350         if ($body_message->header->type1 == 'html') {
 351             if ( $show_html_default <> 1 ) {
 352                 $body = strip_tags( $body );
 353                 translateText($body, $wrap_at,
 354                   $body_message->header->getParameter['charset']);
 355             } else {
 356                 $body = magicHTML( $body, $id, $message, $mailbox );
 357             }
 358         } else {
 359             translateText($body, $wrap_at,
 360                $body_message->header->getParameter('charset'));
 361         }
 362
 363         if ($has_unsafe_images) {
 364             if ($view_unsafe_images) {
 365                 $body .= "<CENTER><SMALL><A HREF=\"read_body.php?passed_id=$id&amp;passed_ent_id=".$message->entity_id."&amp;mailbox=$urlmailbox&amp;sort=$sort&amp;startMessage=$startMessage&amp;show_more=0\">". _("Hide Unsafe Images") ."</A></SMALL></CENTER><BR>\n";
 366             } else {
 367                 $body .= "<CENTER><SMALL><A HREF=\"read_body.php?passed_id=$id&amp;passed_ent_id=".$message->entity_id."&amp;mailbox=$urlmailbox&amp;sort=$sort&amp;startMessage=$startMessage&amp;show_more=0&amp;view_unsafe_images=1\">". _("View Unsafe Images") ."</A></SMALL></CENTER><BR>\n";
 368             }
 369         }
 370     }
 371     return ($body);
 372 }
 373
 374
 375 function formatAttachments($message, $exclude_id, $mailbox, $id) {
 376     global $where, $what;
 377     global $startMessage, $color;
 378     static $ShownHTML = 0;
 379
 380     $att_ar = $message->getAttachments($exclude_id);
 381
 382     if (!count($att_ar)) return '';
 383
 384     $attachments = '';
 385
 386     $urlMailbox = urlencode($mailbox);
 387
 388     foreach ($att_ar as $att) {
 389         $ent = urldecode($att->entity_id);
 390         $header = $att->header;
 391         $type0 = strtolower($header->type0);
 392         $type1 = strtolower($header->type1);
 393         $name = '';
 394         $Links['download link']['text'] = _("download");
 395         $Links['download link']['href'] =
 396                 "../src/download.php?absolute_dl=true&amp;passed_id=$id&amp;mailbox=$urlMailbox&amp;ent_id=$ent";
 397         $ImageURL = '';
 398         if ($type0 =='message' && $type1 == 'rfc822') {
 399             $default_page = '../src/read_body.php';
 400             $rfc822_header = $att->rfc822_header;
 401             $filename = decodeHeader($rfc822_header->subject);
 402
 403             $from_o = $rfc822_header->from;
 404             if (is_object($from_o)) {
 405                 $from_name = $from_o->getAddress(false);
 406             } else {
 407                 $from_name = _("Unknown sender");
 408             }
 409             $from_name = decodeHeader(htmlspecialchars($from_name));
 410             $description = $from_name;
 411         } else {
 412             $default_page = '../src/download.php';
 413             if (is_object($header->disposition)) {
 414                $filename = decodeHeader($header->disposition->getProperty('filename'));
 415                if (trim($filename) == '') {
 416                   $name = decodeHeader($header->disposition->getProperty('name'));
 417                   if (trim($name) == '') {
 418                      if ( trim( $header->id ) == '' )
 419                         $filename = 'untitled-[' . $ent . ']' ;
 420                      else
 421                         $filename = 'cid: ' . $header->id;
 422                   } else {
 423                      $filename = $name;
 424                   }
 425                }
 426             } else {
 427                if ( trim( $header->id ) == '' )
 428                   $filename = 'untitled-[' . $ent . ']' ;
 429                else
 430                   $filename = 'cid: ' . $header->id;
 431             }
 432
 433             if ($header->description) {
 434                 $description = htmlspecialchars($header->description);
 435             } else {
 436                 $description = '';
 437             }
 438         }
 439
 440         $display_filename = $filename;
 441         if (isset($passed_ent_id)) {
 442            $passed_ent_id_link = '&amp;passed_ent_id='.$passed_ent_id;
 443         } else {
 444            $passed_ent_id_link = '';
 445         }
 446         $DefaultLink = $default_page . "?startMessage=$startMessage"
 447                      . "&amp;passed_id=$id&amp;mailbox=$urlMailbox"
 448                      . '&amp;ent_id='.$ent.$passed_ent_id_link;
 449         if ($where && $what) {
 450            $DefaultLink = '&amp;where='. urlencode($where).'&amp;what='.urlencode($what);
 451         }
 452         /* this executes the attachment hook with a specific MIME-type.
 453             * if that doens't have results, it tries if there's a rule
 454             * for a more generic type. */
 455         $HookResults = do_hook("attachment $type0/$type1", $Links,
 456             $startMessage, $id, $urlMailbox, $ent, $DefaultLink,
 457             $display_filename, $where, $what);
 458         if(count($HookResults[1]) <= 1) {
 459             $HookResults = do_hook("attachment $type0/*", $Links,
 460                     $startMessage, $id, $urlMailbox, $ent, $DefaultLink,
 461                     $display_filename, $where, $what);
 462         }
 463
 464         $Links = $HookResults[1];
 465         $DefaultLink = $HookResults[6];
 466
 467         $attachments .= '<TR><TD>' .
 468                         "<A HREF=\"$DefaultLink\">$display_filename</A>&nbsp;</TD>" .
 469                         '<TD><SMALL><b>' . show_readable_size($header->size) .
 470                         '</b>&nbsp;&nbsp;</small></TD>' .
 471                         "<TD><SMALL>[ $type0/$type1 ]&nbsp;</SMALL></TD>" .
 472                         '<TD><SMALL>';
 473         $attachments .= '<b>' . $description . '</b>';
 474         $attachments .= '</SMALL></TD><TD><SMALL>&nbsp;';
 475
 476         $SkipSpaces = 1;
 477         foreach ($Links as $Val) {
 478             if ($SkipSpaces) {
 479                 $SkipSpaces = 0;
 480             } else {
 481                 $attachments .= '&nbsp;&nbsp;|&nbsp;&nbsp;';
 482             }
 483             $attachments .= '<a href="' . $Val['href'] . '">' .  $Val['text'] . '</a>';
 484         }
 485         unset($Links);
 486         $attachments .= "</TD></TR>\n";
 487     }
 488     return $attachments;
 489 }
 490
 491 /** this function decodes the body depending on the encoding type. **/
 492 function decodeBody($body, $encoding) {
 493     global $languages, $squirrelmail_language;
 494
 495   $body = str_replace("\r\n", "\n", $body);
 496   $encoding = strtolower($encoding);
 497
 498   global $show_html_default;
 499
 500   if ($encoding == 'quoted-printable' ||
 501       $encoding == 'quoted_printable') {
 502      $body = quoted_printable_decode($body);
 503
 504      while (ereg("=\n", $body))
 505         $body = ereg_replace ("=\n", "", $body);
 506
 507   } else if ($encoding == 'base64') {
 508      $body = base64_decode($body);
 509   }
 510
 511   if (function_exists($languages[$squirrelmail_language]['XTRA_CODE'])) {
 512       $body = $languages[$squirrelmail_language]['XTRA_CODE']('decode', $body);
 513   }
 514
 515   // All other encodings are returned raw.
 516   return $body;
 517 }
 518
 519 /*
 520  * This functions decode strings that is encoded according to
 521  * RFC1522 (MIME Part Two: Message Header Extensions for Non-ASCII Text).
 522  * Patched by Christian Schmidt <christian@ostenfeld.dk>  23/03/2002
 523  */
 524 function decodeHeader ($string, $utfencode=true) {
 525     global $languages, $squirrelmail_language;
 526     if (is_array($string)) {
 527         $string = implode("\n", $string);
 528     }
 529
 530     if (function_exists($languages[$squirrelmail_language]['XTRA_CODE'])) {
 531         $string = $languages[$squirrelmail_language]['XTRA_CODE']('decodeheader', $string);
 532     }
 533
 534     $i = 0;
 535     while (preg_match('/^(.{' . $i . '})(.*)=\?([^?]*)\?(Q|B)\?([^?]*)\?=/Ui',
 536                       $string, $res)) {
 537         $prefix = $res[1];
 538         // Ignore white-space between consecutive encoded-words
 539         if (strspn($res[2], " \t") != strlen($res[2])) {
 540             $prefix .= $res[2];
 541         }
 542
 543         if (ucfirst($res[4]) == 'B') {
 544             $replace = base64_decode($res[5]);
 545         } else {
 546             $replace = str_replace('_', ' ', $res[5]);
 547             $replace = preg_replace('/=([0-9a-f]{2})/ie', 'chr(hexdec("\1"))',
 548                                     $replace);
 549             /* Only encode into entities by default. Some places
 550                don't need the encoding, like the compose form. */
 551             if ($utfencode) {
 552                 $replace = charset_decode($res[3], $replace);
 553             }
 554         }
 555         $string = $prefix . $replace . substr($string, strlen($res[0]));
 556         $i = strlen($prefix) + strlen($replace);
 557     }
 558     return( $string );
 559 }
 560
 561 /*
 562  * Encode a string according to RFC 1522 for use in headers if it
 563  * contains 8-bit characters or anything that looks like it should
 564  * be encoded.
 565  */
 566 function encodeHeader ($string) {
 567     global $default_charset, $languages, $squirrelmail_language;
 568
 569     if (function_exists($languages[$squirrelmail_language]['XTRA_CODE'])) {
 570         return  $languages[$squirrelmail_language]['XTRA_CODE']('encodeheader', $string);
 571     }
 572
 573     // Encode only if the string contains 8-bit characters or =?
 574     $j = strlen( $string  );
 575     $l = strstr($string, '=?');         // Must be encoded ?
 576     $ret = '';
 577     for( $i=0; $i < $j; ++$i) {
 578         switch( $string{$i} ) {
 579            case '=':
 580           $ret .= '=3D';
 581           break;
 582         case '?':
 583           $ret .= '=3F';
 584           break;
 585         case '_':
 586           $ret .= '=5F';
 587           break;
 588         case ' ':
 589           $ret .= '_';
 590           break;
 591         default:
 592           $k = ord( $string{$i} );
 593           if ( $k > 126 ) {
 594              $ret .= sprintf("=%02X", $k);
 595              $l = TRUE;
 596           } else
 597              $ret .= $string{$i};
 598         }
 599     }
 600
 601     if ( $l ) {
 602         $string = "=?$default_charset?Q?$ret?=";
 603     }
 604
 605     return( $string );
 606 }
 607
 608 /* This function trys to locate the entity_id of a specific mime element */
 609
 610 function find_ent_id( $id, $message ) {
 611     $ret = '';
 612     for ($i=0; $ret == '' && $i < count($message->entities); $i++) {
 613         if ( $message->entities[$i]->header->type0 == 'multipart')  {
 614             $ret = find_ent_id( $id, $message->entities[$i] );
 615         } else {
 616             if ( strcasecmp( $message->entities[$i]->header->id, $id ) == 0 )
 617                 $ret = $message->entities[$i]->entity_id;
 618         }
 619     }
 620     return( $ret );
 621 }
 622
 623 /**
 624  ** HTMLFILTER ROUTINES
 625  */
 626
 627 /**
 628  * This function returns the final tag out of the tag name, an array
 629  * of attributes, and the type of the tag. This function is called by
 630  * sq_sanitize internally.
 631  *
 632  * @param  $tagname  the name of the tag.
 633  * @param  $attary   the array of attributes and their values
 634  * @param  $tagtype  The type of the tag (see in comments).
 635  * @return           a string with the final tag representation.
 636  */
 637 function sq_tagprint($tagname, $attary, $tagtype){
 638     $me = "sq_tagprint";
 639     if ($tagtype == 2){
 640         $fulltag = '</' . $tagname . '>';
 641     } else {
 642         $fulltag = '<' . $tagname;
 643         if (is_array($attary) && sizeof($attary)){
 644             $atts = Array();
 645             while (list($attname, $attvalue) = each($attary)){
 646                 array_push($atts, "$attname=$attvalue");
 647             }
 648             $fulltag .= ' ' . join(" ", $atts);
 649         }
 650         if ($tagtype == 3){
 651             $fulltag .= " /";
 652         }
 653         $fulltag .= ">";
 654     }
 655     return $fulltag;
 656 }
 657
 658 /**
 659  * A small helper function to use with array_walk. Modifies a by-ref
 660  * value and makes it lowercase.
 661  *
 662  * @param  $val a value passed by-ref.
 663  * @return      void since it modifies a by-ref value.
 664  */
 665 function sq_casenormalize(&$val){
 666     $val = strtolower($val);
 667 }
 668
 669 /**
 670  * This function skips any whitespace from the current position within
 671  * a string and to the next non-whitespace value.
 672  *
 673  * @param  $body   the string
 674  * @param  $offset the offset within the string where we should start
 675  *                 looking for the next non-whitespace character.
 676  * @return         the location within the $body where the next
 677  *                 non-whitespace char is located.
 678  */
 679 function sq_skipspace($body, $offset){
 680     $me = "sq_skipspace";
 681     preg_match("/^(\s*)/s", substr($body, $offset), $matches);
 682     if (sizeof($matches{1})){
 683         $count = strlen($matches{1});
 684         $offset += $count;
 685     }
 686     return $offset;
 687 }
 688
 689 /**
 690  * This function looks for the next character within a string.  It's
 691  * really just a glorified "strpos", except it catches if failures
 692  * nicely.
 693  *
 694  * @param  $body   The string to look for needle in.
 695  * @param  $offset Start looking from this position.
 696  * @param  $needle The character/string to look for.
 697  * @return         location of the next occurance of the needle, or
 698  *                 strlen($body) if needle wasn't found.
 699  */
 700 function sq_findnxstr($body, $offset, $needle){
 701     $me = "sq_findnxstr";
 702     $pos = strpos($body, $needle, $offset);
 703     if ($pos === FALSE){
 704         $pos = strlen($body);
 705     }
 706     return $pos;
 707 }
 708
 709 /**
 710  * This function takes a PCRE-style regexp and tries to match it
 711  * within the string.
 712  *
 713  * @param  $body   The string to look for needle in.
 714  * @param  $offset Start looking from here.
 715  * @param  $reg    A PCRE-style regex to match.
 716  * @return         Returns a false if no matches found, or an array
 717  *                 with the following members:
 718  *                 - integer with the location of the match within $body
 719  *                 - string with whatever content between offset and the match
 720  *                 - string with whatever it is we matched
 721  */
 722 function sq_findnxreg($body, $offset, $reg){
 723     $me = "sq_findnxreg";
 724     $matches = Array();
 725     $retarr = Array();
 726     preg_match("%^(.*?)($reg)%s", substr($body, $offset), $matches);
 727     if (!$matches{0}){
 728         $retarr = false;
 729     } else {
 730         $retarr{0} = $offset + strlen($matches{1});
 731         $retarr{1} = $matches{1};
 732         $retarr{2} = $matches{2};
 733     }
 734     return $retarr;
 735 }
 736
 737 /**
 738  * This function looks for the next tag.
 739  *
 740  * @param  $body   String where to look for the next tag.
 741  * @param  $offset Start looking from here.
 742  * @return         false if no more tags exist in the body, or
 743  *                 an array with the following members:
 744  *                 - string with the name of the tag
 745  *                 - array with attributes and their values
 746  *                 - integer with tag type (1, 2, or 3)
 747  *                 - integer where the tag starts (starting "<")
 748  *                 - integer where the tag ends (ending ">")
 749  *                 first three members will be false, if the tag is invalid.
 750  */
 751 function sq_getnxtag($body, $offset){
 752     $me = "sq_getnxtag";
 753     if ($offset > strlen($body)){
 754         return false;
 755     }
 756     $lt = sq_findnxstr($body, $offset, "<");
 757     if ($lt == strlen($body)){
 758         return false;
 759     }
 760     /**
 761      * We are here:
 762      * blah blah <tag attribute="value">
 763      * \---------^
 764      */
 765     $pos = sq_skipspace($body, $lt+1);
 766     if ($pos >= strlen($body)){
 767         return Array(false, false, false, $lt, strlen($body));
 768     }
 769     /**
 770      * There are 3 kinds of tags:
 771      * 1. Opening tag, e.g.:
 772      *    <a href="blah">
 773      * 2. Closing tag, e.g.:
 774      *    </a>
 775      * 3. XHTML-style content-less tag, e.g.:
 776      *    <img src="blah"/>
 777      */
 778     $tagtype = false;
 779     switch (substr($body, $pos, 1)){
 780     case "/":
 781         $tagtype = 2;
 782         $pos++;
 783         break;
 784     case "!":
 785         /**
 786          * A comment or an SGML declaration.
 787          */
 788         if (substr($body, $pos+1, 2) == "--"){
 789             $gt = strpos($body, "-->", $pos);
 790             if ($gt === false){
 791                 $gt = strlen($body);
 792             } else {
 793                 $gt += 2;
 794             }
 795             return Array(false, false, false, $lt, $gt);
 796         } else {
 797             $gt = sq_findnxstr($body, $pos, ">");
 798             return Array(false, false, false, $lt, $gt);
 799         }
 800         break;
 801     default:
 802         /**
 803          * Assume tagtype 1 for now. If it's type 3, we'll switch values
 804          * later.
 805          */
 806         $tagtype = 1;
 807         break;
 808     }
 809
 810     $tag_start = $pos;
 811     $tagname = '';
 812     /**
 813      * Look for next [\W-_], which will indicate the end of the tag name.
 814      */
 815     $regary = sq_findnxreg($body, $pos, "[^\w\-_]");
 816     if ($regary == false){
 817         return Array(false, false, false, $lt, strlen($body));
 818     }
 819     list($pos, $tagname, $match) = $regary;
 820     $tagname = strtolower($tagname);
 821
 822     /**
 823      * $match can be either of these:
 824      * '>'  indicating the end of the tag entirely.
 825      * '\s' indicating the end of the tag name.
 826      * '/'  indicating that this is type-3 xhtml tag.
 827      *
 828      * Whatever else we find there indicates an invalid tag.
 829      */
 830     switch ($match){
 831     case "/":
 832         /**
 833          * This is an xhtml-style tag with a closing / at the
 834          * end, like so: <img src="blah"/>. Check if it's followed
 835          * by the closing bracket. If not, then this tag is invalid
 836          */
 837         if (substr($body, $pos, 2) == "/>"){
 838             $pos++;
 839             $tagtype = 3;
 840         } else {
 841             $gt = sq_findnxstr($body, $pos, ">");
 842             $retary = Array(false, false, false, $lt, $gt);
 843             return $retary;
 844         }
 845     case ">":
 846         return Array($tagname, false, $tagtype, $lt, $pos);
 847         break;
 848     default:
 849         /**
 850          * Check if it's whitespace
 851          */
 852         if (preg_match("/\s/", $match)){
 853         } else {
 854             /**
 855              * This is an invalid tag! Look for the next closing ">".
 856              */
 857             $gt = sq_findnxstr($body, $offset, ">");
 858             return Array(false, false, false, $lt, $gt);
 859         }
 860     }
 861
 862     /**
 863      * At this point we're here:
 864      * <tagname  attribute='blah'>
 865      * \-------^
 866      *
 867      * At this point we loop in order to find all attributes.
 868      */
 869     $attname = '';
 870     $atttype = false;
 871     $attary = Array();
 872
 873     while ($pos <= strlen($body)){
 874         $pos = sq_skipspace($body, $pos);
 875         if ($pos == strlen($body)){
 876             /**
 877              * Non-closed tag.
 878              */
 879             return Array(false, false, false, $lt, $pos);
 880         }
 881         /**
 882          * See if we arrived at a ">" or "/>", which means that we reached
 883          * the end of the tag.
 884          */
 885         $matches = Array();
 886         if (preg_match("%^(\s*)(>|/>)%s", substr($body, $pos), $matches)) {
 887             /**
 888              * Yep. So we did.
 889              */
 890             $pos += strlen($matches{1});
 891             if ($matches{2} == "/>"){
 892                 $tagtype = 3;
 893                 $pos++;
 894             }
 895             return Array($tagname, $attary, $tagtype, $lt, $pos);
 896         }
 897
 898         /**
 899          * There are several types of attributes, with optional
 900          * [:space:] between members.
 901          * Type 1:
 902          *   attrname[:space:]=[:space:]'CDATA'
 903          * Type 2:
 904          *   attrname[:space:]=[:space:]"CDATA"
 905          * Type 3:
 906          *   attr[:space:]=[:space:]CDATA
 907          * Type 4:
 908          *   attrname
 909          *
 910          * We leave types 1 and 2 the same, type 3 we check for
 911          * '"' and convert to "&quot" if needed, then wrap in
 912          * double quotes. Type 4 we convert into:
 913          * attrname="yes".
 914          */
 915         $regary = sq_findnxreg($body, $pos, "[^\w\-_]");
 916         if ($regary == false){
 917             /**
 918              * Looks like body ended before the end of tag.
 919              */
 920             return Array(false, false, false, $lt, strlen($body));
 921         }
 922         list($pos, $attname, $match) = $regary;
 923         $attname = strtolower($attname);
 924         /**
 925          * We arrived at the end of attribute name. Several things possible
 926          * here:
 927          * '>'  means the end of the tag and this is attribute type 4
 928          * '/'  if followed by '>' means the same thing as above
 929          * '\s' means a lot of things -- look what it's followed by.
 930          *      anything else means the attribute is invalid.
 931          */
 932         switch($match){
 933         case "/":
 934             /**
 935              * This is an xhtml-style tag with a closing / at the
 936              * end, like so: <img src="blah"/>. Check if it's followed
 937              * by the closing bracket. If not, then this tag is invalid
 938              */
 939             if (substr($body, $pos, 2) == "/>"){
 940                 $pos++;
 941                 $tagtype = 3;
 942             } else {
 943                 $gt = sq_findnxstr($body, $pos, ">");
 944                 $retary = Array(false, false, false, $lt, $gt);
 945                 return $retary;
 946             }
 947         case ">":
 948             $attary{$attname} = '"yes"';
 949             return Array($tagname, $attary, $tagtype, $lt, $pos);
 950             break;
 951         default:
 952             /**
 953              * Skip whitespace and see what we arrive at.
 954              */
 955             $pos = sq_skipspace($body, $pos);
 956             $char = substr($body, $pos, 1);
 957             /**
 958              * Two things are valid here:
 959              * '=' means this is attribute type 1 2 or 3.
 960              * \w means this was attribute type 4.
 961              * anything else we ignore and re-loop. End of tag and
 962              * invalid stuff will be caught by our checks at the beginning
 963              * of the loop.
 964              */
 965             if ($char == "="){
 966                 $pos++;
 967                 $pos = sq_skipspace($body, $pos);
 968                 /**
 969                  * Here are 3 possibilities:
 970                  * "'"  attribute type 1
 971                  * '"'  attribute type 2
 972                  * everything else is the content of tag type 3
 973                  */
 974                 $quot = substr($body, $pos, 1);
 975                 if ($quot == "'"){
 976                     $regary = sq_findnxreg($body, $pos+1, "\'");
 977                     if ($regary == false){
 978                         return Array(false, false, false, $lt, strlen($body));
 979                     }
 980                     list($pos, $attval, $match) = $regary;
 981                     $pos++;
 982                     $attary{$attname} = "'" . $attval . "'";
 983                 } else if ($quot == '"'){
 984                     $regary = sq_findnxreg($body, $pos+1, '\"');
 985                     if ($regary == false){
 986                         return Array(false, false, false, $lt, strlen($body));
 987                     }
 988                     list($pos, $attval, $match) = $regary;
 989                     $pos++;
 990                     $attary{$attname} = '"' . $attval . '"';
 991                 } else {
 992                     /**
 993                      * These are hateful. Look for \s, or >.
 994                      */
 995                     $regary = sq_findnxreg($body, $pos, "[\s>]");
 996                     if ($regary == false){
 997                         return Array(false, false, false, $lt, strlen($body));
 998                     }
 999                     list($pos, $attval, $match) = $regary;
1000                     /**
1001                      * If it's ">" it will be caught at the top.
1002                      */
1003                     $attval = preg_replace("/\"/s", "&quot;", $attval);
1004                     $attary{$attname} = '"' . $attval . '"';
1005                 }
1006             } else if (preg_match("|[\w/>]|", $char)) {
1007                 /**
1008                  * That was attribute type 4.
1009                  */
1010                 $attary{$attname} = '"yes"';
1011             } else {
1012                 /**
1013                  * An illegal character. Find next '>' and return.
1014                  */
1015                 $gt = sq_findnxstr($body, $pos, ">");
1016                 return Array(false, false, false, $lt, $gt);
1017             }
1018         }
1019     }
1020     /**
1021      * The fact that we got here indicates that the tag end was never
1022      * found. Return invalid tag indication so it gets stripped.
1023      */
1024     return Array(false, false, false, $lt, strlen($body));
1025 }
1026
1027 /**
1028  * This function checks attribute values for entity-encoded values
1029  * and returns them translated into 8-bit strings so we can run
1030  * checks on them.
1031  *
1032  * @param  $attvalue A string to run entity check against.
1033  * @return           Translated value.
1034  */
1035 function sq_deent($attvalue){
1036     $me="sq_deent";
1037     /**
1038      * See if we have to run the checks first. All entities must start
1039      * with "&".
1040      */
1041     if (strpos($attvalue, "&") === false){
1042         return $attvalue;
1043     }
1044     /**
1045      * Check named entities first.
1046      */
1047     $trans = get_html_translation_table(HTML_ENTITIES);
1048     /**
1049      * Leave &quot; in, as it can mess us up.
1050      */
1051     $trans = array_flip($trans);
1052     unset($trans{"&quot;"});
1053     while (list($ent, $val) = each($trans)){
1054         $attvalue = preg_replace("/$ent*(\W)/si", "$val\\1", $attvalue);
1055     }
1056     /**
1057      * Now translate numbered entities from 1 to 255 if needed.
1058      */
1059     if (strpos($attvalue, "#") !== false){
1060         $omit = Array(34, 39);
1061         for ($asc=1; $asc<256; $asc++){
1062             if (!in_array($asc, $omit)){
1063                 $chr = chr($asc);
1064                 $attvalue = preg_replace("/\&#0*$asc;*(\D)/si", "$chr\\1",
1065                                          $attvalue);
1066                 $attvalue = preg_replace("/\&#x0*".dechex($asc).";*(\W)/si",
1067                                          "$chr\\1", $attvalue);
1068             }
1069         }
1070     }
1071     return $attvalue;
1072 }
1073
1074 /**
1075  * This function runs various checks against the attributes.
1076  *
1077  * @param  $tagname         String with the name of the tag.
1078  * @param  $attary          Array with all tag attributes.
1079  * @param  $rm_attnames     See description for sq_sanitize
1080  * @param  $bad_attvals     See description for sq_sanitize
1081  * @param  $add_attr_to_tag See description for sq_sanitize
1082  * @param  $message         message object
1083  * @param  $id              message id
1084  * @return                  Array with modified attributes.
1085  */
1086 function sq_fixatts($tagname,
1087                     $attary,
1088                     $rm_attnames,
1089                     $bad_attvals,
1090                     $add_attr_to_tag,
1091                     $message,
1092                     $id,
1093                     $mailbox
1094                     ){
1095     $me = "sq_fixatts";
1096     while (list($attname, $attvalue) = each($attary)){
1097         /**
1098          * See if this attribute should be removed.
1099          */
1100         foreach ($rm_attnames as $matchtag=>$matchattrs){
1101             if (preg_match($matchtag, $tagname)){
1102                 foreach ($matchattrs as $matchattr){
1103                     if (preg_match($matchattr, $attname)){
1104                         unset($attary{$attname});
1105                         continue;
1106                     }
1107                 }
1108             }
1109         }
1110         /**
1111          * Remove any entities.
1112          */
1113         $attvalue = sq_deent($attvalue);
1114
1115         /**
1116          * Now let's run checks on the attvalues.
1117          * I don't expect anyone to comprehend this. If you do,
1118          * get in touch with me so I can drive to where you live and
1119          * shake your hand personally. :)
1120          */
1121         foreach ($bad_attvals as $matchtag=>$matchattrs){
1122             if (preg_match($matchtag, $tagname)){
1123                 foreach ($matchattrs as $matchattr=>$valary){
1124                     if (preg_match($matchattr, $attname)){
1125                         /**
1126                          * There are two arrays in valary.
1127                          * First is matches.
1128                          * Second one is replacements
1129                          */
1130                         list($valmatch, $valrepl) = $valary;
1131                         $newvalue =
1132                             preg_replace($valmatch, $valrepl, $attvalue);
1133                         if ($newvalue != $attvalue){
1134                             $attary{$attname} = $newvalue;
1135                         }
1136                     }
1137                 }
1138             }
1139         }
1140         /**
1141          * Turn cid: urls into http-friendly ones.
1142          */
1143         if (preg_match("/^[\'\"]\s*cid:/si", $attvalue)){
1144             $attary{$attname} = sq_cid2http($message, $id, $attvalue, $mailbox);
1145         }
1146     }
1147     /**
1148      * See if we need to append any attributes to this tag.
1149      */
1150     foreach ($add_attr_to_tag as $matchtag=>$addattary){
1151         if (preg_match($matchtag, $tagname)){
1152             $attary = array_merge($attary, $addattary);
1153         }
1154     }
1155     return $attary;
1156 }
1157
1158 /**
1159  * This function edits the style definition to make them friendly and
1160  * usable in squirrelmail.
1161  *
1162  * @param  $message  the message object
1163  * @param  $id       the message id
1164  * @param  $content  a string with whatever is between <style> and </style>
1165  * @return           a string with edited content.
1166  */
1167 function sq_fixstyle($message, $id, $content){
1168     global $view_unsafe_images;
1169     $me = "sq_fixstyle";
1170     /**
1171      * First look for general BODY style declaration, which would be
1172      * like so:
1173      * body {background: blah-blah}
1174      * and change it to .bodyclass so we can just assign it to a <div>
1175      */
1176     $content = preg_replace("|body(\s*\{.*?\})|si", ".bodyclass\\1", $content);
1177     $secremoveimg = "../images/" . _("sec_remove_eng.png");
1178     /**
1179      * Fix url('blah') declarations.
1180      */
1181     $content = preg_replace("|url\(([\'\"])\s*\S+script\s*:.*?([\'\"])\)|si",
1182                             "url(\\1$secremoveimg\\2)", $content);
1183     /**
1184      * Fix url('https*://.*) declarations but only if $view_unsafe_images
1185      * is false.
1186      */
1187     if (!$view_unsafe_images){
1188         $content = preg_replace("|url\(([\'\"])\s*https*:.*?([\'\"])\)|si",
1189                                 "url(\\1$secremoveimg\\2)", $content);
1190     }
1191
1192     /**
1193      * Fix urls that refer to cid:
1194      */
1195     while (preg_match("|url\(([\'\"]\s*cid:.*?[\'\"])\)|si", $content,
1196                       $matches)){
1197         $cidurl = $matches{1};
1198         $httpurl = sq_cid2http($message, $id, $cidurl);
1199         $content = preg_replace("|url\($cidurl\)|si",
1200                                 "url($httpurl)", $content);
1201     }
1202
1203     /**
1204      * Fix stupid css declarations which lead to vulnerabilities
1205      * in IE.
1206      */
1207     $match   = Array('/expression/si',
1208                      '/behaviou*r/si',
1209                      '/binding/si');
1210     $replace = Array('idiocy', 'idiocy', 'idiocy');
1211     $content = preg_replace($match, $replace, $content);
1212     return $content;
1213 }
1214
1215 /**
1216  * This function converts cid: url's into the ones that can be viewed in
1217  * the browser.
1218  *
1219  * @param  $message  the message object
1220  * @param  $id       the message id
1221  * @param  $cidurl   the cid: url.
1222  * @return           a string with a http-friendly url
1223  */
1224 function sq_cid2http($message, $id, $cidurl, $mailbox){
1225     /**
1226      * Get rid of quotes.
1227      */
1228     $quotchar = substr($cidurl, 0, 1);
1229     $cidurl = str_replace($quotchar, "", $cidurl);
1230     $cidurl = substr(trim($cidurl), 4);
1231     $httpurl = $quotchar . "../src/download.php?absolute_dl=true&amp;" .
1232         "passed_id=$id&amp;mailbox=" . urlencode($mailbox) .
1233         "&amp;ent_id=" . find_ent_id($cidurl, $message) . $quotchar;
1234     return $httpurl;
1235 }
1236
1237 /**
1238  * This function changes the <body> tag into a <div> tag since we
1239  * can't really have a body-within-body.
1240  *
1241  * @param  $attary  an array of attributes and values of <body>
1242  * @return          a modified array of attributes to be set for <div>
1243  */
1244 function sq_body2div($attary){
1245     $me = "sq_body2div";
1246     $divattary = Array("class"=>"'bodyclass'");
1247     $bgcolor="#ffffff";
1248     $text="#000000";
1249     $styledef="";
1250     if (is_array($attary) && sizeof($attary) > 0){
1251         foreach ($attary as $attname=>$attvalue){
1252             $quotchar = substr($attvalue, 0, 1);
1253             $attvalue = str_replace($quotchar, "", $attvalue);
1254             switch ($attname){
1255             case "background":
1256                 $styledef .= "background-image: url('$attvalue'); ";
1257                 break;
1258             case "bgcolor":
1259                 $styledef .= "background-color: $attvalue; ";
1260                 break;
1261             case "text":
1262                 $styledef .= "color: $attvalue; ";
1263             }
1264         }
1265         if (strlen($styledef) > 0){
1266             $divattary{"style"} = "\"$styledef\"";
1267         }
1268     }
1269     return $divattary;
1270 }
1271
1272 /**
1273  * This is the main function and the one you should actually be calling.
1274  * There are several variables you should be aware of an which need
1275  * special description.
1276  *
1277  * Since the description is quite lengthy, see it here:
1278  * http://www.mricon.com/html/phpfilter.html
1279  *
1280  * @param $body                 the string with HTML you wish to filter
1281  * @param $tag_list             see description above
1282  * @param $rm_tags_with_content see description above
1283  * @param $self_closing_tags    see description above
1284  * @param $force_tag_closing    see description above
1285  * @param $rm_attnames          see description above
1286  * @param $bad_attvals          see description above
1287  * @param $add_attr_to_tag      see description above
1288  * @param $message              message object
1289  * @param $id                   message id
1290  * @return                      sanitized html safe to show on your pages.
1291  */
1292 function sq_sanitize($body,
1293                      $tag_list,
1294                      $rm_tags_with_content,
1295                      $self_closing_tags,
1296                      $force_tag_closing,
1297                      $rm_attnames,
1298                      $bad_attvals,
1299                      $add_attr_to_tag,
1300                      $message,
1301                      $id,
1302                      $mailbox
1303                      ){
1304     $me = "sq_sanitize";
1305     /**
1306      * Normalize rm_tags and rm_tags_with_content.
1307      */
1308     @array_walk($rm_tags, 'sq_casenormalize');
1309     @array_walk($rm_tags_with_content, 'sq_casenormalize');
1310     @array_walk($self_closing_tags, 'sq_casenormalize');
1311     /**
1312      * See if tag_list is of tags to remove or tags to allow.
1313      * false  means remove these tags
1314      * true   means allow these tags
1315      */
1316     $rm_tags = array_shift($tag_list);
1317     $curpos = 0;
1318     $open_tags = Array();
1319     $trusted = "<!-- begin sanitized html -->\n";
1320     $skip_content = false;
1321     /**
1322      * Take care of netscape's stupid javascript entities like
1323      * &{alert('boo')};
1324      */
1325     $body = preg_replace("/&(\{.*?\};)/si", "&amp;\\1", $body);
1326
1327     while (($curtag=sq_getnxtag($body, $curpos)) != FALSE){
1328         list($tagname, $attary, $tagtype, $lt, $gt) = $curtag;
1329         $free_content = substr($body, $curpos, $lt-$curpos);
1330         /**
1331          * Take care of <style>
1332          */
1333         if ($tagname == "style" && $tagtype == 2){
1334             /**
1335              * This is a closing </style>. Edit the
1336              * content before we apply it.
1337              */
1338             $free_content = sq_fixstyle($message, $id, $free_content);
1339         }
1340         if ($skip_content == false){
1341             $trusted .= $free_content;
1342         } else {
1343         }
1344         if ($tagname != FALSE){
1345             if ($tagtype == 2){
1346                 if ($skip_content == $tagname){
1347                     /**
1348                      * Got to the end of tag we needed to remove.
1349                      */
1350                     $tagname = false;
1351                     $skip_content = false;
1352                 } else {
1353                     if ($skip_content == false){
1354                         if ($tagname == "body"){
1355                             $tagname = "div";
1356                         } else {
1357                             if (isset($open_tags{$tagname}) &&
1358                                 $open_tags{$tagname} > 0){
1359                                 $open_tags{$tagname}--;
1360                             } else {
1361                                 $tagname = false;
1362                             }
1363                         }
1364                     } else {
1365                     }
1366                 }
1367             } else {
1368                 /**
1369                  * $rm_tags_with_content
1370                  */
1371                 if ($skip_content == false){
1372                     /**
1373                      * See if this is a self-closing type and change
1374                      * tagtype appropriately.
1375                      */
1376                     if ($tagtype == 1
1377                         && in_array($tagname, $self_closing_tags)){
1378                         $tagtype=3;
1379                     }
1380                     /**
1381                      * See if we should skip this tag and any content
1382                      * inside it.
1383                      */
1384                     if ($tagtype == 1 &&
1385                         in_array($tagname, $rm_tags_with_content)){
1386                         $skip_content = $tagname;
1387                     } else {
1388                         if (($rm_tags == false
1389                              && in_array($tagname, $tag_list)) ||
1390                             ($rm_tags == true &&
1391                              !in_array($tagname, $tag_list))){
1392                             $tagname = false;
1393                         } else {
1394                             if ($tagtype == 1){
1395                                 if (isset($open_tags{$tagname})){
1396                                     $open_tags{$tagname}++;
1397                                 } else {
1398                                     $open_tags{$tagname}=1;
1399                                 }
1400                             }
1401                             /**
1402                              * This is where we run other checks.
1403                              */
1404                             if (is_array($attary) && sizeof($attary) > 0){
1405                                 $attary = sq_fixatts($tagname,
1406                                                      $attary,
1407                                                      $rm_attnames,
1408                                                      $bad_attvals,
1409                                                      $add_attr_to_tag,
1410                                                      $message,
1411                                                      $id,
1412                                                      $mailbox
1413                                                      );
1414                             }
1415                             /**
1416                              * Convert body into div.
1417                              */
1418                             if ($tagname == "body"){
1419                                 $tagname = "div";
1420                                 $attary = sq_body2div($attary, $message, $id);
1421                             }
1422                         }
1423                     }
1424                 } else {
1425                 }
1426             }
1427             if ($tagname != false && $skip_content == false){
1428                 $trusted .= sq_tagprint($tagname, $attary, $tagtype);
1429             }
1430         } else {
1431         }
1432         $curpos = $gt+1;
1433     }
1434     $trusted .= substr($body, $curpos, strlen($body)-$curpos);
1435     if ($force_tag_closing == true){
1436         foreach ($open_tags as $tagname=>$opentimes){
1437             while ($opentimes > 0){
1438                 $trusted .= '</' . $tagname . '>';
1439                 $opentimes--;
1440             }
1441         }
1442         $trusted .= "\n";
1443     }
1444     $trusted .= "<!-- end sanitized html -->\n";
1445     return $trusted;
1446 }
1447
1448 /**
1449  * This is a wrapper function to call html sanitizing routines.
1450  *
1451  * @param  $body  the body of the message
1452  * @param  $id    the id of the message
1453  * @return        a string with html safe to display in the browser.
1454  */
1455 function magicHTML($body, $id, $message, $mailbox = 'INBOX'){
1456     global $attachment_common_show_images, $view_unsafe_images,
1457         $has_unsafe_images;
1458     /**
1459      * Don't display attached images in HTML mode.
1460      */
1461     $attachment_common_show_images = false;
1462     $tag_list = Array(
1463                       false,
1464                       "object",
1465                       "meta",
1466                       "html",
1467                       "head",
1468                       "base",
1469                       "link"
1470                       );
1471
1472     $rm_tags_with_content = Array(
1473                                   "script",
1474                                   "applet",
1475                                   "embed",
1476                                   "title"
1477                                   );
1478
1479     $self_closing_tags =  Array(
1480                                 "img",
1481                                 "br",
1482                                 "hr",
1483                                 "input"
1484                                 );
1485
1486     $force_tag_closing = false;
1487
1488     $rm_attnames = Array(
1489                          "/.*/" =>
1490                          Array(
1491                                "/target/si",
1492                                "/^on.*/si",
1493                                "/^dynsrc/si",
1494                                "/^data.*/si"
1495                                )
1496                          );
1497
1498     $secremoveimg = "../images/" . _("sec_remove_eng.png");
1499     $bad_attvals = Array(
1500         "/.*/" =>
1501             Array(
1502                 "/^src|background/i" =>
1503                     Array(
1504                           Array(
1505                                 "|^([\'\"])\s*\.\./.*([\'\"])|si",
1506                                 "/^([\'\"])\s*\S+script\s*:.*([\'\"])/si",
1507                                 "/^([\'\"])\s*mocha\s*:*.*([\'\"])/si",
1508                                 "/^([\'\"])\s*about\s*:.*([\'\"])/si"
1509                                 ),
1510                           Array(
1511                                 "\\1$secremoveimg\\2",
1512                                 "\\1$secremoveimg\\2",
1513                                 "\\1$secremoveimg\\2",
1514                                 "\\1$secremoveimg\\2"
1515                                 )
1516                         ),
1517                 "/^href|action/i" =>
1518                     Array(
1519                           Array(
1520                                 "|^([\'\"])\s*\.\./.*([\'\"])|si",
1521                                 "/^([\'\"])\s*\S+script\s*:.*([\'\"])/si",
1522                                 "/^([\'\"])\s*mocha\s*:*.*([\'\"])/si",
1523                                 "/^([\'\"])\s*about\s*:.*([\'\"])/si"
1524                                 ),
1525                           Array(
1526                                 "\\1#\\2",
1527                                 "\\1#\\2",
1528                                 "\\1#\\2",
1529                                 "\\1#\\2"
1530                                 )
1531                         ),
1532                 "/^style/si" =>
1533                     Array(
1534                           Array(
1535                                 "/expression/si",
1536                                 "/binding/si",
1537                                 "/behaviou*r/si",
1538                                 "|url\(([\'\"])\s*\.\./.*([\'\"])\)|si",
1539                                 "/url\(([\'\"])\s*\S+script\s*:.*([\'\"])\)/si",
1540                                 "/url\(([\'\"])\s*mocha\s*:.*([\'\"])\)/si",
1541                                 "/url\(([\'\"])\s*about\s*:.*([\'\"])\)/si"
1542                                ),
1543                           Array(
1544                                 "idiocy",
1545                                 "idiocy",
1546                                 "idiocy",
1547                                 "url(\\1#\\2)",
1548                                 "url(\\1#\\2)",
1549                                 "url(\\1#\\2)",
1550                                 "url(\\1#\\2)"
1551                                )
1552                           )
1553                 )
1554         );
1555     if (!$view_unsafe_images){
1556         /**
1557          * Remove any references to http/https if view_unsafe_images set
1558          * to false.
1559          */
1560          array_push($bad_attvals{'/.*/'}{'/^src|background/i'}[0],
1561                     '/^([\'\"])\s*https*:.*([\'\"])/si');
1562          array_push($bad_attvals{'/.*/'}{'/^src|background/i'}[1],
1563                     "\\1$secremoveimg\\2");
1564          array_push($bad_attvals{'/.*/'}{'/^style/si'}[0],
1565                     '/url\(([\'\"])\s*https*:.*([\'\"])\)/si');
1566          array_push($bad_attvals{'/.*/'}{'/^style/si'}[1],
1567                     "url(\\1$secremoveimg\\2)");
1568     }
1569
1570     $add_attr_to_tag = Array(
1571                              "/^a$/si" => Array('target'=>'"_new"')
1572                              );
1573     $trusted = sq_sanitize($body,
1574                            $tag_list,
1575                            $rm_tags_with_content,
1576                            $self_closing_tags,
1577                            $force_tag_closing,
1578                            $rm_attnames,
1579                            $bad_attvals,
1580                            $add_attr_to_tag,
1581                            $message,
1582                            $id,
1583                            $mailbox
1584                            );
1585     if (preg_match("|$secremoveimg|si", $trusted)){
1586         $has_unsafe_images = true;
1587     }
1588     return $trusted;
1589 }
1590
1591 ?>