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