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 (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
 512       function_exists($languages[$squirrelmail_language]['XTRA_CODE'])) {
 513       $body = $languages[$squirrelmail_language]['XTRA_CODE']('decode', $body);
 514   }
 515
 516   // All other encodings are returned raw.
 517   return $body;
 518 }
 519
 520 /*
 521  * This functions decode strings that is encoded according to
 522  * RFC1522 (MIME Part Two: Message Header Extensions for Non-ASCII Text).
 523  * Patched by Christian Schmidt <christian@ostenfeld.dk>  23/03/2002
 524  */
 525 function decodeHeader ($string, $utfencode=true) {
 526     global $languages, $squirrelmail_language;
 527     if (is_array($string)) {
 528         $string = implode("\n", $string);
 529     }
 530
 531     if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
 532         function_exists($languages[$squirrelmail_language]['XTRA_CODE'])) {
 533         $string = $languages[$squirrelmail_language]['XTRA_CODE']('decodeheader', $string);
 534     }
 535
 536     $i = 0;
 537     while (preg_match('/^(.{' . $i . '})(.*)=\?([^?]*)\?(Q|B)\?([^?]*)\?=/Ui',
 538                       $string, $res)) {
 539         $prefix = $res[1];
 540         // Ignore white-space between consecutive encoded-words
 541         if (strspn($res[2], " \t") != strlen($res[2])) {
 542             $prefix .= $res[2];
 543         }
 544
 545         if (ucfirst($res[4]) == 'B') {
 546             $replace = base64_decode($res[5]);
 547         } else {
 548             $replace = str_replace('_', ' ', $res[5]);
 549             $replace = preg_replace('/=([0-9a-f]{2})/ie', 'chr(hexdec("\1"))',
 550                                     $replace);
 551             /* Only encode into entities by default. Some places
 552                don't need the encoding, like the compose form. */
 553             if ($utfencode) {
 554                 $replace = charset_decode($res[3], $replace);
 555             }
 556         }
 557         $string = $prefix . $replace . substr($string, strlen($res[0]));
 558         $i = strlen($prefix) + strlen($replace);
 559     }
 560     return( $string );
 561 }
 562
 563 /*
 564  * Encode a string according to RFC 1522 for use in headers if it
 565  * contains 8-bit characters or anything that looks like it should
 566  * be encoded.
 567  */
 568 function encodeHeader ($string) {
 569     global $default_charset, $languages, $squirrelmail_language;
 570
 571     if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
 572         function_exists($languages[$squirrelmail_language]['XTRA_CODE'])) {
 573         return  $languages[$squirrelmail_language]['XTRA_CODE']('encodeheader', $string);
 574     }
 575
 576     // Encode only if the string contains 8-bit characters or =?
 577     $j = strlen( $string  );
 578     $l = strstr($string, '=?');         // Must be encoded ?
 579     $ret = '';
 580     for( $i=0; $i < $j; ++$i) {
 581         switch( $string{$i} ) {
 582            case '=':
 583           $ret .= '=3D';
 584           break;
 585         case '?':
 586           $ret .= '=3F';
 587           break;
 588         case '_':
 589           $ret .= '=5F';
 590           break;
 591         case ' ':
 592           $ret .= '_';
 593           break;
 594         default:
 595           $k = ord( $string{$i} );
 596           if ( $k > 126 ) {
 597              $ret .= sprintf("=%02X", $k);
 598              $l = TRUE;
 599           } else
 600              $ret .= $string{$i};
 601         }
 602     }
 603
 604     if ( $l ) {
 605         $string = "=?$default_charset?Q?$ret?=";
 606     }
 607
 608     return( $string );
 609 }
 610
 611 /* This function trys to locate the entity_id of a specific mime element */
 612
 613 function find_ent_id( $id, $message ) {
 614     $ret = '';
 615     for ($i=0; $ret == '' && $i < count($message->entities); $i++) {
 616         if ( $message->entities[$i]->header->type0 == 'multipart')  {
 617             $ret = find_ent_id( $id, $message->entities[$i] );
 618         } else {
 619             if ( strcasecmp( $message->entities[$i]->header->id, $id ) == 0 )
 620                 $ret = $message->entities[$i]->entity_id;
 621         }
 622     }
 623     return( $ret );
 624 }
 625
 626 /**
 627  ** HTMLFILTER ROUTINES
 628  */
 629
 630 /**
 631  * This function returns the final tag out of the tag name, an array
 632  * of attributes, and the type of the tag. This function is called by
 633  * sq_sanitize internally.
 634  *
 635  * @param  $tagname  the name of the tag.
 636  * @param  $attary   the array of attributes and their values
 637  * @param  $tagtype  The type of the tag (see in comments).
 638  * @return           a string with the final tag representation.
 639  */
 640 function sq_tagprint($tagname, $attary, $tagtype){
 641     $me = "sq_tagprint";
 642     if ($tagtype == 2){
 643         $fulltag = '</' . $tagname . '>';
 644     } else {
 645         $fulltag = '<' . $tagname;
 646         if (is_array($attary) && sizeof($attary)){
 647             $atts = Array();
 648             while (list($attname, $attvalue) = each($attary)){
 649                 array_push($atts, "$attname=$attvalue");
 650             }
 651             $fulltag .= ' ' . join(" ", $atts);
 652         }
 653         if ($tagtype == 3){
 654             $fulltag .= " /";
 655         }
 656         $fulltag .= ">";
 657     }
 658     return $fulltag;
 659 }
 660
 661 /**
 662  * A small helper function to use with array_walk. Modifies a by-ref
 663  * value and makes it lowercase.
 664  *
 665  * @param  $val a value passed by-ref.
 666  * @return      void since it modifies a by-ref value.
 667  */
 668 function sq_casenormalize(&$val){
 669     $val = strtolower($val);
 670 }
 671
 672 /**
 673  * This function skips any whitespace from the current position within
 674  * a string and to the next non-whitespace value.
 675  *
 676  * @param  $body   the string
 677  * @param  $offset the offset within the string where we should start
 678  *                 looking for the next non-whitespace character.
 679  * @return         the location within the $body where the next
 680  *                 non-whitespace char is located.
 681  */
 682 function sq_skipspace($body, $offset){
 683     $me = "sq_skipspace";
 684     preg_match("/^(\s*)/s", substr($body, $offset), $matches);
 685     if (sizeof($matches{1})){
 686         $count = strlen($matches{1});
 687         $offset += $count;
 688     }
 689     return $offset;
 690 }
 691
 692 /**
 693  * This function looks for the next character within a string.  It's
 694  * really just a glorified "strpos", except it catches if failures
 695  * nicely.
 696  *
 697  * @param  $body   The string to look for needle in.
 698  * @param  $offset Start looking from this position.
 699  * @param  $needle The character/string to look for.
 700  * @return         location of the next occurance of the needle, or
 701  *                 strlen($body) if needle wasn't found.
 702  */
 703 function sq_findnxstr($body, $offset, $needle){
 704     $me = "sq_findnxstr";
 705     $pos = strpos($body, $needle, $offset);
 706     if ($pos === FALSE){
 707         $pos = strlen($body);
 708     }
 709     return $pos;
 710 }
 711
 712 /**
 713  * This function takes a PCRE-style regexp and tries to match it
 714  * within the string.
 715  *
 716  * @param  $body   The string to look for needle in.
 717  * @param  $offset Start looking from here.
 718  * @param  $reg    A PCRE-style regex to match.
 719  * @return         Returns a false if no matches found, or an array
 720  *                 with the following members:
 721  *                 - integer with the location of the match within $body
 722  *                 - string with whatever content between offset and the match
 723  *                 - string with whatever it is we matched
 724  */
 725 function sq_findnxreg($body, $offset, $reg){
 726     $me = "sq_findnxreg";
 727     $matches = Array();
 728     $retarr = Array();
 729     preg_match("%^(.*?)($reg)%s", substr($body, $offset), $matches);
 730     if (!$matches{0}){
 731         $retarr = false;
 732     } else {
 733         $retarr{0} = $offset + strlen($matches{1});
 734         $retarr{1} = $matches{1};
 735         $retarr{2} = $matches{2};
 736     }
 737     return $retarr;
 738 }
 739
 740 /**
 741  * This function looks for the next tag.
 742  *
 743  * @param  $body   String where to look for the next tag.
 744  * @param  $offset Start looking from here.
 745  * @return         false if no more tags exist in the body, or
 746  *                 an array with the following members:
 747  *                 - string with the name of the tag
 748  *                 - array with attributes and their values
 749  *                 - integer with tag type (1, 2, or 3)
 750  *                 - integer where the tag starts (starting "<")
 751  *                 - integer where the tag ends (ending ">")
 752  *                 first three members will be false, if the tag is invalid.
 753  */
 754 function sq_getnxtag($body, $offset){
 755     $me = "sq_getnxtag";
 756     if ($offset > strlen($body)){
 757         return false;
 758     }
 759     $lt = sq_findnxstr($body, $offset, "<");
 760     if ($lt == strlen($body)){
 761         return false;
 762     }
 763     /**
 764      * We are here:
 765      * blah blah <tag attribute="value">
 766      * \---------^
 767      */
 768     $pos = sq_skipspace($body, $lt+1);
 769     if ($pos >= strlen($body)){
 770         return Array(false, false, false, $lt, strlen($body));
 771     }
 772     /**
 773      * There are 3 kinds of tags:
 774      * 1. Opening tag, e.g.:
 775      *    <a href="blah">
 776      * 2. Closing tag, e.g.:
 777      *    </a>
 778      * 3. XHTML-style content-less tag, e.g.:
 779      *    <img src="blah"/>
 780      */
 781     $tagtype = false;
 782     switch (substr($body, $pos, 1)){
 783     case "/":
 784         $tagtype = 2;
 785         $pos++;
 786         break;
 787     case "!":
 788         /**
 789          * A comment or an SGML declaration.
 790          */
 791         if (substr($body, $pos+1, 2) == "--"){
 792             $gt = strpos($body, "-->", $pos);
 793             if ($gt === false){
 794                 $gt = strlen($body);
 795             } else {
 796                 $gt += 2;
 797             }
 798             return Array(false, false, false, $lt, $gt);
 799         } else {
 800             $gt = sq_findnxstr($body, $pos, ">");
 801             return Array(false, false, false, $lt, $gt);
 802         }
 803         break;
 804     default:
 805         /**
 806          * Assume tagtype 1 for now. If it's type 3, we'll switch values
 807          * later.
 808          */
 809         $tagtype = 1;
 810         break;
 811     }
 812
 813     $tag_start = $pos;
 814     $tagname = '';
 815     /**
 816      * Look for next [\W-_], which will indicate the end of the tag name.
 817      */
 818     $regary = sq_findnxreg($body, $pos, "[^\w\-_]");
 819     if ($regary == false){
 820         return Array(false, false, false, $lt, strlen($body));
 821     }
 822     list($pos, $tagname, $match) = $regary;
 823     $tagname = strtolower($tagname);
 824
 825     /**
 826      * $match can be either of these:
 827      * '>'  indicating the end of the tag entirely.
 828      * '\s' indicating the end of the tag name.
 829      * '/'  indicating that this is type-3 xhtml tag.
 830      *
 831      * Whatever else we find there indicates an invalid tag.
 832      */
 833     switch ($match){
 834     case "/":
 835         /**
 836          * This is an xhtml-style tag with a closing / at the
 837          * end, like so: <img src="blah"/>. Check if it's followed
 838          * by the closing bracket. If not, then this tag is invalid
 839          */
 840         if (substr($body, $pos, 2) == "/>"){
 841             $pos++;
 842             $tagtype = 3;
 843         } else {
 844             $gt = sq_findnxstr($body, $pos, ">");
 845             $retary = Array(false, false, false, $lt, $gt);
 846             return $retary;
 847         }
 848     case ">":
 849         return Array($tagname, false, $tagtype, $lt, $pos);
 850         break;
 851     default:
 852         /**
 853          * Check if it's whitespace
 854          */
 855         if (preg_match("/\s/", $match)){
 856         } else {
 857             /**
 858              * This is an invalid tag! Look for the next closing ">".
 859              */
 860             $gt = sq_findnxstr($body, $offset, ">");
 861             return Array(false, false, false, $lt, $gt);
 862         }
 863     }
 864
 865     /**
 866      * At this point we're here:
 867      * <tagname  attribute='blah'>
 868      * \-------^
 869      *
 870      * At this point we loop in order to find all attributes.
 871      */
 872     $attname = '';
 873     $atttype = false;
 874     $attary = Array();
 875
 876     while ($pos <= strlen($body)){
 877         $pos = sq_skipspace($body, $pos);
 878         if ($pos == strlen($body)){
 879             /**
 880              * Non-closed tag.
 881              */
 882             return Array(false, false, false, $lt, $pos);
 883         }
 884         /**
 885          * See if we arrived at a ">" or "/>", which means that we reached
 886          * the end of the tag.
 887          */
 888         $matches = Array();
 889         if (preg_match("%^(\s*)(>|/>)%s", substr($body, $pos), $matches)) {
 890             /**
 891              * Yep. So we did.
 892              */
 893             $pos += strlen($matches{1});
 894             if ($matches{2} == "/>"){
 895                 $tagtype = 3;
 896                 $pos++;
 897             }
 898             return Array($tagname, $attary, $tagtype, $lt, $pos);
 899         }
 900
 901         /**
 902          * There are several types of attributes, with optional
 903          * [:space:] between members.
 904          * Type 1:
 905          *   attrname[:space:]=[:space:]'CDATA'
 906          * Type 2:
 907          *   attrname[:space:]=[:space:]"CDATA"
 908          * Type 3:
 909          *   attr[:space:]=[:space:]CDATA
 910          * Type 4:
 911          *   attrname
 912          *
 913          * We leave types 1 and 2 the same, type 3 we check for
 914          * '"' and convert to "&quot" if needed, then wrap in
 915          * double quotes. Type 4 we convert into:
 916          * attrname="yes".
 917          */
 918         $regary = sq_findnxreg($body, $pos, "[^\w\-_]");
 919         if ($regary == false){
 920             /**
 921              * Looks like body ended before the end of tag.
 922              */
 923             return Array(false, false, false, $lt, strlen($body));
 924         }
 925         list($pos, $attname, $match) = $regary;
 926         $attname = strtolower($attname);
 927         /**
 928          * We arrived at the end of attribute name. Several things possible
 929          * here:
 930          * '>'  means the end of the tag and this is attribute type 4
 931          * '/'  if followed by '>' means the same thing as above
 932          * '\s' means a lot of things -- look what it's followed by.
 933          *      anything else means the attribute is invalid.
 934          */
 935         switch($match){
 936         case "/":
 937             /**
 938              * This is an xhtml-style tag with a closing / at the
 939              * end, like so: <img src="blah"/>. Check if it's followed
 940              * by the closing bracket. If not, then this tag is invalid
 941              */
 942             if (substr($body, $pos, 2) == "/>"){
 943                 $pos++;
 944                 $tagtype = 3;
 945             } else {
 946                 $gt = sq_findnxstr($body, $pos, ">");
 947                 $retary = Array(false, false, false, $lt, $gt);
 948                 return $retary;
 949             }
 950         case ">":
 951             $attary{$attname} = '"yes"';
 952             return Array($tagname, $attary, $tagtype, $lt, $pos);
 953             break;
 954         default:
 955             /**
 956              * Skip whitespace and see what we arrive at.
 957              */
 958             $pos = sq_skipspace($body, $pos);
 959             $char = substr($body, $pos, 1);
 960             /**
 961              * Two things are valid here:
 962              * '=' means this is attribute type 1 2 or 3.
 963              * \w means this was attribute type 4.
 964              * anything else we ignore and re-loop. End of tag and
 965              * invalid stuff will be caught by our checks at the beginning
 966              * of the loop.
 967              */
 968             if ($char == "="){
 969                 $pos++;
 970                 $pos = sq_skipspace($body, $pos);
 971                 /**
 972                  * Here are 3 possibilities:
 973                  * "'"  attribute type 1
 974                  * '"'  attribute type 2
 975                  * everything else is the content of tag type 3
 976                  */
 977                 $quot = substr($body, $pos, 1);
 978                 if ($quot == "'"){
 979                     $regary = sq_findnxreg($body, $pos+1, "\'");
 980                     if ($regary == false){
 981                         return Array(false, false, false, $lt, strlen($body));
 982                     }
 983                     list($pos, $attval, $match) = $regary;
 984                     $pos++;
 985                     $attary{$attname} = "'" . $attval . "'";
 986                 } else if ($quot == '"'){
 987                     $regary = sq_findnxreg($body, $pos+1, '\"');
 988                     if ($regary == false){
 989                         return Array(false, false, false, $lt, strlen($body));
 990                     }
 991                     list($pos, $attval, $match) = $regary;
 992                     $pos++;
 993                     $attary{$attname} = '"' . $attval . '"';
 994                 } else {
 995                     /**
 996                      * These are hateful. Look for \s, or >.
 997                      */
 998                     $regary = sq_findnxreg($body, $pos, "[\s>]");
 999                     if ($regary == false){
1000                         return Array(false, false, false, $lt, strlen($body));
1001                     }
1002                     list($pos, $attval, $match) = $regary;
1003                     /**
1004                      * If it's ">" it will be caught at the top.
1005                      */
1006                     $attval = preg_replace("/\"/s", "&quot;", $attval);
1007                     $attary{$attname} = '"' . $attval . '"';
1008                 }
1009             } else if (preg_match("|[\w/>]|", $char)) {
1010                 /**
1011                  * That was attribute type 4.
1012                  */
1013                 $attary{$attname} = '"yes"';
1014             } else {
1015                 /**
1016                  * An illegal character. Find next '>' and return.
1017                  */
1018                 $gt = sq_findnxstr($body, $pos, ">");
1019                 return Array(false, false, false, $lt, $gt);
1020             }
1021         }
1022     }
1023     /**
1024      * The fact that we got here indicates that the tag end was never
1025      * found. Return invalid tag indication so it gets stripped.
1026      */
1027     return Array(false, false, false, $lt, strlen($body));
1028 }
1029
1030 /**
1031  * This function checks attribute values for entity-encoded values
1032  * and returns them translated into 8-bit strings so we can run
1033  * checks on them.
1034  *
1035  * @param  $attvalue A string to run entity check against.
1036  * @return           Translated value.
1037  */
1038 function sq_deent($attvalue){
1039     $me="sq_deent";
1040     /**
1041      * See if we have to run the checks first. All entities must start
1042      * with "&".
1043      */
1044     if (strpos($attvalue, "&") === false){
1045         return $attvalue;
1046     }
1047     /**
1048      * Check named entities first.
1049      */
1050     $trans = get_html_translation_table(HTML_ENTITIES);
1051     /**
1052      * Leave &quot; in, as it can mess us up.
1053      */
1054     $trans = array_flip($trans);
1055     unset($trans{"&quot;"});
1056     while (list($ent, $val) = each($trans)){
1057         $attvalue = preg_replace("/$ent*(\W)/si", "$val\\1", $attvalue);
1058     }
1059     /**
1060      * Now translate numbered entities from 1 to 255 if needed.
1061      */
1062     if (strpos($attvalue, "#") !== false){
1063         $omit = Array(34, 39);
1064         for ($asc=1; $asc<256; $asc++){
1065             if (!in_array($asc, $omit)){
1066                 $chr = chr($asc);
1067                 $attvalue = preg_replace("/\&#0*$asc;*(\D)/si", "$chr\\1",
1068                                          $attvalue);
1069                 $attvalue = preg_replace("/\&#x0*".dechex($asc).";*(\W)/si",
1070                                          "$chr\\1", $attvalue);
1071             }
1072         }
1073     }
1074     return $attvalue;
1075 }
1076
1077 /**
1078  * This function runs various checks against the attributes.
1079  *
1080  * @param  $tagname         String with the name of the tag.
1081  * @param  $attary          Array with all tag attributes.
1082  * @param  $rm_attnames     See description for sq_sanitize
1083  * @param  $bad_attvals     See description for sq_sanitize
1084  * @param  $add_attr_to_tag See description for sq_sanitize
1085  * @param  $message         message object
1086  * @param  $id              message id
1087  * @return                  Array with modified attributes.
1088  */
1089 function sq_fixatts($tagname,
1090                     $attary,
1091                     $rm_attnames,
1092                     $bad_attvals,
1093                     $add_attr_to_tag,
1094                     $message,
1095                     $id,
1096                     $mailbox
1097                     ){
1098     $me = "sq_fixatts";
1099     while (list($attname, $attvalue) = each($attary)){
1100         /**
1101          * See if this attribute should be removed.
1102          */
1103         foreach ($rm_attnames as $matchtag=>$matchattrs){
1104             if (preg_match($matchtag, $tagname)){
1105                 foreach ($matchattrs as $matchattr){
1106                     if (preg_match($matchattr, $attname)){
1107                         unset($attary{$attname});
1108                         continue;
1109                     }
1110                 }
1111             }
1112         }
1113         /**
1114          * Remove any entities.
1115          */
1116         $attvalue = sq_deent($attvalue);
1117
1118         /**
1119          * Now let's run checks on the attvalues.
1120          * I don't expect anyone to comprehend this. If you do,
1121          * get in touch with me so I can drive to where you live and
1122          * shake your hand personally. :)
1123          */
1124         foreach ($bad_attvals as $matchtag=>$matchattrs){
1125             if (preg_match($matchtag, $tagname)){
1126                 foreach ($matchattrs as $matchattr=>$valary){
1127                     if (preg_match($matchattr, $attname)){
1128                         /**
1129                          * There are two arrays in valary.
1130                          * First is matches.
1131                          * Second one is replacements
1132                          */
1133                         list($valmatch, $valrepl) = $valary;
1134                         $newvalue =
1135                             preg_replace($valmatch, $valrepl, $attvalue);
1136                         if ($newvalue != $attvalue){
1137                             $attary{$attname} = $newvalue;
1138                         }
1139                     }
1140                 }
1141             }
1142         }
1143         /**
1144          * Turn cid: urls into http-friendly ones.
1145          */
1146         if (preg_match("/^[\'\"]\s*cid:/si", $attvalue)){
1147             $attary{$attname} = sq_cid2http($message, $id, $attvalue, $mailbox);
1148         }
1149     }
1150     /**
1151      * See if we need to append any attributes to this tag.
1152      */
1153     foreach ($add_attr_to_tag as $matchtag=>$addattary){
1154         if (preg_match($matchtag, $tagname)){
1155             $attary = array_merge($attary, $addattary);
1156         }
1157     }
1158     return $attary;
1159 }
1160
1161 /**
1162  * This function edits the style definition to make them friendly and
1163  * usable in squirrelmail.
1164  *
1165  * @param  $message  the message object
1166  * @param  $id       the message id
1167  * @param  $content  a string with whatever is between <style> and </style>
1168  * @return           a string with edited content.
1169  */
1170 function sq_fixstyle($message, $id, $content){
1171     global $view_unsafe_images;
1172     $me = "sq_fixstyle";
1173     /**
1174      * First look for general BODY style declaration, which would be
1175      * like so:
1176      * body {background: blah-blah}
1177      * and change it to .bodyclass so we can just assign it to a <div>
1178      */
1179     $content = preg_replace("|body(\s*\{.*?\})|si", ".bodyclass\\1", $content);
1180     $secremoveimg = "../images/" . _("sec_remove_eng.png");
1181     /**
1182      * Fix url('blah') declarations.
1183      */
1184     $content = preg_replace("|url\(([\'\"])\s*\S+script\s*:.*?([\'\"])\)|si",
1185                             "url(\\1$secremoveimg\\2)", $content);
1186     /**
1187      * Fix url('https*://.*) declarations but only if $view_unsafe_images
1188      * is false.
1189      */
1190     if (!$view_unsafe_images){
1191         $content = preg_replace("|url\(([\'\"])\s*https*:.*?([\'\"])\)|si",
1192                                 "url(\\1$secremoveimg\\2)", $content);
1193     }
1194
1195     /**
1196      * Fix urls that refer to cid:
1197      */
1198     while (preg_match("|url\(([\'\"]\s*cid:.*?[\'\"])\)|si", $content,
1199                       $matches)){
1200         $cidurl = $matches{1};
1201         $httpurl = sq_cid2http($message, $id, $cidurl);
1202         $content = preg_replace("|url\($cidurl\)|si",
1203                                 "url($httpurl)", $content);
1204     }
1205
1206     /**
1207      * Fix stupid css declarations which lead to vulnerabilities
1208      * in IE.
1209      */
1210     $match   = Array('/expression/si',
1211                      '/behaviou*r/si',
1212                      '/binding/si');
1213     $replace = Array('idiocy', 'idiocy', 'idiocy');
1214     $content = preg_replace($match, $replace, $content);
1215     return $content;
1216 }
1217
1218 /**
1219  * This function converts cid: url's into the ones that can be viewed in
1220  * the browser.
1221  *
1222  * @param  $message  the message object
1223  * @param  $id       the message id
1224  * @param  $cidurl   the cid: url.
1225  * @return           a string with a http-friendly url
1226  */
1227 function sq_cid2http($message, $id, $cidurl, $mailbox){
1228     /**
1229      * Get rid of quotes.
1230      */
1231     $quotchar = substr($cidurl, 0, 1);
1232     $cidurl = str_replace($quotchar, "", $cidurl);
1233     $cidurl = substr(trim($cidurl), 4);
1234     $httpurl = $quotchar . "../src/download.php?absolute_dl=true&amp;" .
1235         "passed_id=$id&amp;mailbox=" . urlencode($mailbox) .
1236         "&amp;ent_id=" . find_ent_id($cidurl, $message) . $quotchar;
1237     return $httpurl;
1238 }
1239
1240 /**
1241  * This function changes the <body> tag into a <div> tag since we
1242  * can't really have a body-within-body.
1243  *
1244  * @param  $attary  an array of attributes and values of <body>
1245  * @return          a modified array of attributes to be set for <div>
1246  */
1247 function sq_body2div($attary){
1248     $me = "sq_body2div";
1249     $divattary = Array("class"=>"'bodyclass'");
1250     $bgcolor="#ffffff";
1251     $text="#000000";
1252     $styledef="";
1253     if (is_array($attary) && sizeof($attary) > 0){
1254         foreach ($attary as $attname=>$attvalue){
1255             $quotchar = substr($attvalue, 0, 1);
1256             $attvalue = str_replace($quotchar, "", $attvalue);
1257             switch ($attname){
1258             case "background":
1259                 $styledef .= "background-image: url('$attvalue'); ";
1260                 break;
1261             case "bgcolor":
1262                 $styledef .= "background-color: $attvalue; ";
1263                 break;
1264             case "text":
1265                 $styledef .= "color: $attvalue; ";
1266             }
1267         }
1268         if (strlen($styledef) > 0){
1269             $divattary{"style"} = "\"$styledef\"";
1270         }
1271     }
1272     return $divattary;
1273 }
1274
1275 /**
1276  * This is the main function and the one you should actually be calling.
1277  * There are several variables you should be aware of an which need
1278  * special description.
1279  *
1280  * Since the description is quite lengthy, see it here:
1281  * http://www.mricon.com/html/phpfilter.html
1282  *
1283  * @param $body                 the string with HTML you wish to filter
1284  * @param $tag_list             see description above
1285  * @param $rm_tags_with_content see description above
1286  * @param $self_closing_tags    see description above
1287  * @param $force_tag_closing    see description above
1288  * @param $rm_attnames          see description above
1289  * @param $bad_attvals          see description above
1290  * @param $add_attr_to_tag      see description above
1291  * @param $message              message object
1292  * @param $id                   message id
1293  * @return                      sanitized html safe to show on your pages.
1294  */
1295 function sq_sanitize($body,
1296                      $tag_list,
1297                      $rm_tags_with_content,
1298                      $self_closing_tags,
1299                      $force_tag_closing,
1300                      $rm_attnames,
1301                      $bad_attvals,
1302                      $add_attr_to_tag,
1303                      $message,
1304                      $id,
1305                      $mailbox
1306                      ){
1307     $me = "sq_sanitize";
1308     /**
1309      * Normalize rm_tags and rm_tags_with_content.
1310      */
1311     @array_walk($rm_tags, 'sq_casenormalize');
1312     @array_walk($rm_tags_with_content, 'sq_casenormalize');
1313     @array_walk($self_closing_tags, 'sq_casenormalize');
1314     /**
1315      * See if tag_list is of tags to remove or tags to allow.
1316      * false  means remove these tags
1317      * true   means allow these tags
1318      */
1319     $rm_tags = array_shift($tag_list);
1320     $curpos = 0;
1321     $open_tags = Array();
1322     $trusted = "<!-- begin sanitized html -->\n";
1323     $skip_content = false;
1324     /**
1325      * Take care of netscape's stupid javascript entities like
1326      * &{alert('boo')};
1327      */
1328     $body = preg_replace("/&(\{.*?\};)/si", "&amp;\\1", $body);
1329
1330     while (($curtag=sq_getnxtag($body, $curpos)) != FALSE){
1331         list($tagname, $attary, $tagtype, $lt, $gt) = $curtag;
1332         $free_content = substr($body, $curpos, $lt-$curpos);
1333         /**
1334          * Take care of <style>
1335          */
1336         if ($tagname == "style" && $tagtype == 2){
1337             /**
1338              * This is a closing </style>. Edit the
1339              * content before we apply it.
1340              */
1341             $free_content = sq_fixstyle($message, $id, $free_content);
1342         }
1343         if ($skip_content == false){
1344             $trusted .= $free_content;
1345         } else {
1346         }
1347         if ($tagname != FALSE){
1348             if ($tagtype == 2){
1349                 if ($skip_content == $tagname){
1350                     /**
1351                      * Got to the end of tag we needed to remove.
1352                      */
1353                     $tagname = false;
1354                     $skip_content = false;
1355                 } else {
1356                     if ($skip_content == false){
1357                         if ($tagname == "body"){
1358                             $tagname = "div";
1359                         } else {
1360                             if (isset($open_tags{$tagname}) &&
1361                                 $open_tags{$tagname} > 0){
1362                                 $open_tags{$tagname}--;
1363                             } else {
1364                                 $tagname = false;
1365                             }
1366                         }
1367                     } else {
1368                     }
1369                 }
1370             } else {
1371                 /**
1372                  * $rm_tags_with_content
1373                  */
1374                 if ($skip_content == false){
1375                     /**
1376                      * See if this is a self-closing type and change
1377                      * tagtype appropriately.
1378                      */
1379                     if ($tagtype == 1
1380                         && in_array($tagname, $self_closing_tags)){
1381                         $tagtype=3;
1382                     }
1383                     /**
1384                      * See if we should skip this tag and any content
1385                      * inside it.
1386                      */
1387                     if ($tagtype == 1 &&
1388                         in_array($tagname, $rm_tags_with_content)){
1389                         $skip_content = $tagname;
1390                     } else {
1391                         if (($rm_tags == false
1392                              && in_array($tagname, $tag_list)) ||
1393                             ($rm_tags == true &&
1394                              !in_array($tagname, $tag_list))){
1395                             $tagname = false;
1396                         } else {
1397                             if ($tagtype == 1){
1398                                 if (isset($open_tags{$tagname})){
1399                                     $open_tags{$tagname}++;
1400                                 } else {
1401                                     $open_tags{$tagname}=1;
1402                                 }
1403                             }
1404                             /**
1405                              * This is where we run other checks.
1406                              */
1407                             if (is_array($attary) && sizeof($attary) > 0){
1408                                 $attary = sq_fixatts($tagname,
1409                                                      $attary,
1410                                                      $rm_attnames,
1411                                                      $bad_attvals,
1412                                                      $add_attr_to_tag,
1413                                                      $message,
1414                                                      $id,
1415                                                      $mailbox
1416                                                      );
1417                             }
1418                             /**
1419                              * Convert body into div.
1420                              */
1421                             if ($tagname == "body"){
1422                                 $tagname = "div";
1423                                 $attary = sq_body2div($attary, $message, $id);
1424                             }
1425                         }
1426                     }
1427                 } else {
1428                 }
1429             }
1430             if ($tagname != false && $skip_content == false){
1431                 $trusted .= sq_tagprint($tagname, $attary, $tagtype);
1432             }
1433         } else {
1434         }
1435         $curpos = $gt+1;
1436     }
1437     $trusted .= substr($body, $curpos, strlen($body)-$curpos);
1438     if ($force_tag_closing == true){
1439         foreach ($open_tags as $tagname=>$opentimes){
1440             while ($opentimes > 0){
1441                 $trusted .= '</' . $tagname . '>';
1442                 $opentimes--;
1443             }
1444         }
1445         $trusted .= "\n";
1446     }
1447     $trusted .= "<!-- end sanitized html -->\n";
1448     return $trusted;
1449 }
1450
1451 /**
1452  * This is a wrapper function to call html sanitizing routines.
1453  *
1454  * @param  $body  the body of the message
1455  * @param  $id    the id of the message
1456  * @return        a string with html safe to display in the browser.
1457  */
1458 function magicHTML($body, $id, $message, $mailbox = 'INBOX'){
1459     global $attachment_common_show_images, $view_unsafe_images,
1460         $has_unsafe_images;
1461     /**
1462      * Don't display attached images in HTML mode.
1463      */
1464     $attachment_common_show_images = false;
1465     $tag_list = Array(
1466                       false,
1467                       "object",
1468                       "meta",
1469                       "html",
1470                       "head",
1471                       "base",
1472                       "link"
1473                       );
1474
1475     $rm_tags_with_content = Array(
1476                                   "script",
1477                                   "applet",
1478                                   "embed",
1479                                   "title"
1480                                   );
1481
1482     $self_closing_tags =  Array(
1483                                 "img",
1484                                 "br",
1485                                 "hr",
1486                                 "input"
1487                                 );
1488
1489     $force_tag_closing = false;
1490
1491     $rm_attnames = Array(
1492                          "/.*/" =>
1493                          Array(
1494                                "/target/si",
1495                                "/^on.*/si",
1496                                "/^dynsrc/si",
1497                                "/^data.*/si"
1498                                )
1499                          );
1500
1501     $secremoveimg = "../images/" . _("sec_remove_eng.png");
1502     $bad_attvals = Array(
1503         "/.*/" =>
1504             Array(
1505                 "/^src|background/i" =>
1506                     Array(
1507                           Array(
1508                                 "|^([\'\"])\s*\.\./.*([\'\"])|si",
1509                                 "/^([\'\"])\s*\S+script\s*:.*([\'\"])/si",
1510                                 "/^([\'\"])\s*mocha\s*:*.*([\'\"])/si",
1511                                 "/^([\'\"])\s*about\s*:.*([\'\"])/si"
1512                                 ),
1513                           Array(
1514                                 "\\1$secremoveimg\\2",
1515                                 "\\1$secremoveimg\\2",
1516                                 "\\1$secremoveimg\\2",
1517                                 "\\1$secremoveimg\\2"
1518                                 )
1519                         ),
1520                 "/^href|action/i" =>
1521                     Array(
1522                           Array(
1523                                 "|^([\'\"])\s*\.\./.*([\'\"])|si",
1524                                 "/^([\'\"])\s*\S+script\s*:.*([\'\"])/si",
1525                                 "/^([\'\"])\s*mocha\s*:*.*([\'\"])/si",
1526                                 "/^([\'\"])\s*about\s*:.*([\'\"])/si"
1527                                 ),
1528                           Array(
1529                                 "\\1#\\2",
1530                                 "\\1#\\2",
1531                                 "\\1#\\2",
1532                                 "\\1#\\2"
1533                                 )
1534                         ),
1535                 "/^style/si" =>
1536                     Array(
1537                           Array(
1538                                 "/expression/si",
1539                                 "/binding/si",
1540                                 "/behaviou*r/si",
1541                                 "|url\(([\'\"])\s*\.\./.*([\'\"])\)|si",
1542                                 "/url\(([\'\"])\s*\S+script\s*:.*([\'\"])\)/si",
1543                                 "/url\(([\'\"])\s*mocha\s*:.*([\'\"])\)/si",
1544                                 "/url\(([\'\"])\s*about\s*:.*([\'\"])\)/si"
1545                                ),
1546                           Array(
1547                                 "idiocy",
1548                                 "idiocy",
1549                                 "idiocy",
1550                                 "url(\\1#\\2)",
1551                                 "url(\\1#\\2)",
1552                                 "url(\\1#\\2)",
1553                                 "url(\\1#\\2)"
1554                                )
1555                           )
1556                 )
1557         );
1558     if (!$view_unsafe_images){
1559         /**
1560          * Remove any references to http/https if view_unsafe_images set
1561          * to false.
1562          */
1563          array_push($bad_attvals{'/.*/'}{'/^src|background/i'}[0],
1564                     '/^([\'\"])\s*https*:.*([\'\"])/si');
1565          array_push($bad_attvals{'/.*/'}{'/^src|background/i'}[1],
1566                     "\\1$secremoveimg\\2");
1567          array_push($bad_attvals{'/.*/'}{'/^style/si'}[0],
1568                     '/url\(([\'\"])\s*https*:.*([\'\"])\)/si');
1569          array_push($bad_attvals{'/.*/'}{'/^style/si'}[1],
1570                     "url(\\1$secremoveimg\\2)");
1571     }
1572
1573     $add_attr_to_tag = Array(
1574                              "/^a$/si" => Array('target'=>'"_new"')
1575                              );
1576     $trusted = sq_sanitize($body,
1577                            $tag_list,
1578                            $rm_tags_with_content,
1579                            $self_closing_tags,
1580                            $force_tag_closing,
1581                            $rm_attnames,
1582                            $bad_attvals,
1583                            $add_attr_to_tag,
1584                            $message,
1585                            $id,
1586                            $mailbox
1587                            );
1588     if (preg_match("|$secremoveimg|si", $trusted)){
1589         $has_unsafe_images = true;
1590     }
1591     return $trusted;
1592 }
1593
1594 ?>