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             $body = charset_decode_japanese($body);
 352             if ( $show_html_default <> 1 ) {
 353                 $body = strip_tags( $body );
 354                 translateText($body, $wrap_at,
 355                   $body_message->header->getParameter['charset']);
 356             } else {
 357                 $body = magicHTML( $body, $id, $message, $mailbox );
 358             }
 359         } else {
 360             translateText($body, $wrap_at,
 361                $body_message->header->getParameter('charset'));
 362         }
 363
 364         if ($has_unsafe_images) {
 365             if ($view_unsafe_images) {
 366                 $body .= "<CENTER><SMALL><A HREF=\"read_body.php?passed_id=$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";
 367             } else {
 368                 $body .= "<CENTER><SMALL><A HREF=\"read_body.php?passed_id=$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";
 369             }
 370         }
 371     }
 372     return ($body);
 373 }
 374
 375
 376 function formatAttachments($message, $exclude_id, $mailbox, $id) {
 377     global $where, $what;
 378     global $startMessage, $color;
 379     static $ShownHTML = 0;
 380
 381     $att_ar = $message->getAttachments($exclude_id);
 382
 383     if (!count($att_ar)) return '';
 384
 385     $attachments = '';
 386
 387     $urlMailbox = urlencode($mailbox);
 388
 389     foreach ($att_ar as $att) {
 390         $ent = urldecode($att->entity_id);
 391         $header = $att->header;
 392         $type0 = strtolower($header->type0);
 393         $type1 = strtolower($header->type1);
 394         $name = '';
 395         $Links['download link']['text'] = _("download");
 396         $Links['download link']['href'] =
 397                 "../src/download.php?absolute_dl=true&amp;passed_id=$id&amp;mailbox=$urlMailbox&amp;ent_id=$ent";
 398         $ImageURL = '';
 399         if ($type0 =='message' && $type1 == 'rfc822') {
 400             $default_page = '../src/read_body.php';
 401             $rfc822_header = $att->rfc822_header;
 402             $filename = decodeHeader($rfc822_header->subject);
 403
 404             $from_o = $rfc822_header->from;
 405             if (is_object($from_o)) {
 406                 $from_name = $from_o->getAddress(false);
 407             } else {
 408                 $from_name = _("Unknown sender");
 409             }
 410             $from_name = decodeHeader(htmlspecialchars($from_name));
 411             $description = $from_name;
 412         } else {
 413             $default_page = '../src/download.php';
 414             if (is_object($header->disposition)) {
 415                $filename = decodeHeader($header->disposition->getProperty('filename'));
 416                if (trim($filename) == '') {
 417                   $name = decodeHeader($header->disposition->getProperty('name'));
 418                   if (trim($name) == '') {
 419                      if ( trim( $header->id ) == '' )
 420                         $filename = 'untitled-[' . $ent . ']' ;
 421                      else
 422                         $filename = 'cid: ' . $header->id;
 423                   } else {
 424                      $filename = $name;
 425                   }
 426                }
 427             } else {
 428                if ( trim( $header->id ) == '' )
 429                   $filename = 'untitled-[' . $ent . ']' ;
 430                else
 431                   $filename = 'cid: ' . $header->id;
 432             }
 433
 434             if ($header->description) {
 435                 $description = htmlspecialchars($header->description);
 436             } else {
 437                 $description = '';
 438             }
 439         }
 440
 441         $display_filename = $filename;
 442         if (isset($passed_ent_id)) {
 443            $passed_ent_id_link = '&amp;passed_ent_id='.$passed_ent_id;
 444         } else {
 445            $passed_ent_id_link = '';
 446         }
 447         $DefaultLink = $default_page . "?startMessage=$startMessage"
 448                      . "&amp;passed_id=$id&amp;mailbox=$urlMailbox"
 449                      . '&amp;ent_id='.$ent.$passed_ent_id_link;
 450         if ($where && $what) {
 451            $DefaultLink = '&amp;where='. urlencode($where).'&amp;what='.urlencode($what);
 452         }
 453         /* this executes the attachment hook with a specific MIME-type.
 454             * if that doens't have results, it tries if there's a rule
 455             * for a more generic type. */
 456         $HookResults = do_hook("attachment $type0/$type1", $Links,
 457             $startMessage, $id, $urlMailbox, $ent, $DefaultLink,
 458             $display_filename, $where, $what);
 459         if(count($HookResults[1]) <= 1) {
 460             $HookResults = do_hook("attachment $type0/*", $Links,
 461                     $startMessage, $id, $urlMailbox, $ent, $DefaultLink,
 462                     $display_filename, $where, $what);
 463         }
 464
 465         $Links = $HookResults[1];
 466         $DefaultLink = $HookResults[6];
 467
 468         $attachments .= '<TR><TD>' .
 469                         "<A HREF=\"$DefaultLink\">$display_filename</A>&nbsp;</TD>" .
 470                         '<TD><SMALL><b>' . show_readable_size($header->size) .
 471                         '</b>&nbsp;&nbsp;</small></TD>' .
 472                         "<TD><SMALL>[ $type0/$type1 ]&nbsp;</SMALL></TD>" .
 473                         '<TD><SMALL>';
 474         $attachments .= '<b>' . $description . '</b>';
 475         $attachments .= '</SMALL></TD><TD><SMALL>&nbsp;';
 476
 477         $SkipSpaces = 1;
 478         foreach ($Links as $Val) {
 479             if ($SkipSpaces) {
 480                 $SkipSpaces = 0;
 481             } else {
 482                 $attachments .= '&nbsp;&nbsp;|&nbsp;&nbsp;';
 483             }
 484             $attachments .= '<a href="' . $Val['href'] . '">' .  $Val['text'] . '</a>';
 485         }
 486         unset($Links);
 487         $attachments .= "</TD></TR>\n";
 488     }
 489     return $attachments;
 490 }
 491
 492 /** this function decodes the body depending on the encoding type. **/
 493 function decodeBody($body, $encoding) {
 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   // All other encodings are returned raw.
 512   return $body;
 513 }
 514
 515 /*
 516  * This functions decode strings that is encoded according to
 517  * RFC1522 (MIME Part Two: Message Header Extensions for Non-ASCII Text).
 518  * Patched by Christian Schmidt <christian@ostenfeld.dk>  23/03/2002
 519  */
 520 function decodeHeader ($string, $utfencode=true) {
 521     global $squirrelmail_language;
 522     if (is_array($string)) {
 523         $string = implode("\n", $string);
 524     }
 525
 526     if ($squirrelmail_language == 'ja_JP' && function_exists('mb_convert_encoding')) {
 527         $string = str_replace("\t", "", $string);
 528         if (eregi('=\\?([^?]+)\\?(q|b)\\?([^?]+)\\?=', $string))
 529             $string = mb_decode_mimeheader($string);
 530         return @mb_convert_encoding($string, 'EUC-JP', 'AUTO');
 531     }
 532
 533     $i = 0;
 534     while (preg_match('/^(.{' . $i . '})(.*)=\?([^?]*)\?(Q|B)\?([^?]*)\?=/Ui',
 535                       $string, $res)) {
 536         $prefix = $res[1];
 537         // Ignore white-space between consecutive encoded-words
 538         if (strspn($res[2], " \t") != strlen($res[2])) {
 539             $prefix .= $res[2];
 540         }
 541
 542         if (ucfirst($res[4]) == 'B') {
 543             $replace = base64_decode($res[5]);
 544         } else {
 545             $replace = str_replace('_', ' ', $res[5]);
 546             $replace = preg_replace('/=([0-9a-f]{2})/ie', 'chr(hexdec("\1"))',
 547                                     $replace);
 548             /* Only encode into entities by default. Some places
 549                don't need the encoding, like the compose form. */
 550             if ($utfencode) {
 551                 $replace = charset_decode($res[3], $replace);
 552             }
 553         }
 554         $string = $prefix . $replace . substr($string, strlen($res[0]));
 555         $i = strlen($prefix) + strlen($replace);
 556     }
 557     return( $string );
 558 }
 559
 560 /*
 561  * Encode a string according to RFC 1522 for use in headers if it
 562  * contains 8-bit characters or anything that looks like it should
 563  * be encoded.
 564  */
 565 function encodeHeader ($string) {
 566     global $default_charset, $squirrelmail_language;
 567
 568     if ($squirrelmail_language == 'ja_JP' && function_exists('mb_encode_mimeheader')) {
 569         return mb_encode_mimeheader($string);
 570     }
 571
 572     // Encode only if the string contains 8-bit characters or =?
 573     $j = strlen( $string  );
 574     $l = strstr($string, '=?');         // Must be encoded ?
 575     $ret = '';
 576     for( $i=0; $i < $j; ++$i) {
 577         switch( $string{$i} ) {
 578            case '=':
 579           $ret .= '=3D';
 580           break;
 581         case '?':
 582           $ret .= '=3F';
 583           break;
 584         case '_':
 585           $ret .= '=5F';
 586           break;
 587         case ' ':
 588           $ret .= '_';
 589           break;
 590         default:
 591           $k = ord( $string{$i} );
 592           if ( $k > 126 ) {
 593              $ret .= sprintf("=%02X", $k);
 594              $l = TRUE;
 595           } else
 596              $ret .= $string{$i};
 597         }
 598     }
 599
 600     if ( $l ) {
 601         $string = "=?$default_charset?Q?$ret?=";
 602     }
 603
 604     return( $string );
 605 }
 606
 607 /* This function trys to locate the entity_id of a specific mime element */
 608
 609 function find_ent_id( $id, $message ) {
 610     $ret = '';
 611     for ($i=0; $ret == '' && $i < count($message->entities); $i++) {
 612         if ( $message->entities[$i]->header->type0 == 'multipart')  {
 613             $ret = find_ent_id( $id, $message->entities[$i] );
 614         } else {
 615             if ( strcasecmp( $message->entities[$i]->header->id, $id ) == 0 )
 616                 $ret = $message->entities[$i]->entity_id;
 617         }
 618     }
 619     return( $ret );
 620 }
 621
 622 /**
 623  ** HTMLFILTER ROUTINES
 624  */
 625
 626 /**
 627  * This function returns the final tag out of the tag name, an array
 628  * of attributes, and the type of the tag. This function is called by
 629  * sq_sanitize internally.
 630  *
 631  * @param  $tagname  the name of the tag.
 632  * @param  $attary   the array of attributes and their values
 633  * @param  $tagtype  The type of the tag (see in comments).
 634  * @return           a string with the final tag representation.
 635  */
 636 function sq_tagprint($tagname, $attary, $tagtype){
 637     $me = "sq_tagprint";
 638     if ($tagtype == 2){
 639         $fulltag = '</' . $tagname . '>';
 640     } else {
 641         $fulltag = '<' . $tagname;
 642         if (is_array($attary) && sizeof($attary)){
 643             $atts = Array();
 644             while (list($attname, $attvalue) = each($attary)){
 645                 array_push($atts, "$attname=$attvalue");
 646             }
 647             $fulltag .= ' ' . join(" ", $atts);
 648         }
 649         if ($tagtype == 3){
 650             $fulltag .= " /";
 651         }
 652         $fulltag .= ">";
 653     }
 654     return $fulltag;
 655 }
 656
 657 /**
 658  * A small helper function to use with array_walk. Modifies a by-ref
 659  * value and makes it lowercase.
 660  *
 661  * @param  $val a value passed by-ref.
 662  * @return      void since it modifies a by-ref value.
 663  */
 664 function sq_casenormalize(&$val){
 665     $val = strtolower($val);
 666 }
 667
 668 /**
 669  * This function skips any whitespace from the current position within
 670  * a string and to the next non-whitespace value.
 671  *
 672  * @param  $body   the string
 673  * @param  $offset the offset within the string where we should start
 674  *                 looking for the next non-whitespace character.
 675  * @return         the location within the $body where the next
 676  *                 non-whitespace char is located.
 677  */
 678 function sq_skipspace($body, $offset){
 679     $me = "sq_skipspace";
 680     preg_match("/^(\s*)/s", substr($body, $offset), $matches);
 681     if (sizeof($matches{1})){
 682         $count = strlen($matches{1});
 683         $offset += $count;
 684     }
 685     return $offset;
 686 }
 687
 688 /**
 689  * This function looks for the next character within a string.  It's
 690  * really just a glorified "strpos", except it catches if failures
 691  * nicely.
 692  *
 693  * @param  $body   The string to look for needle in.
 694  * @param  $offset Start looking from this position.
 695  * @param  $needle The character/string to look for.
 696  * @return         location of the next occurance of the needle, or
 697  *                 strlen($body) if needle wasn't found.
 698  */
 699 function sq_findnxstr($body, $offset, $needle){
 700     $me = "sq_findnxstr";
 701     $pos = strpos($body, $needle, $offset);
 702     if ($pos === FALSE){
 703         $pos = strlen($body);
 704     }
 705     return $pos;
 706 }
 707
 708 /**
 709  * This function takes a PCRE-style regexp and tries to match it
 710  * within the string.
 711  *
 712  * @param  $body   The string to look for needle in.
 713  * @param  $offset Start looking from here.
 714  * @param  $reg    A PCRE-style regex to match.
 715  * @return         Returns a false if no matches found, or an array
 716  *                 with the following members:
 717  *                 - integer with the location of the match within $body
 718  *                 - string with whatever content between offset and the match
 719  *                 - string with whatever it is we matched
 720  */
 721 function sq_findnxreg($body, $offset, $reg){
 722     $me = "sq_findnxreg";
 723     $matches = Array();
 724     $retarr = Array();
 725     preg_match("%^(.*?)($reg)%s", substr($body, $offset), $matches);
 726     if (!$matches{0}){
 727         $retarr = false;
 728     } else {
 729         $retarr{0} = $offset + strlen($matches{1});
 730         $retarr{1} = $matches{1};
 731         $retarr{2} = $matches{2};
 732     }
 733     return $retarr;
 734 }
 735
 736 /**
 737  * This function looks for the next tag.
 738  *
 739  * @param  $body   String where to look for the next tag.
 740  * @param  $offset Start looking from here.
 741  * @return         false if no more tags exist in the body, or
 742  *                 an array with the following members:
 743  *                 - string with the name of the tag
 744  *                 - array with attributes and their values
 745  *                 - integer with tag type (1, 2, or 3)
 746  *                 - integer where the tag starts (starting "<")
 747  *                 - integer where the tag ends (ending ">")
 748  *                 first three members will be false, if the tag is invalid.
 749  */
 750 function sq_getnxtag($body, $offset){
 751     $me = "sq_getnxtag";
 752     if ($offset > strlen($body)){
 753         return false;
 754     }
 755     $lt = sq_findnxstr($body, $offset, "<");
 756     if ($lt == strlen($body)){
 757         return false;
 758     }
 759     /**
 760      * We are here:
 761      * blah blah <tag attribute="value">
 762      * \---------^
 763      */
 764     $pos = sq_skipspace($body, $lt+1);
 765     if ($pos >= strlen($body)){
 766         return Array(false, false, false, $lt, strlen($body));
 767     }
 768     /**
 769      * There are 3 kinds of tags:
 770      * 1. Opening tag, e.g.:
 771      *    <a href="blah">
 772      * 2. Closing tag, e.g.:
 773      *    </a>
 774      * 3. XHTML-style content-less tag, e.g.:
 775      *    <img src="blah"/>
 776      */
 777     $tagtype = false;
 778     switch (substr($body, $pos, 1)){
 779     case "/":
 780         $tagtype = 2;
 781         $pos++;
 782         break;
 783     case "!":
 784         /**
 785          * A comment or an SGML declaration.
 786          */
 787         if (substr($body, $pos+1, 2) == "--"){
 788             $gt = strpos($body, "-->", $pos);
 789             if ($gt === false){
 790                 $gt = strlen($body);
 791             } else {
 792                 $gt += 2;
 793             }
 794             return Array(false, false, false, $lt, $gt);
 795         } else {
 796             $gt = sq_findnxstr($body, $pos, ">");
 797             return Array(false, false, false, $lt, $gt);
 798         }
 799         break;
 800     default:
 801         /**
 802          * Assume tagtype 1 for now. If it's type 3, we'll switch values
 803          * later.
 804          */
 805         $tagtype = 1;
 806         break;
 807     }
 808
 809     $tag_start = $pos;
 810     $tagname = '';
 811     /**
 812      * Look for next [\W-_], which will indicate the end of the tag name.
 813      */
 814     $regary = sq_findnxreg($body, $pos, "[^\w\-_]");
 815     if ($regary == false){
 816         return Array(false, false, false, $lt, strlen($body));
 817     }
 818     list($pos, $tagname, $match) = $regary;
 819     $tagname = strtolower($tagname);
 820
 821     /**
 822      * $match can be either of these:
 823      * '>'  indicating the end of the tag entirely.
 824      * '\s' indicating the end of the tag name.
 825      * '/'  indicating that this is type-3 xhtml tag.
 826      *
 827      * Whatever else we find there indicates an invalid tag.
 828      */
 829     switch ($match){
 830     case "/":
 831         /**
 832          * This is an xhtml-style tag with a closing / at the
 833          * end, like so: <img src="blah"/>. Check if it's followed
 834          * by the closing bracket. If not, then this tag is invalid
 835          */
 836         if (substr($body, $pos, 2) == "/>"){
 837             $pos++;
 838             $tagtype = 3;
 839         } else {
 840             $gt = sq_findnxstr($body, $pos, ">");
 841             $retary = Array(false, false, false, $lt, $gt);
 842             return $retary;
 843         }
 844     case ">":
 845         return Array($tagname, false, $tagtype, $lt, $pos);
 846         break;
 847     default:
 848         /**
 849          * Check if it's whitespace
 850          */
 851         if (preg_match("/\s/", $match)){
 852         } else {
 853             /**
 854              * This is an invalid tag! Look for the next closing ">".
 855              */
 856             $gt = sq_findnxstr($body, $offset, ">");
 857             return Array(false, false, false, $lt, $gt);
 858         }
 859     }
 860
 861     /**
 862      * At this point we're here:
 863      * <tagname  attribute='blah'>
 864      * \-------^
 865      *
 866      * At this point we loop in order to find all attributes.
 867      */
 868     $attname = '';
 869     $atttype = false;
 870     $attary = Array();
 871
 872     while ($pos <= strlen($body)){
 873         $pos = sq_skipspace($body, $pos);
 874         if ($pos == strlen($body)){
 875             /**
 876              * Non-closed tag.
 877              */
 878             return Array(false, false, false, $lt, $pos);
 879         }
 880         /**
 881          * See if we arrived at a ">" or "/>", which means that we reached
 882          * the end of the tag.
 883          */
 884         $matches = Array();
 885         if (preg_match("%^(\s*)(>|/>)%s", substr($body, $pos), $matches)) {
 886             /**
 887              * Yep. So we did.
 888              */
 889             $pos += strlen($matches{1});
 890             if ($matches{2} == "/>"){
 891                 $tagtype = 3;
 892                 $pos++;
 893             }
 894             return Array($tagname, $attary, $tagtype, $lt, $pos);
 895         }
 896
 897         /**
 898          * There are several types of attributes, with optional
 899          * [:space:] between members.
 900          * Type 1:
 901          *   attrname[:space:]=[:space:]'CDATA'
 902          * Type 2:
 903          *   attrname[:space:]=[:space:]"CDATA"
 904          * Type 3:
 905          *   attr[:space:]=[:space:]CDATA
 906          * Type 4:
 907          *   attrname
 908          *
 909          * We leave types 1 and 2 the same, type 3 we check for
 910          * '"' and convert to "&quot" if needed, then wrap in
 911          * double quotes. Type 4 we convert into:
 912          * attrname="yes".
 913          */
 914         $regary = sq_findnxreg($body, $pos, "[^\w\-_]");
 915         if ($regary == false){
 916             /**
 917              * Looks like body ended before the end of tag.
 918              */
 919             return Array(false, false, false, $lt, strlen($body));
 920         }
 921         list($pos, $attname, $match) = $regary;
 922         $attname = strtolower($attname);
 923         /**
 924          * We arrived at the end of attribute name. Several things possible
 925          * here:
 926          * '>'  means the end of the tag and this is attribute type 4
 927          * '/'  if followed by '>' means the same thing as above
 928          * '\s' means a lot of things -- look what it's followed by.
 929          *      anything else means the attribute is invalid.
 930          */
 931         switch($match){
 932         case "/":
 933             /**
 934              * This is an xhtml-style tag with a closing / at the
 935              * end, like so: <img src="blah"/>. Check if it's followed
 936              * by the closing bracket. If not, then this tag is invalid
 937              */
 938             if (substr($body, $pos, 2) == "/>"){
 939                 $pos++;
 940                 $tagtype = 3;
 941             } else {
 942                 $gt = sq_findnxstr($body, $pos, ">");
 943                 $retary = Array(false, false, false, $lt, $gt);
 944                 return $retary;
 945             }
 946         case ">":
 947             $attary{$attname} = '"yes"';
 948             return Array($tagname, $attary, $tagtype, $lt, $pos);
 949             break;
 950         default:
 951             /**
 952              * Skip whitespace and see what we arrive at.
 953              */
 954             $pos = sq_skipspace($body, $pos);
 955             $char = substr($body, $pos, 1);
 956             /**
 957              * Two things are valid here:
 958              * '=' means this is attribute type 1 2 or 3.
 959              * \w means this was attribute type 4.
 960              * anything else we ignore and re-loop. End of tag and
 961              * invalid stuff will be caught by our checks at the beginning
 962              * of the loop.
 963              */
 964             if ($char == "="){
 965                 $pos++;
 966                 $pos = sq_skipspace($body, $pos);
 967                 /**
 968                  * Here are 3 possibilities:
 969                  * "'"  attribute type 1
 970                  * '"'  attribute type 2
 971                  * everything else is the content of tag type 3
 972                  */
 973                 $quot = substr($body, $pos, 1);
 974                 if ($quot == "'"){
 975                     $regary = sq_findnxreg($body, $pos+1, "\'");
 976                     if ($regary == false){
 977                         return Array(false, false, false, $lt, strlen($body));
 978                     }
 979                     list($pos, $attval, $match) = $regary;
 980                     $pos++;
 981                     $attary{$attname} = "'" . $attval . "'";
 982                 } else if ($quot == '"'){
 983                     $regary = sq_findnxreg($body, $pos+1, '\"');
 984                     if ($regary == false){
 985                         return Array(false, false, false, $lt, strlen($body));
 986                     }
 987                     list($pos, $attval, $match) = $regary;
 988                     $pos++;
 989                     $attary{$attname} = '"' . $attval . '"';
 990                 } else {
 991                     /**
 992                      * These are hateful. Look for \s, or >.
 993                      */
 994                     $regary = sq_findnxreg($body, $pos, "[\s>]");
 995                     if ($regary == false){
 996                         return Array(false, false, false, $lt, strlen($body));
 997                     }
 998                     list($pos, $attval, $match) = $regary;
 999                     /**
1000                      * If it's ">" it will be caught at the top.
1001                      */
1002                     $attval = preg_replace("/\"/s", "&quot;", $attval);
1003                     $attary{$attname} = '"' . $attval . '"';
1004                 }
1005             } else if (preg_match("|[\w/>]|", $char)) {
1006                 /**
1007                  * That was attribute type 4.
1008                  */
1009                 $attary{$attname} = '"yes"';
1010             } else {
1011                 /**
1012                  * An illegal character. Find next '>' and return.
1013                  */
1014                 $gt = sq_findnxstr($body, $pos, ">");
1015                 return Array(false, false, false, $lt, $gt);
1016             }
1017         }
1018     }
1019     /**
1020      * The fact that we got here indicates that the tag end was never
1021      * found. Return invalid tag indication so it gets stripped.
1022      */
1023     return Array(false, false, false, $lt, strlen($body));
1024 }
1025
1026 /**
1027  * This function checks attribute values for entity-encoded values
1028  * and returns them translated into 8-bit strings so we can run
1029  * checks on them.
1030  *
1031  * @param  $attvalue A string to run entity check against.
1032  * @return           Translated value.
1033  */
1034 function sq_deent($attvalue){
1035     $me="sq_deent";
1036     /**
1037      * See if we have to run the checks first. All entities must start
1038      * with "&".
1039      */
1040     if (strpos($attvalue, "&") === false){
1041         return $attvalue;
1042     }
1043     /**
1044      * Check named entities first.
1045      */
1046     $trans = get_html_translation_table(HTML_ENTITIES);
1047     /**
1048      * Leave &quot; in, as it can mess us up.
1049      */
1050     $trans = array_flip($trans);
1051     unset($trans{"&quot;"});
1052     while (list($ent, $val) = each($trans)){
1053         $attvalue = preg_replace("/$ent*(\W)/si", "$val\\1", $attvalue);
1054     }
1055     /**
1056      * Now translate numbered entities from 1 to 255 if needed.
1057      */
1058     if (strpos($attvalue, "#") !== false){
1059         $omit = Array(34, 39);
1060         for ($asc=1; $asc<256; $asc++){
1061             if (!in_array($asc, $omit)){
1062                 $chr = chr($asc);
1063                 $attvalue = preg_replace("/\&#0*$asc;*(\D)/si", "$chr\\1",
1064                                          $attvalue);
1065                 $attvalue = preg_replace("/\&#x0*".dechex($asc).";*(\W)/si",
1066                                          "$chr\\1", $attvalue);
1067             }
1068         }
1069     }
1070     return $attvalue;
1071 }
1072
1073 /**
1074  * This function runs various checks against the attributes.
1075  *
1076  * @param  $tagname         String with the name of the tag.
1077  * @param  $attary          Array with all tag attributes.
1078  * @param  $rm_attnames     See description for sq_sanitize
1079  * @param  $bad_attvals     See description for sq_sanitize
1080  * @param  $add_attr_to_tag See description for sq_sanitize
1081  * @param  $message         message object
1082  * @param  $id              message id
1083  * @return                  Array with modified attributes.
1084  */
1085 function sq_fixatts($tagname,
1086                     $attary,
1087                     $rm_attnames,
1088                     $bad_attvals,
1089                     $add_attr_to_tag,
1090                     $message,
1091                     $id,
1092                     $mailbox
1093                     ){
1094     $me = "sq_fixatts";
1095     while (list($attname, $attvalue) = each($attary)){
1096         /**
1097          * See if this attribute should be removed.
1098          */
1099         foreach ($rm_attnames as $matchtag=>$matchattrs){
1100             if (preg_match($matchtag, $tagname)){
1101                 foreach ($matchattrs as $matchattr){
1102                     if (preg_match($matchattr, $attname)){
1103                         unset($attary{$attname});
1104                         continue;
1105                     }
1106                 }
1107             }
1108         }
1109         /**
1110          * Remove any entities.
1111          */
1112         $attvalue = sq_deent($attvalue);
1113
1114         /**
1115          * Now let's run checks on the attvalues.
1116          * I don't expect anyone to comprehend this. If you do,
1117          * get in touch with me so I can drive to where you live and
1118          * shake your hand personally. :)
1119          */
1120         foreach ($bad_attvals as $matchtag=>$matchattrs){
1121             if (preg_match($matchtag, $tagname)){
1122                 foreach ($matchattrs as $matchattr=>$valary){
1123                     if (preg_match($matchattr, $attname)){
1124                         /**
1125                          * There are two arrays in valary.
1126                          * First is matches.
1127                          * Second one is replacements
1128                          */
1129                         list($valmatch, $valrepl) = $valary;
1130                         $newvalue =
1131                             preg_replace($valmatch, $valrepl, $attvalue);
1132                         if ($newvalue != $attvalue){
1133                             $attary{$attname} = $newvalue;
1134                         }
1135                     }
1136                 }
1137             }
1138         }
1139         /**
1140          * Turn cid: urls into http-friendly ones.
1141          */
1142         if (preg_match("/^[\'\"]\s*cid:/si", $attvalue)){
1143             $attary{$attname} = sq_cid2http($message, $id, $attvalue, $mailbox);
1144         }
1145     }
1146     /**
1147      * See if we need to append any attributes to this tag.
1148      */
1149     foreach ($add_attr_to_tag as $matchtag=>$addattary){
1150         if (preg_match($matchtag, $tagname)){
1151             $attary = array_merge($attary, $addattary);
1152         }
1153     }
1154     return $attary;
1155 }
1156
1157 /**
1158  * This function edits the style definition to make them friendly and
1159  * usable in squirrelmail.
1160  *
1161  * @param  $message  the message object
1162  * @param  $id       the message id
1163  * @param  $content  a string with whatever is between <style> and </style>
1164  * @return           a string with edited content.
1165  */
1166 function sq_fixstyle($message, $id, $content){
1167     global $view_unsafe_images;
1168     $me = "sq_fixstyle";
1169     /**
1170      * First look for general BODY style declaration, which would be
1171      * like so:
1172      * body {background: blah-blah}
1173      * and change it to .bodyclass so we can just assign it to a <div>
1174      */
1175     $content = preg_replace("|body(\s*\{.*?\})|si", ".bodyclass\\1", $content);
1176     $secremoveimg = "../images/" . _("sec_remove_eng.png");
1177     /**
1178      * Fix url('blah') declarations.
1179      */
1180     $content = preg_replace("|url\(([\'\"])\s*\S+script\s*:.*?([\'\"])\)|si",
1181                             "url(\\1$secremoveimg\\2)", $content);
1182     /**
1183      * Fix url('https*://.*) declarations but only if $view_unsafe_images
1184      * is false.
1185      */
1186     if (!$view_unsafe_images){
1187         $content = preg_replace("|url\(([\'\"])\s*https*:.*?([\'\"])\)|si",
1188                                 "url(\\1$secremoveimg\\2)", $content);
1189     }
1190
1191     /**
1192      * Fix urls that refer to cid:
1193      */
1194     while (preg_match("|url\(([\'\"]\s*cid:.*?[\'\"])\)|si", $content,
1195                       $matches)){
1196         $cidurl = $matches{1};
1197         $httpurl = sq_cid2http($message, $id, $cidurl);
1198         $content = preg_replace("|url\($cidurl\)|si",
1199                                 "url($httpurl)", $content);
1200     }
1201
1202     /**
1203      * Fix stupid css declarations which lead to vulnerabilities
1204      * in IE.
1205      */
1206     $match   = Array('/expression/si',
1207                      '/behaviou*r/si',
1208                      '/binding/si');
1209     $replace = Array('idiocy', 'idiocy', 'idiocy');
1210     $content = preg_replace($match, $replace, $content);
1211     return $content;
1212 }
1213
1214 /**
1215  * This function converts cid: url's into the ones that can be viewed in
1216  * the browser.
1217  *
1218  * @param  $message  the message object
1219  * @param  $id       the message id
1220  * @param  $cidurl   the cid: url.
1221  * @return           a string with a http-friendly url
1222  */
1223 function sq_cid2http($message, $id, $cidurl, $mailbox){
1224     /**
1225      * Get rid of quotes.
1226      */
1227     $quotchar = substr($cidurl, 0, 1);
1228     $cidurl = str_replace($quotchar, "", $cidurl);
1229     $cidurl = substr(trim($cidurl), 4);
1230     $httpurl = $quotchar . "../src/download.php?absolute_dl=true&amp;" .
1231         "passed_id=$id&amp;mailbox=" . urlencode($mailbox) .
1232         "&amp;ent_id=" . find_ent_id($cidurl, $message) . $quotchar;
1233     return $httpurl;
1234 }
1235
1236 /**
1237  * This function changes the <body> tag into a <div> tag since we
1238  * can't really have a body-within-body.
1239  *
1240  * @param  $attary  an array of attributes and values of <body>
1241  * @return          a modified array of attributes to be set for <div>
1242  */
1243 function sq_body2div($attary){
1244     $me = "sq_body2div";
1245     $divattary = Array("class"=>"'bodyclass'");
1246     $bgcolor="#ffffff";
1247     $text="#000000";
1248     $styledef="";
1249     if (is_array($attary) && sizeof($attary) > 0){
1250         foreach ($attary as $attname=>$attvalue){
1251             $quotchar = substr($attvalue, 0, 1);
1252             $attvalue = str_replace($quotchar, "", $attvalue);
1253             switch ($attname){
1254             case "background":
1255                 $styledef .= "background-image: url('$attvalue'); ";
1256                 break;
1257             case "bgcolor":
1258                 $styledef .= "background-color: $attvalue; ";
1259                 break;
1260             case "text":
1261                 $styledef .= "color: $attvalue; ";
1262             }
1263         }
1264         if (strlen($styledef) > 0){
1265             $divattary{"style"} = "\"$styledef\"";
1266         }
1267     }
1268     return $divattary;
1269 }
1270
1271 /**
1272  * This is the main function and the one you should actually be calling.
1273  * There are several variables you should be aware of an which need
1274  * special description.
1275  *
1276  * Since the description is quite lengthy, see it here:
1277  * http://www.mricon.com/html/phpfilter.html
1278  *
1279  * @param $body                 the string with HTML you wish to filter
1280  * @param $tag_list             see description above
1281  * @param $rm_tags_with_content see description above
1282  * @param $self_closing_tags    see description above
1283  * @param $force_tag_closing    see description above
1284  * @param $rm_attnames          see description above
1285  * @param $bad_attvals          see description above
1286  * @param $add_attr_to_tag      see description above
1287  * @param $message              message object
1288  * @param $id                   message id
1289  * @return                      sanitized html safe to show on your pages.
1290  */
1291 function sq_sanitize($body,
1292                      $tag_list,
1293                      $rm_tags_with_content,
1294                      $self_closing_tags,
1295                      $force_tag_closing,
1296                      $rm_attnames,
1297                      $bad_attvals,
1298                      $add_attr_to_tag,
1299                      $message,
1300                      $id,
1301                      $mailbox
1302                      ){
1303     $me = "sq_sanitize";
1304     /**
1305      * Normalize rm_tags and rm_tags_with_content.
1306      */
1307     @array_walk($rm_tags, 'sq_casenormalize');
1308     @array_walk($rm_tags_with_content, 'sq_casenormalize');
1309     @array_walk($self_closing_tags, 'sq_casenormalize');
1310     /**
1311      * See if tag_list is of tags to remove or tags to allow.
1312      * false  means remove these tags
1313      * true   means allow these tags
1314      */
1315     $rm_tags = array_shift($tag_list);
1316     $curpos = 0;
1317     $open_tags = Array();
1318     $trusted = "<!-- begin sanitized html -->\n";
1319     $skip_content = false;
1320     /**
1321      * Take care of netscape's stupid javascript entities like
1322      * &{alert('boo')};
1323      */
1324     $body = preg_replace("/&(\{.*?\};)/si", "&amp;\\1", $body);
1325
1326     while (($curtag=sq_getnxtag($body, $curpos)) != FALSE){
1327         list($tagname, $attary, $tagtype, $lt, $gt) = $curtag;
1328         $free_content = substr($body, $curpos, $lt-$curpos);
1329         /**
1330          * Take care of <style>
1331          */
1332         if ($tagname == "style" && $tagtype == 2){
1333             /**
1334              * This is a closing </style>. Edit the
1335              * content before we apply it.
1336              */
1337             $free_content = sq_fixstyle($message, $id, $free_content);
1338         }
1339         if ($skip_content == false){
1340             $trusted .= $free_content;
1341         } else {
1342         }
1343         if ($tagname != FALSE){
1344             if ($tagtype == 2){
1345                 if ($skip_content == $tagname){
1346                     /**
1347                      * Got to the end of tag we needed to remove.
1348                      */
1349                     $tagname = false;
1350                     $skip_content = false;
1351                 } else {
1352                     if ($skip_content == false){
1353                         if ($tagname == "body"){
1354                             $tagname = "div";
1355                         } else {
1356                             if (isset($open_tags{$tagname}) &&
1357                                 $open_tags{$tagname} > 0){
1358                                 $open_tags{$tagname}--;
1359                             } else {
1360                                 $tagname = false;
1361                             }
1362                         }
1363                     } else {
1364                     }
1365                 }
1366             } else {
1367                 /**
1368                  * $rm_tags_with_content
1369                  */
1370                 if ($skip_content == false){
1371                     /**
1372                      * See if this is a self-closing type and change
1373                      * tagtype appropriately.
1374                      */
1375                     if ($tagtype == 1
1376                         && in_array($tagname, $self_closing_tags)){
1377                         $tagtype=3;
1378                     }
1379                     /**
1380                      * See if we should skip this tag and any content
1381                      * inside it.
1382                      */
1383                     if ($tagtype == 1 &&
1384                         in_array($tagname, $rm_tags_with_content)){
1385                         $skip_content = $tagname;
1386                     } else {
1387                         if (($rm_tags == false
1388                              && in_array($tagname, $tag_list)) ||
1389                             ($rm_tags == true &&
1390                              !in_array($tagname, $tag_list))){
1391                             $tagname = false;
1392                         } else {
1393                             if ($tagtype == 1){
1394                                 if (isset($open_tags{$tagname})){
1395                                     $open_tags{$tagname}++;
1396                                 } else {
1397                                     $open_tags{$tagname}=1;
1398                                 }
1399                             }
1400                             /**
1401                              * This is where we run other checks.
1402                              */
1403                             if (is_array($attary) && sizeof($attary) > 0){
1404                                 $attary = sq_fixatts($tagname,
1405                                                      $attary,
1406                                                      $rm_attnames,
1407                                                      $bad_attvals,
1408                                                      $add_attr_to_tag,
1409                                                      $message,
1410                                                      $id,
1411                                                      $mailbox
1412                                                      );
1413                             }
1414                             /**
1415                              * Convert body into div.
1416                              */
1417                             if ($tagname == "body"){
1418                                 $tagname = "div";
1419                                 $attary = sq_body2div($attary, $message, $id);
1420                             }
1421                         }
1422                     }
1423                 } else {
1424                 }
1425             }
1426             if ($tagname != false && $skip_content == false){
1427                 $trusted .= sq_tagprint($tagname, $attary, $tagtype);
1428             }
1429         } else {
1430         }
1431         $curpos = $gt+1;
1432     }
1433     $trusted .= substr($body, $curpos, strlen($body)-$curpos);
1434     if ($force_tag_closing == true){
1435         foreach ($open_tags as $tagname=>$opentimes){
1436             while ($opentimes > 0){
1437                 $trusted .= '</' . $tagname . '>';
1438                 $opentimes--;
1439             }
1440         }
1441         $trusted .= "\n";
1442     }
1443     $trusted .= "<!-- end sanitized html -->\n";
1444     return $trusted;
1445 }
1446
1447 /**
1448  * This is a wrapper function to call html sanitizing routines.
1449  *
1450  * @param  $body  the body of the message
1451  * @param  $id    the id of the message
1452  * @return        a string with html safe to display in the browser.
1453  */
1454 function magicHTML($body, $id, $message, $mailbox = 'INBOX'){
1455     global $attachment_common_show_images, $view_unsafe_images,
1456         $has_unsafe_images;
1457     /**
1458      * Don't display attached images in HTML mode.
1459      */
1460     $attachment_common_show_images = false;
1461     $tag_list = Array(
1462                       false,
1463                       "object",
1464                       "meta",
1465                       "html",
1466                       "head",
1467                       "base",
1468                       "link"
1469                       );
1470
1471     $rm_tags_with_content = Array(
1472                                   "script",
1473                                   "applet",
1474                                   "embed",
1475                                   "title"
1476                                   );
1477
1478     $self_closing_tags =  Array(
1479                                 "img",
1480                                 "br",
1481                                 "hr",
1482                                 "input"
1483                                 );
1484
1485     $force_tag_closing = false;
1486
1487     $rm_attnames = Array(
1488                          "/.*/" =>
1489                          Array(
1490                                "/target/si",
1491                                "/^on.*/si",
1492                                "/^dynsrc/si",
1493                                "/^data.*/si"
1494                                )
1495                          );
1496
1497     $secremoveimg = "../images/" . _("sec_remove_eng.png");
1498     $bad_attvals = Array(
1499         "/.*/" =>
1500             Array(
1501                 "/^src|background/i" =>
1502                     Array(
1503                           Array(
1504                                 "|^([\'\"])\s*\.\./.*([\'\"])|si",
1505                                 "/^([\'\"])\s*\S+script\s*:.*([\'\"])/si",
1506                                 "/^([\'\"])\s*mocha\s*:*.*([\'\"])/si",
1507                                 "/^([\'\"])\s*about\s*:.*([\'\"])/si"
1508                                 ),
1509                           Array(
1510                                 "\\1$secremoveimg\\2",
1511                                 "\\1$secremoveimg\\2",
1512                                 "\\1$secremoveimg\\2",
1513                                 "\\1$secremoveimg\\2"
1514                                 )
1515                         ),
1516                 "/^href|action/i" =>
1517                     Array(
1518                           Array(
1519                                 "|^([\'\"])\s*\.\./.*([\'\"])|si",
1520                                 "/^([\'\"])\s*\S+script\s*:.*([\'\"])/si",
1521                                 "/^([\'\"])\s*mocha\s*:*.*([\'\"])/si",
1522                                 "/^([\'\"])\s*about\s*:.*([\'\"])/si"
1523                                 ),
1524                           Array(
1525                                 "\\1#\\2",
1526                                 "\\1#\\2",
1527                                 "\\1#\\2",
1528                                 "\\1#\\2"
1529                                 )
1530                         ),
1531                 "/^style/si" =>
1532                     Array(
1533                           Array(
1534                                 "/expression/si",
1535                                 "/binding/si",
1536                                 "/behaviou*r/si",
1537                                 "|url\(([\'\"])\s*\.\./.*([\'\"])\)|si",
1538                                 "/url\(([\'\"])\s*\S+script\s*:.*([\'\"])\)/si",
1539                                 "/url\(([\'\"])\s*mocha\s*:.*([\'\"])\)/si",
1540                                 "/url\(([\'\"])\s*about\s*:.*([\'\"])\)/si"
1541                                ),
1542                           Array(
1543                                 "idiocy",
1544                                 "idiocy",
1545                                 "idiocy",
1546                                 "url(\\1#\\2)",
1547                                 "url(\\1#\\2)",
1548                                 "url(\\1#\\2)",
1549                                 "url(\\1#\\2)"
1550                                )
1551                           )
1552                 )
1553         );
1554     if (!$view_unsafe_images){
1555         /**
1556          * Remove any references to http/https if view_unsafe_images set
1557          * to false.
1558          */
1559          array_push($bad_attvals{'/.*/'}{'/^src|background/i'}[0],
1560                     '/^([\'\"])\s*https*:.*([\'\"])/si');
1561          array_push($bad_attvals{'/.*/'}{'/^src|background/i'}[1],
1562                     "\\1$secremoveimg\\2");
1563          array_push($bad_attvals{'/.*/'}{'/^style/si'}[0],
1564                     '/url\(([\'\"])\s*https*:.*([\'\"])\)/si');
1565          array_push($bad_attvals{'/.*/'}{'/^style/si'}[1],
1566                     "url(\\1$secremoveimg\\2)");
1567     }
1568
1569     $add_attr_to_tag = Array(
1570                              "/^a$/si" => Array('target'=>'"_new"')
1571                              );
1572     $trusted = sq_sanitize($body,
1573                            $tag_list,
1574                            $rm_tags_with_content,
1575                            $self_closing_tags,
1576                            $force_tag_closing,
1577                            $rm_attnames,
1578                            $bad_attvals,
1579                            $add_attr_to_tag,
1580                            $message,
1581                            $id,
1582                            $mailbox
1583                            );
1584     if (preg_match("|$secremoveimg|si", $trusted)){
1585         $has_unsafe_images = true;
1586     }
1587     return $trusted;
1588 }
1589
1590 ?>