functions/mime.php

   1 <?php
   2
   3 /**
   4  * mime.php
   5  *
   6  * Copyright (c) 1999-2002 The SquirrelMail Project Team
   7  * Licensed under the GNU GPL. For full terms see the file COPYING.
   8  *
   9  * This contains the functions necessary to detect and decode MIME
  10  * messages.
  11  *
  12  * $Id$
  13  */
  14
  15 require_once('../functions/imap.php');
  16 require_once('../functions/attachment_common.php');
  17
  18 /* --------------------------------------------------------------------------------- */
  19 /* MIME DECODING                                                                     */
  20 /* --------------------------------------------------------------------------------- */
  21
  22 /* This function gets the structure of a message and stores it in the "message" class.
  23  * It will return this object for use with all relevant header information and
  24  * fully parsed into the standard "message" object format.
  25  */
  26
  27 function mime_structure ($bodystructure, $flags=array()) {
  28
  29     // isolate the body structure and remove beginning and end parenthesis
  30     $read = trim(substr ($bodystructure, strpos(strtolower($bodystructure), 'bodystructure') + 13));
  31     $msg = &new message();
  32     $read = trim(substr ($read, 0, -1));
  33     $res = $msg->parseStructure($read);
  34     $msg = $res[0];
  35     $msg->setEnt('0');
  36     if (count($flags)) {
  37        foreach ($flags as $flag) {
  38           $char = strtoupper($flag{1});
  39           switch ($char) {
  40              case 'S':
  41                if (strtolower($flag) == '\\seen') {
  42                   $msg->is_seen = true;
  43                }
  44                break;
  45              case 'A':
  46                if (strtolower($flag) == '\\answered') {
  47                   $msg->is_answered = true;
  48                }
  49                break;
  50              case 'D':
  51                if (strtolower($flag) == '\\deleted') {
  52                   $msg->is_deleted = true;
  53                }
  54                break;
  55              case 'F':
  56                if (strtolower($flag) == '\\flagged') {
  57                   $msg->is_flagged = true;
  58                }
  59                break;
  60              case 'M':
  61                if (strtolower($flag) == '$mdnsent') {
  62                   $msg->is_mdnsent = true;
  63                }
  64                break;
  65              default:
  66                break;
  67           }
  68        }
  69     }
  70 //    listEntities($msg);
  71     return( $msg );
  72 }
  73
  74 /* this starts the parsing of a particular structure.  It is called recursively,
  75  * so it can be passed different structures.  It returns an object of type
  76  * $message.
  77  * First, it checks to see if it is a multipart message.  If it is, then it
  78  * handles that as it sees is necessary.  If it is just a regular entity,
  79  * then it parses it and adds the necessary header information (by calling out
  80  * to mime_get_elements()
  81  */
  82
  83 function mime_fetch_body($imap_stream, $id, $ent_id) {
  84     global $uid_support;
  85     /*
  86      * do a bit of error correction.  If we couldn't find the entity id, just guess
  87      * that it is the first one.  That is usually the case anyway.
  88      */
  89     if (!$ent_id) {
  90         $ent_id = 1;
  91     }
  92     $cmd = "FETCH $id BODY[$ent_id]";
  93
  94     $data = sqimap_run_command ($imap_stream, $cmd, true, $response, $message, $uid_support);
  95     do {
  96         $topline = trim(array_shift( $data ));
  97     } while( $topline && $topline[0] == '*' && !preg_match( '/\* [0-9]+ FETCH.*/i', $topline )) ;
  98
  99     $wholemessage = implode('', $data);
 100     if (ereg('\\{([^\\}]*)\\}', $topline, $regs)) {
 101
 102         $ret = substr( $wholemessage, 0, $regs[1] );
 103         /*
 104             There is some information in the content info header that could be important
 105             in order to parse html messages. Let's get them here.
 106         */
 107         if ( $ret{0} == '<' ) {
 108             $data = sqimap_run_command ($imap_stream, "FETCH $id BODY[$ent_id.MIME]", true, $response, $message, $uid_support);
 109         }
 110     } else if (ereg('"([^"]*)"', $topline, $regs)) {
 111         $ret = $regs[1];
 112     } else {
 113         global $where, $what, $mailbox, $passed_id, $startMessage;
 114         $par = 'mailbox=' . urlencode($mailbox) . "&amp;passed_id=$passed_id";
 115         if (isset($where) && isset($what)) {
 116             $par .= '&amp;where='. urlencode($where) . "&amp;what=" . urlencode($what);
 117         } else {
 118             $par .= "&amp;startMessage=$startMessage&amp;show_more=0";
 119         }
 120         $par .= '&amp;response=' . urlencode($response) .
 121                 '&amp;message=' . urlencode($message).
 122                 '&amp;topline=' . urlencode($topline);
 123
 124         echo   '<tt><br>' .
 125                '<table width="80%"><tr>' .
 126                '<tr><td colspan=2>' .
 127                _("Body retrieval error. The reason for this is most probably that the message is malformed. Please help us making future versions better by submitting this message to the developers knowledgebase!") .
 128                " <A HREF=\"../src/retrievalerror.php?$par\"><br>" .
 129                _("Submit message") . '</A><BR>&nbsp;' .
 130                '</td></tr>' .
 131                '<td><b>' . _("Command:") . "</td><td>$cmd</td></tr>" .
 132                '<td><b>' . _("Response:") . "</td><td>$response</td></tr>" .
 133                '<td><b>' . _("Message:") . "</td><td>$message</td></tr>" .
 134                '<td><b>' . _("FETCH line:") . "</td><td>$topline</td></tr>" .
 135                "</table><BR></tt></font><hr>";
 136
 137         $data = sqimap_run_command ($imap_stream, "FETCH $passed_id BODY[]", true, $response, $message, $uid_support);
 138         array_shift($data);
 139         $wholemessage = implode('', $data);
 140
 141         $ret = $wholemessage;
 142     }
 143     return( $ret );
 144 }
 145
 146 function mime_print_body_lines ($imap_stream, $id, $ent_id, $encoding) {
 147     global $uid_support;
 148     // do a bit of error correction.  If we couldn't find the entity id, just guess
 149     // that it is the first one.  That is usually the case anyway.
 150     if (!$ent_id) {
 151         $ent_id = 1;
 152     }
 153     $sid = sqimap_session_id($uid_support);
 154     // Don't kill the connection if the browser is over a dialup
 155     // and it would take over 30 seconds to download it.
 156
 157     // don´t call set_time_limit in safe mode.
 158     if (!ini_get("safe_mode")) {
 159         set_time_limit(0);
 160     }
 161     if ($uid_support) {
 162        $sid_s = substr($sid,0,strpos($sid, ' '));
 163     } else {
 164        $sid_s = $sid;
 165     }
 166
 167     $body = mime_fetch_body ($imap_stream, $id, $ent_id);
 168     echo decodeBody($body, $encoding);
 169     return;
 170 /*
 171     fputs ($imap_stream, "$sid FETCH $id BODY[$ent_id]\r\n");
 172     $cnt = 0;
 173     $continue = true;
 174     $read = fgets ($imap_stream,8192);
 175
 176
 177     // This could be bad -- if the section has sqimap_session_id() . ' OK'
 178     // or similar, it will kill the download.
 179     while (!ereg("^".$sid_s." (OK|BAD|NO)(.*)$", $read, $regs)) {
 180         if (trim($read) == ')==') {
 181             $read1 = $read;
 182             $read = fgets ($imap_stream,4096);
 183             if (ereg("^".$sid." (OK|BAD|NO)(.*)$", $read, $regs)) {
 184                 return;
 185             } else {
 186                 echo decodeBody($read1, $encoding) .
 187                      decodeBody($read, $encoding);
 188             }
 189         } else if ($cnt) {
 190             echo decodeBody($read, $encoding);
 191         }
 192         $read = fgets ($imap_stream,4096);
 193         $cnt++;
 194 //      break;
 195     }
 196 */
 197 }
 198
 199 /* -[ END MIME DECODING ]----------------------------------------------------------- */
 200
 201 /* 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=$urlMailbox&amp;passed_ent_id=$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             $filename = decodeHeader($header->disposition->getProperty('filename'));
 414             if (trim($filename) == '') {
 415                 $name = decodeHeader($header->disposition->getProperty('name'));
 416                 if (trim($name) == '') {
 417                     if ( trim( $header->id ) == '' )
 418                         $filename = 'untitled-[' . $ent . ']' ;
 419                     else
 420                         $filename = 'cid: ' . $header->id;
 421                 } else {
 422                     $filename = $name;
 423                 }
 424             }
 425             if ($header->description) {
 426                 $description = htmlspecialchars($header->description);
 427             } else {
 428                 $description = '';
 429             }
 430         }
 431
 432         $display_filename = $filename;
 433         $DefaultLink = $default_page . "?startMessage=$startMessage"
 434                      . "&amp;passed_id=$id&amp;mailbox=$urlMailbox"
 435                      . "&amp;passed_ent_id=$ent";
 436         if ($where && $what) {
 437            $DefaultLink = '&amp;where='. urlencode($where).'&amp;what='.urlencode($what);
 438         }
 439         /* this executes the attachment hook with a specific MIME-type.
 440             * if that doens't have results, it tries if there's a rule
 441             * for a more generic type. */
 442         $HookResults = do_hook("attachment $type0/$type1", $Links,
 443             $startMessage, $id, $urlMailbox, $ent, $DefaultLink,
 444             $display_filename, $where, $what);
 445         if(count($HookResults[1]) <= 1) {
 446             $HookResults = do_hook("attachment $type0/*", $Links,
 447                     $startMessage, $id, $urlMailbox, $ent, $DefaultLink,
 448                     $display_filename, $where, $what);
 449         }
 450
 451         $Links = $HookResults[1];
 452         $DefaultLink = $HookResults[6];
 453
 454         $attachments .= '<TR><TD>' .
 455                         "<A HREF=\"$DefaultLink\">$display_filename</A>&nbsp;</TD>" .
 456                         '<TD><SMALL><b>' . show_readable_size($header->size) .
 457                         '</b>&nbsp;&nbsp;</small></TD>' .
 458                         "<TD><SMALL>[ $type0/$type1 ]&nbsp;</SMALL></TD>" .
 459                         '<TD><SMALL>';
 460         $attachments .= '<b>' . $description . '</b>';
 461         $attachments .= '</SMALL></TD><TD><SMALL>&nbsp;';
 462
 463         $SkipSpaces = 1;
 464         foreach ($Links as $Val) {
 465             if ($SkipSpaces) {
 466                 $SkipSpaces = 0;
 467             } else {
 468                 $attachments .= '&nbsp;&nbsp;|&nbsp;&nbsp;';
 469             }
 470             $attachments .= '<a href="' . $Val['href'] . '">' .  $Val['text'] . '</a>';
 471         }
 472         unset($Links);
 473         $attachments .= "</TD></TR>\n";
 474     }
 475     return $attachments;
 476 }
 477
 478 /** this function decodes the body depending on the encoding type. **/
 479 function decodeBody($body, $encoding) {
 480   $body = str_replace("\r\n", "\n", $body);
 481   $encoding = strtolower($encoding);
 482
 483   global $show_html_default;
 484
 485   if ($encoding == 'quoted-printable' ||
 486       $encoding == 'quoted_printable') {
 487      $body = quoted_printable_decode($body);
 488
 489      while (ereg("=\n", $body))
 490         $body = ereg_replace ("=\n", "", $body);
 491
 492   } else if ($encoding == 'base64') {
 493      $body = base64_decode($body);
 494   }
 495
 496   // All other encodings are returned raw.
 497   return $body;
 498 }
 499
 500 /*
 501  * This functions decode strings that is encoded according to
 502  * RFC1522 (MIME Part Two: Message Header Extensions for Non-ASCII Text).
 503  * Patched by Christian Schmidt <christian@ostenfeld.dk>  23/03/2002
 504  */
 505 function decodeHeader ($string, $utfencode=true) {
 506     if (is_array($string)) {
 507         $string = implode("\n", $string);
 508     }
 509     $i = 0;
 510     while (preg_match('/^(.{' . $i . '})(.*)=\?([^?]*)\?(Q|B)\?([^?]*)\?=/Ui',
 511                       $string, $res)) {
 512         $prefix = $res[1];
 513         // Ignore white-space between consecutive encoded-words
 514         if (strspn($res[2], " \t") != strlen($res[2])) {
 515             $prefix .= $res[2];
 516         }
 517
 518         if (ucfirst($res[4]) == 'B') {
 519             $replace = base64_decode($res[5]);
 520         } else {
 521             $replace = str_replace('_', ' ', $res[5]);
 522             $replace = preg_replace('/=([0-9a-f]{2})/ie', 'chr(hexdec("\1"))',
 523                                     $replace);
 524             /* Only encode into entities by default. Some places
 525                don't need the encoding, like the compose form. */
 526             if ($utfencode) {
 527                 $replace = charset_decode($res[3], $replace);
 528             }
 529         }
 530         $string = $prefix . $replace . substr($string, strlen($res[0]));
 531         $i = strlen($prefix) + strlen($replace);
 532     }
 533     return( $string );
 534 }
 535
 536 /*
 537  * Encode a string according to RFC 1522 for use in headers if it
 538  * contains 8-bit characters or anything that looks like it should
 539  * be encoded.
 540  */
 541 function encodeHeader ($string) {
 542     global $default_charset;
 543
 544     // Encode only if the string contains 8-bit characters or =?
 545     $j = strlen( $string  );
 546     $l = strstr($string, '=?');         // Must be encoded ?
 547     $ret = '';
 548     for( $i=0; $i < $j; ++$i) {
 549         switch( $string{$i} ) {
 550            case '=':
 551           $ret .= '=3D';
 552           break;
 553         case '?':
 554           $ret .= '=3F';
 555           break;
 556         case '_':
 557           $ret .= '=5F';
 558           break;
 559         case ' ':
 560           $ret .= '_';
 561           break;
 562         default:
 563           $k = ord( $string{$i} );
 564           if ( $k > 126 ) {
 565              $ret .= sprintf("=%02X", $k);
 566              $l = TRUE;
 567           } else
 568              $ret .= $string{$i};
 569         }
 570     }
 571
 572     if ( $l ) {
 573         $string = "=?$default_charset?Q?$ret?=";
 574     }
 575
 576     return( $string );
 577 }
 578
 579 /* This function trys to locate the entity_id of a specific mime element */
 580
 581 function find_ent_id( $id, $message ) {
 582     $ret = '';
 583     for ($i=0; $ret == '' && $i < count($message->entities); $i++) {
 584         if ( $message->entities[$i]->header->type0 == 'multipart')  {
 585             $ret = find_ent_id( $id, $message->entities[$i] );
 586         } else {
 587             if ( strcasecmp( $message->entities[$i]->header->id, $id ) == 0 )
 588                 $ret = $message->entities[$i]->entity_id;
 589         }
 590     }
 591     return( $ret );
 592 }
 593
 594 /**
 595  ** HTMLFILTER ROUTINES
 596  */
 597
 598 /**
 599  * This function returns the final tag out of the tag name, an array
 600  * of attributes, and the type of the tag. This function is called by
 601  * sq_sanitize internally.
 602  *
 603  * @param  $tagname  the name of the tag.
 604  * @param  $attary   the array of attributes and their values
 605  * @param  $tagtype  The type of the tag (see in comments).
 606  * @return           a string with the final tag representation.
 607  */
 608 function sq_tagprint($tagname, $attary, $tagtype){
 609     $me = "sq_tagprint";
 610     if ($tagtype == 2){
 611         $fulltag = '</' . $tagname . '>';
 612     } else {
 613         $fulltag = '<' . $tagname;
 614         if (is_array($attary) && sizeof($attary)){
 615             $atts = Array();
 616             while (list($attname, $attvalue) = each($attary)){
 617                 array_push($atts, "$attname=$attvalue");
 618             }
 619             $fulltag .= ' ' . join(" ", $atts);
 620         }
 621         if ($tagtype == 3){
 622             $fulltag .= " /";
 623         }
 624         $fulltag .= ">";
 625     }
 626     return $fulltag;
 627 }
 628
 629 /**
 630  * A small helper function to use with array_walk. Modifies a by-ref
 631  * value and makes it lowercase.
 632  *
 633  * @param  $val a value passed by-ref.
 634  * @return      void since it modifies a by-ref value.
 635  */
 636 function sq_casenormalize(&$val){
 637     $val = strtolower($val);
 638 }
 639
 640 /**
 641  * This function skips any whitespace from the current position within
 642  * a string and to the next non-whitespace value.
 643  *
 644  * @param  $body   the string
 645  * @param  $offset the offset within the string where we should start
 646  *                 looking for the next non-whitespace character.
 647  * @return         the location within the $body where the next
 648  *                 non-whitespace char is located.
 649  */
 650 function sq_skipspace($body, $offset){
 651     $me = "sq_skipspace";
 652     preg_match("/^(\s*)/s", substr($body, $offset), $matches);
 653     if (sizeof($matches{1})){
 654         $count = strlen($matches{1});
 655         $offset += $count;
 656     }
 657     return $offset;
 658 }
 659
 660 /**
 661  * This function looks for the next character within a string.  It's
 662  * really just a glorified "strpos", except it catches if failures
 663  * nicely.
 664  *
 665  * @param  $body   The string to look for needle in.
 666  * @param  $offset Start looking from this position.
 667  * @param  $needle The character/string to look for.
 668  * @return         location of the next occurance of the needle, or
 669  *                 strlen($body) if needle wasn't found.
 670  */
 671 function sq_findnxstr($body, $offset, $needle){
 672     $me = "sq_findnxstr";
 673     $pos = strpos($body, $needle, $offset);
 674     if ($pos === FALSE){
 675         $pos = strlen($body);
 676     }
 677     return $pos;
 678 }
 679
 680 /**
 681  * This function takes a PCRE-style regexp and tries to match it
 682  * within the string.
 683  *
 684  * @param  $body   The string to look for needle in.
 685  * @param  $offset Start looking from here.
 686  * @param  $reg    A PCRE-style regex to match.
 687  * @return         Returns a false if no matches found, or an array
 688  *                 with the following members:
 689  *                 - integer with the location of the match within $body
 690  *                 - string with whatever content between offset and the match
 691  *                 - string with whatever it is we matched
 692  */
 693 function sq_findnxreg($body, $offset, $reg){
 694     $me = "sq_findnxreg";
 695     $matches = Array();
 696     $retarr = Array();
 697     preg_match("%^(.*?)($reg)%s", substr($body, $offset), $matches);
 698     if (!$matches{0}){
 699         $retarr = false;
 700     } else {
 701         $retarr{0} = $offset + strlen($matches{1});
 702         $retarr{1} = $matches{1};
 703         $retarr{2} = $matches{2};
 704     }
 705     return $retarr;
 706 }
 707
 708 /**
 709  * This function looks for the next tag.
 710  *
 711  * @param  $body   String where to look for the next tag.
 712  * @param  $offset Start looking from here.
 713  * @return         false if no more tags exist in the body, or
 714  *                 an array with the following members:
 715  *                 - string with the name of the tag
 716  *                 - array with attributes and their values
 717  *                 - integer with tag type (1, 2, or 3)
 718  *                 - integer where the tag starts (starting "<")
 719  *                 - integer where the tag ends (ending ">")
 720  *                 first three members will be false, if the tag is invalid.
 721  */
 722 function sq_getnxtag($body, $offset){
 723     $me = "sq_getnxtag";
 724     if ($offset > strlen($body)){
 725         return false;
 726     }
 727     $lt = sq_findnxstr($body, $offset, "<");
 728     if ($lt == strlen($body)){
 729         return false;
 730     }
 731     /**
 732      * We are here:
 733      * blah blah <tag attribute="value">
 734      * \---------^
 735      */
 736     $pos = sq_skipspace($body, $lt+1);
 737     if ($pos >= strlen($body)){
 738         return Array(false, false, false, $lt, strlen($body));
 739     }
 740     /**
 741      * There are 3 kinds of tags:
 742      * 1. Opening tag, e.g.:
 743      *    <a href="blah">
 744      * 2. Closing tag, e.g.:
 745      *    </a>
 746      * 3. XHTML-style content-less tag, e.g.:
 747      *    <img src="blah"/>
 748      */
 749     $tagtype = false;
 750     switch (substr($body, $pos, 1)){
 751     case "/":
 752         $tagtype = 2;
 753         $pos++;
 754         break;
 755     case "!":
 756         /**
 757          * A comment or an SGML declaration.
 758          */
 759         if (substr($body, $pos+1, 2) == "--"){
 760             $gt = strpos($body, "-->", $pos);
 761             if ($gt === false){
 762                 $gt = strlen($body);
 763             } else {
 764                 $gt += 2;
 765             }
 766             return Array(false, false, false, $lt, $gt);
 767         } else {
 768             $gt = sq_findnxstr($body, $pos, ">");
 769             return Array(false, false, false, $lt, $gt);
 770         }
 771         break;
 772     default:
 773         /**
 774          * Assume tagtype 1 for now. If it's type 3, we'll switch values
 775          * later.
 776          */
 777         $tagtype = 1;
 778         break;
 779     }
 780
 781     $tag_start = $pos;
 782     $tagname = '';
 783     /**
 784      * Look for next [\W-_], which will indicate the end of the tag name.
 785      */
 786     $regary = sq_findnxreg($body, $pos, "[^\w\-_]");
 787     if ($regary == false){
 788         return Array(false, false, false, $lt, strlen($body));
 789     }
 790     list($pos, $tagname, $match) = $regary;
 791     $tagname = strtolower($tagname);
 792
 793     /**
 794      * $match can be either of these:
 795      * '>'  indicating the end of the tag entirely.
 796      * '\s' indicating the end of the tag name.
 797      * '/'  indicating that this is type-3 xhtml tag.
 798      *
 799      * Whatever else we find there indicates an invalid tag.
 800      */
 801     switch ($match){
 802     case "/":
 803         /**
 804          * This is an xhtml-style tag with a closing / at the
 805          * end, like so: <img src="blah"/>. Check if it's followed
 806          * by the closing bracket. If not, then this tag is invalid
 807          */
 808         if (substr($body, $pos, 2) == "/>"){
 809             $pos++;
 810             $tagtype = 3;
 811         } else {
 812             $gt = sq_findnxstr($body, $pos, ">");
 813             $retary = Array(false, false, false, $lt, $gt);
 814             return $retary;
 815         }
 816     case ">":
 817         return Array($tagname, false, $tagtype, $lt, $pos);
 818         break;
 819     default:
 820         /**
 821          * Check if it's whitespace
 822          */
 823         if (preg_match("/\s/", $match)){
 824         } else {
 825             /**
 826              * This is an invalid tag! Look for the next closing ">".
 827              */
 828             $gt = sq_findnxstr($body, $offset, ">");
 829             return Array(false, false, false, $lt, $gt);
 830         }
 831     }
 832
 833     /**
 834      * At this point we're here:
 835      * <tagname  attribute='blah'>
 836      * \-------^
 837      *
 838      * At this point we loop in order to find all attributes.
 839      */
 840     $attname = '';
 841     $atttype = false;
 842     $attary = Array();
 843
 844     while ($pos <= strlen($body)){
 845         $pos = sq_skipspace($body, $pos);
 846         if ($pos == strlen($body)){
 847             /**
 848              * Non-closed tag.
 849              */
 850             return Array(false, false, false, $lt, $pos);
 851         }
 852         /**
 853          * See if we arrived at a ">" or "/>", which means that we reached
 854          * the end of the tag.
 855          */
 856         $matches = Array();
 857         if (preg_match("%^(\s*)(>|/>)%s", substr($body, $pos), $matches)) {
 858             /**
 859              * Yep. So we did.
 860              */
 861             $pos += strlen($matches{1});
 862             if ($matches{2} == "/>"){
 863                 $tagtype = 3;
 864                 $pos++;
 865             }
 866             return Array($tagname, $attary, $tagtype, $lt, $pos);
 867         }
 868
 869         /**
 870          * There are several types of attributes, with optional
 871          * [:space:] between members.
 872          * Type 1:
 873          *   attrname[:space:]=[:space:]'CDATA'
 874          * Type 2:
 875          *   attrname[:space:]=[:space:]"CDATA"
 876          * Type 3:
 877          *   attr[:space:]=[:space:]CDATA
 878          * Type 4:
 879          *   attrname
 880          *
 881          * We leave types 1 and 2 the same, type 3 we check for
 882          * '"' and convert to "&quot" if needed, then wrap in
 883          * double quotes. Type 4 we convert into:
 884          * attrname="yes".
 885          */
 886         $regary = sq_findnxreg($body, $pos, "[^\w\-_]");
 887         if ($regary == false){
 888             /**
 889              * Looks like body ended before the end of tag.
 890              */
 891             return Array(false, false, false, $lt, strlen($body));
 892         }
 893         list($pos, $attname, $match) = $regary;
 894         $attname = strtolower($attname);
 895         /**
 896          * We arrived at the end of attribute name. Several things possible
 897          * here:
 898          * '>'  means the end of the tag and this is attribute type 4
 899          * '/'  if followed by '>' means the same thing as above
 900          * '\s' means a lot of things -- look what it's followed by.
 901          *      anything else means the attribute is invalid.
 902          */
 903         switch($match){
 904         case "/":
 905             /**
 906              * This is an xhtml-style tag with a closing / at the
 907              * end, like so: <img src="blah"/>. Check if it's followed
 908              * by the closing bracket. If not, then this tag is invalid
 909              */
 910             if (substr($body, $pos, 2) == "/>"){
 911                 $pos++;
 912                 $tagtype = 3;
 913             } else {
 914                 $gt = sq_findnxstr($body, $pos, ">");
 915                 $retary = Array(false, false, false, $lt, $gt);
 916                 return $retary;
 917             }
 918         case ">":
 919             $attary{$attname} = '"yes"';
 920             return Array($tagname, $attary, $tagtype, $lt, $pos);
 921             break;
 922         default:
 923             /**
 924              * Skip whitespace and see what we arrive at.
 925              */
 926             $pos = sq_skipspace($body, $pos);
 927             $char = substr($body, $pos, 1);
 928             /**
 929              * Two things are valid here:
 930              * '=' means this is attribute type 1 2 or 3.
 931              * \w means this was attribute type 4.
 932              * anything else we ignore and re-loop. End of tag and
 933              * invalid stuff will be caught by our checks at the beginning
 934              * of the loop.
 935              */
 936             if ($char == "="){
 937                 $pos++;
 938                 $pos = sq_skipspace($body, $pos);
 939                 /**
 940                  * Here are 3 possibilities:
 941                  * "'"  attribute type 1
 942                  * '"'  attribute type 2
 943                  * everything else is the content of tag type 3
 944                  */
 945                 $quot = substr($body, $pos, 1);
 946                 if ($quot == "'"){
 947                     $regary = sq_findnxreg($body, $pos+1, "\'");
 948                     if ($regary == false){
 949                         return Array(false, false, false, $lt, strlen($body));
 950                     }
 951                     list($pos, $attval, $match) = $regary;
 952                     $pos++;
 953                     $attary{$attname} = "'" . $attval . "'";
 954                 } else if ($quot == '"'){
 955                     $regary = sq_findnxreg($body, $pos+1, '\"');
 956                     if ($regary == false){
 957                         return Array(false, false, false, $lt, strlen($body));
 958                     }
 959                     list($pos, $attval, $match) = $regary;
 960                     $pos++;
 961                     $attary{$attname} = '"' . $attval . '"';
 962                 } else {
 963                     /**
 964                      * These are hateful. Look for \s, or >.
 965                      */
 966                     $regary = sq_findnxreg($body, $pos, "[\s>]");
 967                     if ($regary == false){
 968                         return Array(false, false, false, $lt, strlen($body));
 969                     }
 970                     list($pos, $attval, $match) = $regary;
 971                     /**
 972                      * If it's ">" it will be caught at the top.
 973                      */
 974                     $attval = preg_replace("/\"/s", "&quot;", $attval);
 975                     $attary{$attname} = '"' . $attval . '"';
 976                 }
 977             } else if (preg_match("|[\w/>]|", $char)) {
 978                 /**
 979                  * That was attribute type 4.
 980                  */
 981                 $attary{$attname} = '"yes"';
 982             } else {
 983                 /**
 984                  * An illegal character. Find next '>' and return.
 985                  */
 986                 $gt = sq_findnxstr($body, $pos, ">");
 987                 return Array(false, false, false, $lt, $gt);
 988             }
 989         }
 990     }
 991     /**
 992      * The fact that we got here indicates that the tag end was never
 993      * found. Return invalid tag indication so it gets stripped.
 994      */
 995     return Array(false, false, false, $lt, strlen($body));
 996 }
 997
 998 /**
 999  * This function checks attribute values for entity-encoded values
1000  * and returns them translated into 8-bit strings so we can run
1001  * checks on them.
1002  *
1003  * @param  $attvalue A string to run entity check against.
1004  * @return           Translated value.
1005  */
1006 function sq_deent($attvalue){
1007     $me="sq_deent";
1008     /**
1009      * See if we have to run the checks first. All entities must start
1010      * with "&".
1011      */
1012     if (strpos($attvalue, "&") === false){
1013         return $attvalue;
1014     }
1015     /**
1016      * Check named entities first.
1017      */
1018     $trans = get_html_translation_table(HTML_ENTITIES);
1019     /**
1020      * Leave &quot; in, as it can mess us up.
1021      */
1022     $trans = array_flip($trans);
1023     unset($trans{"&quot;"});
1024     while (list($ent, $val) = each($trans)){
1025         $attvalue = preg_replace("/$ent*(\W)/si", "$val\\1", $attvalue);
1026     }
1027     /**
1028      * Now translate numbered entities from 1 to 255 if needed.
1029      */
1030     if (strpos($attvalue, "#") !== false){
1031         $omit = Array(34, 39);
1032         for ($asc=1; $asc<256; $asc++){
1033             if (!in_array($asc, $omit)){
1034                 $chr = chr($asc);
1035                 $attvalue = preg_replace("/\&#0*$asc;*(\D)/si", "$chr\\1",
1036                                          $attvalue);
1037                 $attvalue = preg_replace("/\&#x0*".dechex($asc).";*(\W)/si",
1038                                          "$chr\\1", $attvalue);
1039             }
1040         }
1041     }
1042     return $attvalue;
1043 }
1044
1045 /**
1046  * This function runs various checks against the attributes.
1047  *
1048  * @param  $tagname         String with the name of the tag.
1049  * @param  $attary          Array with all tag attributes.
1050  * @param  $rm_attnames     See description for sq_sanitize
1051  * @param  $bad_attvals     See description for sq_sanitize
1052  * @param  $add_attr_to_tag See description for sq_sanitize
1053  * @param  $message         message object
1054  * @param  $id              message id
1055  * @return                  Array with modified attributes.
1056  */
1057 function sq_fixatts($tagname,
1058                     $attary,
1059                     $rm_attnames,
1060                     $bad_attvals,
1061                     $add_attr_to_tag,
1062                     $message,
1063                     $id,
1064                     $mailbox
1065                     ){
1066     $me = "sq_fixatts";
1067     while (list($attname, $attvalue) = each($attary)){
1068         /**
1069          * See if this attribute should be removed.
1070          */
1071         foreach ($rm_attnames as $matchtag=>$matchattrs){
1072             if (preg_match($matchtag, $tagname)){
1073                 foreach ($matchattrs as $matchattr){
1074                     if (preg_match($matchattr, $attname)){
1075                         unset($attary{$attname});
1076                         continue;
1077                     }
1078                 }
1079             }
1080         }
1081         /**
1082          * Remove any entities.
1083          */
1084         $attvalue = sq_deent($attvalue);
1085
1086         /**
1087          * Now let's run checks on the attvalues.
1088          * I don't expect anyone to comprehend this. If you do,
1089          * get in touch with me so I can drive to where you live and
1090          * shake your hand personally. :)
1091          */
1092         foreach ($bad_attvals as $matchtag=>$matchattrs){
1093             if (preg_match($matchtag, $tagname)){
1094                 foreach ($matchattrs as $matchattr=>$valary){
1095                     if (preg_match($matchattr, $attname)){
1096                         /**
1097                          * There are two arrays in valary.
1098                          * First is matches.
1099                          * Second one is replacements
1100                          */
1101                         list($valmatch, $valrepl) = $valary;
1102                         $newvalue =
1103                             preg_replace($valmatch, $valrepl, $attvalue);
1104                         if ($newvalue != $attvalue){
1105                             $attary{$attname} = $newvalue;
1106                         }
1107                     }
1108                 }
1109             }
1110         }
1111         /**
1112          * Turn cid: urls into http-friendly ones.
1113          */
1114         if (preg_match("/^[\'\"]\s*cid:/si", $attvalue)){
1115             $attary{$attname} = sq_cid2http($message, $id, $attvalue, $mailbox);
1116         }
1117     }
1118     /**
1119      * See if we need to append any attributes to this tag.
1120      */
1121     foreach ($add_attr_to_tag as $matchtag=>$addattary){
1122         if (preg_match($matchtag, $tagname)){
1123             $attary = array_merge($attary, $addattary);
1124         }
1125     }
1126     return $attary;
1127 }
1128
1129 /**
1130  * This function edits the style definition to make them friendly and
1131  * usable in squirrelmail.
1132  *
1133  * @param  $message  the message object
1134  * @param  $id       the message id
1135  * @param  $content  a string with whatever is between <style> and </style>
1136  * @return           a string with edited content.
1137  */
1138 function sq_fixstyle($message, $id, $content){
1139     global $view_unsafe_images;
1140     $me = "sq_fixstyle";
1141     /**
1142      * First look for general BODY style declaration, which would be
1143      * like so:
1144      * body {background: blah-blah}
1145      * and change it to .bodyclass so we can just assign it to a <div>
1146      */
1147     $content = preg_replace("|body(\s*\{.*?\})|si", ".bodyclass\\1", $content);
1148     $secremoveimg = "../images/" . _("sec_remove_eng.png");
1149     /**
1150      * Fix url('blah') declarations.
1151      */
1152     $content = preg_replace("|url\(([\'\"])\s*\S+script\s*:.*?([\'\"])\)|si",
1153                             "url(\\1$secremoveimg\\2)", $content);
1154     /**
1155      * Fix url('https*://.*) declarations but only if $view_unsafe_images
1156      * is false.
1157      */
1158     if (!$view_unsafe_images){
1159         $content = preg_replace("|url\(([\'\"])\s*https*:.*?([\'\"])\)|si",
1160                                 "url(\\1$secremoveimg\\2)", $content);
1161     }
1162
1163     /**
1164      * Fix urls that refer to cid:
1165      */
1166     while (preg_match("|url\(([\'\"]\s*cid:.*?[\'\"])\)|si", $content,
1167                       $matches)){
1168         $cidurl = $matches{1};
1169         $httpurl = sq_cid2http($message, $id, $cidurl);
1170         $content = preg_replace("|url\($cidurl\)|si",
1171                                 "url($httpurl)", $content);
1172     }
1173
1174     /**
1175      * Fix stupid css declarations which lead to vulnerabilities
1176      * in IE.
1177      */
1178     $match   = Array('/expression/si',
1179                      '/behaviou*r/si',
1180                      '/binding/si');
1181     $replace = Array('idiocy', 'idiocy', 'idiocy');
1182     $content = preg_replace($match, $replace, $content);
1183     return $content;
1184 }
1185
1186 /**
1187  * This function converts cid: url's into the ones that can be viewed in
1188  * the browser.
1189  *
1190  * @param  $message  the message object
1191  * @param  $id       the message id
1192  * @param  $cidurl   the cid: url.
1193  * @return           a string with a http-friendly url
1194  */
1195 function sq_cid2http($message, $id, $cidurl, $mailbox){
1196     /**
1197      * Get rid of quotes.
1198      */
1199     $quotchar = substr($cidurl, 0, 1);
1200     $cidurl = str_replace($quotchar, "", $cidurl);
1201     $cidurl = substr(trim($cidurl), 4);
1202     $httpurl = $quotchar . "../src/download.php?absolute_dl=true&amp;" .
1203         "passed_id=$id&amp;mailbox=" . urlencode($mailbox) .
1204         "&amp;passed_ent_id=" . find_ent_id($cidurl, $message) . $quotchar;
1205     return $httpurl;
1206 }
1207
1208 /**
1209  * This function changes the <body> tag into a <div> tag since we
1210  * can't really have a body-within-body.
1211  *
1212  * @param  $attary  an array of attributes and values of <body>
1213  * @return          a modified array of attributes to be set for <div>
1214  */
1215 function sq_body2div($attary){
1216     $me = "sq_body2div";
1217     $divattary = Array("class"=>"'bodyclass'");
1218     $bgcolor="#ffffff";
1219     $text="#000000";
1220     $styledef="";
1221     if (is_array($attary) && sizeof($attary) > 0){
1222         foreach ($attary as $attname=>$attvalue){
1223             $quotchar = substr($attvalue, 0, 1);
1224             $attvalue = str_replace($quotchar, "", $attvalue);
1225             switch ($attname){
1226             case "background":
1227                 $styledef .= "background-image: url('$attvalue'); ";
1228                 break;
1229             case "bgcolor":
1230                 $styledef .= "background-color: $attvalue; ";
1231                 break;
1232             case "text":
1233                 $styledef .= "color: $attvalue; ";
1234             }
1235         }
1236         if (strlen($styledef) > 0){
1237             $divattary{"style"} = "\"$styledef\"";
1238         }
1239     }
1240     return $divattary;
1241 }
1242
1243 /**
1244  * This is the main function and the one you should actually be calling.
1245  * There are several variables you should be aware of an which need
1246  * special description.
1247  *
1248  * Since the description is quite lengthy, see it here:
1249  * http://www.mricon.com/html/phpfilter.html
1250  *
1251  * @param $body                 the string with HTML you wish to filter
1252  * @param $tag_list             see description above
1253  * @param $rm_tags_with_content see description above
1254  * @param $self_closing_tags    see description above
1255  * @param $force_tag_closing    see description above
1256  * @param $rm_attnames          see description above
1257  * @param $bad_attvals          see description above
1258  * @param $add_attr_to_tag      see description above
1259  * @param $message              message object
1260  * @param $id                   message id
1261  * @return                      sanitized html safe to show on your pages.
1262  */
1263 function sq_sanitize($body,
1264                      $tag_list,
1265                      $rm_tags_with_content,
1266                      $self_closing_tags,
1267                      $force_tag_closing,
1268                      $rm_attnames,
1269                      $bad_attvals,
1270                      $add_attr_to_tag,
1271                      $message,
1272                      $id,
1273                      $mailbox
1274                      ){
1275     $me = "sq_sanitize";
1276     /**
1277      * Normalize rm_tags and rm_tags_with_content.
1278      */
1279     @array_walk($rm_tags, 'sq_casenormalize');
1280     @array_walk($rm_tags_with_content, 'sq_casenormalize');
1281     @array_walk($self_closing_tags, 'sq_casenormalize');
1282     /**
1283      * See if tag_list is of tags to remove or tags to allow.
1284      * false  means remove these tags
1285      * true   means allow these tags
1286      */
1287     $rm_tags = array_shift($tag_list);
1288     $curpos = 0;
1289     $open_tags = Array();
1290     $trusted = "<!-- begin sanitized html -->\n";
1291     $skip_content = false;
1292     /**
1293      * Take care of netscape's stupid javascript entities like
1294      * &{alert('boo')};
1295      */
1296     $body = preg_replace("/&(\{.*?\};)/si", "&amp;\\1", $body);
1297
1298     while (($curtag=sq_getnxtag($body, $curpos)) != FALSE){
1299         list($tagname, $attary, $tagtype, $lt, $gt) = $curtag;
1300         $free_content = substr($body, $curpos, $lt-$curpos);
1301         /**
1302          * Take care of <style>
1303          */
1304         if ($tagname == "style" && $tagtype == 2){
1305             /**
1306              * This is a closing </style>. Edit the
1307              * content before we apply it.
1308              */
1309             $free_content = sq_fixstyle($message, $id, $free_content);
1310         }
1311         if ($skip_content == false){
1312             $trusted .= $free_content;
1313         } else {
1314         }
1315         if ($tagname != FALSE){
1316             if ($tagtype == 2){
1317                 if ($skip_content == $tagname){
1318                     /**
1319                      * Got to the end of tag we needed to remove.
1320                      */
1321                     $tagname = false;
1322                     $skip_content = false;
1323                 } else {
1324                     if ($skip_content == false){
1325                         if ($tagname == "body"){
1326                             $tagname = "div";
1327                         } else {
1328                             if (isset($open_tags{$tagname}) &&
1329                                 $open_tags{$tagname} > 0){
1330                                 $open_tags{$tagname}--;
1331                             } else {
1332                                 $tagname = false;
1333                             }
1334                         }
1335                     } else {
1336                     }
1337                 }
1338             } else {
1339                 /**
1340                  * $rm_tags_with_content
1341                  */
1342                 if ($skip_content == false){
1343                     /**
1344                      * See if this is a self-closing type and change
1345                      * tagtype appropriately.
1346                      */
1347                     if ($tagtype == 1
1348                         && in_array($tagname, $self_closing_tags)){
1349                         $tagtype=3;
1350                     }
1351                     /**
1352                      * See if we should skip this tag and any content
1353                      * inside it.
1354                      */
1355                     if ($tagtype == 1 &&
1356                         in_array($tagname, $rm_tags_with_content)){
1357                         $skip_content = $tagname;
1358                     } else {
1359                         if (($rm_tags == false
1360                              && in_array($tagname, $tag_list)) ||
1361                             ($rm_tags == true &&
1362                              !in_array($tagname, $tag_list))){
1363                             $tagname = false;
1364                         } else {
1365                             if ($tagtype == 1){
1366                                 if (isset($open_tags{$tagname})){
1367                                     $open_tags{$tagname}++;
1368                                 } else {
1369                                     $open_tags{$tagname}=1;
1370                                 }
1371                             }
1372                             /**
1373                              * This is where we run other checks.
1374                              */
1375                             if (is_array($attary) && sizeof($attary) > 0){
1376                                 $attary = sq_fixatts($tagname,
1377                                                      $attary,
1378                                                      $rm_attnames,
1379                                                      $bad_attvals,
1380                                                      $add_attr_to_tag,
1381                                                      $message,
1382                                                      $id,
1383                                                      $mailbox
1384                                                      );
1385                             }
1386                             /**
1387                              * Convert body into div.
1388                              */
1389                             if ($tagname == "body"){
1390                                 $tagname = "div";
1391                                 $attary = sq_body2div($attary, $message, $id);
1392                             }
1393                         }
1394                     }
1395                 } else {
1396                 }
1397             }
1398             if ($tagname != false && $skip_content == false){
1399                 $trusted .= sq_tagprint($tagname, $attary, $tagtype);
1400             }
1401         } else {
1402         }
1403         $curpos = $gt+1;
1404     }
1405     $trusted .= substr($body, $curpos, strlen($body)-$curpos);
1406     if ($force_tag_closing == true){
1407         foreach ($open_tags as $tagname=>$opentimes){
1408             while ($opentimes > 0){
1409                 $trusted .= '</' . $tagname . '>';
1410                 $opentimes--;
1411             }
1412         }
1413         $trusted .= "\n";
1414     }
1415     $trusted .= "<!-- end sanitized html -->\n";
1416     return $trusted;
1417 }
1418
1419 /**
1420  * This is a wrapper function to call html sanitizing routines.
1421  *
1422  * @param  $body  the body of the message
1423  * @param  $id    the id of the message
1424  * @return        a string with html safe to display in the browser.
1425  */
1426 function magicHTML($body, $id, $message, $mailbox = 'INBOX'){
1427     global $attachment_common_show_images, $view_unsafe_images,
1428         $has_unsafe_images;
1429     /**
1430      * Don't display attached images in HTML mode.
1431      */
1432     $attachment_common_show_images = false;
1433     $tag_list = Array(
1434                       false,
1435                       "object",
1436                       "meta",
1437                       "html",
1438                       "head",
1439                       "base",
1440                       "link"
1441                       );
1442
1443     $rm_tags_with_content = Array(
1444                                   "script",
1445                                   "applet",
1446                                   "embed",
1447                                   "title"
1448                                   );
1449
1450     $self_closing_tags =  Array(
1451                                 "img",
1452                                 "br",
1453                                 "hr",
1454                                 "input"
1455                                 );
1456
1457     $force_tag_closing = false;
1458
1459     $rm_attnames = Array(
1460                          "/.*/" =>
1461                          Array(
1462                                "/target/si",
1463                                "/^on.*/si",
1464                                "/^dynsrc/si",
1465                                "/^data.*/si"
1466                                )
1467                          );
1468
1469     $secremoveimg = "../images/" . _("sec_remove_eng.png");
1470     $bad_attvals = Array(
1471         "/.*/" =>
1472             Array(
1473                 "/^src|background/i" =>
1474                     Array(
1475                           Array(
1476                                 "|^([\'\"])\s*\.\./.*([\'\"])|si",
1477                                 "/^([\'\"])\s*\S+script\s*:.*([\'\"])/si",
1478                                 "/^([\'\"])\s*mocha\s*:*.*([\'\"])/si",
1479                                 "/^([\'\"])\s*about\s*:.*([\'\"])/si"
1480                                 ),
1481                           Array(
1482                                 "\\1$secremoveimg\\2",
1483                                 "\\1$secremoveimg\\2",
1484                                 "\\1$secremoveimg\\2",
1485                                 "\\1$secremoveimg\\2"
1486                                 )
1487                         ),
1488                 "/^href|action/i" =>
1489                     Array(
1490                           Array(
1491                                 "|^([\'\"])\s*\.\./.*([\'\"])|si",
1492                                 "/^([\'\"])\s*\S+script\s*:.*([\'\"])/si",
1493                                 "/^([\'\"])\s*mocha\s*:*.*([\'\"])/si",
1494                                 "/^([\'\"])\s*about\s*:.*([\'\"])/si"
1495                                 ),
1496                           Array(
1497                                 "\\1#\\2",
1498                                 "\\1#\\2",
1499                                 "\\1#\\2",
1500                                 "\\1#\\2"
1501                                 )
1502                         ),
1503                 "/^style/si" =>
1504                     Array(
1505                           Array(
1506                                 "/expression/si",
1507                                 "/binding/si",
1508                                 "/behaviou*r/si",
1509                                 "|url\(([\'\"])\s*\.\./.*([\'\"])\)|si",
1510                                 "/url\(([\'\"])\s*\S+script\s*:.*([\'\"])\)/si",
1511                                 "/url\(([\'\"])\s*mocha\s*:.*([\'\"])\)/si",
1512                                 "/url\(([\'\"])\s*about\s*:.*([\'\"])\)/si"
1513                                ),
1514                           Array(
1515                                 "idiocy",
1516                                 "idiocy",
1517                                 "idiocy",
1518                                 "url(\\1#\\2)",
1519                                 "url(\\1#\\2)",
1520                                 "url(\\1#\\2)",
1521                                 "url(\\1#\\2)"
1522                                )
1523                           )
1524                 )
1525         );
1526     if (!$view_unsafe_images){
1527         /**
1528          * Remove any references to http/https if view_unsafe_images set
1529          * to false.
1530          */
1531          array_push($bad_attvals{'/.*/'}{'/^src|background/i'}[0],
1532                     '/^([\'\"])\s*https*:.*([\'\"])/si');
1533          array_push($bad_attvals{'/.*/'}{'/^src|background/i'}[1],
1534                     "\\1$secremoveimg\\2");
1535          array_push($bad_attvals{'/.*/'}{'/^style/si'}[0],
1536                     '/url\(([\'\"])\s*https*:.*([\'\"])\)/si');
1537          array_push($bad_attvals{'/.*/'}{'/^style/si'}[1],
1538                     "url(\\1$secremoveimg\\2)");
1539     }
1540
1541     $add_attr_to_tag = Array(
1542                              "/^a$/si" => Array('target'=>'"_new"')
1543                              );
1544     $trusted = sq_sanitize($body,
1545                            $tag_list,
1546                            $rm_tags_with_content,
1547                            $self_closing_tags,
1548                            $force_tag_closing,
1549                            $rm_attnames,
1550                            $bad_attvals,
1551                            $add_attr_to_tag,
1552                            $message,
1553                            $id,
1554                            $mailbox
1555                            );
1556     if (preg_match("|$secremoveimg|si", $trusted)){
1557         $has_unsafe_images = true;
1558     }
1559     return $trusted;
1560 }
1561
1562 ?>