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