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(SM_PATH . 'functions/imap.php');
  16 require_once(SM_PATH . '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     $read = trim(substr ($read, 0, -1));
  32     $msg =& new Message();
  33     $res  = $msg->parseStructure($read);
  34     $msg  = $res[0];
  35     if (!is_object($msg)) {
  36         include_once(SM_PATH . 'functions/display_messages.php');
  37         global $color, $mailbox;
  38         displayPageHeader( $color, urldecode($mailbox) );
  39         echo "<BODY TEXT=\"$color[8]\" BGCOLOR=\"$color[4]\" LINK=\"$color[7]\" VLINK=\"$color[7]\" ALINK=\"$color[7]\">\n\n" .
  40          '<CENTER>';
  41         $errormessage  = _("Squirrelmail could not decode the bodystructure of the message");
  42         $errormessage .= '<BR>'._("the provided bodystructure by your imap-server").':<BR><BR>';
  43         $errormessage .= '<table><tr><td>' . htmlspecialchars($read) . '</td></tr></table>';
  44         plain_error_message( $errormessage, $color );
  45         echo '</body></html>';
  46         exit;
  47     }
  48     $msg->setEnt('0');
  49     if (count($flags)) {
  50         foreach ($flags as $flag) {
  51             $char = strtoupper($flag{1});
  52             switch ($char) {
  53                 case 'S':
  54                     if (strtolower($flag) == '\\seen') {
  55                         $msg->is_seen = true;
  56                     }
  57                     break;
  58                 case 'A':
  59                     if (strtolower($flag) == '\\answered') {
  60                         $msg->is_answered = true;
  61                     }
  62                     break;
  63                 case 'D':
  64                     if (strtolower($flag) == '\\deleted') {
  65                         $msg->is_deleted = true;
  66                     }
  67                     break;
  68                 case 'F':
  69                     if (strtolower($flag) == '\\flagged') {
  70                         $msg->is_flagged = true;
  71                     }
  72                     break;
  73                 case 'M':
  74                     if (strtolower($flag) == '$mdnsent') {
  75                         $msg->is_mdnsent = true;
  76                     }
  77                     break;
  78                 default:
  79                     break;
  80             }
  81         }
  82     }
  83     //    listEntities($msg);
  84     return $msg;
  85 }
  86
  87 /* This starts the parsing of a particular structure.  It is called recursively,
  88  * so it can be passed different structures.  It returns an object of type
  89  * $message.
  90  * First, it checks to see if it is a multipart message.  If it is, then it
  91  * handles that as it sees is necessary.  If it is just a regular entity,
  92  * then it parses it and adds the necessary header information (by calling out
  93  * to mime_get_elements()
  94  */
  95
  96 function mime_fetch_body($imap_stream, $id, $ent_id) {
  97     global $uid_support;
  98     /* Do a bit of error correction.  If we couldn't find the entity id, just guess
  99      * that it is the first one.  That is usually the case anyway.
 100      */
 101     if (!$ent_id) {
 102         $ent_id = 1;
 103     }
 104     $cmd = "FETCH $id BODY[$ent_id]";
 105
 106     $data = sqimap_run_command ($imap_stream, $cmd, true, $response, $message, $uid_support);
 107     do {
 108         $topline = trim(array_shift($data));
 109     } while($topline && ($topline[0] == '*') && !preg_match('/\* [0-9]+ FETCH.*/i', $topline)) ;
 110
 111     $wholemessage = implode('', $data);
 112     if (ereg('\\{([^\\}]*)\\}', $topline, $regs)) {
 113         $ret = substr($wholemessage, 0, $regs[1]);
 114         /* There is some information in the content info header that could be important
 115          * in order to parse html messages. Let's get them here.
 116          */
 117         if ($ret{0} == '<') {
 118             $data = sqimap_run_command ($imap_stream, "FETCH $id BODY[$ent_id.MIME]", true, $response, $message, $uid_support);
 119         }
 120     } else if (ereg('"([^"]*)"', $topline, $regs)) {
 121         $ret = $regs[1];
 122     } else {
 123         global $where, $what, $mailbox, $passed_id, $startMessage;
 124         $par = 'mailbox=' . urlencode($mailbox) . '&amp;passed_id=' . $passed_id;
 125         if (isset($where) && isset($what)) {
 126             $par .= '&amp;where=' . urlencode($where) . '&amp;what=' . urlencode($what);
 127         } else {
 128             $par .= '&amp;startMessage=' . $startMessage . '&amp;show_more=0';
 129         }
 130         $par .= '&amp;response=' . urlencode($response) .
 131                 '&amp;message='  . urlencode($message)  .
 132                 '&amp;topline='  . urlencode($topline);
 133
 134         echo   '<tt><br>' .
 135                '<table width="80%"><tr>' .
 136                '<tr><td colspan=2>' .
 137                _("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!") .
 138                " <A HREF=\"../src/retrievalerror.php?$par\"><br>" .
 139                _("Submit message") . '</A><BR>&nbsp;' .
 140                '</td></tr>' .
 141                '<td><b>' . _("Command:") . "</td><td>$cmd</td></tr>" .
 142                '<td><b>' . _("Response:") . "</td><td>$response</td></tr>" .
 143                '<td><b>' . _("Message:") . "</td><td>$message</td></tr>" .
 144                '<td><b>' . _("FETCH line:") . "</td><td>$topline</td></tr>" .
 145                "</table><BR></tt></font><hr>";
 146
 147         $data = sqimap_run_command ($imap_stream, "FETCH $passed_id BODY[]", true, $response, $message, $uid_support);
 148         array_shift($data);
 149         $wholemessage = implode('', $data);
 150
 151         $ret = $wholemessage;
 152     }
 153     return $ret;
 154 }
 155
 156 function mime_print_body_lines ($imap_stream, $id, $ent_id, $encoding) {
 157     global $uid_support;
 158     /* Do a bit of error correction.  If we couldn't find the entity id, just guess
 159      * that it is the first one.  That is usually the case anyway.
 160      */
 161     if (!$ent_id) {
 162         $ent_id = 1;
 163     }
 164     $sid = sqimap_session_id($uid_support);
 165     /* Don't kill the connection if the browser is over a dialup
 166      * and it would take over 30 seconds to download it.
 167      * Don´t call set_time_limit in safe mode.
 168      */
 169
 170     if (!ini_get('safe_mode')) {
 171         set_time_limit(0);
 172     }
 173     if ($uid_support) {
 174        $sid_s = substr($sid,0,strpos($sid, ' '));
 175     } else {
 176        $sid_s = $sid;
 177     }
 178
 179     $body = mime_fetch_body ($imap_stream, $id, $ent_id);
 180     echo decodeBody($body, $encoding);
 181     return;
 182 /*
 183     fputs ($imap_stream, "$sid FETCH $id BODY[$ent_id]\r\n");
 184     $cnt = 0;
 185     $continue = true;
 186     $read = fgets ($imap_stream,8192);
 187
 188
 189     // This could be bad -- if the section has sqimap_session_id() . ' OK'
 190     // or similar, it will kill the download.
 191     while (!ereg("^".$sid_s." (OK|BAD|NO)(.*)$", $read, $regs)) {
 192         if (trim($read) == ')==') {
 193             $read1 = $read;
 194             $read = fgets ($imap_stream,4096);
 195             if (ereg("^".$sid." (OK|BAD|NO)(.*)$", $read, $regs)) {
 196                 return;
 197             } else {
 198                 echo decodeBody($read1, $encoding) .
 199                      decodeBody($read, $encoding);
 200             }
 201         } else if ($cnt) {
 202             echo decodeBody($read, $encoding);
 203         }
 204         $read = fgets ($imap_stream,4096);
 205         $cnt++;
 206 //      break;
 207     }
 208 */
 209 }
 210
 211 /* -[ END MIME DECODING ]----------------------------------------------------------- */
 212
 213 /* This is here for debugging purposes.  It will print out a list
 214  * of all the entity IDs that are in the $message object.
 215  */
 216 function listEntities ($message) {
 217     if ($message) {
 218         echo "<tt>" . $message->entity_id . ' : ' . $message->type0 . '/' . $message->type1 . ' parent = '. $message->parent->entity_id. '<br>';
 219         for ($i = 0; isset($message->entities[$i]); $i++) {
 220             echo "$i : ";
 221             $msg = listEntities($message->entities[$i]);
 222
 223             if ($msg) {
 224                 echo "return: ";
 225                 return $msg;
 226             }
 227         }
 228     }
 229 }
 230
 231 function getPriorityStr($priority) {
 232     $priority_level = substr($priority,0,1);
 233
 234     switch($priority_level) {
 235         /* Check for a higher then normal priority. */
 236         case '1':
 237         case '2':
 238             $priority_string = _("High");
 239             break;
 240
 241         /* Check for a lower then normal priority. */
 242         case '4':
 243         case '5':
 244             $priority_string = _("Low");
 245             break;
 246
 247         /* Check for a normal priority. */
 248         case '3':
 249         default:
 250             $priority_level = '3';
 251             $priority_string = _("Normal");
 252             break;
 253
 254     }
 255     return $priority_string;
 256 }
 257
 258 /* returns a $message object for a particular entity id */
 259 function getEntity ($message, $ent_id) {
 260     return $message->getEntity($ent_id);
 261 }
 262
 263 /* translateText
 264  * Extracted from strings.php 23/03/2002
 265  */
 266
 267 function translateText(&$body, $wrap_at, $charset) {
 268     global $where, $what;   /* from searching */
 269     global $color;          /* color theme */
 270
 271     require_once(SM_PATH . 'functions/url_parser.php');
 272
 273     $body_ary = explode("\n", $body);
 274     for ($i=0; $i < count($body_ary); $i++) {
 275         $line = $body_ary[$i];
 276         if (strlen($line) - 2 >= $wrap_at) {
 277             sqWordWrap($line, $wrap_at);
 278         }
 279         $line = charset_decode($charset, $line);
 280         $line = str_replace("\t", '        ', $line);
 281
 282         parseUrl ($line);
 283
 284         $quotes = 0;
 285         $pos = 0;
 286         $j = strlen($line);
 287
 288         while ($pos < $j) {
 289             if ($line[$pos] == ' ') {
 290                 $pos++;
 291             } else if (strpos($line, '&gt;', $pos) === $pos) {
 292                 $pos += 4;
 293                 $quotes++;
 294             } else {
 295                 break;
 296             }
 297         }
 298
 299         if ($quotes > 1) {
 300             if (!isset($color[14])) {
 301                 $color[14] = '#FF0000';
 302             }
 303             $line = '<FONT COLOR="' . $color[14] . '">' . $line . '</FONT>';
 304         } elseif ($quotes) {
 305             if (!isset($color[13])) {
 306                 $color[13] = '#800000';
 307             }
 308             $line = '<FONT COLOR="' . $color[13] . '">' . $line . '</FONT>';
 309         }
 310
 311         $body_ary[$i] = $line;
 312     }
 313     $body = '<pre>' . implode("\n", $body_ary) . '</pre>';
 314 }
 315
 316
 317 /* This returns a parsed string called $body. That string can then
 318  * be displayed as the actual message in the HTML. It contains
 319  * everything needed, including HTML Tags, Attachments at the
 320  * bottom, etc.
 321  */
 322 function formatBody($imap_stream, $message, $color, $wrap_at, $ent_num, $id, $mailbox='INBOX') {
 323     /* This if statement checks for the entity to show as the
 324      * primary message. To add more of them, just put them in the
 325      * order that is their priority.
 326      */
 327     global $startMessage, $username, $key, $imapServerAddress, $imapPort,
 328            $show_html_default, $has_unsafe_images, $view_unsafe_images, $sort;
 329
 330     $has_unsafe_images= 0;
 331     $body = '';
 332     $urlmailbox = urlencode($mailbox);
 333     $body_message = getEntity($message, $ent_num);
 334     if (($body_message->header->type0 == 'text') ||
 335         ($body_message->header->type0 == 'rfc822')) {
 336         $body = mime_fetch_body ($imap_stream, $id, $ent_num);
 337         $body = decodeBody($body, $body_message->header->encoding);
 338         $hookResults = do_hook("message_body", $body);
 339         $body = $hookResults[1];
 340
 341         /* If there are other types that shouldn't be formatted, add
 342          * them here.
 343          */
 344
 345         if ($body_message->header->type1 == 'html') {
 346             if ($show_html_default <> 1) {
 347                 $entity_conv = array('&nbsp;' => ' ',
 348                                      '<p>'    => "\n",
 349                                      '<br>'   => "\n",
 350                                      '<P>'    => "\n",
 351                                      '<BR>'   => "\n",
 352                                      '&gt;'   => '>',
 353                                      '&lt;'   => '<');
 354                 $body = strtr($body, $entity_conv);
 355                 $body = strip_tags($body);
 356                 $body = trim($body);
 357                 translateText($body, $wrap_at,
 358                               $body_message->header->getParameter('charset'));
 359             } else {
 360                 $body = magicHTML($body, $id, $message, $mailbox);
 361             }
 362         } else {
 363             translateText($body, $wrap_at,
 364                           $body_message->header->getParameter('charset'));
 365         }
 366
 367         if ($has_unsafe_images) {
 368             if ($view_unsafe_images) {
 369                 $untext = '">' . _("Hide Unsafe Images");
 370             } else {
 371                 $untext = '&amp;view_unsafe_images=1">' . _("View Unsafe Images");
 372             }
 373             $body .= '<center><small><a href="read_body.php?passed_id=' . $id .
 374                      '&amp;passed_ent_id=' . $message->entity_id . '&amp;mailbox=' . $urlmailbox .
 375                      '&amp;sort=' . $sort . '&amp;startMessage=' . $startMessage . '&amp;show_more=0' .
 376                      $untext . '</a></small></center><br>' . "\n";
 377         }
 378     }
 379     return $body;
 380 }
 381
 382
 383 function formatAttachments($message, $exclude_id, $mailbox, $id) {
 384     global $where, $what, $startMessage, $color;
 385     static $ShownHTML = 0;
 386
 387     $att_ar = $message->getAttachments($exclude_id);
 388
 389     if (!count($att_ar)) return '';
 390
 391     $attachments = '';
 392
 393     $urlMailbox = urlencode($mailbox);
 394
 395     foreach ($att_ar as $att) {
 396         $ent = urldecode($att->entity_id);
 397         $header = $att->header;
 398         $type0 = strtolower($header->type0);
 399         $type1 = strtolower($header->type1);
 400         $name = '';
 401         $links['download link']['text'] = _("download");
 402         $links['download link']['href'] =
 403                 "../src/download.php?absolute_dl=true&amp;passed_id=$id&amp;mailbox=$urlMailbox&amp;ent_id=$ent";
 404         $ImageURL = '';
 405         if ($type0 =='message' && $type1 == 'rfc822') {
 406             $default_page = '../src/read_body.php';
 407             $rfc822_header = $att->rfc822_header;
 408             $filename = decodeHeader($rfc822_header->subject);
 409
 410             $from_o = $rfc822_header->from;
 411             if (is_object($from_o)) {
 412                 $from_name = $from_o->getAddress(false);
 413             } else {
 414                 $from_name = _("Unknown sender");
 415             }
 416             $from_name = decodeHeader(htmlspecialchars($from_name));
 417             $description = $from_name;
 418         } else {
 419             $default_page = '../src/download.php';
 420             if (is_object($header->disposition)) {
 421                 $filename = decodeHeader($header->disposition->getProperty('filename'));
 422                 if (trim($filename) == '') {
 423                     $name = decodeHeader($header->disposition->getProperty('name'));
 424                     if (trim($name) == '') {
 425                         if (trim( $header->id ) == '') {
 426                             $filename = 'untitled-[' . $ent . ']' ;
 427                         } else {
 428                             $filename = 'cid: ' . $header->id;
 429                         }
 430                     } else {
 431                         $filename = $name;
 432                     }
 433                 }
 434             } else {
 435                 if (trim( $header->id ) == '') {
 436                     $filename = 'untitled-[' . $ent . ']' ;
 437                 } else {
 438                     $filename = 'cid: ' . $header->id;
 439                 }
 440             }
 441
 442             if ($header->description) {
 443                 $description = htmlspecialchars($header->description);
 444             } else {
 445                 $description = '';
 446             }
 447         }
 448
 449         $display_filename = $filename;
 450         if (isset($passed_ent_id)) {
 451             $passed_ent_id_link = '&amp;passed_ent_id='.$passed_ent_id;
 452         } else {
 453             $passed_ent_id_link = '';
 454         }
 455         $defaultlink = $default_page . "?startMessage=$startMessage"
 456                      . "&amp;passed_id=$id&amp;mailbox=$urlMailbox"
 457                      . '&amp;ent_id='.$ent.$passed_ent_id_link;
 458         if ($where && $what) {
 459            $defaultlink = '&amp;where='. urlencode($where).'&amp;what='.urlencode($what);
 460         }
 461         /* This executes the attachment hook with a specific MIME-type.
 462          * If that doesn't have results, it tries if there's a rule
 463          * for a more generic type.
 464          */
 465         $hookresults = do_hook("attachment $type0/$type1", $links,
 466                                $startMessage, $id, $urlMailbox, $ent, $defaultlink,
 467                                $display_filename, $where, $what);
 468         if(count($hookresults[1]) <= 1) {
 469             $hookresults = do_hook("attachment $type0/*", $links,
 470                                    $startMessage, $id, $urlMailbox, $ent, $defaultlink,
 471                                    $display_filename, $where, $what);
 472         }
 473
 474         $links = $hookresults[1];
 475         $defaultlink = $hookresults[6];
 476
 477         $attachments .= '<TR><TD>' .
 478                         "<A HREF=\"$defaultlink\">$display_filename</A>&nbsp;</TD>" .
 479                         '<TD><SMALL><b>' . show_readable_size($header->size) .
 480                         '</b>&nbsp;&nbsp;</small></TD>' .
 481                         "<TD><SMALL>[ $type0/$type1 ]&nbsp;</SMALL></TD>" .
 482                         '<TD><SMALL>';
 483         $attachments .= '<b>' . $description . '</b>';
 484         $attachments .= '</SMALL></TD><TD><SMALL>&nbsp;';
 485
 486         $skipspaces = 1;
 487         foreach ($links as $val) {
 488             if ($skipspaces) {
 489                 $skipspaces = 0;
 490             } else {
 491                 $attachments .= '&nbsp;&nbsp;|&nbsp;&nbsp;';
 492             }
 493             $attachments .= '<a href="' . $val['href'] . '">' .  $val['text'] . '</a>';
 494         }
 495         unset($links);
 496         $attachments .= "</TD></TR>\n";
 497     }
 498     return $attachments;
 499 }
 500
 501 /* This function decodes the body depending on the encoding type. */
 502 function decodeBody($body, $encoding) {
 503     global $languages, $squirrelmail_language;
 504     global $show_html_default;
 505
 506     $body = str_replace("\r\n", "\n", $body);
 507     $encoding = strtolower($encoding);
 508
 509     if ($encoding == 'quoted-printable' ||
 510         $encoding == 'quoted_printable') {
 511         $body = quoted_printable_decode($body);
 512
 513         while (ereg("=\n", $body)) {
 514             $body = ereg_replace ("=\n", '', $body);
 515         }
 516
 517     } else if ($encoding == 'base64') {
 518         $body = base64_decode($body);
 519     }
 520
 521     if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
 522         function_exists($languages[$squirrelmail_language]['XTRA_CODE'])) {
 523         $body = $languages[$squirrelmail_language]['XTRA_CODE']('decode', $body);
 524     }
 525
 526     // All other encodings are returned raw.
 527     return $body;
 528 }
 529
 530 /*
 531  * This functions decode strings that is encoded according to
 532  * RFC1522 (MIME Part Two: Message Header Extensions for Non-ASCII Text).
 533  * Patched by Christian Schmidt <christian@ostenfeld.dk>  23/03/2002
 534  */
 535 function decodeHeader ($string, $utfencode=true) {
 536     global $languages, $squirrelmail_language;
 537     if (is_array($string)) {
 538         $string = implode("\n", $string);
 539     }
 540
 541     if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
 542         function_exists($languages[$squirrelmail_language]['XTRA_CODE'])) {
 543         $string = $languages[$squirrelmail_language]['XTRA_CODE']('decodeheader', $string);
 544     }
 545
 546     $i = 0;
 547     while (preg_match('/^(.{' . $i . '})(.*)=\?([^?]*)\?(Q|B)\?([^?]*)\?=/Ui',
 548                       $string, $res)) {
 549         $prefix = $res[1];
 550         /* Ignore white-space between consecutive encoded-words. */
 551         if (strspn($res[2], " \t") != strlen($res[2])) {
 552             $prefix .= $res[2];
 553         }
 554
 555         if (ucfirst($res[4]) == 'B') {
 556             $replace = base64_decode($res[5]);
 557         } else {
 558             $replace = str_replace('_', ' ', $res[5]);
 559             $replace = preg_replace('/=([0-9a-f]{2})/ie', 'chr(hexdec("\1"))',
 560                                     $replace);
 561             /* Only encode into entities by default. Some places
 562              * don't need the encoding, like the compose form.
 563              */
 564             if ($utfencode) {
 565                 $replace = charset_decode($res[3], $replace);
 566             }
 567         }
 568         $string = $prefix . $replace . substr($string, strlen($res[0]));
 569         $i = strlen($prefix) + strlen($replace);
 570     }
 571     return $string;
 572 }
 573
 574 /*
 575  * Encode a string according to RFC 1522 for use in headers if it
 576  * contains 8-bit characters or anything that looks like it should
 577  * be encoded.
 578  */
 579 function encodeHeader ($string) {
 580     global $default_charset, $languages, $squirrelmail_language;
 581
 582     if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
 583         function_exists($languages[$squirrelmail_language]['XTRA_CODE'])) {
 584         return  $languages[$squirrelmail_language]['XTRA_CODE']('encodeheader', $string);
 585     }
 586
 587     // Encode only if the string contains 8-bit characters or =?
 588     $j = strlen($string);
 589     $l = strstr($string, '=?');         // Must be encoded ?
 590     $ret = '';
 591     for($i = 0; $i < $j; ++$i) {
 592         switch($string{$i}) {
 593             case '=':
 594                 $ret .= '=3D';
 595                 break;
 596             case '?':
 597                 $ret .= '=3F';
 598                 break;
 599             case '_':
 600                 $ret .= '=5F';
 601                 break;
 602             case ' ':
 603                 $ret .= '_';
 604                 break;
 605             default:
 606                 $k = ord($string{$i});
 607                 if ($k > 126) {
 608                     $ret .= sprintf("=%02X", $k);
 609                     $l = TRUE;
 610                 } else {
 611                     $ret .= $string{$i};
 612                 }
 613                 break;
 614         }
 615     }
 616
 617     if ($l) {
 618         $string = "=?$default_charset?Q?$ret?=";
 619     }
 620
 621     return $string;
 622 }
 623
 624 /* This function trys to locate the entity_id of a specific mime element */
 625 function find_ent_id($id, $message) {
 626     for ($i = 0, $ret = ''; $ret == '' && $i < count($message->entities); $i++) {
 627         if ($message->entities[$i]->header->type0 == 'multipart')  {
 628             $ret = find_ent_id($id, $message->entities[$i]);
 629         } else {
 630             if (strcasecmp($message->entities[$i]->header->id, $id) == 0) {
 631                 if (sq_check_save_extension($message->entities[$i])) {
 632                     return $message->entities[$i]->entity_id;
 633                 }
 634             }
 635         }
 636     }
 637     return $ret;
 638 }
 639
 640 function sq_check_save_extension($message) {
 641     $filename = $message->getFilename();
 642     $ext = substr($filename, strrpos($filename,'.')+1);
 643     $save_extensions = array('jpg','jpeg','gif','png','bmp');
 644     return in_array($ext, $save_extensions);
 645 }
 646
 647
 648 /**
 649  ** HTMLFILTER ROUTINES
 650  */
 651
 652 /**
 653  * This function returns the final tag out of the tag name, an array
 654  * of attributes, and the type of the tag. This function is called by
 655  * sq_sanitize internally.
 656  *
 657  * @param  $tagname  the name of the tag.
 658  * @param  $attary   the array of attributes and their values
 659  * @param  $tagtype  The type of the tag (see in comments).
 660  * @return           a string with the final tag representation.
 661  */
 662 function sq_tagprint($tagname, $attary, $tagtype){
 663     $me = 'sq_tagprint';
 664
 665     if ($tagtype == 2){
 666         $fulltag = '</' . $tagname . '>';
 667     } else {
 668         $fulltag = '<' . $tagname;
 669         if (is_array($attary) && sizeof($attary)){
 670             $atts = Array();
 671             while (list($attname, $attvalue) = each($attary)){
 672                 array_push($atts, "$attname=$attvalue");
 673             }
 674             $fulltag .= ' ' . join(" ", $atts);
 675         }
 676         if ($tagtype == 3){
 677             $fulltag .= ' /';
 678         }
 679         $fulltag .= '>';
 680     }
 681     return $fulltag;
 682 }
 683
 684 /**
 685  * A small helper function to use with array_walk. Modifies a by-ref
 686  * value and makes it lowercase.
 687  *
 688  * @param  $val a value passed by-ref.
 689  * @return      void since it modifies a by-ref value.
 690  */
 691 function sq_casenormalize(&$val){
 692     $val = strtolower($val);
 693 }
 694
 695 /**
 696  * This function skips any whitespace from the current position within
 697  * a string and to the next non-whitespace value.
 698  *
 699  * @param  $body   the string
 700  * @param  $offset the offset within the string where we should start
 701  *                 looking for the next non-whitespace character.
 702  * @return         the location within the $body where the next
 703  *                 non-whitespace char is located.
 704  */
 705 function sq_skipspace($body, $offset){
 706     $me = 'sq_skipspace';
 707     preg_match('/^(\s*)/s', substr($body, $offset), $matches);
 708     if (sizeof($matches{1})){
 709         $count = strlen($matches{1});
 710         $offset += $count;
 711     }
 712     return $offset;
 713 }
 714
 715 /**
 716  * This function looks for the next character within a string.  It's
 717  * really just a glorified "strpos", except it catches if failures
 718  * nicely.
 719  *
 720  * @param  $body   The string to look for needle in.
 721  * @param  $offset Start looking from this position.
 722  * @param  $needle The character/string to look for.
 723  * @return         location of the next occurance of the needle, or
 724  *                 strlen($body) if needle wasn't found.
 725  */
 726 function sq_findnxstr($body, $offset, $needle){
 727     $me  = 'sq_findnxstr';
 728     $pos = strpos($body, $needle, $offset);
 729     if ($pos === FALSE){
 730         $pos = strlen($body);
 731     }
 732     return $pos;
 733 }
 734
 735 /**
 736  * This function takes a PCRE-style regexp and tries to match it
 737  * within the string.
 738  *
 739  * @param  $body   The string to look for needle in.
 740  * @param  $offset Start looking from here.
 741  * @param  $reg    A PCRE-style regex to match.
 742  * @return         Returns a false if no matches found, or an array
 743  *                 with the following members:
 744  *                 - integer with the location of the match within $body
 745  *                 - string with whatever content between offset and the match
 746  *                 - string with whatever it is we matched
 747  */
 748 function sq_findnxreg($body, $offset, $reg){
 749     $me = 'sq_findnxreg';
 750     $matches = Array();
 751     $retarr = Array();
 752     preg_match("%^(.*?)($reg)%s", substr($body, $offset), $matches);
 753     if (!$matches{0}){
 754         $retarr = false;
 755     } else {
 756         $retarr{0} = $offset + strlen($matches{1});
 757         $retarr{1} = $matches{1};
 758         $retarr{2} = $matches{2};
 759     }
 760     return $retarr;
 761 }
 762
 763 /**
 764  * This function looks for the next tag.
 765  *
 766  * @param  $body   String where to look for the next tag.
 767  * @param  $offset Start looking from here.
 768  * @return         false if no more tags exist in the body, or
 769  *                 an array with the following members:
 770  *                 - string with the name of the tag
 771  *                 - array with attributes and their values
 772  *                 - integer with tag type (1, 2, or 3)
 773  *                 - integer where the tag starts (starting "<")
 774  *                 - integer where the tag ends (ending ">")
 775  *                 first three members will be false, if the tag is invalid.
 776  */
 777 function sq_getnxtag($body, $offset){
 778     $me = 'sq_getnxtag';
 779     if ($offset > strlen($body)){
 780         return false;
 781     }
 782     $lt = sq_findnxstr($body, $offset, "<");
 783     if ($lt == strlen($body)){
 784         return false;
 785     }
 786     /**
 787      * We are here:
 788      * blah blah <tag attribute="value">
 789      * \---------^
 790      */
 791     $pos = sq_skipspace($body, $lt+1);
 792     if ($pos >= strlen($body)){
 793         return Array(false, false, false, $lt, strlen($body));
 794     }
 795     /**
 796      * There are 3 kinds of tags:
 797      * 1. Opening tag, e.g.:
 798      *    <a href="blah">
 799      * 2. Closing tag, e.g.:
 800      *    </a>
 801      * 3. XHTML-style content-less tag, e.g.:
 802      *    <img src="blah"/>
 803      */
 804     $tagtype = false;
 805     switch (substr($body, $pos, 1)){
 806         case '/':
 807             $tagtype = 2;
 808             $pos++;
 809             break;
 810         case '!':
 811             /**
 812              * A comment or an SGML declaration.
 813              */
 814             if (substr($body, $pos+1, 2) == "--"){
 815                 $gt = strpos($body, "-->", $pos);
 816                 if ($gt === false){
 817                     $gt = strlen($body);
 818                 } else {
 819                     $gt += 2;
 820                 }
 821                 return Array(false, false, false, $lt, $gt);
 822             } else {
 823                 $gt = sq_findnxstr($body, $pos, ">");
 824                 return Array(false, false, false, $lt, $gt);
 825             }
 826             break;
 827         default:
 828             /**
 829              * Assume tagtype 1 for now. If it's type 3, we'll switch values
 830              * later.
 831              */
 832             $tagtype = 1;
 833             break;
 834     }
 835
 836     $tag_start = $pos;
 837     $tagname = '';
 838     /**
 839      * Look for next [\W-_], which will indicate the end of the tag name.
 840      */
 841     $regary = sq_findnxreg($body, $pos, "[^\w\-_]");
 842     if ($regary == false){
 843         return Array(false, false, false, $lt, strlen($body));
 844     }
 845     list($pos, $tagname, $match) = $regary;
 846     $tagname = strtolower($tagname);
 847
 848     /**
 849      * $match can be either of these:
 850      * '>'  indicating the end of the tag entirely.
 851      * '\s' indicating the end of the tag name.
 852      * '/'  indicating that this is type-3 xhtml tag.
 853      *
 854      * Whatever else we find there indicates an invalid tag.
 855      */
 856     switch ($match){
 857         case '/':
 858             /**
 859              * This is an xhtml-style tag with a closing / at the
 860              * end, like so: <img src="blah"/>. Check if it's followed
 861              * by the closing bracket. If not, then this tag is invalid
 862              */
 863             if (substr($body, $pos, 2) == "/>"){
 864                 $pos++;
 865                 $tagtype = 3;
 866             } else {
 867                 $gt = sq_findnxstr($body, $pos, ">");
 868                 $retary = Array(false, false, false, $lt, $gt);
 869                 return $retary;
 870             }
 871         case '>':
 872             return Array($tagname, false, $tagtype, $lt, $pos);
 873             break;
 874         default:
 875             /**
 876              * Check if it's whitespace
 877              */
 878             if (!preg_match('/\s/', $match)){
 879                 /**
 880                  * This is an invalid tag! Look for the next closing ">".
 881                  */
 882                 $gt = sq_findnxstr($body, $offset, ">");
 883                 return Array(false, false, false, $lt, $gt);
 884             }
 885             break;
 886     }
 887
 888     /**
 889      * At this point we're here:
 890      * <tagname  attribute='blah'>
 891      * \-------^
 892      *
 893      * At this point we loop in order to find all attributes.
 894      */
 895     $attname = '';
 896     $atttype = false;
 897     $attary = Array();
 898
 899     while ($pos <= strlen($body)){
 900         $pos = sq_skipspace($body, $pos);
 901         if ($pos == strlen($body)){
 902             /**
 903              * Non-closed tag.
 904              */
 905             return Array(false, false, false, $lt, $pos);
 906         }
 907         /**
 908          * See if we arrived at a ">" or "/>", which means that we reached
 909          * the end of the tag.
 910          */
 911         $matches = Array();
 912         if (preg_match("%^(\s*)(>|/>)%s", substr($body, $pos), $matches)) {
 913             /**
 914              * Yep. So we did.
 915              */
 916             $pos += strlen($matches{1});
 917             if ($matches{2} == "/>"){
 918                 $tagtype = 3;
 919                 $pos++;
 920             }
 921             return Array($tagname, $attary, $tagtype, $lt, $pos);
 922         }
 923
 924         /**
 925          * There are several types of attributes, with optional
 926          * [:space:] between members.
 927          * Type 1:
 928          *   attrname[:space:]=[:space:]'CDATA'
 929          * Type 2:
 930          *   attrname[:space:]=[:space:]"CDATA"
 931          * Type 3:
 932          *   attr[:space:]=[:space:]CDATA
 933          * Type 4:
 934          *   attrname
 935          *
 936          * We leave types 1 and 2 the same, type 3 we check for
 937          * '"' and convert to "&quot" if needed, then wrap in
 938          * double quotes. Type 4 we convert into:
 939          * attrname="yes".
 940          */
 941         $regary = sq_findnxreg($body, $pos, "[^\w\-_]");
 942         if ($regary == false){
 943             /**
 944              * Looks like body ended before the end of tag.
 945              */
 946             return Array(false, false, false, $lt, strlen($body));
 947         }
 948         list($pos, $attname, $match) = $regary;
 949         $attname = strtolower($attname);
 950         /**
 951          * We arrived at the end of attribute name. Several things possible
 952          * here:
 953          * '>'  means the end of the tag and this is attribute type 4
 954          * '/'  if followed by '>' means the same thing as above
 955          * '\s' means a lot of things -- look what it's followed by.
 956          *      anything else means the attribute is invalid.
 957          */
 958         switch($match){
 959             case '/':
 960                 /**
 961                  * This is an xhtml-style tag with a closing / at the
 962                  * end, like so: <img src="blah"/>. Check if it's followed
 963                  * by the closing bracket. If not, then this tag is invalid
 964                  */
 965                 if (substr($body, $pos, 2) == "/>"){
 966                     $pos++;
 967                     $tagtype = 3;
 968                 } else {
 969                     $gt = sq_findnxstr($body, $pos, ">");
 970                     $retary = Array(false, false, false, $lt, $gt);
 971                     return $retary;
 972                 }
 973             case '>':
 974                 $attary{$attname} = '"yes"';
 975                 return Array($tagname, $attary, $tagtype, $lt, $pos);
 976                 break;
 977             default:
 978                 /**
 979                  * Skip whitespace and see what we arrive at.
 980                  */
 981                 $pos = sq_skipspace($body, $pos);
 982                 $char = substr($body, $pos, 1);
 983                 /**
 984                  * Two things are valid here:
 985                  * '=' means this is attribute type 1 2 or 3.
 986                  * \w means this was attribute type 4.
 987                  * anything else we ignore and re-loop. End of tag and
 988                  * invalid stuff will be caught by our checks at the beginning
 989                  * of the loop.
 990                  */
 991                 if ($char == "="){
 992                     $pos++;
 993                     $pos = sq_skipspace($body, $pos);
 994                     /**
 995                      * Here are 3 possibilities:
 996                      * "'"  attribute type 1
 997                      * '"'  attribute type 2
 998                      * everything else is the content of tag type 3
 999                      */
1000                     $quot = substr($body, $pos, 1);
1001                     if ($quot == "'"){
1002                         $regary = sq_findnxreg($body, $pos+1, "\'");
1003                         if ($regary == false){
1004                             return Array(false, false, false, $lt, strlen($body));
1005                         }
1006                         list($pos, $attval, $match) = $regary;
1007                         $pos++;
1008                         $attary{$attname} = "'" . $attval . "'";
1009                     } else if ($quot == '"'){
1010                         $regary = sq_findnxreg($body, $pos+1, '\"');
1011                         if ($regary == false){
1012                             return Array(false, false, false, $lt, strlen($body));
1013                         }
1014                         list($pos, $attval, $match) = $regary;
1015                         $pos++;
1016                         $attary{$attname} = '"' . $attval . '"';
1017                     } else {
1018                         /**
1019                          * These are hateful. Look for \s, or >.
1020                          */
1021                         $regary = sq_findnxreg($body, $pos, "[\s>]");
1022                         if ($regary == false){
1023                             return Array(false, false, false, $lt, strlen($body));
1024                         }
1025                         list($pos, $attval, $match) = $regary;
1026                         /**
1027                          * If it's ">" it will be caught at the top.
1028                          */
1029                         $attval = preg_replace("/\"/s", "&quot;", $attval);
1030                         $attary{$attname} = '"' . $attval . '"';
1031                     }
1032                 } else if (preg_match("|[\w/>]|", $char)) {
1033                     /**
1034                      * That was attribute type 4.
1035                      */
1036                     $attary{$attname} = '"yes"';
1037                 } else {
1038                     /**
1039                      * An illegal character. Find next '>' and return.
1040                      */
1041                     $gt = sq_findnxstr($body, $pos, ">");
1042                     return Array(false, false, false, $lt, $gt);
1043                 }
1044                 break;
1045         }
1046     }
1047     /**
1048      * The fact that we got here indicates that the tag end was never
1049      * found. Return invalid tag indication so it gets stripped.
1050      */
1051     return Array(false, false, false, $lt, strlen($body));
1052 }
1053
1054 /**
1055  * This function checks attribute values for entity-encoded values
1056  * and returns them translated into 8-bit strings so we can run
1057  * checks on them.
1058  *
1059  * @param  $attvalue A string to run entity check against.
1060  * @return           Translated value.
1061  */
1062 function sq_deent($attvalue){
1063     $me = 'sq_deent';
1064     /**
1065      * See if we have to run the checks first. All entities must start
1066      * with "&".
1067      */
1068     if (strpos($attvalue, "&") === false){
1069         return $attvalue;
1070     }
1071     /**
1072      * Check named entities first.
1073      */
1074     $trans = get_html_translation_table(HTML_ENTITIES);
1075     /**
1076      * Leave &quot; in, as it can mess us up.
1077      */
1078     $trans = array_flip($trans);
1079     unset($trans{"&quot;"});
1080     while (list($ent, $val) = each($trans)){
1081         $attvalue = preg_replace("/$ent*(\W)/si", "$val\\1", $attvalue);
1082     }
1083     /**
1084      * Now translate numbered entities from 1 to 255 if needed.
1085      */
1086     if (strpos($attvalue, "#") !== false){
1087         $omit = Array(34, 39);
1088         for ($asc=1; $asc<256; $asc++){
1089             if (!in_array($asc, $omit)){
1090                 $chr = chr($asc);
1091                 $attvalue = preg_replace("/\&#0*$asc;*(\D)/si", "$chr\\1",
1092                                          $attvalue);
1093                 $attvalue = preg_replace("/\&#x0*".dechex($asc).";*(\W)/si",
1094                                          "$chr\\1", $attvalue);
1095             }
1096         }
1097     }
1098     return $attvalue;
1099 }
1100
1101 /**
1102  * This function runs various checks against the attributes.
1103  *
1104  * @param  $tagname         String with the name of the tag.
1105  * @param  $attary          Array with all tag attributes.
1106  * @param  $rm_attnames     See description for sq_sanitize
1107  * @param  $bad_attvals     See description for sq_sanitize
1108  * @param  $add_attr_to_tag See description for sq_sanitize
1109  * @param  $message         message object
1110  * @param  $id              message id
1111  * @return                  Array with modified attributes.
1112  */
1113 function sq_fixatts($tagname,
1114                     $attary,
1115                     $rm_attnames,
1116                     $bad_attvals,
1117                     $add_attr_to_tag,
1118                     $message,
1119                     $id,
1120                     $mailbox
1121                     ){
1122     $me = 'sq_fixatts';
1123     while (list($attname, $attvalue) = each($attary)){
1124         /**
1125          * See if this attribute should be removed.
1126          */
1127         foreach ($rm_attnames as $matchtag=>$matchattrs){
1128             if (preg_match($matchtag, $tagname)){
1129                 foreach ($matchattrs as $matchattr){
1130                     if (preg_match($matchattr, $attname)){
1131                         unset($attary{$attname});
1132                         continue;
1133                     }
1134                 }
1135             }
1136         }
1137         /**
1138          * Remove any entities.
1139          */
1140         $attvalue = sq_deent($attvalue);
1141
1142         /**
1143          * Now let's run checks on the attvalues.
1144          * I don't expect anyone to comprehend this. If you do,
1145          * get in touch with me so I can drive to where you live and
1146          * shake your hand personally. :)
1147          */
1148         foreach ($bad_attvals as $matchtag=>$matchattrs){
1149             if (preg_match($matchtag, $tagname)){
1150                 foreach ($matchattrs as $matchattr=>$valary){
1151                     if (preg_match($matchattr, $attname)){
1152                         /**
1153                          * There are two arrays in valary.
1154                          * First is matches.
1155                          * Second one is replacements
1156                          */
1157                         list($valmatch, $valrepl) = $valary;
1158                         $newvalue =
1159                             preg_replace($valmatch, $valrepl, $attvalue);
1160                         if ($newvalue != $attvalue){
1161                             $attary{$attname} = $newvalue;
1162                         }
1163                     }
1164                 }
1165             }
1166         }
1167         /**
1168          * Turn cid: urls into http-friendly ones.
1169          */
1170         if (preg_match("/^[\'\"]\s*cid:/si", $attvalue)){
1171             $attary{$attname} = sq_cid2http($message, $id, $attvalue, $mailbox);
1172         }
1173     }
1174     /**
1175      * See if we need to append any attributes to this tag.
1176      */
1177     foreach ($add_attr_to_tag as $matchtag=>$addattary){
1178         if (preg_match($matchtag, $tagname)){
1179             $attary = array_merge($attary, $addattary);
1180         }
1181     }
1182     return $attary;
1183 }
1184
1185 /**
1186  * This function edits the style definition to make them friendly and
1187  * usable in squirrelmail.
1188  *
1189  * @param  $message  the message object
1190  * @param  $id       the message id
1191  * @param  $content  a string with whatever is between <style> and </style>
1192  * @return           a string with edited content.
1193  */
1194 function sq_fixstyle($message, $id, $content){
1195     global $view_unsafe_images;
1196     $me = 'sq_fixstyle';
1197     /**
1198      * First look for general BODY style declaration, which would be
1199      * like so:
1200      * body {background: blah-blah}
1201      * and change it to .bodyclass so we can just assign it to a <div>
1202      */
1203     $content = preg_replace("|body(\s*\{.*?\})|si", ".bodyclass\\1", $content);
1204     $secremoveimg = '../images/' . _("sec_remove_eng.png");
1205     /**
1206      * Fix url('blah') declarations.
1207      */
1208     $content = preg_replace("|url\(([\'\"])\s*\S+script\s*:.*?([\'\"])\)|si",
1209                             "url(\\1$secremoveimg\\2)", $content);
1210     /**
1211      * Fix url('https*://.*) declarations but only if $view_unsafe_images
1212      * is false.
1213      */
1214     if (!$view_unsafe_images){
1215         $content = preg_replace("|url\(([\'\"])\s*https*:.*?([\'\"])\)|si",
1216                                 "url(\\1$secremoveimg\\2)", $content);
1217     }
1218
1219     /**
1220      * Fix urls that refer to cid:
1221      */
1222     while (preg_match("|url\(([\'\"]\s*cid:.*?[\'\"])\)|si", $content,
1223                       $matches)){
1224         $cidurl = $matches{1};
1225         $httpurl = sq_cid2http($message, $id, $cidurl);
1226         $content = preg_replace("|url\($cidurl\)|si",
1227                                 "url($httpurl)", $content);
1228     }
1229
1230     /**
1231      * Fix stupid css declarations which lead to vulnerabilities
1232      * in IE.
1233      */
1234     $match   = Array('/expression/si',
1235                      '/behaviou*r/si',
1236                      '/binding/si');
1237     $replace = Array('idiocy', 'idiocy', 'idiocy');
1238     $content = preg_replace($match, $replace, $content);
1239     return $content;
1240 }
1241
1242 /**
1243  * This function converts cid: url's into the ones that can be viewed in
1244  * the browser.
1245  *
1246  * @param  $message  the message object
1247  * @param  $id       the message id
1248  * @param  $cidurl   the cid: url.
1249  * @return           a string with a http-friendly url
1250  */
1251 function sq_cid2http($message, $id, $cidurl, $mailbox){
1252     /**
1253      * Get rid of quotes.
1254      */
1255     $quotchar = substr($cidurl, 0, 1);
1256     $cidurl = str_replace($quotchar, "", $cidurl);
1257     $cidurl = substr(trim($cidurl), 4);
1258     $linkurl = find_ent_id($cidurl, $message);
1259     /* in case of non-save cid links $httpurl should be replaced by a sort of
1260        unsave link image */
1261     $httpurl = '';
1262     if ($linkurl) {
1263         $httpurl = $quotchar . '../src/download.php?absolute_dl=true&amp;' .
1264                    "passed_id=$id&amp;mailbox=" . urlencode($mailbox) .
1265                    '&amp;ent_id=' . $linkurl . $quotchar;
1266     }
1267     return $httpurl;
1268 }
1269
1270 /**
1271  * This function changes the <body> tag into a <div> tag since we
1272  * can't really have a body-within-body.
1273  *
1274  * @param  $attary  an array of attributes and values of <body>
1275  * @return          a modified array of attributes to be set for <div>
1276  */
1277 function sq_body2div($attary){
1278     $me = 'sq_body2div';
1279     $divattary = Array('class' => "'bodyclass'");
1280     $bgcolor = '#ffffff';
1281     $text = '#000000';
1282     $styledef = '';
1283     if (is_array($attary) && sizeof($attary) > 0){
1284         foreach ($attary as $attname=>$attvalue){
1285             $quotchar = substr($attvalue, 0, 1);
1286             $attvalue = str_replace($quotchar, "", $attvalue);
1287             switch ($attname){
1288                 case 'background':
1289                     $styledef .= "background-image: url('$attvalue'); ";
1290                     break;
1291                 case 'bgcolor':
1292                     $styledef .= "background-color: $attvalue; ";
1293                     break;
1294                 case 'text':
1295                     $styledef .= "color: $attvalue; ";
1296                     break;
1297             }
1298         }
1299         if (strlen($styledef) > 0){
1300             $divattary{"style"} = "\"$styledef\"";
1301         }
1302     }
1303     return $divattary;
1304 }
1305
1306 /**
1307  * This is the main function and the one you should actually be calling.
1308  * There are several variables you should be aware of an which need
1309  * special description.
1310  *
1311  * Since the description is quite lengthy, see it here:
1312  * http://www.mricon.com/html/phpfilter.html
1313  *
1314  * @param $body                 the string with HTML you wish to filter
1315  * @param $tag_list             see description above
1316  * @param $rm_tags_with_content see description above
1317  * @param $self_closing_tags    see description above
1318  * @param $force_tag_closing    see description above
1319  * @param $rm_attnames          see description above
1320  * @param $bad_attvals          see description above
1321  * @param $add_attr_to_tag      see description above
1322  * @param $message              message object
1323  * @param $id                   message id
1324  * @return                      sanitized html safe to show on your pages.
1325  */
1326 function sq_sanitize($body,
1327                      $tag_list,
1328                      $rm_tags_with_content,
1329                      $self_closing_tags,
1330                      $force_tag_closing,
1331                      $rm_attnames,
1332                      $bad_attvals,
1333                      $add_attr_to_tag,
1334                      $message,
1335                      $id,
1336                      $mailbox
1337                      ){
1338     $me = 'sq_sanitize';
1339     /**
1340      * Normalize rm_tags and rm_tags_with_content.
1341      */
1342     @array_walk($rm_tags, 'sq_casenormalize');
1343     @array_walk($rm_tags_with_content, 'sq_casenormalize');
1344     @array_walk($self_closing_tags, 'sq_casenormalize');
1345     /**
1346      * See if tag_list is of tags to remove or tags to allow.
1347      * false  means remove these tags
1348      * true   means allow these tags
1349      */
1350     $rm_tags = array_shift($tag_list);
1351     $curpos = 0;
1352     $open_tags = Array();
1353     $trusted = "<!-- begin sanitized html -->\n";
1354     $skip_content = false;
1355     /**
1356      * Take care of netscape's stupid javascript entities like
1357      * &{alert('boo')};
1358      */
1359     $body = preg_replace("/&(\{.*?\};)/si", "&amp;\\1", $body);
1360
1361     while (($curtag=sq_getnxtag($body, $curpos)) != FALSE){
1362         list($tagname, $attary, $tagtype, $lt, $gt) = $curtag;
1363         $free_content = substr($body, $curpos, $lt-$curpos);
1364         /**
1365          * Take care of <style>
1366          */
1367         if ($tagname == "style" && $tagtype == 2){
1368             /**
1369              * This is a closing </style>. Edit the
1370              * content before we apply it.
1371              */
1372             $free_content = sq_fixstyle($message, $id, $free_content);
1373         }
1374         if ($skip_content == false){
1375             $trusted .= $free_content;
1376         }
1377         if ($tagname != FALSE){
1378             if ($tagtype == 2){
1379                 if ($skip_content == $tagname){
1380                     /**
1381                      * Got to the end of tag we needed to remove.
1382                      */
1383                     $tagname = false;
1384                     $skip_content = false;
1385                 } else {
1386                     if ($skip_content == false){
1387                         if ($tagname == "body"){
1388                             $tagname = "div";
1389                         } else {
1390                             if (isset($open_tags{$tagname}) &&
1391                                 $open_tags{$tagname} > 0){
1392                                 $open_tags{$tagname}--;
1393                             } else {
1394                                 $tagname = false;
1395                             }
1396                         }
1397                     }
1398                 }
1399             } else {
1400                 /**
1401                  * $rm_tags_with_content
1402                  */
1403                 if ($skip_content == false){
1404                     /**
1405                      * See if this is a self-closing type and change
1406                      * tagtype appropriately.
1407                      */
1408                     if ($tagtype == 1
1409                         && in_array($tagname, $self_closing_tags)){
1410                         $tagtype=3;
1411                     }
1412                     /**
1413                      * See if we should skip this tag and any content
1414                      * inside it.
1415                      */
1416                     if ($tagtype == 1 &&
1417                         in_array($tagname, $rm_tags_with_content)){
1418                         $skip_content = $tagname;
1419                     } else {
1420                         if (($rm_tags == false
1421                              && in_array($tagname, $tag_list)) ||
1422                             ($rm_tags == true &&
1423                              !in_array($tagname, $tag_list))){
1424                             $tagname = false;
1425                         } else {
1426                             if ($tagtype == 1){
1427                                 if (isset($open_tags{$tagname})){
1428                                     $open_tags{$tagname}++;
1429                                 } else {
1430                                     $open_tags{$tagname}=1;
1431                                 }
1432                             }
1433                             /**
1434                              * This is where we run other checks.
1435                              */
1436                             if (is_array($attary) && sizeof($attary) > 0){
1437                                 $attary = sq_fixatts($tagname,
1438                                                      $attary,
1439                                                      $rm_attnames,
1440                                                      $bad_attvals,
1441                                                      $add_attr_to_tag,
1442                                                      $message,
1443                                                      $id,
1444                                                      $mailbox
1445                                                      );
1446                             }
1447                             /**
1448                              * Convert body into div.
1449                              */
1450                             if ($tagname == "body"){
1451                                 $tagname = "div";
1452                                 $attary = sq_body2div($attary, $message, $id);
1453                             }
1454                         }
1455                     }
1456                 }
1457             }
1458             if ($tagname != false && $skip_content == false){
1459                 $trusted .= sq_tagprint($tagname, $attary, $tagtype);
1460             }
1461         }
1462         $curpos = $gt+1;
1463     }
1464     $trusted .= substr($body, $curpos, strlen($body)-$curpos);
1465     if ($force_tag_closing == true){
1466         foreach ($open_tags as $tagname=>$opentimes){
1467             while ($opentimes > 0){
1468                 $trusted .= '</' . $tagname . '>';
1469                 $opentimes--;
1470             }
1471         }
1472         $trusted .= "\n";
1473     }
1474     $trusted .= "<!-- end sanitized html -->\n";
1475     return $trusted;
1476 }
1477
1478 /**
1479  * This is a wrapper function to call html sanitizing routines.
1480  *
1481  * @param  $body  the body of the message
1482  * @param  $id    the id of the message
1483  * @return        a string with html safe to display in the browser.
1484  */
1485 function magicHTML($body, $id, $message, $mailbox = 'INBOX'){
1486     global $attachment_common_show_images, $view_unsafe_images,
1487            $has_unsafe_images;
1488     /**
1489      * Don't display attached images in HTML mode.
1490      */
1491     $attachment_common_show_images = false;
1492     $tag_list = Array(
1493                       false,
1494                       "object",
1495                       "meta",
1496                       "html",
1497                       "head",
1498                       "base",
1499                       "link",
1500                       "frame",
1501                       "iframe"
1502                       );
1503
1504     $rm_tags_with_content = Array(
1505                                   "script",
1506                                   "applet",
1507                                   "embed",
1508                                   "title"
1509                                   );
1510
1511     $self_closing_tags =  Array(
1512                                 "img",
1513                                 "br",
1514                                 "hr",
1515                                 "input"
1516                                 );
1517
1518     $force_tag_closing = false;
1519
1520     $rm_attnames = Array(
1521                          "/.*/" =>
1522                          Array(
1523                                "/target/si",
1524                                "/^on.*/si",
1525                                "/^dynsrc/si",
1526                                "/^data.*/si"
1527                                )
1528                          );
1529
1530     $secremoveimg = "../images/" . _("sec_remove_eng.png");
1531     $bad_attvals = Array(
1532         "/.*/" =>
1533             Array(
1534                 "/^src|background/i" =>
1535                     Array(
1536                           Array(
1537                                 "|^([\'\"])\s*\.\./.*([\'\"])|si",
1538                                 "/^([\'\"])\s*\S+script\s*:.*([\'\"])/si",
1539                                 "/^([\'\"])\s*mocha\s*:*.*([\'\"])/si",
1540                                 "/^([\'\"])\s*about\s*:.*([\'\"])/si"
1541                                 ),
1542                           Array(
1543                                 "\\1$secremoveimg\\2",
1544                                 "\\1$secremoveimg\\2",
1545                                 "\\1$secremoveimg\\2",
1546                                 "\\1$secremoveimg\\2"
1547                                 )
1548                         ),
1549                 "/^href|action/i" =>
1550                     Array(
1551                           Array(
1552                                 "|^([\'\"])\s*\.\./.*([\'\"])|si",
1553                                 "/^([\'\"])\s*\S+script\s*:.*([\'\"])/si",
1554                                 "/^([\'\"])\s*mocha\s*:*.*([\'\"])/si",
1555                                 "/^([\'\"])\s*about\s*:.*([\'\"])/si"
1556                                 ),
1557                           Array(
1558                                 "\\1#\\2",
1559                                 "\\1#\\2",
1560                                 "\\1#\\2",
1561                                 "\\1#\\2"
1562                                 )
1563                         ),
1564                 "/^style/si" =>
1565                     Array(
1566                           Array(
1567                                 "/expression/si",
1568                                 "/binding/si",
1569                                 "/behaviou*r/si",
1570                                 "|url\(([\'\"])\s*\.\./.*([\'\"])\)|si",
1571                                 "/url\(([\'\"])\s*\S+script\s*:.*([\'\"])\)/si",
1572                                 "/url\(([\'\"])\s*mocha\s*:.*([\'\"])\)/si",
1573                                 "/url\(([\'\"])\s*about\s*:.*([\'\"])\)/si"
1574                                ),
1575                           Array(
1576                                 "idiocy",
1577                                 "idiocy",
1578                                 "idiocy",
1579                                 "url(\\1#\\2)",
1580                                 "url(\\1#\\2)",
1581                                 "url(\\1#\\2)",
1582                                 "url(\\1#\\2)"
1583                                )
1584                           )
1585                 )
1586         );
1587     if (!$view_unsafe_images){
1588         /**
1589          * Remove any references to http/https if view_unsafe_images set
1590          * to false.
1591          */
1592          array_push($bad_attvals{'/.*/'}{'/^src|background/i'}[0],
1593                     '/^([\'\"])\s*https*:.*([\'\"])/si');
1594          array_push($bad_attvals{'/.*/'}{'/^src|background/i'}[1],
1595                     "\\1$secremoveimg\\2");
1596          array_push($bad_attvals{'/.*/'}{'/^style/si'}[0],
1597                     '/url\(([\'\"])\s*https*:.*([\'\"])\)/si');
1598          array_push($bad_attvals{'/.*/'}{'/^style/si'}[1],
1599                     "url(\\1$secremoveimg\\2)");
1600     }
1601
1602     $add_attr_to_tag = Array(
1603                              "/^a$/si" => Array('target'=>'"_new"')
1604                              );
1605     $trusted = sq_sanitize($body,
1606                            $tag_list,
1607                            $rm_tags_with_content,
1608                            $self_closing_tags,
1609                            $force_tag_closing,
1610                            $rm_attnames,
1611                            $bad_attvals,
1612                            $add_attr_to_tag,
1613                            $message,
1614                            $id,
1615                            $mailbox
1616                            );
1617     if (preg_match("|$secremoveimg|si", $trusted)){
1618         $has_unsafe_images = true;
1619     }
1620     return $trusted;
1621 }
1622
1623 ?>