functions/mime.php

   1 <?php
   2
   3 /**
   4  * mime.php
   5  *
   6  * Copyright (c) 1999-2003 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     $i = 0;
  33     $msg = Message::parseStructure($read,$i);
  34     if (!is_object($msg)) {
  35         include_once(SM_PATH . 'functions/display_messages.php');
  36         global $color, $mailbox;
  37         /* removed urldecode because $_GET is auto urldecoded ??? */
  38         displayPageHeader( $color, $mailbox );
  39         $errormessage  = _("SquirrelMail could not decode the bodystructure of the message");
  40         $errormessage .= '<BR>'._("the provided bodystructure by your imap-server").':<BR><BR>';
  41         $errormessage .= '<pre>' . htmlspecialchars($read) . '</pre>';
  42         plain_error_message( $errormessage, $color );
  43         echo '</body></html>';
  44         exit;
  45     }
  46     if (count($flags)) {
  47         foreach ($flags as $flag) {
  48             $char = strtoupper($flag{1});
  49             switch ($char) {
  50                 case 'S':
  51                     if (strtolower($flag) == '\\seen') {
  52                         $msg->is_seen = true;
  53                     }
  54                     break;
  55                 case 'A':
  56                     if (strtolower($flag) == '\\answered') {
  57                         $msg->is_answered = true;
  58                     }
  59                     break;
  60                 case 'D':
  61                     if (strtolower($flag) == '\\deleted') {
  62                         $msg->is_deleted = true;
  63                     }
  64                     break;
  65                 case 'F':
  66                     if (strtolower($flag) == '\\flagged') {
  67                         $msg->is_flagged = true;
  68                     }
  69                     break;
  70                 case 'M':
  71                     if (strtolower($flag) == '$mdnsent') {
  72                         $msg->is_mdnsent = true;
  73                     }
  74                     break;
  75                 default:
  76                     break;
  77             }
  78         }
  79     }
  80     //    listEntities($msg);
  81     return $msg;
  82 }
  83
  84
  85
  86 /* This starts the parsing of a particular structure.  It is called recursively,
  87  * so it can be passed different structures.  It returns an object of type
  88  * $message.
  89  * First, it checks to see if it is a multipart message.  If it is, then it
  90  * handles that as it sees is necessary.  If it is just a regular entity,
  91  * then it parses it and adds the necessary header information (by calling out
  92  * to mime_get_elements()
  93  */
  94
  95 function mime_fetch_body($imap_stream, $id, $ent_id=1, $fetch_size=0) {
  96     global $uid_support;
  97     /* Do a bit of error correction.  If we couldn't find the entity id, just guess
  98      * that it is the first one.  That is usually the case anyway.
  99      */
 100
 101     if (!$ent_id) {
 102         $cmd = "FETCH $id BODY[]";
 103     } else {
 104         $cmd = "FETCH $id BODY[$ent_id]";
 105     }
 106
 107     if ($fetch_size!=0) $cmd .= "<0.$fetch_size>";
 108
 109     $data = sqimap_run_command ($imap_stream, $cmd, true, $response, $message, $uid_support);
 110     do {
 111         $topline = trim(array_shift($data));
 112     } while($topline && ($topline[0] == '*') && !preg_match('/\* [0-9]+ FETCH.*/i', $topline)) ;
 113
 114     $wholemessage = implode('', $data);
 115     if (ereg('\\{([^\\}]*)\\}', $topline, $regs)) {
 116         $ret = substr($wholemessage, 0, $regs[1]);
 117         /* There is some information in the content info header that could be important
 118          * in order to parse html messages. Let's get them here.
 119          */
 120 //        if ($ret{0} == '<') {
 121 //            $data = sqimap_run_command ($imap_stream, "FETCH $id BODY[$ent_id.MIME]", true, $response, $message, $uid_support);
 122 //        }
 123     } else if (ereg('"([^"]*)"', $topline, $regs)) {
 124         $ret = $regs[1];
 125     } else {
 126         global $where, $what, $mailbox, $passed_id, $startMessage;
 127         $par = 'mailbox=' . urlencode($mailbox) . '&amp;passed_id=' . $passed_id;
 128         if (isset($where) && isset($what)) {
 129             $par .= '&amp;where=' . urlencode($where) . '&amp;what=' . urlencode($what);
 130         } else {
 131             $par .= '&amp;startMessage=' . $startMessage . '&amp;show_more=0';
 132         }
 133         $par .= '&amp;response=' . urlencode($response) .
 134                 '&amp;message='  . urlencode($message)  .
 135                 '&amp;topline='  . urlencode($topline);
 136
 137         echo   '<tt><br>' .
 138                '<table width="80%"><tr>' .
 139                '<tr><td colspan=2>' .
 140                _("Body retrieval error. The reason for this is most probably that the message is malformed.") .
 141                '</td></tr>' .
 142                '<tr><td><b>' . _("Command:") . "</td><td>$cmd</td></tr>" .
 143                '<tr><td><b>' . _("Response:") . "</td><td>$response</td></tr>" .
 144                '<tr><td><b>' . _("Message:") . "</td><td>$message</td></tr>" .
 145                '<tr><td><b>' . _("FETCH line:") . "</td><td>$topline</td></tr>" .
 146                "</table><BR></tt></font><hr>";
 147
 148         $data = sqimap_run_command ($imap_stream, "FETCH $passed_id BODY[]", true, $response, $message, $uid_support);
 149         array_shift($data);
 150         $wholemessage = implode('', $data);
 151
 152         $ret = $wholemessage;
 153     }
 154     return $ret;
 155 }
 156
 157 function mime_print_body_lines ($imap_stream, $id, $ent_id=1, $encoding) {
 158     global $uid_support;
 159
 160     /* Don't kill the connection if the browser is over a dialup
 161      * and it would take over 30 seconds to download it.
 162      * Don´t call set_time_limit in safe mode.
 163      */
 164
 165     if (!ini_get('safe_mode')) {
 166         set_time_limit(0);
 167     }
 168     /* in case of base64 encoded attachments, do not buffer them.
 169        Instead, echo the decoded attachment directly to screen */
 170     if (strtolower($encoding) == 'base64') {
 171         if (!$ent_id) {
 172            $query = "FETCH $id BODY[]";
 173         } else {
 174            $query = "FETCH $id BODY[$ent_id]";
 175         }
 176         sqimap_run_command($imap_stream,$query,true,$response,$message,$uid_support,'sqimap_base64_decode','php://stdout',true);
 177     } else {
 178        $body = mime_fetch_body ($imap_stream, $id, $ent_id);
 179        echo decodeBody($body, $encoding);
 180     }
 181
 182     /*
 183        TODO, use the same method for quoted printable.
 184        However, I assume that quoted printable attachments aren't that large
 185        so the performancegain / memory usage drop will be minimal.
 186        If we decide to add that then we need to adapt sqimap_fread because
 187        we need to split te result on \n and fread doesn't stop at \n. That
 188        means we also should provide $results from sqimap_fread (by ref) to
 189        te function and set $no_return to false. The $filter function for
 190        quoted printable should handle unsetting of $results.
 191     */
 192     /*
 193        TODO 2: find out how we write to the output stream php://stdout. fwrite
 194        doesn't work because 'php://stdout isn't a stream.
 195     */
 196
 197     return;
 198 /*
 199     fputs ($imap_stream, "$sid FETCH $id BODY[$ent_id]\r\n");
 200     $cnt = 0;
 201     $continue = true;
 202     $read = fgets ($imap_stream,8192);
 203
 204
 205     // This could be bad -- if the section has sqimap_session_id() . ' OK'
 206     // or similar, it will kill the download.
 207     while (!ereg("^".$sid_s." (OK|BAD|NO)(.*)$", $read, $regs)) {
 208         if (trim($read) == ')==') {
 209             $read1 = $read;
 210             $read = fgets ($imap_stream,4096);
 211             if (ereg("^".$sid." (OK|BAD|NO)(.*)$", $read, $regs)) {
 212                 return;
 213             } else {
 214                 echo decodeBody($read1, $encoding) .
 215                      decodeBody($read, $encoding);
 216             }
 217         } else if ($cnt) {
 218             echo decodeBody($read, $encoding);
 219         }
 220         $read = fgets ($imap_stream,4096);
 221         $cnt++;
 222 //      break;
 223     }
 224 */
 225 }
 226
 227 /* -[ END MIME DECODING ]----------------------------------------------------------- */
 228
 229 /* This is here for debugging purposes.  It will print out a list
 230  * of all the entity IDs that are in the $message object.
 231  */
 232 function listEntities ($message) {
 233     if ($message) {
 234         echo "<tt>" . $message->entity_id . ' : ' . $message->type0 . '/' . $message->type1 . ' parent = '. $message->parent->entity_id. '<br>';
 235         for ($i = 0; isset($message->entities[$i]); $i++) {
 236             echo "$i : ";
 237             $msg = listEntities($message->entities[$i]);
 238
 239             if ($msg) {
 240                 echo "return: ";
 241                 return $msg;
 242             }
 243         }
 244     }
 245 }
 246
 247 function getPriorityStr($priority) {
 248     $priority_level = substr($priority,0,1);
 249
 250     switch($priority_level) {
 251         /* Check for a higher then normal priority. */
 252         case '1':
 253         case '2':
 254             $priority_string = _("High");
 255             break;
 256
 257         /* Check for a lower then normal priority. */
 258         case '4':
 259         case '5':
 260             $priority_string = _("Low");
 261             break;
 262
 263         /* Check for a normal priority. */
 264         case '3':
 265         default:
 266             $priority_level = '3';
 267             $priority_string = _("Normal");
 268             break;
 269
 270     }
 271     return $priority_string;
 272 }
 273
 274 /* returns a $message object for a particular entity id */
 275 function getEntity ($message, $ent_id) {
 276     return $message->getEntity($ent_id);
 277 }
 278
 279 /* translateText
 280  * Extracted from strings.php 23/03/2002
 281  */
 282
 283 function translateText(&$body, $wrap_at, $charset) {
 284     global $where, $what;   /* from searching */
 285     global $color;          /* color theme */
 286
 287     require_once(SM_PATH . 'functions/url_parser.php');
 288
 289     $body_ary = explode("\n", $body);
 290     for ($i=0; $i < count($body_ary); $i++) {
 291         $line = $body_ary[$i];
 292         if (strlen($line) - 2 >= $wrap_at) {
 293             sqWordWrap($line, $wrap_at);
 294         }
 295         $line = charset_decode($charset, $line);
 296         $line = str_replace("\t", '        ', $line);
 297
 298         parseUrl ($line);
 299
 300         $quotes = 0;
 301         $pos = 0;
 302         $j = strlen($line);
 303
 304         while ($pos < $j) {
 305             if ($line[$pos] == ' ') {
 306                 $pos++;
 307             } else if (strpos($line, '&gt;', $pos) === $pos) {
 308                 $pos += 4;
 309                 $quotes++;
 310             } else {
 311                 break;
 312             }
 313         }
 314
 315         if ($quotes % 2) {
 316             if (!isset($color[13])) {
 317                 $color[13] = '#800000';
 318             }
 319             $line = '<font color="' . $color[13] . '">' . $line . '</font>';
 320         } elseif ($quotes) {
 321             if (!isset($color[14])) {
 322                 $color[14] = '#FF0000';
 323             }
 324             $line = '<font color="' . $color[14] . '">' . $line . '</font>';
 325         }
 326
 327         $body_ary[$i] = $line;
 328     }
 329     $body = '<pre>' . implode("\n", $body_ary) . '</pre>';
 330 }
 331
 332 /* This returns a parsed string called $body. That string can then
 333  * be displayed as the actual message in the HTML. It contains
 334  * everything needed, including HTML Tags, Attachments at the
 335  * bottom, etc.
 336  */
 337 function formatBody($imap_stream, $message, $color, $wrap_at, $ent_num, $id, $mailbox='INBOX') {
 338     /* This if statement checks for the entity to show as the
 339      * primary message. To add more of them, just put them in the
 340      * order that is their priority.
 341      */
 342     global $startMessage, $username, $key, $imapServerAddress, $imapPort,
 343            $show_html_default, $sort, $has_unsafe_images, $passed_ent_id;
 344     global $languages, $squirrelmail_language;
 345
 346     if( !sqgetGlobalVar('view_unsafe_images', $view_unsafe_images, SQ_GET) ) {
 347         $view_unsafe_images = false;
 348     }
 349
 350     $body = '';
 351     $urlmailbox = urlencode($mailbox);
 352     $body_message = getEntity($message, $ent_num);
 353     if (($body_message->header->type0 == 'text') ||
 354         ($body_message->header->type0 == 'rfc822')) {
 355         $body = mime_fetch_body ($imap_stream, $id, $ent_num);
 356         $body = decodeBody($body, $body_message->header->encoding);
 357
 358         if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
 359             function_exists($languages[$squirrelmail_language]['XTRA_CODE'])) {
 360             if (mb_detect_encoding($body) != 'ASCII') {
 361                 $body = $languages[$squirrelmail_language]['XTRA_CODE']('decode', $body);
 362             }
 363         }
 364         $hookResults = do_hook("message_body", $body);
 365         $body = $hookResults[1];
 366
 367         /* If there are other types that shouldn't be formatted, add
 368          * them here.
 369          */
 370
 371         if ($body_message->header->type1 == 'html') {
 372             if ($show_html_default <> 1) {
 373                 $entity_conv = array('&nbsp;' => ' ',
 374                                      '<p>'    => "\n",
 375                                      '<br>'   => "\n",
 376                                      '<P>'    => "\n",
 377                                      '<BR>'   => "\n",
 378                                      '&gt;'   => '>',
 379                                      '&lt;'   => '<');
 380                 $body = strtr($body, $entity_conv);
 381                 $body = strip_tags($body);
 382                 $body = trim($body);
 383                 translateText($body, $wrap_at,
 384                               $body_message->header->getParameter('charset'));
 385             } else {
 386                 $body = magicHTML($body, $id, $message, $mailbox);
 387             }
 388         } else {
 389             translateText($body, $wrap_at,
 390                           $body_message->header->getParameter('charset'));
 391         }
 392         $link = 'read_body.php?passed_id=' . $id . '&amp;ent_id='.$ent_num.
 393                 '&amp;mailbox=' . $urlmailbox .'&amp;sort=' . $sort .
 394                 '&amp;startMessage=' . $startMessage . '&amp;show_more=0';
 395         if (isset($passed_ent_id)) {
 396             $link .= '&amp;passed_ent_id='.$passed_ent_id;
 397         }
 398         if ($view_unsafe_images) {
 399             $text = _("Hide Unsafe Images");
 400         } else {
 401             if (isset($has_unsafe_images) && $has_unsafe_images) {
 402                 $link .= '&amp;view_unsafe_images=1';
 403                 $text = _("View Unsafe Images");
 404             } else {
 405                 $text = '';
 406             }
 407         }
 408         $body .= '<center><small><a href="'.$link.'">'.$text.
 409                  '</a></small></center><br>' . "\n";
 410     }
 411     return $body;
 412 }
 413
 414
 415 function formatAttachments($message, $exclude_id, $mailbox, $id) {
 416     global $where, $what, $startMessage, $color;
 417     static $ShownHTML = 0;
 418
 419     $att_ar = $message->getAttachments($exclude_id);
 420
 421     if (!count($att_ar)) return '';
 422
 423     $attachments = '';
 424
 425     $urlMailbox = urlencode($mailbox);
 426
 427     foreach ($att_ar as $att) {
 428         $ent = $att->entity_id;
 429         $header = $att->header;
 430         $type0 = strtolower($header->type0);
 431         $type1 = strtolower($header->type1);
 432         $name = '';
 433         $links['download link']['text'] = _("download");
 434         $links['download link']['href'] = SM_PATH .
 435                 "src/download.php?absolute_dl=true&amp;passed_id=$id&amp;mailbox=$urlMailbox&amp;ent_id=$ent";
 436         $ImageURL = '';
 437         if ($type0 =='message' && $type1 == 'rfc822') {
 438             $default_page = SM_PATH . 'src/read_body.php';
 439             $rfc822_header = $att->rfc822_header;
 440             $filename = $rfc822_header->subject;
 441             if (trim( $filename ) == '') {
 442                 $filename = 'untitled-[' . $ent . ']' ;
 443             }
 444             $from_o = $rfc822_header->from;
 445             if (is_object($from_o)) {
 446                 $from_name = decodeHeader($from_o->getAddress(false));
 447             } else {
 448                 $from_name = _("Unknown sender");
 449             }
 450             $description = $from_name;
 451         } else {
 452             $default_page = SM_PATH . 'src/download.php';
 453             if (is_object($header->disposition)) {
 454                 $filename = $header->disposition->getProperty('filename');
 455                 if (trim($filename) == '') {
 456                     $name = decodeHeader($header->disposition->getProperty('name'));
 457                     if (trim($name) == '') {
 458                         $name = $header->getParameter('name');
 459                         if(trim($name) == '') {
 460                             if (trim( $header->id ) == '') {
 461                                 $filename = 'untitled-[' . $ent . ']' ;
 462                             } else {
 463                                 $filename = 'cid: ' . $header->id;
 464                             }
 465                         } else {
 466                             $filename = $name;
 467                         }
 468                     } else {
 469                         $filename = $name;
 470                     }
 471                 }
 472             } else {
 473                 $filename = $header->getParameter('name');
 474                 if (!trim($filename)) {
 475                     if (trim( $header->id ) == '') {
 476                         $filename = 'untitled-[' . $ent . ']' ;
 477                     } else {
 478                         $filename = 'cid: ' . $header->id;
 479                     }
 480                 }
 481             }
 482             if ($header->description) {
 483                 $description = decodeHeader($header->description);
 484             } else {
 485                 $description = '';
 486             }
 487         }
 488
 489         $display_filename = $filename;
 490         if (isset($passed_ent_id)) {
 491             $passed_ent_id_link = '&amp;passed_ent_id='.$passed_ent_id;
 492         } else {
 493             $passed_ent_id_link = '';
 494         }
 495         $defaultlink = $default_page . "?startMessage=$startMessage"
 496                      . "&amp;passed_id=$id&amp;mailbox=$urlMailbox"
 497                      . '&amp;ent_id='.$ent.$passed_ent_id_link;
 498         if ($where && $what) {
 499            $defaultlink .= '&amp;where='. urlencode($where).'&amp;what='.urlencode($what);
 500         }
 501         /* This executes the attachment hook with a specific MIME-type.
 502          * If that doesn't have results, it tries if there's a rule
 503          * for a more generic type.
 504          */
 505         $hookresults = do_hook("attachment $type0/$type1", $links,
 506                                $startMessage, $id, $urlMailbox, $ent, $defaultlink,
 507                                $display_filename, $where, $what);
 508         if(count($hookresults[1]) <= 1) {
 509             $hookresults = do_hook("attachment $type0/*", $links,
 510                                    $startMessage, $id, $urlMailbox, $ent, $defaultlink,
 511                                    $display_filename, $where, $what);
 512         }
 513
 514         $links = $hookresults[1];
 515         $defaultlink = $hookresults[6];
 516
 517         $attachments .= '<TR><TD>' .
 518                         '<A HREF="'.$defaultlink.'">'.decodeHeader($display_filename).'</A>&nbsp;</TD>' .
 519                         '<TD><SMALL><b>' . show_readable_size($header->size) .
 520                         '</b>&nbsp;&nbsp;</small></TD>' .
 521                         "<TD><SMALL>[ $type0/$type1 ]&nbsp;</SMALL></TD>" .
 522                         '<TD><SMALL>';
 523         $attachments .= '<b>' . $description . '</b>';
 524         $attachments .= '</SMALL></TD><TD><SMALL>&nbsp;';
 525
 526         $skipspaces = 1;
 527         foreach ($links as $val) {
 528             if ($skipspaces) {
 529                 $skipspaces = 0;
 530             } else {
 531                 $attachments .= '&nbsp;&nbsp;|&nbsp;&nbsp;';
 532             }
 533             $attachments .= '<a href="' . $val['href'] . '">' .  $val['text'] . '</a>';
 534         }
 535         unset($links);
 536         $attachments .= "</TD></TR>\n";
 537     }
 538     $attachmentadd = do_hook_function('attachments_bottom',$attachments);
 539     if ($attachmentadd != '')
 540         $attachments = $attachmentadd;
 541     return $attachments;
 542 }
 543
 544 function sqimap_base64_decode(&$string) {
 545     $string = str_replace("\r\n", "\n", $string);
 546     $string = base64_decode($string);
 547 }
 548
 549 /* This function decodes the body depending on the encoding type. */
 550 function decodeBody($body, $encoding) {
 551     global $show_html_default;
 552
 553     $body = str_replace("\r\n", "\n", $body);
 554     $encoding = strtolower($encoding);
 555
 556     $encoding_handler = do_hook_function('decode_body', $encoding);
 557
 558
 559     // plugins get first shot at decoding the body
 560     //
 561     if (!empty($encoding_handler) && function_exists($encoding_handler)) {
 562         $body = $encoding_handler('decode', $body);
 563
 564     } else if ($encoding == 'quoted-printable' ||
 565         $encoding == 'quoted_printable') {
 566         $body = quoted_printable_decode($body);
 567
 568         while (ereg("=\n", $body)) {
 569             $body = ereg_replace ("=\n", '', $body);
 570         }
 571
 572     } else if ($encoding == 'base64') {
 573         $body = base64_decode($body);
 574     }
 575
 576     // All other encodings are returned raw.
 577     return $body;
 578 }
 579
 580 /*
 581  * This functions decode strings that is encoded according to
 582  * RFC1522 (MIME Part Two: Message Header Extensions for Non-ASCII Text).
 583  * Patched by Christian Schmidt <christian@ostenfeld.dk>  23/03/2002
 584  */
 585 function decodeHeader ($string, $utfencode=true,$htmlsave=true) {
 586     global $languages, $squirrelmail_language;
 587     if (is_array($string)) {
 588         $string = implode("\n", $string);
 589     }
 590
 591     if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
 592         function_exists($languages[$squirrelmail_language]['XTRA_CODE'])) {
 593         $string = $languages[$squirrelmail_language]['XTRA_CODE']('decodeheader', $string);
 594         // Do we need to return at this point?
 595         // return $string;
 596     }
 597     $i = 0;
 598     $iLastMatch = -2;
 599     $encoded = true;
 600
 601     $aString = explode(' ',$string);
 602     $ret = '';
 603     foreach ($aString as $chunk) {
 604         if ($encoded && $chunk === '') {
 605             continue;
 606         } elseif ($chunk === '') {
 607             $ret .= ' ';
 608             continue;
 609         }
 610         $encoded = false;
 611         /* if encoded words are not separated by a linear-space-white we still catch them */
 612         $j = $i-1;
 613 //      if ($chunk{0} === '=') { /* performance, saves an unnessecarry preg call */
 614         while ($match = preg_match('/^(.*)=\?([^?]*)\?(Q|B)\?([^?]*)\?=(.*)$/Ui',$chunk,$res)) {
 615             /* if the last chunk isn't an encoded string then put back the space, otherwise don't */
 616             if ($iLastMatch !== $j) {
 617                 if ($htmlsave) {
 618                     $ret .= '&nbsp;';
 619                 } else {
 620                     $ret .= ' ';
 621                 }
 622             }
 623             $iLastMatch = $i;
 624             $j = $i;
 625             $ret .= $res[1];
 626             $encoding = ucfirst($res[3]);
 627             switch ($encoding)
 628             {
 629             case 'B':
 630                 $replace = base64_decode($res[4]);
 631                 $ret .= charset_decode($res[2],$replace);
 632                 break;
 633             case 'Q':
 634                 $replace = str_replace('_', ' ', $res[4]);
 635                 $replace = preg_replace('/=([0-9a-f]{2})/ie', 'chr(hexdec("\1"))',
 636                                     $replace);
 637                 /* Only encode into entities by default. Some places
 638                  * don't need the encoding, like the compose form.
 639                  */
 640                 if ($utfencode) {
 641                     $replace = charset_decode($res[2], $replace);
 642                 } else {
 643                     if ($htmlsave) {
 644                         $replace = htmlspecialchars($replace);
 645                     }
 646                 }
 647                 $ret .= $replace;
 648                 break;
 649             default:
 650                 break;
 651             }
 652             $chunk = $res[5];
 653             $encoded = true;
 654         }
 655 //      }
 656         if (!$encoded) {
 657             if ($htmlsave) {
 658                 $ret .= '&nbsp;';
 659             } else {
 660                 $ret .= ' ';
 661             }
 662         }
 663
 664         if (!$encoded && $htmlsave) {
 665             $ret .= htmlspecialchars($chunk);
 666         } else {
 667             $ret .= $chunk;
 668         }
 669         ++$i;
 670     }
 671     /* remove the first added space */
 672     if ($ret) {
 673         if ($htmlsave) {
 674             $ret = substr($ret,6);
 675         } else {
 676             $ret = substr($ret,1);
 677         }
 678     }
 679
 680     return $ret;
 681 }
 682
 683 /*
 684  * Encode a string according to RFC 1522 for use in headers if it
 685  * contains 8-bit characters or anything that looks like it should
 686  * be encoded.
 687  */
 688 function encodeHeader ($string) {
 689     global $default_charset, $languages, $squirrelmail_language;
 690
 691     if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
 692         function_exists($languages[$squirrelmail_language]['XTRA_CODE'])) {
 693         return  $languages[$squirrelmail_language]['XTRA_CODE']('encodeheader', $string);
 694     }
 695     if (strtolower($default_charset) == 'iso-8859-1') {
 696         $string = str_replace("\240",' ',$string);
 697     }
 698
 699     // Encode only if the string contains 8-bit characters or =?
 700     $j = strlen($string);
 701     $max_l = 75 - strlen($default_charset) - 7;
 702     $aRet = array();
 703     $ret = '';
 704     $iEncStart = $enc_init = false;
 705     $cur_l = $iOffset = 0;
 706     for($i = 0; $i < $j; ++$i) {
 707         switch($string{$i})
 708         {
 709         case '=':
 710         case '<':
 711         case '>':
 712         case ',':
 713         case '?':
 714         case '_':
 715             if ($iEncStart === false) {
 716                 $iEncStart = $i;
 717             }
 718             $cur_l+=3;
 719             if ($cur_l > ($max_l-2)) {
 720                 /* if there is an stringpart that doesn't need encoding, add it */
 721                 $aRet[] = substr($string,$iOffset,$iEncStart-$iOffset);
 722                 $aRet[] = "=?$default_charset?Q?$ret?=";
 723                 $iOffset = $i;
 724                 $cur_l = 0;
 725                 $ret = '';
 726                 $iEncStart = false;
 727             } else {
 728                 $ret .= sprintf("=%02X",ord($string{$i}));
 729             }
 730             break;
 731         case '(':
 732         case ')':
 733             if ($iEncStart !== false) {
 734                 $aRet[] = substr($string,$iOffset,$iEncStart-$iOffset);
 735                 $aRet[] = "=?$default_charset?Q?$ret?=";
 736                 $iOffset = $i;
 737                 $cur_l = 0;
 738                 $ret = '';
 739                 $iEncStart = false;
 740             }
 741             break;
 742         case ' ':
 743             if ($iEncStart !== false) {
 744                 $cur_l++;
 745                 if ($cur_l > $max_l) {
 746                     $aRet[] = substr($string,$iOffset,$iEncStart-$iOffset);
 747                     $aRet[] = "=?$default_charset?Q?$ret?=";
 748                     $iOffset = $i;
 749                     $cur_l = 0;
 750                     $ret = '';
 751                     $iEncStart = false;
 752                 } else {
 753                     $ret .= '_';
 754                 }
 755             }
 756             break;
 757         default:
 758             $k = ord($string{$i});
 759             if ($k > 126) {
 760                 if ($iEncStart === false) {
 761                     // do not start encoding in the middle of a string, also take the rest of the word.
 762                     $sLeadString = substr($string,0,$i);
 763                     $aLeadString = explode(' ',$sLeadString);
 764                     $sToBeEncoded = array_pop($aLeadString);
 765                     $iEncStart = $i - strlen($sToBeEncoded);
 766                     $ret .= $sToBeEncoded;
 767                     $cur_l += strlen($sToBeEncoded);
 768                 }
 769                 $cur_l += 3;
 770                 /* first we add the encoded string that reached it's max size */
 771                 if ($cur_l > ($max_l-2)) {
 772                     $aRet[] = substr($string,$iOffset,$iEncStart-$iOffset);
 773                     $aRet[] = "=?$default_charset?Q?$ret?= "; /* the next part is also encoded => separate by space */
 774                     $cur_l = 3;
 775                     $ret = '';
 776                     $iOffset = $i;
 777                     $iEncStart = $i;
 778                 }
 779                 $enc_init = true;
 780                 $ret .= sprintf("=%02X", $k);
 781             } else {
 782                 if ($iEncStart !== false) {
 783                     $cur_l++;
 784                     if ($cur_l > $max_l) {
 785                         $aRet[] = substr($string,$iOffset,$iEncStart-$iOffset);
 786                         $aRet[] = "=?$default_charset?Q?$ret?=";
 787                         $iEncStart = false;
 788                         $iOffset = $i;
 789                         $cur_l = 0;
 790                         $ret = '';
 791                     } else {
 792                         $ret .= $string{$i};
 793                     }
 794                 }
 795             }
 796             break;
 797         }
 798     }
 799
 800     if ($enc_init) {
 801         if ($iEncStart !== false) {
 802             $aRet[] = substr($string,$iOffset,$iEncStart-$iOffset);
 803             $aRet[] = "=?$default_charset?Q?$ret?=";
 804         } else {
 805             $aRet[] = substr($string,$iOffset);
 806         }
 807         $string = implode('',$aRet);
 808     }
 809     return $string;
 810 }
 811
 812 /* This function trys to locate the entity_id of a specific mime element */
 813 function find_ent_id($id, $message) {
 814     for ($i = 0, $ret = ''; $ret == '' && $i < count($message->entities); $i++) {
 815         if ($message->entities[$i]->header->type0 == 'multipart')  {
 816             $ret = find_ent_id($id, $message->entities[$i]);
 817         } else {
 818             if (strcasecmp($message->entities[$i]->header->id, $id) == 0) {
 819 //                if (sq_check_save_extension($message->entities[$i])) {
 820                     return $message->entities[$i]->entity_id;
 821 //                }
 822             }
 823         }
 824     }
 825     return $ret;
 826 }
 827
 828 function sq_check_save_extension($message) {
 829     $filename = $message->getFilename();
 830     $ext = substr($filename, strrpos($filename,'.')+1);
 831     $save_extensions = array('jpg','jpeg','gif','png','bmp');
 832     return in_array($ext, $save_extensions);
 833 }
 834
 835
 836 /**
 837  ** HTMLFILTER ROUTINES
 838  */
 839
 840 /**
 841  * This function is more or less a wrapper around stripslashes. Apparently
 842  * Explorer is stupid enough to just remove the backslashes and then
 843  * execute the content of the attribute as if nothing happened.
 844  * Who does that?
 845  *
 846  * @param  attvalue   The value of the attribute
 847  * @return attvalue   The value of the attribute stripslashed.
 848  */
 849 function sq_unbackslash($attvalue){
 850     /**
 851      * Remove any backslashes. See if there are any first.
 852      */
 853     if (strstr($attvalue, '\\') !== false){
 854         $attvalue = stripslashes($attvalue);
 855     }
 856     return $attvalue;
 857 }
 858
 859 /**
 860  * Kill any tabs, newlines, or carriage returns. Our friends the
 861  * makers of the browser with 95% market value decided that it'd
 862  * be funny to make "java[tab]script" be just as good as "javascript".
 863  *
 864  * @param  attvalue  The attribute value before extraneous spaces removed.
 865  * @return attvalue  The attribute value after extraneous spaces removed.
 866  */
 867 function sq_unspace($attvalue){
 868     if (strcspn($attvalue, "\t\r\n") != strlen($attvalue)){
 869         $attvalue = str_replace(Array("\t", "\r", "\n"), Array('', '', ''),
 870                                 $attvalue);
 871     }
 872     return $attvalue;
 873 }
 874
 875 /**
 876  * This function returns the final tag out of the tag name, an array
 877  * of attributes, and the type of the tag. This function is called by
 878  * sq_sanitize internally.
 879  *
 880  * @param  $tagname  the name of the tag.
 881  * @param  $attary   the array of attributes and their values
 882  * @param  $tagtype  The type of the tag (see in comments).
 883  * @return           a string with the final tag representation.
 884  */
 885 function sq_tagprint($tagname, $attary, $tagtype){
 886     $me = 'sq_tagprint';
 887
 888     if ($tagtype == 2){
 889         $fulltag = '</' . $tagname . '>';
 890     } else {
 891         $fulltag = '<' . $tagname;
 892         if (is_array($attary) && sizeof($attary)){
 893             $atts = Array();
 894             while (list($attname, $attvalue) = each($attary)){
 895                 array_push($atts, "$attname=$attvalue");
 896             }
 897             $fulltag .= ' ' . join(" ", $atts);
 898         }
 899         if ($tagtype == 3){
 900             $fulltag .= ' /';
 901         }
 902         $fulltag .= '>';
 903     }
 904     return $fulltag;
 905 }
 906
 907 /**
 908  * A small helper function to use with array_walk. Modifies a by-ref
 909  * value and makes it lowercase.
 910  *
 911  * @param  $val a value passed by-ref.
 912  * @return      void since it modifies a by-ref value.
 913  */
 914 function sq_casenormalize(&$val){
 915     $val = strtolower($val);
 916 }
 917
 918 /**
 919  * This function skips any whitespace from the current position within
 920  * a string and to the next non-whitespace value.
 921  *
 922  * @param  $body   the string
 923  * @param  $offset the offset within the string where we should start
 924  *                 looking for the next non-whitespace character.
 925  * @return         the location within the $body where the next
 926  *                 non-whitespace char is located.
 927  */
 928 function sq_skipspace($body, $offset){
 929     $me = 'sq_skipspace';
 930     preg_match('/^(\s*)/s', substr($body, $offset), $matches);
 931     if (sizeof($matches{1})){
 932         $count = strlen($matches{1});
 933         $offset += $count;
 934     }
 935     return $offset;
 936 }
 937
 938 /**
 939  * This function looks for the next character within a string.  It's
 940  * really just a glorified "strpos", except it catches if failures
 941  * nicely.
 942  *
 943  * @param  $body   The string to look for needle in.
 944  * @param  $offset Start looking from this position.
 945  * @param  $needle The character/string to look for.
 946  * @return         location of the next occurance of the needle, or
 947  *                 strlen($body) if needle wasn't found.
 948  */
 949 function sq_findnxstr($body, $offset, $needle){
 950     $me  = 'sq_findnxstr';
 951     $pos = strpos($body, $needle, $offset);
 952     if ($pos === FALSE){
 953         $pos = strlen($body);
 954     }
 955     return $pos;
 956 }
 957
 958 /**
 959  * This function takes a PCRE-style regexp and tries to match it
 960  * within the string.
 961  *
 962  * @param  $body   The string to look for needle in.
 963  * @param  $offset Start looking from here.
 964  * @param  $reg    A PCRE-style regex to match.
 965  * @return         Returns a false if no matches found, or an array
 966  *                 with the following members:
 967  *                 - integer with the location of the match within $body
 968  *                 - string with whatever content between offset and the match
 969  *                 - string with whatever it is we matched
 970  */
 971 function sq_findnxreg($body, $offset, $reg){
 972     $me = 'sq_findnxreg';
 973     $matches = Array();
 974     $retarr = Array();
 975     preg_match("%^(.*?)($reg)%si", substr($body, $offset), $matches);
 976     if (!isset($matches{0}) || !$matches{0}){
 977         $retarr = false;
 978     } else {
 979         $retarr{0} = $offset + strlen($matches{1});
 980         $retarr{1} = $matches{1};
 981         $retarr{2} = $matches{2};
 982     }
 983     return $retarr;
 984 }
 985
 986 /**
 987  * This function looks for the next tag.
 988  *
 989  * @param  $body   String where to look for the next tag.
 990  * @param  $offset Start looking from here.
 991  * @return         false if no more tags exist in the body, or
 992  *                 an array with the following members:
 993  *                 - string with the name of the tag
 994  *                 - array with attributes and their values
 995  *                 - integer with tag type (1, 2, or 3)
 996  *                 - integer where the tag starts (starting "<")
 997  *                 - integer where the tag ends (ending ">")
 998  *                 first three members will be false, if the tag is invalid.
 999  */
1000 function sq_getnxtag($body, $offset){
1001     $me = 'sq_getnxtag';
1002     if ($offset > strlen($body)){
1003         return false;
1004     }
1005     $lt = sq_findnxstr($body, $offset, "<");
1006     if ($lt == strlen($body)){
1007         return false;
1008     }
1009     /**
1010      * We are here:
1011      * blah blah <tag attribute="value">
1012      * \---------^
1013      */
1014     $pos = sq_skipspace($body, $lt+1);
1015     if ($pos >= strlen($body)){
1016         return Array(false, false, false, $lt, strlen($body));
1017     }
1018     /**
1019      * There are 3 kinds of tags:
1020      * 1. Opening tag, e.g.:
1021      *    <a href="blah">
1022      * 2. Closing tag, e.g.:
1023      *    </a>
1024      * 3. XHTML-style content-less tag, e.g.:
1025      *    <img src="blah"/>
1026      */
1027     $tagtype = false;
1028     switch (substr($body, $pos, 1)){
1029         case '/':
1030             $tagtype = 2;
1031             $pos++;
1032             break;
1033         case '!':
1034             /**
1035              * A comment or an SGML declaration.
1036              */
1037             if (substr($body, $pos+1, 2) == "--"){
1038                 $gt = strpos($body, "-->", $pos);
1039                 if ($gt === false){
1040                     $gt = strlen($body);
1041                 } else {
1042                     $gt += 2;
1043                 }
1044                 return Array(false, false, false, $lt, $gt);
1045             } else {
1046                 $gt = sq_findnxstr($body, $pos, ">");
1047                 return Array(false, false, false, $lt, $gt);
1048             }
1049             break;
1050         default:
1051             /**
1052              * Assume tagtype 1 for now. If it's type 3, we'll switch values
1053              * later.
1054              */
1055             $tagtype = 1;
1056             break;
1057     }
1058
1059     $tag_start = $pos;
1060     $tagname = '';
1061     /**
1062      * Look for next [\W-_], which will indicate the end of the tag name.
1063      */
1064     $regary = sq_findnxreg($body, $pos, "[^\w\-_]");
1065     if ($regary == false){
1066         return Array(false, false, false, $lt, strlen($body));
1067     }
1068     list($pos, $tagname, $match) = $regary;
1069     $tagname = strtolower($tagname);
1070
1071     /**
1072      * $match can be either of these:
1073      * '>'  indicating the end of the tag entirely.
1074      * '\s' indicating the end of the tag name.
1075      * '/'  indicating that this is type-3 xhtml tag.
1076      *
1077      * Whatever else we find there indicates an invalid tag.
1078      */
1079     switch ($match){
1080         case '/':
1081             /**
1082              * This is an xhtml-style tag with a closing / at the
1083              * end, like so: <img src="blah"/>. Check if it's followed
1084              * by the closing bracket. If not, then this tag is invalid
1085              */
1086             if (substr($body, $pos, 2) == "/>"){
1087                 $pos++;
1088                 $tagtype = 3;
1089             } else {
1090                 $gt = sq_findnxstr($body, $pos, ">");
1091                 $retary = Array(false, false, false, $lt, $gt);
1092                 return $retary;
1093             }
1094         case '>':
1095             return Array($tagname, false, $tagtype, $lt, $pos);
1096             break;
1097         default:
1098             /**
1099              * Check if it's whitespace
1100              */
1101             if (!preg_match('/\s/', $match)){
1102                 /**
1103                  * This is an invalid tag! Look for the next closing ">".
1104                  */
1105                 $gt = sq_findnxstr($body, $lt, ">");
1106                 return Array(false, false, false, $lt, $gt);
1107             }
1108             break;
1109     }
1110
1111     /**
1112      * At this point we're here:
1113      * <tagname  attribute='blah'>
1114      * \-------^
1115      *
1116      * At this point we loop in order to find all attributes.
1117      */
1118     $attname = '';
1119     $atttype = false;
1120     $attary = Array();
1121
1122     while ($pos <= strlen($body)){
1123         $pos = sq_skipspace($body, $pos);
1124         if ($pos == strlen($body)){
1125             /**
1126              * Non-closed tag.
1127              */
1128             return Array(false, false, false, $lt, $pos);
1129         }
1130         /**
1131          * See if we arrived at a ">" or "/>", which means that we reached
1132          * the end of the tag.
1133          */
1134         $matches = Array();
1135         if (preg_match("%^(\s*)(>|/>)%s", substr($body, $pos), $matches)) {
1136             /**
1137              * Yep. So we did.
1138              */
1139             $pos += strlen($matches{1});
1140             if ($matches{2} == "/>"){
1141                 $tagtype = 3;
1142                 $pos++;
1143             }
1144             return Array($tagname, $attary, $tagtype, $lt, $pos);
1145         }
1146
1147         /**
1148          * There are several types of attributes, with optional
1149          * [:space:] between members.
1150          * Type 1:
1151          *   attrname[:space:]=[:space:]'CDATA'
1152          * Type 2:
1153          *   attrname[:space:]=[:space:]"CDATA"
1154          * Type 3:
1155          *   attr[:space:]=[:space:]CDATA
1156          * Type 4:
1157          *   attrname
1158          *
1159          * We leave types 1 and 2 the same, type 3 we check for
1160          * '"' and convert to "&quot" if needed, then wrap in
1161          * double quotes. Type 4 we convert into:
1162          * attrname="yes".
1163          */
1164         $regary = sq_findnxreg($body, $pos, "[^:\w\-_]");
1165         if ($regary == false){
1166             /**
1167              * Looks like body ended before the end of tag.
1168              */
1169             return Array(false, false, false, $lt, strlen($body));
1170         }
1171         list($pos, $attname, $match) = $regary;
1172         $attname = strtolower($attname);
1173         /**
1174          * We arrived at the end of attribute name. Several things possible
1175          * here:
1176          * '>'  means the end of the tag and this is attribute type 4
1177          * '/'  if followed by '>' means the same thing as above
1178          * '\s' means a lot of things -- look what it's followed by.
1179          *      anything else means the attribute is invalid.
1180          */
1181         switch($match){
1182             case '/':
1183                 /**
1184                  * This is an xhtml-style tag with a closing / at the
1185                  * end, like so: <img src="blah"/>. Check if it's followed
1186                  * by the closing bracket. If not, then this tag is invalid
1187                  */
1188                 if (substr($body, $pos, 2) == "/>"){
1189                     $pos++;
1190                     $tagtype = 3;
1191                 } else {
1192                     $gt = sq_findnxstr($body, $pos, ">");
1193                     $retary = Array(false, false, false, $lt, $gt);
1194                     return $retary;
1195                 }
1196             case '>':
1197                 $attary{$attname} = '"yes"';
1198                 return Array($tagname, $attary, $tagtype, $lt, $pos);
1199                 break;
1200             default:
1201                 /**
1202                  * Skip whitespace and see what we arrive at.
1203                  */
1204                 $pos = sq_skipspace($body, $pos);
1205                 $char = substr($body, $pos, 1);
1206                 /**
1207                  * Two things are valid here:
1208                  * '=' means this is attribute type 1 2 or 3.
1209                  * \w means this was attribute type 4.
1210                  * anything else we ignore and re-loop. End of tag and
1211                  * invalid stuff will be caught by our checks at the beginning
1212                  * of the loop.
1213                  */
1214                 if ($char == "="){
1215                     $pos++;
1216                     $pos = sq_skipspace($body, $pos);
1217                     /**
1218                      * Here are 3 possibilities:
1219                      * "'"  attribute type 1
1220                      * '"'  attribute type 2
1221                      * everything else is the content of tag type 3
1222                      */
1223                     $quot = substr($body, $pos, 1);
1224                     if ($quot == "'"){
1225                         $regary = sq_findnxreg($body, $pos+1, "\'");
1226                         if ($regary == false){
1227                             return Array(false, false, false, $lt, strlen($body));
1228                         }
1229                         list($pos, $attval, $match) = $regary;
1230                         $pos++;
1231                         $attary{$attname} = "'" . $attval . "'";
1232                     } else if ($quot == '"'){
1233                         $regary = sq_findnxreg($body, $pos+1, '\"');
1234                         if ($regary == false){
1235                             return Array(false, false, false, $lt, strlen($body));
1236                         }
1237                         list($pos, $attval, $match) = $regary;
1238                         $pos++;
1239                         $attary{$attname} = '"' . $attval . '"';
1240                     } else {
1241                         /**
1242                          * These are hateful. Look for \s, or >.
1243                          */
1244                         $regary = sq_findnxreg($body, $pos, "[\s>]");
1245                         if ($regary == false){
1246                             return Array(false, false, false, $lt, strlen($body));
1247                         }
1248                         list($pos, $attval, $match) = $regary;
1249                         /**
1250                          * If it's ">" it will be caught at the top.
1251                          */
1252                         $attval = preg_replace("/\"/s", "&quot;", $attval);
1253                         $attary{$attname} = '"' . $attval . '"';
1254                     }
1255                 } else if (preg_match("|[\w/>]|", $char)) {
1256                     /**
1257                      * That was attribute type 4.
1258                      */
1259                     $attary{$attname} = '"yes"';
1260                 } else {
1261                     /**
1262                      * An illegal character. Find next '>' and return.
1263                      */
1264                     $gt = sq_findnxstr($body, $pos, ">");
1265                     return Array(false, false, false, $lt, $gt);
1266                 }
1267                 break;
1268         }
1269     }
1270     /**
1271      * The fact that we got here indicates that the tag end was never
1272      * found. Return invalid tag indication so it gets stripped.
1273      */
1274     return Array(false, false, false, $lt, strlen($body));
1275 }
1276
1277 /**
1278  * This function checks attribute values for entity-encoded values
1279  * and returns them translated into 8-bit strings so we can run
1280  * checks on them.
1281  *
1282  * @param  $attvalue A string to run entity check against.
1283  * @return           Translated value.
1284  */
1285 function sq_deent($attvalue){
1286     $me = 'sq_deent';
1287     /**
1288      * See if we have to run the checks first. All entities must start
1289      * with "&".
1290      */
1291     if (strpos($attvalue, "&") === false){
1292         return $attvalue;
1293     }
1294     /**
1295      * Check named entities first.
1296      */
1297     $trans = get_html_translation_table(HTML_ENTITIES);
1298     /**
1299      * Leave &quot; in, as it can mess us up.
1300      */
1301     $trans = array_flip($trans);
1302     unset($trans{"&quot;"});
1303     while (list($ent, $val) = each($trans)){
1304         $attvalue = preg_replace("/$ent*(\W)/si", "$val\\1", $attvalue);
1305     }
1306     /**
1307      * Now translate numbered entities from 1 to 255 if needed.
1308      */
1309     if (strpos($attvalue, "#") !== false){
1310         $omit = Array(34, 39);
1311         for ($asc=1; $asc<256; $asc++){
1312             if (!in_array($asc, $omit)){
1313                 $chr = chr($asc);
1314                 $attvalue = preg_replace("/\&#0*$asc;*(\D)/si", "$chr\\1",
1315                                          $attvalue);
1316                 $attvalue = preg_replace("/\&#x0*".dechex($asc).";*(\W)/si",
1317                                          "$chr\\1", $attvalue);
1318             }
1319         }
1320     }
1321     return $attvalue;
1322 }
1323
1324 /**
1325  * This function runs various checks against the attributes.
1326  *
1327  * @param  $tagname         String with the name of the tag.
1328  * @param  $attary          Array with all tag attributes.
1329  * @param  $rm_attnames     See description for sq_sanitize
1330  * @param  $bad_attvals     See description for sq_sanitize
1331  * @param  $add_attr_to_tag See description for sq_sanitize
1332  * @param  $message         message object
1333  * @param  $id              message id
1334  * @return                  Array with modified attributes.
1335  */
1336 function sq_fixatts($tagname,
1337                     $attary,
1338                     $rm_attnames,
1339                     $bad_attvals,
1340                     $add_attr_to_tag,
1341                     $message,
1342                     $id,
1343                     $mailbox
1344                     ){
1345     $me = 'sq_fixatts';
1346     while (list($attname, $attvalue) = each($attary)){
1347         /**
1348          * See if this attribute should be removed.
1349          */
1350         foreach ($rm_attnames as $matchtag=>$matchattrs){
1351             if (preg_match($matchtag, $tagname)){
1352                 foreach ($matchattrs as $matchattr){
1353                     if (preg_match($matchattr, $attname)){
1354                         unset($attary{$attname});
1355                         continue;
1356                     }
1357                 }
1358             }
1359         }
1360         /**
1361          * Remove any backslashes, entities, and extraneous whitespace.
1362          */
1363         $attvalue = sq_unbackslash($attvalue);
1364         $attvalue = sq_deent($attvalue);
1365         $attvalue = sq_unspace($attvalue);
1366
1367         /**
1368          * Now let's run checks on the attvalues.
1369          * I don't expect anyone to comprehend this. If you do,
1370          * get in touch with me so I can drive to where you live and
1371          * shake your hand personally. :)
1372          */
1373         foreach ($bad_attvals as $matchtag=>$matchattrs){
1374             if (preg_match($matchtag, $tagname)){
1375                 foreach ($matchattrs as $matchattr=>$valary){
1376                     if (preg_match($matchattr, $attname)){
1377                         /**
1378                          * There are two arrays in valary.
1379                          * First is matches.
1380                          * Second one is replacements
1381                          */
1382                         list($valmatch, $valrepl) = $valary;
1383                         $newvalue =
1384                             preg_replace($valmatch, $valrepl, $attvalue);
1385                         if ($newvalue != $attvalue){
1386                             $attary{$attname} = $newvalue;
1387                         }
1388                     }
1389                 }
1390             }
1391         }
1392         /**
1393          * Turn cid: urls into http-friendly ones.
1394          */
1395         if (preg_match("/^[\'\"]\s*cid:/si", $attvalue)){
1396             $attary{$attname} = sq_cid2http($message, $id, $attvalue, $mailbox);
1397         }
1398     }
1399     /**
1400      * See if we need to append any attributes to this tag.
1401      */
1402     foreach ($add_attr_to_tag as $matchtag=>$addattary){
1403         if (preg_match($matchtag, $tagname)){
1404             $attary = array_merge($attary, $addattary);
1405         }
1406     }
1407     return $attary;
1408 }
1409
1410 /**
1411  * This function edits the style definition to make them friendly and
1412  * usable in squirrelmail.
1413  *
1414  * @param  $message  the message object
1415  * @param  $id       the message id
1416  * @param  $content  a string with whatever is between <style> and </style>
1417  * @return           a string with edited content.
1418  */
1419 function sq_fixstyle($body, $pos, $message, $id){
1420     global $view_unsafe_images;
1421     $me = 'sq_fixstyle';
1422     $ret = sq_findnxreg($body, $pos, '</\s*style\s*>');
1423     if ($ret == FALSE){
1424         return array(FALSE, strlen($body));
1425     }
1426     $newpos = $ret[0] + strlen($ret[2]);
1427     $content = $ret[1];
1428     /**
1429      * First look for general BODY style declaration, which would be
1430      * like so:
1431      * body {background: blah-blah}
1432      * and change it to .bodyclass so we can just assign it to a <div>
1433      */
1434     $content = preg_replace("|body(\s*\{.*?\})|si", ".bodyclass\\1", $content);
1435     $secremoveimg = '../images/' . _("sec_remove_eng.png");
1436     /**
1437      * Fix url('blah') declarations.
1438      */
1439     $content = preg_replace("|url\s*\(\s*([\'\"])\s*\S+script\s*:.*?([\'\"])\s*\)|si",
1440                             "url(\\1$secremoveimg\\2)", $content);
1441     /**
1442      * Fix url('https*://.*) declarations but only if $view_unsafe_images
1443      * is false.
1444      */
1445     if (!$view_unsafe_images){
1446         $content = preg_replace("|url\s*\(\s*([\'\"])\s*https*:.*?([\'\"])\s*\)|si",
1447                                 "url(\\1$secremoveimg\\2)", $content);
1448     }
1449
1450     /**
1451      * Fix urls that refer to cid:
1452      */
1453     while (preg_match("|url\s*\(\s*([\'\"]\s*cid:.*?[\'\"])\s*\)|si",
1454                       $content, $matches)){
1455         $cidurl = $matches{1};
1456         $httpurl = sq_cid2http($message, $id, $cidurl);
1457         $content = preg_replace("|url\s*\(\s*$cidurl\s*\)|si",
1458                                 "url($httpurl)", $content);
1459     }
1460
1461     /**
1462      * Fix stupid css declarations which lead to vulnerabilities
1463      * in IE.
1464      */
1465     $match   = Array('/expression/i',
1466                      '/behaviou*r/i',
1467                      '/binding/i',
1468                      '/include-source/i');
1469     $replace = Array('idiocy', 'idiocy', 'idiocy', 'idiocy');
1470     $content = preg_replace($match, $replace, $content);
1471     return array($content, $newpos);
1472 }
1473
1474 /**
1475  * This function converts cid: url's into the ones that can be viewed in
1476  * the browser.
1477  *
1478  * @param  $message  the message object
1479  * @param  $id       the message id
1480  * @param  $cidurl   the cid: url.
1481  * @return           a string with a http-friendly url
1482  */
1483 function sq_cid2http($message, $id, $cidurl, $mailbox){
1484     /**
1485      * Get rid of quotes.
1486      */
1487     $quotchar = substr($cidurl, 0, 1);
1488     if ($quotchar == '"' || $quotchar == "'"){
1489         $cidurl = str_replace($quotchar, "", $cidurl);
1490     } else {
1491         $quotchar = '';
1492     }
1493     $cidurl = substr(trim($cidurl), 4);
1494     $linkurl = find_ent_id($cidurl, $message);
1495     /* in case of non-save cid links $httpurl should be replaced by a sort of
1496        unsave link image */
1497     $httpurl = '';
1498     if ($linkurl) {
1499         $httpurl = $quotchar . SM_PATH . 'src/download.php?absolute_dl=true&amp;' .
1500                    "passed_id=$id&amp;mailbox=" . urlencode($mailbox) .
1501                    '&amp;ent_id=' . $linkurl . $quotchar;
1502     }
1503     return $httpurl;
1504 }
1505
1506 /**
1507  * This function changes the <body> tag into a <div> tag since we
1508  * can't really have a body-within-body.
1509  *
1510  * @param  $attary   an array of attributes and values of <body>
1511  * @param  $mailbox  mailbox we're currently reading (for cid2http)
1512  * @param  $message  current message (for cid2http)
1513  * @param  $id       current message id (for cid2http)
1514  * @return           a modified array of attributes to be set for <div>
1515  */
1516 function sq_body2div($attary, $mailbox, $message, $id){
1517     $me = 'sq_body2div';
1518     $divattary = Array('class' => "'bodyclass'");
1519     $bgcolor = '#ffffff';
1520     $text = '#000000';
1521     $styledef = '';
1522     if (is_array($attary) && sizeof($attary) > 0){
1523         foreach ($attary as $attname=>$attvalue){
1524             $quotchar = substr($attvalue, 0, 1);
1525             $attvalue = str_replace($quotchar, "", $attvalue);
1526             switch ($attname){
1527                 case 'background':
1528                     $attvalue = sq_cid2http($message, $id,
1529                                             $attvalue, $mailbox);
1530                     $styledef .= "background-image: url('$attvalue'); ";
1531                     break;
1532                 case 'bgcolor':
1533                     $styledef .= "background-color: $attvalue; ";
1534                     break;
1535                 case 'text':
1536                     $styledef .= "color: $attvalue; ";
1537                     break;
1538             }
1539         }
1540         if (strlen($styledef) > 0){
1541             $divattary{"style"} = "\"$styledef\"";
1542         }
1543     }
1544     return $divattary;
1545 }
1546
1547 /**
1548  * This is the main function and the one you should actually be calling.
1549  * There are several variables you should be aware of an which need
1550  * special description.
1551  *
1552  * Since the description is quite lengthy, see it here:
1553  * http://www.mricon.com/html/phpfilter.html
1554  *
1555  * @param $body                 the string with HTML you wish to filter
1556  * @param $tag_list             see description above
1557  * @param $rm_tags_with_content see description above
1558  * @param $self_closing_tags    see description above
1559  * @param $force_tag_closing    see description above
1560  * @param $rm_attnames          see description above
1561  * @param $bad_attvals          see description above
1562  * @param $add_attr_to_tag      see description above
1563  * @param $message              message object
1564  * @param $id                   message id
1565  * @return                      sanitized html safe to show on your pages.
1566  */
1567 function sq_sanitize($body,
1568                      $tag_list,
1569                      $rm_tags_with_content,
1570                      $self_closing_tags,
1571                      $force_tag_closing,
1572                      $rm_attnames,
1573                      $bad_attvals,
1574                      $add_attr_to_tag,
1575                      $message,
1576                      $id,
1577                      $mailbox
1578                      ){
1579     $me = 'sq_sanitize';
1580     $rm_tags = array_shift($tag_list);
1581     /**
1582      * Normalize rm_tags and rm_tags_with_content.
1583      */
1584     @array_walk($tag_list, 'sq_casenormalize');
1585     @array_walk($rm_tags_with_content, 'sq_casenormalize');
1586     @array_walk($self_closing_tags, 'sq_casenormalize');
1587     /**
1588      * See if tag_list is of tags to remove or tags to allow.
1589      * false  means remove these tags
1590      * true   means allow these tags
1591      */
1592     $curpos = 0;
1593     $open_tags = Array();
1594     $trusted = "\n<!-- begin sanitized html -->\n";
1595     $skip_content = false;
1596     /**
1597      * Take care of netscape's stupid javascript entities like
1598      * &{alert('boo')};
1599      */
1600     $body = preg_replace("/&(\{.*?\};)/si", "&amp;\\1", $body);
1601
1602     while (($curtag = sq_getnxtag($body, $curpos)) != FALSE){
1603         list($tagname, $attary, $tagtype, $lt, $gt) = $curtag;
1604         $free_content = substr($body, $curpos, $lt-$curpos);
1605         /**
1606          * Take care of <style>
1607          */
1608         if ($tagname == "style" && $tagtype == 1){
1609             list($free_content, $curpos) =
1610                 sq_fixstyle($body, $gt+1, $message, $id);
1611             if ($free_content != FALSE){
1612                 $trusted .= sq_tagprint($tagname, $attary, $tagtype);
1613                 $trusted .= $free_content;
1614                 $trusted .= sq_tagprint($tagname, false, 2);
1615             }
1616             continue;
1617         }
1618         if ($skip_content == false){
1619             $trusted .= $free_content;
1620         }
1621         if ($tagname != FALSE){
1622             if ($tagtype == 2){
1623                 if ($skip_content == $tagname){
1624                     /**
1625                      * Got to the end of tag we needed to remove.
1626                      */
1627                     $tagname = false;
1628                     $skip_content = false;
1629                 } else {
1630                     if ($skip_content == false){
1631                         if ($tagname == "body"){
1632                             $tagname = "div";
1633                         }
1634                         if (isset($open_tags{$tagname}) &&
1635                             $open_tags{$tagname} > 0){
1636                             $open_tags{$tagname}--;
1637                         } else {
1638                             $tagname = false;
1639                         }
1640                     }
1641                 }
1642             } else {
1643                 /**
1644                  * $rm_tags_with_content
1645                  */
1646                 if ($skip_content == false){
1647                     /**
1648                      * See if this is a self-closing type and change
1649                      * tagtype appropriately.
1650                      */
1651                     if ($tagtype == 1
1652                         && in_array($tagname, $self_closing_tags)){
1653                         $tagtype = 3;
1654                     }
1655                     /**
1656                      * See if we should skip this tag and any content
1657                      * inside it.
1658                      */
1659                     if ($tagtype == 1 &&
1660                         in_array($tagname, $rm_tags_with_content)){
1661                         $skip_content = $tagname;
1662                     } else {
1663                         if (($rm_tags == false
1664                              && in_array($tagname, $tag_list)) ||
1665                             ($rm_tags == true &&
1666                              !in_array($tagname, $tag_list))){
1667                             $tagname = false;
1668                         } else {
1669                             /**
1670                              * Convert body into div.
1671                              */
1672                             if ($tagname == "body"){
1673                                 $tagname = "div";
1674                                 $attary = sq_body2div($attary, $mailbox,
1675                                                       $message, $id);
1676                             }
1677                             if ($tagtype == 1){
1678                                 if (isset($open_tags{$tagname})){
1679                                     $open_tags{$tagname}++;
1680                                 } else {
1681                                     $open_tags{$tagname}=1;
1682                                 }
1683                             }
1684                             /**
1685                              * This is where we run other checks.
1686                              */
1687                             if (is_array($attary) && sizeof($attary) > 0){
1688                                 $attary = sq_fixatts($tagname,
1689                                                      $attary,
1690                                                      $rm_attnames,
1691                                                      $bad_attvals,
1692                                                      $add_attr_to_tag,
1693                                                      $message,
1694                                                      $id,
1695                                                      $mailbox
1696                                                      );
1697                             }
1698                         }
1699                     }
1700                 }
1701             }
1702             if ($tagname != false && $skip_content == false){
1703                 $trusted .= sq_tagprint($tagname, $attary, $tagtype);
1704             }
1705         }
1706         $curpos = $gt+1;
1707     }
1708     $trusted .= substr($body, $curpos, strlen($body)-$curpos);
1709     if ($force_tag_closing == true){
1710         foreach ($open_tags as $tagname=>$opentimes){
1711             while ($opentimes > 0){
1712                 $trusted .= '</' . $tagname . '>';
1713                 $opentimes--;
1714             }
1715         }
1716         $trusted .= "\n";
1717     }
1718     $trusted .= "<!-- end sanitized html -->\n";
1719     return $trusted;
1720 }
1721
1722 /**
1723  * This is a wrapper function to call html sanitizing routines.
1724  *
1725  * @param  $body  the body of the message
1726  * @param  $id    the id of the message
1727  * @return        a string with html safe to display in the browser.
1728  */
1729 function magicHTML($body, $id, $message, $mailbox = 'INBOX') {
1730     global $attachment_common_show_images, $view_unsafe_images,
1731            $has_unsafe_images;
1732     /**
1733      * Don't display attached images in HTML mode.
1734      */
1735     $attachment_common_show_images = false;
1736     $tag_list = Array(
1737                       false,
1738                       "object",
1739                       "meta",
1740                       "html",
1741                       "head",
1742                       "base",
1743                       "link",
1744                       "frame",
1745                       "iframe",
1746                       "plaintext",
1747                       "marquee"
1748                       );
1749
1750     $rm_tags_with_content = Array(
1751                                   "script",
1752                                   "applet",
1753                                   "embed",
1754                                   "title",
1755                                   "frameset",
1756                                   "xml"
1757                                   );
1758
1759     $self_closing_tags =  Array(
1760                                 "img",
1761                                 "br",
1762                                 "hr",
1763                                 "input"
1764                                 );
1765
1766     $force_tag_closing = true;
1767
1768     $rm_attnames = Array(
1769                          "/.*/" =>
1770                          Array(
1771                                "/target/i",
1772                                "/^on.*/i",
1773                                "/^dynsrc/i",
1774                                "/^data.*/i",
1775                                "/^lowsrc.*/i"
1776                                )
1777                          );
1778
1779     $secremoveimg = "../images/" . _("sec_remove_eng.png");
1780     $bad_attvals = Array(
1781         "/.*/" =>
1782             Array(
1783                 "/^src|background/i" =>
1784                     Array(
1785                           Array(
1786                                 "/^([\'\"])\s*\S+script\s*:.*([\'\"])/si",
1787                                 "/^([\'\"])\s*mocha\s*:*.*([\'\"])/si",
1788                                 "/^([\'\"])\s*about\s*:.*([\'\"])/si"
1789                                 ),
1790                           Array(
1791                                 "\\1$secremoveimg\\2",
1792                                 "\\1$secremoveimg\\2",
1793                                 "\\1$secremoveimg\\2",
1794                                 "\\1$secremoveimg\\2"
1795                                 )
1796                         ),
1797                 "/^href|action/i" =>
1798                     Array(
1799                           Array(
1800                                 "/^([\'\"])\s*\S+script\s*:.*([\'\"])/si",
1801                                 "/^([\'\"])\s*mocha\s*:*.*([\'\"])/si",
1802                                 "/^([\'\"])\s*about\s*:.*([\'\"])/si"
1803                                 ),
1804                           Array(
1805                                 "\\1#\\1",
1806                                 "\\1#\\1",
1807                                 "\\1#\\1",
1808                                 "\\1#\\1"
1809                                 )
1810                         ),
1811                 "/^style/i" =>
1812                     Array(
1813                           Array(
1814                                 "/expression/i",
1815                                 "/binding/i",
1816                                 "/behaviou*r/i",
1817                                 "/include-source/i",
1818                                 "/url\s*\(\s*([\'\"])\s*\S+script\s*:.*([\'\"])\s*\)/si",
1819                                 "/url\s*\(\s*([\'\"])\s*mocha\s*:.*([\'\"])\s*\)/si",
1820                                 "/url\s*\(\s*([\'\"])\s*about\s*:.*([\'\"])\s*\)/si"
1821                                ),
1822                           Array(
1823                                 "idiocy",
1824                                 "idiocy",
1825                                 "idiocy",
1826                                 "idiocy",
1827                                 "url(\\1#\\1)",
1828                                 "url(\\1#\\1)",
1829                                 "url(\\1#\\1)",
1830                                 "url(\\1#\\1)"
1831                                )
1832                           )
1833                 )
1834         );
1835     if( !sqgetGlobalVar('view_unsafe_images', $view_unsafe_images, SQ_GET) ) {
1836         $view_unsafe_images = false;
1837     }
1838     if (!$view_unsafe_images){
1839         /**
1840          * Remove any references to http/https if view_unsafe_images set
1841          * to false.
1842          */
1843          array_push($bad_attvals{'/.*/'}{'/^src|background/i'}[0],
1844                     '/^([\'\"])\s*https*:.*([\'\"])/si');
1845          array_push($bad_attvals{'/.*/'}{'/^src|background/i'}[1],
1846                     "\\1$secremoveimg\\1");
1847          array_push($bad_attvals{'/.*/'}{'/^style/i'}[0],
1848                     '/url\(([\'\"])\s*https*:.*([\'\"])\)/si');
1849          array_push($bad_attvals{'/.*/'}{'/^style/i'}[1],
1850                     "url(\\1$secremoveimg\\1)");
1851     }
1852
1853     $add_attr_to_tag = Array(
1854         "/^a$/i" =>
1855             Array('target'=>'"_new"',
1856                   'title'=>'"'._("This external link will open in a new window").'"'
1857             )
1858     );
1859     $trusted = sq_sanitize($body,
1860                            $tag_list,
1861                            $rm_tags_with_content,
1862                            $self_closing_tags,
1863                            $force_tag_closing,
1864                            $rm_attnames,
1865                            $bad_attvals,
1866                            $add_attr_to_tag,
1867                            $message,
1868                            $id,
1869                            $mailbox
1870                            );
1871     if (preg_match("|$secremoveimg|i", $trusted)){
1872         $has_unsafe_images = true;
1873     }
1874     return $trusted;
1875 }
1876
1877 ?>