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