functions/mime.php

   1 <?php
   2
   3 /**
   4  * mime.php
   5  *
   6  * Copyright (c) 1999-2002 The SquirrelMail Project Team
   7  * Licensed under the GNU GPL. For full terms see the file COPYING.
   8  *
   9  * This contains the functions necessary to detect and decode MIME
  10  * messages.
  11  *
  12  * $Id$
  13  */
  14
  15 require_once('../functions/imap.php');
  16 require_once('../functions/attachment_common.php');
  17
  18 /* --------------------------------------------------------------------------------- */
  19 /* MIME DECODING                                                                     */
  20 /* --------------------------------------------------------------------------------- */
  21
  22 /* This function gets the structure of a message and stores it in the "message" class.
  23  * It will return this object for use with all relevant header information and
  24  * fully parsed into the standard "message" object format.
  25  */
  26
  27 function mime_structure ($bodystructure, $flags=array()) {
  28
  29     // isolate the body structure and remove beginning and end parenthesis
  30     $read = trim(substr ($bodystructure, strpos(strtolower($bodystructure), 'bodystructure') + 13));
  31     $msg = &new message();
  32     $read = trim(substr ($read, 0, -1));
  33     $res = $msg->parseStructure($read);
  34     $msg = $res[0];
  35     $msg->setEnt('0');
  36     if (count($flags)) {
  37        foreach ($flags as $flag) {
  38           $char = strtoupper($flag{1});
  39           switch ($char) {
  40              case 'S':
  41                if (strtolower($flag) == '\\seen') {
  42                   $msg->is_seen = true;
  43                }
  44                break;
  45              case 'A':
  46                if (strtolower($flag) == '\\answered') {
  47                   $msg->is_answered = true;
  48                }
  49                break;
  50              case 'D':
  51                if (strtolower($flag) == '\\deleted') {
  52                   $msg->is_deleted = true;
  53                }
  54                break;
  55              case 'F':
  56                if (strtolower($flag) == '\\flagged') {
  57                   $msg->is_flagged = true;
  58                }
  59                break;
  60              case 'M':
  61                if (strtolower($flag) == '$mdnsent') {
  62                   $msg->is_mdnsent = true;
  63                }
  64                break;
  65              default:
  66                break;
  67           }
  68        }
  69     }
  70 //    listEntities($msg);
  71     return( $msg );
  72 }
  73
  74 /* this starts the parsing of a particular structure.  It is called recursively,
  75  * so it can be passed different structures.  It returns an object of type
  76  * $message.
  77  * First, it checks to see if it is a multipart message.  If it is, then it
  78  * handles that as it sees is necessary.  If it is just a regular entity,
  79  * then it parses it and adds the necessary header information (by calling out
  80  * to mime_get_elements()
  81  */
  82
  83 function mime_fetch_body($imap_stream, $id, $ent_id) {
  84     global $uid_support;
  85     /*
  86      * do a bit of error correction.  If we couldn't find the entity id, just guess
  87      * that it is the first one.  That is usually the case anyway.
  88      */
  89     if (!$ent_id) {
  90         $ent_id = 1;
  91     }
  92     $cmd = "FETCH $id BODY[$ent_id]";
  93
  94     $data = sqimap_run_command ($imap_stream, $cmd, true, $response, $message, $uid_support);
  95     do {
  96         $topline = trim(array_shift( $data ));
  97     } while( $topline && $topline[0] == '*' && !preg_match( '/\* [0-9]+ FETCH.*/i', $topline )) ;
  98
  99     $wholemessage = implode('', $data);
 100     if (ereg('\\{([^\\}]*)\\}', $topline, $regs)) {
 101
 102         $ret = substr( $wholemessage, 0, $regs[1] );
 103         /*
 104             There is some information in the content info header that could be important
 105             in order to parse html messages. Let's get them here.
 106         */
 107         if ( $ret{0} == '<' ) {
 108             $data = sqimap_run_command ($imap_stream, "FETCH $id BODY[$ent_id.MIME]", true, $response, $message, $uid_support);
 109         }
 110     } else if (ereg('"([^"]*)"', $topline, $regs)) {
 111         $ret = $regs[1];
 112     } else {
 113         global $where, $what, $mailbox, $passed_id, $startMessage;
 114         $par = 'mailbox=' . urlencode($mailbox) . "&amp;passed_id=$passed_id";
 115         if (isset($where) && isset($what)) {
 116             $par .= '&amp;where='. urlencode($where) . "&amp;what=" . urlencode($what);
 117         } else {
 118             $par .= "&amp;startMessage=$startMessage&amp;show_more=0";
 119         }
 120         $par .= '&amp;response=' . urlencode($response) .
 121                 '&amp;message=' . urlencode($message).
 122                 '&amp;topline=' . urlencode($topline);
 123
 124         echo   '<tt><br>' .
 125                '<table width="80%"><tr>' .
 126                '<tr><td colspan=2>' .
 127                _("Body retrieval error. The reason for this is most probably that the message is malformed. Please help us making future versions better by submitting this message to the developers knowledgebase!") .
 128                " <A HREF=\"../src/retrievalerror.php?$par\"><br>" .
 129                _("Submit message") . '</A><BR>&nbsp;' .
 130                '</td></tr>' .
 131                '<td><b>' . _("Command:") . "</td><td>$cmd</td></tr>" .
 132                '<td><b>' . _("Response:") . "</td><td>$response</td></tr>" .
 133                '<td><b>' . _("Message:") . "</td><td>$message</td></tr>" .
 134                '<td><b>' . _("FETCH line:") . "</td><td>$topline</td></tr>" .
 135                "</table><BR></tt></font><hr>";
 136
 137         $data = sqimap_run_command ($imap_stream, "FETCH $passed_id BODY[]", true, $response, $message, $uid_support);
 138         array_shift($data);
 139         $wholemessage = implode('', $data);
 140
 141         $ret = $wholemessage;
 142     }
 143     return( $ret );
 144 }
 145
 146 function mime_print_body_lines ($imap_stream, $id, $ent_id, $encoding) {
 147     global $uid_support;
 148     // do a bit of error correction.  If we couldn't find the entity id, just guess
 149     // that it is the first one.  That is usually the case anyway.
 150     if (!$ent_id) {
 151         $ent_id = 1;
 152     }
 153     $sid = sqimap_session_id($uid_support);
 154     // Don't kill the connection if the browser is over a dialup
 155     // and it would take over 30 seconds to download it.
 156
 157     // don´t call set_time_limit in safe mode.
 158     if (!ini_get("safe_mode")) {
 159         set_time_limit(0);
 160     }
 161     if ($uid_support) {
 162        $sid_s = substr($sid,0,strpos($sid, ' '));
 163     } else {
 164        $sid_s = $sid;
 165     }
 166
 167     $body = mime_fetch_body ($imap_stream, $id, $ent_id);
 168     echo decodeBody($body, $encoding);
 169     return;
 170 /*
 171     fputs ($imap_stream, "$sid FETCH $id BODY[$ent_id]\r\n");
 172     $cnt = 0;
 173     $continue = true;
 174     $read = fgets ($imap_stream,8192);
 175
 176
 177     // This could be bad -- if the section has sqimap_session_id() . ' OK'
 178     // or similar, it will kill the download.
 179     while (!ereg("^".$sid_s." (OK|BAD|NO)(.*)$", $read, $regs)) {
 180         if (trim($read) == ')==') {
 181             $read1 = $read;
 182             $read = fgets ($imap_stream,4096);
 183             if (ereg("^".$sid." (OK|BAD|NO)(.*)$", $read, $regs)) {
 184                 return;
 185             } else {
 186                 echo decodeBody($read1, $encoding) .
 187                      decodeBody($read, $encoding);
 188             }
 189         } else if ($cnt) {
 190             echo decodeBody($read, $encoding);
 191         }
 192         $read = fgets ($imap_stream,4096);
 193         $cnt++;
 194 //      break;
 195     }
 196 */
 197 }
 198
 199 /* -[ END MIME DECODING ]----------------------------------------------------------- */
 200
 201 /* findDisplayEntity
 202  * Checks to see if $message contains content of type $type0/$type1
 203  * returns the first entity number it finds of that type, or NULL if
 204  * none is found.  Takes optional argument $start to allow the caller
 205  * to continue where they left off
 206  */
 207 function findDisplayEntity($message, $type0, $type1, $start=0) {
 208     if ($message) {
 209         for ($i = $start;isset($message->entities[$i]); $i++) {
 210             $entity = $message->entities[$i];
 211             if ($entity->type0 == $type0 && $entity->type1 == $type1) {
 212                 return $i;
 213             }
 214         }
 215     }
 216     return NULL;
 217 }
 218
 219 // This is here for debugging purposese.  It will print out a list
 220 // of all the entity IDs that are in the $message object.
 221
 222 function listEntities ($message) {
 223   if ($message) {
 224     echo "<tt>" . $message->entity_id . ' : ' . $message->type0 . '/' . $message->type1 . ' parent = '. $message->parent->entity_id. '<br>';
 225     for ($i = 0;isset($message->entities[$i]); $i++) {
 226        echo "$i : ";
 227        $msg = listEntities($message->entities[$i]);
 228
 229        if ($msg) {
 230           echo "return: ";
 231           return $msg;
 232        }
 233     }
 234   }
 235 }
 236
 237 function getPriorityStr($priority) {
 238    $priority_level = substr($priority,0,1);
 239
 240    switch($priority_level) {
 241      /* check for a higher then normal priority. */
 242      case '1':
 243      case '2':
 244         $priority_string = _("High");
 245         break;
 246
 247       /* check for a lower then normal priority. */
 248      case '4':
 249      case '5':
 250         $priority_string = _("Low");
 251         break;
 252
 253      /* check for a normal priority. */
 254      case '3':
 255      default:
 256         $priority_level = '3';
 257         $priority_string = _("Normal");
 258         break;
 259
 260    }
 261    return $priority_string;
 262 }
 263
 264 /* returns a $message object for a particular entity id */
 265 function getEntity ($message, $ent_id) {
 266     return $message->getEntity($ent_id);
 267 }
 268
 269 /*
 270  * translateText
 271  * Extracted from strings.php 23/03/2002
 272  */
 273
 274 function translateText(&$body, $wrap_at, $charset) {
 275     global $where, $what; /* from searching */
 276     global $color; /* color theme */
 277
 278     require_once('../functions/url_parser.php');
 279
 280     $body_ary = explode("\n", $body);
 281     $PriorQuotes = 0;
 282     for ($i=0; $i < count($body_ary); $i++) {
 283         $line = $body_ary[$i];
 284         if (strlen($line) - 2 >= $wrap_at) {
 285             sqWordWrap($line, $wrap_at);
 286         }
 287         $line = charset_decode($charset, $line);
 288         $line = str_replace("\t", '        ', $line);
 289
 290         parseUrl ($line);
 291
 292         $Quotes = 0;
 293         $pos = 0;
 294         $j = strlen( $line );
 295
 296         while ( $pos < $j ) {
 297             if ($line[$pos] == ' ') {
 298                 $pos ++;
 299             } else if (strpos($line, '&gt;', $pos) === $pos) {
 300                 $pos += 4;
 301                 $Quotes ++;
 302             } else {
 303                 break;
 304             }
 305         }
 306
 307         if ($Quotes > 1) {
 308             if (! isset($color[14])) {
 309                 $color[14] = '#FF0000';
 310             }
 311             $line = '<FONT COLOR="' . $color[14] . '">' . $line . '</FONT>';
 312         } elseif ($Quotes) {
 313             if (! isset($color[13])) {
 314                 $color[13] = '#800000';
 315             }
 316             $line = '<FONT COLOR="' . $color[13] . '">' . $line . '</FONT>';
 317         }
 318
 319         $body_ary[$i] = $line;
 320     }
 321     $body = '<pre>' . implode("\n", $body_ary) . '</pre>';
 322 }
 323
 324
 325 /* This returns a parsed string called $body. That string can then
 326 be displayed as the actual message in the HTML. It contains
 327 everything needed, including HTML Tags, Attachments at the
 328 bottom, etc.
 329 */
 330 function formatBody($imap_stream, $message, $color, $wrap_at, $ent_num, $id, $mailbox='INBOX') {
 331     // this if statement checks for the entity to show as the
 332     // primary message. To add more of them, just put them in the
 333     // order that is their priority.
 334     global $startMessage, $username, $key, $imapServerAddress, $imapPort,
 335            $show_html_default, $has_unsafe_images, $view_unsafe_images, $sort;
 336
 337     $has_unsafe_images= 0;
 338     $body = '';
 339     $urlmailbox = urlencode($mailbox);
 340     $body_message = getEntity($message, $ent_num);
 341     if (($body_message->header->type0 == 'text') ||
 342         ($body_message->header->type0 == 'rfc822')) {
 343         $body = mime_fetch_body ($imap_stream, $id, $ent_num);
 344         $body = decodeBody($body, $body_message->header->encoding);
 345         $hookResults = do_hook("message_body", $body);
 346         $body = $hookResults[1];
 347
 348         // If there are other types that shouldn't be formatted, add
 349         // them here
 350         if ($body_message->header->type1 == 'html') {
 351             if ( $show_html_default <> 1 ) {
 352                 $body = strip_tags( $body );
 353                 translateText($body, $wrap_at,
 354                   $body_message->header->getParameter['charset']);
 355             } else {
 356                 $body = magicHTML( $body, $id, $message, $mailbox );
 357             }
 358         } else {
 359             translateText($body, $wrap_at,
 360                $body_message->header->getParameter('charset'));
 361         }
 362
 363         if ($has_unsafe_images) {
 364             if ($view_unsafe_images) {
 365                 $body .= "<CENTER><SMALL><A HREF=\"read_body.php?passed_id=$id&amp;passed_ent_id=".$message->entity_id."&amp;mailbox=$urlmailbox&amp;sort=$sort&amp;startMessage=$startMessage&amp;show_more=0\">". _("Hide Unsafe Images") ."</A></SMALL></CENTER><BR>\n";
 366             } else {
 367                 $body .= "<CENTER><SMALL><A HREF=\"read_body.php?passed_id=$id&amp;passed_ent_id=".$message->entity_id."&amp;mailbox=$urlmailbox&amp;sort=$sort&amp;startMessage=$startMessage&amp;show_more=0&amp;view_unsafe_images=1\">". _("View Unsafe Images") ."</A></SMALL></CENTER><BR>\n";
 368             }
 369         }
 370     }
 371     return ($body);
 372 }
 373
 374
 375 function formatAttachments($message, $exclude_id, $mailbox, $id) {
 376     global $where, $what;
 377     global $startMessage, $color;
 378     static $ShownHTML = 0;
 379
 380     $att_ar = $message->getAttachments($exclude_id);
 381
 382     if (!count($att_ar)) return '';
 383
 384     $attachments = '';
 385
 386     $urlMailbox = urlencode($mailbox);
 387
 388     foreach ($att_ar as $att) {
 389         $ent = urldecode($att->entity_id);
 390         $header = $att->header;
 391         $type0 = strtolower($header->type0);
 392         $type1 = strtolower($header->type1);
 393         $name = '';
 394         $Links['download link']['text'] = _("download");
 395         $Links['download link']['href'] =
 396                 "../src/download.php?absolute_dl=true&amp;passed_id=$id&amp;mailbox=$urlMailbox&amp;ent_id=$ent";
 397         $ImageURL = '';
 398         if ($type0 =='message' && $type1 == 'rfc822') {
 399             $default_page = '../src/read_body.php';
 400             $rfc822_header = $att->rfc822_header;
 401             $filename = decodeHeader($rfc822_header->subject);
 402
 403             $from_o = $rfc822_header->from;
 404             if (is_object($from_o)) {
 405                 $from_name = $from_o->getAddress(false);
 406             } else {
 407                 $from_name = _("Unknown sender");
 408             }
 409             $from_name = decodeHeader(htmlspecialchars($from_name));
 410             $description = $from_name;
 411         } else {
 412             $default_page = '../src/download.php';
 413             if (is_object($header->disposition)) {
 414                $filename = decodeHeader($header->disposition->getProperty('filename'));
 415                if (trim($filename) == '') {
 416                   $name = decodeHeader($header->disposition->getProperty('name'));
 417                   if (trim($name) == '') {
 418                      if ( trim( $header->id ) == '' )
 419                         $filename = 'untitled-[' . $ent . ']' ;
 420                      else
 421                         $filename = 'cid: ' . $header->id;
 422                   } else {
 423                      $filename = $name;
 424                   }
 425                }
 426             } else {
 427                if ( trim( $header->id ) == '' )
 428                   $filename = 'untitled-[' . $ent . ']' ;
 429                else
 430                   $filename = 'cid: ' . $header->id;
 431             }
 432
 433             if ($header->description) {
 434                 $description = htmlspecialchars($header->description);
 435             } else {
 436                 $description = '';
 437             }
 438         }
 439
 440         $display_filename = $filename;
 441         if (isset($passed_ent_id)) {
 442            $passed_ent_id_link = '&amp;passed_ent_id='.$passed_ent_id;
 443         } else {
 444            $passed_ent_id_link = '';
 445         }
 446         $DefaultLink = $default_page . "?startMessage=$startMessage"
 447                      . "&amp;passed_id=$id&amp;mailbox=$urlMailbox"
 448                      . '&amp;ent_id='.$ent.$passed_ent_id_link;
 449         if ($where && $what) {
 450            $DefaultLink = '&amp;where='. urlencode($where).'&amp;what='.urlencode($what);
 451         }
 452         /* this executes the attachment hook with a specific MIME-type.
 453             * if that doens't have results, it tries if there's a rule
 454             * for a more generic type. */
 455         $HookResults = do_hook("attachment $type0/$type1", $Links,
 456             $startMessage, $id, $urlMailbox, $ent, $DefaultLink,
 457             $display_filename, $where, $what);
 458         if(count($HookResults[1]) <= 1) {
 459             $HookResults = do_hook("attachment $type0/*", $Links,
 460                     $startMessage, $id, $urlMailbox, $ent, $DefaultLink,
 461                     $display_filename, $where, $what);
 462         }
 463
 464         $Links = $HookResults[1];
 465         $DefaultLink = $HookResults[6];
 466
 467         $attachments .= '<TR><TD>' .
 468                         "<A HREF=\"$DefaultLink\">$display_filename</A>&nbsp;</TD>" .
 469                         '<TD><SMALL><b>' . show_readable_size($header->size) .
 470                         '</b>&nbsp;&nbsp;</small></TD>' .
 471                         "<TD><SMALL>[ $type0/$type1 ]&nbsp;</SMALL></TD>" .
 472                         '<TD><SMALL>';
 473         $attachments .= '<b>' . $description . '</b>';
 474         $attachments .= '</SMALL></TD><TD><SMALL>&nbsp;';
 475
 476         $SkipSpaces = 1;
 477         foreach ($Links as $Val) {
 478             if ($SkipSpaces) {
 479                 $SkipSpaces = 0;
 480             } else {
 481                 $attachments .= '&nbsp;&nbsp;|&nbsp;&nbsp;';
 482             }
 483             $attachments .= '<a href="' . $Val['href'] . '">' .  $Val['text'] . '</a>';
 484         }
 485         unset($Links);
 486         $attachments .= "</TD></TR>\n";
 487     }
 488     return $attachments;
 489 }
 490
 491 /** this function decodes the body depending on the encoding type. **/
 492 function decodeBody($body, $encoding) {
 493   $body = str_replace("\r\n", "\n", $body);
 494   $encoding = strtolower($encoding);
 495
 496   global $show_html_default;
 497
 498   if ($encoding == 'quoted-printable' ||
 499       $encoding == 'quoted_printable') {
 500      $body = quoted_printable_decode($body);
 501
 502      while (ereg("=\n", $body))
 503         $body = ereg_replace ("=\n", "", $body);
 504
 505   } else if ($encoding == 'base64') {
 506      $body = base64_decode($body);
 507   }
 508
 509   // All other encodings are returned raw.
 510   return $body;
 511 }
 512
 513 /*
 514  * This functions decode strings that is encoded according to
 515  * RFC1522 (MIME Part Two: Message Header Extensions for Non-ASCII Text).
 516  * Patched by Christian Schmidt <christian@ostenfeld.dk>  23/03/2002
 517  */
 518 function decodeHeader ($string, $utfencode=true) {
 519     if (is_array($string)) {
 520         $string = implode("\n", $string);
 521     }
 522     $i = 0;
 523     while (preg_match('/^(.{' . $i . '})(.*)=\?([^?]*)\?(Q|B)\?([^?]*)\?=/Ui',
 524                       $string, $res)) {
 525         $prefix = $res[1];
 526         // Ignore white-space between consecutive encoded-words
 527         if (strspn($res[2], " \t") != strlen($res[2])) {
 528             $prefix .= $res[2];
 529         }
 530
 531         if (ucfirst($res[4]) == 'B') {
 532             $replace = base64_decode($res[5]);
 533         } else {
 534             $replace = str_replace('_', ' ', $res[5]);
 535             $replace = preg_replace('/=([0-9a-f]{2})/ie', 'chr(hexdec("\1"))',
 536                                     $replace);
 537             /* Only encode into entities by default. Some places
 538                don't need the encoding, like the compose form. */
 539             if ($utfencode) {
 540                 $replace = charset_decode($res[3], $replace);
 541             }
 542         }
 543         $string = $prefix . $replace . substr($string, strlen($res[0]));
 544         $i = strlen($prefix) + strlen($replace);
 545     }
 546     return( $string );
 547 }
 548
 549 /*
 550  * Encode a string according to RFC 1522 for use in headers if it
 551  * contains 8-bit characters or anything that looks like it should
 552  * be encoded.
 553  */
 554 function encodeHeader ($string) {
 555     global $default_charset;
 556
 557     // Encode only if the string contains 8-bit characters or =?
 558     $j = strlen( $string  );
 559     $l = strstr($string, '=?');         // Must be encoded ?
 560     $ret = '';
 561     for( $i=0; $i < $j; ++$i) {
 562         switch( $string{$i} ) {
 563            case '=':
 564           $ret .= '=3D';
 565           break;
 566         case '?':
 567           $ret .= '=3F';
 568           break;
 569         case '_':
 570           $ret .= '=5F';
 571           break;
 572         case ' ':
 573           $ret .= '_';
 574           break;
 575         default:
 576           $k = ord( $string{$i} );
 577           if ( $k > 126 ) {
 578              $ret .= sprintf("=%02X", $k);
 579              $l = TRUE;
 580           } else
 581              $ret .= $string{$i};
 582         }
 583     }
 584
 585     if ( $l ) {
 586         $string = "=?$default_charset?Q?$ret?=";
 587     }
 588
 589     return( $string );
 590 }
 591
 592 /* This function trys to locate the entity_id of a specific mime element */
 593
 594 function find_ent_id( $id, $message ) {
 595     $ret = '';
 596     for ($i=0; $ret == '' && $i < count($message->entities); $i++) {
 597         if ( $message->entities[$i]->header->type0 == 'multipart')  {
 598             $ret = find_ent_id( $id, $message->entities[$i] );
 599         } else {
 600             if ( strcasecmp( $message->entities[$i]->header->id, $id ) == 0 )
 601                 $ret = $message->entities[$i]->entity_id;
 602         }
 603     }
 604     return( $ret );
 605 }
 606
 607 /**
 608  ** HTMLFILTER ROUTINES
 609  */
 610
 611 /**
 612  * This function returns the final tag out of the tag name, an array
 613  * of attributes, and the type of the tag. This function is called by
 614  * sq_sanitize internally.
 615  *
 616  * @param  $tagname  the name of the tag.
 617  * @param  $attary   the array of attributes and their values
 618  * @param  $tagtype  The type of the tag (see in comments).
 619  * @return           a string with the final tag representation.
 620  */
 621 function sq_tagprint($tagname, $attary, $tagtype){
 622     $me = "sq_tagprint";
 623     if ($tagtype == 2){
 624         $fulltag = '</' . $tagname . '>';
 625     } else {
 626         $fulltag = '<' . $tagname;
 627         if (is_array($attary) && sizeof($attary)){
 628             $atts = Array();
 629             while (list($attname, $attvalue) = each($attary)){
 630                 array_push($atts, "$attname=$attvalue");
 631             }
 632             $fulltag .= ' ' . join(" ", $atts);
 633         }
 634         if ($tagtype == 3){
 635             $fulltag .= " /";
 636         }
 637         $fulltag .= ">";
 638     }
 639     return $fulltag;
 640 }
 641
 642 /**
 643  * A small helper function to use with array_walk. Modifies a by-ref
 644  * value and makes it lowercase.
 645  *
 646  * @param  $val a value passed by-ref.
 647  * @return      void since it modifies a by-ref value.
 648  */
 649 function sq_casenormalize(&$val){
 650     $val = strtolower($val);
 651 }
 652
 653 /**
 654  * This function skips any whitespace from the current position within
 655  * a string and to the next non-whitespace value.
 656  *
 657  * @param  $body   the string
 658  * @param  $offset the offset within the string where we should start
 659  *                 looking for the next non-whitespace character.
 660  * @return         the location within the $body where the next
 661  *                 non-whitespace char is located.
 662  */
 663 function sq_skipspace($body, $offset){
 664     $me = "sq_skipspace";
 665     preg_match("/^(\s*)/s", substr($body, $offset), $matches);
 666     if (sizeof($matches{1})){
 667         $count = strlen($matches{1});
 668         $offset += $count;
 669     }
 670     return $offset;
 671 }
 672
 673 /**
 674  * This function looks for the next character within a string.  It's
 675  * really just a glorified "strpos", except it catches if failures
 676  * nicely.
 677  *
 678  * @param  $body   The string to look for needle in.
 679  * @param  $offset Start looking from this position.
 680  * @param  $needle The character/string to look for.
 681  * @return         location of the next occurance of the needle, or
 682  *                 strlen($body) if needle wasn't found.
 683  */
 684 function sq_findnxstr($body, $offset, $needle){
 685     $me = "sq_findnxstr";
 686     $pos = strpos($body, $needle, $offset);
 687     if ($pos === FALSE){
 688         $pos = strlen($body);
 689     }
 690     return $pos;
 691 }
 692
 693 /**
 694  * This function takes a PCRE-style regexp and tries to match it
 695  * within the string.
 696  *
 697  * @param  $body   The string to look for needle in.
 698  * @param  $offset Start looking from here.
 699  * @param  $reg    A PCRE-style regex to match.
 700  * @return         Returns a false if no matches found, or an array
 701  *                 with the following members:
 702  *                 - integer with the location of the match within $body
 703  *                 - string with whatever content between offset and the match
 704  *                 - string with whatever it is we matched
 705  */
 706 function sq_findnxreg($body, $offset, $reg){
 707     $me = "sq_findnxreg";
 708     $matches = Array();
 709     $retarr = Array();
 710     preg_match("%^(.*?)($reg)%s", substr($body, $offset), $matches);
 711     if (!$matches{0}){
 712         $retarr = false;
 713     } else {
 714         $retarr{0} = $offset + strlen($matches{1});
 715         $retarr{1} = $matches{1};
 716         $retarr{2} = $matches{2};
 717     }
 718     return $retarr;
 719 }
 720
 721 /**
 722  * This function looks for the next tag.
 723  *
 724  * @param  $body   String where to look for the next tag.
 725  * @param  $offset Start looking from here.
 726  * @return         false if no more tags exist in the body, or
 727  *                 an array with the following members:
 728  *                 - string with the name of the tag
 729  *                 - array with attributes and their values
 730  *                 - integer with tag type (1, 2, or 3)
 731  *                 - integer where the tag starts (starting "<")
 732  *                 - integer where the tag ends (ending ">")
 733  *                 first three members will be false, if the tag is invalid.
 734  */
 735 function sq_getnxtag($body, $offset){
 736     $me = "sq_getnxtag";
 737     if ($offset > strlen($body)){
 738         return false;
 739     }
 740     $lt = sq_findnxstr($body, $offset, "<");
 741     if ($lt == strlen($body)){
 742         return false;
 743     }
 744     /**
 745      * We are here:
 746      * blah blah <tag attribute="value">
 747      * \---------^
 748      */
 749     $pos = sq_skipspace($body, $lt+1);
 750     if ($pos >= strlen($body)){
 751         return Array(false, false, false, $lt, strlen($body));
 752     }
 753     /**
 754      * There are 3 kinds of tags:
 755      * 1. Opening tag, e.g.:
 756      *    <a href="blah">
 757      * 2. Closing tag, e.g.:
 758      *    </a>
 759      * 3. XHTML-style content-less tag, e.g.:
 760      *    <img src="blah"/>
 761      */
 762     $tagtype = false;
 763     switch (substr($body, $pos, 1)){
 764     case "/":
 765         $tagtype = 2;
 766         $pos++;
 767         break;
 768     case "!":
 769         /**
 770          * A comment or an SGML declaration.
 771          */
 772         if (substr($body, $pos+1, 2) == "--"){
 773             $gt = strpos($body, "-->", $pos);
 774             if ($gt === false){
 775                 $gt = strlen($body);
 776             } else {
 777                 $gt += 2;
 778             }
 779             return Array(false, false, false, $lt, $gt);
 780         } else {
 781             $gt = sq_findnxstr($body, $pos, ">");
 782             return Array(false, false, false, $lt, $gt);
 783         }
 784         break;
 785     default:
 786         /**
 787          * Assume tagtype 1 for now. If it's type 3, we'll switch values
 788          * later.
 789          */
 790         $tagtype = 1;
 791         break;
 792     }
 793
 794     $tag_start = $pos;
 795     $tagname = '';
 796     /**
 797      * Look for next [\W-_], which will indicate the end of the tag name.
 798      */
 799     $regary = sq_findnxreg($body, $pos, "[^\w\-_]");
 800     if ($regary == false){
 801         return Array(false, false, false, $lt, strlen($body));
 802     }
 803     list($pos, $tagname, $match) = $regary;
 804     $tagname = strtolower($tagname);
 805
 806     /**
 807      * $match can be either of these:
 808      * '>'  indicating the end of the tag entirely.
 809      * '\s' indicating the end of the tag name.
 810      * '/'  indicating that this is type-3 xhtml tag.
 811      *
 812      * Whatever else we find there indicates an invalid tag.
 813      */
 814     switch ($match){
 815     case "/":
 816         /**
 817          * This is an xhtml-style tag with a closing / at the
 818          * end, like so: <img src="blah"/>. Check if it's followed
 819          * by the closing bracket. If not, then this tag is invalid
 820          */
 821         if (substr($body, $pos, 2) == "/>"){
 822             $pos++;
 823             $tagtype = 3;
 824         } else {
 825             $gt = sq_findnxstr($body, $pos, ">");
 826             $retary = Array(false, false, false, $lt, $gt);
 827             return $retary;
 828         }
 829     case ">":
 830         return Array($tagname, false, $tagtype, $lt, $pos);
 831         break;
 832     default:
 833         /**
 834          * Check if it's whitespace
 835          */
 836         if (preg_match("/\s/", $match)){
 837         } else {
 838             /**
 839              * This is an invalid tag! Look for the next closing ">".
 840              */
 841             $gt = sq_findnxstr($body, $offset, ">");
 842             return Array(false, false, false, $lt, $gt);
 843         }
 844     }
 845
 846     /**
 847      * At this point we're here:
 848      * <tagname  attribute='blah'>
 849      * \-------^
 850      *
 851      * At this point we loop in order to find all attributes.
 852      */
 853     $attname = '';
 854     $atttype = false;
 855     $attary = Array();
 856
 857     while ($pos <= strlen($body)){
 858         $pos = sq_skipspace($body, $pos);
 859         if ($pos == strlen($body)){
 860             /**
 861              * Non-closed tag.
 862              */
 863             return Array(false, false, false, $lt, $pos);
 864         }
 865         /**
 866          * See if we arrived at a ">" or "/>", which means that we reached
 867          * the end of the tag.
 868          */
 869         $matches = Array();
 870         if (preg_match("%^(\s*)(>|/>)%s", substr($body, $pos), $matches)) {
 871             /**
 872              * Yep. So we did.
 873              */
 874             $pos += strlen($matches{1});
 875             if ($matches{2} == "/>"){
 876                 $tagtype = 3;
 877                 $pos++;
 878             }
 879             return Array($tagname, $attary, $tagtype, $lt, $pos);
 880         }
 881
 882         /**
 883          * There are several types of attributes, with optional
 884          * [:space:] between members.
 885          * Type 1:
 886          *   attrname[:space:]=[:space:]'CDATA'
 887          * Type 2:
 888          *   attrname[:space:]=[:space:]"CDATA"
 889          * Type 3:
 890          *   attr[:space:]=[:space:]CDATA
 891          * Type 4:
 892          *   attrname
 893          *
 894          * We leave types 1 and 2 the same, type 3 we check for
 895          * '"' and convert to "&quot" if needed, then wrap in
 896          * double quotes. Type 4 we convert into:
 897          * attrname="yes".
 898          */
 899         $regary = sq_findnxreg($body, $pos, "[^\w\-_]");
 900         if ($regary == false){
 901             /**
 902              * Looks like body ended before the end of tag.
 903              */
 904             return Array(false, false, false, $lt, strlen($body));
 905         }
 906         list($pos, $attname, $match) = $regary;
 907         $attname = strtolower($attname);
 908         /**
 909          * We arrived at the end of attribute name. Several things possible
 910          * here:
 911          * '>'  means the end of the tag and this is attribute type 4
 912          * '/'  if followed by '>' means the same thing as above
 913          * '\s' means a lot of things -- look what it's followed by.
 914          *      anything else means the attribute is invalid.
 915          */
 916         switch($match){
 917         case "/":
 918             /**
 919              * This is an xhtml-style tag with a closing / at the
 920              * end, like so: <img src="blah"/>. Check if it's followed
 921              * by the closing bracket. If not, then this tag is invalid
 922              */
 923             if (substr($body, $pos, 2) == "/>"){
 924                 $pos++;
 925                 $tagtype = 3;
 926             } else {
 927                 $gt = sq_findnxstr($body, $pos, ">");
 928                 $retary = Array(false, false, false, $lt, $gt);
 929                 return $retary;
 930             }
 931         case ">":
 932             $attary{$attname} = '"yes"';
 933             return Array($tagname, $attary, $tagtype, $lt, $pos);
 934             break;
 935         default:
 936             /**
 937              * Skip whitespace and see what we arrive at.
 938              */
 939             $pos = sq_skipspace($body, $pos);
 940             $char = substr($body, $pos, 1);
 941             /**
 942              * Two things are valid here:
 943              * '=' means this is attribute type 1 2 or 3.
 944              * \w means this was attribute type 4.
 945              * anything else we ignore and re-loop. End of tag and
 946              * invalid stuff will be caught by our checks at the beginning
 947              * of the loop.
 948              */
 949             if ($char == "="){
 950                 $pos++;
 951                 $pos = sq_skipspace($body, $pos);
 952                 /**
 953                  * Here are 3 possibilities:
 954                  * "'"  attribute type 1
 955                  * '"'  attribute type 2
 956                  * everything else is the content of tag type 3
 957                  */
 958                 $quot = substr($body, $pos, 1);
 959                 if ($quot == "'"){
 960                     $regary = sq_findnxreg($body, $pos+1, "\'");
 961                     if ($regary == false){
 962                         return Array(false, false, false, $lt, strlen($body));
 963                     }
 964                     list($pos, $attval, $match) = $regary;
 965                     $pos++;
 966                     $attary{$attname} = "'" . $attval . "'";
 967                 } else if ($quot == '"'){
 968                     $regary = sq_findnxreg($body, $pos+1, '\"');
 969                     if ($regary == false){
 970                         return Array(false, false, false, $lt, strlen($body));
 971                     }
 972                     list($pos, $attval, $match) = $regary;
 973                     $pos++;
 974                     $attary{$attname} = '"' . $attval . '"';
 975                 } else {
 976                     /**
 977                      * These are hateful. Look for \s, or >.
 978                      */
 979                     $regary = sq_findnxreg($body, $pos, "[\s>]");
 980                     if ($regary == false){
 981                         return Array(false, false, false, $lt, strlen($body));
 982                     }
 983                     list($pos, $attval, $match) = $regary;
 984                     /**
 985                      * If it's ">" it will be caught at the top.
 986                      */
 987                     $attval = preg_replace("/\"/s", "&quot;", $attval);
 988                     $attary{$attname} = '"' . $attval . '"';
 989                 }
 990             } else if (preg_match("|[\w/>]|", $char)) {
 991                 /**
 992                  * That was attribute type 4.
 993                  */
 994                 $attary{$attname} = '"yes"';
 995             } else {
 996                 /**
 997                  * An illegal character. Find next '>' and return.
 998                  */
 999                 $gt = sq_findnxstr($body, $pos, ">");
1000                 return Array(false, false, false, $lt, $gt);
1001             }
1002         }
1003     }
1004     /**
1005      * The fact that we got here indicates that the tag end was never
1006      * found. Return invalid tag indication so it gets stripped.
1007      */
1008     return Array(false, false, false, $lt, strlen($body));
1009 }
1010
1011 /**
1012  * This function checks attribute values for entity-encoded values
1013  * and returns them translated into 8-bit strings so we can run
1014  * checks on them.
1015  *
1016  * @param  $attvalue A string to run entity check against.
1017  * @return           Translated value.
1018  */
1019 function sq_deent($attvalue){
1020     $me="sq_deent";
1021     /**
1022      * See if we have to run the checks first. All entities must start
1023      * with "&".
1024      */
1025     if (strpos($attvalue, "&") === false){
1026         return $attvalue;
1027     }
1028     /**
1029      * Check named entities first.
1030      */
1031     $trans = get_html_translation_table(HTML_ENTITIES);
1032     /**
1033      * Leave &quot; in, as it can mess us up.
1034      */
1035     $trans = array_flip($trans);
1036     unset($trans{"&quot;"});
1037     while (list($ent, $val) = each($trans)){
1038         $attvalue = preg_replace("/$ent*(\W)/si", "$val\\1", $attvalue);
1039     }
1040     /**
1041      * Now translate numbered entities from 1 to 255 if needed.
1042      */
1043     if (strpos($attvalue, "#") !== false){
1044         $omit = Array(34, 39);
1045         for ($asc=1; $asc<256; $asc++){
1046             if (!in_array($asc, $omit)){
1047                 $chr = chr($asc);
1048                 $attvalue = preg_replace("/\&#0*$asc;*(\D)/si", "$chr\\1",
1049                                          $attvalue);
1050                 $attvalue = preg_replace("/\&#x0*".dechex($asc).";*(\W)/si",
1051                                          "$chr\\1", $attvalue);
1052             }
1053         }
1054     }
1055     return $attvalue;
1056 }
1057
1058 /**
1059  * This function runs various checks against the attributes.
1060  *
1061  * @param  $tagname         String with the name of the tag.
1062  * @param  $attary          Array with all tag attributes.
1063  * @param  $rm_attnames     See description for sq_sanitize
1064  * @param  $bad_attvals     See description for sq_sanitize
1065  * @param  $add_attr_to_tag See description for sq_sanitize
1066  * @param  $message         message object
1067  * @param  $id              message id
1068  * @return                  Array with modified attributes.
1069  */
1070 function sq_fixatts($tagname,
1071                     $attary,
1072                     $rm_attnames,
1073                     $bad_attvals,
1074                     $add_attr_to_tag,
1075                     $message,
1076                     $id,
1077                     $mailbox
1078                     ){
1079     $me = "sq_fixatts";
1080     while (list($attname, $attvalue) = each($attary)){
1081         /**
1082          * See if this attribute should be removed.
1083          */
1084         foreach ($rm_attnames as $matchtag=>$matchattrs){
1085             if (preg_match($matchtag, $tagname)){
1086                 foreach ($matchattrs as $matchattr){
1087                     if (preg_match($matchattr, $attname)){
1088                         unset($attary{$attname});
1089                         continue;
1090                     }
1091                 }
1092             }
1093         }
1094         /**
1095          * Remove any entities.
1096          */
1097         $attvalue = sq_deent($attvalue);
1098
1099         /**
1100          * Now let's run checks on the attvalues.
1101          * I don't expect anyone to comprehend this. If you do,
1102          * get in touch with me so I can drive to where you live and
1103          * shake your hand personally. :)
1104          */
1105         foreach ($bad_attvals as $matchtag=>$matchattrs){
1106             if (preg_match($matchtag, $tagname)){
1107                 foreach ($matchattrs as $matchattr=>$valary){
1108                     if (preg_match($matchattr, $attname)){
1109                         /**
1110                          * There are two arrays in valary.
1111                          * First is matches.
1112                          * Second one is replacements
1113                          */
1114                         list($valmatch, $valrepl) = $valary;
1115                         $newvalue =
1116                             preg_replace($valmatch, $valrepl, $attvalue);
1117                         if ($newvalue != $attvalue){
1118                             $attary{$attname} = $newvalue;
1119                         }
1120                     }
1121                 }
1122             }
1123         }
1124         /**
1125          * Turn cid: urls into http-friendly ones.
1126          */
1127         if (preg_match("/^[\'\"]\s*cid:/si", $attvalue)){
1128             $attary{$attname} = sq_cid2http($message, $id, $attvalue, $mailbox);
1129         }
1130     }
1131     /**
1132      * See if we need to append any attributes to this tag.
1133      */
1134     foreach ($add_attr_to_tag as $matchtag=>$addattary){
1135         if (preg_match($matchtag, $tagname)){
1136             $attary = array_merge($attary, $addattary);
1137         }
1138     }
1139     return $attary;
1140 }
1141
1142 /**
1143  * This function edits the style definition to make them friendly and
1144  * usable in squirrelmail.
1145  *
1146  * @param  $message  the message object
1147  * @param  $id       the message id
1148  * @param  $content  a string with whatever is between <style> and </style>
1149  * @return           a string with edited content.
1150  */
1151 function sq_fixstyle($message, $id, $content){
1152     global $view_unsafe_images;
1153     $me = "sq_fixstyle";
1154     /**
1155      * First look for general BODY style declaration, which would be
1156      * like so:
1157      * body {background: blah-blah}
1158      * and change it to .bodyclass so we can just assign it to a <div>
1159      */
1160     $content = preg_replace("|body(\s*\{.*?\})|si", ".bodyclass\\1", $content);
1161     $secremoveimg = "../images/" . _("sec_remove_eng.png");
1162     /**
1163      * Fix url('blah') declarations.
1164      */
1165     $content = preg_replace("|url\(([\'\"])\s*\S+script\s*:.*?([\'\"])\)|si",
1166                             "url(\\1$secremoveimg\\2)", $content);
1167     /**
1168      * Fix url('https*://.*) declarations but only if $view_unsafe_images
1169      * is false.
1170      */
1171     if (!$view_unsafe_images){
1172         $content = preg_replace("|url\(([\'\"])\s*https*:.*?([\'\"])\)|si",
1173                                 "url(\\1$secremoveimg\\2)", $content);
1174     }
1175
1176     /**
1177      * Fix urls that refer to cid:
1178      */
1179     while (preg_match("|url\(([\'\"]\s*cid:.*?[\'\"])\)|si", $content,
1180                       $matches)){
1181         $cidurl = $matches{1};
1182         $httpurl = sq_cid2http($message, $id, $cidurl);
1183         $content = preg_replace("|url\($cidurl\)|si",
1184                                 "url($httpurl)", $content);
1185     }
1186
1187     /**
1188      * Fix stupid css declarations which lead to vulnerabilities
1189      * in IE.
1190      */
1191     $match   = Array('/expression/si',
1192                      '/behaviou*r/si',
1193                      '/binding/si');
1194     $replace = Array('idiocy', 'idiocy', 'idiocy');
1195     $content = preg_replace($match, $replace, $content);
1196     return $content;
1197 }
1198
1199 /**
1200  * This function converts cid: url's into the ones that can be viewed in
1201  * the browser.
1202  *
1203  * @param  $message  the message object
1204  * @param  $id       the message id
1205  * @param  $cidurl   the cid: url.
1206  * @return           a string with a http-friendly url
1207  */
1208 function sq_cid2http($message, $id, $cidurl, $mailbox){
1209     /**
1210      * Get rid of quotes.
1211      */
1212     $quotchar = substr($cidurl, 0, 1);
1213     $cidurl = str_replace($quotchar, "", $cidurl);
1214     $cidurl = substr(trim($cidurl), 4);
1215     $httpurl = $quotchar . "../src/download.php?absolute_dl=true&amp;" .
1216         "passed_id=$id&amp;mailbox=" . urlencode($mailbox) .
1217         "&amp;ent_id=" . find_ent_id($cidurl, $message) . $quotchar;
1218     return $httpurl;
1219 }
1220
1221 /**
1222  * This function changes the <body> tag into a <div> tag since we
1223  * can't really have a body-within-body.
1224  *
1225  * @param  $attary  an array of attributes and values of <body>
1226  * @return          a modified array of attributes to be set for <div>
1227  */
1228 function sq_body2div($attary){
1229     $me = "sq_body2div";
1230     $divattary = Array("class"=>"'bodyclass'");
1231     $bgcolor="#ffffff";
1232     $text="#000000";
1233     $styledef="";
1234     if (is_array($attary) && sizeof($attary) > 0){
1235         foreach ($attary as $attname=>$attvalue){
1236             $quotchar = substr($attvalue, 0, 1);
1237             $attvalue = str_replace($quotchar, "", $attvalue);
1238             switch ($attname){
1239             case "background":
1240                 $styledef .= "background-image: url('$attvalue'); ";
1241                 break;
1242             case "bgcolor":
1243                 $styledef .= "background-color: $attvalue; ";
1244                 break;
1245             case "text":
1246                 $styledef .= "color: $attvalue; ";
1247             }
1248         }
1249         if (strlen($styledef) > 0){
1250             $divattary{"style"} = "\"$styledef\"";
1251         }
1252     }
1253     return $divattary;
1254 }
1255
1256 /**
1257  * This is the main function and the one you should actually be calling.
1258  * There are several variables you should be aware of an which need
1259  * special description.
1260  *
1261  * Since the description is quite lengthy, see it here:
1262  * http://www.mricon.com/html/phpfilter.html
1263  *
1264  * @param $body                 the string with HTML you wish to filter
1265  * @param $tag_list             see description above
1266  * @param $rm_tags_with_content see description above
1267  * @param $self_closing_tags    see description above
1268  * @param $force_tag_closing    see description above
1269  * @param $rm_attnames          see description above
1270  * @param $bad_attvals          see description above
1271  * @param $add_attr_to_tag      see description above
1272  * @param $message              message object
1273  * @param $id                   message id
1274  * @return                      sanitized html safe to show on your pages.
1275  */
1276 function sq_sanitize($body,
1277                      $tag_list,
1278                      $rm_tags_with_content,
1279                      $self_closing_tags,
1280                      $force_tag_closing,
1281                      $rm_attnames,
1282                      $bad_attvals,
1283                      $add_attr_to_tag,
1284                      $message,
1285                      $id,
1286                      $mailbox
1287                      ){
1288     $me = "sq_sanitize";
1289     /**
1290      * Normalize rm_tags and rm_tags_with_content.
1291      */
1292     @array_walk($rm_tags, 'sq_casenormalize');
1293     @array_walk($rm_tags_with_content, 'sq_casenormalize');
1294     @array_walk($self_closing_tags, 'sq_casenormalize');
1295     /**
1296      * See if tag_list is of tags to remove or tags to allow.
1297      * false  means remove these tags
1298      * true   means allow these tags
1299      */
1300     $rm_tags = array_shift($tag_list);
1301     $curpos = 0;
1302     $open_tags = Array();
1303     $trusted = "<!-- begin sanitized html -->\n";
1304     $skip_content = false;
1305     /**
1306      * Take care of netscape's stupid javascript entities like
1307      * &{alert('boo')};
1308      */
1309     $body = preg_replace("/&(\{.*?\};)/si", "&amp;\\1", $body);
1310
1311     while (($curtag=sq_getnxtag($body, $curpos)) != FALSE){
1312         list($tagname, $attary, $tagtype, $lt, $gt) = $curtag;
1313         $free_content = substr($body, $curpos, $lt-$curpos);
1314         /**
1315          * Take care of <style>
1316          */
1317         if ($tagname == "style" && $tagtype == 2){
1318             /**
1319              * This is a closing </style>. Edit the
1320              * content before we apply it.
1321              */
1322             $free_content = sq_fixstyle($message, $id, $free_content);
1323         }
1324         if ($skip_content == false){
1325             $trusted .= $free_content;
1326         } else {
1327         }
1328         if ($tagname != FALSE){
1329             if ($tagtype == 2){
1330                 if ($skip_content == $tagname){
1331                     /**
1332                      * Got to the end of tag we needed to remove.
1333                      */
1334                     $tagname = false;
1335                     $skip_content = false;
1336                 } else {
1337                     if ($skip_content == false){
1338                         if ($tagname == "body"){
1339                             $tagname = "div";
1340                         } else {
1341                             if (isset($open_tags{$tagname}) &&
1342                                 $open_tags{$tagname} > 0){
1343                                 $open_tags{$tagname}--;
1344                             } else {
1345                                 $tagname = false;
1346                             }
1347                         }
1348                     } else {
1349                     }
1350                 }
1351             } else {
1352                 /**
1353                  * $rm_tags_with_content
1354                  */
1355                 if ($skip_content == false){
1356                     /**
1357                      * See if this is a self-closing type and change
1358                      * tagtype appropriately.
1359                      */
1360                     if ($tagtype == 1
1361                         && in_array($tagname, $self_closing_tags)){
1362                         $tagtype=3;
1363                     }
1364                     /**
1365                      * See if we should skip this tag and any content
1366                      * inside it.
1367                      */
1368                     if ($tagtype == 1 &&
1369                         in_array($tagname, $rm_tags_with_content)){
1370                         $skip_content = $tagname;
1371                     } else {
1372                         if (($rm_tags == false
1373                              && in_array($tagname, $tag_list)) ||
1374                             ($rm_tags == true &&
1375                              !in_array($tagname, $tag_list))){
1376                             $tagname = false;
1377                         } else {
1378                             if ($tagtype == 1){
1379                                 if (isset($open_tags{$tagname})){
1380                                     $open_tags{$tagname}++;
1381                                 } else {
1382                                     $open_tags{$tagname}=1;
1383                                 }
1384                             }
1385                             /**
1386                              * This is where we run other checks.
1387                              */
1388                             if (is_array($attary) && sizeof($attary) > 0){
1389                                 $attary = sq_fixatts($tagname,
1390                                                      $attary,
1391                                                      $rm_attnames,
1392                                                      $bad_attvals,
1393                                                      $add_attr_to_tag,
1394                                                      $message,
1395                                                      $id,
1396                                                      $mailbox
1397                                                      );
1398                             }
1399                             /**
1400                              * Convert body into div.
1401                              */
1402                             if ($tagname == "body"){
1403                                 $tagname = "div";
1404                                 $attary = sq_body2div($attary, $message, $id);
1405                             }
1406                         }
1407                     }
1408                 } else {
1409                 }
1410             }
1411             if ($tagname != false && $skip_content == false){
1412                 $trusted .= sq_tagprint($tagname, $attary, $tagtype);
1413             }
1414         } else {
1415         }
1416         $curpos = $gt+1;
1417     }
1418     $trusted .= substr($body, $curpos, strlen($body)-$curpos);
1419     if ($force_tag_closing == true){
1420         foreach ($open_tags as $tagname=>$opentimes){
1421             while ($opentimes > 0){
1422                 $trusted .= '</' . $tagname . '>';
1423                 $opentimes--;
1424             }
1425         }
1426         $trusted .= "\n";
1427     }
1428     $trusted .= "<!-- end sanitized html -->\n";
1429     return $trusted;
1430 }
1431
1432 /**
1433  * This is a wrapper function to call html sanitizing routines.
1434  *
1435  * @param  $body  the body of the message
1436  * @param  $id    the id of the message
1437  * @return        a string with html safe to display in the browser.
1438  */
1439 function magicHTML($body, $id, $message, $mailbox = 'INBOX'){
1440     global $attachment_common_show_images, $view_unsafe_images,
1441         $has_unsafe_images;
1442     /**
1443      * Don't display attached images in HTML mode.
1444      */
1445     $attachment_common_show_images = false;
1446     $tag_list = Array(
1447                       false,
1448                       "object",
1449                       "meta",
1450                       "html",
1451                       "head",
1452                       "base",
1453                       "link"
1454                       );
1455
1456     $rm_tags_with_content = Array(
1457                                   "script",
1458                                   "applet",
1459                                   "embed",
1460                                   "title"
1461                                   );
1462
1463     $self_closing_tags =  Array(
1464                                 "img",
1465                                 "br",
1466                                 "hr",
1467                                 "input"
1468                                 );
1469
1470     $force_tag_closing = false;
1471
1472     $rm_attnames = Array(
1473                          "/.*/" =>
1474                          Array(
1475                                "/target/si",
1476                                "/^on.*/si",
1477                                "/^dynsrc/si",
1478                                "/^data.*/si"
1479                                )
1480                          );
1481
1482     $secremoveimg = "../images/" . _("sec_remove_eng.png");
1483     $bad_attvals = Array(
1484         "/.*/" =>
1485             Array(
1486                 "/^src|background/i" =>
1487                     Array(
1488                           Array(
1489                                 "|^([\'\"])\s*\.\./.*([\'\"])|si",
1490                                 "/^([\'\"])\s*\S+script\s*:.*([\'\"])/si",
1491                                 "/^([\'\"])\s*mocha\s*:*.*([\'\"])/si",
1492                                 "/^([\'\"])\s*about\s*:.*([\'\"])/si"
1493                                 ),
1494                           Array(
1495                                 "\\1$secremoveimg\\2",
1496                                 "\\1$secremoveimg\\2",
1497                                 "\\1$secremoveimg\\2",
1498                                 "\\1$secremoveimg\\2"
1499                                 )
1500                         ),
1501                 "/^href|action/i" =>
1502                     Array(
1503                           Array(
1504                                 "|^([\'\"])\s*\.\./.*([\'\"])|si",
1505                                 "/^([\'\"])\s*\S+script\s*:.*([\'\"])/si",
1506                                 "/^([\'\"])\s*mocha\s*:*.*([\'\"])/si",
1507                                 "/^([\'\"])\s*about\s*:.*([\'\"])/si"
1508                                 ),
1509                           Array(
1510                                 "\\1#\\2",
1511                                 "\\1#\\2",
1512                                 "\\1#\\2",
1513                                 "\\1#\\2"
1514                                 )
1515                         ),
1516                 "/^style/si" =>
1517                     Array(
1518                           Array(
1519                                 "/expression/si",
1520                                 "/binding/si",
1521                                 "/behaviou*r/si",
1522                                 "|url\(([\'\"])\s*\.\./.*([\'\"])\)|si",
1523                                 "/url\(([\'\"])\s*\S+script\s*:.*([\'\"])\)/si",
1524                                 "/url\(([\'\"])\s*mocha\s*:.*([\'\"])\)/si",
1525                                 "/url\(([\'\"])\s*about\s*:.*([\'\"])\)/si"
1526                                ),
1527                           Array(
1528                                 "idiocy",
1529                                 "idiocy",
1530                                 "idiocy",
1531                                 "url(\\1#\\2)",
1532                                 "url(\\1#\\2)",
1533                                 "url(\\1#\\2)",
1534                                 "url(\\1#\\2)"
1535                                )
1536                           )
1537                 )
1538         );
1539     if (!$view_unsafe_images){
1540         /**
1541          * Remove any references to http/https if view_unsafe_images set
1542          * to false.
1543          */
1544          array_push($bad_attvals{'/.*/'}{'/^src|background/i'}[0],
1545                     '/^([\'\"])\s*https*:.*([\'\"])/si');
1546          array_push($bad_attvals{'/.*/'}{'/^src|background/i'}[1],
1547                     "\\1$secremoveimg\\2");
1548          array_push($bad_attvals{'/.*/'}{'/^style/si'}[0],
1549                     '/url\(([\'\"])\s*https*:.*([\'\"])\)/si');
1550          array_push($bad_attvals{'/.*/'}{'/^style/si'}[1],
1551                     "url(\\1$secremoveimg\\2)");
1552     }
1553
1554     $add_attr_to_tag = Array(
1555                              "/^a$/si" => Array('target'=>'"_new"')
1556                              );
1557     $trusted = sq_sanitize($body,
1558                            $tag_list,
1559                            $rm_tags_with_content,
1560                            $self_closing_tags,
1561                            $force_tag_closing,
1562                            $rm_attnames,
1563                            $bad_attvals,
1564                            $add_attr_to_tag,
1565                            $message,
1566                            $id,
1567                            $mailbox
1568                            );
1569     if (preg_match("|$secremoveimg|si", $trusted)){
1570         $has_unsafe_images = true;
1571     }
1572     return $trusted;
1573 }
1574
1575 ?>