functions/mime.php

   1 <?php
   2
   3 /**
   4  * mime.php
   5  *
   6  * Copyright (c) 1999-2002 The SquirrelMail Project Team
   7  * Licensed under the GNU GPL. For full terms see the file COPYING.
   8  *
   9  * This contains the functions necessary to detect and decode MIME
  10  * messages.
  11  *
  12  * $Id$
  13  */
  14
  15 require_once('../functions/imap.php');
  16 require_once('../functions/attachment_common.php');
  17
  18 /* --------------------------------------------------------------------------------- */
  19 /* MIME DECODING                                                                     */
  20 /* --------------------------------------------------------------------------------- */
  21
  22 /* This function gets the structure of a message and stores it in the "message" class.
  23  * It will return this object for use with all relevant header information and
  24  * fully parsed into the standard "message" object format.
  25  */
  26
  27 function mime_structure ($bodystructure, $flags=array()) {
  28
  29     // isolate the body structure and remove beginning and end parenthesis
  30     $read = trim(substr ($bodystructure, strpos(strtolower($bodystructure), 'bodystructure') + 13));
  31     $msg = &new message();
  32     $read = trim(substr ($read, 0, -1));
  33     $msg = $msg->parseStructure($read,0);
  34     $msg->setEnt('0');
  35     if (count($flags)) {
  36        foreach ($flags as $flag) {
  37           $char = strtoupper($flag{1});
  38           switch ($char) {
  39              case 'S':
  40                if (strtolower($flag) == '\\seen') {
  41                   $msg->is_seen = true;
  42                }
  43                break;
  44              case 'A':
  45                if (strtolower($flag) == '\\answered') {
  46                   $msg->is_answered = true;
  47                }
  48                break;
  49              case 'D':
  50                if (strtolower($flag) == '\\deleted') {
  51                   $msg->is_deleted = true;
  52                }
  53                break;
  54              case 'F':
  55                if (strtolower($flag) == '\\flagged') {
  56                   $msg->is_flagged = true;
  57                }
  58                break;
  59              case 'M':
  60                if (strtolower($flag) == '$mdnsent') {
  61                   $msg->is_mdnsent = true;
  62                }
  63                break;
  64              default:
  65                break;
  66           }
  67        }
  68     }
  69 //    listEntities($msg);
  70     return( $msg );
  71 }
  72
  73 /* this starts the parsing of a particular structure.  It is called recursively,
  74  * so it can be passed different structures.  It returns an object of type
  75  * $message.
  76  * First, it checks to see if it is a multipart message.  If it is, then it
  77  * handles that as it sees is necessary.  If it is just a regular entity,
  78  * then it parses it and adds the necessary header information (by calling out
  79  * to mime_get_elements()
  80  */
  81
  82 function mime_fetch_body($imap_stream, $id, $ent_id) {
  83     global $uid_support;
  84     /*
  85      * do a bit of error correction.  If we couldn't find the entity id, just guess
  86      * that it is the first one.  That is usually the case anyway.
  87      */
  88     if (!$ent_id) {
  89         $ent_id = 1;
  90     }
  91     $cmd = "FETCH $id BODY[$ent_id]";
  92
  93     $data = sqimap_run_command ($imap_stream, $cmd, true, $response, $message, $uid_support);
  94     do {
  95         $topline = trim(array_shift( $data ));
  96     } while( $topline && $topline[0] == '*' && !preg_match( '/\* [0-9]+ FETCH.*/i', $topline )) ;
  97
  98     $wholemessage = implode('', $data);
  99     if (ereg('\\{([^\\}]*)\\}', $topline, $regs)) {
 100
 101         $ret = substr( $wholemessage, 0, $regs[1] );
 102         /*
 103             There is some information in the content info header that could be important
 104             in order to parse html messages. Let's get them here.
 105         */
 106         if ( $ret{0} == '<' ) {
 107             $data = sqimap_run_command ($imap_stream, "FETCH $id BODY[$ent_id.MIME]", true, $response, $message, $uid_support);
 108         }
 109     } else if (ereg('"([^"]*)"', $topline, $regs)) {
 110         $ret = $regs[1];
 111     } else {
 112         global $where, $what, $mailbox, $passed_id, $startMessage;
 113         $par = 'mailbox=' . urlencode($mailbox) . "&amp;passed_id=$passed_id";
 114         if (isset($where) && isset($what)) {
 115             $par .= '&amp;where='. urlencode($where) . "&amp;what=" . urlencode($what);
 116         } else {
 117             $par .= "&amp;startMessage=$startMessage&amp;show_more=0";
 118         }
 119         $par .= '&amp;response=' . urlencode($response) .
 120                 '&amp;message=' . urlencode($message).
 121                 '&amp;topline=' . urlencode($topline);
 122
 123         echo   '<tt><br>' .
 124                '<table width="80%"><tr>' .
 125                '<tr><td colspan=2>' .
 126                _("Body retrieval error. The reason for this is most probably that the message is malformed. Please help us making future versions better by submitting this message to the developers knowledgebase!") .
 127                " <A HREF=\"../src/retrievalerror.php?$par\"><br>" .
 128                _("Submit message") . '</A><BR>&nbsp;' .
 129                '</td></tr>' .
 130                '<td><b>' . _("Command:") . "</td><td>$cmd</td></tr>" .
 131                '<td><b>' . _("Response:") . "</td><td>$response</td></tr>" .
 132                '<td><b>' . _("Message:") . "</td><td>$message</td></tr>" .
 133                '<td><b>' . _("FETCH line:") . "</td><td>$topline</td></tr>" .
 134                "</table><BR></tt></font><hr>";
 135
 136         $data = sqimap_run_command ($imap_stream, "FETCH $passed_id BODY[]", true, $response, $message, $uid_support);
 137         array_shift($data);
 138         $wholemessage = implode('', $data);
 139
 140         $ret = $wholemessage;
 141     }
 142     return( $ret );
 143 }
 144
 145 function mime_print_body_lines ($imap_stream, $id, $ent_id, $encoding) {
 146     global $uid_support;
 147     // do a bit of error correction.  If we couldn't find the entity id, just guess
 148     // that it is the first one.  That is usually the case anyway.
 149     if (!$ent_id) {
 150         $ent_id = 1;
 151     }
 152     $sid = sqimap_session_id($uid_support);
 153     // Don't kill the connection if the browser is over a dialup
 154     // and it would take over 30 seconds to download it.
 155
 156     // don´t call set_time_limit in safe mode.
 157     if (!ini_get("safe_mode")) {
 158         set_time_limit(0);
 159     }
 160     if ($uid_support) {
 161        $sid_s = substr($sid,0,strpos($sid, ' '));
 162     } else {
 163        $sid_s = $sid;
 164     }
 165
 166     fputs ($imap_stream, "$sid FETCH $id BODY[$ent_id]\r\n");
 167     $cnt = 0;
 168     $continue = true;
 169     $read = fgets ($imap_stream,8192);
 170
 171
 172     // This could be bad -- if the section has sqimap_session_id() . ' OK'
 173     // or similar, it will kill the download.
 174     while (!ereg("^".$sid_s." (OK|BAD|NO)(.*)$", $read, $regs)) {
 175       if (trim($read) == ')==') {
 176           $read1 = $read;
 177           $read = fgets ($imap_stream,4096);
 178           if (ereg("^".$sid." (OK|BAD|NO)(.*)$", $read, $regs)) {
 179               return;
 180           } else {
 181               echo decodeBody($read1, $encoding) .
 182                    decodeBody($read, $encoding);
 183           }
 184       } else if ($cnt) {
 185           echo decodeBody($read, $encoding);
 186       }
 187       $read = fgets ($imap_stream,4096);
 188       $cnt++;
 189 //      break;
 190     }
 191 }
 192
 193 /* -[ END MIME DECODING ]----------------------------------------------------------- */
 194
 195
 196 /* findDisplayEntity
 197  * Checks to see if $message contains content of type $type0/$type1
 198  * returns the first entity number it finds of that type, or NULL if
 199  * none is found.  Takes optional argument $start to allow the caller
 200  * to continue where they left off
 201  */
 202 function findDisplayEntity($message, $type0, $type1, $start=0) {
 203     if ($message) {
 204         for ($i = $start;isset($message->entities[$i]); $i++) {
 205             $entity = $message->entities[$i];
 206             if ($entity->type0 == $type0 && $entity->type1 == $type1) {
 207                 return $i;
 208             }
 209         }
 210     }
 211     return NULL;
 212 }
 213 // This is here for debugging purposese.  It will print out a list
 214 // of all the entity IDs that are in the $message object.
 215
 216 function listEntities ($message) {
 217   if ($message) {
 218     echo "<tt>" . $message->entity_id . ' : ' . $message->type0 . '/' . $message->type1 . ' parent = '. $message->parent->entity_id. '<br>';
 219     for ($i = 0;isset($message->entities[$i]); $i++) {
 220        echo "$i : ";
 221        $msg = listEntities($message->entities[$i]);
 222
 223        if ($msg) {
 224           echo "return: ";
 225           return $msg;
 226        }
 227     }
 228   }
 229 }
 230
 231 function getPriorityStr($priority) {
 232    $priority_level = substr($priority,0,1);
 233
 234    switch($priority_level) {
 235      /* check for a higher then normal priority. */
 236      case '1':
 237      case '2':
 238         $priority_string = _("High");
 239         break;
 240
 241       /* check for a lower then normal priority. */
 242      case '4':
 243      case '5':
 244         $priority_string = _("Low");
 245         break;
 246
 247      /* check for a normal priority. */
 248      case '3':
 249      default:
 250         $priority_level = '3';
 251         $priority_string = _("Normal");
 252         break;
 253
 254    }
 255    return $priority_string;
 256 }
 257
 258 /* returns a $message object for a particular entity id */
 259 function getEntity ($message, $ent_id) {
 260     return $message->getEntity($ent_id);
 261 }
 262
 263 /*
 264  * translateText
 265  * Extracted from strings.php 23/03/2002
 266  */
 267
 268 function translateText(&$body, $wrap_at, $charset) {
 269     global $where, $what; /* from searching */
 270     global $color; /* color theme */
 271
 272     require_once('../functions/url_parser.php');
 273
 274     $body_ary = explode("\n", $body);
 275     $PriorQuotes = 0;
 276     for ($i=0; $i < count($body_ary); $i++) {
 277         $line = $body_ary[$i];
 278         if (strlen($line) - 2 >= $wrap_at) {
 279             sqWordWrap($line, $wrap_at);
 280         }
 281         $line = charset_decode($charset, $line);
 282         $line = str_replace("\t", '        ', $line);
 283
 284         parseUrl ($line);
 285
 286         $Quotes = 0;
 287         $pos = 0;
 288         $j = strlen( $line );
 289
 290         while ( $pos < $j ) {
 291             if ($line[$pos] == ' ') {
 292                 $pos ++;
 293             } else if (strpos($line, '&gt;', $pos) === $pos) {
 294                 $pos += 4;
 295                 $Quotes ++;
 296             } else {
 297                 break;
 298             }
 299         }
 300
 301         if ($Quotes > 1) {
 302             if (! isset($color[14])) {
 303                 $color[14] = '#FF0000';
 304             }
 305             $line = '<FONT COLOR="' . $color[14] . '">' . $line . '</FONT>';
 306         } elseif ($Quotes) {
 307             if (! isset($color[13])) {
 308                 $color[13] = '#800000';
 309             }
 310             $line = '<FONT COLOR="' . $color[13] . '">' . $line . '</FONT>';
 311         }
 312
 313         $body_ary[$i] = $line;
 314     }
 315     $body = '<pre>' . implode("\n", $body_ary) . '</pre>';
 316 }
 317
 318
 319 /* This returns a parsed string called $body. That string can then
 320 be displayed as the actual message in the HTML. It contains
 321 everything needed, including HTML Tags, Attachments at the
 322 bottom, etc.
 323 */
 324 function formatBody($imap_stream, $message, $color, $wrap_at, $ent_num, $id, $mailbox='INBOX') {
 325     // this if statement checks for the entity to show as the
 326     // primary message. To add more of them, just put them in the
 327     // order that is their priority.
 328     global $startMessage, $username, $key, $imapServerAddress, $imapPort,
 329            $show_html_default, $has_unsafe_images, $view_unsafe_images, $sort;
 330
 331     $has_unsafe_images= 0;
 332     $body = '';
 333     $urlmailbox = urlencode($mailbox);
 334     $body_message = getEntity($message, $ent_num);
 335     if (($body_message->header->type0 == 'text') ||
 336         ($body_message->header->type0 == 'rfc822')) {
 337         $body = mime_fetch_body ($imap_stream, $id, $ent_num);
 338         $body = decodeBody($body, $body_message->header->encoding);
 339         $hookResults = do_hook("message_body", $body);
 340         $body = $hookResults[1];
 341
 342         // If there are other types that shouldn't be formatted, add
 343         // them here
 344         if ($body_message->header->type1 == 'html') {
 345             if ( $show_html_default <> 1 ) {
 346                 $body = strip_tags( $body );
 347                 translateText($body, $wrap_at,
 348                   $body_message->header->getParameter['charset']);
 349             } else {
 350                 $body = magicHTML( $body, $id, $message, $mailbox );
 351             }
 352         } else {
 353             translateText($body, $wrap_at,
 354                $body_message->header->getParameter('charset'));
 355         }
 356
 357         if ($has_unsafe_images) {
 358             if ($view_unsafe_images) {
 359                 $body .= "<CENTER><SMALL><A HREF=\"read_body.php?passed_id=$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";
 360             } else {
 361                 $body .= "<CENTER><SMALL><A HREF=\"read_body.php?passed_id=$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";
 362             }
 363         }
 364     }
 365     return ($body);
 366 }
 367
 368
 369 function formatAttachments($message, $exclude_id, $mailbox, $id) {
 370     global $where, $what;
 371     global $startMessage, $color;
 372     static $ShownHTML = 0;
 373
 374     $att_ar = $message->getAttachments($exclude_id);
 375
 376     if (!count($att_ar)) return '';
 377
 378     $attachments = "<TABLE WIDTH=\"100%\" CELLSPACING=0 CELLPADDING=2 BORDER=0 BGCOLOR=\"$color[0]\"><TR>\n" .
 379                 "<TH ALIGN=\"left\" BGCOLOR=\"$color[9]\"><B>\n" .
 380                 _("Attachments") . ':' .
 381                 "</B></TH></TR><TR><TD>\n" .
 382                 "<TABLE CELLSPACING=0 CELLPADDING=1 BORDER=0>\n";
 383
 384     $urlMailbox = urlencode($mailbox);
 385
 386     foreach ($att_ar as $att) {
 387
 388         $ent = urldecode($att->entity_id);
 389         $header = $att->header;
 390         $type0 = strtolower($header->type0);
 391         $type1 = strtolower($header->type1);
 392         $name = '';
 393         if ($type0 =='message' && $type1 == 'rfc822') {
 394             $rfc822_header = $att->rfc822_header;
 395             $filename = decodeHeader($rfc822_header->subject);
 396             $display_filename = $filename;
 397
 398             $DefaultLink =
 399                 "../src/read_body.php?startMessage=$startMessage&amp;passed_id=$id&amp;mailbox=$urlMailbox&amp;passed_ent_id=$ent";
 400             if ($where && $what) {
 401                 $DefaultLink .= '&amp;where=' . urlencode($where) . '&amp;what=' . urlencode($what);
 402             }
 403             $Links['download link']['text'] = _("download");
 404             $Links['download link']['href'] =
 405                 "../src/download.php?absolute_dl=true&amp;passed_id=$id&amp;mailbox=$urlMailbox&amp;passed_ent_id=$ent";
 406             $ImageURL = '';
 407
 408             /* this executes the attachment hook with a specific MIME-type.
 409                 * if that doens't have results, it tries if there's a rule
 410                 * for a more generic type. */
 411             $HookResults = do_hook("attachment $type0/$type1", $Links,
 412                 $startMessage, $id, $urlMailbox, $ent, $DefaultLink, $display_filename, $where, $what);
 413             if(count($HookResults[1]) <= 1) {
 414                 $HookResults = do_hook("attachment $type0/*", $Links,
 415                 $startMessage, $id, $urlMailbox, $ent, $DefaultLink,
 416                 $display_filename, $where, $what);
 417             }
 418
 419             $Links = $HookResults[1];
 420             $DefaultLink = $HookResults[6];
 421
 422             $attachments .= '<TR><TD>&nbsp;&nbsp;</TD><TD>' .
 423                         "<A HREF=\"$DefaultLink\">$display_filename</A>&nbsp;</TD>" .
 424                         '<TD><SMALL><b>' . show_readable_size($header->size) .
 425                         '</b>&nbsp;&nbsp;</small></TD>' .
 426                         "<TD><SMALL>[ $type0/$type1 ]&nbsp;</SMALL></TD>" .
 427                         '<TD><SMALL>';
 428             $from_o = $rfc822_header->from;
 429             if (is_object($from_o)) {
 430                 $from_name = $from_o->getAddress(false);
 431             } else {
 432                 $from_name = _("Unknown sender");
 433             }
 434             $from_name = decodeHeader(htmlspecialchars($from_name));
 435             $attachments .= '<b>' . $from_name . '</b>';
 436             $attachments .= '</SMALL></TD><TD><SMALL>&nbsp;';
 437
 438             $SkipSpaces = 1;
 439             foreach ($Links as $Val) {
 440                 if ($SkipSpaces) {
 441                     $SkipSpaces = 0;
 442                 } else {
 443                     $attachments .= '&nbsp;&nbsp;|&nbsp;&nbsp;';
 444                 }
 445                 $attachments .= '<a href="' . $Val['href'] . '">' .  $Val['text'] . '</a>';
 446             }
 447             unset($Links);
 448         } else {
 449             $filename = decodeHeader($header->getParameter('filename'));
 450             if (trim($filename) == '') {
 451                 $name = decodeHeader($header->getParameter('name'));
 452                 if (trim($name) == '') {
 453                     if ( trim( $header->id ) == '' )
 454                         $display_filename = 'untitled-[' . $ent . ']' ;
 455                     else
 456                         $display_filename = 'cid: ' . $header->id;
 457                 } else {
 458                     $display_filename = $name;
 459                     $filename = $name;
 460                 }
 461             } else {
 462                 $display_filename = $filename;
 463             }
 464
 465             $DefaultLink =
 466                 "../src/download.php?startMessage=$startMessage&amp;passed_id=$id&amp;mailbox=$urlMailbox&amp;passed_ent_id=$ent";
 467             if ($where && $what) {
 468                $DefaultLink = '&amp;where='. urlencode($where).'&amp;what='.urlencode($what);
 469             }
 470             $Links['download link']['text'] = _("download");
 471             $Links['download link']['href'] =
 472                 "../src/download.php?absolute_dl=true&amp;passed_id=$id&amp;mailbox=$urlMailbox&amp;passed_ent_id=$ent";
 473             $ImageURL = '';
 474
 475             /* this executes the attachment hook with a specific MIME-type.
 476                 * if that doens't have results, it tries if there's a rule
 477                 * for a more generic type. */
 478             $HookResults = do_hook("attachment $type0/$type1", $Links,
 479                 $startMessage, $id, $urlMailbox, $ent, $DefaultLink,
 480                 $display_filename, $where, $what);
 481             if(count($HookResults[1]) <= 1) {
 482                 $HookResults = do_hook("attachment $type0/*", $Links,
 483                 $startMessage, $id, $urlMailbox, $ent, $DefaultLink,
 484                 $display_filename, $where, $what);
 485             }
 486
 487             $Links = $HookResults[1];
 488             $DefaultLink = $HookResults[6];
 489
 490             $attachments .= '<TR><TD>&nbsp;&nbsp;</TD><TD>' .
 491                         "<A HREF=\"$DefaultLink\">$display_filename</A>&nbsp;</TD>" .
 492                         '<TD><SMALL><b>' . show_readable_size($header->size) .
 493                         '</b>&nbsp;&nbsp;</small></TD>' .
 494                         "<TD><SMALL>[ $type0/$type1 ]&nbsp;</SMALL></TD>" .
 495                         '<TD><SMALL>';
 496             if ($message->header->description) {
 497                 $attachments .= '<b>' . htmlspecialchars(_($header->description)) . '</b>';
 498             }
 499             $attachments .= '</SMALL></TD><TD><SMALL>&nbsp;';
 500
 501             $SkipSpaces = 1;
 502             foreach ($Links as $Val) {
 503                 if ($SkipSpaces) {
 504                     $SkipSpaces = 0;
 505                 } else {
 506                     $attachments .= '&nbsp;&nbsp;|&nbsp;&nbsp;';
 507                 }
 508                 $attachments .= '<a href="' . $Val['href'] . '">' .  $Val['text'] . '</a>';
 509             }
 510
 511             unset($Links);
 512         }
 513
 514    }
 515    $attachments .= "</SMALL></TD></TR>\n";
 516
 517    $attachments .= "</TABLE></TD></TR></TABLE>";
 518
 519    return $attachments;
 520 }
 521
 522 /** this function decodes the body depending on the encoding type. **/
 523 function decodeBody($body, $encoding) {
 524   $body = str_replace("\r\n", "\n", $body);
 525   $encoding = strtolower($encoding);
 526
 527   global $show_html_default;
 528
 529   if ($encoding == 'quoted-printable' ||
 530       $encoding == 'quoted_printable') {
 531      $body = quoted_printable_decode($body);
 532
 533      while (ereg("=\n", $body))
 534         $body = ereg_replace ("=\n", "", $body);
 535
 536   } else if ($encoding == 'base64') {
 537      $body = base64_decode($body);
 538   }
 539
 540   // All other encodings are returned raw.
 541   return $body;
 542 }
 543
 544 /*
 545  * This functions decode strings that is encoded according to
 546  * RFC1522 (MIME Part Two: Message Header Extensions for Non-ASCII Text).
 547  * Patched by Christian Schmidt <christian@ostenfeld.dk>  23/03/2002
 548  */
 549 function decodeHeader ($string, $utfencode=true) {
 550     if (is_array($string)) {
 551         $string = implode("\n", $string);
 552     }
 553     $i = 0;
 554     while (preg_match('/^(.{' . $i . '})(.*)=\?([^?]*)\?(Q|B)\?([^?]*)\?=/Ui',
 555                       $string, $res)) {
 556         $prefix = $res[1];
 557         // Ignore white-space between consecutive encoded-words
 558         if (strspn($res[2], " \t") != strlen($res[2])) {
 559             $prefix .= $res[2];
 560         }
 561
 562         if (ucfirst($res[4]) == 'B') {
 563             $replace = base64_decode($res[5]);
 564         } else {
 565             $replace = str_replace('_', ' ', $res[5]);
 566             $replace = preg_replace('/=([0-9a-f]{2})/ie', 'chr(hexdec("\1"))',
 567                                     $replace);
 568             /* Only encode into entities by default. Some places
 569                don't need the encoding, like the compose form. */
 570             if ($utfencode) {
 571                 $replace = charset_decode($res[3], $replace);
 572             }
 573         }
 574         $string = $prefix . $replace . substr($string, strlen($res[0]));
 575         $i = strlen($prefix) + strlen($replace);
 576     }
 577     return( $string );
 578 }
 579
 580 /*
 581  * Encode a string according to RFC 1522 for use in headers if it
 582  * contains 8-bit characters or anything that looks like it should
 583  * be encoded.
 584  */
 585 function encodeHeader ($string) {
 586     global $default_charset;
 587
 588     // Encode only if the string contains 8-bit characters or =?
 589     $j = strlen( $string  );
 590     $l = strstr($string, '=?');         // Must be encoded ?
 591     $ret = '';
 592     for( $i=0; $i < $j; ++$i) {
 593         switch( $string{$i} ) {
 594            case '=':
 595           $ret .= '=3D';
 596           break;
 597         case '?':
 598           $ret .= '=3F';
 599           break;
 600         case '_':
 601           $ret .= '=5F';
 602           break;
 603         case ' ':
 604           $ret .= '_';
 605           break;
 606         default:
 607           $k = ord( $string{$i} );
 608           if ( $k > 126 ) {
 609              $ret .= sprintf("=%02X", $k);
 610              $l = TRUE;
 611           } else
 612              $ret .= $string{$i};
 613         }
 614     }
 615
 616     if ( $l ) {
 617         $string = "=?$default_charset?Q?$ret?=";
 618     }
 619
 620     return( $string );
 621 }
 622
 623 /* This function trys to locate the entity_id of a specific mime element */
 624
 625 function find_ent_id( $id, $message ) {
 626     $ret = '';
 627     for ($i=0; $ret == '' && $i < count($message->entities); $i++) {
 628         if ( $message->entities[$i]->header->type0 == 'multipart')  {
 629             $ret = find_ent_id( $id, $message->entities[$i] );
 630         } else {
 631             if ( strcasecmp( $message->entities[$i]->header->id, $id ) == 0 )
 632                 $ret = $message->entities[$i]->entity_id;
 633         }
 634     }
 635     return( $ret );
 636 }
 637
 638 /**
 639  ** HTMLFILTER ROUTINES
 640  */
 641
 642 /**
 643  * This function returns the final tag out of the tag name, an array
 644  * of attributes, and the type of the tag. This function is called by
 645  * sq_sanitize internally.
 646  *
 647  * @param  $tagname  the name of the tag.
 648  * @param  $attary   the array of attributes and their values
 649  * @param  $tagtype  The type of the tag (see in comments).
 650  * @return           a string with the final tag representation.
 651  */
 652 function sq_tagprint($tagname, $attary, $tagtype){
 653     $me = "sq_tagprint";
 654     if ($tagtype == 2){
 655         $fulltag = '</' . $tagname . '>';
 656     } else {
 657         $fulltag = '<' . $tagname;
 658         if (is_array($attary) && sizeof($attary)){
 659             $atts = Array();
 660             while (list($attname, $attvalue) = each($attary)){
 661                 array_push($atts, "$attname=$attvalue");
 662             }
 663             $fulltag .= ' ' . join(" ", $atts);
 664         }
 665         if ($tagtype == 3){
 666             $fulltag .= " /";
 667         }
 668         $fulltag .= ">";
 669     }
 670     return $fulltag;
 671 }
 672
 673 /**
 674  * A small helper function to use with array_walk. Modifies a by-ref
 675  * value and makes it lowercase.
 676  *
 677  * @param  $val a value passed by-ref.
 678  * @return      void since it modifies a by-ref value.
 679  */
 680 function sq_casenormalize(&$val){
 681     $val = strtolower($val);
 682 }
 683
 684 /**
 685  * This function skips any whitespace from the current position within
 686  * a string and to the next non-whitespace value.
 687  *
 688  * @param  $body   the string
 689  * @param  $offset the offset within the string where we should start
 690  *                 looking for the next non-whitespace character.
 691  * @return         the location within the $body where the next
 692  *                 non-whitespace char is located.
 693  */
 694 function sq_skipspace($body, $offset){
 695     $me = "sq_skipspace";
 696     preg_match("/^(\s*)/s", substr($body, $offset), $matches);
 697     if (sizeof($matches{1})){
 698         $count = strlen($matches{1});
 699         $offset += $count;
 700     }
 701     return $offset;
 702 }
 703
 704 /**
 705  * This function looks for the next character within a string.  It's
 706  * really just a glorified "strpos", except it catches if failures
 707  * nicely.
 708  *
 709  * @param  $body   The string to look for needle in.
 710  * @param  $offset Start looking from this position.
 711  * @param  $needle The character/string to look for.
 712  * @return         location of the next occurance of the needle, or
 713  *                 strlen($body) if needle wasn't found.
 714  */
 715 function sq_findnxstr($body, $offset, $needle){
 716     $me = "sq_findnxstr";
 717     $pos = strpos($body, $needle, $offset);
 718     if ($pos === FALSE){
 719         $pos = strlen($body);
 720     }
 721     return $pos;
 722 }
 723
 724 /**
 725  * This function takes a PCRE-style regexp and tries to match it
 726  * within the string.
 727  *
 728  * @param  $body   The string to look for needle in.
 729  * @param  $offset Start looking from here.
 730  * @param  $reg    A PCRE-style regex to match.
 731  * @return         Returns a false if no matches found, or an array
 732  *                 with the following members:
 733  *                 - integer with the location of the match within $body
 734  *                 - string with whatever content between offset and the match
 735  *                 - string with whatever it is we matched
 736  */
 737 function sq_findnxreg($body, $offset, $reg){
 738     $me = "sq_findnxreg";
 739     $matches = Array();
 740     $retarr = Array();
 741     preg_match("%^(.*?)($reg)%s", substr($body, $offset), $matches);
 742     if (!$matches{0}){
 743         $retarr = false;
 744     } else {
 745         $retarr{0} = $offset + strlen($matches{1});
 746         $retarr{1} = $matches{1};
 747         $retarr{2} = $matches{2};
 748     }
 749     return $retarr;
 750 }
 751
 752 /**
 753  * This function looks for the next tag.
 754  *
 755  * @param  $body   String where to look for the next tag.
 756  * @param  $offset Start looking from here.
 757  * @return         false if no more tags exist in the body, or
 758  *                 an array with the following members:
 759  *                 - string with the name of the tag
 760  *                 - array with attributes and their values
 761  *                 - integer with tag type (1, 2, or 3)
 762  *                 - integer where the tag starts (starting "<")
 763  *                 - integer where the tag ends (ending ">")
 764  *                 first three members will be false, if the tag is invalid.
 765  */
 766 function sq_getnxtag($body, $offset){
 767     $me = "sq_getnxtag";
 768     if ($offset > strlen($body)){
 769         return false;
 770     }
 771     $lt = sq_findnxstr($body, $offset, "<");
 772     if ($lt == strlen($body)){
 773         return false;
 774     }
 775     /**
 776      * We are here:
 777      * blah blah <tag attribute="value">
 778      * \---------^
 779      */
 780     $pos = sq_skipspace($body, $lt+1);
 781     if ($pos >= strlen($body)){
 782         return Array(false, false, false, $lt, strlen($body));
 783     }
 784     /**
 785      * There are 3 kinds of tags:
 786      * 1. Opening tag, e.g.:
 787      *    <a href="blah">
 788      * 2. Closing tag, e.g.:
 789      *    </a>
 790      * 3. XHTML-style content-less tag, e.g.:
 791      *    <img src="blah"/>
 792      */
 793     $tagtype = false;
 794     switch (substr($body, $pos, 1)){
 795     case "/":
 796         $tagtype = 2;
 797         $pos++;
 798         break;
 799     case "!":
 800         /**
 801          * A comment or an SGML declaration.
 802          */
 803         if (substr($body, $pos+1, 2) == "--"){
 804             $gt = strpos($body, "-->", $pos);
 805             if ($gt === false){
 806                 $gt = strlen($body);
 807             } else {
 808                 $gt += 2;
 809             }
 810             return Array(false, false, false, $lt, $gt);
 811         } else {
 812             $gt = sq_findnxstr($body, $pos, ">");
 813             return Array(false, false, false, $lt, $gt);
 814         }
 815         break;
 816     default:
 817         /**
 818          * Assume tagtype 1 for now. If it's type 3, we'll switch values
 819          * later.
 820          */
 821         $tagtype = 1;
 822         break;
 823     }
 824
 825     $tag_start = $pos;
 826     $tagname = '';
 827     /**
 828      * Look for next [\W-_], which will indicate the end of the tag name.
 829      */
 830     $regary = sq_findnxreg($body, $pos, "[^\w\-_]");
 831     if ($regary == false){
 832         return Array(false, false, false, $lt, strlen($body));
 833     }
 834     list($pos, $tagname, $match) = $regary;
 835     $tagname = strtolower($tagname);
 836
 837     /**
 838      * $match can be either of these:
 839      * '>'  indicating the end of the tag entirely.
 840      * '\s' indicating the end of the tag name.
 841      * '/'  indicating that this is type-3 xhtml tag.
 842      *
 843      * Whatever else we find there indicates an invalid tag.
 844      */
 845     switch ($match){
 846     case "/":
 847         /**
 848          * This is an xhtml-style tag with a closing / at the
 849          * end, like so: <img src="blah"/>. Check if it's followed
 850          * by the closing bracket. If not, then this tag is invalid
 851          */
 852         if (substr($body, $pos, 2) == "/>"){
 853             $pos++;
 854             $tagtype = 3;
 855         } else {
 856             $gt = sq_findnxstr($body, $pos, ">");
 857             $retary = Array(false, false, false, $lt, $gt);
 858             return $retary;
 859         }
 860     case ">":
 861         return Array($tagname, false, $tagtype, $lt, $pos);
 862         break;
 863     default:
 864         /**
 865          * Check if it's whitespace
 866          */
 867         if (preg_match("/\s/", $match)){
 868         } else {
 869             /**
 870              * This is an invalid tag! Look for the next closing ">".
 871              */
 872             $gt = sq_findnxstr($body, $offset, ">");
 873             return Array(false, false, false, $lt, $gt);
 874         }
 875     }
 876
 877     /**
 878      * At this point we're here:
 879      * <tagname  attribute='blah'>
 880      * \-------^
 881      *
 882      * At this point we loop in order to find all attributes.
 883      */
 884     $attname = '';
 885     $atttype = false;
 886     $attary = Array();
 887
 888     while ($pos <= strlen($body)){
 889         $pos = sq_skipspace($body, $pos);
 890         if ($pos == strlen($body)){
 891             /**
 892              * Non-closed tag.
 893              */
 894             return Array(false, false, false, $lt, $pos);
 895         }
 896         /**
 897          * See if we arrived at a ">" or "/>", which means that we reached
 898          * the end of the tag.
 899          */
 900         $matches = Array();
 901         if (preg_match("%^(\s*)(>|/>)%s", substr($body, $pos), $matches)) {
 902             /**
 903              * Yep. So we did.
 904              */
 905             $pos += strlen($matches{1});
 906             if ($matches{2} == "/>"){
 907                 $tagtype = 3;
 908                 $pos++;
 909             }
 910             return Array($tagname, $attary, $tagtype, $lt, $pos);
 911         }
 912
 913         /**
 914          * There are several types of attributes, with optional
 915          * [:space:] between members.
 916          * Type 1:
 917          *   attrname[:space:]=[:space:]'CDATA'
 918          * Type 2:
 919          *   attrname[:space:]=[:space:]"CDATA"
 920          * Type 3:
 921          *   attr[:space:]=[:space:]CDATA
 922          * Type 4:
 923          *   attrname
 924          *
 925          * We leave types 1 and 2 the same, type 3 we check for
 926          * '"' and convert to "&quot" if needed, then wrap in
 927          * double quotes. Type 4 we convert into:
 928          * attrname="yes".
 929          */
 930         $regary = sq_findnxreg($body, $pos, "[^\w\-_]");
 931         if ($regary == false){
 932             /**
 933              * Looks like body ended before the end of tag.
 934              */
 935             return Array(false, false, false, $lt, strlen($body));
 936         }
 937         list($pos, $attname, $match) = $regary;
 938         $attname = strtolower($attname);
 939         /**
 940          * We arrived at the end of attribute name. Several things possible
 941          * here:
 942          * '>'  means the end of the tag and this is attribute type 4
 943          * '/'  if followed by '>' means the same thing as above
 944          * '\s' means a lot of things -- look what it's followed by.
 945          *      anything else means the attribute is invalid.
 946          */
 947         switch($match){
 948         case "/":
 949             /**
 950              * This is an xhtml-style tag with a closing / at the
 951              * end, like so: <img src="blah"/>. Check if it's followed
 952              * by the closing bracket. If not, then this tag is invalid
 953              */
 954             if (substr($body, $pos, 2) == "/>"){
 955                 $pos++;
 956                 $tagtype = 3;
 957             } else {
 958                 $gt = sq_findnxstr($body, $pos, ">");
 959                 $retary = Array(false, false, false, $lt, $gt);
 960                 return $retary;
 961             }
 962         case ">":
 963             $attary{$attname} = '"yes"';
 964             return Array($tagname, $attary, $tagtype, $lt, $pos);
 965             break;
 966         default:
 967             /**
 968              * Skip whitespace and see what we arrive at.
 969              */
 970             $pos = sq_skipspace($body, $pos);
 971             $char = substr($body, $pos, 1);
 972             /**
 973              * Two things are valid here:
 974              * '=' means this is attribute type 1 2 or 3.
 975              * \w means this was attribute type 4.
 976              * anything else we ignore and re-loop. End of tag and
 977              * invalid stuff will be caught by our checks at the beginning
 978              * of the loop.
 979              */
 980             if ($char == "="){
 981                 $pos++;
 982                 $pos = sq_skipspace($body, $pos);
 983                 /**
 984                  * Here are 3 possibilities:
 985                  * "'"  attribute type 1
 986                  * '"'  attribute type 2
 987                  * everything else is the content of tag type 3
 988                  */
 989                 $quot = substr($body, $pos, 1);
 990                 if ($quot == "'"){
 991                     $regary = sq_findnxreg($body, $pos+1, "\'");
 992                     if ($regary == false){
 993                         return Array(false, false, false, $lt, strlen($body));
 994                     }
 995                     list($pos, $attval, $match) = $regary;
 996                     $pos++;
 997                     $attary{$attname} = "'" . $attval . "'";
 998                 } else if ($quot == '"'){
 999                     $regary = sq_findnxreg($body, $pos+1, '\"');
1000                     if ($regary == false){
1001                         return Array(false, false, false, $lt, strlen($body));
1002                     }
1003                     list($pos, $attval, $match) = $regary;
1004                     $pos++;
1005                     $attary{$attname} = '"' . $attval . '"';
1006                 } else {
1007                     /**
1008                      * These are hateful. Look for \s, or >.
1009                      */
1010                     $regary = sq_findnxreg($body, $pos, "[\s>]");
1011                     if ($regary == false){
1012                         return Array(false, false, false, $lt, strlen($body));
1013                     }
1014                     list($pos, $attval, $match) = $regary;
1015                     /**
1016                      * If it's ">" it will be caught at the top.
1017                      */
1018                     $attval = preg_replace("/\"/s", "&quot;", $attval);
1019                     $attary{$attname} = '"' . $attval . '"';
1020                 }
1021             } else if (preg_match("|[\w/>]|", $char)) {
1022                 /**
1023                  * That was attribute type 4.
1024                  */
1025                 $attary{$attname} = '"yes"';
1026             } else {
1027                 /**
1028                  * An illegal character. Find next '>' and return.
1029                  */
1030                 $gt = sq_findnxstr($body, $pos, ">");
1031                 return Array(false, false, false, $lt, $gt);
1032             }
1033         }
1034     }
1035     /**
1036      * The fact that we got here indicates that the tag end was never
1037      * found. Return invalid tag indication so it gets stripped.
1038      */
1039     return Array(false, false, false, $lt, strlen($body));
1040 }
1041
1042 /**
1043  * This function checks attribute values for entity-encoded values
1044  * and returns them translated into 8-bit strings so we can run
1045  * checks on them.
1046  *
1047  * @param  $attvalue A string to run entity check against.
1048  * @return           Translated value.
1049  */
1050 function sq_deent($attvalue){
1051     $me="sq_deent";
1052     /**
1053      * See if we have to run the checks first. All entities must start
1054      * with "&".
1055      */
1056     if (strpos($attvalue, "&") === false){
1057         return $attvalue;
1058     }
1059     /**
1060      * Check named entities first.
1061      */
1062     $trans = get_html_translation_table(HTML_ENTITIES);
1063     /**
1064      * Leave &quot; in, as it can mess us up.
1065      */
1066     $trans = array_flip($trans);
1067     unset($trans{"&quot;"});
1068     while (list($ent, $val) = each($trans)){
1069         $attvalue = preg_replace("/$ent*(\W)/si", "$val\\1", $attvalue);
1070     }
1071     /**
1072      * Now translate numbered entities from 1 to 255 if needed.
1073      */
1074     if (strpos($attvalue, "#") !== false){
1075         $omit = Array(34, 39);
1076         for ($asc=1; $asc<256; $asc++){
1077             if (!in_array($asc, $omit)){
1078                 $chr = chr($asc);
1079                 $attvalue = preg_replace("/\&#0*$asc;*(\D)/si", "$chr\\1",
1080                                          $attvalue);
1081                 $attvalue = preg_replace("/\&#x0*".dechex($asc).";*(\W)/si",
1082                                          "$chr\\1", $attvalue);
1083             }
1084         }
1085     }
1086     return $attvalue;
1087 }
1088
1089 /**
1090  * This function runs various checks against the attributes.
1091  *
1092  * @param  $tagname         String with the name of the tag.
1093  * @param  $attary          Array with all tag attributes.
1094  * @param  $rm_attnames     See description for sq_sanitize
1095  * @param  $bad_attvals     See description for sq_sanitize
1096  * @param  $add_attr_to_tag See description for sq_sanitize
1097  * @param  $message         message object
1098  * @param  $id              message id
1099  * @return                  Array with modified attributes.
1100  */
1101 function sq_fixatts($tagname,
1102                     $attary,
1103                     $rm_attnames,
1104                     $bad_attvals,
1105                     $add_attr_to_tag,
1106                     $message,
1107                     $id,
1108                     $mailbox
1109                     ){
1110     $me = "sq_fixatts";
1111     while (list($attname, $attvalue) = each($attary)){
1112         /**
1113          * See if this attribute should be removed.
1114          */
1115         foreach ($rm_attnames as $matchtag=>$matchattrs){
1116             if (preg_match($matchtag, $tagname)){
1117                 foreach ($matchattrs as $matchattr){
1118                     if (preg_match($matchattr, $attname)){
1119                         unset($attary{$attname});
1120                         continue;
1121                     }
1122                 }
1123             }
1124         }
1125         /**
1126          * Remove any entities.
1127          */
1128         $attvalue = sq_deent($attvalue);
1129
1130         /**
1131          * Now let's run checks on the attvalues.
1132          * I don't expect anyone to comprehend this. If you do,
1133          * get in touch with me so I can drive to where you live and
1134          * shake your hand personally. :)
1135          */
1136         foreach ($bad_attvals as $matchtag=>$matchattrs){
1137             if (preg_match($matchtag, $tagname)){
1138                 foreach ($matchattrs as $matchattr=>$valary){
1139                     if (preg_match($matchattr, $attname)){
1140                         /**
1141                          * There are two arrays in valary.
1142                          * First is matches.
1143                          * Second one is replacements
1144                          */
1145                         list($valmatch, $valrepl) = $valary;
1146                         $newvalue =
1147                             preg_replace($valmatch, $valrepl, $attvalue);
1148                         if ($newvalue != $attvalue){
1149                             $attary{$attname} = $newvalue;
1150                         }
1151                     }
1152                 }
1153             }
1154         }
1155         /**
1156          * Turn cid: urls into http-friendly ones.
1157          */
1158         if (preg_match("/^[\'\"]\s*cid:/si", $attvalue)){
1159             $attary{$attname} = sq_cid2http($message, $id, $attvalue, $mailbox);
1160         }
1161     }
1162     /**
1163      * See if we need to append any attributes to this tag.
1164      */
1165     foreach ($add_attr_to_tag as $matchtag=>$addattary){
1166         if (preg_match($matchtag, $tagname)){
1167             $attary = array_merge($attary, $addattary);
1168         }
1169     }
1170     return $attary;
1171 }
1172
1173 /**
1174  * This function edits the style definition to make them friendly and
1175  * usable in squirrelmail.
1176  *
1177  * @param  $message  the message object
1178  * @param  $id       the message id
1179  * @param  $content  a string with whatever is between <style> and </style>
1180  * @return           a string with edited content.
1181  */
1182 function sq_fixstyle($message, $id, $content){
1183     global $view_unsafe_images;
1184     $me = "sq_fixstyle";
1185     /**
1186      * First look for general BODY style declaration, which would be
1187      * like so:
1188      * body {background: blah-blah}
1189      * and change it to .bodyclass so we can just assign it to a <div>
1190      */
1191     $content = preg_replace("|body(\s*\{.*?\})|si", ".bodyclass\\1", $content);
1192     $secremoveimg = "../images/" . _("sec_remove_eng.png");
1193     /**
1194      * Fix url('blah') declarations.
1195      */
1196     $content = preg_replace("|url\(([\'\"])\s*\S+script\s*:.*?([\'\"])\)|si",
1197                             "url(\\1$secremoveimg\\2)", $content);
1198     /**
1199      * Fix url('https*://.*) declarations but only if $view_unsafe_images
1200      * is false.
1201      */
1202     if (!$view_unsafe_images){
1203         $content = preg_replace("|url\(([\'\"])\s*https*:.*?([\'\"])\)|si",
1204                                 "url(\\1$secremoveimg\\2)", $content);
1205     }
1206
1207     /**
1208      * Fix urls that refer to cid:
1209      */
1210     while (preg_match("|url\(([\'\"]\s*cid:.*?[\'\"])\)|si", $content,
1211                       $matches)){
1212         $cidurl = $matches{1};
1213         $httpurl = sq_cid2http($message, $id, $cidurl);
1214         $content = preg_replace("|url\($cidurl\)|si",
1215                                 "url($httpurl)", $content);
1216     }
1217
1218     /**
1219      * Fix stupid css declarations which lead to vulnerabilities
1220      * in IE.
1221      */
1222     $match   = Array('/expression/si',
1223                      '/behaviou*r/si',
1224                      '/binding/si');
1225     $replace = Array('idiocy', 'idiocy', 'idiocy');
1226     $content = preg_replace($match, $replace, $content);
1227     return $content;
1228 }
1229
1230 /**
1231  * This function converts cid: url's into the ones that can be viewed in
1232  * the browser.
1233  *
1234  * @param  $message  the message object
1235  * @param  $id       the message id
1236  * @param  $cidurl   the cid: url.
1237  * @return           a string with a http-friendly url
1238  */
1239 function sq_cid2http($message, $id, $cidurl, $mailbox){
1240     /**
1241      * Get rid of quotes.
1242      */
1243     $quotchar = substr($cidurl, 0, 1);
1244     $cidurl = str_replace($quotchar, "", $cidurl);
1245     $cidurl = substr(trim($cidurl), 4);
1246     $httpurl = $quotchar . "../src/download.php?absolute_dl=true&amp;" .
1247         "passed_id=$id&amp;mailbox=" . urlencode($mailbox) .
1248         "&amp;passed_ent_id=" . find_ent_id($cidurl, $message) . $quotchar;
1249     return $httpurl;
1250 }
1251
1252 /**
1253  * This function changes the <body> tag into a <div> tag since we
1254  * can't really have a body-within-body.
1255  *
1256  * @param  $attary  an array of attributes and values of <body>
1257  * @return          a modified array of attributes to be set for <div>
1258  */
1259 function sq_body2div($attary){
1260     $me = "sq_body2div";
1261     $divattary = Array("class"=>"'bodyclass'");
1262     $bgcolor="#ffffff";
1263     $text="#000000";
1264     $styledef="";
1265     if (is_array($attary) && sizeof($attary) > 0){
1266         foreach ($attary as $attname=>$attvalue){
1267             $quotchar = substr($attvalue, 0, 1);
1268             $attvalue = str_replace($quotchar, "", $attvalue);
1269             switch ($attname){
1270             case "background":
1271                 $styledef .= "background-image: url('$attvalue'); ";
1272                 break;
1273             case "bgcolor":
1274                 $styledef .= "background-color: $attvalue; ";
1275                 break;
1276             case "text":
1277                 $styledef .= "color: $attvalue; ";
1278             }
1279         }
1280         if (strlen($styledef) > 0){
1281             $divattary{"style"} = "\"$styledef\"";
1282         }
1283     }
1284     return $divattary;
1285 }
1286
1287 /**
1288  * This is the main function and the one you should actually be calling.
1289  * There are several variables you should be aware of an which need
1290  * special description.
1291  *
1292  * Since the description is quite lengthy, see it here:
1293  * http://www.mricon.com/html/phpfilter.html
1294  *
1295  * @param $body                 the string with HTML you wish to filter
1296  * @param $tag_list             see description above
1297  * @param $rm_tags_with_content see description above
1298  * @param $self_closing_tags    see description above
1299  * @param $force_tag_closing    see description above
1300  * @param $rm_attnames          see description above
1301  * @param $bad_attvals          see description above
1302  * @param $add_attr_to_tag      see description above
1303  * @param $message              message object
1304  * @param $id                   message id
1305  * @return                      sanitized html safe to show on your pages.
1306  */
1307 function sq_sanitize($body,
1308                      $tag_list,
1309                      $rm_tags_with_content,
1310                      $self_closing_tags,
1311                      $force_tag_closing,
1312                      $rm_attnames,
1313                      $bad_attvals,
1314                      $add_attr_to_tag,
1315                      $message,
1316                      $id,
1317                      $mailbox
1318                      ){
1319     $me = "sq_sanitize";
1320     /**
1321      * Normalize rm_tags and rm_tags_with_content.
1322      */
1323     @array_walk($rm_tags, 'sq_casenormalize');
1324     @array_walk($rm_tags_with_content, 'sq_casenormalize');
1325     @array_walk($self_closing_tags, 'sq_casenormalize');
1326     /**
1327      * See if tag_list is of tags to remove or tags to allow.
1328      * false  means remove these tags
1329      * true   means allow these tags
1330      */
1331     $rm_tags = array_shift($tag_list);
1332     $curpos = 0;
1333     $open_tags = Array();
1334     $trusted = "<!-- begin sanitized html -->\n";
1335     $skip_content = false;
1336     /**
1337      * Take care of netscape's stupid javascript entities like
1338      * &{alert('boo')};
1339      */
1340     $body = preg_replace("/&(\{.*?\};)/si", "&amp;\\1", $body);
1341
1342     while (($curtag=sq_getnxtag($body, $curpos)) != FALSE){
1343         list($tagname, $attary, $tagtype, $lt, $gt) = $curtag;
1344         $free_content = substr($body, $curpos, $lt-$curpos);
1345         /**
1346          * Take care of <style>
1347          */
1348         if ($tagname == "style" && $tagtype == 2){
1349             /**
1350              * This is a closing </style>. Edit the
1351              * content before we apply it.
1352              */
1353             $free_content = sq_fixstyle($message, $id, $free_content);
1354         }
1355         if ($skip_content == false){
1356             $trusted .= $free_content;
1357         } else {
1358         }
1359         if ($tagname != FALSE){
1360             if ($tagtype == 2){
1361                 if ($skip_content == $tagname){
1362                     /**
1363                      * Got to the end of tag we needed to remove.
1364                      */
1365                     $tagname = false;
1366                     $skip_content = false;
1367                 } else {
1368                     if ($skip_content == false){
1369                         if ($tagname == "body"){
1370                             $tagname = "div";
1371                         } else {
1372                             if (isset($open_tags{$tagname}) &&
1373                                 $open_tags{$tagname} > 0){
1374                                 $open_tags{$tagname}--;
1375                             } else {
1376                                 $tagname = false;
1377                             }
1378                         }
1379                     } else {
1380                     }
1381                 }
1382             } else {
1383                 /**
1384                  * $rm_tags_with_content
1385                  */
1386                 if ($skip_content == false){
1387                     /**
1388                      * See if this is a self-closing type and change
1389                      * tagtype appropriately.
1390                      */
1391                     if ($tagtype == 1
1392                         && in_array($tagname, $self_closing_tags)){
1393                         $tagtype=3;
1394                     }
1395                     /**
1396                      * See if we should skip this tag and any content
1397                      * inside it.
1398                      */
1399                     if ($tagtype == 1 &&
1400                         in_array($tagname, $rm_tags_with_content)){
1401                         $skip_content = $tagname;
1402                     } else {
1403                         if (($rm_tags == false
1404                              && in_array($tagname, $tag_list)) ||
1405                             ($rm_tags == true &&
1406                              !in_array($tagname, $tag_list))){
1407                             $tagname = false;
1408                         } else {
1409                             if ($tagtype == 1){
1410                                 if (isset($open_tags{$tagname})){
1411                                     $open_tags{$tagname}++;
1412                                 } else {
1413                                     $open_tags{$tagname}=1;
1414                                 }
1415                             }
1416                             /**
1417                              * This is where we run other checks.
1418                              */
1419                             if (is_array($attary) && sizeof($attary) > 0){
1420                                 $attary = sq_fixatts($tagname,
1421                                                      $attary,
1422                                                      $rm_attnames,
1423                                                      $bad_attvals,
1424                                                      $add_attr_to_tag,
1425                                                      $message,
1426                                                      $id,
1427                                                      $mailbox
1428                                                      );
1429                             }
1430                             /**
1431                              * Convert body into div.
1432                              */
1433                             if ($tagname == "body"){
1434                                 $tagname = "div";
1435                                 $attary = sq_body2div($attary, $message, $id);
1436                             }
1437                         }
1438                     }
1439                 } else {
1440                 }
1441             }
1442             if ($tagname != false && $skip_content == false){
1443                 $trusted .= sq_tagprint($tagname, $attary, $tagtype);
1444             }
1445         } else {
1446         }
1447         $curpos = $gt+1;
1448     }
1449     $trusted .= substr($body, $curpos, strlen($body)-$curpos);
1450     if ($force_tag_closing == true){
1451         foreach ($open_tags as $tagname=>$opentimes){
1452             while ($opentimes > 0){
1453                 $trusted .= '</' . $tagname . '>';
1454                 $opentimes--;
1455             }
1456         }
1457         $trusted .= "\n";
1458     }
1459     $trusted .= "<!-- end sanitized html -->\n";
1460     return $trusted;
1461 }
1462
1463 /**
1464  * This is a wrapper function to call html sanitizing routines.
1465  *
1466  * @param  $body  the body of the message
1467  * @param  $id    the id of the message
1468  * @return        a string with html safe to display in the browser.
1469  */
1470 function magicHTML($body, $id, $message, $mailbox = 'INBOX'){
1471     global $attachment_common_show_images, $view_unsafe_images,
1472         $has_unsafe_images;
1473     /**
1474      * Don't display attached images in HTML mode.
1475      */
1476     $attachment_common_show_images = false;
1477     $tag_list = Array(
1478                       false,
1479                       "object",
1480                       "meta",
1481                       "html",
1482                       "head",
1483                       "base",
1484                       "link"
1485                       );
1486
1487     $rm_tags_with_content = Array(
1488                                   "script",
1489                                   "applet",
1490                                   "embed",
1491                                   "title"
1492                                   );
1493
1494     $self_closing_tags =  Array(
1495                                 "img",
1496                                 "br",
1497                                 "hr",
1498                                 "input"
1499                                 );
1500
1501     $force_tag_closing = false;
1502
1503     $rm_attnames = Array(
1504                          "/.*/" =>
1505                          Array(
1506                                "/target/si",
1507                                "/^on.*/si",
1508                                "/^dynsrc/si",
1509                                "/^data.*/si"
1510                                )
1511                          );
1512
1513     $secremoveimg = "../images/" . _("sec_remove_eng.png");
1514     $bad_attvals = Array(
1515         "/.*/" =>
1516             Array(
1517                 "/^src|background/i" =>
1518                     Array(
1519                           Array(
1520                                 "|^([\'\"])\s*\.\./.*([\'\"])|si",
1521                                 "/^([\'\"])\s*\S+script\s*:.*([\'\"])/si",
1522                                 "/^([\'\"])\s*mocha\s*:*.*([\'\"])/si",
1523                                 "/^([\'\"])\s*about\s*:.*([\'\"])/si"
1524                                 ),
1525                           Array(
1526                                 "\\1$secremoveimg\\2",
1527                                 "\\1$secremoveimg\\2",
1528                                 "\\1$secremoveimg\\2",
1529                                 "\\1$secremoveimg\\2"
1530                                 )
1531                         ),
1532                 "/^href|action/i" =>
1533                     Array(
1534                           Array(
1535                                 "|^([\'\"])\s*\.\./.*([\'\"])|si",
1536                                 "/^([\'\"])\s*\S+script\s*:.*([\'\"])/si",
1537                                 "/^([\'\"])\s*mocha\s*:*.*([\'\"])/si",
1538                                 "/^([\'\"])\s*about\s*:.*([\'\"])/si"
1539                                 ),
1540                           Array(
1541                                 "\\1#\\2",
1542                                 "\\1#\\2",
1543                                 "\\1#\\2",
1544                                 "\\1#\\2"
1545                                 )
1546                         ),
1547                 "/^style/si" =>
1548                     Array(
1549                           Array(
1550                                 "/expression/si",
1551                                 "/binding/si",
1552                                 "/behaviou*r/si",
1553                                 "|url\(([\'\"])\s*\.\./.*([\'\"])\)|si",
1554                                 "/url\(([\'\"])\s*\S+script\s*:.*([\'\"])\)/si",
1555                                 "/url\(([\'\"])\s*mocha\s*:.*([\'\"])\)/si",
1556                                 "/url\(([\'\"])\s*about\s*:.*([\'\"])\)/si"
1557                                ),
1558                           Array(
1559                                 "idiocy",
1560                                 "idiocy",
1561                                 "idiocy",
1562                                 "url(\\1#\\2)",
1563                                 "url(\\1#\\2)",
1564                                 "url(\\1#\\2)",
1565                                 "url(\\1#\\2)"
1566                                )
1567                           )
1568                 )
1569         );
1570     if (!$view_unsafe_images){
1571         /**
1572          * Remove any references to http/https if view_unsafe_images set
1573          * to false.
1574          */
1575          array_push($bad_attvals{'/.*/'}{'/^src|background/i'}[0],
1576                     '/^([\'\"])\s*https*:.*([\'\"])/si');
1577          array_push($bad_attvals{'/.*/'}{'/^src|background/i'}[1],
1578                     "\\1$secremoveimg\\2");
1579          array_push($bad_attvals{'/.*/'}{'/^style/si'}[0],
1580                     '/url\(([\'\"])\s*https*:.*([\'\"])\)/si');
1581          array_push($bad_attvals{'/.*/'}{'/^style/si'}[1],
1582                     "url(\\1$secremoveimg\\2)");
1583     }
1584
1585     $add_attr_to_tag = Array(
1586                              "/^a$/si" => Array('target'=>'"_new"')
1587                              );
1588     $trusted = sq_sanitize($body,
1589                            $tag_list,
1590                            $rm_tags_with_content,
1591                            $self_closing_tags,
1592                            $force_tag_closing,
1593                            $rm_attnames,
1594                            $bad_attvals,
1595                            $add_attr_to_tag,
1596                            $message,
1597                            $id,
1598                            $mailbox
1599                            );
1600     if (preg_match("|$secremoveimg|si", $trusted)){
1601         $has_unsafe_images = true;
1602     }
1603     return $trusted;
1604 }
1605
1606 ?>