Norvegian Bokmal code is nb_NO
[squirrelmail.git] / functions / mime.php
CommitLineData
59177427 1<?php
2ba13803 2
35586184 3/**
4 * mime.php
5 *
76911253 6 * Copyright (c) 1999-2003 The SquirrelMail Project Team
35586184 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 * $Id$
d6c32258 13 * @package squirrelmail
35586184 14 */
b74ba498 15
d6c32258 16/** The typical includes... */
b68edc75 17require_once(SM_PATH . 'functions/imap.php');
18require_once(SM_PATH . 'functions/attachment_common.php');
8beafbbc 19
7c7b74b3 20/* -------------------------------------------------------------------------- */
21/* MIME DECODING */
22/* -------------------------------------------------------------------------- */
b74ba498 23
d6c32258 24/**
25 * Get the MIME structure
07c49f57 26 *
d6c32258 27 * This function gets the structure of a message and stores it in the "message" class.
451f74a2 28 * It will return this object for use with all relevant header information and
29 * fully parsed into the standard "message" object format.
30 */
a4a70693 31function mime_structure ($bodystructure, $flags=array()) {
c9d78ab4 32
3d8371be 33 /* Isolate the body structure and remove beginning and end parenthesis. */
a4a70693 34 $read = trim(substr ($bodystructure, strpos(strtolower($bodystructure), 'bodystructure') + 13));
451f74a2 35 $read = trim(substr ($read, 0, -1));
22efa9fb 36 $i = 0;
37 $msg = Message::parseStructure($read,$i);
9de42168 38 if (!is_object($msg)) {
bd9c880b 39 include_once(SM_PATH . 'functions/display_messages.php');
3d8371be 40 global $color, $mailbox;
c96c32f4 41 /* removed urldecode because $_GET is auto urldecoded ??? */
a48eba8f 42 displayPageHeader( $color, $mailbox );
5e8de8b6 43 $errormessage = _("SquirrelMail could not decode the bodystructure of the message");
3d8371be 44 $errormessage .= '<BR>'._("the provided bodystructure by your imap-server").':<BR><BR>';
472e7acb 45 $errormessage .= '<pre>' . htmlspecialchars($read) . '</pre>';
9de42168 46 plain_error_message( $errormessage, $color );
3d8371be 47 echo '</body></html>';
9de42168 48 exit;
49 }
a4a70693 50 if (count($flags)) {
7a9e9c89 51 foreach ($flags as $flag) {
52 $char = strtoupper($flag{1});
53 switch ($char) {
3d8371be 54 case 'S':
55 if (strtolower($flag) == '\\seen') {
56 $msg->is_seen = true;
57 }
58 break;
59 case 'A':
60 if (strtolower($flag) == '\\answered') {
61 $msg->is_answered = true;
62 }
63 break;
64 case 'D':
65 if (strtolower($flag) == '\\deleted') {
66 $msg->is_deleted = true;
67 }
68 break;
69 case 'F':
70 if (strtolower($flag) == '\\flagged') {
71 $msg->is_flagged = true;
72 }
73 break;
74 case 'M':
75 if (strtolower($flag) == '$mdnsent') {
76 $msg->is_mdnsent = true;
77 }
78 break;
79 default:
80 break;
7a9e9c89 81 }
82 }
451f74a2 83 }
7a9e9c89 84 // listEntities($msg);
3d8371be 85 return $msg;
451f74a2 86}
b74ba498 87
22efa9fb 88
89
3d8371be 90/* This starts the parsing of a particular structure. It is called recursively,
451f74a2 91 * so it can be passed different structures. It returns an object of type
92 * $message.
93 * First, it checks to see if it is a multipart message. If it is, then it
94 * handles that as it sees is necessary. If it is just a regular entity,
95 * then it parses it and adds the necessary header information (by calling out
96 * to mime_get_elements()
97 */
451f74a2 98
4d592352 99function mime_fetch_body($imap_stream, $id, $ent_id=1, $fetch_size=0) {
da2415c1 100 global $uid_support;
3d8371be 101 /* Do a bit of error correction. If we couldn't find the entity id, just guess
09a4bde3 102 * that it is the first one. That is usually the case anyway.
103 */
7c7b74b3 104
09a4bde3 105 if (!$ent_id) {
08b7f7cc 106 $cmd = "FETCH $id BODY[]";
1035e159 107 } else {
08b7f7cc 108 $cmd = "FETCH $id BODY[$ent_id]";
09a4bde3 109 }
3d8371be 110
4d592352 111 if ($fetch_size!=0) $cmd .= "<0.$fetch_size>";
da2415c1 112
a4a70693 113 $data = sqimap_run_command ($imap_stream, $cmd, true, $response, $message, $uid_support);
77b88425 114 do {
3d8371be 115 $topline = trim(array_shift($data));
116 } while($topline && ($topline[0] == '*') && !preg_match('/\* [0-9]+ FETCH.*/i', $topline)) ;
a4a70693 117
451f74a2 118 $wholemessage = implode('', $data);
119 if (ereg('\\{([^\\}]*)\\}', $topline, $regs)) {
3d8371be 120 $ret = substr($wholemessage, 0, $regs[1]);
121 /* There is some information in the content info header that could be important
122 * in order to parse html messages. Let's get them here.
123 */
0600bdf1 124// if ($ret{0} == '<') {
125// $data = sqimap_run_command ($imap_stream, "FETCH $id BODY[$ent_id.MIME]", true, $response, $message, $uid_support);
126// }
451f74a2 127 } else if (ereg('"([^"]*)"', $topline, $regs)) {
128 $ret = $regs[1];
129 } else {
130 global $where, $what, $mailbox, $passed_id, $startMessage;
3d8371be 131 $par = 'mailbox=' . urlencode($mailbox) . '&amp;passed_id=' . $passed_id;
451f74a2 132 if (isset($where) && isset($what)) {
3d8371be 133 $par .= '&amp;where=' . urlencode($where) . '&amp;what=' . urlencode($what);
a3daaaf3 134 } else {
3d8371be 135 $par .= '&amp;startMessage=' . $startMessage . '&amp;show_more=0';
451f74a2 136 }
e5ea9327 137 $par .= '&amp;response=' . urlencode($response) .
3d8371be 138 '&amp;message=' . urlencode($message) .
139 '&amp;topline=' . urlencode($topline);
a019eeb8 140
346817d4 141 echo '<tt><br>' .
142 '<table width="80%"><tr>' .
143 '<tr><td colspan=2>' .
0e5b61b4 144 _("Body retrieval error. The reason for this is most probably that the message is malformed.") .
346817d4 145 '</td></tr>' .
0e5b61b4 146 '<tr><td><b>' . _("Command:") . "</td><td>$cmd</td></tr>" .
147 '<tr><td><b>' . _("Response:") . "</td><td>$response</td></tr>" .
148 '<tr><td><b>' . _("Message:") . "</td><td>$message</td></tr>" .
149 '<tr><td><b>' . _("FETCH line:") . "</td><td>$topline</td></tr>" .
346817d4 150 "</table><BR></tt></font><hr>";
151
a4a70693 152 $data = sqimap_run_command ($imap_stream, "FETCH $passed_id BODY[]", true, $response, $message, $uid_support);
451f74a2 153 array_shift($data);
154 $wholemessage = implode('', $data);
a019eeb8 155
346817d4 156 $ret = $wholemessage;
a3daaaf3 157 }
3d8371be 158 return $ret;
451f74a2 159}
d4467150 160
1035e159 161function mime_print_body_lines ($imap_stream, $id, $ent_id=1, $encoding) {
a4a70693 162 global $uid_support;
1035e159 163
3d8371be 164 /* Don't kill the connection if the browser is over a dialup
165 * and it would take over 30 seconds to download it.
7e6ca3e8 166 * Dont call set_time_limit in safe mode.
3d8371be 167 */
b7206e1d 168
3d8371be 169 if (!ini_get('safe_mode')) {
b7206e1d 170 set_time_limit(0);
171 }
7c7b74b3 172 /* in case of base64 encoded attachments, do not buffer them.
173 Instead, echo the decoded attachment directly to screen */
174 if (strtolower($encoding) == 'base64') {
175 if (!$ent_id) {
176 $query = "FETCH $id BODY[]";
177 } else {
178 $query = "FETCH $id BODY[$ent_id]";
179 }
180 sqimap_run_command($imap_stream,$query,true,$response,$message,$uid_support,'sqimap_base64_decode','php://stdout',true);
1d142b8d 181 } else {
7c7b74b3 182 $body = mime_fetch_body ($imap_stream, $id, $ent_id);
183 echo decodeBody($body, $encoding);
1d142b8d 184 }
346817d4 185
da2415c1 186 /*
7c7b74b3 187 TODO, use the same method for quoted printable.
188 However, I assume that quoted printable attachments aren't that large
189 so the performancegain / memory usage drop will be minimal.
190 If we decide to add that then we need to adapt sqimap_fread because
da2415c1 191 we need to split te result on \n and fread doesn't stop at \n. That
7c7b74b3 192 means we also should provide $results from sqimap_fread (by ref) to
193 te function and set $no_return to false. The $filter function for
da2415c1 194 quoted printable should handle unsetting of $results.
7c7b74b3 195 */
da2415c1 196 /*
7c7b74b3 197 TODO 2: find out how we write to the output stream php://stdout. fwrite
198 doesn't work because 'php://stdout isn't a stream.
199 */
200
5d9c6f73 201 return;
202/*
451f74a2 203 fputs ($imap_stream, "$sid FETCH $id BODY[$ent_id]\r\n");
204 $cnt = 0;
205 $continue = true;
1d142b8d 206 $read = fgets ($imap_stream,8192);
207
208
451f74a2 209 // This could be bad -- if the section has sqimap_session_id() . ' OK'
210 // or similar, it will kill the download.
1d142b8d 211 while (!ereg("^".$sid_s." (OK|BAD|NO)(.*)$", $read, $regs)) {
5d9c6f73 212 if (trim($read) == ')==') {
213 $read1 = $read;
214 $read = fgets ($imap_stream,4096);
215 if (ereg("^".$sid." (OK|BAD|NO)(.*)$", $read, $regs)) {
216 return;
217 } else {
218 echo decodeBody($read1, $encoding) .
219 decodeBody($read, $encoding);
220 }
221 } else if ($cnt) {
222 echo decodeBody($read, $encoding);
223 }
224 $read = fgets ($imap_stream,4096);
225 $cnt++;
1d142b8d 226// break;
451f74a2 227 }
5d9c6f73 228*/
451f74a2 229}
beb9e459 230
451f74a2 231/* -[ END MIME DECODING ]----------------------------------------------------------- */
d4467150 232
3d8371be 233/* This is here for debugging purposes. It will print out a list
234 * of all the entity IDs that are in the $message object.
235 */
451f74a2 236function listEntities ($message) {
3d8371be 237 if ($message) {
238 echo "<tt>" . $message->entity_id . ' : ' . $message->type0 . '/' . $message->type1 . ' parent = '. $message->parent->entity_id. '<br>';
239 for ($i = 0; isset($message->entities[$i]); $i++) {
240 echo "$i : ";
241 $msg = listEntities($message->entities[$i]);
242
243 if ($msg) {
244 echo "return: ";
245 return $msg;
246 }
247 }
a4a70693 248 }
451f74a2 249}
f0c4dc12 250
f792c641 251function getPriorityStr($priority) {
3d8371be 252 $priority_level = substr($priority,0,1);
253
254 switch($priority_level) {
255 /* Check for a higher then normal priority. */
256 case '1':
257 case '2':
258 $priority_string = _("High");
259 break;
260
261 /* Check for a lower then normal priority. */
262 case '4':
263 case '5':
264 $priority_string = _("Low");
265 break;
266
267 /* Check for a normal priority. */
268 case '3':
269 default:
270 $priority_level = '3';
271 $priority_string = _("Normal");
272 break;
273
274 }
275 return $priority_string;
f792c641 276}
277
451f74a2 278/* returns a $message object for a particular entity id */
279function getEntity ($message, $ent_id) {
a4a70693 280 return $message->getEntity($ent_id);
451f74a2 281}
8beafbbc 282
3d8371be 283/* translateText
da4c66e8 284 * Extracted from strings.php 23/03/2002
285 */
286
287function translateText(&$body, $wrap_at, $charset) {
3d8371be 288 global $where, $what; /* from searching */
289 global $color; /* color theme */
da4c66e8 290
b68edc75 291 require_once(SM_PATH . 'functions/url_parser.php');
da4c66e8 292
293 $body_ary = explode("\n", $body);
da4c66e8 294 for ($i=0; $i < count($body_ary); $i++) {
295 $line = $body_ary[$i];
296 if (strlen($line) - 2 >= $wrap_at) {
297 sqWordWrap($line, $wrap_at);
298 }
299 $line = charset_decode($charset, $line);
300 $line = str_replace("\t", ' ', $line);
301
302 parseUrl ($line);
303
3d8371be 304 $quotes = 0;
da4c66e8 305 $pos = 0;
3d8371be 306 $j = strlen($line);
da4c66e8 307
3d8371be 308 while ($pos < $j) {
da4c66e8 309 if ($line[$pos] == ' ') {
3d8371be 310 $pos++;
da4c66e8 311 } else if (strpos($line, '&gt;', $pos) === $pos) {
312 $pos += 4;
3d8371be 313 $quotes++;
da4c66e8 314 } else {
315 break;
316 }
317 }
3d8371be 318
83c94382 319 if ($quotes % 2) {
3d8371be 320 if (!isset($color[13])) {
da4c66e8 321 $color[13] = '#800000';
322 }
4c25967c 323 $line = '<font color="' . $color[13] . '">' . $line . '</font>';
324 } elseif ($quotes) {
325 if (!isset($color[14])) {
326 $color[14] = '#FF0000';
327 }
328 $line = '<font color="' . $color[14] . '">' . $line . '</font>';
da4c66e8 329 }
3d8371be 330
da4c66e8 331 $body_ary[$i] = $line;
332 }
333 $body = '<pre>' . implode("\n", $body_ary) . '</pre>';
334}
335
451f74a2 336/* This returns a parsed string called $body. That string can then
3d8371be 337 * be displayed as the actual message in the HTML. It contains
338 * everything needed, including HTML Tags, Attachments at the
339 * bottom, etc.
340 */
b3af12ef 341function formatBody($imap_stream, $message, $color, $wrap_at, $ent_num, $id, $mailbox='INBOX') {
3d8371be 342 /* This if statement checks for the entity to show as the
343 * primary message. To add more of them, just put them in the
344 * order that is their priority.
345 */
008989da 346 global $startMessage, $username, $key, $imapServerAddress, $imapPort,
23f617b8 347 $show_html_default, $sort, $has_unsafe_images, $passed_ent_id;
08b7f7cc 348 global $languages, $squirrelmail_language;
77bfbd2e 349
5262d9a6 350 if( !sqgetGlobalVar('view_unsafe_images', $view_unsafe_images, SQ_GET) ) {
351 $view_unsafe_images = false;
77bfbd2e 352 }
d03c24f4 353
cc34b00d 354 $body = '';
23bcec6f 355 $urlmailbox = urlencode($mailbox);
451f74a2 356 $body_message = getEntity($message, $ent_num);
357 if (($body_message->header->type0 == 'text') ||
358 ($body_message->header->type0 == 'rfc822')) {
3d8371be 359 $body = mime_fetch_body ($imap_stream, $id, $ent_num);
451f74a2 360 $body = decodeBody($body, $body_message->header->encoding);
e842b215 361
362 if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
363 function_exists($languages[$squirrelmail_language]['XTRA_CODE'])) {
364 if (mb_detect_encoding($body) != 'ASCII') {
365 $body = $languages[$squirrelmail_language]['XTRA_CODE']('decode', $body);
366 }
367 }
451f74a2 368 $hookResults = do_hook("message_body", $body);
369 $body = $hookResults[1];
23bcec6f 370
3d8371be 371 /* If there are other types that shouldn't be formatted, add
372 * them here.
373 */
374
451f74a2 375 if ($body_message->header->type1 == 'html') {
3d8371be 376 if ($show_html_default <> 1) {
85015544 377 $entity_conv = array('&nbsp;' => ' ',
3d8371be 378 '<p>' => "\n",
379 '<br>' => "\n",
380 '<P>' => "\n",
381 '<BR>' => "\n",
85015544 382 '&gt;' => '>',
383 '&lt;' => '<');
85015544 384 $body = strtr($body, $entity_conv);
3d8371be 385 $body = strip_tags($body);
85015544 386 $body = trim($body);
387 translateText($body, $wrap_at,
807e884a 388 $body_message->header->getParameter('charset'));
a3daaaf3 389 } else {
3d8371be 390 $body = magicHTML($body, $id, $message, $mailbox);
a3daaaf3 391 }
451f74a2 392 } else {
3d8371be 393 translateText($body, $wrap_at,
394 $body_message->header->getParameter('charset'));
451f74a2 395 }
83cf04bd 396 $link = 'passed_id=' . $id . '&amp;ent_id='.$ent_num.
08b7f7cc 397 '&amp;mailbox=' . $urlmailbox .'&amp;sort=' . $sort .
da2415c1 398 '&amp;startMessage=' . $startMessage . '&amp;show_more=0';
08b7f7cc 399 if (isset($passed_ent_id)) {
400 $link .= '&amp;passed_ent_id='.$passed_ent_id;
401 }
83cf04bd 402 $body .= '<center><small><a href="download.php?absolute_dl=true&amp;' .
403 $link . '">' . _("Download this as a file") . '</a>';
7aad7b77 404 if ($view_unsafe_images) {
23f617b8 405 $text = _("Hide Unsafe Images");
7aad7b77 406 } else {
08b7f7cc 407 if (isset($has_unsafe_images) && $has_unsafe_images) {
408 $link .= '&amp;view_unsafe_images=1';
409 $text = _("View Unsafe Images");
410 } else {
411 $text = '';
412 }
3d8371be 413 }
83cf04bd 414 if($text != '') {
415 $body .= '&nbsp;|&nbsp;<a href="read_body.php?' . $link . '">' . $text . '</a>';
416 }
417 $body .= '</small></center><br>' . "\n";
3d8371be 418 }
419 return $body;
451f74a2 420}
b74ba498 421
23bcec6f 422
423function formatAttachments($message, $exclude_id, $mailbox, $id) {
b583c3e8 424 global $where, $what, $startMessage, $color;
77b88425 425 static $ShownHTML = 0;
451f74a2 426
23bcec6f 427 $att_ar = $message->getAttachments($exclude_id);
451f74a2 428
23bcec6f 429 if (!count($att_ar)) return '';
2e25760a 430
8f5425ec 431 $attachments = '';
2e25760a 432
23bcec6f 433 $urlMailbox = urlencode($mailbox);
434
435 foreach ($att_ar as $att) {
fdc9d9b5 436 $ent = $att->entity_id;
2e25760a 437 $header = $att->header;
f0c4dc12 438 $type0 = strtolower($header->type0);
439 $type1 = strtolower($header->type1);
2e25760a 440 $name = '';
3d8371be 441 $links['download link']['text'] = _("download");
6b04287c 442 $links['download link']['href'] = SM_PATH .
443 "src/download.php?absolute_dl=true&amp;passed_id=$id&amp;mailbox=$urlMailbox&amp;ent_id=$ent";
2e25760a 444 $ImageURL = '';
445 if ($type0 =='message' && $type1 == 'rfc822') {
6b04287c 446 $default_page = SM_PATH . 'src/read_body.php';
2e25760a 447 $rfc822_header = $att->rfc822_header;
098ea084 448 $filename = $rfc822_header->subject;
6cc08d8b 449 if (trim( $filename ) == '') {
450 $filename = 'untitled-[' . $ent . ']' ;
08b7f7cc 451 }
2e25760a 452 $from_o = $rfc822_header->from;
453 if (is_object($from_o)) {
04ea844e 454 $from_name = decodeHeader($from_o->getAddress(false));
2e25760a 455 } else {
456 $from_name = _("Unknown sender");
f0c4dc12 457 }
3d8371be 458 $description = $from_name;
23bcec6f 459 } else {
6b04287c 460 $default_page = SM_PATH . 'src/download.php';
3d8371be 461 if (is_object($header->disposition)) {
098ea084 462 $filename = $header->disposition->getProperty('filename');
3d8371be 463 if (trim($filename) == '') {
464 $name = decodeHeader($header->disposition->getProperty('name'));
465 if (trim($name) == '') {
098ea084 466 $name = $header->getParameter('name');
1035e159 467 if(trim($name) == '') {
468 if (trim( $header->id ) == '') {
469 $filename = 'untitled-[' . $ent . ']' ;
470 } else {
471 $filename = 'cid: ' . $header->id;
08b7f7cc 472 }
3d8371be 473 } else {
08b7f7cc 474 $filename = $name;
3d8371be 475 }
476 } else {
477 $filename = $name;
478 }
479 }
480 } else {
08b7f7cc 481 $filename = $header->getParameter('name');
482 if (!trim($filename)) {
483 if (trim( $header->id ) == '') {
484 $filename = 'untitled-[' . $ent . ']' ;
485 } else {
486 $filename = 'cid: ' . $header->id;
487 }
488 }
489 }
f810c0b2 490 if ($header->description) {
098ea084 491 $description = decodeHeader($header->description);
f810c0b2 492 } else {
3d8371be 493 $description = '';
494 }
2e25760a 495 }
496
497 $display_filename = $filename;
3d8371be 498 if (isset($passed_ent_id)) {
499 $passed_ent_id_link = '&amp;passed_ent_id='.$passed_ent_id;
500 } else {
501 $passed_ent_id_link = '';
502 }
503 $defaultlink = $default_page . "?startMessage=$startMessage"
2e25760a 504 . "&amp;passed_id=$id&amp;mailbox=$urlMailbox"
6b04287c 505 . '&amp;ent_id='.$ent.$passed_ent_id_link;
2e25760a 506 if ($where && $what) {
09f4707e 507 $defaultlink .= '&amp;where='. urlencode($where).'&amp;what='.urlencode($what);
2e25760a 508 }
3d8371be 509 /* This executes the attachment hook with a specific MIME-type.
510 * If that doesn't have results, it tries if there's a rule
511 * for a more generic type.
512 */
513 $hookresults = do_hook("attachment $type0/$type1", $links,
514 $startMessage, $id, $urlMailbox, $ent, $defaultlink,
515 $display_filename, $where, $what);
516 if(count($hookresults[1]) <= 1) {
517 $hookresults = do_hook("attachment $type0/*", $links,
518 $startMessage, $id, $urlMailbox, $ent, $defaultlink,
519 $display_filename, $where, $what);
2e25760a 520 }
451f74a2 521
3d8371be 522 $links = $hookresults[1];
523 $defaultlink = $hookresults[6];
77b88425 524
2e25760a 525 $attachments .= '<TR><TD>' .
098ea084 526 '<A HREF="'.$defaultlink.'">'.decodeHeader($display_filename).'</A>&nbsp;</TD>' .
23bcec6f 527 '<TD><SMALL><b>' . show_readable_size($header->size) .
77b88425 528 '</b>&nbsp;&nbsp;</small></TD>' .
529 "<TD><SMALL>[ $type0/$type1 ]&nbsp;</SMALL></TD>" .
530 '<TD><SMALL>';
f810c0b2 531 $attachments .= '<b>' . $description . '</b>';
2e25760a 532 $attachments .= '</SMALL></TD><TD><SMALL>&nbsp;';
b74ba498 533
3d8371be 534 $skipspaces = 1;
535 foreach ($links as $val) {
536 if ($skipspaces) {
537 $skipspaces = 0;
2e25760a 538 } else {
539 $attachments .= '&nbsp;&nbsp;|&nbsp;&nbsp;';
77b88425 540 }
3d8371be 541 $attachments .= '<a href="' . $val['href'] . '">' . $val['text'] . '</a>';
2e25760a 542 }
3d8371be 543 unset($links);
2e25760a 544 $attachments .= "</TD></TR>\n";
545 }
9ad17edb 546 $attachmentadd = do_hook_function('attachments_bottom',$attachments);
547 if ($attachmentadd != '')
548 $attachments = $attachmentadd;
2e25760a 549 return $attachments;
451f74a2 550}
b74ba498 551
7c7b74b3 552function sqimap_base64_decode(&$string) {
325978ac 553 $string = str_replace("\r\n", "\n", $string);
7c7b74b3 554 $string = base64_decode($string);
555}
556
3d8371be 557/* This function decodes the body depending on the encoding type. */
451f74a2 558function decodeBody($body, $encoding) {
3d8371be 559 global $show_html_default;
83be314a 560
b583c3e8 561 $body = str_replace("\r\n", "\n", $body);
562 $encoding = strtolower($encoding);
3d8371be 563
5166f86a 564 $encoding_handler = do_hook_function('decode_body', $encoding);
565
566
567 // plugins get first shot at decoding the body
568 //
569 if (!empty($encoding_handler) && function_exists($encoding_handler)) {
570 $body = $encoding_handler('decode', $body);
571
572 } else if ($encoding == 'quoted-printable' ||
3d8371be 573 $encoding == 'quoted_printable') {
b583c3e8 574 $body = quoted_printable_decode($body);
3d8371be 575
b583c3e8 576 while (ereg("=\n", $body)) {
577 $body = ereg_replace ("=\n", '', $body);
578 }
3d8371be 579
b583c3e8 580 } else if ($encoding == 'base64') {
581 $body = base64_decode($body);
582 }
3d8371be 583
b583c3e8 584 // All other encodings are returned raw.
3d8371be 585 return $body;
451f74a2 586}
587
9f7f68c3 588/**
589 * Decodes headers
590 *
451f74a2 591 * This functions decode strings that is encoded according to
592 * RFC1522 (MIME Part Two: Message Header Extensions for Non-ASCII Text).
79e07c7e 593 * Patched by Christian Schmidt <christian@ostenfeld.dk> 23/03/2002
9f7f68c3 594 *
595 * @param string $string header string that has to be made readable
596 * @param boolean $utfencode change message in order to be readable on user's charset. defaults to true
597 * @param boolean $htmlsave preserve spaces and sanitize html special characters. defaults to true
598 * @param boolean $decide decide if string can be utfencoded. defaults to false
599 * @return string decoded header string
451f74a2 600 */
9f7f68c3 601function decodeHeader ($string, $utfencode=true,$htmlsave=true,$decide=false) {
d7448126 602 global $languages, $squirrelmail_language;
79e07c7e 603 if (is_array($string)) {
604 $string = implode("\n", $string);
605 }
da2415c1 606
10dec454 607 if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
608 function_exists($languages[$squirrelmail_language]['XTRA_CODE'])) {
6fbd125b 609 $string = $languages[$squirrelmail_language]['XTRA_CODE']('decodeheader', $string);
08b7f7cc 610 // Do we need to return at this point?
611 // return $string;
83be314a 612 }
79e07c7e 613 $i = 0;
08b7f7cc 614 $iLastMatch = -2;
db65b6b0 615 $encoded = true;
0a06275a 616
098ea084 617 $aString = explode(' ',$string);
08b7f7cc 618 $ret = '';
098ea084 619 foreach ($aString as $chunk) {
358a78a1 620 if ($encoded && $chunk === '') {
08b7f7cc 621 continue;
358a78a1 622 } elseif ($chunk === '') {
08b7f7cc 623 $ret .= ' ';
624 continue;
625 }
098ea084 626 $encoded = false;
08b7f7cc 627 /* if encoded words are not separated by a linear-space-white we still catch them */
628 $j = $i-1;
7e6ca3e8 629
08b7f7cc 630 while ($match = preg_match('/^(.*)=\?([^?]*)\?(Q|B)\?([^?]*)\?=(.*)$/Ui',$chunk,$res)) {
631 /* if the last chunk isn't an encoded string then put back the space, otherwise don't */
632 if ($iLastMatch !== $j) {
633 if ($htmlsave) {
9f7f68c3 634 $ret .= '&#32;';
08b7f7cc 635 } else {
636 $ret .= ' ';
637 }
638 }
639 $iLastMatch = $i;
640 $j = $i;
641 $ret .= $res[1];
098ea084 642 $encoding = ucfirst($res[3]);
643 switch ($encoding)
644 {
645 case 'B':
646 $replace = base64_decode($res[4]);
08b7f7cc 647 $ret .= charset_decode($res[2],$replace);
098ea084 648 break;
649 case 'Q':
098ea084 650 $replace = str_replace('_', ' ', $res[4]);
da2415c1 651 $replace = preg_replace('/=([0-9a-f]{2})/ie', 'chr(hexdec("\1"))',
79e07c7e 652 $replace);
9f7f68c3 653 /* decide about valid decoding */
654 if ($decide && is_conversion_safe($res[2])) {
655 $utfencode=true;
656 $can_be_decoded=true;
657 } else {
658 $can_be_decoded=false;
659 }
098ea084 660 /* Only encode into entities by default. Some places
661 * don't need the encoding, like the compose form.
662 */
663 if ($utfencode) {
664 $replace = charset_decode($res[2], $replace);
665 } else {
666 if ($htmlsave) {
c96c32f4 667 $replace = htmlspecialchars($replace);
098ea084 668 }
669 }
08b7f7cc 670 $ret .= $replace;
098ea084 671 break;
672 default:
673 break;
79e07c7e 674 }
098ea084 675 $chunk = $res[5];
676 $encoded = true;
08b7f7cc 677 }
678 if (!$encoded) {
679 if ($htmlsave) {
9f7f68c3 680 $ret .= '&#32;';
08b7f7cc 681 } else {
682 $ret .= ' ';
da2415c1 683 }
08b7f7cc 684 }
dc3d13a7 685
686 if (!$encoded && $htmlsave) {
687 $ret .= htmlspecialchars($chunk);
688 } else {
689 $ret .= $chunk;
690 }
098ea084 691 ++$i;
692 }
fd81e884 693 /* remove the first added space */
694 if ($ret) {
695 if ($htmlsave) {
9f7f68c3 696 $ret = substr($ret,5);
fd81e884 697 } else {
698 $ret = substr($ret,1);
699 }
700 }
da2415c1 701
08b7f7cc 702 return $ret;
451f74a2 703}
704
9f7f68c3 705/**
706 * Encodes header as quoted-printable
707 *
451f74a2 708 * Encode a string according to RFC 1522 for use in headers if it
709 * contains 8-bit characters or anything that looks like it should
710 * be encoded.
9f7f68c3 711 *
712 * @param string $string header string, that has to be encoded
713 * @return string quoted-printable encoded string
451f74a2 714 */
715function encodeHeader ($string) {
6fbd125b 716 global $default_charset, $languages, $squirrelmail_language;
83be314a 717
10dec454 718 if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
719 function_exists($languages[$squirrelmail_language]['XTRA_CODE'])) {
6fbd125b 720 return $languages[$squirrelmail_language]['XTRA_CODE']('encodeheader', $string);
83be314a 721 }
9f7f68c3 722 // instead of removing nbsp here, we don't add it in decodeHeader
723 // if (strtolower($default_charset) == 'iso-8859-1') {
724 // $string = str_replace("\240",' ',$string);
725 //}
793cc001 726
451f74a2 727 // Encode only if the string contains 8-bit characters or =?
3d8371be 728 $j = strlen($string);
098ea084 729 $max_l = 75 - strlen($default_charset) - 7;
730 $aRet = array();
451f74a2 731 $ret = '';
c96c32f4 732 $iEncStart = $enc_init = false;
0d53f0f9 733 $cur_l = $iOffset = 0;
3d8371be 734 for($i = 0; $i < $j; ++$i) {
c96c32f4 735 switch($string{$i})
736 {
737 case '=':
738 case '<':
739 case '>':
740 case ',':
741 case '?':
742 case '_':
743 if ($iEncStart === false) {
744 $iEncStart = $i;
745 }
746 $cur_l+=3;
747 if ($cur_l > ($max_l-2)) {
08b7f7cc 748 /* if there is an stringpart that doesn't need encoding, add it */
c96c32f4 749 $aRet[] = substr($string,$iOffset,$iEncStart-$iOffset);
750 $aRet[] = "=?$default_charset?Q?$ret?=";
751 $iOffset = $i;
752 $cur_l = 0;
753 $ret = '';
754 $iEncStart = false;
755 } else {
756 $ret .= sprintf("=%02X",ord($string{$i}));
757 }
758 break;
759 case '(':
760 case ')':
761 if ($iEncStart !== false) {
08b7f7cc 762 $aRet[] = substr($string,$iOffset,$iEncStart-$iOffset);
c96c32f4 763 $aRet[] = "=?$default_charset?Q?$ret?=";
764 $iOffset = $i;
765 $cur_l = 0;
766 $ret = '';
767 $iEncStart = false;
768 }
769 break;
770 case ' ':
771 if ($iEncStart !== false) {
098ea084 772 $cur_l++;
773 if ($cur_l > $max_l) {
08b7f7cc 774 $aRet[] = substr($string,$iOffset,$iEncStart-$iOffset);
c96c32f4 775 $aRet[] = "=?$default_charset?Q?$ret?=";
776 $iOffset = $i;
777 $cur_l = 0;
778 $ret = '';
779 $iEncStart = false;
08b7f7cc 780 } else {
c96c32f4 781 $ret .= '_';
782 }
783 }
784 break;
785 default:
786 $k = ord($string{$i});
787 if ($k > 126) {
788 if ($iEncStart === false) {
325978ac 789 // do not start encoding in the middle of a string, also take the rest of the word.
790 $sLeadString = substr($string,0,$i);
791 $aLeadString = explode(' ',$sLeadString);
da2415c1 792 $sToBeEncoded = array_pop($aLeadString);
325978ac 793 $iEncStart = $i - strlen($sToBeEncoded);
794 $ret .= $sToBeEncoded;
795 $cur_l += strlen($sToBeEncoded);
c96c32f4 796 }
797 $cur_l += 3;
08b7f7cc 798 /* first we add the encoded string that reached it's max size */
c96c32f4 799 if ($cur_l > ($max_l-2)) {
08b7f7cc 800 $aRet[] = substr($string,$iOffset,$iEncStart-$iOffset);
801 $aRet[] = "=?$default_charset?Q?$ret?= "; /* the next part is also encoded => separate by space */
c96c32f4 802 $cur_l = 3;
803 $ret = '';
804 $iOffset = $i;
08b7f7cc 805 $iEncStart = $i;
c96c32f4 806 }
08b7f7cc 807 $enc_init = true;
c96c32f4 808 $ret .= sprintf("=%02X", $k);
809 } else {
810 if ($iEncStart !== false) {
098ea084 811 $cur_l++;
812 if ($cur_l > $max_l) {
08b7f7cc 813 $aRet[] = substr($string,$iOffset,$iEncStart-$iOffset);
c96c32f4 814 $aRet[] = "=?$default_charset?Q?$ret?=";
815 $iEncStart = false;
816 $iOffset = $i;
817 $cur_l = 0;
098ea084 818 $ret = '';
08b7f7cc 819 } else {
c96c32f4 820 $ret .= $string{$i};
821 }
3d8371be 822 }
c96c32f4 823 }
824 break;
f7b3ba37 825 }
451f74a2 826 }
793cc001 827
c96c32f4 828 if ($enc_init) {
829 if ($iEncStart !== false) {
830 $aRet[] = substr($string,$iOffset,$iEncStart-$iOffset);
831 $aRet[] = "=?$default_charset?Q?$ret?=";
832 } else {
833 $aRet[] = substr($string,$iOffset);
834 }
835 $string = implode('',$aRet);
451f74a2 836 }
3d8371be 837 return $string;
451f74a2 838}
b74ba498 839
691a2d25 840/* This function trys to locate the entity_id of a specific mime element */
3d8371be 841function find_ent_id($id, $message) {
a171b359 842 for ($i = 0, $ret = ''; $ret == '' && $i < count($message->entities); $i++) {
843 if ($message->entities[$i]->header->type0 == 'multipart') {
3d8371be 844 $ret = find_ent_id($id, $message->entities[$i]);
451f74a2 845 } else {
3d8371be 846 if (strcasecmp($message->entities[$i]->header->id, $id) == 0) {
d8cffbab 847// if (sq_check_save_extension($message->entities[$i])) {
a171b359 848 return $message->entities[$i]->entity_id;
da2415c1 849// }
3d8371be 850 }
a3daaaf3 851 }
451f74a2 852 }
3d8371be 853 return $ret;
451f74a2 854}
a3daaaf3 855
e5e9381a 856function sq_check_save_extension($message) {
857 $filename = $message->getFilename();
858 $ext = substr($filename, strrpos($filename,'.')+1);
859 $save_extensions = array('jpg','jpeg','gif','png','bmp');
3d8371be 860 return in_array($ext, $save_extensions);
e5e9381a 861}
862
863
691a2d25 864/**
865 ** HTMLFILTER ROUTINES
866 */
451f74a2 867
2dd879b8 868/**
869 * This function is more or less a wrapper around stripslashes. Apparently
870 * Explorer is stupid enough to just remove the backslashes and then
871 * execute the content of the attribute as if nothing happened.
872 * Who does that?
873 *
874 * @param attvalue The value of the attribute
875 * @return attvalue The value of the attribute stripslashed.
876 */
877function sq_unbackslash($attvalue){
878 /**
879 * Remove any backslashes. See if there are any first.
880 */
91220ed1 881
2dd879b8 882 if (strstr($attvalue, '\\') !== false){
883 $attvalue = stripslashes($attvalue);
884 }
885 return $attvalue;
886}
887
888/**
889 * Kill any tabs, newlines, or carriage returns. Our friends the
890 * makers of the browser with 95% market value decided that it'd
891 * be funny to make "java[tab]script" be just as good as "javascript".
da2415c1 892 *
2dd879b8 893 * @param attvalue The attribute value before extraneous spaces removed.
894 * @return attvalue The attribute value after extraneous spaces removed.
895 */
896function sq_unspace($attvalue){
897 if (strcspn($attvalue, "\t\r\n") != strlen($attvalue)){
898 $attvalue = str_replace(Array("\t", "\r", "\n"), Array('', '', ''),
899 $attvalue);
900 }
901 return $attvalue;
902}
903
691a2d25 904/**
905 * This function returns the final tag out of the tag name, an array
da2415c1 906 * of attributes, and the type of the tag. This function is called by
691a2d25 907 * sq_sanitize internally.
908 *
909 * @param $tagname the name of the tag.
910 * @param $attary the array of attributes and their values
911 * @param $tagtype The type of the tag (see in comments).
912 * @return a string with the final tag representation.
913 */
914function sq_tagprint($tagname, $attary, $tagtype){
b583c3e8 915 $me = 'sq_tagprint';
3d8371be 916
691a2d25 917 if ($tagtype == 2){
918 $fulltag = '</' . $tagname . '>';
919 } else {
920 $fulltag = '<' . $tagname;
921 if (is_array($attary) && sizeof($attary)){
922 $atts = Array();
923 while (list($attname, $attvalue) = each($attary)){
924 array_push($atts, "$attname=$attvalue");
925 }
926 $fulltag .= ' ' . join(" ", $atts);
927 }
928 if ($tagtype == 3){
b583c3e8 929 $fulltag .= ' /';
691a2d25 930 }
b583c3e8 931 $fulltag .= '>';
451f74a2 932 }
691a2d25 933 return $fulltag;
451f74a2 934}
a3daaaf3 935
691a2d25 936/**
937 * A small helper function to use with array_walk. Modifies a by-ref
938 * value and makes it lowercase.
939 *
940 * @param $val a value passed by-ref.
941 * @return void since it modifies a by-ref value.
942 */
943function sq_casenormalize(&$val){
944 $val = strtolower($val);
945}
451f74a2 946
691a2d25 947/**
948 * This function skips any whitespace from the current position within
949 * a string and to the next non-whitespace value.
da2415c1 950 *
691a2d25 951 * @param $body the string
952 * @param $offset the offset within the string where we should start
953 * looking for the next non-whitespace character.
954 * @return the location within the $body where the next
955 * non-whitespace char is located.
956 */
957function sq_skipspace($body, $offset){
b583c3e8 958 $me = 'sq_skipspace';
3d8371be 959 preg_match('/^(\s*)/s', substr($body, $offset), $matches);
691a2d25 960 if (sizeof($matches{1})){
961 $count = strlen($matches{1});
962 $offset += $count;
451f74a2 963 }
691a2d25 964 return $offset;
451f74a2 965}
a3daaaf3 966
691a2d25 967/**
968 * This function looks for the next character within a string. It's
969 * really just a glorified "strpos", except it catches if failures
970 * nicely.
971 *
972 * @param $body The string to look for needle in.
973 * @param $offset Start looking from this position.
974 * @param $needle The character/string to look for.
975 * @return location of the next occurance of the needle, or
976 * strlen($body) if needle wasn't found.
977 */
978function sq_findnxstr($body, $offset, $needle){
3d8371be 979 $me = 'sq_findnxstr';
691a2d25 980 $pos = strpos($body, $needle, $offset);
981 if ($pos === FALSE){
982 $pos = strlen($body);
451f74a2 983 }
691a2d25 984 return $pos;
451f74a2 985}
a3daaaf3 986
691a2d25 987/**
988 * This function takes a PCRE-style regexp and tries to match it
989 * within the string.
990 *
991 * @param $body The string to look for needle in.
992 * @param $offset Start looking from here.
993 * @param $reg A PCRE-style regex to match.
994 * @return Returns a false if no matches found, or an array
995 * with the following members:
996 * - integer with the location of the match within $body
997 * - string with whatever content between offset and the match
998 * - string with whatever it is we matched
999 */
1000function sq_findnxreg($body, $offset, $reg){
b583c3e8 1001 $me = 'sq_findnxreg';
691a2d25 1002 $matches = Array();
1003 $retarr = Array();
7d06541f 1004 preg_match("%^(.*?)($reg)%si", substr($body, $offset), $matches);
1005 if (!isset($matches{0}) || !$matches{0}){
691a2d25 1006 $retarr = false;
1007 } else {
1008 $retarr{0} = $offset + strlen($matches{1});
1009 $retarr{1} = $matches{1};
1010 $retarr{2} = $matches{2};
1011 }
1012 return $retarr;
1013}
a3daaaf3 1014
691a2d25 1015/**
1016 * This function looks for the next tag.
1017 *
1018 * @param $body String where to look for the next tag.
1019 * @param $offset Start looking from here.
1020 * @return false if no more tags exist in the body, or
1021 * an array with the following members:
1022 * - string with the name of the tag
1023 * - array with attributes and their values
1024 * - integer with tag type (1, 2, or 3)
1025 * - integer where the tag starts (starting "<")
1026 * - integer where the tag ends (ending ">")
1027 * first three members will be false, if the tag is invalid.
1028 */
1029function sq_getnxtag($body, $offset){
b583c3e8 1030 $me = 'sq_getnxtag';
691a2d25 1031 if ($offset > strlen($body)){
1032 return false;
1033 }
1034 $lt = sq_findnxstr($body, $offset, "<");
1035 if ($lt == strlen($body)){
1036 return false;
1037 }
1038 /**
1039 * We are here:
1040 * blah blah <tag attribute="value">
1041 * \---------^
1042 */
1043 $pos = sq_skipspace($body, $lt+1);
1044 if ($pos >= strlen($body)){
1045 return Array(false, false, false, $lt, strlen($body));
1046 }
1047 /**
1048 * There are 3 kinds of tags:
1049 * 1. Opening tag, e.g.:
1050 * <a href="blah">
1051 * 2. Closing tag, e.g.:
1052 * </a>
1053 * 3. XHTML-style content-less tag, e.g.:
1054 * <img src="blah"/>
1055 */
1056 $tagtype = false;
1057 switch (substr($body, $pos, 1)){
3d8371be 1058 case '/':
1059 $tagtype = 2;
1060 $pos++;
1061 break;
1062 case '!':
1063 /**
1064 * A comment or an SGML declaration.
1065 */
1066 if (substr($body, $pos+1, 2) == "--"){
1067 $gt = strpos($body, "-->", $pos);
1068 if ($gt === false){
1069 $gt = strlen($body);
1070 } else {
1071 $gt += 2;
1072 }
1073 return Array(false, false, false, $lt, $gt);
bb8d0799 1074 } else {
3d8371be 1075 $gt = sq_findnxstr($body, $pos, ">");
1076 return Array(false, false, false, $lt, $gt);
1077 }
1078 break;
1079 default:
1080 /**
1081 * Assume tagtype 1 for now. If it's type 3, we'll switch values
1082 * later.
1083 */
1084 $tagtype = 1;
1085 break;
691a2d25 1086 }
a3daaaf3 1087
691a2d25 1088 $tag_start = $pos;
1089 $tagname = '';
1090 /**
1091 * Look for next [\W-_], which will indicate the end of the tag name.
1092 */
1093 $regary = sq_findnxreg($body, $pos, "[^\w\-_]");
1094 if ($regary == false){
1095 return Array(false, false, false, $lt, strlen($body));
1096 }
1097 list($pos, $tagname, $match) = $regary;
1098 $tagname = strtolower($tagname);
1099
1100 /**
1101 * $match can be either of these:
1102 * '>' indicating the end of the tag entirely.
1103 * '\s' indicating the end of the tag name.
1104 * '/' indicating that this is type-3 xhtml tag.
da2415c1 1105 *
691a2d25 1106 * Whatever else we find there indicates an invalid tag.
1107 */
1108 switch ($match){
3d8371be 1109 case '/':
691a2d25 1110 /**
3d8371be 1111 * This is an xhtml-style tag with a closing / at the
1112 * end, like so: <img src="blah"/>. Check if it's followed
1113 * by the closing bracket. If not, then this tag is invalid
691a2d25 1114 */
3d8371be 1115 if (substr($body, $pos, 2) == "/>"){
1116 $pos++;
1117 $tagtype = 3;
1118 } else {
1119 $gt = sq_findnxstr($body, $pos, ">");
1120 $retary = Array(false, false, false, $lt, $gt);
1121 return $retary;
1122 }
1123 case '>':
1124 return Array($tagname, false, $tagtype, $lt, $pos);
1125 break;
1126 default:
1127 /**
1128 * Check if it's whitespace
1129 */
1130 if (!preg_match('/\s/', $match)){
1131 /**
1132 * This is an invalid tag! Look for the next closing ">".
1133 */
7d06541f 1134 $gt = sq_findnxstr($body, $lt, ">");
3d8371be 1135 return Array(false, false, false, $lt, $gt);
1136 }
1137 break;
691a2d25 1138 }
3d8371be 1139
691a2d25 1140 /**
1141 * At this point we're here:
1142 * <tagname attribute='blah'>
1143 * \-------^
1144 *
1145 * At this point we loop in order to find all attributes.
1146 */
1147 $attname = '';
1148 $atttype = false;
1149 $attary = Array();
1150
1151 while ($pos <= strlen($body)){
1152 $pos = sq_skipspace($body, $pos);
1153 if ($pos == strlen($body)){
1154 /**
1155 * Non-closed tag.
1156 */
1157 return Array(false, false, false, $lt, $pos);
1158 }
1159 /**
1160 * See if we arrived at a ">" or "/>", which means that we reached
1161 * the end of the tag.
1162 */
1163 $matches = Array();
164800ad 1164 if (preg_match("%^(\s*)(>|/>)%s", substr($body, $pos), $matches)) {
c828931c 1165 /**
1166 * Yep. So we did.
1167 */
1168 $pos += strlen($matches{1});
1169 if ($matches{2} == "/>"){
1170 $tagtype = 3;
1171 $pos++;
1172 }
1173 return Array($tagname, $attary, $tagtype, $lt, $pos);
1174 }
a3daaaf3 1175
cca46357 1176 /**
691a2d25 1177 * There are several types of attributes, with optional
1178 * [:space:] between members.
1179 * Type 1:
1180 * attrname[:space:]=[:space:]'CDATA'
1181 * Type 2:
1182 * attrname[:space:]=[:space:]"CDATA"
1183 * Type 3:
1184 * attr[:space:]=[:space:]CDATA
1185 * Type 4:
1186 * attrname
cca46357 1187 *
691a2d25 1188 * We leave types 1 and 2 the same, type 3 we check for
1189 * '"' and convert to "&quot" if needed, then wrap in
1190 * double quotes. Type 4 we convert into:
1191 * attrname="yes".
cca46357 1192 */
3f7c623f 1193 $regary = sq_findnxreg($body, $pos, "[^:\w\-_]");
691a2d25 1194 if ($regary == false){
1195 /**
1196 * Looks like body ended before the end of tag.
1197 */
1198 return Array(false, false, false, $lt, strlen($body));
cca46357 1199 }
691a2d25 1200 list($pos, $attname, $match) = $regary;
1201 $attname = strtolower($attname);
1202 /**
1203 * We arrived at the end of attribute name. Several things possible
1204 * here:
1205 * '>' means the end of the tag and this is attribute type 4
1206 * '/' if followed by '>' means the same thing as above
1207 * '\s' means a lot of things -- look what it's followed by.
1208 * anything else means the attribute is invalid.
1209 */
1210 switch($match){
3d8371be 1211 case '/':
691a2d25 1212 /**
3d8371be 1213 * This is an xhtml-style tag with a closing / at the
1214 * end, like so: <img src="blah"/>. Check if it's followed
1215 * by the closing bracket. If not, then this tag is invalid
691a2d25 1216 */
3d8371be 1217 if (substr($body, $pos, 2) == "/>"){
691a2d25 1218 $pos++;
3d8371be 1219 $tagtype = 3;
691a2d25 1220 } else {
3d8371be 1221 $gt = sq_findnxstr($body, $pos, ">");
1222 $retary = Array(false, false, false, $lt, $gt);
1223 return $retary;
1224 }
1225 case '>':
1226 $attary{$attname} = '"yes"';
1227 return Array($tagname, $attary, $tagtype, $lt, $pos);
1228 break;
1229 default:
1230 /**
1231 * Skip whitespace and see what we arrive at.
1232 */
1233 $pos = sq_skipspace($body, $pos);
1234 $char = substr($body, $pos, 1);
1235 /**
1236 * Two things are valid here:
1237 * '=' means this is attribute type 1 2 or 3.
1238 * \w means this was attribute type 4.
1239 * anything else we ignore and re-loop. End of tag and
1240 * invalid stuff will be caught by our checks at the beginning
1241 * of the loop.
1242 */
1243 if ($char == "="){
1244 $pos++;
1245 $pos = sq_skipspace($body, $pos);
691a2d25 1246 /**
3d8371be 1247 * Here are 3 possibilities:
1248 * "'" attribute type 1
1249 * '"' attribute type 2
1250 * everything else is the content of tag type 3
691a2d25 1251 */
3d8371be 1252 $quot = substr($body, $pos, 1);
1253 if ($quot == "'"){
1254 $regary = sq_findnxreg($body, $pos+1, "\'");
1255 if ($regary == false){
1256 return Array(false, false, false, $lt, strlen($body));
1257 }
1258 list($pos, $attval, $match) = $regary;
1259 $pos++;
1260 $attary{$attname} = "'" . $attval . "'";
1261 } else if ($quot == '"'){
1262 $regary = sq_findnxreg($body, $pos+1, '\"');
1263 if ($regary == false){
1264 return Array(false, false, false, $lt, strlen($body));
1265 }
1266 list($pos, $attval, $match) = $regary;
1267 $pos++;
1268 $attary{$attname} = '"' . $attval . '"';
1269 } else {
1270 /**
1271 * These are hateful. Look for \s, or >.
1272 */
1273 $regary = sq_findnxreg($body, $pos, "[\s>]");
1274 if ($regary == false){
1275 return Array(false, false, false, $lt, strlen($body));
1276 }
1277 list($pos, $attval, $match) = $regary;
1278 /**
1279 * If it's ">" it will be caught at the top.
1280 */
1281 $attval = preg_replace("/\"/s", "&quot;", $attval);
1282 $attary{$attname} = '"' . $attval . '"';
7e235a1a 1283 }
3d8371be 1284 } else if (preg_match("|[\w/>]|", $char)) {
691a2d25 1285 /**
3d8371be 1286 * That was attribute type 4.
691a2d25 1287 */
3d8371be 1288 $attary{$attname} = '"yes"';
1289 } else {
1290 /**
1291 * An illegal character. Find next '>' and return.
1292 */
1293 $gt = sq_findnxstr($body, $pos, ">");
1294 return Array(false, false, false, $lt, $gt);
451f74a2 1295 }
3d8371be 1296 break;
691a2d25 1297 }
1298 }
1299 /**
1300 * The fact that we got here indicates that the tag end was never
1301 * found. Return invalid tag indication so it gets stripped.
1302 */
1303 return Array(false, false, false, $lt, strlen($body));
1304}
1305
1306/**
1307 * This function checks attribute values for entity-encoded values
1308 * and returns them translated into 8-bit strings so we can run
1309 * checks on them.
1310 *
1311 * @param $attvalue A string to run entity check against.
1312 * @return Translated value.
1313 */
91220ed1 1314
691a2d25 1315function sq_deent($attvalue){
b583c3e8 1316 $me = 'sq_deent';
691a2d25 1317 /**
1318 * See if we have to run the checks first. All entities must start
1319 * with "&".
1320 */
91220ed1 1321 if (strpos($attvalue, '&') === false){
691a2d25 1322 return $attvalue;
1323 }
1324 /**
1325 * Check named entities first.
1326 */
1327 $trans = get_html_translation_table(HTML_ENTITIES);
1328 /**
1329 * Leave &quot; in, as it can mess us up.
1330 */
1331 $trans = array_flip($trans);
91220ed1 1332 unset($trans{'&quot;'});
691a2d25 1333 while (list($ent, $val) = each($trans)){
91220ed1 1334 $attvalue = preg_replace('/' . $ent . '*/si', $val, $attvalue);
691a2d25 1335 }
1336 /**
1337 * Now translate numbered entities from 1 to 255 if needed.
1338 */
91220ed1 1339 if (strpos($attvalue, '#') !== false){
691a2d25 1340 $omit = Array(34, 39);
91220ed1 1341 for ($asc = 256; $asc >= 0; $asc--){
691a2d25 1342 if (!in_array($asc, $omit)){
1343 $chr = chr($asc);
91220ed1 1344 $octrule = '/\&#0*' . $asc . ';*/si';
1345 $hexrule = '/\&#x0*' . dechex($asc) . ';*/si';
1346 $attvalue = preg_replace($octrule, $chr, $attvalue);
1347 $attvalue = preg_replace($hexrule, $chr, $attvalue);
a3daaaf3 1348 }
691a2d25 1349 }
1350 }
1351 return $attvalue;
1352}
1353
1354/**
1355 * This function runs various checks against the attributes.
1356 *
1357 * @param $tagname String with the name of the tag.
1358 * @param $attary Array with all tag attributes.
1359 * @param $rm_attnames See description for sq_sanitize
1360 * @param $bad_attvals See description for sq_sanitize
1361 * @param $add_attr_to_tag See description for sq_sanitize
1362 * @param $message message object
1363 * @param $id message id
1364 * @return Array with modified attributes.
1365 */
da2415c1 1366function sq_fixatts($tagname,
1367 $attary,
691a2d25 1368 $rm_attnames,
1369 $bad_attvals,
1370 $add_attr_to_tag,
1371 $message,
b3af12ef 1372 $id,
3d8371be 1373 $mailbox
691a2d25 1374 ){
b583c3e8 1375 $me = 'sq_fixatts';
691a2d25 1376 while (list($attname, $attvalue) = each($attary)){
1377 /**
1378 * See if this attribute should be removed.
1379 */
1380 foreach ($rm_attnames as $matchtag=>$matchattrs){
1381 if (preg_match($matchtag, $tagname)){
1382 foreach ($matchattrs as $matchattr){
1383 if (preg_match($matchattr, $attname)){
1384 unset($attary{$attname});
1385 continue;
1386 }
451f74a2 1387 }
451f74a2 1388 }
691a2d25 1389 }
1390 /**
2dd879b8 1391 * Remove any backslashes, entities, and extraneous whitespace.
691a2d25 1392 */
2dd879b8 1393 $attvalue = sq_unbackslash($attvalue);
691a2d25 1394 $attvalue = sq_deent($attvalue);
2dd879b8 1395 $attvalue = sq_unspace($attvalue);
691a2d25 1396
1397 /**
1398 * Now let's run checks on the attvalues.
1399 * I don't expect anyone to comprehend this. If you do,
1400 * get in touch with me so I can drive to where you live and
1401 * shake your hand personally. :)
1402 */
1403 foreach ($bad_attvals as $matchtag=>$matchattrs){
1404 if (preg_match($matchtag, $tagname)){
1405 foreach ($matchattrs as $matchattr=>$valary){
1406 if (preg_match($matchattr, $attname)){
1407 /**
1408 * There are two arrays in valary.
1409 * First is matches.
1410 * Second one is replacements
1411 */
1412 list($valmatch, $valrepl) = $valary;
da2415c1 1413 $newvalue =
691a2d25 1414 preg_replace($valmatch, $valrepl, $attvalue);
1415 if ($newvalue != $attvalue){
1416 $attary{$attname} = $newvalue;
1417 }
1418 }
1419 }
451f74a2 1420 }
a3daaaf3 1421 }
691a2d25 1422 /**
1423 * Turn cid: urls into http-friendly ones.
1424 */
1425 if (preg_match("/^[\'\"]\s*cid:/si", $attvalue)){
b3af12ef 1426 $attary{$attname} = sq_cid2http($message, $id, $attvalue, $mailbox);
691a2d25 1427 }
a3daaaf3 1428 }
691a2d25 1429 /**
1430 * See if we need to append any attributes to this tag.
1431 */
1432 foreach ($add_attr_to_tag as $matchtag=>$addattary){
1433 if (preg_match($matchtag, $tagname)){
1434 $attary = array_merge($attary, $addattary);
1435 }
1436 }
1437 return $attary;
451f74a2 1438}
a3daaaf3 1439
691a2d25 1440/**
1441 * This function edits the style definition to make them friendly and
1442 * usable in squirrelmail.
da2415c1 1443 *
691a2d25 1444 * @param $message the message object
1445 * @param $id the message id
1446 * @param $content a string with whatever is between <style> and </style>
e60a299a 1447 * @param $mailbox the message mailbox
691a2d25 1448 * @return a string with edited content.
1449 */
e60a299a 1450function sq_fixstyle($body, $pos, $message, $id, $mailbox){
691a2d25 1451 global $view_unsafe_images;
b583c3e8 1452 $me = 'sq_fixstyle';
7d06541f 1453 $ret = sq_findnxreg($body, $pos, '</\s*style\s*>');
1454 if ($ret == FALSE){
1455 return array(FALSE, strlen($body));
1456 }
1457 $newpos = $ret[0] + strlen($ret[2]);
1458 $content = $ret[1];
691a2d25 1459 /**
1460 * First look for general BODY style declaration, which would be
1461 * like so:
1462 * body {background: blah-blah}
1463 * and change it to .bodyclass so we can just assign it to a <div>
1464 */
1465 $content = preg_replace("|body(\s*\{.*?\})|si", ".bodyclass\\1", $content);
3d8371be 1466 $secremoveimg = '../images/' . _("sec_remove_eng.png");
691a2d25 1467 /**
1468 * Fix url('blah') declarations.
1469 */
7d06541f 1470 $content = preg_replace("|url\s*\(\s*([\'\"])\s*\S+script\s*:.*?([\'\"])\s*\)|si",
691a2d25 1471 "url(\\1$secremoveimg\\2)", $content);
1472 /**
1473 * Fix url('https*://.*) declarations but only if $view_unsafe_images
1474 * is false.
1475 */
1476 if (!$view_unsafe_images){
7d06541f 1477 $content = preg_replace("|url\s*\(\s*([\'\"])\s*https*:.*?([\'\"])\s*\)|si",
691a2d25 1478 "url(\\1$secremoveimg\\2)", $content);
1479 }
da2415c1 1480
691a2d25 1481 /**
1482 * Fix urls that refer to cid:
1483 */
da2415c1 1484 while (preg_match("|url\s*\(\s*([\'\"]\s*cid:.*?[\'\"])\s*\)|si",
7d06541f 1485 $content, $matches)){
691a2d25 1486 $cidurl = $matches{1};
e60a299a 1487 $httpurl = sq_cid2http($message, $id, $cidurl, $mailbox);
7d06541f 1488 $content = preg_replace("|url\s*\(\s*$cidurl\s*\)|si",
691a2d25 1489 "url($httpurl)", $content);
1490 }
a3daaaf3 1491
691a2d25 1492 /**
bb8d0799 1493 * Fix stupid css declarations which lead to vulnerabilities
691a2d25 1494 * in IE.
1495 */
f83c60a2 1496 $match = Array('/expression/i',
1497 '/behaviou*r/i',
2dd879b8 1498 '/binding/i',
1499 '/include-source/i');
1500 $replace = Array('idiocy', 'idiocy', 'idiocy', 'idiocy');
bb8d0799 1501 $content = preg_replace($match, $replace, $content);
7d06541f 1502 return array($content, $newpos);
691a2d25 1503}
a3daaaf3 1504
691a2d25 1505/**
1506 * This function converts cid: url's into the ones that can be viewed in
1507 * the browser.
1508 *
1509 * @param $message the message object
1510 * @param $id the message id
1511 * @param $cidurl the cid: url.
e60a299a 1512 * @param $mailbox the message mailbox
691a2d25 1513 * @return a string with a http-friendly url
1514 */
b3af12ef 1515function sq_cid2http($message, $id, $cidurl, $mailbox){
691a2d25 1516 /**
1517 * Get rid of quotes.
1518 */
1519 $quotchar = substr($cidurl, 0, 1);
2dd879b8 1520 if ($quotchar == '"' || $quotchar == "'"){
1521 $cidurl = str_replace($quotchar, "", $cidurl);
1522 } else {
1523 $quotchar = '';
1524 }
691a2d25 1525 $cidurl = substr(trim($cidurl), 4);
e5e9381a 1526 $linkurl = find_ent_id($cidurl, $message);
1527 /* in case of non-save cid links $httpurl should be replaced by a sort of
1528 unsave link image */
1529 $httpurl = '';
1530 if ($linkurl) {
6b04287c 1531 $httpurl = $quotchar . SM_PATH . 'src/download.php?absolute_dl=true&amp;' .
e5e9381a 1532 "passed_id=$id&amp;mailbox=" . urlencode($mailbox) .
3d8371be 1533 '&amp;ent_id=' . $linkurl . $quotchar;
e5e9381a 1534 }
691a2d25 1535 return $httpurl;
1536}
1537
1538/**
1539 * This function changes the <body> tag into a <div> tag since we
1540 * can't really have a body-within-body.
1541 *
2dd879b8 1542 * @param $attary an array of attributes and values of <body>
1543 * @param $mailbox mailbox we're currently reading (for cid2http)
1544 * @param $message current message (for cid2http)
1545 * @param $id current message id (for cid2http)
1546 * @return a modified array of attributes to be set for <div>
691a2d25 1547 */
2dd879b8 1548function sq_body2div($attary, $mailbox, $message, $id){
b583c3e8 1549 $me = 'sq_body2div';
3d8371be 1550 $divattary = Array('class' => "'bodyclass'");
b583c3e8 1551 $bgcolor = '#ffffff';
1552 $text = '#000000';
80b4debd 1553 $has_bgc_stl = $has_txt_stl = false;
b583c3e8 1554 $styledef = '';
691a2d25 1555 if (is_array($attary) && sizeof($attary) > 0){
1556 foreach ($attary as $attname=>$attvalue){
1557 $quotchar = substr($attvalue, 0, 1);
1558 $attvalue = str_replace($quotchar, "", $attvalue);
1559 switch ($attname){
3d8371be 1560 case 'background':
da2415c1 1561 $attvalue = sq_cid2http($message, $id,
2dd879b8 1562 $attvalue, $mailbox);
3d8371be 1563 $styledef .= "background-image: url('$attvalue'); ";
1564 break;
1565 case 'bgcolor':
80b4debd 1566 $has_bgc_stl = true;
3d8371be 1567 $styledef .= "background-color: $attvalue; ";
1568 break;
1569 case 'text':
80b4debd 1570 $has_txt_stl = true;
3d8371be 1571 $styledef .= "color: $attvalue; ";
1572 break;
691a2d25 1573 }
a3daaaf3 1574 }
80b4debd 1575 // Outlook defines a white bgcolor and no text color. This can lead to
1576 // white text on a white bg with certain themes.
1577 if ($has_bgc_stl && !$has_txt_stl) {
1578 $styledef .= "color: $text; ";
1579 }
691a2d25 1580 if (strlen($styledef) > 0){
1581 $divattary{"style"} = "\"$styledef\"";
1582 }
1583 }
1584 return $divattary;
1585}
a3daaaf3 1586
691a2d25 1587/**
1588 * This is the main function and the one you should actually be calling.
1589 * There are several variables you should be aware of an which need
1590 * special description.
1591 *
1592 * Since the description is quite lengthy, see it here:
1593 * http://www.mricon.com/html/phpfilter.html
1594 *
1595 * @param $body the string with HTML you wish to filter
1596 * @param $tag_list see description above
1597 * @param $rm_tags_with_content see description above
1598 * @param $self_closing_tags see description above
1599 * @param $force_tag_closing see description above
1600 * @param $rm_attnames see description above
1601 * @param $bad_attvals see description above
1602 * @param $add_attr_to_tag see description above
1603 * @param $message message object
1604 * @param $id message id
1605 * @return sanitized html safe to show on your pages.
1606 */
da2415c1 1607function sq_sanitize($body,
1608 $tag_list,
691a2d25 1609 $rm_tags_with_content,
1610 $self_closing_tags,
1611 $force_tag_closing,
1612 $rm_attnames,
1613 $bad_attvals,
1614 $add_attr_to_tag,
1615 $message,
b3af12ef 1616 $id,
3d8371be 1617 $mailbox
691a2d25 1618 ){
b583c3e8 1619 $me = 'sq_sanitize';
7d06541f 1620 $rm_tags = array_shift($tag_list);
691a2d25 1621 /**
1622 * Normalize rm_tags and rm_tags_with_content.
1623 */
7d06541f 1624 @array_walk($tag_list, 'sq_casenormalize');
691a2d25 1625 @array_walk($rm_tags_with_content, 'sq_casenormalize');
1626 @array_walk($self_closing_tags, 'sq_casenormalize');
1627 /**
1628 * See if tag_list is of tags to remove or tags to allow.
1629 * false means remove these tags
1630 * true means allow these tags
1631 */
691a2d25 1632 $curpos = 0;
1633 $open_tags = Array();
2dd879b8 1634 $trusted = "\n<!-- begin sanitized html -->\n";
691a2d25 1635 $skip_content = false;
bb8d0799 1636 /**
1637 * Take care of netscape's stupid javascript entities like
1638 * &{alert('boo')};
1639 */
1640 $body = preg_replace("/&(\{.*?\};)/si", "&amp;\\1", $body);
691a2d25 1641
7d06541f 1642 while (($curtag = sq_getnxtag($body, $curpos)) != FALSE){
691a2d25 1643 list($tagname, $attary, $tagtype, $lt, $gt) = $curtag;
1644 $free_content = substr($body, $curpos, $lt-$curpos);
1645 /**
1646 * Take care of <style>
1647 */
7d06541f 1648 if ($tagname == "style" && $tagtype == 1){
da2415c1 1649 list($free_content, $curpos) =
e60a299a 1650 sq_fixstyle($body, $gt+1, $message, $id, $mailbox);
7d06541f 1651 if ($free_content != FALSE){
1652 $trusted .= sq_tagprint($tagname, $attary, $tagtype);
1653 $trusted .= $free_content;
1654 $trusted .= sq_tagprint($tagname, false, 2);
1655 }
1656 continue;
691a2d25 1657 }
1658 if ($skip_content == false){
1659 $trusted .= $free_content;
691a2d25 1660 }
1661 if ($tagname != FALSE){
1662 if ($tagtype == 2){
1663 if ($skip_content == $tagname){
1664 /**
1665 * Got to the end of tag we needed to remove.
1666 */
1667 $tagname = false;
1668 $skip_content = false;
1669 } else {
1670 if ($skip_content == false){
c828931c 1671 if ($tagname == "body"){
1672 $tagname = "div";
2dd879b8 1673 }
da2415c1 1674 if (isset($open_tags{$tagname}) &&
2dd879b8 1675 $open_tags{$tagname} > 0){
1676 $open_tags{$tagname}--;
691a2d25 1677 } else {
2dd879b8 1678 $tagname = false;
691a2d25 1679 }
691a2d25 1680 }
1681 }
1682 } else {
1683 /**
1684 * $rm_tags_with_content
1685 */
1686 if ($skip_content == false){
1687 /**
1688 * See if this is a self-closing type and change
1689 * tagtype appropriately.
1690 */
1691 if ($tagtype == 1
1692 && in_array($tagname, $self_closing_tags)){
2dd879b8 1693 $tagtype = 3;
691a2d25 1694 }
1695 /**
1696 * See if we should skip this tag and any content
1697 * inside it.
1698 */
1699 if ($tagtype == 1 &&
1700 in_array($tagname, $rm_tags_with_content)){
1701 $skip_content = $tagname;
1702 } else {
da2415c1 1703 if (($rm_tags == false
691a2d25 1704 && in_array($tagname, $tag_list)) ||
1705 ($rm_tags == true &&
1706 !in_array($tagname, $tag_list))){
1707 $tagname = false;
1708 } else {
2dd879b8 1709 /**
1710 * Convert body into div.
1711 */
1712 if ($tagname == "body"){
1713 $tagname = "div";
da2415c1 1714 $attary = sq_body2div($attary, $mailbox,
2dd879b8 1715 $message, $id);
1716 }
691a2d25 1717 if ($tagtype == 1){
1718 if (isset($open_tags{$tagname})){
1719 $open_tags{$tagname}++;
1720 } else {
1721 $open_tags{$tagname}=1;
1722 }
1723 }
1724 /**
1725 * This is where we run other checks.
1726 */
1727 if (is_array($attary) && sizeof($attary) > 0){
1728 $attary = sq_fixatts($tagname,
1729 $attary,
1730 $rm_attnames,
1731 $bad_attvals,
1732 $add_attr_to_tag,
1733 $message,
b3af12ef 1734 $id,
3d8371be 1735 $mailbox
691a2d25 1736 );
1737 }
1738 }
1739 }
691a2d25 1740 }
1741 }
1742 if ($tagname != false && $skip_content == false){
1743 $trusted .= sq_tagprint($tagname, $attary, $tagtype);
1744 }
691a2d25 1745 }
1746 $curpos = $gt+1;
a3daaaf3 1747 }
691a2d25 1748 $trusted .= substr($body, $curpos, strlen($body)-$curpos);
1749 if ($force_tag_closing == true){
1750 foreach ($open_tags as $tagname=>$opentimes){
1751 while ($opentimes > 0){
1752 $trusted .= '</' . $tagname . '>';
1753 $opentimes--;
1754 }
1755 }
1756 $trusted .= "\n";
1757 }
1758 $trusted .= "<!-- end sanitized html -->\n";
1759 return $trusted;
1760}
451f74a2 1761
691a2d25 1762/**
1763 * This is a wrapper function to call html sanitizing routines.
1764 *
1765 * @param $body the body of the message
1766 * @param $id the id of the message
1767 * @return a string with html safe to display in the browser.
1768 */
7aad7b77 1769function magicHTML($body, $id, $message, $mailbox = 'INBOX') {
691a2d25 1770 global $attachment_common_show_images, $view_unsafe_images,
3d8371be 1771 $has_unsafe_images;
691a2d25 1772 /**
1773 * Don't display attached images in HTML mode.
1774 */
1775 $attachment_common_show_images = false;
1776 $tag_list = Array(
1777 false,
1778 "object",
1779 "meta",
1780 "html",
1781 "head",
cc34b00d 1782 "base",
e18cb31b 1783 "link",
45071bd6 1784 "frame",
2dd879b8 1785 "iframe",
1786 "plaintext",
1787 "marquee"
691a2d25 1788 );
1789
1790 $rm_tags_with_content = Array(
1791 "script",
1792 "applet",
1793 "embed",
2dd879b8 1794 "title",
1795 "frameset",
1796 "xml"
691a2d25 1797 );
1798
1799 $self_closing_tags = Array(
1800 "img",
1801 "br",
1802 "hr",
1803 "input"
1804 );
1805
2dd879b8 1806 $force_tag_closing = true;
691a2d25 1807
1808 $rm_attnames = Array(
1809 "/.*/" =>
1810 Array(
3a50c8d2 1811 "/target/i",
1812 "/^on.*/i",
1813 "/^dynsrc/i",
1814 "/^data.*/i",
1815 "/^lowsrc.*/i"
691a2d25 1816 )
1817 );
1818
1819 $secremoveimg = "../images/" . _("sec_remove_eng.png");
1820 $bad_attvals = Array(
1821 "/.*/" =>
1822 Array(
0a6ec9b5 1823 "/^src|background/i" =>
691a2d25 1824 Array(
1825 Array(
f83c60a2 1826 "/^([\'\"])\s*\S+script\s*:.*([\'\"])/si",
1827 "/^([\'\"])\s*mocha\s*:*.*([\'\"])/si",
1828 "/^([\'\"])\s*about\s*:.*([\'\"])/si"
691a2d25 1829 ),
1830 Array(
1831 "\\1$secremoveimg\\2",
bb8d0799 1832 "\\1$secremoveimg\\2",
3d8371be 1833 "\\1$secremoveimg\\2",
1834 "\\1$secremoveimg\\2"
691a2d25 1835 )
1836 ),
0a6ec9b5 1837 "/^href|action/i" =>
1838 Array(
1839 Array(
f83c60a2 1840 "/^([\'\"])\s*\S+script\s*:.*([\'\"])/si",
1841 "/^([\'\"])\s*mocha\s*:*.*([\'\"])/si",
1842 "/^([\'\"])\s*about\s*:.*([\'\"])/si"
0a6ec9b5 1843 ),
1844 Array(
2dd879b8 1845 "\\1#\\1",
1846 "\\1#\\1",
1847 "\\1#\\1",
1848 "\\1#\\1"
0a6ec9b5 1849 )
1850 ),
f83c60a2 1851 "/^style/i" =>
691a2d25 1852 Array(
1853 Array(
f83c60a2 1854 "/expression/i",
1855 "/binding/i",
1856 "/behaviou*r/i",
2dd879b8 1857 "/include-source/i",
7d06541f 1858 "/url\s*\(\s*([\'\"])\s*\S+script\s*:.*([\'\"])\s*\)/si",
1859 "/url\s*\(\s*([\'\"])\s*mocha\s*:.*([\'\"])\s*\)/si",
1860 "/url\s*\(\s*([\'\"])\s*about\s*:.*([\'\"])\s*\)/si"
691a2d25 1861 ),
1862 Array(
bb8d0799 1863 "idiocy",
3d8371be 1864 "idiocy",
1865 "idiocy",
2dd879b8 1866 "idiocy",
1867 "url(\\1#\\1)",
1868 "url(\\1#\\1)",
1869 "url(\\1#\\1)",
1870 "url(\\1#\\1)"
691a2d25 1871 )
1872 )
1873 )
1874 );
5262d9a6 1875 if( !sqgetGlobalVar('view_unsafe_images', $view_unsafe_images, SQ_GET) ) {
2dd879b8 1876 $view_unsafe_images = false;
45071bd6 1877 }
691a2d25 1878 if (!$view_unsafe_images){
1879 /**
1880 * Remove any references to http/https if view_unsafe_images set
1881 * to false.
1882 */
0a6ec9b5 1883 array_push($bad_attvals{'/.*/'}{'/^src|background/i'}[0],
f83c60a2 1884 '/^([\'\"])\s*https*:.*([\'\"])/si');
0a6ec9b5 1885 array_push($bad_attvals{'/.*/'}{'/^src|background/i'}[1],
2dd879b8 1886 "\\1$secremoveimg\\1");
3a50c8d2 1887 array_push($bad_attvals{'/.*/'}{'/^style/i'}[0],
43d678b6 1888 '/url\(([\'\"])\s*https*:.*([\'\"])\)/si');
3a50c8d2 1889 array_push($bad_attvals{'/.*/'}{'/^style/i'}[1],
2dd879b8 1890 "url(\\1$secremoveimg\\1)");
691a2d25 1891 }
451f74a2 1892
691a2d25 1893 $add_attr_to_tag = Array(
da2415c1 1894 "/^a$/i" =>
2dd879b8 1895 Array('target'=>'"_new"',
1896 'title'=>'"'._("This external link will open in a new window").'"'
1897 )
1898 );
da2415c1 1899 $trusted = sq_sanitize($body,
1900 $tag_list,
691a2d25 1901 $rm_tags_with_content,
1902 $self_closing_tags,
1903 $force_tag_closing,
1904 $rm_attnames,
1905 $bad_attvals,
1906 $add_attr_to_tag,
1907 $message,
b3af12ef 1908 $id,
3d8371be 1909 $mailbox
691a2d25 1910 );
f83c60a2 1911 if (preg_match("|$secremoveimg|i", $trusted)){
691a2d25 1912 $has_unsafe_images = true;
da2415c1 1913 }
691a2d25 1914 return $trusted;
451f74a2 1915}
a4a70693 1916
da2415c1 1917/**
1918 * function SendDownloadHeaders - send file to the browser
1919 *
1920 * Original Source: SM core src/download.php
1921 * moved here to make it available to other code, and separate
1922 * front end from back end functionality.
1923 *
1924 * @param string $type0 first half of mime type
1925 * @param string $type1 second half of mime type
1926 * @param string $filename filename to tell the browser for downloaded file
1927 * @param boolean $force whether to force the download dialog to pop
07c49f57 1928 * @param integer $filesize optional, send the Content-Header and length to the browser
da2415c1 1929 * @return void
1930 */
07c49f57 1931 function SendDownloadHeaders($type0, $type1, $filename, $force, $filesize=0) {
da2415c1 1932 global $languages, $squirrelmail_language;
1933 $isIE = $isIE6 = 0;
1934
1935 sqgetGlobalVar('HTTP_USER_AGENT', $HTTP_USER_AGENT, SQ_SERVER);
1936
1937 if (strstr($HTTP_USER_AGENT, 'compatible; MSIE ') !== false &&
1938 strstr($HTTP_USER_AGENT, 'Opera') === false) {
1939 $isIE = 1;
1940 }
1941
1942 if (strstr($HTTP_USER_AGENT, 'compatible; MSIE 6') !== false &&
1943 strstr($HTTP_USER_AGENT, 'Opera') === false) {
1944 $isIE6 = 1;
1945 }
1946
1947 if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
1948 function_exists($languages[$squirrelmail_language]['XTRA_CODE'])) {
1949 $filename =
1950 $languages[$squirrelmail_language]['XTRA_CODE']('downloadfilename', $filename, $HTTP_USER_AGENT);
1951 } else {
1952 $filename = ereg_replace('[\\/:\*\?"<>\|;]', '_', str_replace('&nbsp;', ' ', $filename));
1953 }
1954
1955 // A Pox on Microsoft and it's Office!
1956 if (!$force) {
1957 // Try to show in browser window
1958 header("Content-Disposition: inline; filename=\"$filename\"");
1959 header("Content-Type: $type0/$type1; name=\"$filename\"");
1960 } else {
1961 // Try to pop up the "save as" box
1962 // IE makes this hard. It pops up 2 save boxes, or none.
1963 // http://support.microsoft.com/support/kb/articles/Q238/5/88.ASP
1964 // But, accordint to Microsoft, it is "RFC compliant but doesn't
1965 // take into account some deviations that allowed within the
1966 // specification." Doesn't that mean RFC non-compliant?
1967 // http://support.microsoft.com/support/kb/articles/Q258/4/52.ASP
1968 //
1969 // The best thing you can do for IE is to upgrade to the latest
1970 // version
1971 if ($isIE && !$isIE6) {
1972 // http://support.microsoft.com/support/kb/articles/Q182/3/15.asp
1973 // Do not have quotes around filename, but that applied to
1974 // "attachment"... does it apply to inline too?
1975 //
1976 // This combination seems to work mostly. IE 5.5 SP 1 has
1977 // known issues (see the Microsoft Knowledge Base)
1978 header("Content-Disposition: inline; filename=$filename");
1979 // This works for most types, but doesn't work with Word files
1980 header("Content-Type: application/download; name=\"$filename\"");
1981
1982 // These are spares, just in case. :-)
1983 //header("Content-Type: $type0/$type1; name=\"$filename\"");
1984 //header("Content-Type: application/x-msdownload; name=\"$filename\"");
1985 //header("Content-Type: application/octet-stream; name=\"$filename\"");
1986 } else {
1987 header("Content-Disposition: attachment; filename=\"$filename\"");
1988 // application/octet-stream forces download for Netscape
1989 header("Content-Type: application/octet-stream; name=\"$filename\"");
1990 }
1991 }
07c49f57 1992
1993 //send the content-length header if the calling function provides it
1994 if ($filesize > 0) {
1995 header("Content-Length: $filesize");
1996 }
1997
1998 } // end fn SendDownlaodHeaders
da2415c1 1999
d6c32258 2000?>