Add error handling to abook db lookups
[squirrelmail.git] / functions / mime.php
1 <?php
2
3 /**
4 * mime.php
5 *
6 * This contains the functions necessary to detect and decode MIME
7 * messages.
8 *
9 * @copyright &copy; 1999-2007 The SquirrelMail Project Team
10 * @license http://opensource.org/licenses/gpl-license.php GNU Public License
11 * @version $Id$
12 * @package squirrelmail
13 */
14
15 /**
16 * dependency information
17 functions dependency
18 mime_structure
19 class/mime/Message.class.php
20 Message::parseStructure
21 functions/page_header.php
22 displayPageHeader
23 functions/display_messages.php
24 plain_error_message
25 mime_fetch_body
26 functions/imap_general.php
27 sqimap_run_command
28 mime_print_body_lines
29
30
31
32 functions/imap.php
33 functions/attachment_common.php
34 functions/display_messages.php
35
36 magicHtml => url_parser
37 translateText => url_parser
38
39 */
40
41
42 /* -------------------------------------------------------------------------- */
43 /* MIME DECODING */
44 /* -------------------------------------------------------------------------- */
45
46 /**
47 * Get the MIME structure
48 *
49 * This function gets the structure of a message and stores it in the "message" class.
50 * It will return this object for use with all relevant header information and
51 * fully parsed into the standard "message" object format.
52 */
53 function mime_structure ($bodystructure, $flags=array()) {
54
55 /* Isolate the body structure and remove beginning and end parenthesis. */
56 $read = trim(substr ($bodystructure, strpos(strtolower($bodystructure), 'bodystructure') + 13));
57 $read = trim(substr ($read, 0, -1));
58 $i = 0;
59 $msg = Message::parseStructure($read,$i);
60
61 if (!is_object($msg)) {
62 global $color, $mailbox;
63 displayPageHeader( $color, $mailbox );
64 $errormessage = _("SquirrelMail could not decode the bodystructure of the message");
65 $errormessage .= '<br />'._("The bodystructure provided by your IMAP server:").'<br /><br />';
66 $errormessage .= '<pre>' . htmlspecialchars($read) . '</pre>';
67 plain_error_message( $errormessage );
68 echo '</body></html>';
69 exit;
70 }
71 if (count($flags)) {
72 foreach ($flags as $flag) {
73 //FIXME: please document why it is we have to check the first char of the flag but we then go ahead and do a full string comparison anyway. Is this a speed enhancement? If not, let's keep it simple and just compare the full string and forget the switch block.
74 $char = strtoupper($flag{1});
75 switch ($char) {
76 case 'S':
77 if (strtolower($flag) == '\\seen') {
78 $msg->is_seen = true;
79 }
80 break;
81 case 'A':
82 if (strtolower($flag) == '\\answered') {
83 $msg->is_answered = true;
84 }
85 break;
86 case 'D':
87 if (strtolower($flag) == '\\deleted') {
88 $msg->is_deleted = true;
89 }
90 break;
91 case 'F':
92 if (strtolower($flag) == '\\flagged') {
93 $msg->is_flagged = true;
94 }
95 break;
96 case 'M':
97 if (strtolower($flag) == '$mdnsent') {
98 $msg->is_mdnsent = true;
99 }
100 break;
101 default:
102 break;
103 }
104 }
105 }
106 // listEntities($msg);
107 return $msg;
108 }
109
110
111
112 /* This starts the parsing of a particular structure. It is called recursively,
113 * so it can be passed different structures. It returns an object of type
114 * $message.
115 * First, it checks to see if it is a multipart message. If it is, then it
116 * handles that as it sees is necessary. If it is just a regular entity,
117 * then it parses it and adds the necessary header information (by calling out
118 * to mime_get_elements()
119 */
120
121 function mime_fetch_body($imap_stream, $id, $ent_id=1, $fetch_size=0) {
122 /* Do a bit of error correction. If we couldn't find the entity id, just guess
123 * that it is the first one. That is usually the case anyway.
124 */
125
126 if (!$ent_id) {
127 $cmd = "FETCH $id BODY[]";
128 } else {
129 $cmd = "FETCH $id BODY[$ent_id]";
130 }
131
132 if ($fetch_size!=0) $cmd .= "<0.$fetch_size>";
133
134 $data = sqimap_run_command ($imap_stream, $cmd, true, $response, $message, TRUE);
135 do {
136 $topline = trim(array_shift($data));
137 } while($topline && ($topline[0] == '*') && !preg_match('/\* [0-9]+ FETCH.*/i', $topline)) ;
138
139 $wholemessage = implode('', $data);
140 if (ereg('\\{([^\\}]*)\\}', $topline, $regs)) {
141 $ret = substr($wholemessage, 0, $regs[1]);
142 /* There is some information in the content info header that could be important
143 * in order to parse html messages. Let's get them here.
144 */
145 // if ($ret{0} == '<') {
146 // $data = sqimap_run_command ($imap_stream, "FETCH $id BODY[$ent_id.MIME]", true, $response, $message, TRUE);
147 // }
148 } else if (ereg('"([^"]*)"', $topline, $regs)) {
149 $ret = $regs[1];
150 } else if ((stristr($topline, 'nil') !== false) && (empty($wholemessage))) {
151 $ret = $wholemessage;
152 } else {
153 global $where, $what, $mailbox, $passed_id, $startMessage;
154 $par = 'mailbox=' . urlencode($mailbox) . '&amp;passed_id=' . $passed_id;
155 if (isset($where) && isset($what)) {
156 $par .= '&amp;where=' . urlencode($where) . '&amp;what=' . urlencode($what);
157 } else {
158 $par .= '&amp;startMessage=' . $startMessage . '&amp;show_more=0';
159 }
160 $par .= '&amp;response=' . urlencode($response) .
161 '&amp;message=' . urlencode($message) .
162 '&amp;topline=' . urlencode($topline);
163
164 echo '<tt><br />' .
165 '<table width="80%"><tr>' .
166 '<tr><td colspan="2">' .
167 _("Body retrieval error. The reason for this is most probably that the message is malformed.") .
168 '</td></tr>' .
169 '<tr><td><b>' . _("Command:") . "</td><td>$cmd</td></tr>" .
170 '<tr><td><b>' . _("Response:") . "</td><td>$response</td></tr>" .
171 '<tr><td><b>' . _("Message:") . "</td><td>$message</td></tr>" .
172 '<tr><td><b>' . _("FETCH line:") . "</td><td>$topline</td></tr>" .
173 "</table><br /></tt></font><hr />";
174
175 $data = sqimap_run_command ($imap_stream, "FETCH $passed_id BODY[]", true, $response, $message, TRUE);
176 array_shift($data);
177 $wholemessage = implode('', $data);
178
179 $ret = $wholemessage;
180 }
181 return $ret;
182 }
183
184 function mime_print_body_lines ($imap_stream, $id, $ent_id=1, $encoding, $rStream='php://stdout') {
185
186 /* Don't kill the connection if the browser is over a dialup
187 * and it would take over 30 seconds to download it.
188 * Don't call set_time_limit in safe mode.
189 */
190
191 if (!ini_get('safe_mode')) {
192 set_time_limit(0);
193 }
194 /* in case of base64 encoded attachments, do not buffer them.
195 Instead, echo the decoded attachment directly to screen */
196 if (strtolower($encoding) == 'base64') {
197 if (!$ent_id) {
198 $query = "FETCH $id BODY[]";
199 } else {
200 $query = "FETCH $id BODY[$ent_id]";
201 }
202 sqimap_run_command($imap_stream,$query,true,$response,$message,TRUE,'sqimap_base64_decode',$rStream,true);
203 } else {
204 $body = mime_fetch_body ($imap_stream, $id, $ent_id);
205 if (is_resource($rStream)) {
206 fputs($rStream,decodeBody($body,$encoding));
207 } else {
208 echo decodeBody($body, $encoding);
209 }
210 }
211
212 /*
213 TODO, use the same method for quoted printable.
214 However, I assume that quoted printable attachments aren't that large
215 so the performancegain / memory usage drop will be minimal.
216 If we decide to add that then we need to adapt sqimap_fread because
217 we need to split te result on \n and fread doesn't stop at \n. That
218 means we also should provide $results from sqimap_fread (by ref) to
219 te function and set $no_return to false. The $filter function for
220 quoted printable should handle unsetting of $results.
221 */
222 /*
223 TODO 2: find out how we write to the output stream php://stdout. fwrite
224 doesn't work because 'php://stdout isn't a stream.
225 */
226
227 return;
228 }
229
230 /* -[ END MIME DECODING ]----------------------------------------------------------- */
231
232 /* This is here for debugging purposes. It will print out a list
233 * of all the entity IDs that are in the $message object.
234 */
235 function listEntities ($message) {
236 if ($message) {
237 echo "<tt>" . $message->entity_id . ' : ' . $message->type0 . '/' . $message->type1 . ' parent = '. $message->parent->entity_id. '<br />';
238 for ($i = 0; isset($message->entities[$i]); $i++) {
239 echo "$i : ";
240 $msg = listEntities($message->entities[$i]);
241
242 if ($msg) {
243 echo "return: ";
244 return $msg;
245 }
246 }
247 }
248 }
249
250 function getPriorityStr($priority) {
251 $priority_level = substr($priority,0,1);
252
253 switch($priority_level) {
254 /* Check for a higher then normal priority. */
255 case '1':
256 case '2':
257 $priority_string = _("High");
258 break;
259
260 /* Check for a lower then normal priority. */
261 case '4':
262 case '5':
263 $priority_string = _("Low");
264 break;
265
266 /* Check for a normal priority. */
267 case '3':
268 default:
269 $priority_level = '3';
270 $priority_string = _("Normal");
271 break;
272
273 }
274 return $priority_string;
275 }
276
277 /* returns a $message object for a particular entity id */
278 function getEntity ($message, $ent_id) {
279 return $message->getEntity($ent_id);
280 }
281
282 /* translateText
283 * Extracted from strings.php 23/03/2002
284 */
285
286 function translateText(&$body, $wrap_at, $charset) {
287 global $where, $what; /* from searching */
288 global $color; /* color theme */
289
290 // require_once(SM_PATH . 'functions/url_parser.php');
291
292 $body_ary = explode("\n", $body);
293 for ($i=0; $i < count($body_ary); $i++) {
294 $line = $body_ary[$i];
295 if (strlen($line) - 2 >= $wrap_at) {
296 sqWordWrap($line, $wrap_at, $charset);
297 }
298 $line = charset_decode($charset, $line);
299 $line = str_replace("\t", ' ', $line);
300
301 parseUrl ($line);
302
303 $quotes = 0;
304 $pos = 0;
305 $j = strlen($line);
306
307 while ($pos < $j) {
308 if ($line[$pos] == ' ') {
309 $pos++;
310 } else if (strpos($line, '&gt;', $pos) === $pos) {
311 $pos += 4;
312 $quotes++;
313 } else {
314 break;
315 }
316 }
317
318 if ($quotes % 2) {
319 $line = '<span class="quote1">' . $line . '</span>';
320 } elseif ($quotes) {
321 $line = '<span class="quote2">' . $line . '</span>';
322 }
323
324 $body_ary[$i] = $line;
325 }
326 $body = '<pre>' . implode("\n", $body_ary) . '</pre>';
327 }
328
329 /**
330 * This returns a parsed string called $body. That string can then
331 * be displayed as the actual message in the HTML. It contains
332 * everything needed, including HTML Tags, Attachments at the
333 * bottom, etc.
334 *
335 * Since 1.2.0 function uses message_body hook.
336 * Till 1.3.0 function included output of formatAttachments().
337 *
338 * @param resource $imap_stream imap connection resource
339 * @param object $message squirrelmail message object
340 * @param array $color squirrelmail color theme array
341 * @param integer $wrap_at number of characters per line
342 * @param string $ent_num (since 1.3.0) message part id
343 * @param integer $id (since 1.3.0) message id
344 * @param string $mailbox (since 1.3.0) imap folder name
345 * @return string html formated message text
346 */
347 function formatBody($imap_stream, $message, $color, $wrap_at, $ent_num, $id, $mailbox='INBOX') {
348 /* This if statement checks for the entity to show as the
349 * primary message. To add more of them, just put them in the
350 * order that is their priority.
351 */
352 global $startMessage, $languages, $squirrelmail_language,
353 $show_html_default, $sort, $has_unsafe_images, $passed_ent_id,
354 $use_iframe, $iframe_height, $download_and_unsafe_link,
355 $download_href, $unsafe_image_toggle_href, $unsafe_image_toggle_text,
356 $oTemplate, $nbsp;
357
358 // workaround for not updated config.php
359 if (! isset($use_iframe)) $use_iframe = false;
360
361 // If there's no "view_unsafe_images" variable in the URL, turn unsafe
362 // images off by default.
363 sqgetGlobalVar('view_unsafe_images', $view_unsafe_images, SQ_GET, FALSE);
364
365 $body = '';
366 $urlmailbox = urlencode($mailbox);
367 $body_message = getEntity($message, $ent_num);
368 if (($body_message->header->type0 == 'text') ||
369 ($body_message->header->type0 == 'rfc822')) {
370 $body = mime_fetch_body ($imap_stream, $id, $ent_num);
371 $body = decodeBody($body, $body_message->header->encoding);
372
373 if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
374 function_exists($languages[$squirrelmail_language]['XTRA_CODE'] . '_decode')) {
375 if (mb_detect_encoding($body) != 'ASCII') {
376 $body = call_user_func($languages[$squirrelmail_language]['XTRA_CODE'] . '_decode',$body);
377 }
378 }
379
380 /* As of 1.5.2, $body is passed (and modified) by reference */
381 do_hook('message_body', $body);
382
383 /* If there are other types that shouldn't be formatted, add
384 * them here.
385 */
386
387 if ($body_message->header->type1 == 'html') {
388 if ($show_html_default <> 1) {
389 $entity_conv = array('&nbsp;' => ' ',
390 '<p>' => "\n",
391 '<P>' => "\n",
392 '<br>' => "\n",
393 '<BR>' => "\n",
394 '<br />' => "\n",
395 '<BR />' => "\n",
396 '&gt;' => '>',
397 '&lt;' => '<');
398 $body = strtr($body, $entity_conv);
399 $body = strip_tags($body);
400 $body = trim($body);
401 translateText($body, $wrap_at,
402 $body_message->header->getParameter('charset'));
403 } elseif ($use_iframe) {
404 /**
405 * If we don't add html message between iframe tags,
406 * we must detect unsafe images and modify $has_unsafe_images.
407 */
408 $html_body = magicHTML($body, $id, $message, $mailbox);
409 // Convert character set in order to display html mails in different character set
410 $html_body = charset_decode($body_message->header->getParameter('charset'),$html_body,false,true);
411
412 // creating iframe url
413 $iframeurl=sqm_baseuri().'src/view_html.php?'
414 . 'mailbox=' . $urlmailbox
415 . '&amp;passed_id=' . $id
416 . '&amp;ent_id=' . $ent_num
417 . '&amp;view_unsafe_images=' . (int) $view_unsafe_images;
418
419 global $oTemplate;
420 $oTemplate->assign('iframe_url', $iframeurl);
421 $oTemplate->assign('iframe_height', $iframe_height);
422 $oTemplate->assign('html_body', $html_body);
423
424 $body = $oTemplate->fetch('read_html_iframe.tpl');
425 } else {
426 // old way of html rendering
427 /**
428 * convert character set. charset_decode does not remove html special chars
429 * applied by magicHTML functions and does not sanitize them second time if
430 * fourth argument is true.
431 */
432 $charset = $body_message->header->getParameter('charset');
433 if (!empty($charset)) {
434 $body = charset_decode($charset,$body,false,true);
435 }
436 $body = magicHTML($body, $id, $message, $mailbox);
437 }
438 } else {
439 translateText($body, $wrap_at,
440 $body_message->header->getParameter('charset'));
441 }
442
443 /*
444 * Previously the links for downloading and unsafe images were printed
445 * under the mail. By putting the links in a global variable we can
446 * print it in the toolbar where it belongs. Since the original code was
447 * in this place it's left here. It might be possible to move it to some
448 * other place if that makes sense. The possibility to do so has not
449 * been evaluated yet.
450 */
451
452 // Initialize the global variable to an empty string.
453 // FIXME: To have $download_and_unsafe_link as a global variable might not be needed since the use of separate variables ($download_href, $unsafe_image_toggle_href, and $unsafe_image_toggle_text) for the templates was introduced.
454 $download_and_unsafe_link = '';
455
456 // Prepare and build a link for downloading the mail.
457 $link = 'passed_id=' . $id . '&amp;ent_id='.$ent_num.
458 '&amp;mailbox=' . $urlmailbox .'&amp;sort=' . $sort .
459 '&amp;startMessage=' . $startMessage . '&amp;show_more=0';
460 if (isset($passed_ent_id)) {
461 $link .= '&amp;passed_ent_id='.$passed_ent_id;
462 }
463 $download_href = SM_PATH . 'src/download.php?absolute_dl=true&amp;' . $link;
464
465 // Always add the link for downloading the mail as a file to the global
466 // variable.
467 $download_and_unsafe_link .= "$nbsp|$nbsp"
468 . create_hyperlink($download_href, _("Download this as a file"));
469
470 // Find out the right text to use in the link depending on the
471 // circumstances. If the unsafe images are displayed the link should
472 // hide them, if they aren't displayed the link should only appear if
473 // the mail really contains unsafe images.
474 if ($view_unsafe_images) {
475 $text = _("Hide Unsafe Images");
476 } else {
477 if (isset($has_unsafe_images) && $has_unsafe_images) {
478 $link .= '&amp;view_unsafe_images=1';
479 $text = _("View Unsafe Images");
480 } else {
481 $text = '';
482 }
483 }
484
485 // Only create a link for unsafe images if there's need for one. If so:
486 // add it to the global variable.
487 if($text != '') {
488 $unsafe_image_toggle_href = SM_PATH . 'src/read_body.php?'.$link;
489 $unsafe_image_toggle_text = $text;
490 $download_and_unsafe_link .= "$nbsp|$nbsp"
491 . create_hyperlink($unsafe_image_toggle_href, $text);
492 }
493 }
494 return $body;
495 }
496
497 /**
498 * Generate attachments array for passing to templates.
499 *
500 * @since 1.5.2
501 * @param object $message SquirrelMail message object
502 * @param array $exclude_id message parts that are not attachments.
503 * @param string $mailbox mailbox name
504 * @param integer $id message id
505 */
506 function buildAttachmentArray($message, $exclude_id, $mailbox, $id) {
507 global $where, $what, $startMessage, $color, $passed_ent_id, $base_uri;
508
509 $att_ar = $message->getAttachments($exclude_id);
510 $urlMailbox = urlencode($mailbox);
511
512 $attachments = array();
513 foreach ($att_ar as $att) {
514 $ent = $att->entity_id;
515 $header = $att->header;
516 $type0 = strtolower($header->type0);
517 $type1 = strtolower($header->type1);
518 $name = '';
519 $links = array();
520 $links['download link']['text'] = _("Download");
521 $links['download link']['href'] = $base_uri .
522 "src/download.php?absolute_dl=true&amp;passed_id=$id&amp;mailbox=$urlMailbox&amp;ent_id=$ent";
523
524 if ($type0 =='message' && $type1 == 'rfc822') {
525 $default_page = $base_uri . 'src/read_body.php';
526 $rfc822_header = $att->rfc822_header;
527 $filename = $rfc822_header->subject;
528 if (trim( $filename ) == '') {
529 $filename = 'untitled-[' . $ent . ']' ;
530 }
531 $from_o = $rfc822_header->from;
532 if (is_object($from_o)) {
533 $from_name = decodeHeader($from_o->getAddress(false));
534 } elseif (is_array($from_o) && count($from_o) && is_object($from_o[0])) {
535 // something weird happens when a digest message is opened and you return to the digest
536 // now the from object is part of an array. Probably the parseHeader call overwrites the info
537 // retrieved from the bodystructure in a different way. We need to fix this later.
538 // possible starting point, do not fetch header we already have and inspect how
539 // the rfc822_header object behaves.
540 $from_name = decodeHeader($from_o[0]->getAddress(false));
541 } else {
542 $from_name = _("Unknown sender");
543 }
544 $description = _("From").': '.$from_name;
545 } else {
546 $default_page = $base_uri . 'src/download.php';
547 $filename = $att->getFilename();
548 if ($header->description) {
549 $description = decodeHeader($header->description);
550 } else {
551 $description = '';
552 }
553 }
554
555 $display_filename = $filename;
556 if (isset($passed_ent_id)) {
557 $passed_ent_id_link = '&amp;passed_ent_id='.$passed_ent_id;
558 } else {
559 $passed_ent_id_link = '';
560 }
561 $defaultlink = $default_page . "?startMessage=$startMessage"
562 . "&amp;passed_id=$id&amp;mailbox=$urlMailbox"
563 . '&amp;ent_id='.$ent.$passed_ent_id_link;
564 if ($where && $what) {
565 $defaultlink .= '&amp;where='. urlencode($where).'&amp;what='.urlencode($what);
566 }
567 // IE does make use of mime content sniffing. Forcing a download
568 // prohibit execution of XSS inside an application/octet-stream attachment
569 if ($type0 == 'application' && $type1 == 'octet-stream') {
570 $defaultlink .= '&amp;absolute_dl=true';
571 }
572
573 /* This executes the attachment hook with a specific MIME-type.
574 * If that doesn't have results, it tries if there's a rule
575 * for a more generic type. Finally, a hook for ALL attachment
576 * types is run as well.
577 */
578 // First remember the default link.
579 $defaultlink_orig = $defaultlink;
580
581 /* The API for this hook has changed as of 1.5.2 so that all plugin
582 arguments are passed in an array instead of each their own plugin
583 argument, and arguments are passed by reference, so instead of
584 returning any changes, changes should simply be made to the original
585 arguments themselves. */
586 $temp = array(&$links, &$startMessage, &$id, &$urlMailbox, &$ent,
587 &$defaultlink, &$display_filename, &$where, &$what);
588 do_hook("attachment $type0/$type1", $temp);
589 if(count($links) <= 1 && $defaultlink == $defaultlink_orig) {
590 /* The API for this hook has changed as of 1.5.2 so that all plugin
591 arguments are passed in an array instead of each their own plugin
592 argument, and arguments are passed by reference, so instead of
593 returning any changes, changes should simply be made to the original
594 arguments themselves. */
595 $temp = array(&$links, &$startMessage, &$id, &$urlMailbox, &$ent,
596 &$defaultlink, &$display_filename, &$where, &$what);
597 do_hook("attachment $type0/*", $temp);
598 }
599 /* The API for this hook has changed as of 1.5.2 so that all plugin
600 arguments are passed in an array instead of each their own plugin
601 argument, and arguments are passed by reference, so instead of
602 returning any changes, changes should simply be made to the original
603 arguments themselves. */
604 $temp = array(&$links, &$startMessage, &$id, &$urlMailbox, &$ent,
605 &$defaultlink, &$display_filename, &$where, &$what);
606 // Do not let a generic plugin change the default link if a more
607 // specialized one already did it...
608 if ($defaultlink != $defaultlink_orig) {
609 $dummy = '';
610 $temp[5] = &$dummy;
611 }
612 do_hook("attachment */*", $temp);
613
614 $this_attachment = array();
615 $this_attachment['Name'] = decodeHeader($display_filename);
616 $this_attachment['Description'] = $description;
617 $this_attachment['DefaultHREF'] = $defaultlink;
618 $this_attachment['DownloadHREF'] = $links['download link']['href'];
619 $this_attachment['ViewHREF'] = isset($links['attachment_common']) ? $links['attachment_common']['href'] : '';
620 $this_attachment['Size'] = $header->size;
621 $this_attachment['ContentType'] = htmlspecialchars($type0 .'/'. $type1);
622 $this_attachment['OtherLinks'] = array();
623 foreach ($links as $val) {
624 if ($val['text']==_("Download") || $val['text'] == _("View"))
625 continue;
626 if (empty($val['text']) && empty($val['extra']))
627 continue;
628
629 $temp = array();
630 $temp['HREF'] = $val['href'];
631 $temp['Text'] = (empty($val['text']) ? '' : $val['text']) . (empty($val['extra']) ? '' : $val['extra']);
632 $this_attachment['OtherLinks'][] = $temp;
633 }
634 $attachments[] = $this_attachment;
635
636 unset($links);
637 }
638
639 return $attachments;
640 }
641
642 /**
643 * Displays attachment links and information
644 *
645 * Since 1.3.0 function is not included in formatBody() call.
646 *
647 * Since 1.0.2 uses attachment $type0/$type1 hook.
648 * Since 1.2.5 uses attachment $type0/* hook.
649 * Since 1.5.0 uses attachments_bottom hook.
650 * Since 1.5.2 uses templates and does *not* return a value.
651 *
652 * @param object $message SquirrelMail message object
653 * @param array $exclude_id message parts that are not attachments.
654 * @param string $mailbox mailbox name
655 * @param integer $id message id
656 */
657 function formatAttachments($message, $exclude_id, $mailbox, $id) {
658 global $oTemplate;
659
660 $attach = buildAttachmentArray($message, $exclude_id, $mailbox, $id);
661
662 $oTemplate->assign('attachments', $attach);
663 $oTemplate->display('read_attachments.tpl');
664 }
665
666 function sqimap_base64_decode(&$string) {
667
668 // Base64 encoded data goes in pairs of 4 bytes. To achieve on the
669 // fly decoding (to reduce memory usage) you have to check if the
670 // data has incomplete pairs
671
672 // Remove the noise in order to check if the 4 bytes pairs are complete
673 $string = str_replace(array("\r\n","\n", "\r", " "),array('','','',''),$string);
674
675 $sStringRem = '';
676 $iMod = strlen($string) % 4;
677 if ($iMod) {
678 $sStringRem = substr($string,-$iMod);
679 // Check if $sStringRem contains padding characters
680 if (substr($sStringRem,-1) != '=') {
681 $string = substr($string,0,-$iMod);
682 } else {
683 $sStringRem = '';
684 }
685 }
686 $string = base64_decode($string);
687 return $sStringRem;
688 }
689
690 /**
691 * Decodes encoded message body
692 *
693 * This function decodes the body depending on the encoding type.
694 * Currently quoted-printable and base64 encodings are supported.
695 * decode_body hook was added to this function in 1.4.2/1.5.0
696 * @param string $body encoded message body
697 * @param string $encoding used encoding
698 * @return string decoded string
699 * @since 1.0
700 */
701 function decodeBody($body, $encoding) {
702
703 $body = str_replace("\r\n", "\n", $body);
704 $encoding = strtolower($encoding);
705
706 $encoding_handler = do_hook('decode_body', $encoding);
707
708
709 // plugins get first shot at decoding the body
710 //
711 if (!empty($encoding_handler) && function_exists($encoding_handler)) {
712 $body = $encoding_handler('decode', $body);
713
714 } elseif ($encoding == 'quoted-printable' ||
715 $encoding == 'quoted_printable') {
716 /**
717 * quoted_printable_decode() function is broken in older
718 * php versions. Text with \r\n decoding was fixed only
719 * in php 4.3.0. Minimal code requirement 4.0.4 +
720 * str_replace("\r\n", "\n", $body); call.
721 */
722 $body = quoted_printable_decode($body);
723 } elseif ($encoding == 'base64') {
724 $body = base64_decode($body);
725 }
726
727 // All other encodings are returned raw.
728 return $body;
729 }
730
731 /**
732 * Decodes headers
733 *
734 * This function decodes strings that are encoded according to
735 * RFC1522 (MIME Part Two: Message Header Extensions for Non-ASCII Text).
736 * Patched by Christian Schmidt <christian@ostenfeld.dk> 23/03/2002
737 *
738 * @param string $string header string that has to be made readable
739 * @param boolean $utfencode change message in order to be readable on user's charset. defaults to true
740 * @param boolean $htmlsafe preserve spaces and sanitize html special characters. defaults to true
741 * @param boolean $decide decide if string can be utfencoded. defaults to false
742 * @return string decoded header string
743 */
744 function decodeHeader ($string, $utfencode=true,$htmlsafe=true,$decide=false) {
745 global $languages, $squirrelmail_language,$default_charset;
746 if (is_array($string)) {
747 $string = implode("\n", $string);
748 }
749
750 if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
751 function_exists($languages[$squirrelmail_language]['XTRA_CODE'] . '_decodeheader')) {
752 $string = call_user_func($languages[$squirrelmail_language]['XTRA_CODE'] . '_decodeheader', $string);
753 // Do we need to return at this point?
754 // return $string;
755 }
756 $i = 0;
757 $iLastMatch = -2;
758 $encoded = true;
759
760 $aString = explode(' ',$string);
761 $ret = '';
762 foreach ($aString as $chunk) {
763 if ($encoded && $chunk === '') {
764 continue;
765 } elseif ($chunk === '') {
766 $ret .= ' ';
767 continue;
768 }
769 $encoded = false;
770 /* if encoded words are not separated by a linear-space-white we still catch them */
771 $j = $i-1;
772
773 while ($match = preg_match('/^(.*)=\?([^?]*)\?(Q|B)\?([^?]*)\?=(.*)$/Ui',$chunk,$res)) {
774 /* if the last chunk isn't an encoded string then put back the space, otherwise don't */
775 if ($iLastMatch !== $j) {
776 if ($htmlsafe) {
777 $ret .= '&#32;';
778 } else {
779 $ret .= ' ';
780 }
781 }
782 $iLastMatch = $i;
783 $j = $i;
784 if ($htmlsafe) {
785 $ret .= htmlspecialchars($res[1]);
786 } else {
787 $ret .= $res[1];
788 }
789 $encoding = ucfirst($res[3]);
790
791 /* decide about valid decoding */
792 if ($decide && is_conversion_safe($res[2])) {
793 $utfencode=true;
794 $can_be_encoded=true;
795 } else {
796 $can_be_encoded=false;
797 }
798 switch ($encoding)
799 {
800 case 'B':
801 $replace = base64_decode($res[4]);
802 if ($utfencode) {
803 if ($can_be_encoded) {
804 /* convert string to different charset,
805 * if functions asks for it (usually in compose)
806 */
807 $ret .= charset_convert($res[2],$replace,$default_charset,$htmlsafe);
808 } else {
809 // convert string to html codes in order to display it
810 $ret .= charset_decode($res[2],$replace);
811 }
812 } else {
813 if ($htmlsafe) {
814 $replace = htmlspecialchars($replace);
815 }
816 $ret.= $replace;
817 }
818 break;
819 case 'Q':
820 $replace = str_replace('_', ' ', $res[4]);
821 $replace = preg_replace('/=([0-9a-f]{2})/ie', 'chr(hexdec("\1"))',
822 $replace);
823 if ($utfencode) {
824 if ($can_be_encoded) {
825 /* convert string to different charset,
826 * if functions asks for it (usually in compose)
827 */
828 $replace = charset_convert($res[2], $replace,$default_charset,$htmlsafe);
829 } else {
830 // convert string to html codes in order to display it
831 $replace = charset_decode($res[2], $replace);
832 }
833 } else {
834 if ($htmlsafe) {
835 $replace = htmlspecialchars($replace);
836 }
837 }
838 $ret .= $replace;
839 break;
840 default:
841 break;
842 }
843 $chunk = $res[5];
844 $encoded = true;
845 }
846 if (!$encoded) {
847 if ($htmlsafe) {
848 $ret .= '&#32;';
849 } else {
850 $ret .= ' ';
851 }
852 }
853
854 if (!$encoded && $htmlsafe) {
855 $ret .= htmlspecialchars($chunk);
856 } else {
857 $ret .= $chunk;
858 }
859 ++$i;
860 }
861 /* remove the first added space */
862 if ($ret) {
863 if ($htmlsafe) {
864 $ret = substr($ret,5);
865 } else {
866 $ret = substr($ret,1);
867 }
868 }
869
870 return $ret;
871 }
872
873 /**
874 * Encodes header
875 *
876 * Function uses XTRA_CODE _encodeheader function, if such function exists.
877 *
878 * Function uses Q encoding by default and encodes a string according to RFC
879 * 1522 for use in headers if it contains 8-bit characters or anything that
880 * looks like it should be encoded.
881 *
882 * Function switches to B encoding and encodeHeaderBase64() function, if
883 * string is 8bit and multibyte character set supported by mbstring extension
884 * is used. It can cause E_USER_NOTICE errors, if interface is used with
885 * multibyte character set unsupported by mbstring extension.
886 *
887 * @param string $string header string, that has to be encoded
888 * @return string quoted-printable encoded string
889 * @todo make $mb_charsets system wide constant
890 */
891 function encodeHeader ($string) {
892 global $default_charset, $languages, $squirrelmail_language;
893
894 if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
895 function_exists($languages[$squirrelmail_language]['XTRA_CODE'] . '_encodeheader')) {
896 return call_user_func($languages[$squirrelmail_language]['XTRA_CODE'] . '_encodeheader', $string);
897 }
898
899 // Use B encoding for multibyte charsets
900 $mb_charsets = array('utf-8','big5','gb2313','euc-kr');
901 if (in_array($default_charset,$mb_charsets) &&
902 in_array($default_charset,sq_mb_list_encodings()) &&
903 sq_is8bit($string)) {
904 return encodeHeaderBase64($string,$default_charset);
905 } elseif (in_array($default_charset,$mb_charsets) &&
906 sq_is8bit($string) &&
907 ! in_array($default_charset,sq_mb_list_encodings())) {
908 // Add E_USER_NOTICE error here (can cause 'Cannot add header information' warning in compose.php)
909 // trigger_error('encodeHeader: Multibyte character set unsupported by mbstring extension.',E_USER_NOTICE);
910 }
911
912 // Encode only if the string contains 8-bit characters or =?
913 $j = strlen($string);
914 $max_l = 75 - strlen($default_charset) - 7;
915 $aRet = array();
916 $ret = '';
917 $iEncStart = $enc_init = false;
918 $cur_l = $iOffset = 0;
919 for($i = 0; $i < $j; ++$i) {
920 switch($string{$i})
921 {
922 case '=':
923 case '<':
924 case '>':
925 case ',':
926 case '?':
927 case '_':
928 if ($iEncStart === false) {
929 $iEncStart = $i;
930 }
931 $cur_l+=3;
932 if ($cur_l > ($max_l-2)) {
933 /* if there is an stringpart that doesn't need encoding, add it */
934 $aRet[] = substr($string,$iOffset,$iEncStart-$iOffset);
935 $aRet[] = "=?$default_charset?Q?$ret?=";
936 $iOffset = $i;
937 $cur_l = 0;
938 $ret = '';
939 $iEncStart = false;
940 } else {
941 $ret .= sprintf("=%02X",ord($string{$i}));
942 }
943 break;
944 case '(':
945 case ')':
946 if ($iEncStart !== false) {
947 $aRet[] = substr($string,$iOffset,$iEncStart-$iOffset);
948 $aRet[] = "=?$default_charset?Q?$ret?=";
949 $iOffset = $i;
950 $cur_l = 0;
951 $ret = '';
952 $iEncStart = false;
953 }
954 break;
955 case ' ':
956 if ($iEncStart !== false) {
957 $cur_l++;
958 if ($cur_l > $max_l) {
959 $aRet[] = substr($string,$iOffset,$iEncStart-$iOffset);
960 $aRet[] = "=?$default_charset?Q?$ret?=";
961 $iOffset = $i;
962 $cur_l = 0;
963 $ret = '';
964 $iEncStart = false;
965 } else {
966 $ret .= '_';
967 }
968 }
969 break;
970 default:
971 $k = ord($string{$i});
972 if ($k > 126) {
973 if ($iEncStart === false) {
974 // do not start encoding in the middle of a string, also take the rest of the word.
975 $sLeadString = substr($string,0,$i);
976 $aLeadString = explode(' ',$sLeadString);
977 $sToBeEncoded = array_pop($aLeadString);
978 $iEncStart = $i - strlen($sToBeEncoded);
979 $ret .= $sToBeEncoded;
980 $cur_l += strlen($sToBeEncoded);
981 }
982 $cur_l += 3;
983 /* first we add the encoded string that reached it's max size */
984 if ($cur_l > ($max_l-2)) {
985 $aRet[] = substr($string,$iOffset,$iEncStart-$iOffset);
986 $aRet[] = "=?$default_charset?Q?$ret?= "; /* the next part is also encoded => separate by space */
987 $cur_l = 3;
988 $ret = '';
989 $iOffset = $i;
990 $iEncStart = $i;
991 }
992 $enc_init = true;
993 $ret .= sprintf("=%02X", $k);
994 } else {
995 if ($iEncStart !== false) {
996 $cur_l++;
997 if ($cur_l > $max_l) {
998 $aRet[] = substr($string,$iOffset,$iEncStart-$iOffset);
999 $aRet[] = "=?$default_charset?Q?$ret?=";
1000 $iEncStart = false;
1001 $iOffset = $i;
1002 $cur_l = 0;
1003 $ret = '';
1004 } else {
1005 $ret .= $string{$i};
1006 }
1007 }
1008 }
1009 break;
1010 }
1011 }
1012
1013 if ($enc_init) {
1014 if ($iEncStart !== false) {
1015 $aRet[] = substr($string,$iOffset,$iEncStart-$iOffset);
1016 $aRet[] = "=?$default_charset?Q?$ret?=";
1017 } else {
1018 $aRet[] = substr($string,$iOffset);
1019 }
1020 $string = implode('',$aRet);
1021 }
1022 return $string;
1023 }
1024
1025 /**
1026 * Encodes string according to rfc2047 B encoding header formating rules
1027 *
1028 * It is recommended way to encode headers with character sets that store
1029 * symbols in more than one byte.
1030 *
1031 * Function requires mbstring support. If required mbstring functions are missing,
1032 * function returns false and sets E_USER_WARNING level error message.
1033 *
1034 * Minimal requirements - php 4.0.6 with mbstring extension. Please note,
1035 * that mbstring functions will generate E_WARNING errors, if unsupported
1036 * character set is used. mb_encode_mimeheader function provided by php
1037 * mbstring extension is not used in order to get better control of header
1038 * encoding.
1039 *
1040 * Used php code functions - function_exists(), trigger_error(), strlen()
1041 * (is used with charset names and base64 strings). Used php mbstring
1042 * functions - mb_strlen and mb_substr.
1043 *
1044 * Related documents: rfc 2045 (BASE64 encoding), rfc 2047 (mime header
1045 * encoding), rfc 2822 (header folding)
1046 *
1047 * @param string $string header string that must be encoded
1048 * @param string $charset character set. Must be supported by mbstring extension.
1049 * Use sq_mb_list_encodings() to detect supported charsets.
1050 * @return string string encoded according to rfc2047 B encoding formating rules
1051 * @since 1.5.1
1052 * @todo First header line can be wrapped to $iMaxLength - $HeaderFieldLength - 1
1053 * @todo Do we want to control max length of header?
1054 * @todo Do we want to control EOL (end-of-line) marker?
1055 * @todo Do we want to translate error message?
1056 */
1057 function encodeHeaderBase64($string,$charset) {
1058 /**
1059 * Check mbstring function requirements.
1060 */
1061 if (! function_exists('mb_strlen') ||
1062 ! function_exists('mb_substr')) {
1063 // set E_USER_WARNING
1064 trigger_error('encodeHeaderBase64: Required mbstring functions are missing.',E_USER_WARNING);
1065 // return false
1066 return false;
1067 }
1068
1069 // initial return array
1070 $aRet = array();
1071
1072 /**
1073 * header length = 75 symbols max (same as in encodeHeader)
1074 * remove $charset length
1075 * remove =? ? ?= (5 chars)
1076 * remove 2 more chars (\r\n ?)
1077 */
1078 $iMaxLength = 75 - strlen($charset) - 7;
1079
1080 // set first character position
1081 $iStartCharNum = 0;
1082
1083 // loop through all characters. count characters and not bytes.
1084 for ($iCharNum=1; $iCharNum<=mb_strlen($string,$charset); $iCharNum++) {
1085 // encode string from starting character to current character.
1086 $encoded_string = base64_encode(mb_substr($string,$iStartCharNum,$iCharNum-$iStartCharNum,$charset));
1087
1088 // Check encoded string length
1089 if(strlen($encoded_string)>$iMaxLength) {
1090 // if string exceeds max length, reduce number of encoded characters and add encoded string part to array
1091 $aRet[] = base64_encode(mb_substr($string,$iStartCharNum,$iCharNum-$iStartCharNum-1,$charset));
1092
1093 // set new starting character
1094 $iStartCharNum = $iCharNum-1;
1095
1096 // encode last char (in case it is last character in string)
1097 $encoded_string = base64_encode(mb_substr($string,$iStartCharNum,$iCharNum-$iStartCharNum,$charset));
1098 } // if string is shorter than max length - add next character
1099 }
1100
1101 // add last encoded string to array
1102 $aRet[] = $encoded_string;
1103
1104 // set initial return string
1105 $sRet = '';
1106
1107 // loop through encoded strings
1108 foreach($aRet as $string) {
1109 // TODO: Do we want to control EOL (end-of-line) marker
1110 if ($sRet!='') $sRet.= " ";
1111
1112 // add header tags and encoded string to return string
1113 $sRet.= '=?'.$charset.'?B?'.$string.'?=';
1114 }
1115
1116 return $sRet;
1117 }
1118
1119 /* This function trys to locate the entity_id of a specific mime element */
1120 function find_ent_id($id, $message) {
1121 for ($i = 0, $ret = ''; $ret == '' && $i < count($message->entities); $i++) {
1122 if ($message->entities[$i]->header->type0 == 'multipart') {
1123 $ret = find_ent_id($id, $message->entities[$i]);
1124 } else {
1125 if (strcasecmp($message->entities[$i]->header->id, $id) == 0) {
1126 // if (sq_check_save_extension($message->entities[$i])) {
1127 return $message->entities[$i]->entity_id;
1128 // }
1129 } elseif (!empty($message->entities[$i]->header->parameters['name'])) {
1130 /**
1131 * This is part of a fix for Outlook Express 6.x generating
1132 * cid URLs without creating content-id headers
1133 * @@JA - 20050207
1134 */
1135 if (strcasecmp($message->entities[$i]->header->parameters['name'], $id) == 0) {
1136 return $message->entities[$i]->entity_id;
1137 }
1138 }
1139 }
1140 }
1141 return $ret;
1142 }
1143
1144 function sq_check_save_extension($message) {
1145 $filename = $message->getFilename();
1146 $ext = substr($filename, strrpos($filename,'.')+1);
1147 $save_extensions = array('jpg','jpeg','gif','png','bmp');
1148 return in_array($ext, $save_extensions);
1149 }
1150
1151
1152 /**
1153 ** HTMLFILTER ROUTINES
1154 */
1155
1156 /**
1157 * This function checks attribute values for entity-encoded values
1158 * and returns them translated into 8-bit strings so we can run
1159 * checks on them.
1160 *
1161 * @param $attvalue A string to run entity check against.
1162 * @return Nothing, modifies a reference value.
1163 */
1164 function sq_defang(&$attvalue){
1165 $me = 'sq_defang';
1166 /**
1167 * Skip this if there aren't ampersands or backslashes.
1168 */
1169 if (strpos($attvalue, '&') === false
1170 && strpos($attvalue, '\\') === false){
1171 return;
1172 }
1173 $m = false;
1174 // before deent, translate the dangerous unicode characters and ... to safe values
1175 // otherwise the regular expressions do not match.
1176
1177
1178
1179 do {
1180 $m = false;
1181 $m = $m || sq_deent($attvalue, '/\&#0*(\d+);*/s');
1182 $m = $m || sq_deent($attvalue, '/\&#x0*((\d|[a-f])+);*/si', true);
1183 $m = $m || sq_deent($attvalue, '/\\\\(\d+)/s', true);
1184 } while ($m == true);
1185 $attvalue = stripslashes($attvalue);
1186 }
1187
1188 /**
1189 * Kill any tabs, newlines, or carriage returns. Our friends the
1190 * makers of the browser with 95% market value decided that it'd
1191 * be funny to make "java[tab]script" be just as good as "javascript".
1192 *
1193 * @param attvalue The attribute value before extraneous spaces removed.
1194 * @return attvalue Nothing, modifies a reference value.
1195 */
1196 function sq_unspace(&$attvalue){
1197 $me = 'sq_unspace';
1198 if (strcspn($attvalue, "\t\r\n\0 ") != strlen($attvalue)){
1199 $attvalue = str_replace(Array("\t", "\r", "\n", "\0", " "),
1200 Array('', '', '', '', ''), $attvalue);
1201 }
1202 }
1203
1204 /**
1205 * Translate all dangerous Unicode or Shift_JIS characters which are accepted by
1206 * IE as regular characters.
1207 *
1208 * @param attvalue The attribute value before dangerous characters are translated.
1209 * @return attvalue Nothing, modifies a reference value.
1210 * @author Marc Groot Koerkamp.
1211 */
1212 function sq_fixIE_idiocy(&$attvalue) {
1213 // remove NUL
1214 $attvalue = str_replace("\0", "", $attvalue);
1215 // remove comments
1216 $attvalue = preg_replace("/(\/\*.*?\*\/)/","",$attvalue);
1217
1218 // IE has the evil habit of accepting every possible value for the attribute expression.
1219 // The table below contains characters which are parsed by IE if they are used in the "expression"
1220 // attribute value.
1221 $aDangerousCharsReplacementTable = array(
1222 array('&#x029F;', '&#0671;' ,/* L UNICODE IPA Extension */
1223 '&#x0280;', '&#0640;' ,/* R UNICODE IPA Extension */
1224 '&#x0274;', '&#0628;' ,/* N UNICODE IPA Extension */
1225 '&#xFF25;', '&#65317;' ,/* Unicode FULLWIDTH LATIN CAPITAL LETTER E */
1226 '&#xFF45;', '&#65349;' ,/* Unicode FULLWIDTH LATIN SMALL LETTER E */
1227 '&#xFF38;', '&#65336;',/* Unicode FULLWIDTH LATIN CAPITAL LETTER X */
1228 '&#xFF58;', '&#65368;',/* Unicode FULLWIDTH LATIN SMALL LETTER X */
1229 '&#xFF30;', '&#65328;',/* Unicode FULLWIDTH LATIN CAPITAL LETTER P */
1230 '&#xFF50;', '&#65360;',/* Unicode FULLWIDTH LATIN SMALL LETTER P */
1231 '&#xFF32;', '&#65330;',/* Unicode FULLWIDTH LATIN CAPITAL LETTER R */
1232 '&#xFF52;', '&#65362;',/* Unicode FULLWIDTH LATIN SMALL LETTER R */
1233 '&#xFF33;', '&#65331;',/* Unicode FULLWIDTH LATIN CAPITAL LETTER S */
1234 '&#xFF53;', '&#65363;',/* Unicode FULLWIDTH LATIN SMALL LETTER S */
1235 '&#xFF29;', '&#65321;',/* Unicode FULLWIDTH LATIN CAPITAL LETTER I */
1236 '&#xFF49;', '&#65353;',/* Unicode FULLWIDTH LATIN SMALL LETTER I */
1237 '&#xFF2F;', '&#65327;',/* Unicode FULLWIDTH LATIN CAPITAL LETTER O */
1238 '&#xFF4F;', '&#65359;',/* Unicode FULLWIDTH LATIN SMALL LETTER O */
1239 '&#xFF2E;', '&#65326;',/* Unicode FULLWIDTH LATIN CAPITAL LETTER N */
1240 '&#xFF4E;', '&#65358;',/* Unicode FULLWIDTH LATIN SMALL LETTER N */
1241 '&#xFF2C;', '&#65324;',/* Unicode FULLWIDTH LATIN CAPITAL LETTER L */
1242 '&#xFF4C;', '&#65356;',/* Unicode FULLWIDTH LATIN SMALL LETTER L */
1243 '&#xFF35;', '&#65333;',/* Unicode FULLWIDTH LATIN CAPITAL LETTER U */
1244 '&#xFF55;', '&#65365;',/* Unicode FULLWIDTH LATIN SMALL LETTER U */
1245 '&#x207F;', '&#8319;' ,/* Unicode SUPERSCRIPT LATIN SMALL LETTER N */
1246 "\xEF\xBC\xA5", /* Shift JIS FULLWIDTH LATIN CAPITAL LETTER E */ // in unicode this is some Chinese char range
1247 "\xEF\xBD\x85", /* Shift JIS FULLWIDTH LATIN SMALL LETTER E */
1248 "\xEF\xBC\xB8", /* Shift JIS FULLWIDTH LATIN CAPITAL LETTER X */
1249 "\xEF\xBD\x98", /* Shift JIS FULLWIDTH LATIN SMALL LETTER X */
1250 "\xEF\xBC\xB0", /* Shift JIS FULLWIDTH LATIN CAPITAL LETTER P */
1251 "\xEF\xBD\x90", /* Shift JIS FULLWIDTH LATIN SMALL LETTER P */
1252 "\xEF\xBC\xB2", /* Shift JIS FULLWIDTH LATIN CAPITAL LETTER R */
1253 "\xEF\xBD\x92", /* Shift JIS FULLWIDTH LATIN SMALL LETTER R */
1254 "\xEF\xBC\xB3", /* Shift JIS FULLWIDTH LATIN CAPITAL LETTER S */
1255 "\xEF\xBD\x93", /* Shift JIS FULLWIDTH LATIN SMALL LETTER S */
1256 "\xEF\xBC\xA9", /* Shift JIS FULLWIDTH LATIN CAPITAL LETTER I */
1257 "\xEF\xBD\x89", /* Shift JIS FULLWIDTH LATIN SMALL LETTER I */
1258 "\xEF\xBC\xAF", /* Shift JIS FULLWIDTH LATIN CAPITAL LETTER O */
1259 "\xEF\xBD\x8F", /* Shift JIS FULLWIDTH LATIN SMALL LETTER O */
1260 "\xEF\xBC\xAE", /* Shift JIS FULLWIDTH LATIN CAPITAL LETTER N */
1261 "\xEF\xBD\x8E", /* Shift JIS FULLWIDTH LATIN SMALL LETTER N */
1262 "\xEF\xBC\xAC", /* Shift JIS FULLWIDTH LATIN CAPITAL LETTER L */
1263 "\xEF\xBD\x8C", /* Shift JIS FULLWIDTH LATIN SMALL LETTER L */
1264 "\xEF\xBC\xB5", /* Shift JIS FULLWIDTH LATIN CAPITAL LETTER U */
1265 "\xEF\xBD\x95", /* Shift JIS FULLWIDTH LATIN SMALL LETTER U */
1266 "\xE2\x81\xBF", /* Shift JIS FULLWIDTH SUPERSCRIPT N */
1267 "\xCA\x9F", /* L UNICODE IPA Extension */
1268 "\xCA\x80", /* R UNICODE IPA Extension */
1269 "\xC9\xB4"), /* N UNICODE IPA Extension */
1270 array('l', 'l', 'r','r','n','n',
1271 'E','E','e','e','X','X','x','x','P','P','p','p','R','R','r','r','S','S','s','s','I','I',
1272 'i','i','O','O','o','o','N','N','n','n','L','L','l','l','U','U','u','u','n','n',
1273 'E','e','X','x','P','p','R','r','S','s','I','i','O','o','N','n','L','l','U','u','n','l','r','n'));
1274 $attvalue = str_replace($aDangerousCharsReplacementTable[0],$aDangerousCharsReplacementTable[1],$attvalue);
1275
1276 // Escapes are useful for special characters like "{}[]()'&. In other cases they are
1277 // used for XSS.
1278 $attvalue = preg_replace("/(\\\\)([a-zA-Z]{1})/",'$2',$attvalue);
1279 }
1280
1281 /**
1282 * This function returns the final tag out of the tag name, an array
1283 * of attributes, and the type of the tag. This function is called by
1284 * sq_sanitize internally.
1285 *
1286 * @param $tagname the name of the tag.
1287 * @param $attary the array of attributes and their values
1288 * @param $tagtype The type of the tag (see in comments).
1289 * @return a string with the final tag representation.
1290 */
1291 function sq_tagprint($tagname, $attary, $tagtype){
1292 $me = 'sq_tagprint';
1293
1294 if ($tagtype == 2){
1295 $fulltag = '</' . $tagname . '>';
1296 } else {
1297 $fulltag = '<' . $tagname;
1298 if (is_array($attary) && sizeof($attary)){
1299 $atts = Array();
1300 while (list($attname, $attvalue) = each($attary)){
1301 array_push($atts, "$attname=$attvalue");
1302 }
1303 $fulltag .= ' ' . join(" ", $atts);
1304 }
1305 if ($tagtype == 3){
1306 $fulltag .= ' /';
1307 }
1308 $fulltag .= '>';
1309 }
1310 return $fulltag;
1311 }
1312
1313 /**
1314 * A small helper function to use with array_walk. Modifies a by-ref
1315 * value and makes it lowercase.
1316 *
1317 * @param $val a value passed by-ref.
1318 * @return void since it modifies a by-ref value.
1319 */
1320 function sq_casenormalize(&$val){
1321 $val = strtolower($val);
1322 }
1323
1324 /**
1325 * This function skips any whitespace from the current position within
1326 * a string and to the next non-whitespace value.
1327 *
1328 * @param $body the string
1329 * @param $offset the offset within the string where we should start
1330 * looking for the next non-whitespace character.
1331 * @return the location within the $body where the next
1332 * non-whitespace char is located.
1333 */
1334 function sq_skipspace($body, $offset){
1335 $me = 'sq_skipspace';
1336 preg_match('/^(\s*)/s', substr($body, $offset), $matches);
1337 if (sizeof($matches{1})){
1338 $count = strlen($matches{1});
1339 $offset += $count;
1340 }
1341 return $offset;
1342 }
1343
1344 /**
1345 * This function looks for the next character within a string. It's
1346 * really just a glorified "strpos", except it catches if failures
1347 * nicely.
1348 *
1349 * @param $body The string to look for needle in.
1350 * @param $offset Start looking from this position.
1351 * @param $needle The character/string to look for.
1352 * @return location of the next occurance of the needle, or
1353 * strlen($body) if needle wasn't found.
1354 */
1355 function sq_findnxstr($body, $offset, $needle){
1356 $me = 'sq_findnxstr';
1357 $pos = strpos($body, $needle, $offset);
1358 if ($pos === FALSE){
1359 $pos = strlen($body);
1360 }
1361 return $pos;
1362 }
1363
1364 /**
1365 * This function takes a PCRE-style regexp and tries to match it
1366 * within the string.
1367 *
1368 * @param $body The string to look for needle in.
1369 * @param $offset Start looking from here.
1370 * @param $reg A PCRE-style regex to match.
1371 * @return Returns a false if no matches found, or an array
1372 * with the following members:
1373 * - integer with the location of the match within $body
1374 * - string with whatever content between offset and the match
1375 * - string with whatever it is we matched
1376 */
1377 function sq_findnxreg($body, $offset, $reg){
1378 $me = 'sq_findnxreg';
1379 $matches = Array();
1380 $retarr = Array();
1381 preg_match("%^(.*?)($reg)%si", substr($body, $offset), $matches);
1382 if (!isset($matches{0}) || !$matches{0}){
1383 $retarr = false;
1384 } else {
1385 $retarr{0} = $offset + strlen($matches{1});
1386 $retarr{1} = $matches{1};
1387 $retarr{2} = $matches{2};
1388 }
1389 return $retarr;
1390 }
1391
1392 /**
1393 * This function looks for the next tag.
1394 *
1395 * @param $body String where to look for the next tag.
1396 * @param $offset Start looking from here.
1397 * @return false if no more tags exist in the body, or
1398 * an array with the following members:
1399 * - string with the name of the tag
1400 * - array with attributes and their values
1401 * - integer with tag type (1, 2, or 3)
1402 * - integer where the tag starts (starting "<")
1403 * - integer where the tag ends (ending ">")
1404 * first three members will be false, if the tag is invalid.
1405 */
1406 function sq_getnxtag($body, $offset){
1407 $me = 'sq_getnxtag';
1408 if ($offset > strlen($body)){
1409 return false;
1410 }
1411 $lt = sq_findnxstr($body, $offset, "<");
1412 if ($lt == strlen($body)){
1413 return false;
1414 }
1415 /**
1416 * We are here:
1417 * blah blah <tag attribute="value">
1418 * \---------^
1419 */
1420 $pos = sq_skipspace($body, $lt+1);
1421 if ($pos >= strlen($body)){
1422 return Array(false, false, false, $lt, strlen($body));
1423 }
1424 /**
1425 * There are 3 kinds of tags:
1426 * 1. Opening tag, e.g.:
1427 * <a href="blah">
1428 * 2. Closing tag, e.g.:
1429 * </a>
1430 * 3. XHTML-style content-less tag, e.g.:
1431 * <img src="blah" />
1432 */
1433 $tagtype = false;
1434 switch (substr($body, $pos, 1)){
1435 case '/':
1436 $tagtype = 2;
1437 $pos++;
1438 break;
1439 case '!':
1440 /**
1441 * A comment or an SGML declaration.
1442 */
1443 if (substr($body, $pos+1, 2) == "--"){
1444 $gt = strpos($body, "-->", $pos);
1445 if ($gt === false){
1446 $gt = strlen($body);
1447 } else {
1448 $gt += 2;
1449 }
1450 return Array(false, false, false, $lt, $gt);
1451 } else {
1452 $gt = sq_findnxstr($body, $pos, ">");
1453 return Array(false, false, false, $lt, $gt);
1454 }
1455 break;
1456 default:
1457 /**
1458 * Assume tagtype 1 for now. If it's type 3, we'll switch values
1459 * later.
1460 */
1461 $tagtype = 1;
1462 break;
1463 }
1464
1465 $tag_start = $pos;
1466 $tagname = '';
1467 /**
1468 * Look for next [\W-_], which will indicate the end of the tag name.
1469 */
1470 $regary = sq_findnxreg($body, $pos, "[^\w\-_]");
1471 if ($regary == false){
1472 return Array(false, false, false, $lt, strlen($body));
1473 }
1474 list($pos, $tagname, $match) = $regary;
1475 $tagname = strtolower($tagname);
1476
1477 /**
1478 * $match can be either of these:
1479 * '>' indicating the end of the tag entirely.
1480 * '\s' indicating the end of the tag name.
1481 * '/' indicating that this is type-3 xhtml tag.
1482 *
1483 * Whatever else we find there indicates an invalid tag.
1484 */
1485 switch ($match){
1486 case '/':
1487 /**
1488 * This is an xhtml-style tag with a closing / at the
1489 * end, like so: <img src="blah" />. Check if it's followed
1490 * by the closing bracket. If not, then this tag is invalid
1491 */
1492 if (substr($body, $pos, 2) == "/>"){
1493 $pos++;
1494 $tagtype = 3;
1495 } else {
1496 $gt = sq_findnxstr($body, $pos, ">");
1497 $retary = Array(false, false, false, $lt, $gt);
1498 return $retary;
1499 }
1500 case '>':
1501 return Array($tagname, false, $tagtype, $lt, $pos);
1502 break;
1503 default:
1504 /**
1505 * Check if it's whitespace
1506 */
1507 if (!preg_match('/\s/', $match)){
1508 /**
1509 * This is an invalid tag! Look for the next closing ">".
1510 */
1511 $gt = sq_findnxstr($body, $lt, ">");
1512 return Array(false, false, false, $lt, $gt);
1513 }
1514 break;
1515 }
1516
1517 /**
1518 * At this point we're here:
1519 * <tagname attribute='blah'>
1520 * \-------^
1521 *
1522 * At this point we loop in order to find all attributes.
1523 */
1524 $attname = '';
1525 $atttype = false;
1526 $attary = Array();
1527
1528 while ($pos <= strlen($body)){
1529 $pos = sq_skipspace($body, $pos);
1530 if ($pos == strlen($body)){
1531 /**
1532 * Non-closed tag.
1533 */
1534 return Array(false, false, false, $lt, $pos);
1535 }
1536 /**
1537 * See if we arrived at a ">" or "/>", which means that we reached
1538 * the end of the tag.
1539 */
1540 $matches = Array();
1541 if (preg_match("%^(\s*)(>|/>)%s", substr($body, $pos), $matches)) {
1542 /**
1543 * Yep. So we did.
1544 */
1545 $pos += strlen($matches{1});
1546 if ($matches{2} == "/>"){
1547 $tagtype = 3;
1548 $pos++;
1549 }
1550 return Array($tagname, $attary, $tagtype, $lt, $pos);
1551 }
1552
1553 /**
1554 * There are several types of attributes, with optional
1555 * [:space:] between members.
1556 * Type 1:
1557 * attrname[:space:]=[:space:]'CDATA'
1558 * Type 2:
1559 * attrname[:space:]=[:space:]"CDATA"
1560 * Type 3:
1561 * attr[:space:]=[:space:]CDATA
1562 * Type 4:
1563 * attrname
1564 *
1565 * We leave types 1 and 2 the same, type 3 we check for
1566 * '"' and convert to "&quot" if needed, then wrap in
1567 * double quotes. Type 4 we convert into:
1568 * attrname="yes".
1569 */
1570 $regary = sq_findnxreg($body, $pos, "[^:\w\-_]");
1571 if ($regary == false){
1572 /**
1573 * Looks like body ended before the end of tag.
1574 */
1575 return Array(false, false, false, $lt, strlen($body));
1576 }
1577 list($pos, $attname, $match) = $regary;
1578 $attname = strtolower($attname);
1579 /**
1580 * We arrived at the end of attribute name. Several things possible
1581 * here:
1582 * '>' means the end of the tag and this is attribute type 4
1583 * '/' if followed by '>' means the same thing as above
1584 * '\s' means a lot of things -- look what it's followed by.
1585 * anything else means the attribute is invalid.
1586 */
1587 switch($match){
1588 case '/':
1589 /**
1590 * This is an xhtml-style tag with a closing / at the
1591 * end, like so: <img src="blah" />. Check if it's followed
1592 * by the closing bracket. If not, then this tag is invalid
1593 */
1594 if (substr($body, $pos, 2) == "/>"){
1595 $pos++;
1596 $tagtype = 3;
1597 } else {
1598 $gt = sq_findnxstr($body, $pos, ">");
1599 $retary = Array(false, false, false, $lt, $gt);
1600 return $retary;
1601 }
1602 case '>':
1603 $attary{$attname} = '"yes"';
1604 return Array($tagname, $attary, $tagtype, $lt, $pos);
1605 break;
1606 default:
1607 /**
1608 * Skip whitespace and see what we arrive at.
1609 */
1610 $pos = sq_skipspace($body, $pos);
1611 $char = substr($body, $pos, 1);
1612 /**
1613 * Two things are valid here:
1614 * '=' means this is attribute type 1 2 or 3.
1615 * \w means this was attribute type 4.
1616 * anything else we ignore and re-loop. End of tag and
1617 * invalid stuff will be caught by our checks at the beginning
1618 * of the loop.
1619 */
1620 if ($char == "="){
1621 $pos++;
1622 $pos = sq_skipspace($body, $pos);
1623 /**
1624 * Here are 3 possibilities:
1625 * "'" attribute type 1
1626 * '"' attribute type 2
1627 * everything else is the content of tag type 3
1628 */
1629 $quot = substr($body, $pos, 1);
1630 if ($quot == "'"){
1631 $regary = sq_findnxreg($body, $pos+1, "\'");
1632 if ($regary == false){
1633 return Array(false, false, false, $lt, strlen($body));
1634 }
1635 list($pos, $attval, $match) = $regary;
1636 $pos++;
1637 $attary{$attname} = "'" . $attval . "'";
1638 } else if ($quot == '"'){
1639 $regary = sq_findnxreg($body, $pos+1, '\"');
1640 if ($regary == false){
1641 return Array(false, false, false, $lt, strlen($body));
1642 }
1643 list($pos, $attval, $match) = $regary;
1644 $pos++;
1645 $attary{$attname} = '"' . $attval . '"';
1646 } else {
1647 /**
1648 * These are hateful. Look for \s, or >.
1649 */
1650 $regary = sq_findnxreg($body, $pos, "[\s>]");
1651 if ($regary == false){
1652 return Array(false, false, false, $lt, strlen($body));
1653 }
1654 list($pos, $attval, $match) = $regary;
1655 /**
1656 * If it's ">" it will be caught at the top.
1657 */
1658 $attval = preg_replace("/\"/s", "&quot;", $attval);
1659 $attary{$attname} = '"' . $attval . '"';
1660 }
1661 } else if (preg_match("|[\w/>]|", $char)) {
1662 /**
1663 * That was attribute type 4.
1664 */
1665 $attary{$attname} = '"yes"';
1666 } else {
1667 /**
1668 * An illegal character. Find next '>' and return.
1669 */
1670 $gt = sq_findnxstr($body, $pos, ">");
1671 return Array(false, false, false, $lt, $gt);
1672 }
1673 break;
1674 }
1675 }
1676 /**
1677 * The fact that we got here indicates that the tag end was never
1678 * found. Return invalid tag indication so it gets stripped.
1679 */
1680 return Array(false, false, false, $lt, strlen($body));
1681 }
1682
1683 /**
1684 * Translates entities into literal values so they can be checked.
1685 *
1686 * @param $attvalue the by-ref value to check.
1687 * @param $regex the regular expression to check against.
1688 * @param $hex whether the entites are hexadecimal.
1689 * @return True or False depending on whether there were matches.
1690 */
1691 function sq_deent(&$attvalue, $regex, $hex=false){
1692 $me = 'sq_deent';
1693 $ret_match = false;
1694 // remove comments
1695 //$attvalue = preg_replace("/(\/\*.*\*\/)/","",$attvalue);
1696 preg_match_all($regex, $attvalue, $matches);
1697 if (is_array($matches) && sizeof($matches[0]) > 0){
1698 $repl = Array();
1699 for ($i = 0; $i < sizeof($matches[0]); $i++){
1700 $numval = $matches[1][$i];
1701 if ($hex){
1702 $numval = hexdec($numval);
1703 }
1704 $repl{$matches[0][$i]} = chr($numval);
1705 }
1706 $attvalue = strtr($attvalue, $repl);
1707 return true;
1708 } else {
1709 return false;
1710 }
1711 }
1712
1713 /**
1714 * This function runs various checks against the attributes.
1715 *
1716 * @param $tagname String with the name of the tag.
1717 * @param $attary Array with all tag attributes.
1718 * @param $rm_attnames See description for sq_sanitize
1719 * @param $bad_attvals See description for sq_sanitize
1720 * @param $add_attr_to_tag See description for sq_sanitize
1721 * @param $message message object
1722 * @param $id message id
1723 * @return Array with modified attributes.
1724 */
1725 function sq_fixatts($tagname,
1726 $attary,
1727 $rm_attnames,
1728 $bad_attvals,
1729 $add_attr_to_tag,
1730 $message,
1731 $id,
1732 $mailbox
1733 ){
1734 $me = 'sq_fixatts';
1735 while (list($attname, $attvalue) = each($attary)){
1736 /**
1737 * See if this attribute should be removed.
1738 */
1739 foreach ($rm_attnames as $matchtag=>$matchattrs){
1740 if (preg_match($matchtag, $tagname)){
1741 foreach ($matchattrs as $matchattr){
1742 if (preg_match($matchattr, $attname)){
1743 unset($attary{$attname});
1744 continue;
1745 }
1746 }
1747 }
1748 }
1749 /**
1750 * Workaround for IE quirks
1751 */
1752 sq_fixIE_idiocy($attvalue);
1753
1754 /**
1755 * Remove any backslashes, entities, and extraneous whitespace.
1756 */
1757
1758 $oldattvalue = $attvalue;
1759 sq_defang($attvalue);
1760 if ($attname == 'style' && $attvalue !== $oldattvalue) {
1761 // entities are used in the attribute value. In 99% of the cases it's there as XSS
1762 // i.e.<div style="{ left:exp&#x0280;essio&#x0274;( alert('XSS') ) }">
1763 $attvalue = "idiocy";
1764 $attary{$attname} = $attvalue;
1765 }
1766 sq_unspace($attvalue);
1767
1768 /**
1769 * Now let's run checks on the attvalues.
1770 * I don't expect anyone to comprehend this. If you do,
1771 * get in touch with me so I can drive to where you live and
1772 * shake your hand personally. :)
1773 */
1774 foreach ($bad_attvals as $matchtag=>$matchattrs){
1775 if (preg_match($matchtag, $tagname)){
1776 foreach ($matchattrs as $matchattr=>$valary){
1777 if (preg_match($matchattr, $attname)){
1778 /**
1779 * There are two arrays in valary.
1780 * First is matches.
1781 * Second one is replacements
1782 */
1783 list($valmatch, $valrepl) = $valary;
1784 $newvalue =
1785 preg_replace($valmatch, $valrepl, $attvalue);
1786 if ($newvalue != $attvalue){
1787 $attary{$attname} = $newvalue;
1788 $attvalue = $newvalue;
1789 }
1790 }
1791 }
1792 }
1793 }
1794 if ($attname == 'style') {
1795 if (preg_match('/[\0-\37\200-\377]+/',$attvalue)) {
1796 // 8bit and control characters in style attribute values can be used for XSS, remove them
1797 $attary{$attname} = '"disallowed character"';
1798 }
1799 preg_match_all("/url\s*\((.+)\)/si",$attvalue,$aMatch);
1800 if (count($aMatch)) {
1801 foreach($aMatch[1] as $sMatch) {
1802 // url value
1803 $urlvalue = $sMatch;
1804 sq_fix_url($attname, $urlvalue, $message, $id, $mailbox,"'");
1805 $attary{$attname} = str_replace($sMatch,$urlvalue,$attvalue);
1806 }
1807 }
1808 }
1809 /**
1810 * Use white list based filtering on attributes which can contain url's
1811 */
1812 else if ($attname == 'href' || $attname == 'src' || $attname == 'background') {
1813 sq_fix_url($attname, $attvalue, $message, $id, $mailbox);
1814 $attary{$attname} = $attvalue;
1815 }
1816 }
1817 /**
1818 * See if we need to append any attributes to this tag.
1819 */
1820 foreach ($add_attr_to_tag as $matchtag=>$addattary){
1821 if (preg_match($matchtag, $tagname)){
1822 $attary = array_merge($attary, $addattary);
1823 }
1824 }
1825 return $attary;
1826 }
1827
1828 /**
1829 * This function filters url's
1830 *
1831 * @param $attvalue String with attribute value to filter
1832 * @param $message message object
1833 * @param $id message id
1834 * @param $mailbox mailbox
1835 * @param $sQuote quoting characters around url's
1836 */
1837 function sq_fix_url($attname, &$attvalue, $message, $id, $mailbox,$sQuote = '"') {
1838 $attvalue = trim($attvalue);
1839 if ($attvalue && ($attvalue[0] =='"'|| $attvalue[0] == "'")) {
1840 // remove the double quotes
1841 $sQuote = $attvalue[0];
1842 $attvalue = trim(substr($attvalue,1,-1));
1843 }
1844
1845 // If there's no "view_unsafe_images" variable in the URL, turn unsafe
1846 // images off by default.
1847 sqgetGlobalVar('view_unsafe_images', $view_unsafe_images, SQ_GET, FALSE);
1848
1849 $secremoveimg = '../images/' . _("sec_remove_eng.png");
1850
1851 /**
1852 * Replace empty src tags with the blank image. src is only used
1853 * for frames, images, and image inputs. Doing a replace should
1854 * not affect them working as should be, however it will stop
1855 * IE from being kicked off when src for img tags are not set
1856 */
1857 if ($attvalue == '') {
1858 $attvalue = '"' . SM_PATH . 'images/blank.png"';
1859 } else {
1860 // first, disallow 8 bit characters and control characters
1861 if (preg_match('/[\0-\37\200-\377]+/',$attvalue)) {
1862 switch ($attname) {
1863 case 'href':
1864 $attvalue = $sQuote . 'http://invalid-stuff-detected.example.com' . $sQuote;
1865 break;
1866 default:
1867 $attvalue = $sQuote . SM_PATH . 'images/blank.png'. $sQuote;
1868 break;
1869 }
1870 } else {
1871 $aUrl = parse_url($attvalue);
1872 if (isset($aUrl['scheme'])) {
1873 switch(strtolower($aUrl['scheme'])) {
1874 case 'mailto':
1875 case 'http':
1876 case 'https':
1877 case 'ftp':
1878 if ($attname != 'href') {
1879 if ($view_unsafe_images == false) {
1880 $attvalue = $sQuote . $secremoveimg . $sQuote;
1881 } else {
1882 if (isset($aUrl['path'])) {
1883 // validate image extension.
1884 $ext = strtolower(substr($aUrl['path'],strrpos($aUrl['path'],'.')));
1885 if (!in_array($ext,array('.jpeg','.jpg','xjpeg','.gif','.bmp','.jpe','.png','.xbm'))) {
1886 $attvalue = $sQuote . SM_PATH . 'images/blank.png'. $sQuote;
1887 }
1888 } else {
1889 $attvalue = $sQuote . SM_PATH . 'images/blank.png'. $sQuote;
1890 }
1891 }
1892 }
1893 break;
1894 case 'outbind':
1895 /**
1896 * "Hack" fix for Outlook using propriatary outbind:// protocol in img tags.
1897 * One day MS might actually make it match something useful, for now, falling
1898 * back to using cid2http, so we can grab the blank.png.
1899 */
1900 $attvalue = sq_cid2http($message, $id, $attvalue, $mailbox);
1901 break;
1902 case 'cid':
1903 /**
1904 * Turn cid: urls into http-friendly ones.
1905 */
1906 $attvalue = sq_cid2http($message, $id, $attvalue, $mailbox);
1907 break;
1908 default:
1909 $attvalue = $sQuote . SM_PATH . 'images/blank.png' . $sQuote;
1910 break;
1911 }
1912 } else {
1913 if (!(isset($aUrl['path']) && $aUrl['path'] == $secremoveimg)) {
1914 // parse_url did not lead to satisfying result
1915 $attvalue = $sQuote . SM_PATH . 'images/blank.png' . $sQuote;
1916 }
1917 }
1918 }
1919 }
1920 }
1921
1922 /**
1923 * This function edits the style definition to make them friendly and
1924 * usable in SquirrelMail.
1925 *
1926 * @param $message the message object
1927 * @param $id the message id
1928 * @param $content a string with whatever is between <style> and </style>
1929 * @param $mailbox the message mailbox
1930 * @return a string with edited content.
1931 */
1932 function sq_fixstyle($body, $pos, $message, $id, $mailbox){
1933 $me = 'sq_fixstyle';
1934 // workaround for </style> in between comments
1935 $iCurrentPos = $pos;
1936 $content = '';
1937 $sToken = '';
1938 $bSucces = false;
1939 $bEndTag = false;
1940 for ($i=$pos,$iCount=strlen($body);$i<$iCount;++$i) {
1941 $char = $body{$i};
1942 switch ($char) {
1943 case '<':
1944 $sToken = $char;
1945 break;
1946 case '/':
1947 if ($sToken == '<') {
1948 $sToken .= $char;
1949 $bEndTag = true;
1950 } else {
1951 $content .= $char;
1952 }
1953 break;
1954 case '>':
1955 if ($bEndTag) {
1956 $sToken .= $char;
1957 if (preg_match('/\<\/\s*style\s*\>/i',$sToken,$aMatch)) {
1958 $newpos = $i + 1;
1959 $bSucces = true;
1960 break 2;
1961 } else {
1962 $content .= $sToken;
1963 }
1964 $bEndTag = false;
1965 } else {
1966 $content .= $char;
1967 }
1968 break;
1969 case '!':
1970 if ($sToken == '<') {
1971 // possible comment
1972 if (isset($body{$i+2}) && substr($body,$i,3) == '!--') {
1973 $i = strpos($body,'-->',$i+3);
1974 if ($i === false) { // no end comment
1975 $i = strlen($body);
1976 }
1977 $sToken = '';
1978 }
1979 } else {
1980 $content .= $char;
1981 }
1982 break;
1983 default:
1984 if ($bEndTag) {
1985 $sToken .= $char;
1986 } else {
1987 $content .= $char;
1988 }
1989 break;
1990 }
1991 }
1992 if ($bSucces == FALSE){
1993 return array(FALSE, strlen($body));
1994 }
1995
1996
1997
1998 /**
1999 * First look for general BODY style declaration, which would be
2000 * like so:
2001 * body {background: blah-blah}
2002 * and change it to .bodyclass so we can just assign it to a <div>
2003 */
2004 $content = preg_replace("|body(\s*\{.*?\})|si", ".bodyclass\\1", $content);
2005 $secremoveimg = '../images/' . _("sec_remove_eng.png");
2006 /**
2007 * Fix url('blah') declarations.
2008 */
2009 // $content = preg_replace("|url\s*\(\s*([\'\"])\s*\S+script\s*:.*?([\'\"])\s*\)|si",
2010 // "url(\\1$secremoveimg\\2)", $content);
2011
2012 // first check for 8bit sequences and disallowed control characters
2013 if (preg_match('/[\16-\37\200-\377]+/',$content)) {
2014 $content = '<!-- style block removed by html filter due to presence of 8bit characters -->';
2015 return array($content, $newpos);
2016 }
2017
2018 // IE Sucks hard. We have a special function for it.
2019 sq_fixIE_idiocy($content);
2020
2021 // remove @import line
2022 $content = preg_replace("/^\s*(@import.*)$/mi","\n<!-- @import rules forbidden -->\n",$content);
2023
2024 // translate ur\l and variations (IE parses that)
2025 // TODO check if the sq_fixIE_idiocy function already handles this.
2026 $content = preg_replace("/(\\\\)?u(\\\\)?r(\\\\)?l(\\\\)?/i", 'url', $content);
2027 preg_match_all("/url\s*\((.+)\)/si",$content,$aMatch);
2028 if (count($aMatch)) {
2029 $aValue = $aReplace = array();
2030 foreach($aMatch[1] as $sMatch) {
2031 // url value
2032 $urlvalue = $sMatch;
2033 sq_fix_url('style',$urlvalue, $message, $id, $mailbox,"'");
2034 $aValue[] = $sMatch;
2035 $aReplace[] = $urlvalue;
2036 }
2037 $content = str_replace($aValue,$aReplace,$content);
2038 }
2039
2040 /**
2041 * Remove any backslashes, entities, and extraneous whitespace.
2042 */
2043 $contentTemp = $content;
2044 sq_defang($contentTemp);
2045 sq_unspace($contentTemp);
2046
2047 /**
2048 * Fix stupid css declarations which lead to vulnerabilities
2049 * in IE.
2050 */
2051 $match = Array('/\/\*.*\*\//',
2052 '/expression/i',
2053 '/behaviou*r/i',
2054 '/binding/i',
2055 '/include-source/i',
2056 '/javascript/i',
2057 '/script/i');
2058 $replace = Array('','idiocy', 'idiocy', 'idiocy', 'idiocy', 'idiocy', 'idiocy');
2059 $contentNew = preg_replace($match, $replace, $contentTemp);
2060 if ($contentNew !== $contentTemp) {
2061 // insecure css declarations are used. From now on we don't care
2062 // anymore if the css is destroyed by sq_deent, sq_unspace or sq_unbackslash
2063 $content = $contentNew;
2064 }
2065 return array($content, $newpos);
2066 }
2067
2068
2069 /**
2070 * This function converts cid: url's into the ones that can be viewed in
2071 * the browser.
2072 *
2073 * @param $message the message object
2074 * @param $id the message id
2075 * @param $cidurl the cid: url.
2076 * @param $mailbox the message mailbox
2077 * @return a string with a http-friendly url
2078 */
2079 function sq_cid2http($message, $id, $cidurl, $mailbox){
2080 /**
2081 * Get rid of quotes.
2082 */
2083 $quotchar = substr($cidurl, 0, 1);
2084 if ($quotchar == '"' || $quotchar == "'"){
2085 $cidurl = str_replace($quotchar, "", $cidurl);
2086 } else {
2087 $quotchar = '';
2088 }
2089 $cidurl = substr(trim($cidurl), 4);
2090
2091 $match_str = '/\{.*?\}\//';
2092 $str_rep = '';
2093 $cidurl = preg_replace($match_str, $str_rep, $cidurl);
2094
2095 $linkurl = find_ent_id($cidurl, $message);
2096 /* in case of non-safe cid links $httpurl should be replaced by a sort of
2097 unsafe link image */
2098 $httpurl = '';
2099
2100 /**
2101 * This is part of a fix for Outlook Express 6.x generating
2102 * cid URLs without creating content-id headers. These images are
2103 * not part of the multipart/related html mail. The html contains
2104 * <img src="cid:{some_id}/image_filename.ext"> references to
2105 * attached images with as goal to render them inline although
2106 * the attachment disposition property is not inline.
2107 */
2108
2109 if (empty($linkurl)) {
2110 if (preg_match('/{.*}\//', $cidurl)) {
2111 $cidurl = preg_replace('/{.*}\//','', $cidurl);
2112 if (!empty($cidurl)) {
2113 $linkurl = find_ent_id($cidurl, $message);
2114 }
2115 }
2116 }
2117
2118 if (!empty($linkurl)) {
2119 $httpurl = $quotchar . SM_PATH . 'src/download.php?absolute_dl=true&amp;' .
2120 "passed_id=$id&amp;mailbox=" . urlencode($mailbox) .
2121 '&amp;ent_id=' . $linkurl . $quotchar;
2122 } else {
2123 /**
2124 * If we couldn't generate a proper img url, drop in a blank image
2125 * instead of sending back empty, otherwise it causes unusual behaviour
2126 */
2127 $httpurl = $quotchar . SM_PATH . 'images/blank.png' . $quotchar;
2128 }
2129
2130 return $httpurl;
2131 }
2132
2133 /**
2134 * This function changes the <body> tag into a <div> tag since we
2135 * can't really have a body-within-body.
2136 *
2137 * @param $attary an array of attributes and values of <body>
2138 * @param $mailbox mailbox we're currently reading (for cid2http)
2139 * @param $message current message (for cid2http)
2140 * @param $id current message id (for cid2http)
2141 * @return a modified array of attributes to be set for <div>
2142 */
2143 function sq_body2div($attary, $mailbox, $message, $id){
2144 $me = 'sq_body2div';
2145 $divattary = Array('class' => "'bodyclass'");
2146 $text = '#000000';
2147 $has_bgc_stl = $has_txt_stl = false;
2148 $styledef = '';
2149 if (is_array($attary) && sizeof($attary) > 0){
2150 foreach ($attary as $attname=>$attvalue){
2151 $quotchar = substr($attvalue, 0, 1);
2152 $attvalue = str_replace($quotchar, "", $attvalue);
2153 switch ($attname){
2154 case 'background':
2155 $attvalue = sq_cid2http($message, $id, $attvalue, $mailbox);
2156 $styledef .= "background-image: url('$attvalue'); ";
2157 break;
2158 case 'bgcolor':
2159 $has_bgc_stl = true;
2160 $styledef .= "background-color: $attvalue; ";
2161 break;
2162 case 'text':
2163 $has_txt_stl = true;
2164 $styledef .= "color: $attvalue; ";
2165 break;
2166 }
2167 }
2168 // Outlook defines a white bgcolor and no text color. This can lead to
2169 // white text on a white bg with certain themes.
2170 if ($has_bgc_stl && !$has_txt_stl) {
2171 $styledef .= "color: $text; ";
2172 }
2173 if (strlen($styledef) > 0){
2174 $divattary{"style"} = "\"$styledef\"";
2175 }
2176 }
2177 return $divattary;
2178 }
2179
2180 /**
2181 * This is the main function and the one you should actually be calling.
2182 * There are several variables you should be aware of an which need
2183 * special description.
2184 *
2185 * Since the description is quite lengthy, see it here:
2186 * http://linux.duke.edu/projects/mini/htmlfilter/
2187 *
2188 * @param $body the string with HTML you wish to filter
2189 * @param $tag_list see description above
2190 * @param $rm_tags_with_content see description above
2191 * @param $self_closing_tags see description above
2192 * @param $force_tag_closing see description above
2193 * @param $rm_attnames see description above
2194 * @param $bad_attvals see description above
2195 * @param $add_attr_to_tag see description above
2196 * @param $message message object
2197 * @param $id message id
2198 * @return sanitized html safe to show on your pages.
2199 */
2200 function sq_sanitize($body,
2201 $tag_list,
2202 $rm_tags_with_content,
2203 $self_closing_tags,
2204 $force_tag_closing,
2205 $rm_attnames,
2206 $bad_attvals,
2207 $add_attr_to_tag,
2208 $message,
2209 $id,
2210 $mailbox
2211 ){
2212 $me = 'sq_sanitize';
2213 $rm_tags = array_shift($tag_list);
2214 /**
2215 * Normalize rm_tags and rm_tags_with_content.
2216 */
2217 @array_walk($tag_list, 'sq_casenormalize');
2218 @array_walk($rm_tags_with_content, 'sq_casenormalize');
2219 @array_walk($self_closing_tags, 'sq_casenormalize');
2220 /**
2221 * See if tag_list is of tags to remove or tags to allow.
2222 * false means remove these tags
2223 * true means allow these tags
2224 */
2225 $curpos = 0;
2226 $open_tags = Array();
2227 $trusted = "\n<!-- begin sanitized html -->\n";
2228 $skip_content = false;
2229 /**
2230 * Take care of netscape's stupid javascript entities like
2231 * &{alert('boo')};
2232 */
2233 $body = preg_replace("/&(\{.*?\};)/si", "&amp;\\1", $body);
2234
2235 while (($curtag = sq_getnxtag($body, $curpos)) != FALSE){
2236 list($tagname, $attary, $tagtype, $lt, $gt) = $curtag;
2237 $free_content = substr($body, $curpos, $lt-$curpos);
2238 /**
2239 * Take care of <style>
2240 */
2241 if ($tagname == "style" && $tagtype == 1){
2242 list($free_content, $curpos) =
2243 sq_fixstyle($body, $gt+1, $message, $id, $mailbox);
2244 if ($free_content != FALSE){
2245 $trusted .= sq_tagprint($tagname, $attary, $tagtype);
2246 $trusted .= $free_content;
2247 $trusted .= sq_tagprint($tagname, false, 2);
2248 }
2249 continue;
2250 }
2251 if ($skip_content == false){
2252 $trusted .= $free_content;
2253 }
2254 if ($tagname != FALSE){
2255 if ($tagtype == 2){
2256 if ($skip_content == $tagname){
2257 /**
2258 * Got to the end of tag we needed to remove.
2259 */
2260 $tagname = false;
2261 $skip_content = false;
2262 } else {
2263 if ($skip_content == false){
2264 if ($tagname == "body"){
2265 $tagname = "div";
2266 }
2267 if (isset($open_tags{$tagname}) &&
2268 $open_tags{$tagname} > 0){
2269 $open_tags{$tagname}--;
2270 } else {
2271 $tagname = false;
2272 }
2273 }
2274 }
2275 } else {
2276 /**
2277 * $rm_tags_with_content
2278 */
2279 if ($skip_content == false){
2280 /**
2281 * See if this is a self-closing type and change
2282 * tagtype appropriately.
2283 */
2284 if ($tagtype == 1
2285 && in_array($tagname, $self_closing_tags)){
2286 $tagtype = 3;
2287 }
2288 /**
2289 * See if we should skip this tag and any content
2290 * inside it.
2291 */
2292 if ($tagtype == 1 &&
2293 in_array($tagname, $rm_tags_with_content)){
2294 $skip_content = $tagname;
2295 } else {
2296 if (($rm_tags == false
2297 && in_array($tagname, $tag_list)) ||
2298 ($rm_tags == true &&
2299 !in_array($tagname, $tag_list))){
2300 $tagname = false;
2301 } else {
2302 /**
2303 * Convert body into div.
2304 */
2305 if ($tagname == "body"){
2306 $tagname = "div";
2307 $attary = sq_body2div($attary, $mailbox,
2308 $message, $id);
2309 }
2310 if ($tagtype == 1){
2311 if (isset($open_tags{$tagname})){
2312 $open_tags{$tagname}++;
2313 } else {
2314 $open_tags{$tagname}=1;
2315 }
2316 }
2317 /**
2318 * This is where we run other checks.
2319 */
2320 if (is_array($attary) && sizeof($attary) > 0){
2321 $attary = sq_fixatts($tagname,
2322 $attary,
2323 $rm_attnames,
2324 $bad_attvals,
2325 $add_attr_to_tag,
2326 $message,
2327 $id,
2328 $mailbox
2329 );
2330 }
2331 }
2332 }
2333 }
2334 }
2335 if ($tagname != false && $skip_content == false){
2336 $trusted .= sq_tagprint($tagname, $attary, $tagtype);
2337 }
2338 }
2339 $curpos = $gt+1;
2340 }
2341 $trusted .= substr($body, $curpos, strlen($body)-$curpos);
2342 if ($force_tag_closing == true){
2343 foreach ($open_tags as $tagname=>$opentimes){
2344 while ($opentimes > 0){
2345 $trusted .= '</' . $tagname . '>';
2346 $opentimes--;
2347 }
2348 }
2349 $trusted .= "\n";
2350 }
2351 $trusted .= "<!-- end sanitized html -->\n";
2352 return $trusted;
2353 }
2354
2355 /**
2356 * This is a wrapper function to call html sanitizing routines.
2357 *
2358 * @param $body the body of the message
2359 * @param $id the id of the message
2360
2361 * @param $message
2362 * @param $mailbox
2363 * @param boolean $take_mailto_links When TRUE, converts mailto: links
2364 * into internal SM compose links
2365 * (optional; default = TRUE)
2366 * @return a string with html safe to display in the browser.
2367 */
2368 function magicHTML($body, $id, $message, $mailbox = 'INBOX', $take_mailto_links =true) {
2369
2370 // require_once(SM_PATH . 'functions/url_parser.php'); // for $MailTo_PReg_Match
2371
2372 global $attachment_common_show_images, $view_unsafe_images,
2373 $has_unsafe_images;
2374 /**
2375 * Don't display attached images in HTML mode.
2376 *
2377 * SB: why?
2378 */
2379 $attachment_common_show_images = false;
2380 $tag_list = Array(
2381 false,
2382 "object",
2383 "meta",
2384 "html",
2385 "head",
2386 "base",
2387 "link",
2388 "frame",
2389 "iframe",
2390 "plaintext",
2391 "marquee"
2392 );
2393
2394 $rm_tags_with_content = Array(
2395 "script",
2396 "applet",
2397 "embed",
2398 "title",
2399 "frameset",
2400 "xmp",
2401 "xml"
2402 );
2403
2404 $self_closing_tags = Array(
2405 "img",
2406 "br",
2407 "hr",
2408 "input",
2409 "outbind"
2410 );
2411
2412 $force_tag_closing = true;
2413
2414 $rm_attnames = Array(
2415 "/.*/" =>
2416 Array(
2417 "/target/i",
2418 "/^on.*/i",
2419 "/^dynsrc/i",
2420 "/^data.*/i",
2421 "/^lowsrc.*/i"
2422 )
2423 );
2424
2425 $secremoveimg = "../images/" . _("sec_remove_eng.png");
2426 $bad_attvals = Array(
2427 "/.*/" =>
2428 Array(
2429 "/^src|background/i" =>
2430 Array(
2431 Array(
2432 "/^([\'\"])\s*\S+script\s*:.*([\'\"])/si",
2433 "/^([\'\"])\s*mocha\s*:*.*([\'\"])/si",
2434 "/^([\'\"])\s*about\s*:.*([\'\"])/si"
2435 ),
2436 Array(
2437 "\\1$secremoveimg\\2",
2438 "\\1$secremoveimg\\2",
2439 "\\1$secremoveimg\\2",
2440 )
2441 ),
2442 "/^href|action/i" =>
2443 Array(
2444 Array(
2445 "/^([\'\"])\s*\S+script\s*:.*([\'\"])/si",
2446 "/^([\'\"])\s*mocha\s*:*.*([\'\"])/si",
2447 "/^([\'\"])\s*about\s*:.*([\'\"])/si"
2448 ),
2449 Array(
2450 "\\1#\\1",
2451 "\\1#\\1",
2452 "\\1#\\1"
2453 )
2454 ),
2455 "/^style/i" =>
2456 Array(
2457 Array(
2458 "/\/\*.*\*\//",
2459 "/expression/i",
2460 "/binding/i",
2461 "/behaviou*r/i",
2462 "/include-source/i",
2463 "/position\s*:\s*absolute/i",
2464 "/(\\\\)?u(\\\\)?r(\\\\)?l(\\\\)?/i",
2465 "/url\s*\(\s*([\'\"])\s*\S+script\s*:.*([\'\"])\s*\)/si",
2466 "/url\s*\(\s*([\'\"])\s*mocha\s*:.*([\'\"])\s*\)/si",
2467 "/url\s*\(\s*([\'\"])\s*about\s*:.*([\'\"])\s*\)/si",
2468 "/(.*)\s*:\s*url\s*\(\s*([\'\"]*)\s*\S+script\s*:.*([\'\"]*)\s*\)/si"
2469 ),
2470 Array(
2471 "",
2472 "idiocy",
2473 "idiocy",
2474 "idiocy",
2475 "idiocy",
2476 "idiocy",
2477 "url",
2478 "url(\\1#\\1)",
2479 "url(\\1#\\1)",
2480 "url(\\1#\\1)",
2481 "\\1:url(\\2#\\3)"
2482 )
2483 )
2484 )
2485 );
2486
2487 // If there's no "view_unsafe_images" variable in the URL, turn unsafe
2488 // images off by default.
2489 sqgetGlobalVar('view_unsafe_images', $view_unsafe_images, SQ_GET, FALSE);
2490
2491 if (!$view_unsafe_images){
2492 /**
2493 * Remove any references to http/https if view_unsafe_images set
2494 * to false.
2495 */
2496 array_push($bad_attvals{'/.*/'}{'/^src|background/i'}[0],
2497 '/^([\'\"])\s*https*:.*([\'\"])/si');
2498 array_push($bad_attvals{'/.*/'}{'/^src|background/i'}[1],
2499 "\\1$secremoveimg\\1");
2500 array_push($bad_attvals{'/.*/'}{'/^style/i'}[0],
2501 '/url\([\'\"]?https?:[^\)]*[\'\"]?\)/si');
2502 array_push($bad_attvals{'/.*/'}{'/^style/i'}[1],
2503 "url(\\1$secremoveimg\\1)");
2504 }
2505
2506 $add_attr_to_tag = Array(
2507 "/^a$/i" =>
2508 Array('target'=>'"_blank"',
2509 'title'=>'"'._("This external link will open in a new window").'"'
2510 )
2511 );
2512 $trusted = sq_sanitize($body,
2513 $tag_list,
2514 $rm_tags_with_content,
2515 $self_closing_tags,
2516 $force_tag_closing,
2517 $rm_attnames,
2518 $bad_attvals,
2519 $add_attr_to_tag,
2520 $message,
2521 $id,
2522 $mailbox
2523 );
2524 if (strpos($trusted,$secremoveimg)){
2525 $has_unsafe_images = true;
2526 }
2527
2528 // we want to parse mailto's in HTML output, change to SM compose links
2529 // this is a modified version of code from url_parser.php... but Marc is
2530 // right: we need a better filtering implementation; adding this randomly
2531 // here is not a great solution
2532 //
2533 if ($take_mailto_links) {
2534 // parseUrl($trusted); // this even parses URLs inside of tags... too aggressive
2535 global $MailTo_PReg_Match;
2536 $MailTo_PReg_Match = '/mailto:' . substr($MailTo_PReg_Match, 1) ;
2537 if ((preg_match_all($MailTo_PReg_Match, $trusted, $regs)) && ($regs[0][0] != '')) {
2538 foreach ($regs[0] as $i => $mailto_before) {
2539 $mailto_params = $regs[10][$i];
2540 // get rid of any tailing quote since we have to add send_to to the end
2541 //
2542 if (substr($mailto_before, strlen($mailto_before) - 1) == '"')
2543 $mailto_before = substr($mailto_before, 0, strlen($mailto_before) - 1);
2544 if (substr($mailto_params, strlen($mailto_params) - 1) == '"')
2545 $mailto_params = substr($mailto_params, 0, strlen($mailto_params) - 1);
2546
2547 if ($regs[1][$i]) { //if there is an email addr before '?', we need to merge it with the params
2548 $to = 'to=' . $regs[1][$i];
2549 if (strpos($mailto_params, 'to=') > -1) //already a 'to='
2550 $mailto_params = str_replace('to=', $to . '%2C%20', $mailto_params);
2551 else {
2552 if ($mailto_params) //already some params, append to them
2553 $mailto_params .= '&amp;' . $to;
2554 else
2555 $mailto_params .= '?' . $to;
2556 }
2557 }
2558
2559 $url_str = preg_replace(array('/to=/i', '/(?<!b)cc=/i', '/bcc=/i'), array('send_to=', 'send_to_cc=', 'send_to_bcc='), $mailto_params);
2560
2561 // we'll already have target=_blank, no need to allow comp_in_new
2562 // here (which would be a lot more work anyway)
2563 //
2564 global $compose_new_win;
2565 $temp_comp_in_new = $compose_new_win;
2566 $compose_new_win = 0;
2567 $comp_uri = makeComposeLink('src/compose.php' . $url_str, $mailto_before);
2568 $compose_new_win = $temp_comp_in_new;
2569
2570 // remove <a href=" and anything after the next quote (we only
2571 // need the uri, not the link HTML) in compose uri
2572 //
2573 $comp_uri = substr($comp_uri, 9);
2574 $comp_uri = substr($comp_uri, 0, strpos($comp_uri, '"', 1));
2575 $trusted = str_replace($mailto_before, $comp_uri, $trusted);
2576 }
2577 }
2578 }
2579
2580 return $trusted;
2581 }
2582
2583 /**
2584 * function SendDownloadHeaders - send file to the browser
2585 *
2586 * Original Source: SM core src/download.php
2587 * moved here to make it available to other code, and separate
2588 * front end from back end functionality.
2589 *
2590 * @param string $type0 first half of mime type
2591 * @param string $type1 second half of mime type
2592 * @param string $filename filename to tell the browser for downloaded file
2593 * @param boolean $force whether to force the download dialog to pop
2594 * @param optional integer $filesize send the Content-Header and length to the browser
2595 * @return void
2596 */
2597 function SendDownloadHeaders($type0, $type1, $filename, $force, $filesize=0) {
2598 global $languages, $squirrelmail_language;
2599 $isIE = $isIE6plus = false;
2600
2601 sqgetGlobalVar('HTTP_USER_AGENT', $HTTP_USER_AGENT, SQ_SERVER);
2602
2603 if (strstr($HTTP_USER_AGENT, 'compatible; MSIE ') !== false &&
2604 strstr($HTTP_USER_AGENT, 'Opera') === false) {
2605 $isIE = true;
2606 }
2607
2608 if (preg_match('/compatible; MSIE ([0-9]+)/', $HTTP_USER_AGENT, $match) &&
2609 ((int)$match[1]) >= 6 && strstr($HTTP_USER_AGENT, 'Opera') === false) {
2610 $isIE6plus = true;
2611 }
2612
2613 if (isset($languages[$squirrelmail_language]['XTRA_CODE']) &&
2614 function_exists($languages[$squirrelmail_language]['XTRA_CODE'] . '_downloadfilename')) {
2615 $filename =
2616 call_user_func($languages[$squirrelmail_language]['XTRA_CODE'] . '_downloadfilename', $filename, $HTTP_USER_AGENT);
2617 } else {
2618 $filename = ereg_replace('[\\/:\*\?"<>\|;]', '_', str_replace('&nbsp;', ' ', $filename));
2619 }
2620
2621 // A Pox on Microsoft and it's Internet Explorer!
2622 //
2623 // IE has lots of bugs with file downloads.
2624 // It also has problems with SSL. Both of these cause problems
2625 // for us in this function.
2626 //
2627 // See this article on Cache Control headers and SSL
2628 // http://support.microsoft.com/default.aspx?scid=kb;en-us;323308
2629 //
2630 // The best thing you can do for IE is to upgrade to the latest
2631 // version
2632 //set all the Cache Control Headers for IE
2633 if ($isIE) {
2634 $filename=rawurlencode($filename);
2635 header ("Pragma: public");
2636 header ("Cache-Control: no-store, max-age=0, no-cache, must-revalidate"); // HTTP/1.1
2637 header ("Cache-Control: post-check=0, pre-check=0", false);
2638 header ("Cache-Control: private");
2639
2640 //set the inline header for IE, we'll add the attachment header later if we need it
2641 header ("Content-Disposition: inline; filename=$filename");
2642 }
2643
2644 if (!$force) {
2645 // Try to show in browser window
2646 header ("Content-Disposition: inline; filename=\"$filename\"");
2647 header ("Content-Type: $type0/$type1; name=\"$filename\"");
2648 } else {
2649 // Try to pop up the "save as" box
2650
2651 // IE makes this hard. It pops up 2 save boxes, or none.
2652 // http://support.microsoft.com/support/kb/articles/Q238/5/88.ASP
2653 // http://support.microsoft.com/default.aspx?scid=kb;EN-US;260519
2654 // But, according to Microsoft, it is "RFC compliant but doesn't
2655 // take into account some deviations that allowed within the
2656 // specification." Doesn't that mean RFC non-compliant?
2657 // http://support.microsoft.com/support/kb/articles/Q258/4/52.ASP
2658
2659 // all browsers need the application/octet-stream header for this
2660 header ("Content-Type: application/octet-stream; name=\"$filename\"");
2661
2662 // http://support.microsoft.com/support/kb/articles/Q182/3/15.asp
2663 // Do not have quotes around filename, but that applied to
2664 // "attachment"... does it apply to inline too?
2665 header ("Content-Disposition: attachment; filename=\"$filename\"");
2666
2667 if ($isIE && !$isIE6plus) {
2668 // This combination seems to work mostly. IE 5.5 SP 1 has
2669 // known issues (see the Microsoft Knowledge Base)
2670
2671 // This works for most types, but doesn't work with Word files
2672 header ("Content-Type: application/download; name=\"$filename\"");
2673 header ("Content-Type: application/force-download; name=\"$filename\"");
2674 // These are spares, just in case. :-)
2675 //header("Content-Type: $type0/$type1; name=\"$filename\"");
2676 //header("Content-Type: application/x-msdownload; name=\"$filename\"");
2677 //header("Content-Type: application/octet-stream; name=\"$filename\"");
2678 } else if ($isIE) {
2679 // This is to prevent IE for MIME sniffing and auto open a file in IE
2680 header ("Content-Type: application/force-download; name=\"$filename\"");
2681 } else {
2682 // another application/octet-stream forces download for Netscape
2683 header ("Content-Type: application/octet-stream; name=\"$filename\"");
2684 }
2685 }
2686
2687 //send the content-length header if the calling function provides it
2688 if ($filesize > 0) {
2689 header("Content-Length: $filesize");
2690 }
2691
2692 } // end fn SendDownloadHeaders