Merge "Add new parserTests for table cell attributes coming from templates."
[lhc/web/wiklou.git] / includes / WebRequest.php
1 <?php
2 /**
3 * Deal with importing all those nasty globals and things
4 *
5 * Copyright © 2003 Brion Vibber <brion@pobox.com>
6 * http://www.mediawiki.org/
7 *
8 * This program is free software; you can redistribute it and/or modify
9 * it under the terms of the GNU General Public License as published by
10 * the Free Software Foundation; either version 2 of the License, or
11 * (at your option) any later version.
12 *
13 * This program is distributed in the hope that it will be useful,
14 * but WITHOUT ANY WARRANTY; without even the implied warranty of
15 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
16 * GNU General Public License for more details.
17 *
18 * You should have received a copy of the GNU General Public License along
19 * with this program; if not, write to the Free Software Foundation, Inc.,
20 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
21 * http://www.gnu.org/copyleft/gpl.html
22 *
23 * @file
24 */
25
26 /**
27 * The WebRequest class encapsulates getting at data passed in the
28 * URL or via a POSTed form, handling remove of "magic quotes" slashes,
29 * stripping illegal input characters and normalizing Unicode sequences.
30 *
31 * Usually this is used via a global singleton, $wgRequest. You should
32 * not create a second WebRequest object; make a FauxRequest object if
33 * you want to pass arbitrary data to some function in place of the web
34 * input.
35 *
36 * @ingroup HTTP
37 */
38 class WebRequest {
39 protected $data, $headers = array();
40
41 /**
42 * Lazy-init response object
43 * @var WebResponse
44 */
45 private $response;
46
47 /**
48 * Cached client IP address
49 * @var String
50 */
51 private $ip;
52
53 public function __construct() {
54 /// @todo FIXME: This preemptive de-quoting can interfere with other web libraries
55 /// and increases our memory footprint. It would be cleaner to do on
56 /// demand; but currently we have no wrapper for $_SERVER etc.
57 $this->checkMagicQuotes();
58
59 // POST overrides GET data
60 // We don't use $_REQUEST here to avoid interference from cookies...
61 $this->data = $_POST + $_GET;
62 }
63
64 /**
65 * Extract relevant query arguments from the http request uri's path
66 * to be merged with the normal php provided query arguments.
67 * Tries to use the REQUEST_URI data if available and parses it
68 * according to the wiki's configuration looking for any known pattern.
69 *
70 * If the REQUEST_URI is not provided we'll fall back on the PATH_INFO
71 * provided by the server if any and use that to set a 'title' parameter.
72 *
73 * @param string $want If this is not 'all', then the function
74 * will return an empty array if it determines that the URL is
75 * inside a rewrite path.
76 *
77 * @return Array: Any query arguments found in path matches.
78 */
79 public static function getPathInfo( $want = 'all' ) {
80 global $wgUsePathInfo;
81 // PATH_INFO is mangled due to http://bugs.php.net/bug.php?id=31892
82 // And also by Apache 2.x, double slashes are converted to single slashes.
83 // So we will use REQUEST_URI if possible.
84 $matches = array();
85 if ( !empty( $_SERVER['REQUEST_URI'] ) ) {
86 // Slurp out the path portion to examine...
87 $url = $_SERVER['REQUEST_URI'];
88 if ( !preg_match( '!^https?://!', $url ) ) {
89 $url = 'http://unused' . $url;
90 }
91 wfSuppressWarnings();
92 $a = parse_url( $url );
93 wfRestoreWarnings();
94 if ( $a ) {
95 $path = isset( $a['path'] ) ? $a['path'] : '';
96
97 global $wgScript;
98 if ( $path == $wgScript && $want !== 'all' ) {
99 // Script inside a rewrite path?
100 // Abort to keep from breaking...
101 return $matches;
102 }
103
104 $router = new PathRouter;
105
106 // Raw PATH_INFO style
107 $router->add( "$wgScript/$1" );
108
109 if ( isset( $_SERVER['SCRIPT_NAME'] )
110 && preg_match( '/\.php5?/', $_SERVER['SCRIPT_NAME'] ) )
111 {
112 # Check for SCRIPT_NAME, we handle index.php explicitly
113 # But we do have some other .php files such as img_auth.php
114 # Don't let root article paths clober the parsing for them
115 $router->add( $_SERVER['SCRIPT_NAME'] . "/$1" );
116 }
117
118 global $wgArticlePath;
119 if ( $wgArticlePath ) {
120 $router->add( $wgArticlePath );
121 }
122
123 global $wgActionPaths;
124 if ( $wgActionPaths ) {
125 $router->add( $wgActionPaths, array( 'action' => '$key' ) );
126 }
127
128 global $wgVariantArticlePath, $wgContLang;
129 if ( $wgVariantArticlePath ) {
130 $router->add( $wgVariantArticlePath,
131 array( 'variant' => '$2' ),
132 array( '$2' => $wgContLang->getVariants() )
133 );
134 }
135
136 wfRunHooks( 'WebRequestPathInfoRouter', array( $router ) );
137
138 $matches = $router->parse( $path );
139 }
140 } elseif ( $wgUsePathInfo ) {
141 if ( isset( $_SERVER['ORIG_PATH_INFO'] ) && $_SERVER['ORIG_PATH_INFO'] != '' ) {
142 // Mangled PATH_INFO
143 // http://bugs.php.net/bug.php?id=31892
144 // Also reported when ini_get('cgi.fix_pathinfo')==false
145 $matches['title'] = substr( $_SERVER['ORIG_PATH_INFO'], 1 );
146
147 } elseif ( isset( $_SERVER['PATH_INFO'] ) && $_SERVER['PATH_INFO'] != '' ) {
148 // Regular old PATH_INFO yay
149 $matches['title'] = substr( $_SERVER['PATH_INFO'], 1 );
150 }
151 }
152
153 return $matches;
154 }
155
156 /**
157 * Work out an appropriate URL prefix containing scheme and host, based on
158 * information detected from $_SERVER
159 *
160 * @return string
161 */
162 public static function detectServer() {
163 list( $proto, $stdPort ) = self::detectProtocolAndStdPort();
164
165 $varNames = array( 'HTTP_HOST', 'SERVER_NAME', 'HOSTNAME', 'SERVER_ADDR' );
166 $host = 'localhost';
167 $port = $stdPort;
168 foreach ( $varNames as $varName ) {
169 if ( !isset( $_SERVER[$varName] ) ) {
170 continue;
171 }
172 $parts = IP::splitHostAndPort( $_SERVER[$varName] );
173 if ( !$parts ) {
174 // Invalid, do not use
175 continue;
176 }
177 $host = $parts[0];
178 if ( $parts[1] === false ) {
179 if ( isset( $_SERVER['SERVER_PORT'] ) ) {
180 $port = $_SERVER['SERVER_PORT'];
181 } // else leave it as $stdPort
182 } else {
183 $port = $parts[1];
184 }
185 break;
186 }
187
188 return $proto . '://' . IP::combineHostAndPort( $host, $port, $stdPort );
189 }
190
191 /**
192 * @return array
193 */
194 public static function detectProtocolAndStdPort() {
195 if ( ( isset( $_SERVER['HTTPS'] ) && $_SERVER['HTTPS'] == 'on' ) ||
196 ( isset( $_SERVER['HTTP_X_FORWARDED_PROTO'] ) &&
197 $_SERVER['HTTP_X_FORWARDED_PROTO'] == 'https' ) ) {
198 $arr = array( 'https', 443 );
199 } else {
200 $arr = array( 'http', 80 );
201 }
202 return $arr;
203 }
204
205 /**
206 * @return string
207 */
208 public static function detectProtocol() {
209 list( $proto, ) = self::detectProtocolAndStdPort();
210 return $proto;
211 }
212
213 /**
214 * Check for title, action, and/or variant data in the URL
215 * and interpolate it into the GET variables.
216 * This should only be run after $wgContLang is available,
217 * as we may need the list of language variants to determine
218 * available variant URLs.
219 */
220 public function interpolateTitle() {
221 // bug 16019: title interpolation on API queries is useless and sometimes harmful
222 if ( defined( 'MW_API' ) ) {
223 return;
224 }
225
226 $matches = self::getPathInfo( 'title' );
227 foreach ( $matches as $key => $val ) {
228 $this->data[$key] = $_GET[$key] = $_REQUEST[$key] = $val;
229 }
230 }
231
232 /**
233 * URL rewriting function; tries to extract page title and,
234 * optionally, one other fixed parameter value from a URL path.
235 *
236 * @param string $path the URL path given from the client
237 * @param array $bases one or more URLs, optionally with $1 at the end
238 * @param string $key if provided, the matching key in $bases will be
239 * passed on as the value of this URL parameter
240 * @return array of URL variables to interpolate; empty if no match
241 */
242 static function extractTitle( $path, $bases, $key = false ) {
243 foreach ( (array)$bases as $keyValue => $base ) {
244 // Find the part after $wgArticlePath
245 $base = str_replace( '$1', '', $base );
246 $baseLen = strlen( $base );
247 if ( substr( $path, 0, $baseLen ) == $base ) {
248 $raw = substr( $path, $baseLen );
249 if ( $raw !== '' ) {
250 $matches = array( 'title' => rawurldecode( $raw ) );
251 if ( $key ) {
252 $matches[$key] = $keyValue;
253 }
254 return $matches;
255 }
256 }
257 }
258 return array();
259 }
260
261 /**
262 * Recursively strips slashes from the given array;
263 * used for undoing the evil that is magic_quotes_gpc.
264 *
265 * @param array $arr will be modified
266 * @param bool $topLevel Specifies if the array passed is from the top
267 * level of the source. In PHP5 magic_quotes only escapes the first level
268 * of keys that belong to an array.
269 * @return array the original array
270 * @see http://www.php.net/manual/en/function.get-magic-quotes-gpc.php#49612
271 */
272 private function &fix_magic_quotes( &$arr, $topLevel = true ) {
273 $clean = array();
274 foreach ( $arr as $key => $val ) {
275 if ( is_array( $val ) ) {
276 $cleanKey = $topLevel ? stripslashes( $key ) : $key;
277 $clean[$cleanKey] = $this->fix_magic_quotes( $arr[$key], false );
278 } else {
279 $cleanKey = stripslashes( $key );
280 $clean[$cleanKey] = stripslashes( $val );
281 }
282 }
283 $arr = $clean;
284 return $arr;
285 }
286
287 /**
288 * If magic_quotes_gpc option is on, run the global arrays
289 * through fix_magic_quotes to strip out the stupid slashes.
290 * WARNING: This should only be done once! Running a second
291 * time could damage the values.
292 */
293 private function checkMagicQuotes() {
294 $mustFixQuotes = function_exists( 'get_magic_quotes_gpc' )
295 && get_magic_quotes_gpc();
296 if ( $mustFixQuotes ) {
297 $this->fix_magic_quotes( $_COOKIE );
298 $this->fix_magic_quotes( $_ENV );
299 $this->fix_magic_quotes( $_GET );
300 $this->fix_magic_quotes( $_POST );
301 $this->fix_magic_quotes( $_REQUEST );
302 $this->fix_magic_quotes( $_SERVER );
303 }
304 }
305
306 /**
307 * Recursively normalizes UTF-8 strings in the given array.
308 *
309 * @param $data string|array
310 * @return array|string cleaned-up version of the given
311 * @private
312 */
313 function normalizeUnicode( $data ) {
314 if ( is_array( $data ) ) {
315 foreach ( $data as $key => $val ) {
316 $data[$key] = $this->normalizeUnicode( $val );
317 }
318 } else {
319 global $wgContLang;
320 $data = isset( $wgContLang ) ? $wgContLang->normalize( $data ) : UtfNormal::cleanUp( $data );
321 }
322 return $data;
323 }
324
325 /**
326 * Fetch a value from the given array or return $default if it's not set.
327 *
328 * @param $arr Array
329 * @param $name String
330 * @param $default Mixed
331 * @return mixed
332 */
333 private function getGPCVal( $arr, $name, $default ) {
334 # PHP is so nice to not touch input data, except sometimes:
335 # http://us2.php.net/variables.external#language.variables.external.dot-in-names
336 # Work around PHP *feature* to avoid *bugs* elsewhere.
337 $name = strtr( $name, '.', '_' );
338 if ( isset( $arr[$name] ) ) {
339 global $wgContLang;
340 $data = $arr[$name];
341 if ( isset( $_GET[$name] ) && !is_array( $data ) ) {
342 # Check for alternate/legacy character encoding.
343 if ( isset( $wgContLang ) ) {
344 $data = $wgContLang->checkTitleEncoding( $data );
345 }
346 }
347 $data = $this->normalizeUnicode( $data );
348 return $data;
349 } else {
350 return $default;
351 }
352 }
353
354 /**
355 * Fetch a scalar from the input or return $default if it's not set.
356 * Returns a string. Arrays are discarded. Useful for
357 * non-freeform text inputs (e.g. predefined internal text keys
358 * selected by a drop-down menu). For freeform input, see getText().
359 *
360 * @param $name String
361 * @param string $default optional default (or NULL)
362 * @return String
363 */
364 public function getVal( $name, $default = null ) {
365 $val = $this->getGPCVal( $this->data, $name, $default );
366 if ( is_array( $val ) ) {
367 $val = $default;
368 }
369 if ( is_null( $val ) ) {
370 return $val;
371 } else {
372 return (string)$val;
373 }
374 }
375
376 /**
377 * Set an arbitrary value into our get/post data.
378 *
379 * @param string $key key name to use
380 * @param $value Mixed: value to set
381 * @return Mixed: old value if one was present, null otherwise
382 */
383 public function setVal( $key, $value ) {
384 $ret = isset( $this->data[$key] ) ? $this->data[$key] : null;
385 $this->data[$key] = $value;
386 return $ret;
387 }
388
389 /**
390 * Unset an arbitrary value from our get/post data.
391 *
392 * @param string $key key name to use
393 * @return Mixed: old value if one was present, null otherwise
394 */
395 public function unsetVal( $key ) {
396 if ( !isset( $this->data[$key] ) ) {
397 $ret = null;
398 } else {
399 $ret = $this->data[$key];
400 unset( $this->data[$key] );
401 }
402 return $ret;
403 }
404
405 /**
406 * Fetch an array from the input or return $default if it's not set.
407 * If source was scalar, will return an array with a single element.
408 * If no source and no default, returns NULL.
409 *
410 * @param $name String
411 * @param array $default optional default (or NULL)
412 * @return Array
413 */
414 public function getArray( $name, $default = null ) {
415 $val = $this->getGPCVal( $this->data, $name, $default );
416 if ( is_null( $val ) ) {
417 return null;
418 } else {
419 return (array)$val;
420 }
421 }
422
423 /**
424 * Fetch an array of integers, or return $default if it's not set.
425 * If source was scalar, will return an array with a single element.
426 * If no source and no default, returns NULL.
427 * If an array is returned, contents are guaranteed to be integers.
428 *
429 * @param $name String
430 * @param array $default option default (or NULL)
431 * @return Array of ints
432 */
433 public function getIntArray( $name, $default = null ) {
434 $val = $this->getArray( $name, $default );
435 if ( is_array( $val ) ) {
436 $val = array_map( 'intval', $val );
437 }
438 return $val;
439 }
440
441 /**
442 * Fetch an integer value from the input or return $default if not set.
443 * Guaranteed to return an integer; non-numeric input will typically
444 * return 0.
445 *
446 * @param $name String
447 * @param $default Integer
448 * @return Integer
449 */
450 public function getInt( $name, $default = 0 ) {
451 return intval( $this->getVal( $name, $default ) );
452 }
453
454 /**
455 * Fetch an integer value from the input or return null if empty.
456 * Guaranteed to return an integer or null; non-numeric input will
457 * typically return null.
458 *
459 * @param $name String
460 * @return Integer
461 */
462 public function getIntOrNull( $name ) {
463 $val = $this->getVal( $name );
464 return is_numeric( $val )
465 ? intval( $val )
466 : null;
467 }
468
469 /**
470 * Fetch a boolean value from the input or return $default if not set.
471 * Guaranteed to return true or false, with normal PHP semantics for
472 * boolean interpretation of strings.
473 *
474 * @param $name String
475 * @param $default Boolean
476 * @return Boolean
477 */
478 public function getBool( $name, $default = false ) {
479 return (bool)$this->getVal( $name, $default );
480 }
481
482 /**
483 * Fetch a boolean value from the input or return $default if not set.
484 * Unlike getBool, the string "false" will result in boolean false, which is
485 * useful when interpreting information sent from JavaScript.
486 *
487 * @param $name String
488 * @param $default Boolean
489 * @return Boolean
490 */
491 public function getFuzzyBool( $name, $default = false ) {
492 return $this->getBool( $name, $default ) && strcasecmp( $this->getVal( $name ), 'false' ) !== 0;
493 }
494
495 /**
496 * Return true if the named value is set in the input, whatever that
497 * value is (even "0"). Return false if the named value is not set.
498 * Example use is checking for the presence of check boxes in forms.
499 *
500 * @param $name String
501 * @return Boolean
502 */
503 public function getCheck( $name ) {
504 # Checkboxes and buttons are only present when clicked
505 # Presence connotes truth, absence false
506 return $this->getVal( $name, null ) !== null;
507 }
508
509 /**
510 * Fetch a text string from the given array or return $default if it's not
511 * set. Carriage returns are stripped from the text, and with some language
512 * modules there is an input transliteration applied. This should generally
513 * be used for form "<textarea>" and "<input>" fields. Used for
514 * user-supplied freeform text input (for which input transformations may
515 * be required - e.g. Esperanto x-coding).
516 *
517 * @param $name String
518 * @param string $default optional
519 * @return String
520 */
521 public function getText( $name, $default = '' ) {
522 global $wgContLang;
523 $val = $this->getVal( $name, $default );
524 return str_replace( "\r\n", "\n",
525 $wgContLang->recodeInput( $val ) );
526 }
527
528 /**
529 * Extracts the given named values into an array.
530 * If no arguments are given, returns all input values.
531 * No transformation is performed on the values.
532 *
533 * @return array
534 */
535 public function getValues() {
536 $names = func_get_args();
537 if ( count( $names ) == 0 ) {
538 $names = array_keys( $this->data );
539 }
540
541 $retVal = array();
542 foreach ( $names as $name ) {
543 $value = $this->getGPCVal( $this->data, $name, null );
544 if ( !is_null( $value ) ) {
545 $retVal[$name] = $value;
546 }
547 }
548 return $retVal;
549 }
550
551 /**
552 * Returns the names of all input values excluding those in $exclude.
553 *
554 * @param $exclude Array
555 * @return array
556 */
557 public function getValueNames( $exclude = array() ) {
558 return array_diff( array_keys( $this->getValues() ), $exclude );
559 }
560
561 /**
562 * Get the values passed in the query string.
563 * No transformation is performed on the values.
564 *
565 * @return Array
566 */
567 public function getQueryValues() {
568 return $_GET;
569 }
570
571 /**
572 * Get the HTTP method used for this request.
573 *
574 * @return String
575 */
576 public function getMethod() {
577 return isset( $_SERVER['REQUEST_METHOD'] ) ? $_SERVER['REQUEST_METHOD'] : 'GET';
578 }
579
580 /**
581 * Returns true if the present request was reached by a POST operation,
582 * false otherwise (GET, HEAD, or command-line).
583 *
584 * Note that values retrieved by the object may come from the
585 * GET URL etc even on a POST request.
586 *
587 * @return Boolean
588 */
589 public function wasPosted() {
590 return $this->getMethod() == 'POST';
591 }
592
593 /**
594 * Returns true if there is a session cookie set.
595 * This does not necessarily mean that the user is logged in!
596 *
597 * If you want to check for an open session, use session_id()
598 * instead; that will also tell you if the session was opened
599 * during the current request (in which case the cookie will
600 * be sent back to the client at the end of the script run).
601 *
602 * @return Boolean
603 */
604 public function checkSessionCookie() {
605 return isset( $_COOKIE[session_name()] );
606 }
607
608 /**
609 * Get a cookie from the $_COOKIE jar
610 *
611 * @param string $key the name of the cookie
612 * @param string $prefix a prefix to use for the cookie name, if not $wgCookiePrefix
613 * @param $default Mixed: what to return if the value isn't found
614 * @return Mixed: cookie value or $default if the cookie not set
615 */
616 public function getCookie( $key, $prefix = null, $default = null ) {
617 if ( $prefix === null ) {
618 global $wgCookiePrefix;
619 $prefix = $wgCookiePrefix;
620 }
621 return $this->getGPCVal( $_COOKIE, $prefix . $key, $default );
622 }
623
624 /**
625 * Return the path and query string portion of the request URI.
626 * This will be suitable for use as a relative link in HTML output.
627 *
628 * @throws MWException
629 * @return String
630 */
631 public function getRequestURL() {
632 if ( isset( $_SERVER['REQUEST_URI'] ) && strlen( $_SERVER['REQUEST_URI'] ) ) {
633 $base = $_SERVER['REQUEST_URI'];
634 } elseif ( isset( $_SERVER['HTTP_X_ORIGINAL_URL'] ) && strlen( $_SERVER['HTTP_X_ORIGINAL_URL'] ) ) {
635 // Probably IIS; doesn't set REQUEST_URI
636 $base = $_SERVER['HTTP_X_ORIGINAL_URL'];
637 } elseif ( isset( $_SERVER['SCRIPT_NAME'] ) ) {
638 $base = $_SERVER['SCRIPT_NAME'];
639 if ( isset( $_SERVER['QUERY_STRING'] ) && $_SERVER['QUERY_STRING'] != '' ) {
640 $base .= '?' . $_SERVER['QUERY_STRING'];
641 }
642 } else {
643 // This shouldn't happen!
644 throw new MWException( "Web server doesn't provide either " .
645 "REQUEST_URI, HTTP_X_ORIGINAL_URL or SCRIPT_NAME. Report details " .
646 "of your web server configuration to http://bugzilla.wikimedia.org/" );
647 }
648 // User-agents should not send a fragment with the URI, but
649 // if they do, and the web server passes it on to us, we
650 // need to strip it or we get false-positive redirect loops
651 // or weird output URLs
652 $hash = strpos( $base, '#' );
653 if ( $hash !== false ) {
654 $base = substr( $base, 0, $hash );
655 }
656
657 if ( $base[0] == '/' ) {
658 // More than one slash will look like it is protocol relative
659 return preg_replace( '!^/+!', '/', $base );
660 } else {
661 // We may get paths with a host prepended; strip it.
662 return preg_replace( '!^[^:]+://[^/]+/+!', '/', $base );
663 }
664 }
665
666 /**
667 * Return the request URI with the canonical service and hostname, path,
668 * and query string. This will be suitable for use as an absolute link
669 * in HTML or other output.
670 *
671 * If $wgServer is protocol-relative, this will return a fully
672 * qualified URL with the protocol that was used for this request.
673 *
674 * @return String
675 */
676 public function getFullRequestURL() {
677 return wfExpandUrl( $this->getRequestURL(), PROTO_CURRENT );
678 }
679
680 /**
681 * Take an arbitrary query and rewrite the present URL to include it
682 * @param string $query query string fragment; do not include initial '?'
683 *
684 * @return String
685 */
686 public function appendQuery( $query ) {
687 return $this->appendQueryArray( wfCgiToArray( $query ) );
688 }
689
690 /**
691 * HTML-safe version of appendQuery().
692 * @deprecated: Deprecated in 1.20, warnings in 1.21, remove in 1.22.
693 *
694 * @param string $query query string fragment; do not include initial '?'
695 * @return String
696 */
697 public function escapeAppendQuery( $query ) {
698 return htmlspecialchars( $this->appendQuery( $query ) );
699 }
700
701 /**
702 * @param $key
703 * @param $value
704 * @param $onlyquery bool
705 * @return String
706 */
707 public function appendQueryValue( $key, $value, $onlyquery = false ) {
708 return $this->appendQueryArray( array( $key => $value ), $onlyquery );
709 }
710
711 /**
712 * Appends or replaces value of query variables.
713 *
714 * @param array $array of values to replace/add to query
715 * @param bool $onlyquery whether to only return the query string and not
716 * the complete URL
717 * @return String
718 */
719 public function appendQueryArray( $array, $onlyquery = false ) {
720 global $wgTitle;
721 $newquery = $this->getQueryValues();
722 unset( $newquery['title'] );
723 $newquery = array_merge( $newquery, $array );
724 $query = wfArrayToCgi( $newquery );
725 return $onlyquery ? $query : $wgTitle->getLocalURL( $query );
726 }
727
728 /**
729 * Check for limit and offset parameters on the input, and return sensible
730 * defaults if not given. The limit must be positive and is capped at 5000.
731 * Offset must be positive but is not capped.
732 *
733 * @param $deflimit Integer: limit to use if no input and the user hasn't set the option.
734 * @param string $optionname to specify an option other than rclimit to pull from.
735 * @return array first element is limit, second is offset
736 */
737 public function getLimitOffset( $deflimit = 50, $optionname = 'rclimit' ) {
738 global $wgUser;
739
740 $limit = $this->getInt( 'limit', 0 );
741 if ( $limit < 0 ) {
742 $limit = 0;
743 }
744 if ( ( $limit == 0 ) && ( $optionname != '' ) ) {
745 $limit = $wgUser->getIntOption( $optionname );
746 }
747 if ( $limit <= 0 ) {
748 $limit = $deflimit;
749 }
750 if ( $limit > 5000 ) {
751 $limit = 5000; # We have *some* limits...
752 }
753
754 $offset = $this->getInt( 'offset', 0 );
755 if ( $offset < 0 ) {
756 $offset = 0;
757 }
758
759 return array( $limit, $offset );
760 }
761
762 /**
763 * Return the path to the temporary file where PHP has stored the upload.
764 *
765 * @param $key String:
766 * @return string or NULL if no such file.
767 */
768 public function getFileTempname( $key ) {
769 $file = new WebRequestUpload( $this, $key );
770 return $file->getTempName();
771 }
772
773 /**
774 * Return the size of the upload, or 0.
775 *
776 * @deprecated since 1.17
777 * @param $key String:
778 * @return integer
779 */
780 public function getFileSize( $key ) {
781 wfDeprecated( __METHOD__, '1.17' );
782 $file = new WebRequestUpload( $this, $key );
783 return $file->getSize();
784 }
785
786 /**
787 * Return the upload error or 0
788 *
789 * @param $key String:
790 * @return integer
791 */
792 public function getUploadError( $key ) {
793 $file = new WebRequestUpload( $this, $key );
794 return $file->getError();
795 }
796
797 /**
798 * Return the original filename of the uploaded file, as reported by
799 * the submitting user agent. HTML-style character entities are
800 * interpreted and normalized to Unicode normalization form C, in part
801 * to deal with weird input from Safari with non-ASCII filenames.
802 *
803 * Other than this the name is not verified for being a safe filename.
804 *
805 * @param $key String:
806 * @return string or NULL if no such file.
807 */
808 public function getFileName( $key ) {
809 $file = new WebRequestUpload( $this, $key );
810 return $file->getName();
811 }
812
813 /**
814 * Return a WebRequestUpload object corresponding to the key
815 *
816 * @param $key string
817 * @return WebRequestUpload
818 */
819 public function getUpload( $key ) {
820 return new WebRequestUpload( $this, $key );
821 }
822
823 /**
824 * Return a handle to WebResponse style object, for setting cookies,
825 * headers and other stuff, for Request being worked on.
826 *
827 * @return WebResponse
828 */
829 public function response() {
830 /* Lazy initialization of response object for this request */
831 if ( !is_object( $this->response ) ) {
832 $class = ( $this instanceof FauxRequest ) ? 'FauxResponse' : 'WebResponse';
833 $this->response = new $class();
834 }
835 return $this->response;
836 }
837
838 /**
839 * Initialise the header list
840 */
841 private function initHeaders() {
842 if ( count( $this->headers ) ) {
843 return;
844 }
845
846 if ( function_exists( 'apache_request_headers' ) ) {
847 foreach ( apache_request_headers() as $tempName => $tempValue ) {
848 $this->headers[strtoupper( $tempName )] = $tempValue;
849 }
850 } else {
851 foreach ( $_SERVER as $name => $value ) {
852 if ( substr( $name, 0, 5 ) === 'HTTP_' ) {
853 $name = str_replace( '_', '-', substr( $name, 5 ) );
854 $this->headers[$name] = $value;
855 } elseif ( $name === 'CONTENT_LENGTH' ) {
856 $this->headers['CONTENT-LENGTH'] = $value;
857 }
858 }
859 }
860 }
861
862 /**
863 * Get an array containing all request headers
864 *
865 * @return Array mapping header name to its value
866 */
867 public function getAllHeaders() {
868 $this->initHeaders();
869 return $this->headers;
870 }
871
872 /**
873 * Get a request header, or false if it isn't set
874 * @param string $name case-insensitive header name
875 *
876 * @return string|bool False on failure
877 */
878 public function getHeader( $name ) {
879 $this->initHeaders();
880 $name = strtoupper( $name );
881 if ( isset( $this->headers[$name] ) ) {
882 return $this->headers[$name];
883 } else {
884 return false;
885 }
886 }
887
888 /**
889 * Get data from $_SESSION
890 *
891 * @param string $key name of key in $_SESSION
892 * @return Mixed
893 */
894 public function getSessionData( $key ) {
895 if ( !isset( $_SESSION[$key] ) ) {
896 return null;
897 }
898 return $_SESSION[$key];
899 }
900
901 /**
902 * Set session data
903 *
904 * @param string $key name of key in $_SESSION
905 * @param $data Mixed
906 */
907 public function setSessionData( $key, $data ) {
908 $_SESSION[$key] = $data;
909 }
910
911 /**
912 * Check if Internet Explorer will detect an incorrect cache extension in
913 * PATH_INFO or QUERY_STRING. If the request can't be allowed, show an error
914 * message or redirect to a safer URL. Returns true if the URL is OK, and
915 * false if an error message has been shown and the request should be aborted.
916 *
917 * @param $extWhitelist array
918 * @throws HttpError
919 * @return bool
920 */
921 public function checkUrlExtension( $extWhitelist = array() ) {
922 global $wgScriptExtension;
923 $extWhitelist[] = ltrim( $wgScriptExtension, '.' );
924 if ( IEUrlExtension::areServerVarsBad( $_SERVER, $extWhitelist ) ) {
925 if ( !$this->wasPosted() ) {
926 $newUrl = IEUrlExtension::fixUrlForIE6(
927 $this->getFullRequestURL(), $extWhitelist );
928 if ( $newUrl !== false ) {
929 $this->doSecurityRedirect( $newUrl );
930 return false;
931 }
932 }
933 throw new HttpError( 403,
934 'Invalid file extension found in the path info or query string.' );
935 }
936 return true;
937 }
938
939 /**
940 * Attempt to redirect to a URL with a QUERY_STRING that's not dangerous in
941 * IE 6. Returns true if it was successful, false otherwise.
942 *
943 * @param $url string
944 * @return bool
945 */
946 protected function doSecurityRedirect( $url ) {
947 header( 'Location: ' . $url );
948 header( 'Content-Type: text/html' );
949 $encUrl = htmlspecialchars( $url );
950 echo <<<HTML
951 <html>
952 <head>
953 <title>Security redirect</title>
954 </head>
955 <body>
956 <h1>Security redirect</h1>
957 <p>
958 We can't serve non-HTML content from the URL you have requested, because
959 Internet Explorer would interpret it as an incorrect and potentially dangerous
960 content type.</p>
961 <p>Instead, please use <a href="$encUrl">this URL</a>, which is the same as the URL you have requested, except that
962 "&amp;*" is appended. This prevents Internet Explorer from seeing a bogus file
963 extension.
964 </p>
965 </body>
966 </html>
967 HTML;
968 echo "\n";
969 return true;
970 }
971
972 /**
973 * Returns true if the PATH_INFO ends with an extension other than a script
974 * extension. This could confuse IE for scripts that send arbitrary data which
975 * is not HTML but may be detected as such.
976 *
977 * Various past attempts to use the URL to make this check have generally
978 * run up against the fact that CGI does not provide a standard method to
979 * determine the URL. PATH_INFO may be mangled (e.g. if cgi.fix_pathinfo=0),
980 * but only by prefixing it with the script name and maybe some other stuff,
981 * the extension is not mangled. So this should be a reasonably portable
982 * way to perform this security check.
983 *
984 * Also checks for anything that looks like a file extension at the end of
985 * QUERY_STRING, since IE 6 and earlier will use this to get the file type
986 * if there was no dot before the question mark (bug 28235).
987 *
988 * @deprecated Use checkUrlExtension().
989 *
990 * @param $extWhitelist array
991 *
992 * @return bool
993 */
994 public function isPathInfoBad( $extWhitelist = array() ) {
995 wfDeprecated( __METHOD__, '1.17' );
996 global $wgScriptExtension;
997 $extWhitelist[] = ltrim( $wgScriptExtension, '.' );
998 return IEUrlExtension::areServerVarsBad( $_SERVER, $extWhitelist );
999 }
1000
1001 /**
1002 * Parse the Accept-Language header sent by the client into an array
1003 * @return array array( languageCode => q-value ) sorted by q-value in descending order then
1004 * appearing time in the header in ascending order.
1005 * May contain the "language" '*', which applies to languages other than those explicitly listed.
1006 * This is aligned with rfc2616 section 14.4
1007 * Preference for earlier languages appears in rfc3282 as an extension to HTTP/1.1.
1008 */
1009 public function getAcceptLang() {
1010 // Modified version of code found at http://www.thefutureoftheweb.com/blog/use-accept-language-header
1011 $acceptLang = $this->getHeader( 'Accept-Language' );
1012 if ( !$acceptLang ) {
1013 return array();
1014 }
1015
1016 // Return the language codes in lower case
1017 $acceptLang = strtolower( $acceptLang );
1018
1019 // Break up string into pieces (languages and q factors)
1020 $lang_parse = null;
1021 preg_match_all( '/([a-z]{1,8}(-[a-z]{1,8})*|\*)\s*(;\s*q\s*=\s*(1(\.0{0,3})?|0(\.[0-9]{0,3})?)?)?/',
1022 $acceptLang, $lang_parse );
1023
1024 if ( !count( $lang_parse[1] ) ) {
1025 return array();
1026 }
1027
1028 $langcodes = $lang_parse[1];
1029 $qvalues = $lang_parse[4];
1030 $indices = range( 0, count( $lang_parse[1] ) - 1 );
1031
1032 // Set default q factor to 1
1033 foreach ( $indices as $index ) {
1034 if ( $qvalues[$index] === '' ) {
1035 $qvalues[$index] = 1;
1036 } elseif ( $qvalues[$index] == 0 ) {
1037 unset( $langcodes[$index], $qvalues[$index], $indices[$index] );
1038 }
1039 }
1040
1041 // Sort list. First by $qvalues, then by order. Reorder $langcodes the same way
1042 array_multisort( $qvalues, SORT_DESC, SORT_NUMERIC, $indices, $langcodes );
1043
1044 // Create a list like "en" => 0.8
1045 $langs = array_combine( $langcodes, $qvalues );
1046
1047 return $langs;
1048 }
1049
1050 /**
1051 * Fetch the raw IP from the request
1052 *
1053 * @since 1.19
1054 *
1055 * @throws MWException
1056 * @return String
1057 */
1058 protected function getRawIP() {
1059 if ( !isset( $_SERVER['REMOTE_ADDR'] ) ) {
1060 return null;
1061 }
1062
1063 if ( is_array( $_SERVER['REMOTE_ADDR'] ) || strpos( $_SERVER['REMOTE_ADDR'], ',' ) !== false ) {
1064 throw new MWException( __METHOD__ . " : Could not determine the remote IP address due to multiple values." );
1065 } else {
1066 $ipchain = $_SERVER['REMOTE_ADDR'];
1067 }
1068
1069 return IP::canonicalize( $ipchain );
1070 }
1071
1072 /**
1073 * Work out the IP address based on various globals
1074 * For trusted proxies, use the XFF client IP (first of the chain)
1075 *
1076 * @since 1.19
1077 *
1078 * @throws MWException
1079 * @return string
1080 */
1081 public function getIP() {
1082 global $wgUsePrivateIPs;
1083
1084 # Return cached result
1085 if ( $this->ip !== null ) {
1086 return $this->ip;
1087 }
1088
1089 # collect the originating ips
1090 $ip = $this->getRawIP();
1091
1092 # Append XFF
1093 $forwardedFor = $this->getHeader( 'X-Forwarded-For' );
1094 if ( $forwardedFor !== false ) {
1095 $ipchain = array_map( 'trim', explode( ',', $forwardedFor ) );
1096 $ipchain = array_reverse( $ipchain );
1097 if ( $ip ) {
1098 array_unshift( $ipchain, $ip );
1099 }
1100
1101 # Step through XFF list and find the last address in the list which is a
1102 # trusted server. Set $ip to the IP address given by that trusted server,
1103 # unless the address is not sensible (e.g. private). However, prefer private
1104 # IP addresses over proxy servers controlled by this site (more sensible).
1105 foreach ( $ipchain as $i => $curIP ) {
1106 $curIP = IP::canonicalize( $curIP );
1107 if ( wfIsTrustedProxy( $curIP ) && isset( $ipchain[$i + 1] ) ) {
1108 if ( wfIsConfiguredProxy( $curIP ) || // bug 48919
1109 ( IP::isPublic( $ipchain[$i + 1] ) || $wgUsePrivateIPs )
1110 ) {
1111 $ip = IP::canonicalize( $ipchain[$i + 1] );
1112 continue;
1113 }
1114 }
1115 break;
1116 }
1117 }
1118
1119 # Allow extensions to improve our guess
1120 wfRunHooks( 'GetIP', array( &$ip ) );
1121
1122 if ( !$ip ) {
1123 throw new MWException( "Unable to determine IP" );
1124 }
1125
1126 wfDebug( "IP: $ip\n" );
1127 $this->ip = $ip;
1128 return $ip;
1129 }
1130
1131 /**
1132 * @param string $ip
1133 * @return void
1134 * @since 1.21
1135 */
1136 public function setIP( $ip ) {
1137 $this->ip = $ip;
1138 }
1139 }
1140
1141 /**
1142 * Object to access the $_FILES array
1143 */
1144 class WebRequestUpload {
1145 protected $request;
1146 protected $doesExist;
1147 protected $fileInfo;
1148
1149 /**
1150 * Constructor. Should only be called by WebRequest
1151 *
1152 * @param $request WebRequest The associated request
1153 * @param string $key Key in $_FILES array (name of form field)
1154 */
1155 public function __construct( $request, $key ) {
1156 $this->request = $request;
1157 $this->doesExist = isset( $_FILES[$key] );
1158 if ( $this->doesExist ) {
1159 $this->fileInfo = $_FILES[$key];
1160 }
1161 }
1162
1163 /**
1164 * Return whether a file with this name was uploaded.
1165 *
1166 * @return bool
1167 */
1168 public function exists() {
1169 return $this->doesExist;
1170 }
1171
1172 /**
1173 * Return the original filename of the uploaded file
1174 *
1175 * @return mixed Filename or null if non-existent
1176 */
1177 public function getName() {
1178 if ( !$this->exists() ) {
1179 return null;
1180 }
1181
1182 global $wgContLang;
1183 $name = $this->fileInfo['name'];
1184
1185 # Safari sends filenames in HTML-encoded Unicode form D...
1186 # Horrid and evil! Let's try to make some kind of sense of it.
1187 $name = Sanitizer::decodeCharReferences( $name );
1188 $name = $wgContLang->normalize( $name );
1189 wfDebug( __METHOD__ . ": {$this->fileInfo['name']} normalized to '$name'\n" );
1190 return $name;
1191 }
1192
1193 /**
1194 * Return the file size of the uploaded file
1195 *
1196 * @return int File size or zero if non-existent
1197 */
1198 public function getSize() {
1199 if ( !$this->exists() ) {
1200 return 0;
1201 }
1202
1203 return $this->fileInfo['size'];
1204 }
1205
1206 /**
1207 * Return the path to the temporary file
1208 *
1209 * @return mixed Path or null if non-existent
1210 */
1211 public function getTempName() {
1212 if ( !$this->exists() ) {
1213 return null;
1214 }
1215
1216 return $this->fileInfo['tmp_name'];
1217 }
1218
1219 /**
1220 * Return the upload error. See link for explanation
1221 * http://www.php.net/manual/en/features.file-upload.errors.php
1222 *
1223 * @return int One of the UPLOAD_ constants, 0 if non-existent
1224 */
1225 public function getError() {
1226 if ( !$this->exists() ) {
1227 return 0; # UPLOAD_ERR_OK
1228 }
1229
1230 return $this->fileInfo['error'];
1231 }
1232
1233 /**
1234 * Returns whether this upload failed because of overflow of a maximum set
1235 * in php.ini
1236 *
1237 * @return bool
1238 */
1239 public function isIniSizeOverflow() {
1240 if ( $this->getError() == UPLOAD_ERR_INI_SIZE ) {
1241 # PHP indicated that upload_max_filesize is exceeded
1242 return true;
1243 }
1244
1245 $contentLength = $this->request->getHeader( 'CONTENT_LENGTH' );
1246 if ( $contentLength > wfShorthandToInteger( ini_get( 'post_max_size' ) ) ) {
1247 # post_max_size is exceeded
1248 return true;
1249 }
1250
1251 return false;
1252 }
1253 }
1254
1255 /**
1256 * WebRequest clone which takes values from a provided array.
1257 *
1258 * @ingroup HTTP
1259 */
1260 class FauxRequest extends WebRequest {
1261 private $wasPosted = false;
1262 private $session = array();
1263
1264 /**
1265 * @param array $data of *non*-urlencoded key => value pairs, the
1266 * fake GET/POST values
1267 * @param bool $wasPosted whether to treat the data as POST
1268 * @param $session Mixed: session array or null
1269 * @throws MWException
1270 */
1271 public function __construct( $data = array(), $wasPosted = false, $session = null ) {
1272 if ( is_array( $data ) ) {
1273 $this->data = $data;
1274 } else {
1275 throw new MWException( "FauxRequest() got bogus data" );
1276 }
1277 $this->wasPosted = $wasPosted;
1278 if ( $session ) {
1279 $this->session = $session;
1280 }
1281 }
1282
1283 /**
1284 * @param $method string
1285 * @throws MWException
1286 */
1287 private function notImplemented( $method ) {
1288 throw new MWException( "{$method}() not implemented" );
1289 }
1290
1291 /**
1292 * @param $name string
1293 * @param $default string
1294 * @return string
1295 */
1296 public function getText( $name, $default = '' ) {
1297 # Override; don't recode since we're using internal data
1298 return (string)$this->getVal( $name, $default );
1299 }
1300
1301 /**
1302 * @return Array
1303 */
1304 public function getValues() {
1305 return $this->data;
1306 }
1307
1308 /**
1309 * @return array
1310 */
1311 public function getQueryValues() {
1312 if ( $this->wasPosted ) {
1313 return array();
1314 } else {
1315 return $this->data;
1316 }
1317 }
1318
1319 public function getMethod() {
1320 return $this->wasPosted ? 'POST' : 'GET';
1321 }
1322
1323 /**
1324 * @return bool
1325 */
1326 public function wasPosted() {
1327 return $this->wasPosted;
1328 }
1329
1330 public function getCookie( $key, $prefix = null, $default = null ) {
1331 return $default;
1332 }
1333
1334 public function checkSessionCookie() {
1335 return false;
1336 }
1337
1338 public function getRequestURL() {
1339 $this->notImplemented( __METHOD__ );
1340 }
1341
1342 /**
1343 * @param string $name The name of the header to get (case insensitive).
1344 * @return bool|string
1345 */
1346 public function getHeader( $name ) {
1347 $name = strtoupper( $name );
1348 return isset( $this->headers[$name] ) ? $this->headers[$name] : false;
1349 }
1350
1351 /**
1352 * @param $name string
1353 * @param $val string
1354 */
1355 public function setHeader( $name, $val ) {
1356 $name = strtoupper( $name );
1357 $this->headers[$name] = $val;
1358 }
1359
1360 /**
1361 * @param $key
1362 * @return mixed
1363 */
1364 public function getSessionData( $key ) {
1365 if ( isset( $this->session[$key] ) ) {
1366 return $this->session[$key];
1367 }
1368 return null;
1369 }
1370
1371 /**
1372 * @param $key
1373 * @param $data
1374 */
1375 public function setSessionData( $key, $data ) {
1376 $this->session[$key] = $data;
1377 }
1378
1379 /**
1380 * @return array|Mixed|null
1381 */
1382 public function getSessionArray() {
1383 return $this->session;
1384 }
1385
1386 /**
1387 * @param array $extWhitelist
1388 * @return bool
1389 */
1390 public function isPathInfoBad( $extWhitelist = array() ) {
1391 return false;
1392 }
1393
1394 /**
1395 * @param array $extWhitelist
1396 * @return bool
1397 */
1398 public function checkUrlExtension( $extWhitelist = array() ) {
1399 return true;
1400 }
1401
1402 /**
1403 * @return string
1404 */
1405 protected function getRawIP() {
1406 return '127.0.0.1';
1407 }
1408 }
1409
1410 /**
1411 * Similar to FauxRequest, but only fakes URL parameters and method
1412 * (POST or GET) and use the base request for the remaining stuff
1413 * (cookies, session and headers).
1414 *
1415 * @ingroup HTTP
1416 * @since 1.19
1417 */
1418 class DerivativeRequest extends FauxRequest {
1419 private $base;
1420
1421 public function __construct( WebRequest $base, $data, $wasPosted = false ) {
1422 $this->base = $base;
1423 parent::__construct( $data, $wasPosted );
1424 }
1425
1426 public function getCookie( $key, $prefix = null, $default = null ) {
1427 return $this->base->getCookie( $key, $prefix, $default );
1428 }
1429
1430 public function checkSessionCookie() {
1431 return $this->base->checkSessionCookie();
1432 }
1433
1434 public function getHeader( $name ) {
1435 return $this->base->getHeader( $name );
1436 }
1437
1438 public function getAllHeaders() {
1439 return $this->base->getAllHeaders();
1440 }
1441
1442 public function getSessionData( $key ) {
1443 return $this->base->getSessionData( $key );
1444 }
1445
1446 public function setSessionData( $key, $data ) {
1447 $this->base->setSessionData( $key, $data );
1448 }
1449
1450 public function getAcceptLang() {
1451 return $this->base->getAcceptLang();
1452 }
1453
1454 public function getIP() {
1455 return $this->base->getIP();
1456 }
1457 }