Merge "Allow local interwiki links with an empty title part"
[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 * https://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 /**
54 * Cached URL protocol
55 * @var string
56 */
57 protected $protocol;
58
59 public function __construct() {
60 /// @todo FIXME: This preemptive de-quoting can interfere with other web libraries
61 /// and increases our memory footprint. It would be cleaner to do on
62 /// demand; but currently we have no wrapper for $_SERVER etc.
63 $this->checkMagicQuotes();
64
65 // POST overrides GET data
66 // We don't use $_REQUEST here to avoid interference from cookies...
67 $this->data = $_POST + $_GET;
68 }
69
70 /**
71 * Extract relevant query arguments from the http request uri's path
72 * to be merged with the normal php provided query arguments.
73 * Tries to use the REQUEST_URI data if available and parses it
74 * according to the wiki's configuration looking for any known pattern.
75 *
76 * If the REQUEST_URI is not provided we'll fall back on the PATH_INFO
77 * provided by the server if any and use that to set a 'title' parameter.
78 *
79 * @param string $want If this is not 'all', then the function
80 * will return an empty array if it determines that the URL is
81 * inside a rewrite path.
82 *
83 * @return array Any query arguments found in path matches.
84 */
85 public static function getPathInfo( $want = 'all' ) {
86 global $wgUsePathInfo;
87 // PATH_INFO is mangled due to http://bugs.php.net/bug.php?id=31892
88 // And also by Apache 2.x, double slashes are converted to single slashes.
89 // So we will use REQUEST_URI if possible.
90 $matches = array();
91 if ( !empty( $_SERVER['REQUEST_URI'] ) ) {
92 // Slurp out the path portion to examine...
93 $url = $_SERVER['REQUEST_URI'];
94 if ( !preg_match( '!^https?://!', $url ) ) {
95 $url = 'http://unused' . $url;
96 }
97 wfSuppressWarnings();
98 $a = parse_url( $url );
99 wfRestoreWarnings();
100 if ( $a ) {
101 $path = isset( $a['path'] ) ? $a['path'] : '';
102
103 global $wgScript;
104 if ( $path == $wgScript && $want !== 'all' ) {
105 // Script inside a rewrite path?
106 // Abort to keep from breaking...
107 return $matches;
108 }
109
110 $router = new PathRouter;
111
112 // Raw PATH_INFO style
113 $router->add( "$wgScript/$1" );
114
115 if ( isset( $_SERVER['SCRIPT_NAME'] )
116 && preg_match( '/\.php5?/', $_SERVER['SCRIPT_NAME'] )
117 ) {
118 # Check for SCRIPT_NAME, we handle index.php explicitly
119 # But we do have some other .php files such as img_auth.php
120 # Don't let root article paths clober the parsing for them
121 $router->add( $_SERVER['SCRIPT_NAME'] . "/$1" );
122 }
123
124 global $wgArticlePath;
125 if ( $wgArticlePath ) {
126 $router->add( $wgArticlePath );
127 }
128
129 global $wgActionPaths;
130 if ( $wgActionPaths ) {
131 $router->add( $wgActionPaths, array( 'action' => '$key' ) );
132 }
133
134 global $wgVariantArticlePath, $wgContLang;
135 if ( $wgVariantArticlePath ) {
136 $router->add( $wgVariantArticlePath,
137 array( 'variant' => '$2' ),
138 array( '$2' => $wgContLang->getVariants() )
139 );
140 }
141
142 wfRunHooks( 'WebRequestPathInfoRouter', array( $router ) );
143
144 $matches = $router->parse( $path );
145 }
146 } elseif ( $wgUsePathInfo ) {
147 if ( isset( $_SERVER['ORIG_PATH_INFO'] ) && $_SERVER['ORIG_PATH_INFO'] != '' ) {
148 // Mangled PATH_INFO
149 // http://bugs.php.net/bug.php?id=31892
150 // Also reported when ini_get('cgi.fix_pathinfo')==false
151 $matches['title'] = substr( $_SERVER['ORIG_PATH_INFO'], 1 );
152
153 } elseif ( isset( $_SERVER['PATH_INFO'] ) && $_SERVER['PATH_INFO'] != '' ) {
154 // Regular old PATH_INFO yay
155 $matches['title'] = substr( $_SERVER['PATH_INFO'], 1 );
156 }
157 }
158
159 return $matches;
160 }
161
162 /**
163 * Work out an appropriate URL prefix containing scheme and host, based on
164 * information detected from $_SERVER
165 *
166 * @return string
167 */
168 public static function detectServer() {
169 $proto = self::detectProtocol();
170 $stdPort = $proto === 'https' ? 443 : 80;
171
172 $varNames = array( 'HTTP_HOST', 'SERVER_NAME', 'HOSTNAME', 'SERVER_ADDR' );
173 $host = 'localhost';
174 $port = $stdPort;
175 foreach ( $varNames as $varName ) {
176 if ( !isset( $_SERVER[$varName] ) ) {
177 continue;
178 }
179 $parts = IP::splitHostAndPort( $_SERVER[$varName] );
180 if ( !$parts ) {
181 // Invalid, do not use
182 continue;
183 }
184 $host = $parts[0];
185 if ( $parts[1] === false ) {
186 if ( isset( $_SERVER['SERVER_PORT'] ) ) {
187 $port = $_SERVER['SERVER_PORT'];
188 } // else leave it as $stdPort
189 } else {
190 $port = $parts[1];
191 }
192 break;
193 }
194
195 return $proto . '://' . IP::combineHostAndPort( $host, $port, $stdPort );
196 }
197
198 /**
199 * Detect the protocol from $_SERVER.
200 * This is for use prior to Setup.php, when no WebRequest object is available.
201 * At other times, use the non-static function getProtocol().
202 *
203 * @return array
204 */
205 public static function detectProtocol() {
206 if ( ( isset( $_SERVER['HTTPS'] ) && $_SERVER['HTTPS'] == 'on' ) ||
207 ( isset( $_SERVER['HTTP_X_FORWARDED_PROTO'] ) &&
208 $_SERVER['HTTP_X_FORWARDED_PROTO'] == 'https' ) ) {
209 return 'https';
210 } else {
211 return 'http';
212 }
213 }
214
215 /**
216 * Get the current URL protocol (http or https)
217 * @return string
218 */
219 public function getProtocol() {
220 if ( $this->protocol === null ) {
221 $this->protocol = self::detectProtocol();
222 }
223 return $this->protocol;
224 }
225
226 /**
227 * Check for title, action, and/or variant data in the URL
228 * and interpolate it into the GET variables.
229 * This should only be run after $wgContLang is available,
230 * as we may need the list of language variants to determine
231 * available variant URLs.
232 */
233 public function interpolateTitle() {
234 // bug 16019: title interpolation on API queries is useless and sometimes harmful
235 if ( defined( 'MW_API' ) ) {
236 return;
237 }
238
239 $matches = self::getPathInfo( 'title' );
240 foreach ( $matches as $key => $val ) {
241 $this->data[$key] = $_GET[$key] = $_REQUEST[$key] = $val;
242 }
243 }
244
245 /**
246 * URL rewriting function; tries to extract page title and,
247 * optionally, one other fixed parameter value from a URL path.
248 *
249 * @param string $path The URL path given from the client
250 * @param array $bases One or more URLs, optionally with $1 at the end
251 * @param string $key If provided, the matching key in $bases will be
252 * passed on as the value of this URL parameter
253 * @return array Array of URL variables to interpolate; empty if no match
254 */
255 static function extractTitle( $path, $bases, $key = false ) {
256 foreach ( (array)$bases as $keyValue => $base ) {
257 // Find the part after $wgArticlePath
258 $base = str_replace( '$1', '', $base );
259 $baseLen = strlen( $base );
260 if ( substr( $path, 0, $baseLen ) == $base ) {
261 $raw = substr( $path, $baseLen );
262 if ( $raw !== '' ) {
263 $matches = array( 'title' => rawurldecode( $raw ) );
264 if ( $key ) {
265 $matches[$key] = $keyValue;
266 }
267 return $matches;
268 }
269 }
270 }
271 return array();
272 }
273
274 /**
275 * Recursively strips slashes from the given array;
276 * used for undoing the evil that is magic_quotes_gpc.
277 *
278 * @param array $arr will be modified
279 * @param bool $topLevel Specifies if the array passed is from the top
280 * level of the source. In PHP5 magic_quotes only escapes the first level
281 * of keys that belong to an array.
282 * @return array The original array
283 * @see http://www.php.net/manual/en/function.get-magic-quotes-gpc.php#49612
284 */
285 private function &fix_magic_quotes( &$arr, $topLevel = true ) {
286 $clean = array();
287 foreach ( $arr as $key => $val ) {
288 if ( is_array( $val ) ) {
289 $cleanKey = $topLevel ? stripslashes( $key ) : $key;
290 $clean[$cleanKey] = $this->fix_magic_quotes( $arr[$key], false );
291 } else {
292 $cleanKey = stripslashes( $key );
293 $clean[$cleanKey] = stripslashes( $val );
294 }
295 }
296 $arr = $clean;
297 return $arr;
298 }
299
300 /**
301 * If magic_quotes_gpc option is on, run the global arrays
302 * through fix_magic_quotes to strip out the stupid slashes.
303 * WARNING: This should only be done once! Running a second
304 * time could damage the values.
305 */
306 private function checkMagicQuotes() {
307 $mustFixQuotes = function_exists( 'get_magic_quotes_gpc' )
308 && get_magic_quotes_gpc();
309 if ( $mustFixQuotes ) {
310 $this->fix_magic_quotes( $_COOKIE );
311 $this->fix_magic_quotes( $_ENV );
312 $this->fix_magic_quotes( $_GET );
313 $this->fix_magic_quotes( $_POST );
314 $this->fix_magic_quotes( $_REQUEST );
315 $this->fix_magic_quotes( $_SERVER );
316 }
317 }
318
319 /**
320 * Recursively normalizes UTF-8 strings in the given array.
321 *
322 * @param string|array $data
323 * @return array|string Cleaned-up version of the given
324 * @private
325 */
326 function normalizeUnicode( $data ) {
327 if ( is_array( $data ) ) {
328 foreach ( $data as $key => $val ) {
329 $data[$key] = $this->normalizeUnicode( $val );
330 }
331 } else {
332 global $wgContLang;
333 $data = isset( $wgContLang ) ? $wgContLang->normalize( $data ) : UtfNormal::cleanUp( $data );
334 }
335 return $data;
336 }
337
338 /**
339 * Fetch a value from the given array or return $default if it's not set.
340 *
341 * @param array $arr
342 * @param string $name
343 * @param mixed $default
344 * @return mixed
345 */
346 private function getGPCVal( $arr, $name, $default ) {
347 # PHP is so nice to not touch input data, except sometimes:
348 # http://us2.php.net/variables.external#language.variables.external.dot-in-names
349 # Work around PHP *feature* to avoid *bugs* elsewhere.
350 $name = strtr( $name, '.', '_' );
351 if ( isset( $arr[$name] ) ) {
352 global $wgContLang;
353 $data = $arr[$name];
354 if ( isset( $_GET[$name] ) && !is_array( $data ) ) {
355 # Check for alternate/legacy character encoding.
356 if ( isset( $wgContLang ) ) {
357 $data = $wgContLang->checkTitleEncoding( $data );
358 }
359 }
360 $data = $this->normalizeUnicode( $data );
361 return $data;
362 } else {
363 return $default;
364 }
365 }
366
367 /**
368 * Fetch a scalar from the input or return $default if it's not set.
369 * Returns a string. Arrays are discarded. Useful for
370 * non-freeform text inputs (e.g. predefined internal text keys
371 * selected by a drop-down menu). For freeform input, see getText().
372 *
373 * @param string $name
374 * @param string $default Optional default (or null)
375 * @return string
376 */
377 public function getVal( $name, $default = null ) {
378 $val = $this->getGPCVal( $this->data, $name, $default );
379 if ( is_array( $val ) ) {
380 $val = $default;
381 }
382 if ( is_null( $val ) ) {
383 return $val;
384 } else {
385 return (string)$val;
386 }
387 }
388
389 /**
390 * Set an arbitrary value into our get/post data.
391 *
392 * @param string $key Key name to use
393 * @param mixed $value Value to set
394 * @return mixed Old value if one was present, null otherwise
395 */
396 public function setVal( $key, $value ) {
397 $ret = isset( $this->data[$key] ) ? $this->data[$key] : null;
398 $this->data[$key] = $value;
399 return $ret;
400 }
401
402 /**
403 * Unset an arbitrary value from our get/post data.
404 *
405 * @param string $key Key name to use
406 * @return mixed Old value if one was present, null otherwise
407 */
408 public function unsetVal( $key ) {
409 if ( !isset( $this->data[$key] ) ) {
410 $ret = null;
411 } else {
412 $ret = $this->data[$key];
413 unset( $this->data[$key] );
414 }
415 return $ret;
416 }
417
418 /**
419 * Fetch an array from the input or return $default if it's not set.
420 * If source was scalar, will return an array with a single element.
421 * If no source and no default, returns null.
422 *
423 * @param string $name
424 * @param array $default Optional default (or null)
425 * @return array
426 */
427 public function getArray( $name, $default = null ) {
428 $val = $this->getGPCVal( $this->data, $name, $default );
429 if ( is_null( $val ) ) {
430 return null;
431 } else {
432 return (array)$val;
433 }
434 }
435
436 /**
437 * Fetch an array of integers, or return $default if it's not set.
438 * If source was scalar, will return an array with a single element.
439 * If no source and no default, returns null.
440 * If an array is returned, contents are guaranteed to be integers.
441 *
442 * @param string $name
443 * @param array $default Option default (or null)
444 * @return array Array of ints
445 */
446 public function getIntArray( $name, $default = null ) {
447 $val = $this->getArray( $name, $default );
448 if ( is_array( $val ) ) {
449 $val = array_map( 'intval', $val );
450 }
451 return $val;
452 }
453
454 /**
455 * Fetch an integer value from the input or return $default if not set.
456 * Guaranteed to return an integer; non-numeric input will typically
457 * return 0.
458 *
459 * @param string $name
460 * @param int $default
461 * @return int
462 */
463 public function getInt( $name, $default = 0 ) {
464 return intval( $this->getVal( $name, $default ) );
465 }
466
467 /**
468 * Fetch an integer value from the input or return null if empty.
469 * Guaranteed to return an integer or null; non-numeric input will
470 * typically return null.
471 *
472 * @param string $name
473 * @return int|null
474 */
475 public function getIntOrNull( $name ) {
476 $val = $this->getVal( $name );
477 return is_numeric( $val )
478 ? intval( $val )
479 : null;
480 }
481
482 /**
483 * Fetch a floating point value from the input or return $default if not set.
484 * Guaranteed to return a float; non-numeric input will typically
485 * return 0.
486 *
487 * @since 1.23
488 * @param string $name
489 * @param float $default
490 * @return float
491 */
492 public function getFloat( $name, $default = 0.0 ) {
493 return floatval( $this->getVal( $name, $default ) );
494 }
495
496 /**
497 * Fetch a boolean value from the input or return $default if not set.
498 * Guaranteed to return true or false, with normal PHP semantics for
499 * boolean interpretation of strings.
500 *
501 * @param string $name
502 * @param bool $default
503 * @return bool
504 */
505 public function getBool( $name, $default = false ) {
506 return (bool)$this->getVal( $name, $default );
507 }
508
509 /**
510 * Fetch a boolean value from the input or return $default if not set.
511 * Unlike getBool, the string "false" will result in boolean false, which is
512 * useful when interpreting information sent from JavaScript.
513 *
514 * @param string $name
515 * @param bool $default
516 * @return bool
517 */
518 public function getFuzzyBool( $name, $default = false ) {
519 return $this->getBool( $name, $default ) && strcasecmp( $this->getVal( $name ), 'false' ) !== 0;
520 }
521
522 /**
523 * Return true if the named value is set in the input, whatever that
524 * value is (even "0"). Return false if the named value is not set.
525 * Example use is checking for the presence of check boxes in forms.
526 *
527 * @param string $name
528 * @return bool
529 */
530 public function getCheck( $name ) {
531 # Checkboxes and buttons are only present when clicked
532 # Presence connotes truth, absence false
533 return $this->getVal( $name, null ) !== null;
534 }
535
536 /**
537 * Fetch a text string from the given array or return $default if it's not
538 * set. Carriage returns are stripped from the text, and with some language
539 * modules there is an input transliteration applied. This should generally
540 * be used for form "<textarea>" and "<input>" fields. Used for
541 * user-supplied freeform text input (for which input transformations may
542 * be required - e.g. Esperanto x-coding).
543 *
544 * @param string $name
545 * @param string $default Optional
546 * @return string
547 */
548 public function getText( $name, $default = '' ) {
549 global $wgContLang;
550 $val = $this->getVal( $name, $default );
551 return str_replace( "\r\n", "\n",
552 $wgContLang->recodeInput( $val ) );
553 }
554
555 /**
556 * Extracts the given named values into an array.
557 * If no arguments are given, returns all input values.
558 * No transformation is performed on the values.
559 *
560 * @return array
561 */
562 public function getValues() {
563 $names = func_get_args();
564 if ( count( $names ) == 0 ) {
565 $names = array_keys( $this->data );
566 }
567
568 $retVal = array();
569 foreach ( $names as $name ) {
570 $value = $this->getGPCVal( $this->data, $name, null );
571 if ( !is_null( $value ) ) {
572 $retVal[$name] = $value;
573 }
574 }
575 return $retVal;
576 }
577
578 /**
579 * Returns the names of all input values excluding those in $exclude.
580 *
581 * @param array $exclude
582 * @return array
583 */
584 public function getValueNames( $exclude = array() ) {
585 return array_diff( array_keys( $this->getValues() ), $exclude );
586 }
587
588 /**
589 * Get the values passed in the query string.
590 * No transformation is performed on the values.
591 *
592 * @return array
593 */
594 public function getQueryValues() {
595 return $_GET;
596 }
597
598 /**
599 * Return the contents of the Query with no decoding. Use when you need to
600 * know exactly what was sent, e.g. for an OAuth signature over the elements.
601 *
602 * @return string
603 */
604 public function getRawQueryString() {
605 return $_SERVER['QUERY_STRING'];
606 }
607
608 /**
609 * Return the contents of the POST with no decoding. Use when you need to
610 * know exactly what was sent, e.g. for an OAuth signature over the elements.
611 *
612 * @return string
613 */
614 public function getRawPostString() {
615 if ( !$this->wasPosted() ) {
616 return '';
617 }
618 return $this->getRawInput();
619 }
620
621 /**
622 * Return the raw request body, with no processing. Cached since some methods
623 * disallow reading the stream more than once. As stated in the php docs, this
624 * does not work with enctype="multipart/form-data".
625 *
626 * @return string
627 */
628 public function getRawInput() {
629 static $input = null;
630 if ( $input === null ) {
631 $input = file_get_contents( 'php://input' );
632 }
633 return $input;
634 }
635
636 /**
637 * Get the HTTP method used for this request.
638 *
639 * @return string
640 */
641 public function getMethod() {
642 return isset( $_SERVER['REQUEST_METHOD'] ) ? $_SERVER['REQUEST_METHOD'] : 'GET';
643 }
644
645 /**
646 * Returns true if the present request was reached by a POST operation,
647 * false otherwise (GET, HEAD, or command-line).
648 *
649 * Note that values retrieved by the object may come from the
650 * GET URL etc even on a POST request.
651 *
652 * @return bool
653 */
654 public function wasPosted() {
655 return $this->getMethod() == 'POST';
656 }
657
658 /**
659 * Returns true if there is a session cookie set.
660 * This does not necessarily mean that the user is logged in!
661 *
662 * If you want to check for an open session, use session_id()
663 * instead; that will also tell you if the session was opened
664 * during the current request (in which case the cookie will
665 * be sent back to the client at the end of the script run).
666 *
667 * @return bool
668 */
669 public function checkSessionCookie() {
670 return isset( $_COOKIE[session_name()] );
671 }
672
673 /**
674 * Get a cookie from the $_COOKIE jar
675 *
676 * @param string $key The name of the cookie
677 * @param string $prefix A prefix to use for the cookie name, if not $wgCookiePrefix
678 * @param mixed $default What to return if the value isn't found
679 * @return mixed Cookie value or $default if the cookie not set
680 */
681 public function getCookie( $key, $prefix = null, $default = null ) {
682 if ( $prefix === null ) {
683 global $wgCookiePrefix;
684 $prefix = $wgCookiePrefix;
685 }
686 return $this->getGPCVal( $_COOKIE, $prefix . $key, $default );
687 }
688
689 /**
690 * Return the path and query string portion of the request URI.
691 * This will be suitable for use as a relative link in HTML output.
692 *
693 * @throws MWException
694 * @return string
695 */
696 public function getRequestURL() {
697 if ( isset( $_SERVER['REQUEST_URI'] ) && strlen( $_SERVER['REQUEST_URI'] ) ) {
698 $base = $_SERVER['REQUEST_URI'];
699 } elseif ( isset( $_SERVER['HTTP_X_ORIGINAL_URL'] )
700 && strlen( $_SERVER['HTTP_X_ORIGINAL_URL'] )
701 ) {
702 // Probably IIS; doesn't set REQUEST_URI
703 $base = $_SERVER['HTTP_X_ORIGINAL_URL'];
704 } elseif ( isset( $_SERVER['SCRIPT_NAME'] ) ) {
705 $base = $_SERVER['SCRIPT_NAME'];
706 if ( isset( $_SERVER['QUERY_STRING'] ) && $_SERVER['QUERY_STRING'] != '' ) {
707 $base .= '?' . $_SERVER['QUERY_STRING'];
708 }
709 } else {
710 // This shouldn't happen!
711 throw new MWException( "Web server doesn't provide either " .
712 "REQUEST_URI, HTTP_X_ORIGINAL_URL or SCRIPT_NAME. Report details " .
713 "of your web server configuration to http://bugzilla.wikimedia.org/" );
714 }
715 // User-agents should not send a fragment with the URI, but
716 // if they do, and the web server passes it on to us, we
717 // need to strip it or we get false-positive redirect loops
718 // or weird output URLs
719 $hash = strpos( $base, '#' );
720 if ( $hash !== false ) {
721 $base = substr( $base, 0, $hash );
722 }
723
724 if ( $base[0] == '/' ) {
725 // More than one slash will look like it is protocol relative
726 return preg_replace( '!^/+!', '/', $base );
727 } else {
728 // We may get paths with a host prepended; strip it.
729 return preg_replace( '!^[^:]+://[^/]+/+!', '/', $base );
730 }
731 }
732
733 /**
734 * Return the request URI with the canonical service and hostname, path,
735 * and query string. This will be suitable for use as an absolute link
736 * in HTML or other output.
737 *
738 * If $wgServer is protocol-relative, this will return a fully
739 * qualified URL with the protocol that was used for this request.
740 *
741 * @return string
742 */
743 public function getFullRequestURL() {
744 return wfExpandUrl( $this->getRequestURL(), PROTO_CURRENT );
745 }
746
747 /**
748 * Take an arbitrary query and rewrite the present URL to include it
749 * @param string $query query string fragment; do not include initial '?'
750 *
751 * @return string
752 */
753 public function appendQuery( $query ) {
754 return $this->appendQueryArray( wfCgiToArray( $query ) );
755 }
756
757 /**
758 * @param string $key
759 * @param string $value
760 * @param bool $onlyquery
761 * @return string
762 */
763 public function appendQueryValue( $key, $value, $onlyquery = false ) {
764 return $this->appendQueryArray( array( $key => $value ), $onlyquery );
765 }
766
767 /**
768 * Appends or replaces value of query variables.
769 *
770 * @param array $array Array of values to replace/add to query
771 * @param bool $onlyquery Whether to only return the query string and not the complete URL
772 * @return string
773 */
774 public function appendQueryArray( $array, $onlyquery = false ) {
775 global $wgTitle;
776 $newquery = $this->getQueryValues();
777 unset( $newquery['title'] );
778 $newquery = array_merge( $newquery, $array );
779 $query = wfArrayToCgi( $newquery );
780 return $onlyquery ? $query : $wgTitle->getLocalURL( $query );
781 }
782
783 /**
784 * Check for limit and offset parameters on the input, and return sensible
785 * defaults if not given. The limit must be positive and is capped at 5000.
786 * Offset must be positive but is not capped.
787 *
788 * @param int $deflimit Limit to use if no input and the user hasn't set the option.
789 * @param string $optionname To specify an option other than rclimit to pull from.
790 * @return array First element is limit, second is offset
791 */
792 public function getLimitOffset( $deflimit = 50, $optionname = 'rclimit' ) {
793 global $wgUser;
794
795 $limit = $this->getInt( 'limit', 0 );
796 if ( $limit < 0 ) {
797 $limit = 0;
798 }
799 if ( ( $limit == 0 ) && ( $optionname != '' ) ) {
800 $limit = $wgUser->getIntOption( $optionname );
801 }
802 if ( $limit <= 0 ) {
803 $limit = $deflimit;
804 }
805 if ( $limit > 5000 ) {
806 $limit = 5000; # We have *some* limits...
807 }
808
809 $offset = $this->getInt( 'offset', 0 );
810 if ( $offset < 0 ) {
811 $offset = 0;
812 }
813
814 return array( $limit, $offset );
815 }
816
817 /**
818 * Return the path to the temporary file where PHP has stored the upload.
819 *
820 * @param string $key
821 * @return string|null string or null if no such file.
822 */
823 public function getFileTempname( $key ) {
824 $file = new WebRequestUpload( $this, $key );
825 return $file->getTempName();
826 }
827
828 /**
829 * Return the upload error or 0
830 *
831 * @param string $key
832 * @return int
833 */
834 public function getUploadError( $key ) {
835 $file = new WebRequestUpload( $this, $key );
836 return $file->getError();
837 }
838
839 /**
840 * Return the original filename of the uploaded file, as reported by
841 * the submitting user agent. HTML-style character entities are
842 * interpreted and normalized to Unicode normalization form C, in part
843 * to deal with weird input from Safari with non-ASCII filenames.
844 *
845 * Other than this the name is not verified for being a safe filename.
846 *
847 * @param string $key
848 * @return string|null string or null if no such file.
849 */
850 public function getFileName( $key ) {
851 $file = new WebRequestUpload( $this, $key );
852 return $file->getName();
853 }
854
855 /**
856 * Return a WebRequestUpload object corresponding to the key
857 *
858 * @param string $key
859 * @return WebRequestUpload
860 */
861 public function getUpload( $key ) {
862 return new WebRequestUpload( $this, $key );
863 }
864
865 /**
866 * Return a handle to WebResponse style object, for setting cookies,
867 * headers and other stuff, for Request being worked on.
868 *
869 * @return WebResponse
870 */
871 public function response() {
872 /* Lazy initialization of response object for this request */
873 if ( !is_object( $this->response ) ) {
874 $class = ( $this instanceof FauxRequest ) ? 'FauxResponse' : 'WebResponse';
875 $this->response = new $class();
876 }
877 return $this->response;
878 }
879
880 /**
881 * Initialise the header list
882 */
883 private function initHeaders() {
884 if ( count( $this->headers ) ) {
885 return;
886 }
887
888 $apacheHeaders = function_exists( 'apache_request_headers' ) ? apache_request_headers() : false;
889 if ( $apacheHeaders ) {
890 foreach ( $apacheHeaders as $tempName => $tempValue ) {
891 $this->headers[strtoupper( $tempName )] = $tempValue;
892 }
893 } else {
894 foreach ( $_SERVER as $name => $value ) {
895 if ( substr( $name, 0, 5 ) === 'HTTP_' ) {
896 $name = str_replace( '_', '-', substr( $name, 5 ) );
897 $this->headers[$name] = $value;
898 } elseif ( $name === 'CONTENT_LENGTH' ) {
899 $this->headers['CONTENT-LENGTH'] = $value;
900 }
901 }
902 }
903 }
904
905 /**
906 * Get an array containing all request headers
907 *
908 * @return array Mapping header name to its value
909 */
910 public function getAllHeaders() {
911 $this->initHeaders();
912 return $this->headers;
913 }
914
915 /**
916 * Get a request header, or false if it isn't set
917 * @param string $name case-insensitive header name
918 *
919 * @return string|bool False on failure
920 */
921 public function getHeader( $name ) {
922 $this->initHeaders();
923 $name = strtoupper( $name );
924 if ( isset( $this->headers[$name] ) ) {
925 return $this->headers[$name];
926 } else {
927 return false;
928 }
929 }
930
931 /**
932 * Get data from $_SESSION
933 *
934 * @param string $key Name of key in $_SESSION
935 * @return mixed
936 */
937 public function getSessionData( $key ) {
938 if ( !isset( $_SESSION[$key] ) ) {
939 return null;
940 }
941 return $_SESSION[$key];
942 }
943
944 /**
945 * Set session data
946 *
947 * @param string $key Name of key in $_SESSION
948 * @param mixed $data
949 */
950 public function setSessionData( $key, $data ) {
951 $_SESSION[$key] = $data;
952 }
953
954 /**
955 * Check if Internet Explorer will detect an incorrect cache extension in
956 * PATH_INFO or QUERY_STRING. If the request can't be allowed, show an error
957 * message or redirect to a safer URL. Returns true if the URL is OK, and
958 * false if an error message has been shown and the request should be aborted.
959 *
960 * @param array $extWhitelist
961 * @throws HttpError
962 * @return bool
963 */
964 public function checkUrlExtension( $extWhitelist = array() ) {
965 global $wgScriptExtension;
966 $extWhitelist[] = ltrim( $wgScriptExtension, '.' );
967 if ( IEUrlExtension::areServerVarsBad( $_SERVER, $extWhitelist ) ) {
968 if ( !$this->wasPosted() ) {
969 $newUrl = IEUrlExtension::fixUrlForIE6(
970 $this->getFullRequestURL(), $extWhitelist );
971 if ( $newUrl !== false ) {
972 $this->doSecurityRedirect( $newUrl );
973 return false;
974 }
975 }
976 throw new HttpError( 403,
977 'Invalid file extension found in the path info or query string.' );
978 }
979 return true;
980 }
981
982 /**
983 * Attempt to redirect to a URL with a QUERY_STRING that's not dangerous in
984 * IE 6. Returns true if it was successful, false otherwise.
985 *
986 * @param string $url
987 * @return bool
988 */
989 protected function doSecurityRedirect( $url ) {
990 header( 'Location: ' . $url );
991 header( 'Content-Type: text/html' );
992 $encUrl = htmlspecialchars( $url );
993 echo <<<HTML
994 <html>
995 <head>
996 <title>Security redirect</title>
997 </head>
998 <body>
999 <h1>Security redirect</h1>
1000 <p>
1001 We can't serve non-HTML content from the URL you have requested, because
1002 Internet Explorer would interpret it as an incorrect and potentially dangerous
1003 content type.</p>
1004 <p>Instead, please use <a href="$encUrl">this URL</a>, which is the same as the
1005 URL you have requested, except that "&amp;*" is appended. This prevents Internet
1006 Explorer from seeing a bogus file extension.
1007 </p>
1008 </body>
1009 </html>
1010 HTML;
1011 echo "\n";
1012 return true;
1013 }
1014
1015 /**
1016 * Parse the Accept-Language header sent by the client into an array
1017 *
1018 * @return array array( languageCode => q-value ) sorted by q-value in
1019 * descending order then appearing time in the header in ascending order.
1020 * May contain the "language" '*', which applies to languages other than those explicitly listed.
1021 * This is aligned with rfc2616 section 14.4
1022 * Preference for earlier languages appears in rfc3282 as an extension to HTTP/1.1.
1023 */
1024 public function getAcceptLang() {
1025 // Modified version of code found at
1026 // http://www.thefutureoftheweb.com/blog/use-accept-language-header
1027 $acceptLang = $this->getHeader( 'Accept-Language' );
1028 if ( !$acceptLang ) {
1029 return array();
1030 }
1031
1032 // Return the language codes in lower case
1033 $acceptLang = strtolower( $acceptLang );
1034
1035 // Break up string into pieces (languages and q factors)
1036 $lang_parse = null;
1037 preg_match_all(
1038 '/([a-z]{1,8}(-[a-z]{1,8})*|\*)\s*(;\s*q\s*=\s*(1(\.0{0,3})?|0(\.[0-9]{0,3})?)?)?/',
1039 $acceptLang,
1040 $lang_parse
1041 );
1042
1043 if ( !count( $lang_parse[1] ) ) {
1044 return array();
1045 }
1046
1047 $langcodes = $lang_parse[1];
1048 $qvalues = $lang_parse[4];
1049 $indices = range( 0, count( $lang_parse[1] ) - 1 );
1050
1051 // Set default q factor to 1
1052 foreach ( $indices as $index ) {
1053 if ( $qvalues[$index] === '' ) {
1054 $qvalues[$index] = 1;
1055 } elseif ( $qvalues[$index] == 0 ) {
1056 unset( $langcodes[$index], $qvalues[$index], $indices[$index] );
1057 }
1058 }
1059
1060 // Sort list. First by $qvalues, then by order. Reorder $langcodes the same way
1061 array_multisort( $qvalues, SORT_DESC, SORT_NUMERIC, $indices, $langcodes );
1062
1063 // Create a list like "en" => 0.8
1064 $langs = array_combine( $langcodes, $qvalues );
1065
1066 return $langs;
1067 }
1068
1069 /**
1070 * Fetch the raw IP from the request
1071 *
1072 * @since 1.19
1073 *
1074 * @throws MWException
1075 * @return string
1076 */
1077 protected function getRawIP() {
1078 if ( !isset( $_SERVER['REMOTE_ADDR'] ) ) {
1079 return null;
1080 }
1081
1082 if ( is_array( $_SERVER['REMOTE_ADDR'] ) || strpos( $_SERVER['REMOTE_ADDR'], ',' ) !== false ) {
1083 throw new MWException( __METHOD__
1084 . " : Could not determine the remote IP address due to multiple values." );
1085 } else {
1086 $ipchain = $_SERVER['REMOTE_ADDR'];
1087 }
1088
1089 return IP::canonicalize( $ipchain );
1090 }
1091
1092 /**
1093 * Work out the IP address based on various globals
1094 * For trusted proxies, use the XFF client IP (first of the chain)
1095 *
1096 * @since 1.19
1097 *
1098 * @throws MWException
1099 * @return string
1100 */
1101 public function getIP() {
1102 global $wgUsePrivateIPs;
1103
1104 # Return cached result
1105 if ( $this->ip !== null ) {
1106 return $this->ip;
1107 }
1108
1109 # collect the originating ips
1110 $ip = $this->getRawIP();
1111 if ( !$ip ) {
1112 throw new MWException( 'Unable to determine IP.' );
1113 }
1114
1115 # Append XFF
1116 $forwardedFor = $this->getHeader( 'X-Forwarded-For' );
1117 if ( $forwardedFor !== false ) {
1118 $isConfigured = IP::isConfiguredProxy( $ip );
1119 $ipchain = array_map( 'trim', explode( ',', $forwardedFor ) );
1120 $ipchain = array_reverse( $ipchain );
1121 array_unshift( $ipchain, $ip );
1122
1123 # Step through XFF list and find the last address in the list which is a
1124 # trusted server. Set $ip to the IP address given by that trusted server,
1125 # unless the address is not sensible (e.g. private). However, prefer private
1126 # IP addresses over proxy servers controlled by this site (more sensible).
1127 # Note that some XFF values might be "unknown" with Squid/Varnish.
1128 foreach ( $ipchain as $i => $curIP ) {
1129 $curIP = IP::sanitizeIP( IP::canonicalize( $curIP ) );
1130 if ( !$curIP || !isset( $ipchain[$i + 1] ) || $ipchain[$i + 1] === 'unknown'
1131 || !IP::isTrustedProxy( $curIP )
1132 ) {
1133 break; // IP is not valid/trusted or does not point to anything
1134 }
1135 if (
1136 IP::isPublic( $ipchain[$i + 1] ) ||
1137 $wgUsePrivateIPs ||
1138 IP::isConfiguredProxy( $curIP ) // bug 48919; treat IP as sane
1139 ) {
1140 // Follow the next IP according to the proxy
1141 $nextIP = IP::canonicalize( $ipchain[$i + 1] );
1142 if ( !$nextIP && $isConfigured ) {
1143 // We have not yet made it past CDN/proxy servers of this site,
1144 // so either they are misconfigured or there is some IP spoofing.
1145 throw new MWException( "Invalid IP given in XFF '$forwardedFor'." );
1146 }
1147 $ip = $nextIP;
1148 // keep traversing the chain
1149 continue;
1150 }
1151 break;
1152 }
1153 }
1154
1155 # Allow extensions to improve our guess
1156 wfRunHooks( 'GetIP', array( &$ip ) );
1157
1158 if ( !$ip ) {
1159 throw new MWException( "Unable to determine IP." );
1160 }
1161
1162 wfDebug( "IP: $ip\n" );
1163 $this->ip = $ip;
1164 return $ip;
1165 }
1166
1167 /**
1168 * @param string $ip
1169 * @return void
1170 * @since 1.21
1171 */
1172 public function setIP( $ip ) {
1173 $this->ip = $ip;
1174 }
1175 }
1176
1177 /**
1178 * Object to access the $_FILES array
1179 */
1180 class WebRequestUpload {
1181 protected $request;
1182 protected $doesExist;
1183 protected $fileInfo;
1184
1185 /**
1186 * Constructor. Should only be called by WebRequest
1187 *
1188 * @param WebRequest $request The associated request
1189 * @param string $key Key in $_FILES array (name of form field)
1190 */
1191 public function __construct( $request, $key ) {
1192 $this->request = $request;
1193 $this->doesExist = isset( $_FILES[$key] );
1194 if ( $this->doesExist ) {
1195 $this->fileInfo = $_FILES[$key];
1196 }
1197 }
1198
1199 /**
1200 * Return whether a file with this name was uploaded.
1201 *
1202 * @return bool
1203 */
1204 public function exists() {
1205 return $this->doesExist;
1206 }
1207
1208 /**
1209 * Return the original filename of the uploaded file
1210 *
1211 * @return string|null Filename or null if non-existent
1212 */
1213 public function getName() {
1214 if ( !$this->exists() ) {
1215 return null;
1216 }
1217
1218 global $wgContLang;
1219 $name = $this->fileInfo['name'];
1220
1221 # Safari sends filenames in HTML-encoded Unicode form D...
1222 # Horrid and evil! Let's try to make some kind of sense of it.
1223 $name = Sanitizer::decodeCharReferences( $name );
1224 $name = $wgContLang->normalize( $name );
1225 wfDebug( __METHOD__ . ": {$this->fileInfo['name']} normalized to '$name'\n" );
1226 return $name;
1227 }
1228
1229 /**
1230 * Return the file size of the uploaded file
1231 *
1232 * @return int File size or zero if non-existent
1233 */
1234 public function getSize() {
1235 if ( !$this->exists() ) {
1236 return 0;
1237 }
1238
1239 return $this->fileInfo['size'];
1240 }
1241
1242 /**
1243 * Return the path to the temporary file
1244 *
1245 * @return string|null Path or null if non-existent
1246 */
1247 public function getTempName() {
1248 if ( !$this->exists() ) {
1249 return null;
1250 }
1251
1252 return $this->fileInfo['tmp_name'];
1253 }
1254
1255 /**
1256 * Return the upload error. See link for explanation
1257 * http://www.php.net/manual/en/features.file-upload.errors.php
1258 *
1259 * @return int One of the UPLOAD_ constants, 0 if non-existent
1260 */
1261 public function getError() {
1262 if ( !$this->exists() ) {
1263 return 0; # UPLOAD_ERR_OK
1264 }
1265
1266 return $this->fileInfo['error'];
1267 }
1268
1269 /**
1270 * Returns whether this upload failed because of overflow of a maximum set
1271 * in php.ini
1272 *
1273 * @return bool
1274 */
1275 public function isIniSizeOverflow() {
1276 if ( $this->getError() == UPLOAD_ERR_INI_SIZE ) {
1277 # PHP indicated that upload_max_filesize is exceeded
1278 return true;
1279 }
1280
1281 $contentLength = $this->request->getHeader( 'CONTENT_LENGTH' );
1282 if ( $contentLength > wfShorthandToInteger( ini_get( 'post_max_size' ) ) ) {
1283 # post_max_size is exceeded
1284 return true;
1285 }
1286
1287 return false;
1288 }
1289 }
1290
1291 /**
1292 * WebRequest clone which takes values from a provided array.
1293 *
1294 * @ingroup HTTP
1295 */
1296 class FauxRequest extends WebRequest {
1297 private $wasPosted = false;
1298 private $session = array();
1299
1300 /**
1301 * @param array $data of *non*-urlencoded key => value pairs, the
1302 * fake GET/POST values
1303 * @param bool $wasPosted Whether to treat the data as POST
1304 * @param array|null $session Session array or null
1305 * @param string $protocol 'http' or 'https'
1306 * @throws MWException
1307 */
1308 public function __construct( $data = array(), $wasPosted = false,
1309 $session = null, $protocol = 'http'
1310 ) {
1311 if ( is_array( $data ) ) {
1312 $this->data = $data;
1313 } else {
1314 throw new MWException( "FauxRequest() got bogus data" );
1315 }
1316 $this->wasPosted = $wasPosted;
1317 if ( $session ) {
1318 $this->session = $session;
1319 }
1320 $this->protocol = $protocol;
1321 }
1322
1323 /**
1324 * @param string $method
1325 * @throws MWException
1326 */
1327 private function notImplemented( $method ) {
1328 throw new MWException( "{$method}() not implemented" );
1329 }
1330
1331 /**
1332 * @param string $name
1333 * @param string $default
1334 * @return string
1335 */
1336 public function getText( $name, $default = '' ) {
1337 # Override; don't recode since we're using internal data
1338 return (string)$this->getVal( $name, $default );
1339 }
1340
1341 /**
1342 * @return array
1343 */
1344 public function getValues() {
1345 return $this->data;
1346 }
1347
1348 /**
1349 * @return array
1350 */
1351 public function getQueryValues() {
1352 if ( $this->wasPosted ) {
1353 return array();
1354 } else {
1355 return $this->data;
1356 }
1357 }
1358
1359 public function getMethod() {
1360 return $this->wasPosted ? 'POST' : 'GET';
1361 }
1362
1363 /**
1364 * @return bool
1365 */
1366 public function wasPosted() {
1367 return $this->wasPosted;
1368 }
1369
1370 public function getCookie( $key, $prefix = null, $default = null ) {
1371 return $default;
1372 }
1373
1374 public function checkSessionCookie() {
1375 return false;
1376 }
1377
1378 public function getRequestURL() {
1379 $this->notImplemented( __METHOD__ );
1380 }
1381
1382 public function getProtocol() {
1383 return $this->protocol;
1384 }
1385
1386 /**
1387 * @param string $name The name of the header to get (case insensitive).
1388 * @return bool|string
1389 */
1390 public function getHeader( $name ) {
1391 $name = strtoupper( $name );
1392 return isset( $this->headers[$name] ) ? $this->headers[$name] : false;
1393 }
1394
1395 /**
1396 * @param string $name
1397 * @param string $val
1398 */
1399 public function setHeader( $name, $val ) {
1400 $name = strtoupper( $name );
1401 $this->headers[$name] = $val;
1402 }
1403
1404 /**
1405 * @param string $key
1406 * @return array|null
1407 */
1408 public function getSessionData( $key ) {
1409 if ( isset( $this->session[$key] ) ) {
1410 return $this->session[$key];
1411 }
1412 return null;
1413 }
1414
1415 /**
1416 * @param string $key
1417 * @param array $data
1418 */
1419 public function setSessionData( $key, $data ) {
1420 $this->session[$key] = $data;
1421 }
1422
1423 /**
1424 * @return array|mixed|null
1425 */
1426 public function getSessionArray() {
1427 return $this->session;
1428 }
1429
1430 /**
1431 * FauxRequests shouldn't depend on raw request data (but that could be implemented here)
1432 * @return string
1433 */
1434 public function getRawQueryString() {
1435 return '';
1436 }
1437
1438 /**
1439 * FauxRequests shouldn't depend on raw request data (but that could be implemented here)
1440 * @return string
1441 */
1442 public function getRawPostString() {
1443 return '';
1444 }
1445
1446 /**
1447 * FauxRequests shouldn't depend on raw request data (but that could be implemented here)
1448 * @return string
1449 */
1450 public function getRawInput() {
1451 return '';
1452 }
1453
1454 /**
1455 * @param array $extWhitelist
1456 * @return bool
1457 */
1458 public function checkUrlExtension( $extWhitelist = array() ) {
1459 return true;
1460 }
1461
1462 /**
1463 * @return string
1464 */
1465 protected function getRawIP() {
1466 return '127.0.0.1';
1467 }
1468 }
1469
1470 /**
1471 * Similar to FauxRequest, but only fakes URL parameters and method
1472 * (POST or GET) and use the base request for the remaining stuff
1473 * (cookies, session and headers).
1474 *
1475 * @ingroup HTTP
1476 * @since 1.19
1477 */
1478 class DerivativeRequest extends FauxRequest {
1479 private $base;
1480
1481 /**
1482 * @param WebRequest $base
1483 * @param array $data Array of *non*-urlencoded key => value pairs, the
1484 * fake GET/POST values
1485 * @param bool $wasPosted Whether to treat the data as POST
1486 */
1487 public function __construct( WebRequest $base, $data, $wasPosted = false ) {
1488 $this->base = $base;
1489 parent::__construct( $data, $wasPosted );
1490 }
1491
1492 public function getCookie( $key, $prefix = null, $default = null ) {
1493 return $this->base->getCookie( $key, $prefix, $default );
1494 }
1495
1496 public function checkSessionCookie() {
1497 return $this->base->checkSessionCookie();
1498 }
1499
1500 public function getHeader( $name ) {
1501 return $this->base->getHeader( $name );
1502 }
1503
1504 public function getAllHeaders() {
1505 return $this->base->getAllHeaders();
1506 }
1507
1508 public function getSessionData( $key ) {
1509 return $this->base->getSessionData( $key );
1510 }
1511
1512 public function setSessionData( $key, $data ) {
1513 $this->base->setSessionData( $key, $data );
1514 }
1515
1516 public function getAcceptLang() {
1517 return $this->base->getAcceptLang();
1518 }
1519
1520 public function getIP() {
1521 return $this->base->getIP();
1522 }
1523
1524 public function getProtocol() {
1525 return $this->base->getProtocol();
1526 }
1527 }