(bug 34939) Handle mixed-case URL protocols in wikitext
[lhc/web/wiklou.git] / includes / GlobalFunctions.php
1 <?php
2 /**
3 * Global functions used everywhere.
4 *
5 * This program is free software; you can redistribute it and/or modify
6 * it under the terms of the GNU General Public License as published by
7 * the Free Software Foundation; either version 2 of the License, or
8 * (at your option) any later version.
9 *
10 * This program is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 * GNU General Public License for more details.
14 *
15 * You should have received a copy of the GNU General Public License along
16 * with this program; if not, write to the Free Software Foundation, Inc.,
17 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
18 * http://www.gnu.org/copyleft/gpl.html
19 *
20 * @file
21 */
22
23 if ( !defined( 'MEDIAWIKI' ) ) {
24 die( "This file is part of MediaWiki, it is not a valid entry point" );
25 }
26
27 // Hide compatibility functions from Doxygen
28 /// @cond
29
30 /**
31 * Compatibility functions
32 *
33 * We support PHP 5.3.2 and up.
34 * Re-implementations of newer functions or functions in non-standard
35 * PHP extensions may be included here.
36 */
37
38 if( !function_exists( 'iconv' ) ) {
39 /**
40 * @codeCoverageIgnore
41 * @return string
42 */
43 function iconv( $from, $to, $string ) {
44 return Fallback::iconv( $from, $to, $string );
45 }
46 }
47
48 if ( !function_exists( 'mb_substr' ) ) {
49 /**
50 * @codeCoverageIgnore
51 * @return string
52 */
53 function mb_substr( $str, $start, $count='end' ) {
54 return Fallback::mb_substr( $str, $start, $count );
55 }
56
57 /**
58 * @codeCoverageIgnore
59 * @return int
60 */
61 function mb_substr_split_unicode( $str, $splitPos ) {
62 return Fallback::mb_substr_split_unicode( $str, $splitPos );
63 }
64 }
65
66 if ( !function_exists( 'mb_strlen' ) ) {
67 /**
68 * @codeCoverageIgnore
69 * @return int
70 */
71 function mb_strlen( $str, $enc = '' ) {
72 return Fallback::mb_strlen( $str, $enc );
73 }
74 }
75
76 if( !function_exists( 'mb_strpos' ) ) {
77 /**
78 * @codeCoverageIgnore
79 * @return int
80 */
81 function mb_strpos( $haystack, $needle, $offset = 0, $encoding = '' ) {
82 return Fallback::mb_strpos( $haystack, $needle, $offset, $encoding );
83 }
84
85 }
86
87 if( !function_exists( 'mb_strrpos' ) ) {
88 /**
89 * @codeCoverageIgnore
90 * @return int
91 */
92 function mb_strrpos( $haystack, $needle, $offset = 0, $encoding = '' ) {
93 return Fallback::mb_strrpos( $haystack, $needle, $offset, $encoding );
94 }
95 }
96
97
98 // Support for Wietse Venema's taint feature
99 if ( !function_exists( 'istainted' ) ) {
100 /**
101 * @codeCoverageIgnore
102 * @return int
103 */
104 function istainted( $var ) {
105 return 0;
106 }
107 /** @codeCoverageIgnore */
108 function taint( $var, $level = 0 ) {}
109 /** @codeCoverageIgnore */
110 function untaint( $var, $level = 0 ) {}
111 define( 'TC_HTML', 1 );
112 define( 'TC_SHELL', 1 );
113 define( 'TC_MYSQL', 1 );
114 define( 'TC_PCRE', 1 );
115 define( 'TC_SELF', 1 );
116 }
117 /// @endcond
118
119 /**
120 * Like array_diff( $a, $b ) except that it works with two-dimensional arrays.
121 * @param $a array
122 * @param $b array
123 * @return array
124 */
125 function wfArrayDiff2( $a, $b ) {
126 return array_udiff( $a, $b, 'wfArrayDiff2_cmp' );
127 }
128
129 /**
130 * @param $a
131 * @param $b
132 * @return int
133 */
134 function wfArrayDiff2_cmp( $a, $b ) {
135 if ( !is_array( $a ) ) {
136 return strcmp( $a, $b );
137 } elseif ( count( $a ) !== count( $b ) ) {
138 return count( $a ) < count( $b ) ? -1 : 1;
139 } else {
140 reset( $a );
141 reset( $b );
142 while( ( list( , $valueA ) = each( $a ) ) && ( list( , $valueB ) = each( $b ) ) ) {
143 $cmp = strcmp( $valueA, $valueB );
144 if ( $cmp !== 0 ) {
145 return $cmp;
146 }
147 }
148 return 0;
149 }
150 }
151
152 /**
153 * Array lookup
154 * Returns an array where the values in the first array are replaced by the
155 * values in the second array with the corresponding keys
156 *
157 * @param $a Array
158 * @param $b Array
159 * @return array
160 */
161 function wfArrayLookup( $a, $b ) {
162 return array_flip( array_intersect( array_flip( $a ), array_keys( $b ) ) );
163 }
164
165 /**
166 * Appends to second array if $value differs from that in $default
167 *
168 * @param $key String|Int
169 * @param $value Mixed
170 * @param $default Mixed
171 * @param $changed Array to alter
172 */
173 function wfAppendToArrayIfNotDefault( $key, $value, $default, &$changed ) {
174 if ( is_null( $changed ) ) {
175 throw new MWException( 'GlobalFunctions::wfAppendToArrayIfNotDefault got null' );
176 }
177 if ( $default[$key] !== $value ) {
178 $changed[$key] = $value;
179 }
180 }
181
182 /**
183 * Backwards array plus for people who haven't bothered to read the PHP manual
184 * XXX: will not darn your socks for you.
185 *
186 * @param $array1 Array
187 * @param [$array2, [...]] Arrays
188 * @return Array
189 */
190 function wfArrayMerge( $array1/* ... */ ) {
191 $args = func_get_args();
192 $args = array_reverse( $args, true );
193 $out = array();
194 foreach ( $args as $arg ) {
195 $out += $arg;
196 }
197 return $out;
198 }
199
200 /**
201 * Merge arrays in the style of getUserPermissionsErrors, with duplicate removal
202 * e.g.
203 * wfMergeErrorArrays(
204 * array( array( 'x' ) ),
205 * array( array( 'x', '2' ) ),
206 * array( array( 'x' ) ),
207 * array( array( 'y' ) )
208 * );
209 * returns:
210 * array(
211 * array( 'x', '2' ),
212 * array( 'x' ),
213 * array( 'y' )
214 * )
215 * @param varargs
216 * @return Array
217 */
218 function wfMergeErrorArrays( /*...*/ ) {
219 $args = func_get_args();
220 $out = array();
221 foreach ( $args as $errors ) {
222 foreach ( $errors as $params ) {
223 # @todo FIXME: Sometimes get nested arrays for $params,
224 # which leads to E_NOTICEs
225 $spec = implode( "\t", $params );
226 $out[$spec] = $params;
227 }
228 }
229 return array_values( $out );
230 }
231
232 /**
233 * Insert array into another array after the specified *KEY*
234 *
235 * @param $array Array: The array.
236 * @param $insert Array: The array to insert.
237 * @param $after Mixed: The key to insert after
238 * @return Array
239 */
240 function wfArrayInsertAfter( array $array, array $insert, $after ) {
241 // Find the offset of the element to insert after.
242 $keys = array_keys( $array );
243 $offsetByKey = array_flip( $keys );
244
245 $offset = $offsetByKey[$after];
246
247 // Insert at the specified offset
248 $before = array_slice( $array, 0, $offset + 1, true );
249 $after = array_slice( $array, $offset + 1, count( $array ) - $offset, true );
250
251 $output = $before + $insert + $after;
252
253 return $output;
254 }
255
256 /**
257 * Recursively converts the parameter (an object) to an array with the same data
258 *
259 * @param $objOrArray Object|Array
260 * @param $recursive Bool
261 * @return Array
262 */
263 function wfObjectToArray( $objOrArray, $recursive = true ) {
264 $array = array();
265 if( is_object( $objOrArray ) ) {
266 $objOrArray = get_object_vars( $objOrArray );
267 }
268 foreach ( $objOrArray as $key => $value ) {
269 if ( $recursive && ( is_object( $value ) || is_array( $value ) ) ) {
270 $value = wfObjectToArray( $value );
271 }
272
273 $array[$key] = $value;
274 }
275
276 return $array;
277 }
278
279 /**
280 * Wrapper around array_map() which also taints variables
281 *
282 * @param $function Callback
283 * @param $input Array
284 * @return Array
285 */
286 function wfArrayMap( $function, $input ) {
287 $ret = array_map( $function, $input );
288 foreach ( $ret as $key => $value ) {
289 $taint = istainted( $input[$key] );
290 if ( $taint ) {
291 taint( $ret[$key], $taint );
292 }
293 }
294 return $ret;
295 }
296
297 /**
298 * Get a random decimal value between 0 and 1, in a way
299 * not likely to give duplicate values for any realistic
300 * number of articles.
301 *
302 * @return string
303 */
304 function wfRandom() {
305 # The maximum random value is "only" 2^31-1, so get two random
306 # values to reduce the chance of dupes
307 $max = mt_getrandmax() + 1;
308 $rand = number_format( ( mt_rand() * $max + mt_rand() )
309 / $max / $max, 12, '.', '' );
310 return $rand;
311 }
312
313 /**
314 * Get a random string containing a number of pesudo-random hex
315 * characters.
316 * @note This is not secure, if you are trying to generate some sort
317 * of token please use MWCryptRand instead.
318 *
319 * @param $length int The length of the string to generate
320 * @return String
321 * @since 1.20
322 */
323 function wfRandomString( $length = 32 ) {
324 $str = '';
325 while ( strlen( $str ) < $length ) {
326 $str .= dechex( mt_rand() );
327 }
328 return substr( $str, 0, $length );
329 }
330
331 /**
332 * We want some things to be included as literal characters in our title URLs
333 * for prettiness, which urlencode encodes by default. According to RFC 1738,
334 * all of the following should be safe:
335 *
336 * ;:@&=$-_.+!*'(),
337 *
338 * But + is not safe because it's used to indicate a space; &= are only safe in
339 * paths and not in queries (and we don't distinguish here); ' seems kind of
340 * scary; and urlencode() doesn't touch -_. to begin with. Plus, although /
341 * is reserved, we don't care. So the list we unescape is:
342 *
343 * ;:@$!*(),/
344 *
345 * However, IIS7 redirects fail when the url contains a colon (Bug 22709),
346 * so no fancy : for IIS7.
347 *
348 * %2F in the page titles seems to fatally break for some reason.
349 *
350 * @param $s String:
351 * @return string
352 */
353 function wfUrlencode( $s ) {
354 static $needle;
355 if ( is_null( $s ) ) {
356 $needle = null;
357 return '';
358 }
359
360 if ( is_null( $needle ) ) {
361 $needle = array( '%3B', '%40', '%24', '%21', '%2A', '%28', '%29', '%2C', '%2F' );
362 if ( !isset( $_SERVER['SERVER_SOFTWARE'] ) || ( strpos( $_SERVER['SERVER_SOFTWARE'], 'Microsoft-IIS/7' ) === false ) ) {
363 $needle[] = '%3A';
364 }
365 }
366
367 $s = urlencode( $s );
368 $s = str_ireplace(
369 $needle,
370 array( ';', '@', '$', '!', '*', '(', ')', ',', '/', ':' ),
371 $s
372 );
373
374 return $s;
375 }
376
377 /**
378 * This function takes two arrays as input, and returns a CGI-style string, e.g.
379 * "days=7&limit=100". Options in the first array override options in the second.
380 * Options set to null or false will not be output.
381 *
382 * @param $array1 Array ( String|Array )
383 * @param $array2 Array ( String|Array )
384 * @param $prefix String
385 * @return String
386 */
387 function wfArrayToCgi( $array1, $array2 = null, $prefix = '' ) {
388 if ( !is_null( $array2 ) ) {
389 $array1 = $array1 + $array2;
390 }
391
392 $cgi = '';
393 foreach ( $array1 as $key => $value ) {
394 if ( !is_null($value) && $value !== false ) {
395 if ( $cgi != '' ) {
396 $cgi .= '&';
397 }
398 if ( $prefix !== '' ) {
399 $key = $prefix . "[$key]";
400 }
401 if ( is_array( $value ) ) {
402 $firstTime = true;
403 foreach ( $value as $k => $v ) {
404 $cgi .= $firstTime ? '' : '&';
405 if ( is_array( $v ) ) {
406 $cgi .= wfArrayToCgi( $v, null, $key . "[$k]" );
407 } else {
408 $cgi .= urlencode( $key . "[$k]" ) . '=' . urlencode( $v );
409 }
410 $firstTime = false;
411 }
412 } else {
413 if ( is_object( $value ) ) {
414 $value = $value->__toString();
415 }
416 $cgi .= urlencode( $key ) . '=' . urlencode( $value );
417 }
418 }
419 }
420 return $cgi;
421 }
422
423 /**
424 * This is the logical opposite of wfArrayToCgi(): it accepts a query string as
425 * its argument and returns the same string in array form. This allows compa-
426 * tibility with legacy functions that accept raw query strings instead of nice
427 * arrays. Of course, keys and values are urldecode()d.
428 *
429 * @param $query String: query string
430 * @return array Array version of input
431 */
432 function wfCgiToArray( $query ) {
433 if ( isset( $query[0] ) && $query[0] == '?' ) {
434 $query = substr( $query, 1 );
435 }
436 $bits = explode( '&', $query );
437 $ret = array();
438 foreach ( $bits as $bit ) {
439 if ( $bit === '' ) {
440 continue;
441 }
442 if ( strpos( $bit, '=' ) === false ) {
443 // Pieces like &qwerty become 'qwerty' => '' (at least this is what php does)
444 $key = $bit;
445 $value = '';
446 } else {
447 list( $key, $value ) = explode( '=', $bit );
448 }
449 $key = urldecode( $key );
450 $value = urldecode( $value );
451 if ( strpos( $key, '[' ) !== false ) {
452 $keys = array_reverse( explode( '[', $key ) );
453 $key = array_pop( $keys );
454 $temp = $value;
455 foreach ( $keys as $k ) {
456 $k = substr( $k, 0, -1 );
457 $temp = array( $k => $temp );
458 }
459 if ( isset( $ret[$key] ) ) {
460 $ret[$key] = array_merge( $ret[$key], $temp );
461 } else {
462 $ret[$key] = $temp;
463 }
464 } else {
465 $ret[$key] = $value;
466 }
467 }
468 return $ret;
469 }
470
471 /**
472 * Append a query string to an existing URL, which may or may not already
473 * have query string parameters already. If so, they will be combined.
474 *
475 * @param $url String
476 * @param $query Mixed: string or associative array
477 * @return string
478 */
479 function wfAppendQuery( $url, $query ) {
480 if ( is_array( $query ) ) {
481 $query = wfArrayToCgi( $query );
482 }
483 if( $query != '' ) {
484 if( false === strpos( $url, '?' ) ) {
485 $url .= '?';
486 } else {
487 $url .= '&';
488 }
489 $url .= $query;
490 }
491 return $url;
492 }
493
494 /**
495 * Expand a potentially local URL to a fully-qualified URL. Assumes $wgServer
496 * is correct.
497 *
498 * The meaning of the PROTO_* constants is as follows:
499 * PROTO_HTTP: Output a URL starting with http://
500 * PROTO_HTTPS: Output a URL starting with https://
501 * PROTO_RELATIVE: Output a URL starting with // (protocol-relative URL)
502 * PROTO_CURRENT: Output a URL starting with either http:// or https:// , depending on which protocol was used for the current incoming request
503 * PROTO_CANONICAL: For URLs without a domain, like /w/index.php , use $wgCanonicalServer. For protocol-relative URLs, use the protocol of $wgCanonicalServer
504 * PROTO_INTERNAL: Like PROTO_CANONICAL, but uses $wgInternalServer instead of $wgCanonicalServer
505 *
506 * @todo this won't work with current-path-relative URLs
507 * like "subdir/foo.html", etc.
508 *
509 * @param $url String: either fully-qualified or a local path + query
510 * @param $defaultProto Mixed: one of the PROTO_* constants. Determines the
511 * protocol to use if $url or $wgServer is
512 * protocol-relative
513 * @return string Fully-qualified URL, current-path-relative URL or false if
514 * no valid URL can be constructed
515 */
516 function wfExpandUrl( $url, $defaultProto = PROTO_CURRENT ) {
517 global $wgServer, $wgCanonicalServer, $wgInternalServer;
518 $serverUrl = $wgServer;
519 if ( $defaultProto === PROTO_CANONICAL ) {
520 $serverUrl = $wgCanonicalServer;
521 }
522 // Make $wgInternalServer fall back to $wgServer if not set
523 if ( $defaultProto === PROTO_INTERNAL && $wgInternalServer !== false ) {
524 $serverUrl = $wgInternalServer;
525 }
526 if ( $defaultProto === PROTO_CURRENT ) {
527 $defaultProto = WebRequest::detectProtocol() . '://';
528 }
529
530 // Analyze $serverUrl to obtain its protocol
531 $bits = wfParseUrl( $serverUrl );
532 $serverHasProto = $bits && $bits['scheme'] != '';
533
534 if ( $defaultProto === PROTO_CANONICAL || $defaultProto === PROTO_INTERNAL ) {
535 if ( $serverHasProto ) {
536 $defaultProto = $bits['scheme'] . '://';
537 } else {
538 // $wgCanonicalServer or $wgInternalServer doesn't have a protocol. This really isn't supposed to happen
539 // Fall back to HTTP in this ridiculous case
540 $defaultProto = PROTO_HTTP;
541 }
542 }
543
544 $defaultProtoWithoutSlashes = substr( $defaultProto, 0, -2 );
545
546 if ( substr( $url, 0, 2 ) == '//' ) {
547 $url = $defaultProtoWithoutSlashes . $url;
548 } elseif ( substr( $url, 0, 1 ) == '/' ) {
549 // If $serverUrl is protocol-relative, prepend $defaultProtoWithoutSlashes, otherwise leave it alone
550 $url = ( $serverHasProto ? '' : $defaultProtoWithoutSlashes ) . $serverUrl . $url;
551 }
552
553 $bits = wfParseUrl( $url );
554 if ( $bits && isset( $bits['path'] ) ) {
555 $bits['path'] = wfRemoveDotSegments( $bits['path'] );
556 return wfAssembleUrl( $bits );
557 } elseif ( $bits ) {
558 # No path to expand
559 return $url;
560 } elseif ( substr( $url, 0, 1 ) != '/' ) {
561 # URL is a relative path
562 return wfRemoveDotSegments( $url );
563 }
564
565 # Expanded URL is not valid.
566 return false;
567 }
568
569 /**
570 * This function will reassemble a URL parsed with wfParseURL. This is useful
571 * if you need to edit part of a URL and put it back together.
572 *
573 * This is the basic structure used (brackets contain keys for $urlParts):
574 * [scheme][delimiter][user]:[pass]@[host]:[port][path]?[query]#[fragment]
575 *
576 * @todo Need to integrate this into wfExpandUrl (bug 32168)
577 *
578 * @since 1.19
579 * @param $urlParts Array URL parts, as output from wfParseUrl
580 * @return string URL assembled from its component parts
581 */
582 function wfAssembleUrl( $urlParts ) {
583 $result = '';
584
585 if ( isset( $urlParts['delimiter'] ) ) {
586 if ( isset( $urlParts['scheme'] ) ) {
587 $result .= $urlParts['scheme'];
588 }
589
590 $result .= $urlParts['delimiter'];
591 }
592
593 if ( isset( $urlParts['host'] ) ) {
594 if ( isset( $urlParts['user'] ) ) {
595 $result .= $urlParts['user'];
596 if ( isset( $urlParts['pass'] ) ) {
597 $result .= ':' . $urlParts['pass'];
598 }
599 $result .= '@';
600 }
601
602 $result .= $urlParts['host'];
603
604 if ( isset( $urlParts['port'] ) ) {
605 $result .= ':' . $urlParts['port'];
606 }
607 }
608
609 if ( isset( $urlParts['path'] ) ) {
610 $result .= $urlParts['path'];
611 }
612
613 if ( isset( $urlParts['query'] ) ) {
614 $result .= '?' . $urlParts['query'];
615 }
616
617 if ( isset( $urlParts['fragment'] ) ) {
618 $result .= '#' . $urlParts['fragment'];
619 }
620
621 return $result;
622 }
623
624 /**
625 * Remove all dot-segments in the provided URL path. For example,
626 * '/a/./b/../c/' becomes '/a/c/'. For details on the algorithm, please see
627 * RFC3986 section 5.2.4.
628 *
629 * @todo Need to integrate this into wfExpandUrl (bug 32168)
630 *
631 * @param $urlPath String URL path, potentially containing dot-segments
632 * @return string URL path with all dot-segments removed
633 */
634 function wfRemoveDotSegments( $urlPath ) {
635 $output = '';
636 $inputOffset = 0;
637 $inputLength = strlen( $urlPath );
638
639 while ( $inputOffset < $inputLength ) {
640 $prefixLengthOne = substr( $urlPath, $inputOffset, 1 );
641 $prefixLengthTwo = substr( $urlPath, $inputOffset, 2 );
642 $prefixLengthThree = substr( $urlPath, $inputOffset, 3 );
643 $prefixLengthFour = substr( $urlPath, $inputOffset, 4 );
644 $trimOutput = false;
645
646 if ( $prefixLengthTwo == './' ) {
647 # Step A, remove leading "./"
648 $inputOffset += 2;
649 } elseif ( $prefixLengthThree == '../' ) {
650 # Step A, remove leading "../"
651 $inputOffset += 3;
652 } elseif ( ( $prefixLengthTwo == '/.' ) && ( $inputOffset + 2 == $inputLength ) ) {
653 # Step B, replace leading "/.$" with "/"
654 $inputOffset += 1;
655 $urlPath[$inputOffset] = '/';
656 } elseif ( $prefixLengthThree == '/./' ) {
657 # Step B, replace leading "/./" with "/"
658 $inputOffset += 2;
659 } elseif ( $prefixLengthThree == '/..' && ( $inputOffset + 3 == $inputLength ) ) {
660 # Step C, replace leading "/..$" with "/" and
661 # remove last path component in output
662 $inputOffset += 2;
663 $urlPath[$inputOffset] = '/';
664 $trimOutput = true;
665 } elseif ( $prefixLengthFour == '/../' ) {
666 # Step C, replace leading "/../" with "/" and
667 # remove last path component in output
668 $inputOffset += 3;
669 $trimOutput = true;
670 } elseif ( ( $prefixLengthOne == '.' ) && ( $inputOffset + 1 == $inputLength ) ) {
671 # Step D, remove "^.$"
672 $inputOffset += 1;
673 } elseif ( ( $prefixLengthTwo == '..' ) && ( $inputOffset + 2 == $inputLength ) ) {
674 # Step D, remove "^..$"
675 $inputOffset += 2;
676 } else {
677 # Step E, move leading path segment to output
678 if ( $prefixLengthOne == '/' ) {
679 $slashPos = strpos( $urlPath, '/', $inputOffset + 1 );
680 } else {
681 $slashPos = strpos( $urlPath, '/', $inputOffset );
682 }
683 if ( $slashPos === false ) {
684 $output .= substr( $urlPath, $inputOffset );
685 $inputOffset = $inputLength;
686 } else {
687 $output .= substr( $urlPath, $inputOffset, $slashPos - $inputOffset );
688 $inputOffset += $slashPos - $inputOffset;
689 }
690 }
691
692 if ( $trimOutput ) {
693 $slashPos = strrpos( $output, '/' );
694 if ( $slashPos === false ) {
695 $output = '';
696 } else {
697 $output = substr( $output, 0, $slashPos );
698 }
699 }
700 }
701
702 return $output;
703 }
704
705 /**
706 * Returns a regular expression of url protocols
707 *
708 * @param $includeProtocolRelative bool If false, remove '//' from the returned protocol list.
709 * DO NOT USE this directly, use wfUrlProtocolsWithoutProtRel() instead
710 * @return String
711 */
712 function wfUrlProtocols( $includeProtocolRelative = true ) {
713 global $wgUrlProtocols;
714
715 // Cache return values separately based on $includeProtocolRelative
716 static $withProtRel = null, $withoutProtRel = null;
717 $cachedValue = $includeProtocolRelative ? $withProtRel : $withoutProtRel;
718 if ( !is_null( $cachedValue ) ) {
719 return $cachedValue;
720 }
721
722 // Support old-style $wgUrlProtocols strings, for backwards compatibility
723 // with LocalSettings files from 1.5
724 if ( is_array( $wgUrlProtocols ) ) {
725 $protocols = array();
726 foreach ( $wgUrlProtocols as $protocol ) {
727 // Filter out '//' if !$includeProtocolRelative
728 if ( $includeProtocolRelative || $protocol !== '//' ) {
729 $protocols[] = preg_quote( $protocol, '/' );
730 }
731 }
732
733 $retval = implode( '|', $protocols );
734 } else {
735 // Ignore $includeProtocolRelative in this case
736 // This case exists for pre-1.6 compatibility, and we can safely assume
737 // that '//' won't appear in a pre-1.6 config because protocol-relative
738 // URLs weren't supported until 1.18
739 $retval = $wgUrlProtocols;
740 }
741
742 // Cache return value
743 if ( $includeProtocolRelative ) {
744 $withProtRel = $retval;
745 } else {
746 $withoutProtRel = $retval;
747 }
748 return $retval;
749 }
750
751 /**
752 * Like wfUrlProtocols(), but excludes '//' from the protocol list. Use this if
753 * you need a regex that matches all URL protocols but does not match protocol-
754 * relative URLs
755 * @return String
756 */
757 function wfUrlProtocolsWithoutProtRel() {
758 return wfUrlProtocols( false );
759 }
760
761 /**
762 * parse_url() work-alike, but non-broken. Differences:
763 *
764 * 1) Does not raise warnings on bad URLs (just returns false)
765 * 2) Handles protocols that don't use :// (e.g., mailto: and news: , as well as protocol-relative URLs) correctly
766 * 3) Adds a "delimiter" element to the array, either '://', ':' or '//' (see (2))
767 *
768 * @param $url String: a URL to parse
769 * @return Array: bits of the URL in an associative array, per PHP docs
770 */
771 function wfParseUrl( $url ) {
772 global $wgUrlProtocols; // Allow all protocols defined in DefaultSettings/LocalSettings.php
773
774 // Protocol-relative URLs are handled really badly by parse_url(). It's so bad that the easiest
775 // way to handle them is to just prepend 'http:' and strip the protocol out later
776 $wasRelative = substr( $url, 0, 2 ) == '//';
777 if ( $wasRelative ) {
778 $url = "http:$url";
779 }
780 wfSuppressWarnings();
781 $bits = parse_url( $url );
782 wfRestoreWarnings();
783 // parse_url() returns an array without scheme for some invalid URLs, e.g.
784 // parse_url("%0Ahttp://example.com") == array( 'host' => '%0Ahttp', 'path' => 'example.com' )
785 if ( !$bits || !isset( $bits['scheme'] ) ) {
786 return false;
787 }
788
789 // parse_url() incorrectly handles schemes case-sensitively. Convert it to lowercase.
790 $bits['scheme'] = strtolower( $bits['scheme'] );
791
792 // most of the protocols are followed by ://, but mailto: and sometimes news: not, check for it
793 if ( in_array( $bits['scheme'] . '://', $wgUrlProtocols ) ) {
794 $bits['delimiter'] = '://';
795 } elseif ( in_array( $bits['scheme'] . ':', $wgUrlProtocols ) ) {
796 $bits['delimiter'] = ':';
797 // parse_url detects for news: and mailto: the host part of an url as path
798 // We have to correct this wrong detection
799 if ( isset( $bits['path'] ) ) {
800 $bits['host'] = $bits['path'];
801 $bits['path'] = '';
802 }
803 } else {
804 return false;
805 }
806
807 /* Provide an empty host for eg. file:/// urls (see bug 28627) */
808 if ( !isset( $bits['host'] ) ) {
809 $bits['host'] = '';
810
811 /* parse_url loses the third / for file:///c:/ urls (but not on variants) */
812 if ( substr( $bits['path'], 0, 1 ) !== '/' ) {
813 $bits['path'] = '/' . $bits['path'];
814 }
815 }
816
817 // If the URL was protocol-relative, fix scheme and delimiter
818 if ( $wasRelative ) {
819 $bits['scheme'] = '';
820 $bits['delimiter'] = '//';
821 }
822 return $bits;
823 }
824
825 /**
826 * Take a URL, make sure it's expanded to fully qualified, and replace any
827 * encoded non-ASCII Unicode characters with their UTF-8 original forms
828 * for more compact display and legibility for local audiences.
829 *
830 * @todo handle punycode domains too
831 *
832 * @param $url string
833 * @return string
834 */
835 function wfExpandIRI( $url ) {
836 return preg_replace_callback( '/((?:%[89A-F][0-9A-F])+)/i', 'wfExpandIRI_callback', wfExpandUrl( $url ) );
837 }
838
839 /**
840 * Private callback for wfExpandIRI
841 * @param array $matches
842 * @return string
843 */
844 function wfExpandIRI_callback( $matches ) {
845 return urldecode( $matches[1] );
846 }
847
848
849
850 /**
851 * Make URL indexes, appropriate for the el_index field of externallinks.
852 *
853 * @param $url String
854 * @return array
855 */
856 function wfMakeUrlIndexes( $url ) {
857 $bits = wfParseUrl( $url );
858
859 // Reverse the labels in the hostname, convert to lower case
860 // For emails reverse domainpart only
861 if ( $bits['scheme'] == 'mailto' ) {
862 $mailparts = explode( '@', $bits['host'], 2 );
863 if ( count( $mailparts ) === 2 ) {
864 $domainpart = strtolower( implode( '.', array_reverse( explode( '.', $mailparts[1] ) ) ) );
865 } else {
866 // No domain specified, don't mangle it
867 $domainpart = '';
868 }
869 $reversedHost = $domainpart . '@' . $mailparts[0];
870 } else {
871 $reversedHost = strtolower( implode( '.', array_reverse( explode( '.', $bits['host'] ) ) ) );
872 }
873 // Add an extra dot to the end
874 // Why? Is it in wrong place in mailto links?
875 if ( substr( $reversedHost, -1, 1 ) !== '.' ) {
876 $reversedHost .= '.';
877 }
878 // Reconstruct the pseudo-URL
879 $prot = $bits['scheme'];
880 $index = $prot . $bits['delimiter'] . $reversedHost;
881 // Leave out user and password. Add the port, path, query and fragment
882 if ( isset( $bits['port'] ) ) {
883 $index .= ':' . $bits['port'];
884 }
885 if ( isset( $bits['path'] ) ) {
886 $index .= $bits['path'];
887 } else {
888 $index .= '/';
889 }
890 if ( isset( $bits['query'] ) ) {
891 $index .= '?' . $bits['query'];
892 }
893 if ( isset( $bits['fragment'] ) ) {
894 $index .= '#' . $bits['fragment'];
895 }
896
897 if ( $prot == '' ) {
898 return array( "http:$index", "https:$index" );
899 } else {
900 return array( $index );
901 }
902 }
903
904 /**
905 * Check whether a given URL has a domain that occurs in a given set of domains
906 * @param $url string URL
907 * @param $domains array Array of domains (strings)
908 * @return bool True if the host part of $url ends in one of the strings in $domains
909 */
910 function wfMatchesDomainList( $url, $domains ) {
911 $bits = wfParseUrl( $url );
912 if ( is_array( $bits ) && isset( $bits['host'] ) ) {
913 foreach ( (array)$domains as $domain ) {
914 // FIXME: This gives false positives. http://nds-nl.wikipedia.org will match nl.wikipedia.org
915 // We should use something that interprets dots instead
916 if ( substr( $bits['host'], -strlen( $domain ) ) === $domain ) {
917 return true;
918 }
919 }
920 }
921 return false;
922 }
923
924 /**
925 * Sends a line to the debug log if enabled or, optionally, to a comment in output.
926 * In normal operation this is a NOP.
927 *
928 * Controlling globals:
929 * $wgDebugLogFile - points to the log file
930 * $wgProfileOnly - if set, normal debug messages will not be recorded.
931 * $wgDebugRawPage - if false, 'action=raw' hits will not result in debug output.
932 * $wgDebugComments - if on, some debug items may appear in comments in the HTML output.
933 *
934 * @param $text String
935 * @param $logonly Bool: set true to avoid appearing in HTML when $wgDebugComments is set
936 */
937 function wfDebug( $text, $logonly = false ) {
938 global $wgDebugLogFile, $wgProfileOnly, $wgDebugRawPage, $wgDebugLogPrefix;
939
940 if ( !$wgDebugRawPage && wfIsDebugRawPage() ) {
941 return;
942 }
943
944 $timer = wfDebugTimer();
945 if ( $timer !== '' ) {
946 $text = preg_replace( '/[^\n]/', $timer . '\0', $text, 1 );
947 }
948
949 if ( !$logonly ) {
950 MWDebug::debugMsg( $text );
951 }
952
953 if ( wfRunHooks( 'Debug', array( $text, null /* no log group */ ) ) ) {
954 if ( $wgDebugLogFile != '' && !$wgProfileOnly ) {
955 # Strip unprintables; they can switch terminal modes when binary data
956 # gets dumped, which is pretty annoying.
957 $text = preg_replace( '![\x00-\x08\x0b\x0c\x0e-\x1f]!', ' ', $text );
958 $text = $wgDebugLogPrefix . $text;
959 wfErrorLog( $text, $wgDebugLogFile );
960 }
961 }
962 }
963
964 /**
965 * Returns true if debug logging should be suppressed if $wgDebugRawPage = false
966 * @return bool
967 */
968 function wfIsDebugRawPage() {
969 static $cache;
970 if ( $cache !== null ) {
971 return $cache;
972 }
973 # Check for raw action using $_GET not $wgRequest, since the latter might not be initialised yet
974 if ( ( isset( $_GET['action'] ) && $_GET['action'] == 'raw' )
975 || (
976 isset( $_SERVER['SCRIPT_NAME'] )
977 && substr( $_SERVER['SCRIPT_NAME'], -8 ) == 'load.php'
978 ) )
979 {
980 $cache = true;
981 } else {
982 $cache = false;
983 }
984 return $cache;
985 }
986
987 /**
988 * Get microsecond timestamps for debug logs
989 *
990 * @return string
991 */
992 function wfDebugTimer() {
993 global $wgDebugTimestamps, $wgRequestTime;
994
995 if ( !$wgDebugTimestamps ) {
996 return '';
997 }
998
999 $prefix = sprintf( "%6.4f", microtime( true ) - $wgRequestTime );
1000 $mem = sprintf( "%5.1fM", ( memory_get_usage( true ) / ( 1024 * 1024 ) ) );
1001 return "$prefix $mem ";
1002 }
1003
1004 /**
1005 * Send a line giving PHP memory usage.
1006 *
1007 * @param $exact Bool: print exact values instead of kilobytes (default: false)
1008 */
1009 function wfDebugMem( $exact = false ) {
1010 $mem = memory_get_usage();
1011 if( !$exact ) {
1012 $mem = floor( $mem / 1024 ) . ' kilobytes';
1013 } else {
1014 $mem .= ' bytes';
1015 }
1016 wfDebug( "Memory usage: $mem\n" );
1017 }
1018
1019 /**
1020 * Send a line to a supplementary debug log file, if configured, or main debug log if not.
1021 * $wgDebugLogGroups[$logGroup] should be set to a filename to send to a separate log.
1022 *
1023 * @param $logGroup String
1024 * @param $text String
1025 * @param $public Bool: whether to log the event in the public log if no private
1026 * log file is specified, (default true)
1027 */
1028 function wfDebugLog( $logGroup, $text, $public = true ) {
1029 global $wgDebugLogGroups;
1030 $text = trim( $text ) . "\n";
1031 if( isset( $wgDebugLogGroups[$logGroup] ) ) {
1032 $time = wfTimestamp( TS_DB );
1033 $wiki = wfWikiID();
1034 $host = wfHostname();
1035 if ( wfRunHooks( 'Debug', array( $text, $logGroup ) ) ) {
1036 wfErrorLog( "$time $host $wiki: $text", $wgDebugLogGroups[$logGroup] );
1037 }
1038 } elseif ( $public === true ) {
1039 wfDebug( $text, true );
1040 }
1041 }
1042
1043 /**
1044 * Log for database errors
1045 *
1046 * @param $text String: database error message.
1047 */
1048 function wfLogDBError( $text ) {
1049 global $wgDBerrorLog, $wgDBerrorLogTZ;
1050 static $logDBErrorTimeZoneObject = null;
1051
1052 if ( $wgDBerrorLog ) {
1053 $host = wfHostname();
1054 $wiki = wfWikiID();
1055
1056 if ( $wgDBerrorLogTZ && !$logDBErrorTimeZoneObject ) {
1057 $logDBErrorTimeZoneObject = new DateTimeZone( $wgDBerrorLogTZ );
1058 }
1059
1060 // Workaround for https://bugs.php.net/bug.php?id=52063
1061 // Can be removed when min PHP > 5.3.2
1062 if ( $logDBErrorTimeZoneObject === null ) {
1063 $d = date_create( "now" );
1064 } else {
1065 $d = date_create( "now", $logDBErrorTimeZoneObject );
1066 }
1067
1068 $date = $d->format( 'D M j G:i:s T Y' );
1069
1070 $text = "$date\t$host\t$wiki\t$text";
1071 wfErrorLog( $text, $wgDBerrorLog );
1072 }
1073 }
1074
1075 /**
1076 * Throws a warning that $function is deprecated
1077 *
1078 * @param $function String
1079 * @param $version String|bool: Version of MediaWiki that the function was deprecated in (Added in 1.19).
1080 * @param $component String|bool: Added in 1.19.
1081 * @param $callerOffset integer: How far up the callstack is the original
1082 * caller. 2 = function that called the function that called
1083 * wfDeprecated (Added in 1.20)
1084 *
1085 * @return null
1086 */
1087 function wfDeprecated( $function, $version = false, $component = false, $callerOffset = 2 ) {
1088 MWDebug::deprecated( $function, $version, $component, $callerOffset + 1 );
1089 }
1090
1091 /**
1092 * Send a warning either to the debug log or in a PHP error depending on
1093 * $wgDevelopmentWarnings
1094 *
1095 * @param $msg String: message to send
1096 * @param $callerOffset Integer: number of items to go back in the backtrace to
1097 * find the correct caller (1 = function calling wfWarn, ...)
1098 * @param $level Integer: PHP error level; only used when $wgDevelopmentWarnings
1099 * is true
1100 */
1101 function wfWarn( $msg, $callerOffset = 1, $level = E_USER_NOTICE ) {
1102 MWDebug::warning( $msg, $callerOffset + 1, $level );
1103 }
1104
1105 /**
1106 * Log to a file without getting "file size exceeded" signals.
1107 *
1108 * Can also log to TCP or UDP with the syntax udp://host:port/prefix. This will
1109 * send lines to the specified port, prefixed by the specified prefix and a space.
1110 *
1111 * @param $text String
1112 * @param $file String filename
1113 */
1114 function wfErrorLog( $text, $file ) {
1115 if ( substr( $file, 0, 4 ) == 'udp:' ) {
1116 # Needs the sockets extension
1117 if ( preg_match( '!^(tcp|udp):(?://)?\[([0-9a-fA-F:]+)\]:(\d+)(?:/(.*))?$!', $file, $m ) ) {
1118 // IPv6 bracketed host
1119 $host = $m[2];
1120 $port = intval( $m[3] );
1121 $prefix = isset( $m[4] ) ? $m[4] : false;
1122 $domain = AF_INET6;
1123 } elseif ( preg_match( '!^(tcp|udp):(?://)?([a-zA-Z0-9.-]+):(\d+)(?:/(.*))?$!', $file, $m ) ) {
1124 $host = $m[2];
1125 if ( !IP::isIPv4( $host ) ) {
1126 $host = gethostbyname( $host );
1127 }
1128 $port = intval( $m[3] );
1129 $prefix = isset( $m[4] ) ? $m[4] : false;
1130 $domain = AF_INET;
1131 } else {
1132 throw new MWException( __METHOD__ . ': Invalid UDP specification' );
1133 }
1134
1135 // Clean it up for the multiplexer
1136 if ( strval( $prefix ) !== '' ) {
1137 $text = preg_replace( '/^/m', $prefix . ' ', $text );
1138
1139 // Limit to 64KB
1140 if ( strlen( $text ) > 65506 ) {
1141 $text = substr( $text, 0, 65506 );
1142 }
1143
1144 if ( substr( $text, -1 ) != "\n" ) {
1145 $text .= "\n";
1146 }
1147 } elseif ( strlen( $text ) > 65507 ) {
1148 $text = substr( $text, 0, 65507 );
1149 }
1150
1151 $sock = socket_create( $domain, SOCK_DGRAM, SOL_UDP );
1152 if ( !$sock ) {
1153 return;
1154 }
1155
1156 socket_sendto( $sock, $text, strlen( $text ), 0, $host, $port );
1157 socket_close( $sock );
1158 } else {
1159 wfSuppressWarnings();
1160 $exists = file_exists( $file );
1161 $size = $exists ? filesize( $file ) : false;
1162 if ( !$exists || ( $size !== false && $size + strlen( $text ) < 0x7fffffff ) ) {
1163 file_put_contents( $file, $text, FILE_APPEND );
1164 }
1165 wfRestoreWarnings();
1166 }
1167 }
1168
1169 /**
1170 * @todo document
1171 */
1172 function wfLogProfilingData() {
1173 global $wgRequestTime, $wgDebugLogFile, $wgDebugRawPage, $wgRequest;
1174 global $wgProfileLimit, $wgUser;
1175
1176 $profiler = Profiler::instance();
1177
1178 # Profiling must actually be enabled...
1179 if ( $profiler->isStub() ) {
1180 return;
1181 }
1182
1183 // Get total page request time and only show pages that longer than
1184 // $wgProfileLimit time (default is 0)
1185 $elapsed = microtime( true ) - $wgRequestTime;
1186 if ( $elapsed <= $wgProfileLimit ) {
1187 return;
1188 }
1189
1190 $profiler->logData();
1191
1192 // Check whether this should be logged in the debug file.
1193 if ( $wgDebugLogFile == '' || ( !$wgDebugRawPage && wfIsDebugRawPage() ) ) {
1194 return;
1195 }
1196
1197 $forward = '';
1198 if ( !empty( $_SERVER['HTTP_X_FORWARDED_FOR'] ) ) {
1199 $forward = ' forwarded for ' . $_SERVER['HTTP_X_FORWARDED_FOR'];
1200 }
1201 if ( !empty( $_SERVER['HTTP_CLIENT_IP'] ) ) {
1202 $forward .= ' client IP ' . $_SERVER['HTTP_CLIENT_IP'];
1203 }
1204 if ( !empty( $_SERVER['HTTP_FROM'] ) ) {
1205 $forward .= ' from ' . $_SERVER['HTTP_FROM'];
1206 }
1207 if ( $forward ) {
1208 $forward = "\t(proxied via {$_SERVER['REMOTE_ADDR']}{$forward})";
1209 }
1210 // Don't load $wgUser at this late stage just for statistics purposes
1211 // @todo FIXME: We can detect some anons even if it is not loaded. See User::getId()
1212 if ( $wgUser->isItemLoaded( 'id' ) && $wgUser->isAnon() ) {
1213 $forward .= ' anon';
1214 }
1215 $log = sprintf( "%s\t%04.3f\t%s\n",
1216 gmdate( 'YmdHis' ), $elapsed,
1217 urldecode( $wgRequest->getRequestURL() . $forward ) );
1218
1219 wfErrorLog( $log . $profiler->getOutput(), $wgDebugLogFile );
1220 }
1221
1222 /**
1223 * Increment a statistics counter
1224 *
1225 * @param $key String
1226 * @param $count Int
1227 */
1228 function wfIncrStats( $key, $count = 1 ) {
1229 global $wgStatsMethod;
1230
1231 $count = intval( $count );
1232
1233 if( $wgStatsMethod == 'udp' ) {
1234 global $wgUDPProfilerHost, $wgUDPProfilerPort, $wgDBname, $wgAggregateStatsID;
1235 static $socket;
1236
1237 $id = $wgAggregateStatsID !== false ? $wgAggregateStatsID : $wgDBname;
1238
1239 if ( !$socket ) {
1240 $socket = socket_create( AF_INET, SOCK_DGRAM, SOL_UDP );
1241 $statline = "stats/{$id} - 1 1 1 1 1 -total\n";
1242 socket_sendto(
1243 $socket,
1244 $statline,
1245 strlen( $statline ),
1246 0,
1247 $wgUDPProfilerHost,
1248 $wgUDPProfilerPort
1249 );
1250 }
1251 $statline = "stats/{$id} - {$count} 1 1 1 1 {$key}\n";
1252 wfSuppressWarnings();
1253 socket_sendto(
1254 $socket,
1255 $statline,
1256 strlen( $statline ),
1257 0,
1258 $wgUDPProfilerHost,
1259 $wgUDPProfilerPort
1260 );
1261 wfRestoreWarnings();
1262 } elseif( $wgStatsMethod == 'cache' ) {
1263 global $wgMemc;
1264 $key = wfMemcKey( 'stats', $key );
1265 if ( is_null( $wgMemc->incr( $key, $count ) ) ) {
1266 $wgMemc->add( $key, $count );
1267 }
1268 } else {
1269 // Disabled
1270 }
1271 }
1272
1273 /**
1274 * Check if the wiki read-only lock file is present. This can be used to lock
1275 * off editing functions, but doesn't guarantee that the database will not be
1276 * modified.
1277 *
1278 * @return bool
1279 */
1280 function wfReadOnly() {
1281 global $wgReadOnlyFile, $wgReadOnly;
1282
1283 if ( !is_null( $wgReadOnly ) ) {
1284 return (bool)$wgReadOnly;
1285 }
1286 if ( $wgReadOnlyFile == '' ) {
1287 return false;
1288 }
1289 // Set $wgReadOnly for faster access next time
1290 if ( is_file( $wgReadOnlyFile ) ) {
1291 $wgReadOnly = file_get_contents( $wgReadOnlyFile );
1292 } else {
1293 $wgReadOnly = false;
1294 }
1295 return (bool)$wgReadOnly;
1296 }
1297
1298 /**
1299 * @return bool
1300 */
1301 function wfReadOnlyReason() {
1302 global $wgReadOnly;
1303 wfReadOnly();
1304 return $wgReadOnly;
1305 }
1306
1307 /**
1308 * Return a Language object from $langcode
1309 *
1310 * @param $langcode Mixed: either:
1311 * - a Language object
1312 * - code of the language to get the message for, if it is
1313 * a valid code create a language for that language, if
1314 * it is a string but not a valid code then make a basic
1315 * language object
1316 * - a boolean: if it's false then use the global object for
1317 * the current user's language (as a fallback for the old parameter
1318 * functionality), or if it is true then use global object
1319 * for the wiki's content language.
1320 * @return Language object
1321 */
1322 function wfGetLangObj( $langcode = false ) {
1323 # Identify which language to get or create a language object for.
1324 # Using is_object here due to Stub objects.
1325 if( is_object( $langcode ) ) {
1326 # Great, we already have the object (hopefully)!
1327 return $langcode;
1328 }
1329
1330 global $wgContLang, $wgLanguageCode;
1331 if( $langcode === true || $langcode === $wgLanguageCode ) {
1332 # $langcode is the language code of the wikis content language object.
1333 # or it is a boolean and value is true
1334 return $wgContLang;
1335 }
1336
1337 global $wgLang;
1338 if( $langcode === false || $langcode === $wgLang->getCode() ) {
1339 # $langcode is the language code of user language object.
1340 # or it was a boolean and value is false
1341 return $wgLang;
1342 }
1343
1344 $validCodes = array_keys( Language::fetchLanguageNames() );
1345 if( in_array( $langcode, $validCodes ) ) {
1346 # $langcode corresponds to a valid language.
1347 return Language::factory( $langcode );
1348 }
1349
1350 # $langcode is a string, but not a valid language code; use content language.
1351 wfDebug( "Invalid language code passed to wfGetLangObj, falling back to content language.\n" );
1352 return $wgContLang;
1353 }
1354
1355 /**
1356 * Old function when $wgBetterDirectionality existed
1357 * All usage removed, wfUILang can be removed in near future
1358 *
1359 * @deprecated since 1.18
1360 * @return Language
1361 */
1362 function wfUILang() {
1363 wfDeprecated( __METHOD__, '1.18' );
1364 global $wgLang;
1365 return $wgLang;
1366 }
1367
1368 /**
1369 * This is the new function for getting translated interface messages.
1370 * See the Message class for documentation how to use them.
1371 * The intention is that this function replaces all old wfMsg* functions.
1372 * @param $key \string Message key.
1373 * Varargs: normal message parameters.
1374 * @return Message
1375 * @since 1.17
1376 */
1377 function wfMessage( $key /*...*/) {
1378 $params = func_get_args();
1379 array_shift( $params );
1380 if ( isset( $params[0] ) && is_array( $params[0] ) ) {
1381 $params = $params[0];
1382 }
1383 return new Message( $key, $params );
1384 }
1385
1386 /**
1387 * This function accepts multiple message keys and returns a message instance
1388 * for the first message which is non-empty. If all messages are empty then an
1389 * instance of the first message key is returned.
1390 * @param varargs: message keys
1391 * @return Message
1392 * @since 1.18
1393 */
1394 function wfMessageFallback( /*...*/ ) {
1395 $args = func_get_args();
1396 return MWFunction::callArray( 'Message::newFallbackSequence', $args );
1397 }
1398
1399 /**
1400 * Get a message from anywhere, for the current user language.
1401 *
1402 * Use wfMsgForContent() instead if the message should NOT
1403 * change depending on the user preferences.
1404 *
1405 * @deprecated since 1.18
1406 *
1407 * @param $key String: lookup key for the message, usually
1408 * defined in languages/Language.php
1409 *
1410 * Parameters to the message, which can be used to insert variable text into
1411 * it, can be passed to this function in the following formats:
1412 * - One per argument, starting at the second parameter
1413 * - As an array in the second parameter
1414 * These are not shown in the function definition.
1415 *
1416 * @return String
1417 */
1418 function wfMsg( $key ) {
1419 $args = func_get_args();
1420 array_shift( $args );
1421 return wfMsgReal( $key, $args );
1422 }
1423
1424 /**
1425 * Same as above except doesn't transform the message
1426 *
1427 * @deprecated since 1.18
1428 *
1429 * @param $key String
1430 * @return String
1431 */
1432 function wfMsgNoTrans( $key ) {
1433 $args = func_get_args();
1434 array_shift( $args );
1435 return wfMsgReal( $key, $args, true, false, false );
1436 }
1437
1438 /**
1439 * Get a message from anywhere, for the current global language
1440 * set with $wgLanguageCode.
1441 *
1442 * Use this if the message should NOT change dependent on the
1443 * language set in the user's preferences. This is the case for
1444 * most text written into logs, as well as link targets (such as
1445 * the name of the copyright policy page). Link titles, on the
1446 * other hand, should be shown in the UI language.
1447 *
1448 * Note that MediaWiki allows users to change the user interface
1449 * language in their preferences, but a single installation
1450 * typically only contains content in one language.
1451 *
1452 * Be wary of this distinction: If you use wfMsg() where you should
1453 * use wfMsgForContent(), a user of the software may have to
1454 * customize potentially hundreds of messages in
1455 * order to, e.g., fix a link in every possible language.
1456 *
1457 * @deprecated since 1.18
1458 *
1459 * @param $key String: lookup key for the message, usually
1460 * defined in languages/Language.php
1461 * @return String
1462 */
1463 function wfMsgForContent( $key ) {
1464 global $wgForceUIMsgAsContentMsg;
1465 $args = func_get_args();
1466 array_shift( $args );
1467 $forcontent = true;
1468 if( is_array( $wgForceUIMsgAsContentMsg ) &&
1469 in_array( $key, $wgForceUIMsgAsContentMsg ) )
1470 {
1471 $forcontent = false;
1472 }
1473 return wfMsgReal( $key, $args, true, $forcontent );
1474 }
1475
1476 /**
1477 * Same as above except doesn't transform the message
1478 *
1479 * @deprecated since 1.18
1480 *
1481 * @param $key String
1482 * @return String
1483 */
1484 function wfMsgForContentNoTrans( $key ) {
1485 global $wgForceUIMsgAsContentMsg;
1486 $args = func_get_args();
1487 array_shift( $args );
1488 $forcontent = true;
1489 if( is_array( $wgForceUIMsgAsContentMsg ) &&
1490 in_array( $key, $wgForceUIMsgAsContentMsg ) )
1491 {
1492 $forcontent = false;
1493 }
1494 return wfMsgReal( $key, $args, true, $forcontent, false );
1495 }
1496
1497 /**
1498 * Really get a message
1499 *
1500 * @deprecated since 1.18
1501 *
1502 * @param $key String: key to get.
1503 * @param $args
1504 * @param $useDB Boolean
1505 * @param $forContent Mixed: Language code, or false for user lang, true for content lang.
1506 * @param $transform Boolean: Whether or not to transform the message.
1507 * @return String: the requested message.
1508 */
1509 function wfMsgReal( $key, $args, $useDB = true, $forContent = false, $transform = true ) {
1510 wfProfileIn( __METHOD__ );
1511 $message = wfMsgGetKey( $key, $useDB, $forContent, $transform );
1512 $message = wfMsgReplaceArgs( $message, $args );
1513 wfProfileOut( __METHOD__ );
1514 return $message;
1515 }
1516
1517 /**
1518 * Fetch a message string value, but don't replace any keys yet.
1519 *
1520 * @deprecated since 1.18
1521 *
1522 * @param $key String
1523 * @param $useDB Bool
1524 * @param $langCode String: Code of the language to get the message for, or
1525 * behaves as a content language switch if it is a boolean.
1526 * @param $transform Boolean: whether to parse magic words, etc.
1527 * @return string
1528 */
1529 function wfMsgGetKey( $key, $useDB = true, $langCode = false, $transform = true ) {
1530 wfRunHooks( 'NormalizeMessageKey', array( &$key, &$useDB, &$langCode, &$transform ) );
1531
1532 $cache = MessageCache::singleton();
1533 $message = $cache->get( $key, $useDB, $langCode );
1534 if( $message === false ) {
1535 $message = '&lt;' . htmlspecialchars( $key ) . '&gt;';
1536 } elseif ( $transform ) {
1537 $message = $cache->transform( $message );
1538 }
1539 return $message;
1540 }
1541
1542 /**
1543 * Replace message parameter keys on the given formatted output.
1544 *
1545 * @deprecated since 1.18
1546 *
1547 * @param $message String
1548 * @param $args Array
1549 * @return string
1550 * @private
1551 */
1552 function wfMsgReplaceArgs( $message, $args ) {
1553 # Fix windows line-endings
1554 # Some messages are split with explode("\n", $msg)
1555 $message = str_replace( "\r", '', $message );
1556
1557 // Replace arguments
1558 if ( count( $args ) ) {
1559 if ( is_array( $args[0] ) ) {
1560 $args = array_values( $args[0] );
1561 }
1562 $replacementKeys = array();
1563 foreach( $args as $n => $param ) {
1564 $replacementKeys['$' . ( $n + 1 )] = $param;
1565 }
1566 $message = strtr( $message, $replacementKeys );
1567 }
1568
1569 return $message;
1570 }
1571
1572 /**
1573 * Return an HTML-escaped version of a message.
1574 * Parameter replacements, if any, are done *after* the HTML-escaping,
1575 * so parameters may contain HTML (eg links or form controls). Be sure
1576 * to pre-escape them if you really do want plaintext, or just wrap
1577 * the whole thing in htmlspecialchars().
1578 *
1579 * @deprecated since 1.18
1580 *
1581 * @param $key String
1582 * @param string ... parameters
1583 * @return string
1584 */
1585 function wfMsgHtml( $key ) {
1586 $args = func_get_args();
1587 array_shift( $args );
1588 return wfMsgReplaceArgs( htmlspecialchars( wfMsgGetKey( $key ) ), $args );
1589 }
1590
1591 /**
1592 * Return an HTML version of message
1593 * Parameter replacements, if any, are done *after* parsing the wiki-text message,
1594 * so parameters may contain HTML (eg links or form controls). Be sure
1595 * to pre-escape them if you really do want plaintext, or just wrap
1596 * the whole thing in htmlspecialchars().
1597 *
1598 * @deprecated since 1.18
1599 *
1600 * @param $key String
1601 * @param string ... parameters
1602 * @return string
1603 */
1604 function wfMsgWikiHtml( $key ) {
1605 $args = func_get_args();
1606 array_shift( $args );
1607 return wfMsgReplaceArgs(
1608 MessageCache::singleton()->parse( wfMsgGetKey( $key ), null,
1609 /* can't be set to false */ true, /* interface */ true )->getText(),
1610 $args );
1611 }
1612
1613 /**
1614 * Returns message in the requested format
1615 *
1616 * @deprecated since 1.18
1617 *
1618 * @param $key String: key of the message
1619 * @param $options Array: processing rules. Can take the following options:
1620 * <i>parse</i>: parses wikitext to HTML
1621 * <i>parseinline</i>: parses wikitext to HTML and removes the surrounding
1622 * p's added by parser or tidy
1623 * <i>escape</i>: filters message through htmlspecialchars
1624 * <i>escapenoentities</i>: same, but allows entity references like &#160; through
1625 * <i>replaceafter</i>: parameters are substituted after parsing or escaping
1626 * <i>parsemag</i>: transform the message using magic phrases
1627 * <i>content</i>: fetch message for content language instead of interface
1628 * Also can accept a single associative argument, of the form 'language' => 'xx':
1629 * <i>language</i>: Language object or language code to fetch message for
1630 * (overriden by <i>content</i>).
1631 * Behavior for conflicting options (e.g., parse+parseinline) is undefined.
1632 *
1633 * @return String
1634 */
1635 function wfMsgExt( $key, $options ) {
1636 $args = func_get_args();
1637 array_shift( $args );
1638 array_shift( $args );
1639 $options = (array)$options;
1640
1641 foreach( $options as $arrayKey => $option ) {
1642 if( !preg_match( '/^[0-9]+|language$/', $arrayKey ) ) {
1643 # An unknown index, neither numeric nor "language"
1644 wfWarn( "wfMsgExt called with incorrect parameter key $arrayKey", 1, E_USER_WARNING );
1645 } elseif( preg_match( '/^[0-9]+$/', $arrayKey ) && !in_array( $option,
1646 array( 'parse', 'parseinline', 'escape', 'escapenoentities',
1647 'replaceafter', 'parsemag', 'content' ) ) ) {
1648 # A numeric index with unknown value
1649 wfWarn( "wfMsgExt called with incorrect parameter $option", 1, E_USER_WARNING );
1650 }
1651 }
1652
1653 if( in_array( 'content', $options, true ) ) {
1654 $forContent = true;
1655 $langCode = true;
1656 $langCodeObj = null;
1657 } elseif( array_key_exists( 'language', $options ) ) {
1658 $forContent = false;
1659 $langCode = wfGetLangObj( $options['language'] );
1660 $langCodeObj = $langCode;
1661 } else {
1662 $forContent = false;
1663 $langCode = false;
1664 $langCodeObj = null;
1665 }
1666
1667 $string = wfMsgGetKey( $key, /*DB*/true, $langCode, /*Transform*/false );
1668
1669 if( !in_array( 'replaceafter', $options, true ) ) {
1670 $string = wfMsgReplaceArgs( $string, $args );
1671 }
1672
1673 $messageCache = MessageCache::singleton();
1674 $parseInline = in_array( 'parseinline', $options, true );
1675 if( in_array( 'parse', $options, true ) || $parseInline ) {
1676 $string = $messageCache->parse( $string, null, true, !$forContent, $langCodeObj );
1677 if ( $string instanceof ParserOutput ) {
1678 $string = $string->getText();
1679 }
1680
1681 if ( $parseInline ) {
1682 $m = array();
1683 if( preg_match( '/^<p>(.*)\n?<\/p>\n?$/sU', $string, $m ) ) {
1684 $string = $m[1];
1685 }
1686 }
1687 } elseif ( in_array( 'parsemag', $options, true ) ) {
1688 $string = $messageCache->transform( $string,
1689 !$forContent, $langCodeObj );
1690 }
1691
1692 if ( in_array( 'escape', $options, true ) ) {
1693 $string = htmlspecialchars ( $string );
1694 } elseif ( in_array( 'escapenoentities', $options, true ) ) {
1695 $string = Sanitizer::escapeHtmlAllowEntities( $string );
1696 }
1697
1698 if( in_array( 'replaceafter', $options, true ) ) {
1699 $string = wfMsgReplaceArgs( $string, $args );
1700 }
1701
1702 return $string;
1703 }
1704
1705 /**
1706 * Since wfMsg() and co suck, they don't return false if the message key they
1707 * looked up didn't exist but a XHTML string, this function checks for the
1708 * nonexistance of messages by checking the MessageCache::get() result directly.
1709 *
1710 * @deprecated since 1.18. Use Message::isDisabled().
1711 *
1712 * @param $key String: the message key looked up
1713 * @return Boolean True if the message *doesn't* exist.
1714 */
1715 function wfEmptyMsg( $key ) {
1716 return MessageCache::singleton()->get( $key, /*useDB*/true, /*content*/false ) === false;
1717 }
1718
1719 /**
1720 * Throw a debugging exception. This function previously once exited the process,
1721 * but now throws an exception instead, with similar results.
1722 *
1723 * @param $msg String: message shown when dying.
1724 */
1725 function wfDebugDieBacktrace( $msg = '' ) {
1726 throw new MWException( $msg );
1727 }
1728
1729 /**
1730 * Fetch server name for use in error reporting etc.
1731 * Use real server name if available, so we know which machine
1732 * in a server farm generated the current page.
1733 *
1734 * @return string
1735 */
1736 function wfHostname() {
1737 static $host;
1738 if ( is_null( $host ) ) {
1739
1740 # Hostname overriding
1741 global $wgOverrideHostname;
1742 if( $wgOverrideHostname !== false ) {
1743 # Set static and skip any detection
1744 $host = $wgOverrideHostname;
1745 return $host;
1746 }
1747
1748 if ( function_exists( 'posix_uname' ) ) {
1749 // This function not present on Windows
1750 $uname = posix_uname();
1751 } else {
1752 $uname = false;
1753 }
1754 if( is_array( $uname ) && isset( $uname['nodename'] ) ) {
1755 $host = $uname['nodename'];
1756 } elseif ( getenv( 'COMPUTERNAME' ) ) {
1757 # Windows computer name
1758 $host = getenv( 'COMPUTERNAME' );
1759 } else {
1760 # This may be a virtual server.
1761 $host = $_SERVER['SERVER_NAME'];
1762 }
1763 }
1764 return $host;
1765 }
1766
1767 /**
1768 * Returns a HTML comment with the elapsed time since request.
1769 * This method has no side effects.
1770 *
1771 * @return string
1772 */
1773 function wfReportTime() {
1774 global $wgRequestTime, $wgShowHostnames;
1775
1776 $elapsed = microtime( true ) - $wgRequestTime;
1777
1778 return $wgShowHostnames
1779 ? sprintf( '<!-- Served by %s in %01.3f secs. -->', wfHostname(), $elapsed )
1780 : sprintf( '<!-- Served in %01.3f secs. -->', $elapsed );
1781 }
1782
1783 /**
1784 * Safety wrapper for debug_backtrace().
1785 *
1786 * With Zend Optimizer 3.2.0 loaded, this causes segfaults under somewhat
1787 * murky circumstances, which may be triggered in part by stub objects
1788 * or other fancy talkin'.
1789 *
1790 * Will return an empty array if Zend Optimizer is detected or if
1791 * debug_backtrace is disabled, otherwise the output from
1792 * debug_backtrace() (trimmed).
1793 *
1794 * @param $limit int This parameter can be used to limit the number of stack frames returned
1795 *
1796 * @return array of backtrace information
1797 */
1798 function wfDebugBacktrace( $limit = 0 ) {
1799 static $disabled = null;
1800
1801 if( extension_loaded( 'Zend Optimizer' ) ) {
1802 wfDebug( "Zend Optimizer detected; skipping debug_backtrace for safety.\n" );
1803 return array();
1804 }
1805
1806 if ( is_null( $disabled ) ) {
1807 $disabled = false;
1808 $functions = explode( ',', ini_get( 'disable_functions' ) );
1809 $functions = array_map( 'trim', $functions );
1810 $functions = array_map( 'strtolower', $functions );
1811 if ( in_array( 'debug_backtrace', $functions ) ) {
1812 wfDebug( "debug_backtrace is in disabled_functions\n" );
1813 $disabled = true;
1814 }
1815 }
1816 if ( $disabled ) {
1817 return array();
1818 }
1819
1820 if ( $limit && version_compare( PHP_VERSION, '5.4.0', '>=' ) ) {
1821 return array_slice( debug_backtrace( DEBUG_BACKTRACE_PROVIDE_OBJECT, $limit + 1 ), 1 );
1822 } else {
1823 return array_slice( debug_backtrace(), 1 );
1824 }
1825 }
1826
1827 /**
1828 * Get a debug backtrace as a string
1829 *
1830 * @return string
1831 */
1832 function wfBacktrace() {
1833 global $wgCommandLineMode;
1834
1835 if ( $wgCommandLineMode ) {
1836 $msg = '';
1837 } else {
1838 $msg = "<ul>\n";
1839 }
1840 $backtrace = wfDebugBacktrace();
1841 foreach( $backtrace as $call ) {
1842 if( isset( $call['file'] ) ) {
1843 $f = explode( DIRECTORY_SEPARATOR, $call['file'] );
1844 $file = $f[count( $f ) - 1];
1845 } else {
1846 $file = '-';
1847 }
1848 if( isset( $call['line'] ) ) {
1849 $line = $call['line'];
1850 } else {
1851 $line = '-';
1852 }
1853 if ( $wgCommandLineMode ) {
1854 $msg .= "$file line $line calls ";
1855 } else {
1856 $msg .= '<li>' . $file . ' line ' . $line . ' calls ';
1857 }
1858 if( !empty( $call['class'] ) ) {
1859 $msg .= $call['class'] . $call['type'];
1860 }
1861 $msg .= $call['function'] . '()';
1862
1863 if ( $wgCommandLineMode ) {
1864 $msg .= "\n";
1865 } else {
1866 $msg .= "</li>\n";
1867 }
1868 }
1869 if ( $wgCommandLineMode ) {
1870 $msg .= "\n";
1871 } else {
1872 $msg .= "</ul>\n";
1873 }
1874
1875 return $msg;
1876 }
1877
1878 /**
1879 * Get the name of the function which called this function
1880 * wfGetCaller( 1 ) is the function with the wfGetCaller() call (ie. __FUNCTION__)
1881 * wfGetCaller( 2 ) [default] is the caller of the function running wfGetCaller()
1882 * wfGetCaller( 3 ) is the parent of that.
1883 *
1884 * @param $level Int
1885 * @return string
1886 */
1887 function wfGetCaller( $level = 2 ) {
1888 $backtrace = wfDebugBacktrace( $level + 1 );
1889 if ( isset( $backtrace[$level] ) ) {
1890 return wfFormatStackFrame( $backtrace[$level] );
1891 } else {
1892 return 'unknown';
1893 }
1894 }
1895
1896 /**
1897 * Return a string consisting of callers in the stack. Useful sometimes
1898 * for profiling specific points.
1899 *
1900 * @param $limit int The maximum depth of the stack frame to return, or false for
1901 * the entire stack.
1902 * @return String
1903 */
1904 function wfGetAllCallers( $limit = 3 ) {
1905 $trace = array_reverse( wfDebugBacktrace() );
1906 if ( !$limit || $limit > count( $trace ) - 1 ) {
1907 $limit = count( $trace ) - 1;
1908 }
1909 $trace = array_slice( $trace, -$limit - 1, $limit );
1910 return implode( '/', array_map( 'wfFormatStackFrame', $trace ) );
1911 }
1912
1913 /**
1914 * Return a string representation of frame
1915 *
1916 * @param $frame Array
1917 * @return string
1918 */
1919 function wfFormatStackFrame( $frame ) {
1920 return isset( $frame['class'] ) ?
1921 $frame['class'] . '::' . $frame['function'] :
1922 $frame['function'];
1923 }
1924
1925
1926 /* Some generic result counters, pulled out of SearchEngine */
1927
1928
1929 /**
1930 * @todo document
1931 *
1932 * @param $offset Int
1933 * @param $limit Int
1934 * @return String
1935 */
1936 function wfShowingResults( $offset, $limit ) {
1937 return wfMessage( 'showingresults' )->numParams( $limit, $offset + 1 )->parse();
1938 }
1939
1940 /**
1941 * Generate (prev x| next x) (20|50|100...) type links for paging
1942 *
1943 * @param $offset String
1944 * @param $limit Integer
1945 * @param $link String
1946 * @param $query String: optional URL query parameter string
1947 * @param $atend Bool: optional param for specified if this is the last page
1948 * @return String
1949 * @deprecated in 1.19; use Language::viewPrevNext() instead
1950 */
1951 function wfViewPrevNext( $offset, $limit, $link, $query = '', $atend = false ) {
1952 wfDeprecated( __METHOD__, '1.19' );
1953
1954 global $wgLang;
1955
1956 $query = wfCgiToArray( $query );
1957
1958 if( is_object( $link ) ) {
1959 $title = $link;
1960 } else {
1961 $title = Title::newFromText( $link );
1962 if( is_null( $title ) ) {
1963 return false;
1964 }
1965 }
1966
1967 return $wgLang->viewPrevNext( $title, $offset, $limit, $query, $atend );
1968 }
1969
1970 /**
1971 * Make a list item, used by various special pages
1972 *
1973 * @param $page String Page link
1974 * @param $details String Text between brackets
1975 * @param $oppositedm Boolean Add the direction mark opposite to your
1976 * language, to display text properly
1977 * @return String
1978 * @deprecated since 1.19; use Language::specialList() instead
1979 */
1980 function wfSpecialList( $page, $details, $oppositedm = true ) {
1981 wfDeprecated( __METHOD__, '1.19' );
1982
1983 global $wgLang;
1984 return $wgLang->specialList( $page, $details, $oppositedm );
1985 }
1986
1987 /**
1988 * @todo document
1989 * @todo FIXME: We may want to blacklist some broken browsers
1990 *
1991 * @param $force Bool
1992 * @return bool Whereas client accept gzip compression
1993 */
1994 function wfClientAcceptsGzip( $force = false ) {
1995 static $result = null;
1996 if ( $result === null || $force ) {
1997 $result = false;
1998 if( isset( $_SERVER['HTTP_ACCEPT_ENCODING'] ) ) {
1999 # @todo FIXME: We may want to blacklist some broken browsers
2000 $m = array();
2001 if( preg_match(
2002 '/\bgzip(?:;(q)=([0-9]+(?:\.[0-9]+)))?\b/',
2003 $_SERVER['HTTP_ACCEPT_ENCODING'],
2004 $m )
2005 )
2006 {
2007 if( isset( $m[2] ) && ( $m[1] == 'q' ) && ( $m[2] == 0 ) ) {
2008 $result = false;
2009 return $result;
2010 }
2011 wfDebug( "wfClientAcceptsGzip: client accepts gzip.\n" );
2012 $result = true;
2013 }
2014 }
2015 }
2016 return $result;
2017 }
2018
2019 /**
2020 * Obtain the offset and limit values from the request string;
2021 * used in special pages
2022 *
2023 * @param $deflimit Int default limit if none supplied
2024 * @param $optionname String Name of a user preference to check against
2025 * @return array
2026 *
2027 */
2028 function wfCheckLimits( $deflimit = 50, $optionname = 'rclimit' ) {
2029 global $wgRequest;
2030 return $wgRequest->getLimitOffset( $deflimit, $optionname );
2031 }
2032
2033 /**
2034 * Escapes the given text so that it may be output using addWikiText()
2035 * without any linking, formatting, etc. making its way through. This
2036 * is achieved by substituting certain characters with HTML entities.
2037 * As required by the callers, "<nowiki>" is not used.
2038 *
2039 * @param $text String: text to be escaped
2040 * @return String
2041 */
2042 function wfEscapeWikiText( $text ) {
2043 $text = strtr( "\n$text", array(
2044 '"' => '&#34;', '&' => '&#38;', "'" => '&#39;', '<' => '&#60;',
2045 '=' => '&#61;', '>' => '&#62;', '[' => '&#91;', ']' => '&#93;',
2046 '{' => '&#123;', '|' => '&#124;', '}' => '&#125;',
2047 "\n#" => "\n&#35;", "\n*" => "\n&#42;",
2048 "\n:" => "\n&#58;", "\n;" => "\n&#59;",
2049 '://' => '&#58;//', 'ISBN ' => 'ISBN&#32;', 'RFC ' => 'RFC&#32;',
2050 ) );
2051 return substr( $text, 1 );
2052 }
2053
2054 /**
2055 * Get the current unix timetstamp with microseconds. Useful for profiling
2056 * @return Float
2057 */
2058 function wfTime() {
2059 return microtime( true );
2060 }
2061
2062 /**
2063 * Sets dest to source and returns the original value of dest
2064 * If source is NULL, it just returns the value, it doesn't set the variable
2065 * If force is true, it will set the value even if source is NULL
2066 *
2067 * @param $dest Mixed
2068 * @param $source Mixed
2069 * @param $force Bool
2070 * @return Mixed
2071 */
2072 function wfSetVar( &$dest, $source, $force = false ) {
2073 $temp = $dest;
2074 if ( !is_null( $source ) || $force ) {
2075 $dest = $source;
2076 }
2077 return $temp;
2078 }
2079
2080 /**
2081 * As for wfSetVar except setting a bit
2082 *
2083 * @param $dest Int
2084 * @param $bit Int
2085 * @param $state Bool
2086 *
2087 * @return bool
2088 */
2089 function wfSetBit( &$dest, $bit, $state = true ) {
2090 $temp = (bool)( $dest & $bit );
2091 if ( !is_null( $state ) ) {
2092 if ( $state ) {
2093 $dest |= $bit;
2094 } else {
2095 $dest &= ~$bit;
2096 }
2097 }
2098 return $temp;
2099 }
2100
2101 /**
2102 * A wrapper around the PHP function var_export().
2103 * Either print it or add it to the regular output ($wgOut).
2104 *
2105 * @param $var mixed A PHP variable to dump.
2106 */
2107 function wfVarDump( $var ) {
2108 global $wgOut;
2109 $s = str_replace( "\n", "<br />\n", var_export( $var, true ) . "\n" );
2110 if ( headers_sent() || !isset( $wgOut ) || !is_object( $wgOut ) ) {
2111 print $s;
2112 } else {
2113 $wgOut->addHTML( $s );
2114 }
2115 }
2116
2117 /**
2118 * Provide a simple HTTP error.
2119 *
2120 * @param $code Int|String
2121 * @param $label String
2122 * @param $desc String
2123 */
2124 function wfHttpError( $code, $label, $desc ) {
2125 global $wgOut;
2126 $wgOut->disable();
2127 header( "HTTP/1.0 $code $label" );
2128 header( "Status: $code $label" );
2129 $wgOut->sendCacheControl();
2130
2131 header( 'Content-type: text/html; charset=utf-8' );
2132 print "<!doctype html>" .
2133 '<html><head><title>' .
2134 htmlspecialchars( $label ) .
2135 '</title></head><body><h1>' .
2136 htmlspecialchars( $label ) .
2137 '</h1><p>' .
2138 nl2br( htmlspecialchars( $desc ) ) .
2139 "</p></body></html>\n";
2140 }
2141
2142 /**
2143 * Clear away any user-level output buffers, discarding contents.
2144 *
2145 * Suitable for 'starting afresh', for instance when streaming
2146 * relatively large amounts of data without buffering, or wanting to
2147 * output image files without ob_gzhandler's compression.
2148 *
2149 * The optional $resetGzipEncoding parameter controls suppression of
2150 * the Content-Encoding header sent by ob_gzhandler; by default it
2151 * is left. See comments for wfClearOutputBuffers() for why it would
2152 * be used.
2153 *
2154 * Note that some PHP configuration options may add output buffer
2155 * layers which cannot be removed; these are left in place.
2156 *
2157 * @param $resetGzipEncoding Bool
2158 */
2159 function wfResetOutputBuffers( $resetGzipEncoding = true ) {
2160 if( $resetGzipEncoding ) {
2161 // Suppress Content-Encoding and Content-Length
2162 // headers from 1.10+s wfOutputHandler
2163 global $wgDisableOutputCompression;
2164 $wgDisableOutputCompression = true;
2165 }
2166 while( $status = ob_get_status() ) {
2167 if( $status['type'] == 0 /* PHP_OUTPUT_HANDLER_INTERNAL */ ) {
2168 // Probably from zlib.output_compression or other
2169 // PHP-internal setting which can't be removed.
2170 //
2171 // Give up, and hope the result doesn't break
2172 // output behavior.
2173 break;
2174 }
2175 if( !ob_end_clean() ) {
2176 // Could not remove output buffer handler; abort now
2177 // to avoid getting in some kind of infinite loop.
2178 break;
2179 }
2180 if( $resetGzipEncoding ) {
2181 if( $status['name'] == 'ob_gzhandler' ) {
2182 // Reset the 'Content-Encoding' field set by this handler
2183 // so we can start fresh.
2184 header_remove( 'Content-Encoding' );
2185 break;
2186 }
2187 }
2188 }
2189 }
2190
2191 /**
2192 * More legible than passing a 'false' parameter to wfResetOutputBuffers():
2193 *
2194 * Clear away output buffers, but keep the Content-Encoding header
2195 * produced by ob_gzhandler, if any.
2196 *
2197 * This should be used for HTTP 304 responses, where you need to
2198 * preserve the Content-Encoding header of the real result, but
2199 * also need to suppress the output of ob_gzhandler to keep to spec
2200 * and avoid breaking Firefox in rare cases where the headers and
2201 * body are broken over two packets.
2202 */
2203 function wfClearOutputBuffers() {
2204 wfResetOutputBuffers( false );
2205 }
2206
2207 /**
2208 * Converts an Accept-* header into an array mapping string values to quality
2209 * factors
2210 *
2211 * @param $accept String
2212 * @param $def String default
2213 * @return Array
2214 */
2215 function wfAcceptToPrefs( $accept, $def = '*/*' ) {
2216 # No arg means accept anything (per HTTP spec)
2217 if( !$accept ) {
2218 return array( $def => 1.0 );
2219 }
2220
2221 $prefs = array();
2222
2223 $parts = explode( ',', $accept );
2224
2225 foreach( $parts as $part ) {
2226 # @todo FIXME: Doesn't deal with params like 'text/html; level=1'
2227 $values = explode( ';', trim( $part ) );
2228 $match = array();
2229 if ( count( $values ) == 1 ) {
2230 $prefs[$values[0]] = 1.0;
2231 } elseif ( preg_match( '/q\s*=\s*(\d*\.\d+)/', $values[1], $match ) ) {
2232 $prefs[$values[0]] = floatval( $match[1] );
2233 }
2234 }
2235
2236 return $prefs;
2237 }
2238
2239 /**
2240 * Checks if a given MIME type matches any of the keys in the given
2241 * array. Basic wildcards are accepted in the array keys.
2242 *
2243 * Returns the matching MIME type (or wildcard) if a match, otherwise
2244 * NULL if no match.
2245 *
2246 * @param $type String
2247 * @param $avail Array
2248 * @return string
2249 * @private
2250 */
2251 function mimeTypeMatch( $type, $avail ) {
2252 if( array_key_exists( $type, $avail ) ) {
2253 return $type;
2254 } else {
2255 $parts = explode( '/', $type );
2256 if( array_key_exists( $parts[0] . '/*', $avail ) ) {
2257 return $parts[0] . '/*';
2258 } elseif( array_key_exists( '*/*', $avail ) ) {
2259 return '*/*';
2260 } else {
2261 return null;
2262 }
2263 }
2264 }
2265
2266 /**
2267 * Returns the 'best' match between a client's requested internet media types
2268 * and the server's list of available types. Each list should be an associative
2269 * array of type to preference (preference is a float between 0.0 and 1.0).
2270 * Wildcards in the types are acceptable.
2271 *
2272 * @param $cprefs Array: client's acceptable type list
2273 * @param $sprefs Array: server's offered types
2274 * @return string
2275 *
2276 * @todo FIXME: Doesn't handle params like 'text/plain; charset=UTF-8'
2277 * XXX: generalize to negotiate other stuff
2278 */
2279 function wfNegotiateType( $cprefs, $sprefs ) {
2280 $combine = array();
2281
2282 foreach( array_keys( $sprefs ) as $type ) {
2283 $parts = explode( '/', $type );
2284 if( $parts[1] != '*' ) {
2285 $ckey = mimeTypeMatch( $type, $cprefs );
2286 if( $ckey ) {
2287 $combine[$type] = $sprefs[$type] * $cprefs[$ckey];
2288 }
2289 }
2290 }
2291
2292 foreach( array_keys( $cprefs ) as $type ) {
2293 $parts = explode( '/', $type );
2294 if( $parts[1] != '*' && !array_key_exists( $type, $sprefs ) ) {
2295 $skey = mimeTypeMatch( $type, $sprefs );
2296 if( $skey ) {
2297 $combine[$type] = $sprefs[$skey] * $cprefs[$type];
2298 }
2299 }
2300 }
2301
2302 $bestq = 0;
2303 $besttype = null;
2304
2305 foreach( array_keys( $combine ) as $type ) {
2306 if( $combine[$type] > $bestq ) {
2307 $besttype = $type;
2308 $bestq = $combine[$type];
2309 }
2310 }
2311
2312 return $besttype;
2313 }
2314
2315 /**
2316 * Reference-counted warning suppression
2317 *
2318 * @param $end Bool
2319 */
2320 function wfSuppressWarnings( $end = false ) {
2321 static $suppressCount = 0;
2322 static $originalLevel = false;
2323
2324 if ( $end ) {
2325 if ( $suppressCount ) {
2326 --$suppressCount;
2327 if ( !$suppressCount ) {
2328 error_reporting( $originalLevel );
2329 }
2330 }
2331 } else {
2332 if ( !$suppressCount ) {
2333 $originalLevel = error_reporting( E_ALL & ~( E_WARNING | E_NOTICE | E_USER_WARNING | E_USER_NOTICE | E_DEPRECATED | E_USER_DEPRECATED ) );
2334 }
2335 ++$suppressCount;
2336 }
2337 }
2338
2339 /**
2340 * Restore error level to previous value
2341 */
2342 function wfRestoreWarnings() {
2343 wfSuppressWarnings( true );
2344 }
2345
2346 # Autodetect, convert and provide timestamps of various types
2347
2348 /**
2349 * Unix time - the number of seconds since 1970-01-01 00:00:00 UTC
2350 */
2351 define( 'TS_UNIX', 0 );
2352
2353 /**
2354 * MediaWiki concatenated string timestamp (YYYYMMDDHHMMSS)
2355 */
2356 define( 'TS_MW', 1 );
2357
2358 /**
2359 * MySQL DATETIME (YYYY-MM-DD HH:MM:SS)
2360 */
2361 define( 'TS_DB', 2 );
2362
2363 /**
2364 * RFC 2822 format, for E-mail and HTTP headers
2365 */
2366 define( 'TS_RFC2822', 3 );
2367
2368 /**
2369 * ISO 8601 format with no timezone: 1986-02-09T20:00:00Z
2370 *
2371 * This is used by Special:Export
2372 */
2373 define( 'TS_ISO_8601', 4 );
2374
2375 /**
2376 * An Exif timestamp (YYYY:MM:DD HH:MM:SS)
2377 *
2378 * @see http://exif.org/Exif2-2.PDF The Exif 2.2 spec, see page 28 for the
2379 * DateTime tag and page 36 for the DateTimeOriginal and
2380 * DateTimeDigitized tags.
2381 */
2382 define( 'TS_EXIF', 5 );
2383
2384 /**
2385 * Oracle format time.
2386 */
2387 define( 'TS_ORACLE', 6 );
2388
2389 /**
2390 * Postgres format time.
2391 */
2392 define( 'TS_POSTGRES', 7 );
2393
2394 /**
2395 * DB2 format time
2396 */
2397 define( 'TS_DB2', 8 );
2398
2399 /**
2400 * ISO 8601 basic format with no timezone: 19860209T200000Z. This is used by ResourceLoader
2401 */
2402 define( 'TS_ISO_8601_BASIC', 9 );
2403
2404 /**
2405 * Get a timestamp string in one of various formats
2406 *
2407 * @deprecated
2408 * @param $outputtype Mixed: A timestamp in one of the supported formats, the
2409 * function will autodetect which format is supplied and act
2410 * accordingly.
2411 * @param $ts Mixed: the timestamp to convert or 0 for the current timestamp
2412 * @return Mixed: String / false The same date in the format specified in $outputtype or false
2413 */
2414 function wfTimestamp( $outputtype = TS_UNIX, $ts = 0 ) {
2415 $timestamp = new MWTimestamp( $ts );
2416 return $timestamp->getTimestamp( $outputtype );
2417 }
2418
2419 /**
2420 * Return a formatted timestamp, or null if input is null.
2421 * For dealing with nullable timestamp columns in the database.
2422 *
2423 * @param $outputtype Integer
2424 * @param $ts String
2425 * @return String
2426 */
2427 function wfTimestampOrNull( $outputtype = TS_UNIX, $ts = null ) {
2428 if( is_null( $ts ) ) {
2429 return null;
2430 } else {
2431 return wfTimestamp( $outputtype, $ts );
2432 }
2433 }
2434
2435 /**
2436 * Convenience function; returns MediaWiki timestamp for the present time.
2437 *
2438 * @return string
2439 */
2440 function wfTimestampNow() {
2441 # return NOW
2442 return wfTimestamp( TS_MW, time() );
2443 }
2444
2445 /**
2446 * Check if the operating system is Windows
2447 *
2448 * @return Bool: true if it's Windows, False otherwise.
2449 */
2450 function wfIsWindows() {
2451 static $isWindows = null;
2452 if ( $isWindows === null ) {
2453 $isWindows = substr( php_uname(), 0, 7 ) == 'Windows';
2454 }
2455 return $isWindows;
2456 }
2457
2458 /**
2459 * Check if we are running under HipHop
2460 *
2461 * @return Bool
2462 */
2463 function wfIsHipHop() {
2464 return function_exists( 'hphp_thread_set_warmup_enabled' );
2465 }
2466
2467 /**
2468 * Swap two variables
2469 *
2470 * @param $x Mixed
2471 * @param $y Mixed
2472 */
2473 function swap( &$x, &$y ) {
2474 $z = $x;
2475 $x = $y;
2476 $y = $z;
2477 }
2478
2479 /**
2480 * Tries to get the system directory for temporary files. First
2481 * $wgTmpDirectory is checked, and then the TMPDIR, TMP, and TEMP
2482 * environment variables are then checked in sequence, and if none are
2483 * set try sys_get_temp_dir().
2484 *
2485 * NOTE: When possible, use instead the tmpfile() function to create
2486 * temporary files to avoid race conditions on file creation, etc.
2487 *
2488 * @return String
2489 */
2490 function wfTempDir() {
2491 global $wgTmpDirectory;
2492
2493 if ( $wgTmpDirectory !== false ) {
2494 return $wgTmpDirectory;
2495 }
2496
2497 $tmpDir = array_map( "getenv", array( 'TMPDIR', 'TMP', 'TEMP' ) );
2498
2499 foreach( $tmpDir as $tmp ) {
2500 if( $tmp && file_exists( $tmp ) && is_dir( $tmp ) && is_writable( $tmp ) ) {
2501 return $tmp;
2502 }
2503 }
2504 return sys_get_temp_dir();
2505 }
2506
2507 /**
2508 * Make directory, and make all parent directories if they don't exist
2509 *
2510 * @param $dir String: full path to directory to create
2511 * @param $mode Integer: chmod value to use, default is $wgDirectoryMode
2512 * @param $caller String: optional caller param for debugging.
2513 * @return bool
2514 */
2515 function wfMkdirParents( $dir, $mode = null, $caller = null ) {
2516 global $wgDirectoryMode;
2517
2518 if ( FileBackend::isStoragePath( $dir ) ) { // sanity
2519 throw new MWException( __FUNCTION__ . " given storage path '$dir'." );
2520 }
2521
2522 if ( !is_null( $caller ) ) {
2523 wfDebug( "$caller: called wfMkdirParents($dir)\n" );
2524 }
2525
2526 if( strval( $dir ) === '' || file_exists( $dir ) ) {
2527 return true;
2528 }
2529
2530 $dir = str_replace( array( '\\', '/' ), DIRECTORY_SEPARATOR, $dir );
2531
2532 if ( is_null( $mode ) ) {
2533 $mode = $wgDirectoryMode;
2534 }
2535
2536 // Turn off the normal warning, we're doing our own below
2537 wfSuppressWarnings();
2538 $ok = mkdir( $dir, $mode, true ); // PHP5 <3
2539 wfRestoreWarnings();
2540
2541 if( !$ok ) {
2542 // PHP doesn't report the path in its warning message, so add our own to aid in diagnosis.
2543 trigger_error( sprintf( "%s: failed to mkdir \"%s\" mode 0%o", __FUNCTION__, $dir, $mode ),
2544 E_USER_WARNING );
2545 }
2546 return $ok;
2547 }
2548
2549 /**
2550 * Remove a directory and all its content.
2551 * Does not hide error.
2552 */
2553 function wfRecursiveRemoveDir( $dir ) {
2554 wfDebug( __FUNCTION__ . "( $dir )\n" );
2555 // taken from http://de3.php.net/manual/en/function.rmdir.php#98622
2556 if ( is_dir( $dir ) ) {
2557 $objects = scandir( $dir );
2558 foreach ( $objects as $object ) {
2559 if ( $object != "." && $object != ".." ) {
2560 if ( filetype( $dir . '/' . $object ) == "dir" ) {
2561 wfRecursiveRemoveDir( $dir . '/' . $object );
2562 } else {
2563 unlink( $dir . '/' . $object );
2564 }
2565 }
2566 }
2567 reset( $objects );
2568 rmdir( $dir );
2569 }
2570 }
2571
2572 /**
2573 * @param $nr Mixed: the number to format
2574 * @param $acc Integer: the number of digits after the decimal point, default 2
2575 * @param $round Boolean: whether or not to round the value, default true
2576 * @return float
2577 */
2578 function wfPercent( $nr, $acc = 2, $round = true ) {
2579 $ret = sprintf( "%.${acc}f", $nr );
2580 return $round ? round( $ret, $acc ) . '%' : "$ret%";
2581 }
2582
2583 /**
2584 * Find out whether or not a mixed variable exists in a string
2585 *
2586 * @param $needle String
2587 * @param $str String
2588 * @param $insensitive Boolean
2589 * @return Boolean
2590 */
2591 function in_string( $needle, $str, $insensitive = false ) {
2592 $func = 'strpos';
2593 if( $insensitive ) $func = 'stripos';
2594
2595 return $func( $str, $needle ) !== false;
2596 }
2597
2598 /**
2599 * Safety wrapper around ini_get() for boolean settings.
2600 * The values returned from ini_get() are pre-normalized for settings
2601 * set via php.ini or php_flag/php_admin_flag... but *not*
2602 * for those set via php_value/php_admin_value.
2603 *
2604 * It's fairly common for people to use php_value instead of php_flag,
2605 * which can leave you with an 'off' setting giving a false positive
2606 * for code that just takes the ini_get() return value as a boolean.
2607 *
2608 * To make things extra interesting, setting via php_value accepts
2609 * "true" and "yes" as true, but php.ini and php_flag consider them false. :)
2610 * Unrecognized values go false... again opposite PHP's own coercion
2611 * from string to bool.
2612 *
2613 * Luckily, 'properly' set settings will always come back as '0' or '1',
2614 * so we only have to worry about them and the 'improper' settings.
2615 *
2616 * I frickin' hate PHP... :P
2617 *
2618 * @param $setting String
2619 * @return Bool
2620 */
2621 function wfIniGetBool( $setting ) {
2622 $val = ini_get( $setting );
2623 // 'on' and 'true' can't have whitespace around them, but '1' can.
2624 return strtolower( $val ) == 'on'
2625 || strtolower( $val ) == 'true'
2626 || strtolower( $val ) == 'yes'
2627 || preg_match( "/^\s*[+-]?0*[1-9]/", $val ); // approx C atoi() function
2628 }
2629
2630 /**
2631 * Wrapper function for PHP's dl(). This doesn't work in most situations from
2632 * PHP 5.3 onward, and is usually disabled in shared environments anyway.
2633 *
2634 * @param $extension String A PHP extension. The file suffix (.so or .dll)
2635 * should be omitted
2636 * @param $fileName String Name of the library, if not $extension.suffix
2637 * @return Bool - Whether or not the extension is loaded
2638 */
2639 function wfDl( $extension, $fileName = null ) {
2640 if( extension_loaded( $extension ) ) {
2641 return true;
2642 }
2643
2644 $canDl = false;
2645 $sapi = php_sapi_name();
2646 if( $sapi == 'cli' || $sapi == 'cgi' || $sapi == 'embed' ) {
2647 $canDl = ( function_exists( 'dl' ) && is_callable( 'dl' )
2648 && wfIniGetBool( 'enable_dl' ) && !wfIniGetBool( 'safe_mode' ) );
2649 }
2650
2651 if( $canDl ) {
2652 $fileName = $fileName ? $fileName : $extension;
2653 if( wfIsWindows() ) {
2654 $fileName = 'php_' . $fileName;
2655 }
2656 wfSuppressWarnings();
2657 dl( $fileName . '.' . PHP_SHLIB_SUFFIX );
2658 wfRestoreWarnings();
2659 }
2660 return extension_loaded( $extension );
2661 }
2662
2663 /**
2664 * Windows-compatible version of escapeshellarg()
2665 * Windows doesn't recognise single-quotes in the shell, but the escapeshellarg()
2666 * function puts single quotes in regardless of OS.
2667 *
2668 * Also fixes the locale problems on Linux in PHP 5.2.6+ (bug backported to
2669 * earlier distro releases of PHP)
2670 *
2671 * @param varargs
2672 * @return String
2673 */
2674 function wfEscapeShellArg( ) {
2675 wfInitShellLocale();
2676
2677 $args = func_get_args();
2678 $first = true;
2679 $retVal = '';
2680 foreach ( $args as $arg ) {
2681 if ( !$first ) {
2682 $retVal .= ' ';
2683 } else {
2684 $first = false;
2685 }
2686
2687 if ( wfIsWindows() ) {
2688 // Escaping for an MSVC-style command line parser and CMD.EXE
2689 // Refs:
2690 // * http://web.archive.org/web/20020708081031/http://mailman.lyra.org/pipermail/scite-interest/2002-March/000436.html
2691 // * http://technet.microsoft.com/en-us/library/cc723564.aspx
2692 // * Bug #13518
2693 // * CR r63214
2694 // Double the backslashes before any double quotes. Escape the double quotes.
2695 $tokens = preg_split( '/(\\\\*")/', $arg, -1, PREG_SPLIT_DELIM_CAPTURE );
2696 $arg = '';
2697 $iteration = 0;
2698 foreach ( $tokens as $token ) {
2699 if ( $iteration % 2 == 1 ) {
2700 // Delimiter, a double quote preceded by zero or more slashes
2701 $arg .= str_replace( '\\', '\\\\', substr( $token, 0, -1 ) ) . '\\"';
2702 } elseif ( $iteration % 4 == 2 ) {
2703 // ^ in $token will be outside quotes, need to be escaped
2704 $arg .= str_replace( '^', '^^', $token );
2705 } else { // $iteration % 4 == 0
2706 // ^ in $token will appear inside double quotes, so leave as is
2707 $arg .= $token;
2708 }
2709 $iteration++;
2710 }
2711 // Double the backslashes before the end of the string, because
2712 // we will soon add a quote
2713 $m = array();
2714 if ( preg_match( '/^(.*?)(\\\\+)$/', $arg, $m ) ) {
2715 $arg = $m[1] . str_replace( '\\', '\\\\', $m[2] );
2716 }
2717
2718 // Add surrounding quotes
2719 $retVal .= '"' . $arg . '"';
2720 } else {
2721 $retVal .= escapeshellarg( $arg );
2722 }
2723 }
2724 return $retVal;
2725 }
2726
2727 /**
2728 * Execute a shell command, with time and memory limits mirrored from the PHP
2729 * configuration if supported.
2730 * @param $cmd String Command line, properly escaped for shell.
2731 * @param &$retval null|Mixed optional, will receive the program's exit code.
2732 * (non-zero is usually failure)
2733 * @param $environ Array optional environment variables which should be
2734 * added to the executed command environment.
2735 * @param $limits Array optional array with limits(filesize, memory, time)
2736 * this overwrites the global wgShellMax* limits.
2737 * @return string collected stdout as a string (trailing newlines stripped)
2738 */
2739 function wfShellExec( $cmd, &$retval = null, $environ = array(), $limits = array() ) {
2740 global $IP, $wgMaxShellMemory, $wgMaxShellFileSize, $wgMaxShellTime;
2741
2742 static $disabled;
2743 if ( is_null( $disabled ) ) {
2744 $disabled = false;
2745 if( wfIniGetBool( 'safe_mode' ) ) {
2746 wfDebug( "wfShellExec can't run in safe_mode, PHP's exec functions are too broken.\n" );
2747 $disabled = 'safemode';
2748 } else {
2749 $functions = explode( ',', ini_get( 'disable_functions' ) );
2750 $functions = array_map( 'trim', $functions );
2751 $functions = array_map( 'strtolower', $functions );
2752 if ( in_array( 'passthru', $functions ) ) {
2753 wfDebug( "passthru is in disabled_functions\n" );
2754 $disabled = 'passthru';
2755 }
2756 }
2757 }
2758 if ( $disabled ) {
2759 $retval = 1;
2760 return $disabled == 'safemode' ?
2761 'Unable to run external programs in safe mode.' :
2762 'Unable to run external programs, passthru() is disabled.';
2763 }
2764
2765 wfInitShellLocale();
2766
2767 $envcmd = '';
2768 foreach( $environ as $k => $v ) {
2769 if ( wfIsWindows() ) {
2770 /* Surrounding a set in quotes (method used by wfEscapeShellArg) makes the quotes themselves
2771 * appear in the environment variable, so we must use carat escaping as documented in
2772 * http://technet.microsoft.com/en-us/library/cc723564.aspx
2773 * Note however that the quote isn't listed there, but is needed, and the parentheses
2774 * are listed there but doesn't appear to need it.
2775 */
2776 $envcmd .= "set $k=" . preg_replace( '/([&|()<>^"])/', '^\\1', $v ) . '&& ';
2777 } else {
2778 /* Assume this is a POSIX shell, thus required to accept variable assignments before the command
2779 * http://www.opengroup.org/onlinepubs/009695399/utilities/xcu_chap02.html#tag_02_09_01
2780 */
2781 $envcmd .= "$k=" . escapeshellarg( $v ) . ' ';
2782 }
2783 }
2784 $cmd = $envcmd . $cmd;
2785
2786 if ( php_uname( 's' ) == 'Linux' ) {
2787 $time = intval ( isset($limits['time']) ? $limits['time'] : $wgMaxShellTime );
2788 $mem = intval ( isset($limits['memory']) ? $limits['memory'] : $wgMaxShellMemory );
2789 $filesize = intval ( isset($limits['filesize']) ? $limits['filesize'] : $wgMaxShellFileSize );
2790
2791 if ( $time > 0 && $mem > 0 ) {
2792 $script = "$IP/bin/ulimit4.sh";
2793 if ( is_executable( $script ) ) {
2794 $cmd = '/bin/bash ' . escapeshellarg( $script ) . " $time $mem $filesize " . escapeshellarg( $cmd );
2795 }
2796 }
2797 }
2798 wfDebug( "wfShellExec: $cmd\n" );
2799
2800 $retval = 1; // error by default?
2801 ob_start();
2802 passthru( $cmd, $retval );
2803 $output = ob_get_contents();
2804 ob_end_clean();
2805
2806 if ( $retval == 127 ) {
2807 wfDebugLog( 'exec', "Possibly missing executable file: $cmd\n" );
2808 }
2809 return $output;
2810 }
2811
2812 /**
2813 * Workaround for http://bugs.php.net/bug.php?id=45132
2814 * escapeshellarg() destroys non-ASCII characters if LANG is not a UTF-8 locale
2815 */
2816 function wfInitShellLocale() {
2817 static $done = false;
2818 if ( $done ) {
2819 return;
2820 }
2821 $done = true;
2822 global $wgShellLocale;
2823 if ( !wfIniGetBool( 'safe_mode' ) ) {
2824 putenv( "LC_CTYPE=$wgShellLocale" );
2825 setlocale( LC_CTYPE, $wgShellLocale );
2826 }
2827 }
2828
2829 /**
2830 * Alias to wfShellWikiCmd()
2831 * @see wfShellWikiCmd()
2832 */
2833 function wfShellMaintenanceCmd( $script, array $parameters = array(), array $options = array() ) {
2834 return wfShellWikiCmd( $script, $parameters, $options );
2835 }
2836
2837 /**
2838 * Generate a shell-escaped command line string to run a MediaWiki cli script.
2839 * Note that $parameters should be a flat array and an option with an argument
2840 * should consist of two consecutive items in the array (do not use "--option value").
2841 * @param $script string MediaWiki cli script path
2842 * @param $parameters Array Arguments and options to the script
2843 * @param $options Array Associative array of options:
2844 * 'php': The path to the php executable
2845 * 'wrapper': Path to a PHP wrapper to handle the maintenance script
2846 * @return Array
2847 */
2848 function wfShellWikiCmd( $script, array $parameters = array(), array $options = array() ) {
2849 global $wgPhpCli;
2850 // Give site config file a chance to run the script in a wrapper.
2851 // The caller may likely want to call wfBasename() on $script.
2852 wfRunHooks( 'wfShellWikiCmd', array( &$script, &$parameters, &$options ) );
2853 $cmd = isset( $options['php'] ) ? array( $options['php'] ) : array( $wgPhpCli );
2854 if ( isset( $options['wrapper'] ) ) {
2855 $cmd[] = $options['wrapper'];
2856 }
2857 $cmd[] = $script;
2858 // Escape each parameter for shell
2859 return implode( " ", array_map( 'wfEscapeShellArg', array_merge( $cmd, $parameters ) ) );
2860 }
2861
2862 /**
2863 * wfMerge attempts to merge differences between three texts.
2864 * Returns true for a clean merge and false for failure or a conflict.
2865 *
2866 * @param $old String
2867 * @param $mine String
2868 * @param $yours String
2869 * @param $result String
2870 * @return Bool
2871 */
2872 function wfMerge( $old, $mine, $yours, &$result ) {
2873 global $wgDiff3;
2874
2875 # This check may also protect against code injection in
2876 # case of broken installations.
2877 wfSuppressWarnings();
2878 $haveDiff3 = $wgDiff3 && file_exists( $wgDiff3 );
2879 wfRestoreWarnings();
2880
2881 if( !$haveDiff3 ) {
2882 wfDebug( "diff3 not found\n" );
2883 return false;
2884 }
2885
2886 # Make temporary files
2887 $td = wfTempDir();
2888 $oldtextFile = fopen( $oldtextName = tempnam( $td, 'merge-old-' ), 'w' );
2889 $mytextFile = fopen( $mytextName = tempnam( $td, 'merge-mine-' ), 'w' );
2890 $yourtextFile = fopen( $yourtextName = tempnam( $td, 'merge-your-' ), 'w' );
2891
2892 fwrite( $oldtextFile, $old );
2893 fclose( $oldtextFile );
2894 fwrite( $mytextFile, $mine );
2895 fclose( $mytextFile );
2896 fwrite( $yourtextFile, $yours );
2897 fclose( $yourtextFile );
2898
2899 # Check for a conflict
2900 $cmd = $wgDiff3 . ' -a --overlap-only ' .
2901 wfEscapeShellArg( $mytextName ) . ' ' .
2902 wfEscapeShellArg( $oldtextName ) . ' ' .
2903 wfEscapeShellArg( $yourtextName );
2904 $handle = popen( $cmd, 'r' );
2905
2906 if( fgets( $handle, 1024 ) ) {
2907 $conflict = true;
2908 } else {
2909 $conflict = false;
2910 }
2911 pclose( $handle );
2912
2913 # Merge differences
2914 $cmd = $wgDiff3 . ' -a -e --merge ' .
2915 wfEscapeShellArg( $mytextName, $oldtextName, $yourtextName );
2916 $handle = popen( $cmd, 'r' );
2917 $result = '';
2918 do {
2919 $data = fread( $handle, 8192 );
2920 if ( strlen( $data ) == 0 ) {
2921 break;
2922 }
2923 $result .= $data;
2924 } while ( true );
2925 pclose( $handle );
2926 unlink( $mytextName );
2927 unlink( $oldtextName );
2928 unlink( $yourtextName );
2929
2930 if ( $result === '' && $old !== '' && !$conflict ) {
2931 wfDebug( "Unexpected null result from diff3. Command: $cmd\n" );
2932 $conflict = true;
2933 }
2934 return !$conflict;
2935 }
2936
2937 /**
2938 * Returns unified plain-text diff of two texts.
2939 * Useful for machine processing of diffs.
2940 *
2941 * @param $before String: the text before the changes.
2942 * @param $after String: the text after the changes.
2943 * @param $params String: command-line options for the diff command.
2944 * @return String: unified diff of $before and $after
2945 */
2946 function wfDiff( $before, $after, $params = '-u' ) {
2947 if ( $before == $after ) {
2948 return '';
2949 }
2950
2951 global $wgDiff;
2952 wfSuppressWarnings();
2953 $haveDiff = $wgDiff && file_exists( $wgDiff );
2954 wfRestoreWarnings();
2955
2956 # This check may also protect against code injection in
2957 # case of broken installations.
2958 if( !$haveDiff ) {
2959 wfDebug( "diff executable not found\n" );
2960 $diffs = new Diff( explode( "\n", $before ), explode( "\n", $after ) );
2961 $format = new UnifiedDiffFormatter();
2962 return $format->format( $diffs );
2963 }
2964
2965 # Make temporary files
2966 $td = wfTempDir();
2967 $oldtextFile = fopen( $oldtextName = tempnam( $td, 'merge-old-' ), 'w' );
2968 $newtextFile = fopen( $newtextName = tempnam( $td, 'merge-your-' ), 'w' );
2969
2970 fwrite( $oldtextFile, $before );
2971 fclose( $oldtextFile );
2972 fwrite( $newtextFile, $after );
2973 fclose( $newtextFile );
2974
2975 // Get the diff of the two files
2976 $cmd = "$wgDiff " . $params . ' ' . wfEscapeShellArg( $oldtextName, $newtextName );
2977
2978 $h = popen( $cmd, 'r' );
2979
2980 $diff = '';
2981
2982 do {
2983 $data = fread( $h, 8192 );
2984 if ( strlen( $data ) == 0 ) {
2985 break;
2986 }
2987 $diff .= $data;
2988 } while ( true );
2989
2990 // Clean up
2991 pclose( $h );
2992 unlink( $oldtextName );
2993 unlink( $newtextName );
2994
2995 // Kill the --- and +++ lines. They're not useful.
2996 $diff_lines = explode( "\n", $diff );
2997 if ( strpos( $diff_lines[0], '---' ) === 0 ) {
2998 unset( $diff_lines[0] );
2999 }
3000 if ( strpos( $diff_lines[1], '+++' ) === 0 ) {
3001 unset( $diff_lines[1] );
3002 }
3003
3004 $diff = implode( "\n", $diff_lines );
3005
3006 return $diff;
3007 }
3008
3009 /**
3010 * This function works like "use VERSION" in Perl, the program will die with a
3011 * backtrace if the current version of PHP is less than the version provided
3012 *
3013 * This is useful for extensions which due to their nature are not kept in sync
3014 * with releases, and might depend on other versions of PHP than the main code
3015 *
3016 * Note: PHP might die due to parsing errors in some cases before it ever
3017 * manages to call this function, such is life
3018 *
3019 * @see perldoc -f use
3020 *
3021 * @param $req_ver Mixed: the version to check, can be a string, an integer, or
3022 * a float
3023 */
3024 function wfUsePHP( $req_ver ) {
3025 $php_ver = PHP_VERSION;
3026
3027 if ( version_compare( $php_ver, (string)$req_ver, '<' ) ) {
3028 throw new MWException( "PHP $req_ver required--this is only $php_ver" );
3029 }
3030 }
3031
3032 /**
3033 * This function works like "use VERSION" in Perl except it checks the version
3034 * of MediaWiki, the program will die with a backtrace if the current version
3035 * of MediaWiki is less than the version provided.
3036 *
3037 * This is useful for extensions which due to their nature are not kept in sync
3038 * with releases
3039 *
3040 * @see perldoc -f use
3041 *
3042 * @param $req_ver Mixed: the version to check, can be a string, an integer, or
3043 * a float
3044 */
3045 function wfUseMW( $req_ver ) {
3046 global $wgVersion;
3047
3048 if ( version_compare( $wgVersion, (string)$req_ver, '<' ) ) {
3049 throw new MWException( "MediaWiki $req_ver required--this is only $wgVersion" );
3050 }
3051 }
3052
3053 /**
3054 * Return the final portion of a pathname.
3055 * Reimplemented because PHP5's "basename()" is buggy with multibyte text.
3056 * http://bugs.php.net/bug.php?id=33898
3057 *
3058 * PHP's basename() only considers '\' a pathchar on Windows and Netware.
3059 * We'll consider it so always, as we don't want '\s' in our Unix paths either.
3060 *
3061 * @param $path String
3062 * @param $suffix String: to remove if present
3063 * @return String
3064 */
3065 function wfBaseName( $path, $suffix = '' ) {
3066 $encSuffix = ( $suffix == '' )
3067 ? ''
3068 : ( '(?:' . preg_quote( $suffix, '#' ) . ')?' );
3069 $matches = array();
3070 if( preg_match( "#([^/\\\\]*?){$encSuffix}[/\\\\]*$#", $path, $matches ) ) {
3071 return $matches[1];
3072 } else {
3073 return '';
3074 }
3075 }
3076
3077 /**
3078 * Generate a relative path name to the given file.
3079 * May explode on non-matching case-insensitive paths,
3080 * funky symlinks, etc.
3081 *
3082 * @param $path String: absolute destination path including target filename
3083 * @param $from String: Absolute source path, directory only
3084 * @return String
3085 */
3086 function wfRelativePath( $path, $from ) {
3087 // Normalize mixed input on Windows...
3088 $path = str_replace( '/', DIRECTORY_SEPARATOR, $path );
3089 $from = str_replace( '/', DIRECTORY_SEPARATOR, $from );
3090
3091 // Trim trailing slashes -- fix for drive root
3092 $path = rtrim( $path, DIRECTORY_SEPARATOR );
3093 $from = rtrim( $from, DIRECTORY_SEPARATOR );
3094
3095 $pieces = explode( DIRECTORY_SEPARATOR, dirname( $path ) );
3096 $against = explode( DIRECTORY_SEPARATOR, $from );
3097
3098 if( $pieces[0] !== $against[0] ) {
3099 // Non-matching Windows drive letters?
3100 // Return a full path.
3101 return $path;
3102 }
3103
3104 // Trim off common prefix
3105 while( count( $pieces ) && count( $against )
3106 && $pieces[0] == $against[0] ) {
3107 array_shift( $pieces );
3108 array_shift( $against );
3109 }
3110
3111 // relative dots to bump us to the parent
3112 while( count( $against ) ) {
3113 array_unshift( $pieces, '..' );
3114 array_shift( $against );
3115 }
3116
3117 array_push( $pieces, wfBaseName( $path ) );
3118
3119 return implode( DIRECTORY_SEPARATOR, $pieces );
3120 }
3121
3122 /**
3123 * Do any deferred updates and clear the list
3124 *
3125 * @deprecated since 1.19
3126 * @see DeferredUpdates::doUpdate()
3127 * @param $commit string
3128 */
3129 function wfDoUpdates( $commit = '' ) {
3130 wfDeprecated( __METHOD__, '1.19' );
3131 DeferredUpdates::doUpdates( $commit );
3132 }
3133
3134 /**
3135 * Convert an arbitrarily-long digit string from one numeric base
3136 * to another, optionally zero-padding to a minimum column width.
3137 *
3138 * Supports base 2 through 36; digit values 10-36 are represented
3139 * as lowercase letters a-z. Input is case-insensitive.
3140 *
3141 * @param $input String: of digits
3142 * @param $sourceBase Integer: 2-36
3143 * @param $destBase Integer: 2-36
3144 * @param $pad Integer: 1 or greater
3145 * @param $lowercase Boolean
3146 * @return String or false on invalid input
3147 */
3148 function wfBaseConvert( $input, $sourceBase, $destBase, $pad = 1, $lowercase = true ) {
3149 $input = strval( $input );
3150 if( $sourceBase < 2 ||
3151 $sourceBase > 36 ||
3152 $destBase < 2 ||
3153 $destBase > 36 ||
3154 $pad < 1 ||
3155 $sourceBase != intval( $sourceBase ) ||
3156 $destBase != intval( $destBase ) ||
3157 $pad != intval( $pad ) ||
3158 !is_string( $input ) ||
3159 $input == '' ) {
3160 return false;
3161 }
3162 $digitChars = ( $lowercase ) ? '0123456789abcdefghijklmnopqrstuvwxyz' : '0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ';
3163 $inDigits = array();
3164 $outChars = '';
3165
3166 // Decode and validate input string
3167 $input = strtolower( $input );
3168 for( $i = 0; $i < strlen( $input ); $i++ ) {
3169 $n = strpos( $digitChars, $input[$i] );
3170 if( $n === false || $n > $sourceBase ) {
3171 return false;
3172 }
3173 $inDigits[] = $n;
3174 }
3175
3176 // Iterate over the input, modulo-ing out an output digit
3177 // at a time until input is gone.
3178 while( count( $inDigits ) ) {
3179 $work = 0;
3180 $workDigits = array();
3181
3182 // Long division...
3183 foreach( $inDigits as $digit ) {
3184 $work *= $sourceBase;
3185 $work += $digit;
3186
3187 if( $work < $destBase ) {
3188 // Gonna need to pull another digit.
3189 if( count( $workDigits ) ) {
3190 // Avoid zero-padding; this lets us find
3191 // the end of the input very easily when
3192 // length drops to zero.
3193 $workDigits[] = 0;
3194 }
3195 } else {
3196 // Finally! Actual division!
3197 $workDigits[] = intval( $work / $destBase );
3198
3199 // Isn't it annoying that most programming languages
3200 // don't have a single divide-and-remainder operator,
3201 // even though the CPU implements it that way?
3202 $work = $work % $destBase;
3203 }
3204 }
3205
3206 // All that division leaves us with a remainder,
3207 // which is conveniently our next output digit.
3208 $outChars .= $digitChars[$work];
3209
3210 // And we continue!
3211 $inDigits = $workDigits;
3212 }
3213
3214 while( strlen( $outChars ) < $pad ) {
3215 $outChars .= '0';
3216 }
3217
3218 return strrev( $outChars );
3219 }
3220
3221 /**
3222 * Create an object with a given name and an array of construct parameters
3223 *
3224 * @param $name String
3225 * @param $p Array: parameters
3226 * @return object
3227 * @deprecated since 1.18, warnings in 1.18, removal in 1.20
3228 */
3229 function wfCreateObject( $name, $p ) {
3230 wfDeprecated( __FUNCTION__, '1.18' );
3231 return MWFunction::newObj( $name, $p );
3232 }
3233
3234 /**
3235 * @return bool
3236 */
3237 function wfHttpOnlySafe() {
3238 global $wgHttpOnlyBlacklist;
3239
3240 if( isset( $_SERVER['HTTP_USER_AGENT'] ) ) {
3241 foreach( $wgHttpOnlyBlacklist as $regex ) {
3242 if( preg_match( $regex, $_SERVER['HTTP_USER_AGENT'] ) ) {
3243 return false;
3244 }
3245 }
3246 }
3247
3248 return true;
3249 }
3250
3251 /**
3252 * Override session_id before session startup if php's built-in
3253 * session generation code is not secure.
3254 */
3255 function wfFixSessionID() {
3256 // If the cookie or session id is already set we already have a session and should abort
3257 if ( isset( $_COOKIE[ session_name() ] ) || session_id() ) {
3258 return;
3259 }
3260
3261 // PHP's built-in session entropy is enabled if:
3262 // - entropy_file is set or you're on Windows with php 5.3.3+
3263 // - AND entropy_length is > 0
3264 // We treat it as disabled if it doesn't have an entropy length of at least 32
3265 $entropyEnabled = (
3266 ( wfIsWindows() && version_compare( PHP_VERSION, '5.3.3', '>=' ) )
3267 || ini_get( 'session.entropy_file' )
3268 )
3269 && intval( ini_get( 'session.entropy_length' ) ) >= 32;
3270
3271 // If built-in entropy is not enabled or not sufficient override php's built in session id generation code
3272 if ( !$entropyEnabled ) {
3273 wfDebug( __METHOD__ . ": PHP's built in entropy is disabled or not sufficient, overriding session id generation using our cryptrand source.\n" );
3274 session_id( MWCryptRand::generateHex( 32 ) );
3275 }
3276 }
3277
3278 /**
3279 * Initialise php session
3280 *
3281 * @param $sessionId Bool
3282 */
3283 function wfSetupSession( $sessionId = false ) {
3284 global $wgSessionsInMemcached, $wgSessionsInObjectCache, $wgCookiePath, $wgCookieDomain,
3285 $wgCookieSecure, $wgCookieHttpOnly, $wgSessionHandler;
3286 if( $wgSessionsInObjectCache || $wgSessionsInMemcached ) {
3287 ObjectCacheSessionHandler::install();
3288 } elseif( $wgSessionHandler && $wgSessionHandler != ini_get( 'session.save_handler' ) ) {
3289 # Only set this if $wgSessionHandler isn't null and session.save_handler
3290 # hasn't already been set to the desired value (that causes errors)
3291 ini_set( 'session.save_handler', $wgSessionHandler );
3292 }
3293 $httpOnlySafe = wfHttpOnlySafe() && $wgCookieHttpOnly;
3294 wfDebugLog( 'cookie',
3295 'session_set_cookie_params: "' . implode( '", "',
3296 array(
3297 0,
3298 $wgCookiePath,
3299 $wgCookieDomain,
3300 $wgCookieSecure,
3301 $httpOnlySafe ) ) . '"' );
3302 session_set_cookie_params( 0, $wgCookiePath, $wgCookieDomain, $wgCookieSecure, $httpOnlySafe );
3303 session_cache_limiter( 'private, must-revalidate' );
3304 if ( $sessionId ) {
3305 session_id( $sessionId );
3306 } else {
3307 wfFixSessionID();
3308 }
3309 wfSuppressWarnings();
3310 session_start();
3311 wfRestoreWarnings();
3312 }
3313
3314 /**
3315 * Get an object from the precompiled serialized directory
3316 *
3317 * @param $name String
3318 * @return Mixed: the variable on success, false on failure
3319 */
3320 function wfGetPrecompiledData( $name ) {
3321 global $IP;
3322
3323 $file = "$IP/serialized/$name";
3324 if ( file_exists( $file ) ) {
3325 $blob = file_get_contents( $file );
3326 if ( $blob ) {
3327 return unserialize( $blob );
3328 }
3329 }
3330 return false;
3331 }
3332
3333 /**
3334 * Get a cache key
3335 *
3336 * @param varargs
3337 * @return String
3338 */
3339 function wfMemcKey( /*... */ ) {
3340 global $wgCachePrefix;
3341 $prefix = $wgCachePrefix === false ? wfWikiID() : $wgCachePrefix;
3342 $args = func_get_args();
3343 $key = $prefix . ':' . implode( ':', $args );
3344 $key = str_replace( ' ', '_', $key );
3345 return $key;
3346 }
3347
3348 /**
3349 * Get a cache key for a foreign DB
3350 *
3351 * @param $db String
3352 * @param $prefix String
3353 * @param varargs String
3354 * @return String
3355 */
3356 function wfForeignMemcKey( $db, $prefix /*, ... */ ) {
3357 $args = array_slice( func_get_args(), 2 );
3358 if ( $prefix ) {
3359 $key = "$db-$prefix:" . implode( ':', $args );
3360 } else {
3361 $key = $db . ':' . implode( ':', $args );
3362 }
3363 return $key;
3364 }
3365
3366 /**
3367 * Get an ASCII string identifying this wiki
3368 * This is used as a prefix in memcached keys
3369 *
3370 * @return String
3371 */
3372 function wfWikiID() {
3373 global $wgDBprefix, $wgDBname;
3374 if ( $wgDBprefix ) {
3375 return "$wgDBname-$wgDBprefix";
3376 } else {
3377 return $wgDBname;
3378 }
3379 }
3380
3381 /**
3382 * Split a wiki ID into DB name and table prefix
3383 *
3384 * @param $wiki String
3385 *
3386 * @return array
3387 */
3388 function wfSplitWikiID( $wiki ) {
3389 $bits = explode( '-', $wiki, 2 );
3390 if ( count( $bits ) < 2 ) {
3391 $bits[] = '';
3392 }
3393 return $bits;
3394 }
3395
3396 /**
3397 * Get a Database object.
3398 *
3399 * @param $db Integer: index of the connection to get. May be DB_MASTER for the
3400 * master (for write queries), DB_SLAVE for potentially lagged read
3401 * queries, or an integer >= 0 for a particular server.
3402 *
3403 * @param $groups Mixed: query groups. An array of group names that this query
3404 * belongs to. May contain a single string if the query is only
3405 * in one group.
3406 *
3407 * @param $wiki String: the wiki ID, or false for the current wiki
3408 *
3409 * Note: multiple calls to wfGetDB(DB_SLAVE) during the course of one request
3410 * will always return the same object, unless the underlying connection or load
3411 * balancer is manually destroyed.
3412 *
3413 * Note 2: use $this->getDB() in maintenance scripts that may be invoked by
3414 * updater to ensure that a proper database is being updated.
3415 *
3416 * @return DatabaseBase
3417 */
3418 function &wfGetDB( $db, $groups = array(), $wiki = false ) {
3419 return wfGetLB( $wiki )->getConnection( $db, $groups, $wiki );
3420 }
3421
3422 /**
3423 * Get a load balancer object.
3424 *
3425 * @param $wiki String: wiki ID, or false for the current wiki
3426 * @return LoadBalancer
3427 */
3428 function wfGetLB( $wiki = false ) {
3429 return wfGetLBFactory()->getMainLB( $wiki );
3430 }
3431
3432 /**
3433 * Get the load balancer factory object
3434 *
3435 * @return LBFactory
3436 */
3437 function &wfGetLBFactory() {
3438 return LBFactory::singleton();
3439 }
3440
3441 /**
3442 * Find a file.
3443 * Shortcut for RepoGroup::singleton()->findFile()
3444 *
3445 * @param $title String or Title object
3446 * @param $options array Associative array of options:
3447 * time: requested time for an archived image, or false for the
3448 * current version. An image object will be returned which was
3449 * created at the specified time.
3450 *
3451 * ignoreRedirect: If true, do not follow file redirects
3452 *
3453 * private: If true, return restricted (deleted) files if the current
3454 * user is allowed to view them. Otherwise, such files will not
3455 * be found.
3456 *
3457 * bypassCache: If true, do not use the process-local cache of File objects
3458 *
3459 * @return File, or false if the file does not exist
3460 */
3461 function wfFindFile( $title, $options = array() ) {
3462 return RepoGroup::singleton()->findFile( $title, $options );
3463 }
3464
3465 /**
3466 * Get an object referring to a locally registered file.
3467 * Returns a valid placeholder object if the file does not exist.
3468 *
3469 * @param $title Title|String
3470 * @return LocalFile|null A File, or null if passed an invalid Title
3471 */
3472 function wfLocalFile( $title ) {
3473 return RepoGroup::singleton()->getLocalRepo()->newFile( $title );
3474 }
3475
3476 /**
3477 * Stream a file to the browser. Back-compat alias for StreamFile::stream()
3478 * @deprecated since 1.19
3479 */
3480 function wfStreamFile( $fname, $headers = array() ) {
3481 wfDeprecated( __FUNCTION__, '1.19' );
3482 StreamFile::stream( $fname, $headers );
3483 }
3484
3485 /**
3486 * Should low-performance queries be disabled?
3487 *
3488 * @return Boolean
3489 * @codeCoverageIgnore
3490 */
3491 function wfQueriesMustScale() {
3492 global $wgMiserMode;
3493 return $wgMiserMode
3494 || ( SiteStats::pages() > 100000
3495 && SiteStats::edits() > 1000000
3496 && SiteStats::users() > 10000 );
3497 }
3498
3499 /**
3500 * Get the path to a specified script file, respecting file
3501 * extensions; this is a wrapper around $wgScriptExtension etc.
3502 * except for 'index' and 'load' which use $wgScript/$wgLoadScript
3503 *
3504 * @param $script String: script filename, sans extension
3505 * @return String
3506 */
3507 function wfScript( $script = 'index' ) {
3508 global $wgScriptPath, $wgScriptExtension, $wgScript, $wgLoadScript;
3509 if ( $script === 'index' ) {
3510 return $wgScript;
3511 } else if ( $script === 'load' ) {
3512 return $wgLoadScript;
3513 } else {
3514 return "{$wgScriptPath}/{$script}{$wgScriptExtension}";
3515 }
3516 }
3517
3518 /**
3519 * Get the script URL.
3520 *
3521 * @return string script URL
3522 */
3523 function wfGetScriptUrl() {
3524 if( isset( $_SERVER['SCRIPT_NAME'] ) ) {
3525 #
3526 # as it was called, minus the query string.
3527 #
3528 # Some sites use Apache rewrite rules to handle subdomains,
3529 # and have PHP set up in a weird way that causes PHP_SELF
3530 # to contain the rewritten URL instead of the one that the
3531 # outside world sees.
3532 #
3533 # If in this mode, use SCRIPT_URL instead, which mod_rewrite
3534 # provides containing the "before" URL.
3535 return $_SERVER['SCRIPT_NAME'];
3536 } else {
3537 return $_SERVER['URL'];
3538 }
3539 }
3540
3541 /**
3542 * Convenience function converts boolean values into "true"
3543 * or "false" (string) values
3544 *
3545 * @param $value Boolean
3546 * @return String
3547 */
3548 function wfBoolToStr( $value ) {
3549 return $value ? 'true' : 'false';
3550 }
3551
3552 /**
3553 * Load an extension messages file
3554 *
3555 * @deprecated since 1.16, warnings in 1.18, remove in 1.20
3556 * @codeCoverageIgnore
3557 */
3558 function wfLoadExtensionMessages() {
3559 wfDeprecated( __FUNCTION__, '1.16' );
3560 }
3561
3562 /**
3563 * Get a platform-independent path to the null file, e.g. /dev/null
3564 *
3565 * @return string
3566 */
3567 function wfGetNull() {
3568 return wfIsWindows()
3569 ? 'NUL'
3570 : '/dev/null';
3571 }
3572
3573 /**
3574 * Modern version of wfWaitForSlaves(). Instead of looking at replication lag
3575 * and waiting for it to go down, this waits for the slaves to catch up to the
3576 * master position. Use this when updating very large numbers of rows, as
3577 * in maintenance scripts, to avoid causing too much lag. Of course, this is
3578 * a no-op if there are no slaves.
3579 *
3580 * @param $maxLag Integer (deprecated)
3581 * @param $wiki mixed Wiki identifier accepted by wfGetLB
3582 */
3583 function wfWaitForSlaves( $maxLag = false, $wiki = false ) {
3584 $lb = wfGetLB( $wiki );
3585 // bug 27975 - Don't try to wait for slaves if there are none
3586 // Prevents permission error when getting master position
3587 if ( $lb->getServerCount() > 1 ) {
3588 $dbw = $lb->getConnection( DB_MASTER );
3589 $pos = $dbw->getMasterPos();
3590 $lb->waitForAll( $pos );
3591 }
3592 }
3593
3594 /**
3595 * Used to be used for outputting text in the installer/updater
3596 * @deprecated since 1.18, warnings in 1.18, remove in 1.20
3597 */
3598 function wfOut( $s ) {
3599 wfDeprecated( __FUNCTION__, '1.18' );
3600 global $wgCommandLineMode;
3601 if ( $wgCommandLineMode ) {
3602 echo $s;
3603 } else {
3604 echo htmlspecialchars( $s );
3605 }
3606 flush();
3607 }
3608
3609 /**
3610 * Count down from $n to zero on the terminal, with a one-second pause
3611 * between showing each number. For use in command-line scripts.
3612 * @codeCoverageIgnore
3613 * @param $n int
3614 */
3615 function wfCountDown( $n ) {
3616 for ( $i = $n; $i >= 0; $i-- ) {
3617 if ( $i != $n ) {
3618 echo str_repeat( "\x08", strlen( $i + 1 ) );
3619 }
3620 echo $i;
3621 flush();
3622 if ( $i ) {
3623 sleep( 1 );
3624 }
3625 }
3626 echo "\n";
3627 }
3628
3629 /**
3630 * Generate a random 32-character hexadecimal token.
3631 * @param $salt Mixed: some sort of salt, if necessary, to add to random
3632 * characters before hashing.
3633 * @return string
3634 * @codeCoverageIgnore
3635 * @deprecated since 1.20; Please use MWCryptRand for security purposes and wfRandomString for pesudo-random strings
3636 * @warning This method is NOT secure. Additionally it has many callers that use it for pesudo-random purposes.
3637 */
3638 function wfGenerateToken( $salt = '' ) {
3639 wfDeprecated( __METHOD__, '1.20' );
3640 $salt = serialize( $salt );
3641 return md5( mt_rand( 0, 0x7fffffff ) . $salt );
3642 }
3643
3644 /**
3645 * Replace all invalid characters with -
3646 * Additional characters can be defined in $wgIllegalFileChars (see bug 20489)
3647 * By default, $wgIllegalFileChars = ':'
3648 *
3649 * @param $name Mixed: filename to process
3650 * @return String
3651 */
3652 function wfStripIllegalFilenameChars( $name ) {
3653 global $wgIllegalFileChars;
3654 $illegalFileChars = $wgIllegalFileChars ? "|[" . $wgIllegalFileChars . "]" : '';
3655 $name = wfBaseName( $name );
3656 $name = preg_replace(
3657 "/[^" . Title::legalChars() . "]" . $illegalFileChars . "/",
3658 '-',
3659 $name
3660 );
3661 return $name;
3662 }
3663
3664 /**
3665 * Set PHP's memory limit to the larger of php.ini or $wgMemoryLimit;
3666 *
3667 * @return Integer value memory was set to.
3668 */
3669 function wfMemoryLimit() {
3670 global $wgMemoryLimit;
3671 $memlimit = wfShorthandToInteger( ini_get( 'memory_limit' ) );
3672 if( $memlimit != -1 ) {
3673 $conflimit = wfShorthandToInteger( $wgMemoryLimit );
3674 if( $conflimit == -1 ) {
3675 wfDebug( "Removing PHP's memory limit\n" );
3676 wfSuppressWarnings();
3677 ini_set( 'memory_limit', $conflimit );
3678 wfRestoreWarnings();
3679 return $conflimit;
3680 } elseif ( $conflimit > $memlimit ) {
3681 wfDebug( "Raising PHP's memory limit to $conflimit bytes\n" );
3682 wfSuppressWarnings();
3683 ini_set( 'memory_limit', $conflimit );
3684 wfRestoreWarnings();
3685 return $conflimit;
3686 }
3687 }
3688 return $memlimit;
3689 }
3690
3691 /**
3692 * Converts shorthand byte notation to integer form
3693 *
3694 * @param $string String
3695 * @return Integer
3696 */
3697 function wfShorthandToInteger( $string = '' ) {
3698 $string = trim( $string );
3699 if( $string === '' ) {
3700 return -1;
3701 }
3702 $last = $string[strlen( $string ) - 1];
3703 $val = intval( $string );
3704 switch( $last ) {
3705 case 'g':
3706 case 'G':
3707 $val *= 1024;
3708 // break intentionally missing
3709 case 'm':
3710 case 'M':
3711 $val *= 1024;
3712 // break intentionally missing
3713 case 'k':
3714 case 'K':
3715 $val *= 1024;
3716 }
3717
3718 return $val;
3719 }
3720
3721 /**
3722 * Get the normalised IETF language tag
3723 * See unit test for examples.
3724 *
3725 * @param $code String: The language code.
3726 * @return String: The language code which complying with BCP 47 standards.
3727 */
3728 function wfBCP47( $code ) {
3729 $codeSegment = explode( '-', $code );
3730 $codeBCP = array();
3731 foreach ( $codeSegment as $segNo => $seg ) {
3732 if ( count( $codeSegment ) > 0 ) {
3733 // when previous segment is x, it is a private segment and should be lc
3734 if( $segNo > 0 && strtolower( $codeSegment[( $segNo - 1 )] ) == 'x' ) {
3735 $codeBCP[$segNo] = strtolower( $seg );
3736 // ISO 3166 country code
3737 } elseif ( ( strlen( $seg ) == 2 ) && ( $segNo > 0 ) ) {
3738 $codeBCP[$segNo] = strtoupper( $seg );
3739 // ISO 15924 script code
3740 } elseif ( ( strlen( $seg ) == 4 ) && ( $segNo > 0 ) ) {
3741 $codeBCP[$segNo] = ucfirst( strtolower( $seg ) );
3742 // Use lowercase for other cases
3743 } else {
3744 $codeBCP[$segNo] = strtolower( $seg );
3745 }
3746 } else {
3747 // Use lowercase for single segment
3748 $codeBCP[$segNo] = strtolower( $seg );
3749 }
3750 }
3751 $langCode = implode( '-', $codeBCP );
3752 return $langCode;
3753 }
3754
3755 /**
3756 * Get a cache object.
3757 *
3758 * @param $inputType integer Cache type, one the the CACHE_* constants.
3759 * @return BagOStuff
3760 */
3761 function wfGetCache( $inputType ) {
3762 return ObjectCache::getInstance( $inputType );
3763 }
3764
3765 /**
3766 * Get the main cache object
3767 *
3768 * @return BagOStuff
3769 */
3770 function wfGetMainCache() {
3771 global $wgMainCacheType;
3772 return ObjectCache::getInstance( $wgMainCacheType );
3773 }
3774
3775 /**
3776 * Get the cache object used by the message cache
3777 *
3778 * @return BagOStuff
3779 */
3780 function wfGetMessageCacheStorage() {
3781 global $wgMessageCacheType;
3782 return ObjectCache::getInstance( $wgMessageCacheType );
3783 }
3784
3785 /**
3786 * Get the cache object used by the parser cache
3787 *
3788 * @return BagOStuff
3789 */
3790 function wfGetParserCacheStorage() {
3791 global $wgParserCacheType;
3792 return ObjectCache::getInstance( $wgParserCacheType );
3793 }
3794
3795 /**
3796 * Get the cache object used by the language converter
3797 *
3798 * @return BagOStuff
3799 */
3800 function wfGetLangConverterCacheStorage() {
3801 global $wgLanguageConverterCacheType;
3802 return ObjectCache::getInstance( $wgLanguageConverterCacheType );
3803 }
3804
3805 /**
3806 * Call hook functions defined in $wgHooks
3807 *
3808 * @param $event String: event name
3809 * @param $args Array: parameters passed to hook functions
3810 * @return Boolean True if no handler aborted the hook
3811 */
3812 function wfRunHooks( $event, $args = array() ) {
3813 return Hooks::run( $event, $args );
3814 }
3815
3816 /**
3817 * Wrapper around php's unpack.
3818 *
3819 * @param $format String: The format string (See php's docs)
3820 * @param $data: A binary string of binary data
3821 * @param $length integer or false: The minimun length of $data. This is to
3822 * prevent reading beyond the end of $data. false to disable the check.
3823 *
3824 * Also be careful when using this function to read unsigned 32 bit integer
3825 * because php might make it negative.
3826 *
3827 * @throws MWException if $data not long enough, or if unpack fails
3828 * @return array Associative array of the extracted data
3829 */
3830 function wfUnpack( $format, $data, $length=false ) {
3831 if ( $length !== false ) {
3832 $realLen = strlen( $data );
3833 if ( $realLen < $length ) {
3834 throw new MWException( "Tried to use wfUnpack on a "
3835 . "string of length $realLen, but needed one "
3836 . "of at least length $length."
3837 );
3838 }
3839 }
3840
3841 wfSuppressWarnings();
3842 $result = unpack( $format, $data );
3843 wfRestoreWarnings();
3844
3845 if ( $result === false ) {
3846 // If it cannot extract the packed data.
3847 throw new MWException( "unpack could not unpack binary data" );
3848 }
3849 return $result;
3850 }
3851
3852 /**
3853 * Determine if an image exists on the 'bad image list'.
3854 *
3855 * The format of MediaWiki:Bad_image_list is as follows:
3856 * * Only list items (lines starting with "*") are considered
3857 * * The first link on a line must be a link to a bad image
3858 * * Any subsequent links on the same line are considered to be exceptions,
3859 * i.e. articles where the image may occur inline.
3860 *
3861 * @param $name string the image name to check
3862 * @param $contextTitle Title|bool the page on which the image occurs, if known
3863 * @param $blacklist string wikitext of a file blacklist
3864 * @return bool
3865 */
3866 function wfIsBadImage( $name, $contextTitle = false, $blacklist = null ) {
3867 static $badImageCache = null; // based on bad_image_list msg
3868 wfProfileIn( __METHOD__ );
3869
3870 # Handle redirects
3871 $redirectTitle = RepoGroup::singleton()->checkRedirect( Title::makeTitle( NS_FILE, $name ) );
3872 if( $redirectTitle ) {
3873 $name = $redirectTitle->getDbKey();
3874 }
3875
3876 # Run the extension hook
3877 $bad = false;
3878 if( !wfRunHooks( 'BadImage', array( $name, &$bad ) ) ) {
3879 wfProfileOut( __METHOD__ );
3880 return $bad;
3881 }
3882
3883 $cacheable = ( $blacklist === null );
3884 if( $cacheable && $badImageCache !== null ) {
3885 $badImages = $badImageCache;
3886 } else { // cache miss
3887 if ( $blacklist === null ) {
3888 $blacklist = wfMessage( 'bad_image_list' )->inContentLanguage()->plain(); // site list
3889 }
3890 # Build the list now
3891 $badImages = array();
3892 $lines = explode( "\n", $blacklist );
3893 foreach( $lines as $line ) {
3894 # List items only
3895 if ( substr( $line, 0, 1 ) !== '*' ) {
3896 continue;
3897 }
3898
3899 # Find all links
3900 $m = array();
3901 if ( !preg_match_all( '/\[\[:?(.*?)\]\]/', $line, $m ) ) {
3902 continue;
3903 }
3904
3905 $exceptions = array();
3906 $imageDBkey = false;
3907 foreach ( $m[1] as $i => $titleText ) {
3908 $title = Title::newFromText( $titleText );
3909 if ( !is_null( $title ) ) {
3910 if ( $i == 0 ) {
3911 $imageDBkey = $title->getDBkey();
3912 } else {
3913 $exceptions[$title->getPrefixedDBkey()] = true;
3914 }
3915 }
3916 }
3917
3918 if ( $imageDBkey !== false ) {
3919 $badImages[$imageDBkey] = $exceptions;
3920 }
3921 }
3922 if ( $cacheable ) {
3923 $badImageCache = $badImages;
3924 }
3925 }
3926
3927 $contextKey = $contextTitle ? $contextTitle->getPrefixedDBkey() : false;
3928 $bad = isset( $badImages[$name] ) && !isset( $badImages[$name][$contextKey] );
3929 wfProfileOut( __METHOD__ );
3930 return $bad;
3931 }