Add a way for packagers to override some installation details
[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 // most of the protocols are followed by ://, but mailto: and sometimes news: not, check for it
790 if ( in_array( $bits['scheme'] . '://', $wgUrlProtocols ) ) {
791 $bits['delimiter'] = '://';
792 } elseif ( in_array( $bits['scheme'] . ':', $wgUrlProtocols ) ) {
793 $bits['delimiter'] = ':';
794 // parse_url detects for news: and mailto: the host part of an url as path
795 // We have to correct this wrong detection
796 if ( isset( $bits['path'] ) ) {
797 $bits['host'] = $bits['path'];
798 $bits['path'] = '';
799 }
800 } else {
801 return false;
802 }
803
804 /* Provide an empty host for eg. file:/// urls (see bug 28627) */
805 if ( !isset( $bits['host'] ) ) {
806 $bits['host'] = '';
807
808 /* parse_url loses the third / for file:///c:/ urls (but not on variants) */
809 if ( substr( $bits['path'], 0, 1 ) !== '/' ) {
810 $bits['path'] = '/' . $bits['path'];
811 }
812 }
813
814 // If the URL was protocol-relative, fix scheme and delimiter
815 if ( $wasRelative ) {
816 $bits['scheme'] = '';
817 $bits['delimiter'] = '//';
818 }
819 return $bits;
820 }
821
822 /**
823 * Take a URL, make sure it's expanded to fully qualified, and replace any
824 * encoded non-ASCII Unicode characters with their UTF-8 original forms
825 * for more compact display and legibility for local audiences.
826 *
827 * @todo handle punycode domains too
828 *
829 * @param $url string
830 * @return string
831 */
832 function wfExpandIRI( $url ) {
833 return preg_replace_callback( '/((?:%[89A-F][0-9A-F])+)/i', 'wfExpandIRI_callback', wfExpandUrl( $url ) );
834 }
835
836 /**
837 * Private callback for wfExpandIRI
838 * @param array $matches
839 * @return string
840 */
841 function wfExpandIRI_callback( $matches ) {
842 return urldecode( $matches[1] );
843 }
844
845
846
847 /**
848 * Make URL indexes, appropriate for the el_index field of externallinks.
849 *
850 * @param $url String
851 * @return array
852 */
853 function wfMakeUrlIndexes( $url ) {
854 $bits = wfParseUrl( $url );
855
856 // Reverse the labels in the hostname, convert to lower case
857 // For emails reverse domainpart only
858 if ( $bits['scheme'] == 'mailto' ) {
859 $mailparts = explode( '@', $bits['host'], 2 );
860 if ( count( $mailparts ) === 2 ) {
861 $domainpart = strtolower( implode( '.', array_reverse( explode( '.', $mailparts[1] ) ) ) );
862 } else {
863 // No domain specified, don't mangle it
864 $domainpart = '';
865 }
866 $reversedHost = $domainpart . '@' . $mailparts[0];
867 } else {
868 $reversedHost = strtolower( implode( '.', array_reverse( explode( '.', $bits['host'] ) ) ) );
869 }
870 // Add an extra dot to the end
871 // Why? Is it in wrong place in mailto links?
872 if ( substr( $reversedHost, -1, 1 ) !== '.' ) {
873 $reversedHost .= '.';
874 }
875 // Reconstruct the pseudo-URL
876 $prot = $bits['scheme'];
877 $index = $prot . $bits['delimiter'] . $reversedHost;
878 // Leave out user and password. Add the port, path, query and fragment
879 if ( isset( $bits['port'] ) ) {
880 $index .= ':' . $bits['port'];
881 }
882 if ( isset( $bits['path'] ) ) {
883 $index .= $bits['path'];
884 } else {
885 $index .= '/';
886 }
887 if ( isset( $bits['query'] ) ) {
888 $index .= '?' . $bits['query'];
889 }
890 if ( isset( $bits['fragment'] ) ) {
891 $index .= '#' . $bits['fragment'];
892 }
893
894 if ( $prot == '' ) {
895 return array( "http:$index", "https:$index" );
896 } else {
897 return array( $index );
898 }
899 }
900
901 /**
902 * Check whether a given URL has a domain that occurs in a given set of domains
903 * @param $url string URL
904 * @param $domains array Array of domains (strings)
905 * @return bool True if the host part of $url ends in one of the strings in $domains
906 */
907 function wfMatchesDomainList( $url, $domains ) {
908 $bits = wfParseUrl( $url );
909 if ( is_array( $bits ) && isset( $bits['host'] ) ) {
910 foreach ( (array)$domains as $domain ) {
911 // FIXME: This gives false positives. http://nds-nl.wikipedia.org will match nl.wikipedia.org
912 // We should use something that interprets dots instead
913 if ( substr( $bits['host'], -strlen( $domain ) ) === $domain ) {
914 return true;
915 }
916 }
917 }
918 return false;
919 }
920
921 /**
922 * Sends a line to the debug log if enabled or, optionally, to a comment in output.
923 * In normal operation this is a NOP.
924 *
925 * Controlling globals:
926 * $wgDebugLogFile - points to the log file
927 * $wgProfileOnly - if set, normal debug messages will not be recorded.
928 * $wgDebugRawPage - if false, 'action=raw' hits will not result in debug output.
929 * $wgDebugComments - if on, some debug items may appear in comments in the HTML output.
930 *
931 * @param $text String
932 * @param $logonly Bool: set true to avoid appearing in HTML when $wgDebugComments is set
933 */
934 function wfDebug( $text, $logonly = false ) {
935 global $wgOut, $wgDebugLogFile, $wgDebugComments, $wgProfileOnly, $wgDebugRawPage;
936 global $wgDebugLogPrefix, $wgShowDebug;
937
938 static $cache = array(); // Cache of unoutputted messages
939
940 if ( !$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 ( ( $wgDebugComments || $wgShowDebug ) && !$logonly ) {
950 $cache[] = $text;
951
952 if ( isset( $wgOut ) && is_object( $wgOut ) ) {
953 // add the message and any cached messages to the output
954 array_map( array( $wgOut, 'debug' ), $cache );
955 $cache = array();
956 }
957 }
958 if ( wfRunHooks( 'Debug', array( $text, null /* no log group */ ) ) ) {
959 if ( $wgDebugLogFile != '' && !$wgProfileOnly ) {
960 # Strip unprintables; they can switch terminal modes when binary data
961 # gets dumped, which is pretty annoying.
962 $text = preg_replace( '![\x00-\x08\x0b\x0c\x0e-\x1f]!', ' ', $text );
963 $text = $wgDebugLogPrefix . $text;
964 wfErrorLog( $text, $wgDebugLogFile );
965 }
966 }
967
968 MWDebug::debugMsg( $text );
969 }
970
971 /**
972 * Returns true if debug logging should be suppressed if $wgDebugRawPage = false
973 * @return bool
974 */
975 function wfIsDebugRawPage() {
976 static $cache;
977 if ( $cache !== null ) {
978 return $cache;
979 }
980 # Check for raw action using $_GET not $wgRequest, since the latter might not be initialised yet
981 if ( ( isset( $_GET['action'] ) && $_GET['action'] == 'raw' )
982 || (
983 isset( $_SERVER['SCRIPT_NAME'] )
984 && substr( $_SERVER['SCRIPT_NAME'], -8 ) == 'load.php'
985 ) )
986 {
987 $cache = true;
988 } else {
989 $cache = false;
990 }
991 return $cache;
992 }
993
994 /**
995 * Get microsecond timestamps for debug logs
996 *
997 * @return string
998 */
999 function wfDebugTimer() {
1000 global $wgDebugTimestamps, $wgRequestTime;
1001
1002 if ( !$wgDebugTimestamps ) {
1003 return '';
1004 }
1005
1006 $prefix = sprintf( "%6.4f", microtime( true ) - $wgRequestTime );
1007 $mem = sprintf( "%5.1fM", ( memory_get_usage( true ) / ( 1024 * 1024 ) ) );
1008 return "$prefix $mem ";
1009 }
1010
1011 /**
1012 * Send a line giving PHP memory usage.
1013 *
1014 * @param $exact Bool: print exact values instead of kilobytes (default: false)
1015 */
1016 function wfDebugMem( $exact = false ) {
1017 $mem = memory_get_usage();
1018 if( !$exact ) {
1019 $mem = floor( $mem / 1024 ) . ' kilobytes';
1020 } else {
1021 $mem .= ' bytes';
1022 }
1023 wfDebug( "Memory usage: $mem\n" );
1024 }
1025
1026 /**
1027 * Send a line to a supplementary debug log file, if configured, or main debug log if not.
1028 * $wgDebugLogGroups[$logGroup] should be set to a filename to send to a separate log.
1029 *
1030 * @param $logGroup String
1031 * @param $text String
1032 * @param $public Bool: whether to log the event in the public log if no private
1033 * log file is specified, (default true)
1034 */
1035 function wfDebugLog( $logGroup, $text, $public = true ) {
1036 global $wgDebugLogGroups;
1037 $text = trim( $text ) . "\n";
1038 if( isset( $wgDebugLogGroups[$logGroup] ) ) {
1039 $time = wfTimestamp( TS_DB );
1040 $wiki = wfWikiID();
1041 $host = wfHostname();
1042 if ( wfRunHooks( 'Debug', array( $text, $logGroup ) ) ) {
1043 wfErrorLog( "$time $host $wiki: $text", $wgDebugLogGroups[$logGroup] );
1044 }
1045 } elseif ( $public === true ) {
1046 wfDebug( $text, true );
1047 }
1048 }
1049
1050 /**
1051 * Log for database errors
1052 *
1053 * @param $text String: database error message.
1054 */
1055 function wfLogDBError( $text ) {
1056 global $wgDBerrorLog;
1057 if ( $wgDBerrorLog ) {
1058 $host = wfHostname();
1059 $wiki = wfWikiID();
1060 $text = date( 'D M j G:i:s T Y' ) . "\t$host\t$wiki\t$text";
1061 wfErrorLog( $text, $wgDBerrorLog );
1062 }
1063 }
1064
1065 /**
1066 * Throws a warning that $function is deprecated
1067 *
1068 * @param $function String
1069 * @param $version String|bool: Version of MediaWiki that the function was deprecated in (Added in 1.19).
1070 * @param $component String|bool: Added in 1.19.
1071 * @param $callerOffset integer: How far up the callstack is the original
1072 * caller. 2 = function that called the function that called
1073 * wfDeprecated (Added in 1.20)
1074 *
1075 * @return null
1076 */
1077 function wfDeprecated( $function, $version = false, $component = false, $callerOffset = 2 ) {
1078 static $functionsWarned = array();
1079
1080 MWDebug::deprecated( $function, $version, $component );
1081
1082 if ( !isset( $functionsWarned[$function] ) ) {
1083 $functionsWarned[$function] = true;
1084
1085 if ( $version ) {
1086 global $wgDeprecationReleaseLimit;
1087
1088 if ( $wgDeprecationReleaseLimit && $component === false ) {
1089 # Strip -* off the end of $version so that branches can use the
1090 # format #.##-branchname to avoid issues if the branch is merged into
1091 # a version of MediaWiki later than what it was branched from
1092 $comparableVersion = preg_replace( '/-.*$/', '', $version );
1093
1094 # If the comparableVersion is larger than our release limit then
1095 # skip the warning message for the deprecation
1096 if ( version_compare( $wgDeprecationReleaseLimit, $comparableVersion, '<' ) ) {
1097 return;
1098 }
1099 }
1100
1101 $component = $component === false ? 'MediaWiki' : $component;
1102 wfWarn( "Use of $function was deprecated in $component $version.", $callerOffset );
1103 } else {
1104 wfWarn( "Use of $function is deprecated.", $callerOffset );
1105 }
1106 }
1107 }
1108
1109 /**
1110 * Send a warning either to the debug log or in a PHP error depending on
1111 * $wgDevelopmentWarnings
1112 *
1113 * @param $msg String: message to send
1114 * @param $callerOffset Integer: number of items to go back in the backtrace to
1115 * find the correct caller (1 = function calling wfWarn, ...)
1116 * @param $level Integer: PHP error level; only used when $wgDevelopmentWarnings
1117 * is true
1118 */
1119 function wfWarn( $msg, $callerOffset = 1, $level = E_USER_NOTICE ) {
1120 global $wgDevelopmentWarnings;
1121
1122 MWDebug::warning( $msg, $callerOffset + 2 );
1123
1124 $callers = wfDebugBacktrace();
1125 if ( isset( $callers[$callerOffset + 1] ) ) {
1126 $callerfunc = $callers[$callerOffset + 1];
1127 $callerfile = $callers[$callerOffset];
1128 if ( isset( $callerfile['file'] ) && isset( $callerfile['line'] ) ) {
1129 $file = $callerfile['file'] . ' at line ' . $callerfile['line'];
1130 } else {
1131 $file = '(internal function)';
1132 }
1133 $func = '';
1134 if ( isset( $callerfunc['class'] ) ) {
1135 $func .= $callerfunc['class'] . '::';
1136 }
1137 if ( isset( $callerfunc['function'] ) ) {
1138 $func .= $callerfunc['function'];
1139 }
1140 $msg .= " [Called from $func in $file]";
1141 }
1142
1143 if ( $wgDevelopmentWarnings ) {
1144 trigger_error( $msg, $level );
1145 } else {
1146 wfDebug( "$msg\n" );
1147 }
1148 }
1149
1150 /**
1151 * Log to a file without getting "file size exceeded" signals.
1152 *
1153 * Can also log to TCP or UDP with the syntax udp://host:port/prefix. This will
1154 * send lines to the specified port, prefixed by the specified prefix and a space.
1155 *
1156 * @param $text String
1157 * @param $file String filename
1158 */
1159 function wfErrorLog( $text, $file ) {
1160 if ( substr( $file, 0, 4 ) == 'udp:' ) {
1161 # Needs the sockets extension
1162 if ( preg_match( '!^(tcp|udp):(?://)?\[([0-9a-fA-F:]+)\]:(\d+)(?:/(.*))?$!', $file, $m ) ) {
1163 // IPv6 bracketed host
1164 $host = $m[2];
1165 $port = intval( $m[3] );
1166 $prefix = isset( $m[4] ) ? $m[4] : false;
1167 $domain = AF_INET6;
1168 } elseif ( preg_match( '!^(tcp|udp):(?://)?([a-zA-Z0-9.-]+):(\d+)(?:/(.*))?$!', $file, $m ) ) {
1169 $host = $m[2];
1170 if ( !IP::isIPv4( $host ) ) {
1171 $host = gethostbyname( $host );
1172 }
1173 $port = intval( $m[3] );
1174 $prefix = isset( $m[4] ) ? $m[4] : false;
1175 $domain = AF_INET;
1176 } else {
1177 throw new MWException( __METHOD__ . ': Invalid UDP specification' );
1178 }
1179
1180 // Clean it up for the multiplexer
1181 if ( strval( $prefix ) !== '' ) {
1182 $text = preg_replace( '/^/m', $prefix . ' ', $text );
1183
1184 // Limit to 64KB
1185 if ( strlen( $text ) > 65506 ) {
1186 $text = substr( $text, 0, 65506 );
1187 }
1188
1189 if ( substr( $text, -1 ) != "\n" ) {
1190 $text .= "\n";
1191 }
1192 } elseif ( strlen( $text ) > 65507 ) {
1193 $text = substr( $text, 0, 65507 );
1194 }
1195
1196 $sock = socket_create( $domain, SOCK_DGRAM, SOL_UDP );
1197 if ( !$sock ) {
1198 return;
1199 }
1200
1201 socket_sendto( $sock, $text, strlen( $text ), 0, $host, $port );
1202 socket_close( $sock );
1203 } else {
1204 wfSuppressWarnings();
1205 $exists = file_exists( $file );
1206 $size = $exists ? filesize( $file ) : false;
1207 if ( !$exists || ( $size !== false && $size + strlen( $text ) < 0x7fffffff ) ) {
1208 file_put_contents( $file, $text, FILE_APPEND );
1209 }
1210 wfRestoreWarnings();
1211 }
1212 }
1213
1214 /**
1215 * @todo document
1216 */
1217 function wfLogProfilingData() {
1218 global $wgRequestTime, $wgDebugLogFile, $wgDebugRawPage, $wgRequest;
1219 global $wgProfileLimit, $wgUser;
1220
1221 $profiler = Profiler::instance();
1222
1223 # Profiling must actually be enabled...
1224 if ( $profiler->isStub() ) {
1225 return;
1226 }
1227
1228 // Get total page request time and only show pages that longer than
1229 // $wgProfileLimit time (default is 0)
1230 $elapsed = microtime( true ) - $wgRequestTime;
1231 if ( $elapsed <= $wgProfileLimit ) {
1232 return;
1233 }
1234
1235 $profiler->logData();
1236
1237 // Check whether this should be logged in the debug file.
1238 if ( $wgDebugLogFile == '' || ( !$wgDebugRawPage && wfIsDebugRawPage() ) ) {
1239 return;
1240 }
1241
1242 $forward = '';
1243 if ( !empty( $_SERVER['HTTP_X_FORWARDED_FOR'] ) ) {
1244 $forward = ' forwarded for ' . $_SERVER['HTTP_X_FORWARDED_FOR'];
1245 }
1246 if ( !empty( $_SERVER['HTTP_CLIENT_IP'] ) ) {
1247 $forward .= ' client IP ' . $_SERVER['HTTP_CLIENT_IP'];
1248 }
1249 if ( !empty( $_SERVER['HTTP_FROM'] ) ) {
1250 $forward .= ' from ' . $_SERVER['HTTP_FROM'];
1251 }
1252 if ( $forward ) {
1253 $forward = "\t(proxied via {$_SERVER['REMOTE_ADDR']}{$forward})";
1254 }
1255 // Don't load $wgUser at this late stage just for statistics purposes
1256 // @todo FIXME: We can detect some anons even if it is not loaded. See User::getId()
1257 if ( $wgUser->isItemLoaded( 'id' ) && $wgUser->isAnon() ) {
1258 $forward .= ' anon';
1259 }
1260 $log = sprintf( "%s\t%04.3f\t%s\n",
1261 gmdate( 'YmdHis' ), $elapsed,
1262 urldecode( $wgRequest->getRequestURL() . $forward ) );
1263
1264 wfErrorLog( $log . $profiler->getOutput(), $wgDebugLogFile );
1265 }
1266
1267 /**
1268 * Increment a statistics counter
1269 *
1270 * @param $key String
1271 * @param $count Int
1272 */
1273 function wfIncrStats( $key, $count = 1 ) {
1274 global $wgStatsMethod;
1275
1276 $count = intval( $count );
1277
1278 if( $wgStatsMethod == 'udp' ) {
1279 global $wgUDPProfilerHost, $wgUDPProfilerPort, $wgDBname, $wgAggregateStatsID;
1280 static $socket;
1281
1282 $id = $wgAggregateStatsID !== false ? $wgAggregateStatsID : $wgDBname;
1283
1284 if ( !$socket ) {
1285 $socket = socket_create( AF_INET, SOCK_DGRAM, SOL_UDP );
1286 $statline = "stats/{$id} - 1 1 1 1 1 -total\n";
1287 socket_sendto(
1288 $socket,
1289 $statline,
1290 strlen( $statline ),
1291 0,
1292 $wgUDPProfilerHost,
1293 $wgUDPProfilerPort
1294 );
1295 }
1296 $statline = "stats/{$id} - {$count} 1 1 1 1 {$key}\n";
1297 wfSuppressWarnings();
1298 socket_sendto(
1299 $socket,
1300 $statline,
1301 strlen( $statline ),
1302 0,
1303 $wgUDPProfilerHost,
1304 $wgUDPProfilerPort
1305 );
1306 wfRestoreWarnings();
1307 } elseif( $wgStatsMethod == 'cache' ) {
1308 global $wgMemc;
1309 $key = wfMemcKey( 'stats', $key );
1310 if ( is_null( $wgMemc->incr( $key, $count ) ) ) {
1311 $wgMemc->add( $key, $count );
1312 }
1313 } else {
1314 // Disabled
1315 }
1316 }
1317
1318 /**
1319 * Check if the wiki read-only lock file is present. This can be used to lock
1320 * off editing functions, but doesn't guarantee that the database will not be
1321 * modified.
1322 *
1323 * @return bool
1324 */
1325 function wfReadOnly() {
1326 global $wgReadOnlyFile, $wgReadOnly;
1327
1328 if ( !is_null( $wgReadOnly ) ) {
1329 return (bool)$wgReadOnly;
1330 }
1331 if ( $wgReadOnlyFile == '' ) {
1332 return false;
1333 }
1334 // Set $wgReadOnly for faster access next time
1335 if ( is_file( $wgReadOnlyFile ) ) {
1336 $wgReadOnly = file_get_contents( $wgReadOnlyFile );
1337 } else {
1338 $wgReadOnly = false;
1339 }
1340 return (bool)$wgReadOnly;
1341 }
1342
1343 /**
1344 * @return bool
1345 */
1346 function wfReadOnlyReason() {
1347 global $wgReadOnly;
1348 wfReadOnly();
1349 return $wgReadOnly;
1350 }
1351
1352 /**
1353 * Return a Language object from $langcode
1354 *
1355 * @param $langcode Mixed: either:
1356 * - a Language object
1357 * - code of the language to get the message for, if it is
1358 * a valid code create a language for that language, if
1359 * it is a string but not a valid code then make a basic
1360 * language object
1361 * - a boolean: if it's false then use the global object for
1362 * the current user's language (as a fallback for the old parameter
1363 * functionality), or if it is true then use global object
1364 * for the wiki's content language.
1365 * @return Language object
1366 */
1367 function wfGetLangObj( $langcode = false ) {
1368 # Identify which language to get or create a language object for.
1369 # Using is_object here due to Stub objects.
1370 if( is_object( $langcode ) ) {
1371 # Great, we already have the object (hopefully)!
1372 return $langcode;
1373 }
1374
1375 global $wgContLang, $wgLanguageCode;
1376 if( $langcode === true || $langcode === $wgLanguageCode ) {
1377 # $langcode is the language code of the wikis content language object.
1378 # or it is a boolean and value is true
1379 return $wgContLang;
1380 }
1381
1382 global $wgLang;
1383 if( $langcode === false || $langcode === $wgLang->getCode() ) {
1384 # $langcode is the language code of user language object.
1385 # or it was a boolean and value is false
1386 return $wgLang;
1387 }
1388
1389 $validCodes = array_keys( Language::fetchLanguageNames() );
1390 if( in_array( $langcode, $validCodes ) ) {
1391 # $langcode corresponds to a valid language.
1392 return Language::factory( $langcode );
1393 }
1394
1395 # $langcode is a string, but not a valid language code; use content language.
1396 wfDebug( "Invalid language code passed to wfGetLangObj, falling back to content language.\n" );
1397 return $wgContLang;
1398 }
1399
1400 /**
1401 * Old function when $wgBetterDirectionality existed
1402 * All usage removed, wfUILang can be removed in near future
1403 *
1404 * @deprecated since 1.18
1405 * @return Language
1406 */
1407 function wfUILang() {
1408 wfDeprecated( __METHOD__, '1.18' );
1409 global $wgLang;
1410 return $wgLang;
1411 }
1412
1413 /**
1414 * This is the new function for getting translated interface messages.
1415 * See the Message class for documentation how to use them.
1416 * The intention is that this function replaces all old wfMsg* functions.
1417 * @param $key \string Message key.
1418 * Varargs: normal message parameters.
1419 * @return Message
1420 * @since 1.17
1421 */
1422 function wfMessage( $key /*...*/) {
1423 $params = func_get_args();
1424 array_shift( $params );
1425 if ( isset( $params[0] ) && is_array( $params[0] ) ) {
1426 $params = $params[0];
1427 }
1428 return new Message( $key, $params );
1429 }
1430
1431 /**
1432 * This function accepts multiple message keys and returns a message instance
1433 * for the first message which is non-empty. If all messages are empty then an
1434 * instance of the first message key is returned.
1435 * @param varargs: message keys
1436 * @return Message
1437 * @since 1.18
1438 */
1439 function wfMessageFallback( /*...*/ ) {
1440 $args = func_get_args();
1441 return MWFunction::callArray( 'Message::newFallbackSequence', $args );
1442 }
1443
1444 /**
1445 * Get a message from anywhere, for the current user language.
1446 *
1447 * Use wfMsgForContent() instead if the message should NOT
1448 * change depending on the user preferences.
1449 *
1450 * @param $key String: lookup key for the message, usually
1451 * defined in languages/Language.php
1452 *
1453 * Parameters to the message, which can be used to insert variable text into
1454 * it, can be passed to this function in the following formats:
1455 * - One per argument, starting at the second parameter
1456 * - As an array in the second parameter
1457 * These are not shown in the function definition.
1458 *
1459 * @return String
1460 */
1461 function wfMsg( $key ) {
1462 $args = func_get_args();
1463 array_shift( $args );
1464 return wfMsgReal( $key, $args );
1465 }
1466
1467 /**
1468 * Same as above except doesn't transform the message
1469 *
1470 * @param $key String
1471 * @return String
1472 */
1473 function wfMsgNoTrans( $key ) {
1474 $args = func_get_args();
1475 array_shift( $args );
1476 return wfMsgReal( $key, $args, true, false, false );
1477 }
1478
1479 /**
1480 * Get a message from anywhere, for the current global language
1481 * set with $wgLanguageCode.
1482 *
1483 * Use this if the message should NOT change dependent on the
1484 * language set in the user's preferences. This is the case for
1485 * most text written into logs, as well as link targets (such as
1486 * the name of the copyright policy page). Link titles, on the
1487 * other hand, should be shown in the UI language.
1488 *
1489 * Note that MediaWiki allows users to change the user interface
1490 * language in their preferences, but a single installation
1491 * typically only contains content in one language.
1492 *
1493 * Be wary of this distinction: If you use wfMsg() where you should
1494 * use wfMsgForContent(), a user of the software may have to
1495 * customize potentially hundreds of messages in
1496 * order to, e.g., fix a link in every possible language.
1497 *
1498 * @param $key String: lookup key for the message, usually
1499 * defined in languages/Language.php
1500 * @return String
1501 */
1502 function wfMsgForContent( $key ) {
1503 global $wgForceUIMsgAsContentMsg;
1504 $args = func_get_args();
1505 array_shift( $args );
1506 $forcontent = true;
1507 if( is_array( $wgForceUIMsgAsContentMsg ) &&
1508 in_array( $key, $wgForceUIMsgAsContentMsg ) )
1509 {
1510 $forcontent = false;
1511 }
1512 return wfMsgReal( $key, $args, true, $forcontent );
1513 }
1514
1515 /**
1516 * Same as above except doesn't transform the message
1517 *
1518 * @param $key String
1519 * @return String
1520 */
1521 function wfMsgForContentNoTrans( $key ) {
1522 global $wgForceUIMsgAsContentMsg;
1523 $args = func_get_args();
1524 array_shift( $args );
1525 $forcontent = true;
1526 if( is_array( $wgForceUIMsgAsContentMsg ) &&
1527 in_array( $key, $wgForceUIMsgAsContentMsg ) )
1528 {
1529 $forcontent = false;
1530 }
1531 return wfMsgReal( $key, $args, true, $forcontent, false );
1532 }
1533
1534 /**
1535 * Really get a message
1536 *
1537 * @param $key String: key to get.
1538 * @param $args
1539 * @param $useDB Boolean
1540 * @param $forContent Mixed: Language code, or false for user lang, true for content lang.
1541 * @param $transform Boolean: Whether or not to transform the message.
1542 * @return String: the requested message.
1543 */
1544 function wfMsgReal( $key, $args, $useDB = true, $forContent = false, $transform = true ) {
1545 wfProfileIn( __METHOD__ );
1546 $message = wfMsgGetKey( $key, $useDB, $forContent, $transform );
1547 $message = wfMsgReplaceArgs( $message, $args );
1548 wfProfileOut( __METHOD__ );
1549 return $message;
1550 }
1551
1552 /**
1553 * Fetch a message string value, but don't replace any keys yet.
1554 *
1555 * @param $key String
1556 * @param $useDB Bool
1557 * @param $langCode String: Code of the language to get the message for, or
1558 * behaves as a content language switch if it is a boolean.
1559 * @param $transform Boolean: whether to parse magic words, etc.
1560 * @return string
1561 */
1562 function wfMsgGetKey( $key, $useDB = true, $langCode = false, $transform = true ) {
1563 wfRunHooks( 'NormalizeMessageKey', array( &$key, &$useDB, &$langCode, &$transform ) );
1564
1565 $cache = MessageCache::singleton();
1566 $message = $cache->get( $key, $useDB, $langCode );
1567 if( $message === false ) {
1568 $message = '&lt;' . htmlspecialchars( $key ) . '&gt;';
1569 } elseif ( $transform ) {
1570 $message = $cache->transform( $message );
1571 }
1572 return $message;
1573 }
1574
1575 /**
1576 * Replace message parameter keys on the given formatted output.
1577 *
1578 * @param $message String
1579 * @param $args Array
1580 * @return string
1581 * @private
1582 */
1583 function wfMsgReplaceArgs( $message, $args ) {
1584 # Fix windows line-endings
1585 # Some messages are split with explode("\n", $msg)
1586 $message = str_replace( "\r", '', $message );
1587
1588 // Replace arguments
1589 if ( count( $args ) ) {
1590 if ( is_array( $args[0] ) ) {
1591 $args = array_values( $args[0] );
1592 }
1593 $replacementKeys = array();
1594 foreach( $args as $n => $param ) {
1595 $replacementKeys['$' . ( $n + 1 )] = $param;
1596 }
1597 $message = strtr( $message, $replacementKeys );
1598 }
1599
1600 return $message;
1601 }
1602
1603 /**
1604 * Return an HTML-escaped version of a message.
1605 * Parameter replacements, if any, are done *after* the HTML-escaping,
1606 * so parameters may contain HTML (eg links or form controls). Be sure
1607 * to pre-escape them if you really do want plaintext, or just wrap
1608 * the whole thing in htmlspecialchars().
1609 *
1610 * @param $key String
1611 * @param string ... parameters
1612 * @return string
1613 */
1614 function wfMsgHtml( $key ) {
1615 $args = func_get_args();
1616 array_shift( $args );
1617 return wfMsgReplaceArgs( htmlspecialchars( wfMsgGetKey( $key ) ), $args );
1618 }
1619
1620 /**
1621 * Return an HTML version of message
1622 * Parameter replacements, if any, are done *after* parsing the wiki-text message,
1623 * so parameters may contain HTML (eg links or form controls). Be sure
1624 * to pre-escape them if you really do want plaintext, or just wrap
1625 * the whole thing in htmlspecialchars().
1626 *
1627 * @param $key String
1628 * @param string ... parameters
1629 * @return string
1630 */
1631 function wfMsgWikiHtml( $key ) {
1632 $args = func_get_args();
1633 array_shift( $args );
1634 return wfMsgReplaceArgs(
1635 MessageCache::singleton()->parse( wfMsgGetKey( $key ), null,
1636 /* can't be set to false */ true, /* interface */ true )->getText(),
1637 $args );
1638 }
1639
1640 /**
1641 * Returns message in the requested format
1642 * @param $key String: key of the message
1643 * @param $options Array: processing rules. Can take the following options:
1644 * <i>parse</i>: parses wikitext to HTML
1645 * <i>parseinline</i>: parses wikitext to HTML and removes the surrounding
1646 * p's added by parser or tidy
1647 * <i>escape</i>: filters message through htmlspecialchars
1648 * <i>escapenoentities</i>: same, but allows entity references like &#160; through
1649 * <i>replaceafter</i>: parameters are substituted after parsing or escaping
1650 * <i>parsemag</i>: transform the message using magic phrases
1651 * <i>content</i>: fetch message for content language instead of interface
1652 * Also can accept a single associative argument, of the form 'language' => 'xx':
1653 * <i>language</i>: Language object or language code to fetch message for
1654 * (overriden by <i>content</i>).
1655 * Behavior for conflicting options (e.g., parse+parseinline) is undefined.
1656 *
1657 * @return String
1658 */
1659 function wfMsgExt( $key, $options ) {
1660 $args = func_get_args();
1661 array_shift( $args );
1662 array_shift( $args );
1663 $options = (array)$options;
1664
1665 foreach( $options as $arrayKey => $option ) {
1666 if( !preg_match( '/^[0-9]+|language$/', $arrayKey ) ) {
1667 # An unknown index, neither numeric nor "language"
1668 wfWarn( "wfMsgExt called with incorrect parameter key $arrayKey", 1, E_USER_WARNING );
1669 } elseif( preg_match( '/^[0-9]+$/', $arrayKey ) && !in_array( $option,
1670 array( 'parse', 'parseinline', 'escape', 'escapenoentities',
1671 'replaceafter', 'parsemag', 'content' ) ) ) {
1672 # A numeric index with unknown value
1673 wfWarn( "wfMsgExt called with incorrect parameter $option", 1, E_USER_WARNING );
1674 }
1675 }
1676
1677 if( in_array( 'content', $options, true ) ) {
1678 $forContent = true;
1679 $langCode = true;
1680 $langCodeObj = null;
1681 } elseif( array_key_exists( 'language', $options ) ) {
1682 $forContent = false;
1683 $langCode = wfGetLangObj( $options['language'] );
1684 $langCodeObj = $langCode;
1685 } else {
1686 $forContent = false;
1687 $langCode = false;
1688 $langCodeObj = null;
1689 }
1690
1691 $string = wfMsgGetKey( $key, /*DB*/true, $langCode, /*Transform*/false );
1692
1693 if( !in_array( 'replaceafter', $options, true ) ) {
1694 $string = wfMsgReplaceArgs( $string, $args );
1695 }
1696
1697 $messageCache = MessageCache::singleton();
1698 $parseInline = in_array( 'parseinline', $options, true );
1699 if( in_array( 'parse', $options, true ) || $parseInline ) {
1700 $string = $messageCache->parse( $string, null, true, !$forContent, $langCodeObj );
1701 if ( $string instanceof ParserOutput ) {
1702 $string = $string->getText();
1703 }
1704
1705 if ( $parseInline ) {
1706 $m = array();
1707 if( preg_match( '/^<p>(.*)\n?<\/p>\n?$/sU', $string, $m ) ) {
1708 $string = $m[1];
1709 }
1710 }
1711 } elseif ( in_array( 'parsemag', $options, true ) ) {
1712 $string = $messageCache->transform( $string,
1713 !$forContent, $langCodeObj );
1714 }
1715
1716 if ( in_array( 'escape', $options, true ) ) {
1717 $string = htmlspecialchars ( $string );
1718 } elseif ( in_array( 'escapenoentities', $options, true ) ) {
1719 $string = Sanitizer::escapeHtmlAllowEntities( $string );
1720 }
1721
1722 if( in_array( 'replaceafter', $options, true ) ) {
1723 $string = wfMsgReplaceArgs( $string, $args );
1724 }
1725
1726 return $string;
1727 }
1728
1729 /**
1730 * Since wfMsg() and co suck, they don't return false if the message key they
1731 * looked up didn't exist but a XHTML string, this function checks for the
1732 * nonexistance of messages by checking the MessageCache::get() result directly.
1733 *
1734 * @param $key String: the message key looked up
1735 * @return Boolean True if the message *doesn't* exist.
1736 */
1737 function wfEmptyMsg( $key ) {
1738 return MessageCache::singleton()->get( $key, /*useDB*/true, /*content*/false ) === false;
1739 }
1740
1741 /**
1742 * Throw a debugging exception. This function previously once exited the process,
1743 * but now throws an exception instead, with similar results.
1744 *
1745 * @param $msg String: message shown when dying.
1746 */
1747 function wfDebugDieBacktrace( $msg = '' ) {
1748 throw new MWException( $msg );
1749 }
1750
1751 /**
1752 * Fetch server name for use in error reporting etc.
1753 * Use real server name if available, so we know which machine
1754 * in a server farm generated the current page.
1755 *
1756 * @return string
1757 */
1758 function wfHostname() {
1759 static $host;
1760 if ( is_null( $host ) ) {
1761 if ( function_exists( 'posix_uname' ) ) {
1762 // This function not present on Windows
1763 $uname = posix_uname();
1764 } else {
1765 $uname = false;
1766 }
1767 if( is_array( $uname ) && isset( $uname['nodename'] ) ) {
1768 $host = $uname['nodename'];
1769 } elseif ( getenv( 'COMPUTERNAME' ) ) {
1770 # Windows computer name
1771 $host = getenv( 'COMPUTERNAME' );
1772 } else {
1773 # This may be a virtual server.
1774 $host = $_SERVER['SERVER_NAME'];
1775 }
1776 }
1777 return $host;
1778 }
1779
1780 /**
1781 * Returns a HTML comment with the elapsed time since request.
1782 * This method has no side effects.
1783 *
1784 * @return string
1785 */
1786 function wfReportTime() {
1787 global $wgRequestTime, $wgShowHostnames;
1788
1789 $elapsed = microtime( true ) - $wgRequestTime;
1790
1791 return $wgShowHostnames
1792 ? sprintf( '<!-- Served by %s in %01.3f secs. -->', wfHostname(), $elapsed )
1793 : sprintf( '<!-- Served in %01.3f secs. -->', $elapsed );
1794 }
1795
1796 /**
1797 * Safety wrapper for debug_backtrace().
1798 *
1799 * With Zend Optimizer 3.2.0 loaded, this causes segfaults under somewhat
1800 * murky circumstances, which may be triggered in part by stub objects
1801 * or other fancy talkin'.
1802 *
1803 * Will return an empty array if Zend Optimizer is detected or if
1804 * debug_backtrace is disabled, otherwise the output from
1805 * debug_backtrace() (trimmed).
1806 *
1807 * @param $limit int This parameter can be used to limit the number of stack frames returned
1808 *
1809 * @return array of backtrace information
1810 */
1811 function wfDebugBacktrace( $limit = 0 ) {
1812 static $disabled = null;
1813
1814 if( extension_loaded( 'Zend Optimizer' ) ) {
1815 wfDebug( "Zend Optimizer detected; skipping debug_backtrace for safety.\n" );
1816 return array();
1817 }
1818
1819 if ( is_null( $disabled ) ) {
1820 $disabled = false;
1821 $functions = explode( ',', ini_get( 'disable_functions' ) );
1822 $functions = array_map( 'trim', $functions );
1823 $functions = array_map( 'strtolower', $functions );
1824 if ( in_array( 'debug_backtrace', $functions ) ) {
1825 wfDebug( "debug_backtrace is in disabled_functions\n" );
1826 $disabled = true;
1827 }
1828 }
1829 if ( $disabled ) {
1830 return array();
1831 }
1832
1833 if ( $limit && version_compare( PHP_VERSION, '5.4.0', '>=' ) ) {
1834 return array_slice( debug_backtrace( DEBUG_BACKTRACE_PROVIDE_OBJECT, $limit + 1 ), 1 );
1835 } else {
1836 return array_slice( debug_backtrace(), 1 );
1837 }
1838 }
1839
1840 /**
1841 * Get a debug backtrace as a string
1842 *
1843 * @return string
1844 */
1845 function wfBacktrace() {
1846 global $wgCommandLineMode;
1847
1848 if ( $wgCommandLineMode ) {
1849 $msg = '';
1850 } else {
1851 $msg = "<ul>\n";
1852 }
1853 $backtrace = wfDebugBacktrace();
1854 foreach( $backtrace as $call ) {
1855 if( isset( $call['file'] ) ) {
1856 $f = explode( DIRECTORY_SEPARATOR, $call['file'] );
1857 $file = $f[count( $f ) - 1];
1858 } else {
1859 $file = '-';
1860 }
1861 if( isset( $call['line'] ) ) {
1862 $line = $call['line'];
1863 } else {
1864 $line = '-';
1865 }
1866 if ( $wgCommandLineMode ) {
1867 $msg .= "$file line $line calls ";
1868 } else {
1869 $msg .= '<li>' . $file . ' line ' . $line . ' calls ';
1870 }
1871 if( !empty( $call['class'] ) ) {
1872 $msg .= $call['class'] . $call['type'];
1873 }
1874 $msg .= $call['function'] . '()';
1875
1876 if ( $wgCommandLineMode ) {
1877 $msg .= "\n";
1878 } else {
1879 $msg .= "</li>\n";
1880 }
1881 }
1882 if ( $wgCommandLineMode ) {
1883 $msg .= "\n";
1884 } else {
1885 $msg .= "</ul>\n";
1886 }
1887
1888 return $msg;
1889 }
1890
1891 /**
1892 * Get the name of the function which called this function
1893 * wfGetCaller( 1 ) is the function with the wfGetCaller() call (ie. __FUNCTION__)
1894 * wfGetCaller( 2 ) [default] is the caller of the function running wfGetCaller()
1895 * wfGetCaller( 3 ) is the parent of that.
1896 *
1897 * @param $level Int
1898 * @return Bool|string
1899 */
1900 function wfGetCaller( $level = 2 ) {
1901 $backtrace = wfDebugBacktrace( $level + 1 );
1902 if ( isset( $backtrace[$level] ) ) {
1903 return wfFormatStackFrame( $backtrace[$level] );
1904 } else {
1905 $caller = 'unknown';
1906 }
1907 return $caller;
1908 }
1909
1910 /**
1911 * Return a string consisting of callers in the stack. Useful sometimes
1912 * for profiling specific points.
1913 *
1914 * @param $limit int The maximum depth of the stack frame to return, or false for
1915 * the entire stack.
1916 * @return String
1917 */
1918 function wfGetAllCallers( $limit = 3 ) {
1919 $trace = array_reverse( wfDebugBacktrace() );
1920 if ( !$limit || $limit > count( $trace ) - 1 ) {
1921 $limit = count( $trace ) - 1;
1922 }
1923 $trace = array_slice( $trace, -$limit - 1, $limit );
1924 return implode( '/', array_map( 'wfFormatStackFrame', $trace ) );
1925 }
1926
1927 /**
1928 * Return a string representation of frame
1929 *
1930 * @param $frame Array
1931 * @return Bool
1932 */
1933 function wfFormatStackFrame( $frame ) {
1934 return isset( $frame['class'] ) ?
1935 $frame['class'] . '::' . $frame['function'] :
1936 $frame['function'];
1937 }
1938
1939
1940 /* Some generic result counters, pulled out of SearchEngine */
1941
1942
1943 /**
1944 * @todo document
1945 *
1946 * @param $offset Int
1947 * @param $limit Int
1948 * @return String
1949 */
1950 function wfShowingResults( $offset, $limit ) {
1951 global $wgLang;
1952 return wfMsgExt(
1953 'showingresults',
1954 array( 'parseinline' ),
1955 $wgLang->formatNum( $limit ),
1956 $wgLang->formatNum( $offset + 1 )
1957 );
1958 }
1959
1960 /**
1961 * Generate (prev x| next x) (20|50|100...) type links for paging
1962 *
1963 * @param $offset String
1964 * @param $limit Integer
1965 * @param $link String
1966 * @param $query String: optional URL query parameter string
1967 * @param $atend Bool: optional param for specified if this is the last page
1968 * @return String
1969 * @deprecated in 1.19; use Language::viewPrevNext() instead
1970 */
1971 function wfViewPrevNext( $offset, $limit, $link, $query = '', $atend = false ) {
1972 wfDeprecated( __METHOD__, '1.19' );
1973
1974 global $wgLang;
1975
1976 $query = wfCgiToArray( $query );
1977
1978 if( is_object( $link ) ) {
1979 $title = $link;
1980 } else {
1981 $title = Title::newFromText( $link );
1982 if( is_null( $title ) ) {
1983 return false;
1984 }
1985 }
1986
1987 return $wgLang->viewPrevNext( $title, $offset, $limit, $query, $atend );
1988 }
1989
1990 /**
1991 * Make a list item, used by various special pages
1992 *
1993 * @param $page String Page link
1994 * @param $details String Text between brackets
1995 * @param $oppositedm Boolean Add the direction mark opposite to your
1996 * language, to display text properly
1997 * @return String
1998 * @deprecated since 1.19; use Language::specialList() instead
1999 */
2000 function wfSpecialList( $page, $details, $oppositedm = true ) {
2001 wfDeprecated( __METHOD__, '1.19' );
2002
2003 global $wgLang;
2004 return $wgLang->specialList( $page, $details, $oppositedm );
2005 }
2006
2007 /**
2008 * @todo document
2009 * @todo FIXME: We may want to blacklist some broken browsers
2010 *
2011 * @param $force Bool
2012 * @return bool Whereas client accept gzip compression
2013 */
2014 function wfClientAcceptsGzip( $force = false ) {
2015 static $result = null;
2016 if ( $result === null || $force ) {
2017 $result = false;
2018 if( isset( $_SERVER['HTTP_ACCEPT_ENCODING'] ) ) {
2019 # @todo FIXME: We may want to blacklist some broken browsers
2020 $m = array();
2021 if( preg_match(
2022 '/\bgzip(?:;(q)=([0-9]+(?:\.[0-9]+)))?\b/',
2023 $_SERVER['HTTP_ACCEPT_ENCODING'],
2024 $m )
2025 )
2026 {
2027 if( isset( $m[2] ) && ( $m[1] == 'q' ) && ( $m[2] == 0 ) ) {
2028 $result = false;
2029 return $result;
2030 }
2031 wfDebug( "wfClientAcceptsGzip: client accepts gzip.\n" );
2032 $result = true;
2033 }
2034 }
2035 }
2036 return $result;
2037 }
2038
2039 /**
2040 * Obtain the offset and limit values from the request string;
2041 * used in special pages
2042 *
2043 * @param $deflimit Int default limit if none supplied
2044 * @param $optionname String Name of a user preference to check against
2045 * @return array
2046 *
2047 */
2048 function wfCheckLimits( $deflimit = 50, $optionname = 'rclimit' ) {
2049 global $wgRequest;
2050 return $wgRequest->getLimitOffset( $deflimit, $optionname );
2051 }
2052
2053 /**
2054 * Escapes the given text so that it may be output using addWikiText()
2055 * without any linking, formatting, etc. making its way through. This
2056 * is achieved by substituting certain characters with HTML entities.
2057 * As required by the callers, <nowiki> is not used.
2058 *
2059 * @param $text String: text to be escaped
2060 * @return String
2061 */
2062 function wfEscapeWikiText( $text ) {
2063 $text = strtr( "\n$text", array(
2064 '"' => '&#34;', '&' => '&#38;', "'" => '&#39;', '<' => '&#60;',
2065 '=' => '&#61;', '>' => '&#62;', '[' => '&#91;', ']' => '&#93;',
2066 '{' => '&#123;', '|' => '&#124;', '}' => '&#125;',
2067 "\n#" => "\n&#35;", "\n*" => "\n&#42;",
2068 "\n:" => "\n&#58;", "\n;" => "\n&#59;",
2069 '://' => '&#58;//', 'ISBN ' => 'ISBN&#32;', 'RFC ' => 'RFC&#32;',
2070 ) );
2071 return substr( $text, 1 );
2072 }
2073
2074 /**
2075 * Get the current unix timetstamp with microseconds. Useful for profiling
2076 * @return Float
2077 */
2078 function wfTime() {
2079 return microtime( true );
2080 }
2081
2082 /**
2083 * Sets dest to source and returns the original value of dest
2084 * If source is NULL, it just returns the value, it doesn't set the variable
2085 * If force is true, it will set the value even if source is NULL
2086 *
2087 * @param $dest Mixed
2088 * @param $source Mixed
2089 * @param $force Bool
2090 * @return Mixed
2091 */
2092 function wfSetVar( &$dest, $source, $force = false ) {
2093 $temp = $dest;
2094 if ( !is_null( $source ) || $force ) {
2095 $dest = $source;
2096 }
2097 return $temp;
2098 }
2099
2100 /**
2101 * As for wfSetVar except setting a bit
2102 *
2103 * @param $dest Int
2104 * @param $bit Int
2105 * @param $state Bool
2106 *
2107 * @return bool
2108 */
2109 function wfSetBit( &$dest, $bit, $state = true ) {
2110 $temp = (bool)( $dest & $bit );
2111 if ( !is_null( $state ) ) {
2112 if ( $state ) {
2113 $dest |= $bit;
2114 } else {
2115 $dest &= ~$bit;
2116 }
2117 }
2118 return $temp;
2119 }
2120
2121 /**
2122 * A wrapper around the PHP function var_export().
2123 * Either print it or add it to the regular output ($wgOut).
2124 *
2125 * @param $var mixed A PHP variable to dump.
2126 */
2127 function wfVarDump( $var ) {
2128 global $wgOut;
2129 $s = str_replace( "\n", "<br />\n", var_export( $var, true ) . "\n" );
2130 if ( headers_sent() || !isset( $wgOut ) || !is_object( $wgOut ) ) {
2131 print $s;
2132 } else {
2133 $wgOut->addHTML( $s );
2134 }
2135 }
2136
2137 /**
2138 * Provide a simple HTTP error.
2139 *
2140 * @param $code Int|String
2141 * @param $label String
2142 * @param $desc String
2143 */
2144 function wfHttpError( $code, $label, $desc ) {
2145 global $wgOut;
2146 $wgOut->disable();
2147 header( "HTTP/1.0 $code $label" );
2148 header( "Status: $code $label" );
2149 $wgOut->sendCacheControl();
2150
2151 header( 'Content-type: text/html; charset=utf-8' );
2152 print "<!doctype html>" .
2153 '<html><head><title>' .
2154 htmlspecialchars( $label ) .
2155 '</title></head><body><h1>' .
2156 htmlspecialchars( $label ) .
2157 '</h1><p>' .
2158 nl2br( htmlspecialchars( $desc ) ) .
2159 "</p></body></html>\n";
2160 }
2161
2162 /**
2163 * Clear away any user-level output buffers, discarding contents.
2164 *
2165 * Suitable for 'starting afresh', for instance when streaming
2166 * relatively large amounts of data without buffering, or wanting to
2167 * output image files without ob_gzhandler's compression.
2168 *
2169 * The optional $resetGzipEncoding parameter controls suppression of
2170 * the Content-Encoding header sent by ob_gzhandler; by default it
2171 * is left. See comments for wfClearOutputBuffers() for why it would
2172 * be used.
2173 *
2174 * Note that some PHP configuration options may add output buffer
2175 * layers which cannot be removed; these are left in place.
2176 *
2177 * @param $resetGzipEncoding Bool
2178 */
2179 function wfResetOutputBuffers( $resetGzipEncoding = true ) {
2180 if( $resetGzipEncoding ) {
2181 // Suppress Content-Encoding and Content-Length
2182 // headers from 1.10+s wfOutputHandler
2183 global $wgDisableOutputCompression;
2184 $wgDisableOutputCompression = true;
2185 }
2186 while( $status = ob_get_status() ) {
2187 if( $status['type'] == 0 /* PHP_OUTPUT_HANDLER_INTERNAL */ ) {
2188 // Probably from zlib.output_compression or other
2189 // PHP-internal setting which can't be removed.
2190 //
2191 // Give up, and hope the result doesn't break
2192 // output behavior.
2193 break;
2194 }
2195 if( !ob_end_clean() ) {
2196 // Could not remove output buffer handler; abort now
2197 // to avoid getting in some kind of infinite loop.
2198 break;
2199 }
2200 if( $resetGzipEncoding ) {
2201 if( $status['name'] == 'ob_gzhandler' ) {
2202 // Reset the 'Content-Encoding' field set by this handler
2203 // so we can start fresh.
2204 header_remove( 'Content-Encoding' );
2205 break;
2206 }
2207 }
2208 }
2209 }
2210
2211 /**
2212 * More legible than passing a 'false' parameter to wfResetOutputBuffers():
2213 *
2214 * Clear away output buffers, but keep the Content-Encoding header
2215 * produced by ob_gzhandler, if any.
2216 *
2217 * This should be used for HTTP 304 responses, where you need to
2218 * preserve the Content-Encoding header of the real result, but
2219 * also need to suppress the output of ob_gzhandler to keep to spec
2220 * and avoid breaking Firefox in rare cases where the headers and
2221 * body are broken over two packets.
2222 */
2223 function wfClearOutputBuffers() {
2224 wfResetOutputBuffers( false );
2225 }
2226
2227 /**
2228 * Converts an Accept-* header into an array mapping string values to quality
2229 * factors
2230 *
2231 * @param $accept String
2232 * @param $def String default
2233 * @return Array
2234 */
2235 function wfAcceptToPrefs( $accept, $def = '*/*' ) {
2236 # No arg means accept anything (per HTTP spec)
2237 if( !$accept ) {
2238 return array( $def => 1.0 );
2239 }
2240
2241 $prefs = array();
2242
2243 $parts = explode( ',', $accept );
2244
2245 foreach( $parts as $part ) {
2246 # @todo FIXME: Doesn't deal with params like 'text/html; level=1'
2247 $values = explode( ';', trim( $part ) );
2248 $match = array();
2249 if ( count( $values ) == 1 ) {
2250 $prefs[$values[0]] = 1.0;
2251 } elseif ( preg_match( '/q\s*=\s*(\d*\.\d+)/', $values[1], $match ) ) {
2252 $prefs[$values[0]] = floatval( $match[1] );
2253 }
2254 }
2255
2256 return $prefs;
2257 }
2258
2259 /**
2260 * Checks if a given MIME type matches any of the keys in the given
2261 * array. Basic wildcards are accepted in the array keys.
2262 *
2263 * Returns the matching MIME type (or wildcard) if a match, otherwise
2264 * NULL if no match.
2265 *
2266 * @param $type String
2267 * @param $avail Array
2268 * @return string
2269 * @private
2270 */
2271 function mimeTypeMatch( $type, $avail ) {
2272 if( array_key_exists( $type, $avail ) ) {
2273 return $type;
2274 } else {
2275 $parts = explode( '/', $type );
2276 if( array_key_exists( $parts[0] . '/*', $avail ) ) {
2277 return $parts[0] . '/*';
2278 } elseif( array_key_exists( '*/*', $avail ) ) {
2279 return '*/*';
2280 } else {
2281 return null;
2282 }
2283 }
2284 }
2285
2286 /**
2287 * Returns the 'best' match between a client's requested internet media types
2288 * and the server's list of available types. Each list should be an associative
2289 * array of type to preference (preference is a float between 0.0 and 1.0).
2290 * Wildcards in the types are acceptable.
2291 *
2292 * @param $cprefs Array: client's acceptable type list
2293 * @param $sprefs Array: server's offered types
2294 * @return string
2295 *
2296 * @todo FIXME: Doesn't handle params like 'text/plain; charset=UTF-8'
2297 * XXX: generalize to negotiate other stuff
2298 */
2299 function wfNegotiateType( $cprefs, $sprefs ) {
2300 $combine = array();
2301
2302 foreach( array_keys( $sprefs ) as $type ) {
2303 $parts = explode( '/', $type );
2304 if( $parts[1] != '*' ) {
2305 $ckey = mimeTypeMatch( $type, $cprefs );
2306 if( $ckey ) {
2307 $combine[$type] = $sprefs[$type] * $cprefs[$ckey];
2308 }
2309 }
2310 }
2311
2312 foreach( array_keys( $cprefs ) as $type ) {
2313 $parts = explode( '/', $type );
2314 if( $parts[1] != '*' && !array_key_exists( $type, $sprefs ) ) {
2315 $skey = mimeTypeMatch( $type, $sprefs );
2316 if( $skey ) {
2317 $combine[$type] = $sprefs[$skey] * $cprefs[$type];
2318 }
2319 }
2320 }
2321
2322 $bestq = 0;
2323 $besttype = null;
2324
2325 foreach( array_keys( $combine ) as $type ) {
2326 if( $combine[$type] > $bestq ) {
2327 $besttype = $type;
2328 $bestq = $combine[$type];
2329 }
2330 }
2331
2332 return $besttype;
2333 }
2334
2335 /**
2336 * Reference-counted warning suppression
2337 *
2338 * @param $end Bool
2339 */
2340 function wfSuppressWarnings( $end = false ) {
2341 static $suppressCount = 0;
2342 static $originalLevel = false;
2343
2344 if ( $end ) {
2345 if ( $suppressCount ) {
2346 --$suppressCount;
2347 if ( !$suppressCount ) {
2348 error_reporting( $originalLevel );
2349 }
2350 }
2351 } else {
2352 if ( !$suppressCount ) {
2353 // E_DEPRECATED is undefined in PHP 5.2
2354 if( !defined( 'E_DEPRECATED' ) ) {
2355 define( 'E_DEPRECATED', 8192 );
2356 }
2357 $originalLevel = error_reporting( E_ALL & ~( E_WARNING | E_NOTICE | E_USER_WARNING | E_USER_NOTICE | E_DEPRECATED ) );
2358 }
2359 ++$suppressCount;
2360 }
2361 }
2362
2363 /**
2364 * Restore error level to previous value
2365 */
2366 function wfRestoreWarnings() {
2367 wfSuppressWarnings( true );
2368 }
2369
2370 # Autodetect, convert and provide timestamps of various types
2371
2372 /**
2373 * Unix time - the number of seconds since 1970-01-01 00:00:00 UTC
2374 */
2375 define( 'TS_UNIX', 0 );
2376
2377 /**
2378 * MediaWiki concatenated string timestamp (YYYYMMDDHHMMSS)
2379 */
2380 define( 'TS_MW', 1 );
2381
2382 /**
2383 * MySQL DATETIME (YYYY-MM-DD HH:MM:SS)
2384 */
2385 define( 'TS_DB', 2 );
2386
2387 /**
2388 * RFC 2822 format, for E-mail and HTTP headers
2389 */
2390 define( 'TS_RFC2822', 3 );
2391
2392 /**
2393 * ISO 8601 format with no timezone: 1986-02-09T20:00:00Z
2394 *
2395 * This is used by Special:Export
2396 */
2397 define( 'TS_ISO_8601', 4 );
2398
2399 /**
2400 * An Exif timestamp (YYYY:MM:DD HH:MM:SS)
2401 *
2402 * @see http://exif.org/Exif2-2.PDF The Exif 2.2 spec, see page 28 for the
2403 * DateTime tag and page 36 for the DateTimeOriginal and
2404 * DateTimeDigitized tags.
2405 */
2406 define( 'TS_EXIF', 5 );
2407
2408 /**
2409 * Oracle format time.
2410 */
2411 define( 'TS_ORACLE', 6 );
2412
2413 /**
2414 * Postgres format time.
2415 */
2416 define( 'TS_POSTGRES', 7 );
2417
2418 /**
2419 * DB2 format time
2420 */
2421 define( 'TS_DB2', 8 );
2422
2423 /**
2424 * ISO 8601 basic format with no timezone: 19860209T200000Z. This is used by ResourceLoader
2425 */
2426 define( 'TS_ISO_8601_BASIC', 9 );
2427
2428 /**
2429 * Get a timestamp string in one of various formats
2430 *
2431 * @param $outputtype Mixed: A timestamp in one of the supported formats, the
2432 * function will autodetect which format is supplied and act
2433 * accordingly.
2434 * @param $ts Mixed: the timestamp to convert or 0 for the current timestamp
2435 * @return Mixed: String / false The same date in the format specified in $outputtype or false
2436 */
2437 function wfTimestamp( $outputtype = TS_UNIX, $ts = 0 ) {
2438 $uts = 0;
2439 $da = array();
2440 $strtime = '';
2441
2442 if ( !$ts ) { // We want to catch 0, '', null... but not date strings starting with a letter.
2443 $uts = time();
2444 $strtime = "@$uts";
2445 } elseif ( preg_match( '/^(\d{4})\-(\d\d)\-(\d\d) (\d\d):(\d\d):(\d\d)$/D', $ts, $da ) ) {
2446 # TS_DB
2447 } elseif ( preg_match( '/^(\d{4}):(\d\d):(\d\d) (\d\d):(\d\d):(\d\d)$/D', $ts, $da ) ) {
2448 # TS_EXIF
2449 } elseif ( preg_match( '/^(\d{4})(\d\d)(\d\d)(\d\d)(\d\d)(\d\d)$/D', $ts, $da ) ) {
2450 # TS_MW
2451 } elseif ( preg_match( '/^-?\d{1,13}$/D', $ts ) ) {
2452 # TS_UNIX
2453 $uts = $ts;
2454 $strtime = "@$ts"; // http://php.net/manual/en/datetime.formats.compound.php
2455 } elseif ( preg_match( '/^\d{2}-\d{2}-\d{4} \d{2}:\d{2}:\d{2}.\d{6}$/', $ts ) ) {
2456 # TS_ORACLE // session altered to DD-MM-YYYY HH24:MI:SS.FF6
2457 $strtime = preg_replace( '/(\d\d)\.(\d\d)\.(\d\d)(\.(\d+))?/', "$1:$2:$3",
2458 str_replace( '+00:00', 'UTC', $ts ) );
2459 } elseif ( preg_match( '/^(\d{4})-(\d{2})-(\d{2})T(\d{2}):(\d{2}):(\d{2})(?:\.*\d*)?Z$/', $ts, $da ) ) {
2460 # TS_ISO_8601
2461 } elseif ( preg_match( '/^(\d{4})(\d{2})(\d{2})T(\d{2})(\d{2})(\d{2})(?:\.*\d*)?Z$/', $ts, $da ) ) {
2462 #TS_ISO_8601_BASIC
2463 } elseif ( preg_match( '/^(\d{4})\-(\d\d)\-(\d\d) (\d\d):(\d\d):(\d\d)\.*\d*[\+\- ](\d\d)$/', $ts, $da ) ) {
2464 # TS_POSTGRES
2465 } elseif ( preg_match( '/^(\d{4})\-(\d\d)\-(\d\d) (\d\d):(\d\d):(\d\d)\.*\d* GMT$/', $ts, $da ) ) {
2466 # TS_POSTGRES
2467 } elseif (preg_match( '/^(\d{4})\-(\d\d)\-(\d\d) (\d\d):(\d\d):(\d\d)\.\d\d\d$/', $ts, $da ) ) {
2468 # TS_DB2
2469 } elseif ( preg_match( '/^[ \t\r\n]*([A-Z][a-z]{2},[ \t\r\n]*)?' . # Day of week
2470 '\d\d?[ \t\r\n]*[A-Z][a-z]{2}[ \t\r\n]*\d{2}(?:\d{2})?' . # dd Mon yyyy
2471 '[ \t\r\n]*\d\d[ \t\r\n]*:[ \t\r\n]*\d\d[ \t\r\n]*:[ \t\r\n]*\d\d/S', $ts ) ) { # hh:mm:ss
2472 # TS_RFC2822, accepting a trailing comment. See http://www.squid-cache.org/mail-archive/squid-users/200307/0122.html / r77171
2473 # The regex is a superset of rfc2822 for readability
2474 $strtime = strtok( $ts, ';' );
2475 } elseif ( preg_match( '/^[A-Z][a-z]{5,8}, \d\d-[A-Z][a-z]{2}-\d{2} \d\d:\d\d:\d\d/', $ts ) ) {
2476 # TS_RFC850
2477 $strtime = $ts;
2478 } elseif ( preg_match( '/^[A-Z][a-z]{2} [A-Z][a-z]{2} +\d{1,2} \d\d:\d\d:\d\d \d{4}/', $ts ) ) {
2479 # asctime
2480 $strtime = $ts;
2481 } else {
2482 # Bogus value...
2483 wfDebug("wfTimestamp() fed bogus time value: TYPE=$outputtype; VALUE=$ts\n");
2484
2485 return false;
2486 }
2487
2488 static $formats = array(
2489 TS_UNIX => 'U',
2490 TS_MW => 'YmdHis',
2491 TS_DB => 'Y-m-d H:i:s',
2492 TS_ISO_8601 => 'Y-m-d\TH:i:s\Z',
2493 TS_ISO_8601_BASIC => 'Ymd\THis\Z',
2494 TS_EXIF => 'Y:m:d H:i:s', // This shouldn't ever be used, but is included for completeness
2495 TS_RFC2822 => 'D, d M Y H:i:s',
2496 TS_ORACLE => 'd-m-Y H:i:s.000000', // Was 'd-M-y h.i.s A' . ' +00:00' before r51500
2497 TS_POSTGRES => 'Y-m-d H:i:s',
2498 TS_DB2 => 'Y-m-d H:i:s',
2499 );
2500
2501 if ( !isset( $formats[$outputtype] ) ) {
2502 throw new MWException( 'wfTimestamp() called with illegal output type.' );
2503 }
2504
2505 if ( function_exists( "date_create" ) ) {
2506 if ( count( $da ) ) {
2507 $ds = sprintf("%04d-%02d-%02dT%02d:%02d:%02d.00+00:00",
2508 (int)$da[1], (int)$da[2], (int)$da[3],
2509 (int)$da[4], (int)$da[5], (int)$da[6]);
2510
2511 $d = date_create( $ds, new DateTimeZone( 'GMT' ) );
2512 } elseif ( $strtime ) {
2513 $d = date_create( $strtime, new DateTimeZone( 'GMT' ) );
2514 } else {
2515 return false;
2516 }
2517
2518 if ( !$d ) {
2519 wfDebug("wfTimestamp() fed bogus time value: $outputtype; $ts\n");
2520 return false;
2521 }
2522
2523 $output = $d->format( $formats[$outputtype] );
2524 } else {
2525 if ( count( $da ) ) {
2526 // Warning! gmmktime() acts oddly if the month or day is set to 0
2527 // We may want to handle that explicitly at some point
2528 $uts = gmmktime( (int)$da[4], (int)$da[5], (int)$da[6],
2529 (int)$da[2], (int)$da[3], (int)$da[1] );
2530 } elseif ( $strtime ) {
2531 $uts = strtotime( $strtime );
2532 }
2533
2534 if ( $uts === false ) {
2535 wfDebug("wfTimestamp() can't parse the timestamp (non 32-bit time? Update php): $outputtype; $ts\n");
2536 return false;
2537 }
2538
2539 if ( TS_UNIX == $outputtype ) {
2540 return $uts;
2541 }
2542 $output = gmdate( $formats[$outputtype], $uts );
2543 }
2544
2545 if ( ( $outputtype == TS_RFC2822 ) || ( $outputtype == TS_POSTGRES ) ) {
2546 $output .= ' GMT';
2547 }
2548
2549 return $output;
2550 }
2551
2552 /**
2553 * Return a formatted timestamp, or null if input is null.
2554 * For dealing with nullable timestamp columns in the database.
2555 *
2556 * @param $outputtype Integer
2557 * @param $ts String
2558 * @return String
2559 */
2560 function wfTimestampOrNull( $outputtype = TS_UNIX, $ts = null ) {
2561 if( is_null( $ts ) ) {
2562 return null;
2563 } else {
2564 return wfTimestamp( $outputtype, $ts );
2565 }
2566 }
2567
2568 /**
2569 * Convenience function; returns MediaWiki timestamp for the present time.
2570 *
2571 * @return string
2572 */
2573 function wfTimestampNow() {
2574 # return NOW
2575 return wfTimestamp( TS_MW, time() );
2576 }
2577
2578 /**
2579 * Check if the operating system is Windows
2580 *
2581 * @return Bool: true if it's Windows, False otherwise.
2582 */
2583 function wfIsWindows() {
2584 static $isWindows = null;
2585 if ( $isWindows === null ) {
2586 $isWindows = substr( php_uname(), 0, 7 ) == 'Windows';
2587 }
2588 return $isWindows;
2589 }
2590
2591 /**
2592 * Check if we are running under HipHop
2593 *
2594 * @return Bool
2595 */
2596 function wfIsHipHop() {
2597 return function_exists( 'hphp_thread_set_warmup_enabled' );
2598 }
2599
2600 /**
2601 * Swap two variables
2602 *
2603 * @param $x Mixed
2604 * @param $y Mixed
2605 */
2606 function swap( &$x, &$y ) {
2607 $z = $x;
2608 $x = $y;
2609 $y = $z;
2610 }
2611
2612 /**
2613 * Tries to get the system directory for temporary files. The TMPDIR, TMP, and
2614 * TEMP environment variables are then checked in sequence, and if none are set
2615 * try sys_get_temp_dir() for PHP >= 5.2.1. All else fails, return /tmp for Unix
2616 * or C:\Windows\Temp for Windows and hope for the best.
2617 * It is common to call it with tempnam().
2618 *
2619 * NOTE: When possible, use instead the tmpfile() function to create
2620 * temporary files to avoid race conditions on file creation, etc.
2621 *
2622 * @return String
2623 */
2624 function wfTempDir() {
2625 foreach( array( 'TMPDIR', 'TMP', 'TEMP' ) as $var ) {
2626 $tmp = getenv( $var );
2627 if( $tmp && file_exists( $tmp ) && is_dir( $tmp ) && is_writable( $tmp ) ) {
2628 return $tmp;
2629 }
2630 }
2631 return sys_get_temp_dir();
2632 }
2633
2634 /**
2635 * Make directory, and make all parent directories if they don't exist
2636 *
2637 * @param $dir String: full path to directory to create
2638 * @param $mode Integer: chmod value to use, default is $wgDirectoryMode
2639 * @param $caller String: optional caller param for debugging.
2640 * @return bool
2641 */
2642 function wfMkdirParents( $dir, $mode = null, $caller = null ) {
2643 global $wgDirectoryMode;
2644
2645 if ( FileBackend::isStoragePath( $dir ) ) { // sanity
2646 throw new MWException( __FUNCTION__ . " given storage path `$dir`.");
2647 }
2648
2649 if ( !is_null( $caller ) ) {
2650 wfDebug( "$caller: called wfMkdirParents($dir)\n" );
2651 }
2652
2653 if( strval( $dir ) === '' || file_exists( $dir ) ) {
2654 return true;
2655 }
2656
2657 $dir = str_replace( array( '\\', '/' ), DIRECTORY_SEPARATOR, $dir );
2658
2659 if ( is_null( $mode ) ) {
2660 $mode = $wgDirectoryMode;
2661 }
2662
2663 // Turn off the normal warning, we're doing our own below
2664 wfSuppressWarnings();
2665 $ok = mkdir( $dir, $mode, true ); // PHP5 <3
2666 wfRestoreWarnings();
2667
2668 if( !$ok ) {
2669 // PHP doesn't report the path in its warning message, so add our own to aid in diagnosis.
2670 trigger_error( sprintf( "%s: failed to mkdir \"%s\" mode 0%o", __FUNCTION__, $dir, $mode ),
2671 E_USER_WARNING );
2672 }
2673 return $ok;
2674 }
2675
2676 /**
2677 * Remove a directory and all its content.
2678 * Does not hide error.
2679 */
2680 function wfRecursiveRemoveDir( $dir ) {
2681 wfDebug( __FUNCTION__ . "( $dir )\n" );
2682 // taken from http://de3.php.net/manual/en/function.rmdir.php#98622
2683 if ( is_dir( $dir ) ) {
2684 $objects = scandir( $dir );
2685 foreach ( $objects as $object ) {
2686 if ( $object != "." && $object != ".." ) {
2687 if ( filetype( $dir . '/' . $object ) == "dir" ) {
2688 wfRecursiveRemoveDir( $dir . '/' . $object );
2689 } else {
2690 unlink( $dir . '/' . $object );
2691 }
2692 }
2693 }
2694 reset( $objects );
2695 rmdir( $dir );
2696 }
2697 }
2698
2699 /**
2700 * @param $nr Mixed: the number to format
2701 * @param $acc Integer: the number of digits after the decimal point, default 2
2702 * @param $round Boolean: whether or not to round the value, default true
2703 * @return float
2704 */
2705 function wfPercent( $nr, $acc = 2, $round = true ) {
2706 $ret = sprintf( "%.${acc}f", $nr );
2707 return $round ? round( $ret, $acc ) . '%' : "$ret%";
2708 }
2709
2710 /**
2711 * Find out whether or not a mixed variable exists in a string
2712 *
2713 * @param $needle String
2714 * @param $str String
2715 * @param $insensitive Boolean
2716 * @return Boolean
2717 */
2718 function in_string( $needle, $str, $insensitive = false ) {
2719 $func = 'strpos';
2720 if( $insensitive ) $func = 'stripos';
2721
2722 return $func( $str, $needle ) !== false;
2723 }
2724
2725 /**
2726 * Safety wrapper around ini_get() for boolean settings.
2727 * The values returned from ini_get() are pre-normalized for settings
2728 * set via php.ini or php_flag/php_admin_flag... but *not*
2729 * for those set via php_value/php_admin_value.
2730 *
2731 * It's fairly common for people to use php_value instead of php_flag,
2732 * which can leave you with an 'off' setting giving a false positive
2733 * for code that just takes the ini_get() return value as a boolean.
2734 *
2735 * To make things extra interesting, setting via php_value accepts
2736 * "true" and "yes" as true, but php.ini and php_flag consider them false. :)
2737 * Unrecognized values go false... again opposite PHP's own coercion
2738 * from string to bool.
2739 *
2740 * Luckily, 'properly' set settings will always come back as '0' or '1',
2741 * so we only have to worry about them and the 'improper' settings.
2742 *
2743 * I frickin' hate PHP... :P
2744 *
2745 * @param $setting String
2746 * @return Bool
2747 */
2748 function wfIniGetBool( $setting ) {
2749 $val = ini_get( $setting );
2750 // 'on' and 'true' can't have whitespace around them, but '1' can.
2751 return strtolower( $val ) == 'on'
2752 || strtolower( $val ) == 'true'
2753 || strtolower( $val ) == 'yes'
2754 || preg_match( "/^\s*[+-]?0*[1-9]/", $val ); // approx C atoi() function
2755 }
2756
2757 /**
2758 * Wrapper function for PHP's dl(). This doesn't work in most situations from
2759 * PHP 5.3 onward, and is usually disabled in shared environments anyway.
2760 *
2761 * @param $extension String A PHP extension. The file suffix (.so or .dll)
2762 * should be omitted
2763 * @param $fileName String Name of the library, if not $extension.suffix
2764 * @return Bool - Whether or not the extension is loaded
2765 */
2766 function wfDl( $extension, $fileName = null ) {
2767 if( extension_loaded( $extension ) ) {
2768 return true;
2769 }
2770
2771 $canDl = false;
2772 $sapi = php_sapi_name();
2773 if( $sapi == 'cli' || $sapi == 'cgi' || $sapi == 'embed' ) {
2774 $canDl = ( function_exists( 'dl' ) && is_callable( 'dl' )
2775 && wfIniGetBool( 'enable_dl' ) && !wfIniGetBool( 'safe_mode' ) );
2776 }
2777
2778 if( $canDl ) {
2779 $fileName = $fileName ? $fileName : $extension;
2780 if( wfIsWindows() ) {
2781 $fileName = 'php_' . $fileName;
2782 }
2783 wfSuppressWarnings();
2784 dl( $fileName . '.' . PHP_SHLIB_SUFFIX );
2785 wfRestoreWarnings();
2786 }
2787 return extension_loaded( $extension );
2788 }
2789
2790 /**
2791 * Windows-compatible version of escapeshellarg()
2792 * Windows doesn't recognise single-quotes in the shell, but the escapeshellarg()
2793 * function puts single quotes in regardless of OS.
2794 *
2795 * Also fixes the locale problems on Linux in PHP 5.2.6+ (bug backported to
2796 * earlier distro releases of PHP)
2797 *
2798 * @param varargs
2799 * @return String
2800 */
2801 function wfEscapeShellArg( ) {
2802 wfInitShellLocale();
2803
2804 $args = func_get_args();
2805 $first = true;
2806 $retVal = '';
2807 foreach ( $args as $arg ) {
2808 if ( !$first ) {
2809 $retVal .= ' ';
2810 } else {
2811 $first = false;
2812 }
2813
2814 if ( wfIsWindows() ) {
2815 // Escaping for an MSVC-style command line parser and CMD.EXE
2816 // Refs:
2817 // * http://web.archive.org/web/20020708081031/http://mailman.lyra.org/pipermail/scite-interest/2002-March/000436.html
2818 // * http://technet.microsoft.com/en-us/library/cc723564.aspx
2819 // * Bug #13518
2820 // * CR r63214
2821 // Double the backslashes before any double quotes. Escape the double quotes.
2822 $tokens = preg_split( '/(\\\\*")/', $arg, -1, PREG_SPLIT_DELIM_CAPTURE );
2823 $arg = '';
2824 $iteration = 0;
2825 foreach ( $tokens as $token ) {
2826 if ( $iteration % 2 == 1 ) {
2827 // Delimiter, a double quote preceded by zero or more slashes
2828 $arg .= str_replace( '\\', '\\\\', substr( $token, 0, -1 ) ) . '\\"';
2829 } elseif ( $iteration % 4 == 2 ) {
2830 // ^ in $token will be outside quotes, need to be escaped
2831 $arg .= str_replace( '^', '^^', $token );
2832 } else { // $iteration % 4 == 0
2833 // ^ in $token will appear inside double quotes, so leave as is
2834 $arg .= $token;
2835 }
2836 $iteration++;
2837 }
2838 // Double the backslashes before the end of the string, because
2839 // we will soon add a quote
2840 $m = array();
2841 if ( preg_match( '/^(.*?)(\\\\+)$/', $arg, $m ) ) {
2842 $arg = $m[1] . str_replace( '\\', '\\\\', $m[2] );
2843 }
2844
2845 // Add surrounding quotes
2846 $retVal .= '"' . $arg . '"';
2847 } else {
2848 $retVal .= escapeshellarg( $arg );
2849 }
2850 }
2851 return $retVal;
2852 }
2853
2854 /**
2855 * Execute a shell command, with time and memory limits mirrored from the PHP
2856 * configuration if supported.
2857 * @param $cmd String Command line, properly escaped for shell.
2858 * @param &$retval null|Mixed optional, will receive the program's exit code.
2859 * (non-zero is usually failure)
2860 * @param $environ Array optional environment variables which should be
2861 * added to the executed command environment.
2862 * @return string collected stdout as a string (trailing newlines stripped)
2863 */
2864 function wfShellExec( $cmd, &$retval = null, $environ = array() ) {
2865 global $IP, $wgMaxShellMemory, $wgMaxShellFileSize, $wgMaxShellTime;
2866
2867 static $disabled;
2868 if ( is_null( $disabled ) ) {
2869 $disabled = false;
2870 if( wfIniGetBool( 'safe_mode' ) ) {
2871 wfDebug( "wfShellExec can't run in safe_mode, PHP's exec functions are too broken.\n" );
2872 $disabled = 'safemode';
2873 } else {
2874 $functions = explode( ',', ini_get( 'disable_functions' ) );
2875 $functions = array_map( 'trim', $functions );
2876 $functions = array_map( 'strtolower', $functions );
2877 if ( in_array( 'passthru', $functions ) ) {
2878 wfDebug( "passthru is in disabled_functions\n" );
2879 $disabled = 'passthru';
2880 }
2881 }
2882 }
2883 if ( $disabled ) {
2884 $retval = 1;
2885 return $disabled == 'safemode' ?
2886 'Unable to run external programs in safe mode.' :
2887 'Unable to run external programs, passthru() is disabled.';
2888 }
2889
2890 wfInitShellLocale();
2891
2892 $envcmd = '';
2893 foreach( $environ as $k => $v ) {
2894 if ( wfIsWindows() ) {
2895 /* Surrounding a set in quotes (method used by wfEscapeShellArg) makes the quotes themselves
2896 * appear in the environment variable, so we must use carat escaping as documented in
2897 * http://technet.microsoft.com/en-us/library/cc723564.aspx
2898 * Note however that the quote isn't listed there, but is needed, and the parentheses
2899 * are listed there but doesn't appear to need it.
2900 */
2901 $envcmd .= "set $k=" . preg_replace( '/([&|()<>^"])/', '^\\1', $v ) . '&& ';
2902 } else {
2903 /* Assume this is a POSIX shell, thus required to accept variable assignments before the command
2904 * http://www.opengroup.org/onlinepubs/009695399/utilities/xcu_chap02.html#tag_02_09_01
2905 */
2906 $envcmd .= "$k=" . escapeshellarg( $v ) . ' ';
2907 }
2908 }
2909 $cmd = $envcmd . $cmd;
2910
2911 if ( php_uname( 's' ) == 'Linux' ) {
2912 $time = intval( $wgMaxShellTime );
2913 $mem = intval( $wgMaxShellMemory );
2914 $filesize = intval( $wgMaxShellFileSize );
2915
2916 if ( $time > 0 && $mem > 0 ) {
2917 $script = "$IP/bin/ulimit4.sh";
2918 if ( is_executable( $script ) ) {
2919 $cmd = '/bin/bash ' . escapeshellarg( $script ) . " $time $mem $filesize " . escapeshellarg( $cmd );
2920 }
2921 }
2922 }
2923 wfDebug( "wfShellExec: $cmd\n" );
2924
2925 $retval = 1; // error by default?
2926 ob_start();
2927 passthru( $cmd, $retval );
2928 $output = ob_get_contents();
2929 ob_end_clean();
2930
2931 if ( $retval == 127 ) {
2932 wfDebugLog( 'exec', "Possibly missing executable file: $cmd\n" );
2933 }
2934 return $output;
2935 }
2936
2937 /**
2938 * Workaround for http://bugs.php.net/bug.php?id=45132
2939 * escapeshellarg() destroys non-ASCII characters if LANG is not a UTF-8 locale
2940 */
2941 function wfInitShellLocale() {
2942 static $done = false;
2943 if ( $done ) {
2944 return;
2945 }
2946 $done = true;
2947 global $wgShellLocale;
2948 if ( !wfIniGetBool( 'safe_mode' ) ) {
2949 putenv( "LC_CTYPE=$wgShellLocale" );
2950 setlocale( LC_CTYPE, $wgShellLocale );
2951 }
2952 }
2953
2954 /**
2955 * Alias to wfShellWikiCmd()
2956 * @see wfShellWikiCmd()
2957 */
2958 function wfShellMaintenanceCmd( $script, array $parameters = array(), array $options = array() ) {
2959 return wfShellWikiCmd( $script, $parameters, $options );
2960 }
2961
2962 /**
2963 * Generate a shell-escaped command line string to run a MediaWiki cli script.
2964 * Note that $parameters should be a flat array and an option with an argument
2965 * should consist of two consecutive items in the array (do not use "--option value").
2966 * @param $script string MediaWiki cli script path
2967 * @param $parameters Array Arguments and options to the script
2968 * @param $options Array Associative array of options:
2969 * 'php': The path to the php executable
2970 * 'wrapper': Path to a PHP wrapper to handle the maintenance script
2971 * @return Array
2972 */
2973 function wfShellWikiCmd( $script, array $parameters = array(), array $options = array() ) {
2974 global $wgPhpCli;
2975 // Give site config file a chance to run the script in a wrapper.
2976 // The caller may likely want to call wfBasename() on $script.
2977 wfRunHooks( 'wfShellWikiCmd', array( &$script, &$parameters, &$options ) );
2978 $cmd = isset( $options['php'] ) ? array( $options['php'] ) : array( $wgPhpCli );
2979 if ( isset( $options['wrapper'] ) ) {
2980 $cmd[] = $options['wrapper'];
2981 }
2982 $cmd[] = $script;
2983 // Escape each parameter for shell
2984 return implode( " ", array_map( 'wfEscapeShellArg', array_merge( $cmd, $parameters ) ) );
2985 }
2986
2987 /**
2988 * wfMerge attempts to merge differences between three texts.
2989 * Returns true for a clean merge and false for failure or a conflict.
2990 *
2991 * @param $old String
2992 * @param $mine String
2993 * @param $yours String
2994 * @param $result String
2995 * @return Bool
2996 */
2997 function wfMerge( $old, $mine, $yours, &$result ) {
2998 global $wgDiff3;
2999
3000 # This check may also protect against code injection in
3001 # case of broken installations.
3002 wfSuppressWarnings();
3003 $haveDiff3 = $wgDiff3 && file_exists( $wgDiff3 );
3004 wfRestoreWarnings();
3005
3006 if( !$haveDiff3 ) {
3007 wfDebug( "diff3 not found\n" );
3008 return false;
3009 }
3010
3011 # Make temporary files
3012 $td = wfTempDir();
3013 $oldtextFile = fopen( $oldtextName = tempnam( $td, 'merge-old-' ), 'w' );
3014 $mytextFile = fopen( $mytextName = tempnam( $td, 'merge-mine-' ), 'w' );
3015 $yourtextFile = fopen( $yourtextName = tempnam( $td, 'merge-your-' ), 'w' );
3016
3017 fwrite( $oldtextFile, $old );
3018 fclose( $oldtextFile );
3019 fwrite( $mytextFile, $mine );
3020 fclose( $mytextFile );
3021 fwrite( $yourtextFile, $yours );
3022 fclose( $yourtextFile );
3023
3024 # Check for a conflict
3025 $cmd = $wgDiff3 . ' -a --overlap-only ' .
3026 wfEscapeShellArg( $mytextName ) . ' ' .
3027 wfEscapeShellArg( $oldtextName ) . ' ' .
3028 wfEscapeShellArg( $yourtextName );
3029 $handle = popen( $cmd, 'r' );
3030
3031 if( fgets( $handle, 1024 ) ) {
3032 $conflict = true;
3033 } else {
3034 $conflict = false;
3035 }
3036 pclose( $handle );
3037
3038 # Merge differences
3039 $cmd = $wgDiff3 . ' -a -e --merge ' .
3040 wfEscapeShellArg( $mytextName, $oldtextName, $yourtextName );
3041 $handle = popen( $cmd, 'r' );
3042 $result = '';
3043 do {
3044 $data = fread( $handle, 8192 );
3045 if ( strlen( $data ) == 0 ) {
3046 break;
3047 }
3048 $result .= $data;
3049 } while ( true );
3050 pclose( $handle );
3051 unlink( $mytextName );
3052 unlink( $oldtextName );
3053 unlink( $yourtextName );
3054
3055 if ( $result === '' && $old !== '' && !$conflict ) {
3056 wfDebug( "Unexpected null result from diff3. Command: $cmd\n" );
3057 $conflict = true;
3058 }
3059 return !$conflict;
3060 }
3061
3062 /**
3063 * Returns unified plain-text diff of two texts.
3064 * Useful for machine processing of diffs.
3065 *
3066 * @param $before String: the text before the changes.
3067 * @param $after String: the text after the changes.
3068 * @param $params String: command-line options for the diff command.
3069 * @return String: unified diff of $before and $after
3070 */
3071 function wfDiff( $before, $after, $params = '-u' ) {
3072 if ( $before == $after ) {
3073 return '';
3074 }
3075
3076 global $wgDiff;
3077 wfSuppressWarnings();
3078 $haveDiff = $wgDiff && file_exists( $wgDiff );
3079 wfRestoreWarnings();
3080
3081 # This check may also protect against code injection in
3082 # case of broken installations.
3083 if( !$haveDiff ) {
3084 wfDebug( "diff executable not found\n" );
3085 $diffs = new Diff( explode( "\n", $before ), explode( "\n", $after ) );
3086 $format = new UnifiedDiffFormatter();
3087 return $format->format( $diffs );
3088 }
3089
3090 # Make temporary files
3091 $td = wfTempDir();
3092 $oldtextFile = fopen( $oldtextName = tempnam( $td, 'merge-old-' ), 'w' );
3093 $newtextFile = fopen( $newtextName = tempnam( $td, 'merge-your-' ), 'w' );
3094
3095 fwrite( $oldtextFile, $before );
3096 fclose( $oldtextFile );
3097 fwrite( $newtextFile, $after );
3098 fclose( $newtextFile );
3099
3100 // Get the diff of the two files
3101 $cmd = "$wgDiff " . $params . ' ' . wfEscapeShellArg( $oldtextName, $newtextName );
3102
3103 $h = popen( $cmd, 'r' );
3104
3105 $diff = '';
3106
3107 do {
3108 $data = fread( $h, 8192 );
3109 if ( strlen( $data ) == 0 ) {
3110 break;
3111 }
3112 $diff .= $data;
3113 } while ( true );
3114
3115 // Clean up
3116 pclose( $h );
3117 unlink( $oldtextName );
3118 unlink( $newtextName );
3119
3120 // Kill the --- and +++ lines. They're not useful.
3121 $diff_lines = explode( "\n", $diff );
3122 if ( strpos( $diff_lines[0], '---' ) === 0 ) {
3123 unset( $diff_lines[0] );
3124 }
3125 if ( strpos( $diff_lines[1], '+++' ) === 0 ) {
3126 unset( $diff_lines[1] );
3127 }
3128
3129 $diff = implode( "\n", $diff_lines );
3130
3131 return $diff;
3132 }
3133
3134 /**
3135 * This function works like "use VERSION" in Perl, the program will die with a
3136 * backtrace if the current version of PHP is less than the version provided
3137 *
3138 * This is useful for extensions which due to their nature are not kept in sync
3139 * with releases, and might depend on other versions of PHP than the main code
3140 *
3141 * Note: PHP might die due to parsing errors in some cases before it ever
3142 * manages to call this function, such is life
3143 *
3144 * @see perldoc -f use
3145 *
3146 * @param $req_ver Mixed: the version to check, can be a string, an integer, or
3147 * a float
3148 */
3149 function wfUsePHP( $req_ver ) {
3150 $php_ver = PHP_VERSION;
3151
3152 if ( version_compare( $php_ver, (string)$req_ver, '<' ) ) {
3153 throw new MWException( "PHP $req_ver required--this is only $php_ver" );
3154 }
3155 }
3156
3157 /**
3158 * This function works like "use VERSION" in Perl except it checks the version
3159 * of MediaWiki, the program will die with a backtrace if the current version
3160 * of MediaWiki is less than the version provided.
3161 *
3162 * This is useful for extensions which due to their nature are not kept in sync
3163 * with releases
3164 *
3165 * @see perldoc -f use
3166 *
3167 * @param $req_ver Mixed: the version to check, can be a string, an integer, or
3168 * a float
3169 */
3170 function wfUseMW( $req_ver ) {
3171 global $wgVersion;
3172
3173 if ( version_compare( $wgVersion, (string)$req_ver, '<' ) ) {
3174 throw new MWException( "MediaWiki $req_ver required--this is only $wgVersion" );
3175 }
3176 }
3177
3178 /**
3179 * Return the final portion of a pathname.
3180 * Reimplemented because PHP5's basename() is buggy with multibyte text.
3181 * http://bugs.php.net/bug.php?id=33898
3182 *
3183 * PHP's basename() only considers '\' a pathchar on Windows and Netware.
3184 * We'll consider it so always, as we don't want \s in our Unix paths either.
3185 *
3186 * @param $path String
3187 * @param $suffix String: to remove if present
3188 * @return String
3189 */
3190 function wfBaseName( $path, $suffix = '' ) {
3191 $encSuffix = ( $suffix == '' )
3192 ? ''
3193 : ( '(?:' . preg_quote( $suffix, '#' ) . ')?' );
3194 $matches = array();
3195 if( preg_match( "#([^/\\\\]*?){$encSuffix}[/\\\\]*$#", $path, $matches ) ) {
3196 return $matches[1];
3197 } else {
3198 return '';
3199 }
3200 }
3201
3202 /**
3203 * Generate a relative path name to the given file.
3204 * May explode on non-matching case-insensitive paths,
3205 * funky symlinks, etc.
3206 *
3207 * @param $path String: absolute destination path including target filename
3208 * @param $from String: Absolute source path, directory only
3209 * @return String
3210 */
3211 function wfRelativePath( $path, $from ) {
3212 // Normalize mixed input on Windows...
3213 $path = str_replace( '/', DIRECTORY_SEPARATOR, $path );
3214 $from = str_replace( '/', DIRECTORY_SEPARATOR, $from );
3215
3216 // Trim trailing slashes -- fix for drive root
3217 $path = rtrim( $path, DIRECTORY_SEPARATOR );
3218 $from = rtrim( $from, DIRECTORY_SEPARATOR );
3219
3220 $pieces = explode( DIRECTORY_SEPARATOR, dirname( $path ) );
3221 $against = explode( DIRECTORY_SEPARATOR, $from );
3222
3223 if( $pieces[0] !== $against[0] ) {
3224 // Non-matching Windows drive letters?
3225 // Return a full path.
3226 return $path;
3227 }
3228
3229 // Trim off common prefix
3230 while( count( $pieces ) && count( $against )
3231 && $pieces[0] == $against[0] ) {
3232 array_shift( $pieces );
3233 array_shift( $against );
3234 }
3235
3236 // relative dots to bump us to the parent
3237 while( count( $against ) ) {
3238 array_unshift( $pieces, '..' );
3239 array_shift( $against );
3240 }
3241
3242 array_push( $pieces, wfBaseName( $path ) );
3243
3244 return implode( DIRECTORY_SEPARATOR, $pieces );
3245 }
3246
3247 /**
3248 * Do any deferred updates and clear the list
3249 *
3250 * @deprecated since 1.19
3251 * @see DeferredUpdates::doUpdate()
3252 * @param $commit string
3253 */
3254 function wfDoUpdates( $commit = '' ) {
3255 wfDeprecated( __METHOD__, '1.19' );
3256 DeferredUpdates::doUpdates( $commit );
3257 }
3258
3259 /**
3260 * Convert an arbitrarily-long digit string from one numeric base
3261 * to another, optionally zero-padding to a minimum column width.
3262 *
3263 * Supports base 2 through 36; digit values 10-36 are represented
3264 * as lowercase letters a-z. Input is case-insensitive.
3265 *
3266 * @param $input String: of digits
3267 * @param $sourceBase Integer: 2-36
3268 * @param $destBase Integer: 2-36
3269 * @param $pad Integer: 1 or greater
3270 * @param $lowercase Boolean
3271 * @return String or false on invalid input
3272 */
3273 function wfBaseConvert( $input, $sourceBase, $destBase, $pad = 1, $lowercase = true ) {
3274 $input = strval( $input );
3275 if( $sourceBase < 2 ||
3276 $sourceBase > 36 ||
3277 $destBase < 2 ||
3278 $destBase > 36 ||
3279 $pad < 1 ||
3280 $sourceBase != intval( $sourceBase ) ||
3281 $destBase != intval( $destBase ) ||
3282 $pad != intval( $pad ) ||
3283 !is_string( $input ) ||
3284 $input == '' ) {
3285 return false;
3286 }
3287 $digitChars = ( $lowercase ) ? '0123456789abcdefghijklmnopqrstuvwxyz' : '0123456789ABCDEFGHIJKLMNOPQRSTUVWXYZ';
3288 $inDigits = array();
3289 $outChars = '';
3290
3291 // Decode and validate input string
3292 $input = strtolower( $input );
3293 for( $i = 0; $i < strlen( $input ); $i++ ) {
3294 $n = strpos( $digitChars, $input[$i] );
3295 if( $n === false || $n > $sourceBase ) {
3296 return false;
3297 }
3298 $inDigits[] = $n;
3299 }
3300
3301 // Iterate over the input, modulo-ing out an output digit
3302 // at a time until input is gone.
3303 while( count( $inDigits ) ) {
3304 $work = 0;
3305 $workDigits = array();
3306
3307 // Long division...
3308 foreach( $inDigits as $digit ) {
3309 $work *= $sourceBase;
3310 $work += $digit;
3311
3312 if( $work < $destBase ) {
3313 // Gonna need to pull another digit.
3314 if( count( $workDigits ) ) {
3315 // Avoid zero-padding; this lets us find
3316 // the end of the input very easily when
3317 // length drops to zero.
3318 $workDigits[] = 0;
3319 }
3320 } else {
3321 // Finally! Actual division!
3322 $workDigits[] = intval( $work / $destBase );
3323
3324 // Isn't it annoying that most programming languages
3325 // don't have a single divide-and-remainder operator,
3326 // even though the CPU implements it that way?
3327 $work = $work % $destBase;
3328 }
3329 }
3330
3331 // All that division leaves us with a remainder,
3332 // which is conveniently our next output digit.
3333 $outChars .= $digitChars[$work];
3334
3335 // And we continue!
3336 $inDigits = $workDigits;
3337 }
3338
3339 while( strlen( $outChars ) < $pad ) {
3340 $outChars .= '0';
3341 }
3342
3343 return strrev( $outChars );
3344 }
3345
3346 /**
3347 * Create an object with a given name and an array of construct parameters
3348 *
3349 * @param $name String
3350 * @param $p Array: parameters
3351 * @return object
3352 * @deprecated since 1.18, warnings in 1.18, removal in 1.20
3353 */
3354 function wfCreateObject( $name, $p ) {
3355 wfDeprecated( __FUNCTION__, '1.18' );
3356 return MWFunction::newObj( $name, $p );
3357 }
3358
3359 /**
3360 * @return bool
3361 */
3362 function wfHttpOnlySafe() {
3363 global $wgHttpOnlyBlacklist;
3364
3365 if( isset( $_SERVER['HTTP_USER_AGENT'] ) ) {
3366 foreach( $wgHttpOnlyBlacklist as $regex ) {
3367 if( preg_match( $regex, $_SERVER['HTTP_USER_AGENT'] ) ) {
3368 return false;
3369 }
3370 }
3371 }
3372
3373 return true;
3374 }
3375
3376 /**
3377 * Override session_id before session startup if php's built-in
3378 * session generation code is not secure.
3379 */
3380 function wfFixSessionID() {
3381 // If the cookie or session id is already set we already have a session and should abort
3382 if ( isset( $_COOKIE[ session_name() ] ) || session_id() ) {
3383 return;
3384 }
3385
3386 // PHP's built-in session entropy is enabled if:
3387 // - entropy_file is set or you're on Windows with php 5.3.3+
3388 // - AND entropy_length is > 0
3389 // We treat it as disabled if it doesn't have an entropy length of at least 32
3390 $entropyEnabled = (
3391 ( wfIsWindows() && version_compare( PHP_VERSION, '5.3.3', '>=' ) )
3392 || ini_get( 'session.entropy_file' )
3393 )
3394 && intval( ini_get( 'session.entropy_length' ) ) >= 32;
3395
3396 // If built-in entropy is not enabled or not sufficient override php's built in session id generation code
3397 if ( !$entropyEnabled ) {
3398 wfDebug( __METHOD__ . ": PHP's built in entropy is disabled or not sufficient, overriding session id generation using our cryptrand source.\n" );
3399 session_id( MWCryptRand::generateHex( 32 ) );
3400 }
3401 }
3402
3403 /**
3404 * Initialise php session
3405 *
3406 * @param $sessionId Bool
3407 */
3408 function wfSetupSession( $sessionId = false ) {
3409 global $wgSessionsInMemcached, $wgCookiePath, $wgCookieDomain,
3410 $wgCookieSecure, $wgCookieHttpOnly, $wgSessionHandler;
3411 if( $wgSessionsInMemcached ) {
3412 if ( !defined( 'MW_COMPILED' ) ) {
3413 global $IP;
3414 require_once( "$IP/includes/cache/MemcachedSessions.php" );
3415 }
3416 session_set_save_handler( 'memsess_open', 'memsess_close', 'memsess_read',
3417 'memsess_write', 'memsess_destroy', 'memsess_gc' );
3418
3419 // It's necessary to register a shutdown function to call session_write_close(),
3420 // because by the time the request shutdown function for the session module is
3421 // called, $wgMemc has already been destroyed. Shutdown functions registered
3422 // this way are called before object destruction.
3423 register_shutdown_function( 'memsess_write_close' );
3424 } elseif( $wgSessionHandler && $wgSessionHandler != ini_get( 'session.save_handler' ) ) {
3425 # Only set this if $wgSessionHandler isn't null and session.save_handler
3426 # hasn't already been set to the desired value (that causes errors)
3427 ini_set( 'session.save_handler', $wgSessionHandler );
3428 }
3429 $httpOnlySafe = wfHttpOnlySafe() && $wgCookieHttpOnly;
3430 wfDebugLog( 'cookie',
3431 'session_set_cookie_params: "' . implode( '", "',
3432 array(
3433 0,
3434 $wgCookiePath,
3435 $wgCookieDomain,
3436 $wgCookieSecure,
3437 $httpOnlySafe ) ) . '"' );
3438 session_set_cookie_params( 0, $wgCookiePath, $wgCookieDomain, $wgCookieSecure, $httpOnlySafe );
3439 session_cache_limiter( 'private, must-revalidate' );
3440 if ( $sessionId ) {
3441 session_id( $sessionId );
3442 } else {
3443 wfFixSessionID();
3444 }
3445 wfSuppressWarnings();
3446 session_start();
3447 wfRestoreWarnings();
3448 }
3449
3450 /**
3451 * Get an object from the precompiled serialized directory
3452 *
3453 * @param $name String
3454 * @return Mixed: the variable on success, false on failure
3455 */
3456 function wfGetPrecompiledData( $name ) {
3457 global $IP;
3458
3459 $file = "$IP/serialized/$name";
3460 if ( file_exists( $file ) ) {
3461 $blob = file_get_contents( $file );
3462 if ( $blob ) {
3463 return unserialize( $blob );
3464 }
3465 }
3466 return false;
3467 }
3468
3469 /**
3470 * Get a cache key
3471 *
3472 * @param varargs
3473 * @return String
3474 */
3475 function wfMemcKey( /*... */ ) {
3476 global $wgCachePrefix;
3477 $prefix = $wgCachePrefix === false ? wfWikiID() : $wgCachePrefix;
3478 $args = func_get_args();
3479 $key = $prefix . ':' . implode( ':', $args );
3480 $key = str_replace( ' ', '_', $key );
3481 return $key;
3482 }
3483
3484 /**
3485 * Get a cache key for a foreign DB
3486 *
3487 * @param $db String
3488 * @param $prefix String
3489 * @param varargs String
3490 * @return String
3491 */
3492 function wfForeignMemcKey( $db, $prefix /*, ... */ ) {
3493 $args = array_slice( func_get_args(), 2 );
3494 if ( $prefix ) {
3495 $key = "$db-$prefix:" . implode( ':', $args );
3496 } else {
3497 $key = $db . ':' . implode( ':', $args );
3498 }
3499 return $key;
3500 }
3501
3502 /**
3503 * Get an ASCII string identifying this wiki
3504 * This is used as a prefix in memcached keys
3505 *
3506 * @return String
3507 */
3508 function wfWikiID() {
3509 global $wgDBprefix, $wgDBname;
3510 if ( $wgDBprefix ) {
3511 return "$wgDBname-$wgDBprefix";
3512 } else {
3513 return $wgDBname;
3514 }
3515 }
3516
3517 /**
3518 * Split a wiki ID into DB name and table prefix
3519 *
3520 * @param $wiki String
3521 *
3522 * @return array
3523 */
3524 function wfSplitWikiID( $wiki ) {
3525 $bits = explode( '-', $wiki, 2 );
3526 if ( count( $bits ) < 2 ) {
3527 $bits[] = '';
3528 }
3529 return $bits;
3530 }
3531
3532 /**
3533 * Get a Database object.
3534 *
3535 * @param $db Integer: index of the connection to get. May be DB_MASTER for the
3536 * master (for write queries), DB_SLAVE for potentially lagged read
3537 * queries, or an integer >= 0 for a particular server.
3538 *
3539 * @param $groups Mixed: query groups. An array of group names that this query
3540 * belongs to. May contain a single string if the query is only
3541 * in one group.
3542 *
3543 * @param $wiki String: the wiki ID, or false for the current wiki
3544 *
3545 * Note: multiple calls to wfGetDB(DB_SLAVE) during the course of one request
3546 * will always return the same object, unless the underlying connection or load
3547 * balancer is manually destroyed.
3548 *
3549 * Note 2: use $this->getDB() in maintenance scripts that may be invoked by
3550 * updater to ensure that a proper database is being updated.
3551 *
3552 * @return DatabaseBase
3553 */
3554 function &wfGetDB( $db, $groups = array(), $wiki = false ) {
3555 return wfGetLB( $wiki )->getConnection( $db, $groups, $wiki );
3556 }
3557
3558 /**
3559 * Get a load balancer object.
3560 *
3561 * @param $wiki String: wiki ID, or false for the current wiki
3562 * @return LoadBalancer
3563 */
3564 function wfGetLB( $wiki = false ) {
3565 return wfGetLBFactory()->getMainLB( $wiki );
3566 }
3567
3568 /**
3569 * Get the load balancer factory object
3570 *
3571 * @return LBFactory
3572 */
3573 function &wfGetLBFactory() {
3574 return LBFactory::singleton();
3575 }
3576
3577 /**
3578 * Find a file.
3579 * Shortcut for RepoGroup::singleton()->findFile()
3580 *
3581 * @param $title String or Title object
3582 * @param $options array Associative array of options:
3583 * time: requested time for an archived image, or false for the
3584 * current version. An image object will be returned which was
3585 * created at the specified time.
3586 *
3587 * ignoreRedirect: If true, do not follow file redirects
3588 *
3589 * private: If true, return restricted (deleted) files if the current
3590 * user is allowed to view them. Otherwise, such files will not
3591 * be found.
3592 *
3593 * bypassCache: If true, do not use the process-local cache of File objects
3594 *
3595 * @return File, or false if the file does not exist
3596 */
3597 function wfFindFile( $title, $options = array() ) {
3598 return RepoGroup::singleton()->findFile( $title, $options );
3599 }
3600
3601 /**
3602 * Get an object referring to a locally registered file.
3603 * Returns a valid placeholder object if the file does not exist.
3604 *
3605 * @param $title Title|String
3606 * @return LocalFile|null A File, or null if passed an invalid Title
3607 */
3608 function wfLocalFile( $title ) {
3609 return RepoGroup::singleton()->getLocalRepo()->newFile( $title );
3610 }
3611
3612 /**
3613 * Stream a file to the browser. Back-compat alias for StreamFile::stream()
3614 * @deprecated since 1.19
3615 */
3616 function wfStreamFile( $fname, $headers = array() ) {
3617 wfDeprecated( __FUNCTION__, '1.19' );
3618 StreamFile::stream( $fname, $headers );
3619 }
3620
3621 /**
3622 * Should low-performance queries be disabled?
3623 *
3624 * @return Boolean
3625 * @codeCoverageIgnore
3626 */
3627 function wfQueriesMustScale() {
3628 global $wgMiserMode;
3629 return $wgMiserMode
3630 || ( SiteStats::pages() > 100000
3631 && SiteStats::edits() > 1000000
3632 && SiteStats::users() > 10000 );
3633 }
3634
3635 /**
3636 * Get the path to a specified script file, respecting file
3637 * extensions; this is a wrapper around $wgScriptExtension etc.
3638 *
3639 * @param $script String: script filename, sans extension
3640 * @return String
3641 */
3642 function wfScript( $script = 'index' ) {
3643 global $wgScriptPath, $wgScriptExtension;
3644 return "{$wgScriptPath}/{$script}{$wgScriptExtension}";
3645 }
3646
3647 /**
3648 * Get the script URL.
3649 *
3650 * @return string script URL
3651 */
3652 function wfGetScriptUrl() {
3653 if( isset( $_SERVER['SCRIPT_NAME'] ) ) {
3654 #
3655 # as it was called, minus the query string.
3656 #
3657 # Some sites use Apache rewrite rules to handle subdomains,
3658 # and have PHP set up in a weird way that causes PHP_SELF
3659 # to contain the rewritten URL instead of the one that the
3660 # outside world sees.
3661 #
3662 # If in this mode, use SCRIPT_URL instead, which mod_rewrite
3663 # provides containing the "before" URL.
3664 return $_SERVER['SCRIPT_NAME'];
3665 } else {
3666 return $_SERVER['URL'];
3667 }
3668 }
3669
3670 /**
3671 * Convenience function converts boolean values into "true"
3672 * or "false" (string) values
3673 *
3674 * @param $value Boolean
3675 * @return String
3676 */
3677 function wfBoolToStr( $value ) {
3678 return $value ? 'true' : 'false';
3679 }
3680
3681 /**
3682 * Load an extension messages file
3683 *
3684 * @deprecated since 1.16, warnings in 1.18, remove in 1.20
3685 * @codeCoverageIgnore
3686 */
3687 function wfLoadExtensionMessages() {
3688 wfDeprecated( __FUNCTION__, '1.16' );
3689 }
3690
3691 /**
3692 * Get a platform-independent path to the null file, e.g. /dev/null
3693 *
3694 * @return string
3695 */
3696 function wfGetNull() {
3697 return wfIsWindows()
3698 ? 'NUL'
3699 : '/dev/null';
3700 }
3701
3702 /**
3703 * Modern version of wfWaitForSlaves(). Instead of looking at replication lag
3704 * and waiting for it to go down, this waits for the slaves to catch up to the
3705 * master position. Use this when updating very large numbers of rows, as
3706 * in maintenance scripts, to avoid causing too much lag. Of course, this is
3707 * a no-op if there are no slaves.
3708 *
3709 * @param $maxLag Integer (deprecated)
3710 * @param $wiki mixed Wiki identifier accepted by wfGetLB
3711 */
3712 function wfWaitForSlaves( $maxLag = false, $wiki = false ) {
3713 $lb = wfGetLB( $wiki );
3714 // bug 27975 - Don't try to wait for slaves if there are none
3715 // Prevents permission error when getting master position
3716 if ( $lb->getServerCount() > 1 ) {
3717 $dbw = $lb->getConnection( DB_MASTER );
3718 $pos = $dbw->getMasterPos();
3719 $lb->waitForAll( $pos );
3720 }
3721 }
3722
3723 /**
3724 * Used to be used for outputting text in the installer/updater
3725 * @deprecated since 1.18, warnings in 1.18, remove in 1.20
3726 */
3727 function wfOut( $s ) {
3728 wfDeprecated( __FUNCTION__, '1.18' );
3729 global $wgCommandLineMode;
3730 if ( $wgCommandLineMode ) {
3731 echo $s;
3732 } else {
3733 echo htmlspecialchars( $s );
3734 }
3735 flush();
3736 }
3737
3738 /**
3739 * Count down from $n to zero on the terminal, with a one-second pause
3740 * between showing each number. For use in command-line scripts.
3741 * @codeCoverageIgnore
3742 * @param $n int
3743 */
3744 function wfCountDown( $n ) {
3745 for ( $i = $n; $i >= 0; $i-- ) {
3746 if ( $i != $n ) {
3747 echo str_repeat( "\x08", strlen( $i + 1 ) );
3748 }
3749 echo $i;
3750 flush();
3751 if ( $i ) {
3752 sleep( 1 );
3753 }
3754 }
3755 echo "\n";
3756 }
3757
3758 /**
3759 * Generate a random 32-character hexadecimal token.
3760 * @param $salt Mixed: some sort of salt, if necessary, to add to random
3761 * characters before hashing.
3762 * @return string
3763 * @codeCoverageIgnore
3764 * @deprecated since 1.20; Please use MWCryptRand for security purposes and wfRandomString for pesudo-random strings
3765 * @warning This method is NOT secure. Additionally it has many callers that use it for pesudo-random purposes.
3766 */
3767 function wfGenerateToken( $salt = '' ) {
3768 wfDeprecated( __METHOD__, '1.20' );
3769 $salt = serialize( $salt );
3770 return md5( mt_rand( 0, 0x7fffffff ) . $salt );
3771 }
3772
3773 /**
3774 * Replace all invalid characters with -
3775 *
3776 * @param $name Mixed: filename to process
3777 * @return String
3778 */
3779 function wfStripIllegalFilenameChars( $name ) {
3780 global $wgIllegalFileChars;
3781 $name = wfBaseName( $name );
3782 $name = preg_replace(
3783 "/[^" . Title::legalChars() . "]" .
3784 ( $wgIllegalFileChars ? "|[" . $wgIllegalFileChars . "]" : '' ) .
3785 "/",
3786 '-',
3787 $name
3788 );
3789 return $name;
3790 }
3791
3792 /**
3793 * Set PHP's memory limit to the larger of php.ini or $wgMemoryLimit;
3794 *
3795 * @return Integer value memory was set to.
3796 */
3797 function wfMemoryLimit() {
3798 global $wgMemoryLimit;
3799 $memlimit = wfShorthandToInteger( ini_get( 'memory_limit' ) );
3800 if( $memlimit != -1 ) {
3801 $conflimit = wfShorthandToInteger( $wgMemoryLimit );
3802 if( $conflimit == -1 ) {
3803 wfDebug( "Removing PHP's memory limit\n" );
3804 wfSuppressWarnings();
3805 ini_set( 'memory_limit', $conflimit );
3806 wfRestoreWarnings();
3807 return $conflimit;
3808 } elseif ( $conflimit > $memlimit ) {
3809 wfDebug( "Raising PHP's memory limit to $conflimit bytes\n" );
3810 wfSuppressWarnings();
3811 ini_set( 'memory_limit', $conflimit );
3812 wfRestoreWarnings();
3813 return $conflimit;
3814 }
3815 }
3816 return $memlimit;
3817 }
3818
3819 /**
3820 * Converts shorthand byte notation to integer form
3821 *
3822 * @param $string String
3823 * @return Integer
3824 */
3825 function wfShorthandToInteger( $string = '' ) {
3826 $string = trim( $string );
3827 if( $string === '' ) {
3828 return -1;
3829 }
3830 $last = $string[strlen( $string ) - 1];
3831 $val = intval( $string );
3832 switch( $last ) {
3833 case 'g':
3834 case 'G':
3835 $val *= 1024;
3836 // break intentionally missing
3837 case 'm':
3838 case 'M':
3839 $val *= 1024;
3840 // break intentionally missing
3841 case 'k':
3842 case 'K':
3843 $val *= 1024;
3844 }
3845
3846 return $val;
3847 }
3848
3849 /**
3850 * Get the normalised IETF language tag
3851 * See unit test for examples.
3852 *
3853 * @param $code String: The language code.
3854 * @return String: The language code which complying with BCP 47 standards.
3855 */
3856 function wfBCP47( $code ) {
3857 $codeSegment = explode( '-', $code );
3858 $codeBCP = array();
3859 foreach ( $codeSegment as $segNo => $seg ) {
3860 if ( count( $codeSegment ) > 0 ) {
3861 // when previous segment is x, it is a private segment and should be lc
3862 if( $segNo > 0 && strtolower( $codeSegment[( $segNo - 1 )] ) == 'x' ) {
3863 $codeBCP[$segNo] = strtolower( $seg );
3864 // ISO 3166 country code
3865 } elseif ( ( strlen( $seg ) == 2 ) && ( $segNo > 0 ) ) {
3866 $codeBCP[$segNo] = strtoupper( $seg );
3867 // ISO 15924 script code
3868 } elseif ( ( strlen( $seg ) == 4 ) && ( $segNo > 0 ) ) {
3869 $codeBCP[$segNo] = ucfirst( strtolower( $seg ) );
3870 // Use lowercase for other cases
3871 } else {
3872 $codeBCP[$segNo] = strtolower( $seg );
3873 }
3874 } else {
3875 // Use lowercase for single segment
3876 $codeBCP[$segNo] = strtolower( $seg );
3877 }
3878 }
3879 $langCode = implode( '-', $codeBCP );
3880 return $langCode;
3881 }
3882
3883 /**
3884 * Get a cache object.
3885 *
3886 * @param $inputType integer Cache type, one the the CACHE_* constants.
3887 * @return BagOStuff
3888 */
3889 function wfGetCache( $inputType ) {
3890 return ObjectCache::getInstance( $inputType );
3891 }
3892
3893 /**
3894 * Get the main cache object
3895 *
3896 * @return BagOStuff
3897 */
3898 function wfGetMainCache() {
3899 global $wgMainCacheType;
3900 return ObjectCache::getInstance( $wgMainCacheType );
3901 }
3902
3903 /**
3904 * Get the cache object used by the message cache
3905 *
3906 * @return BagOStuff
3907 */
3908 function wfGetMessageCacheStorage() {
3909 global $wgMessageCacheType;
3910 return ObjectCache::getInstance( $wgMessageCacheType );
3911 }
3912
3913 /**
3914 * Get the cache object used by the parser cache
3915 *
3916 * @return BagOStuff
3917 */
3918 function wfGetParserCacheStorage() {
3919 global $wgParserCacheType;
3920 return ObjectCache::getInstance( $wgParserCacheType );
3921 }
3922
3923 /**
3924 * Get the cache object used by the language converter
3925 *
3926 * @return BagOStuff
3927 */
3928 function wfGetLangConverterCacheStorage() {
3929 global $wgLanguageConverterCacheType;
3930 return ObjectCache::getInstance( $wgLanguageConverterCacheType );
3931 }
3932
3933 /**
3934 * Call hook functions defined in $wgHooks
3935 *
3936 * @param $event String: event name
3937 * @param $args Array: parameters passed to hook functions
3938 * @return Boolean True if no handler aborted the hook
3939 */
3940 function wfRunHooks( $event, $args = array() ) {
3941 return Hooks::run( $event, $args );
3942 }
3943
3944 /**
3945 * Wrapper around php's unpack.
3946 *
3947 * @param $format String: The format string (See php's docs)
3948 * @param $data: A binary string of binary data
3949 * @param $length integer or false: The minimun length of $data. This is to
3950 * prevent reading beyond the end of $data. false to disable the check.
3951 *
3952 * Also be careful when using this function to read unsigned 32 bit integer
3953 * because php might make it negative.
3954 *
3955 * @throws MWException if $data not long enough, or if unpack fails
3956 * @return array Associative array of the extracted data
3957 */
3958 function wfUnpack( $format, $data, $length=false ) {
3959 if ( $length !== false ) {
3960 $realLen = strlen( $data );
3961 if ( $realLen < $length ) {
3962 throw new MWException( "Tried to use wfUnpack on a "
3963 . "string of length $realLen, but needed one "
3964 . "of at least length $length."
3965 );
3966 }
3967 }
3968
3969 wfSuppressWarnings();
3970 $result = unpack( $format, $data );
3971 wfRestoreWarnings();
3972
3973 if ( $result === false ) {
3974 // If it cannot extract the packed data.
3975 throw new MWException( "unpack could not unpack binary data" );
3976 }
3977 return $result;
3978 }
3979
3980 /**
3981 * Determine if an image exists on the 'bad image list'.
3982 *
3983 * The format of MediaWiki:Bad_image_list is as follows:
3984 * * Only list items (lines starting with "*") are considered
3985 * * The first link on a line must be a link to a bad image
3986 * * Any subsequent links on the same line are considered to be exceptions,
3987 * i.e. articles where the image may occur inline.
3988 *
3989 * @param $name string the image name to check
3990 * @param $contextTitle Title|bool the page on which the image occurs, if known
3991 * @param $blacklist string wikitext of a file blacklist
3992 * @return bool
3993 */
3994 function wfIsBadImage( $name, $contextTitle = false, $blacklist = null ) {
3995 static $badImageCache = null; // based on bad_image_list msg
3996 wfProfileIn( __METHOD__ );
3997
3998 # Handle redirects
3999 $redirectTitle = RepoGroup::singleton()->checkRedirect( Title::makeTitle( NS_FILE, $name ) );
4000 if( $redirectTitle ) {
4001 $name = $redirectTitle->getDbKey();
4002 }
4003
4004 # Run the extension hook
4005 $bad = false;
4006 if( !wfRunHooks( 'BadImage', array( $name, &$bad ) ) ) {
4007 wfProfileOut( __METHOD__ );
4008 return $bad;
4009 }
4010
4011 $cacheable = ( $blacklist === null );
4012 if( $cacheable && $badImageCache !== null ) {
4013 $badImages = $badImageCache;
4014 } else { // cache miss
4015 if ( $blacklist === null ) {
4016 $blacklist = wfMsgForContentNoTrans( 'bad_image_list' ); // site list
4017 }
4018 # Build the list now
4019 $badImages = array();
4020 $lines = explode( "\n", $blacklist );
4021 foreach( $lines as $line ) {
4022 # List items only
4023 if ( substr( $line, 0, 1 ) !== '*' ) {
4024 continue;
4025 }
4026
4027 # Find all links
4028 $m = array();
4029 if ( !preg_match_all( '/\[\[:?(.*?)\]\]/', $line, $m ) ) {
4030 continue;
4031 }
4032
4033 $exceptions = array();
4034 $imageDBkey = false;
4035 foreach ( $m[1] as $i => $titleText ) {
4036 $title = Title::newFromText( $titleText );
4037 if ( !is_null( $title ) ) {
4038 if ( $i == 0 ) {
4039 $imageDBkey = $title->getDBkey();
4040 } else {
4041 $exceptions[$title->getPrefixedDBkey()] = true;
4042 }
4043 }
4044 }
4045
4046 if ( $imageDBkey !== false ) {
4047 $badImages[$imageDBkey] = $exceptions;
4048 }
4049 }
4050 if ( $cacheable ) {
4051 $badImageCache = $badImages;
4052 }
4053 }
4054
4055 $contextKey = $contextTitle ? $contextTitle->getPrefixedDBkey() : false;
4056 $bad = isset( $badImages[$name] ) && !isset( $badImages[$name][$contextKey] );
4057 wfProfileOut( __METHOD__ );
4058 return $bad;
4059 }