Merge "Move wfEscapeWikiText() to Parser::escapeWikitext()"
[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 use Liuggio\StatsdClient\Sender\SocketSender;
28 use MediaWiki\Logger\LoggerFactory;
29 use MediaWiki\Session\SessionManager;
30
31 // Hide compatibility functions from Doxygen
32 /// @cond
33 /**
34 * Compatibility functions
35 *
36 * We support PHP 5.5.9 and up.
37 * Re-implementations of newer functions or functions in non-standard
38 * PHP extensions may be included here.
39 */
40
41 // hash_equals function only exists in PHP >= 5.6.0
42 // http://php.net/hash_equals
43 if ( !function_exists( 'hash_equals' ) ) {
44 /**
45 * Check whether a user-provided string is equal to a fixed-length secret string
46 * without revealing bytes of the secret string through timing differences.
47 *
48 * The usual way to compare strings (PHP's === operator or the underlying memcmp()
49 * function in C) is to compare corresponding bytes and stop at the first difference,
50 * which would take longer for a partial match than for a complete mismatch. This
51 * is not secure when one of the strings (e.g. an HMAC or token) must remain secret
52 * and the other may come from an attacker. Statistical analysis of timing measurements
53 * over many requests may allow the attacker to guess the string's bytes one at a time
54 * (and check his guesses) even if the timing differences are extremely small.
55 *
56 * When making such a security-sensitive comparison, it is essential that the sequence
57 * in which instructions are executed and memory locations are accessed not depend on
58 * the secret string's value. HOWEVER, for simplicity, we do not attempt to minimize
59 * the inevitable leakage of the string's length. That is generally known anyway as
60 * a chararacteristic of the hash function used to compute the secret value.
61 *
62 * Longer explanation: http://www.emerose.com/timing-attacks-explained
63 *
64 * @codeCoverageIgnore
65 * @param string $known_string Fixed-length secret string to compare against
66 * @param string $user_string User-provided string
67 * @return bool True if the strings are the same, false otherwise
68 */
69 function hash_equals( $known_string, $user_string ) {
70 // Strict type checking as in PHP's native implementation
71 if ( !is_string( $known_string ) ) {
72 trigger_error( 'hash_equals(): Expected known_string to be a string, ' .
73 gettype( $known_string ) . ' given', E_USER_WARNING );
74
75 return false;
76 }
77
78 if ( !is_string( $user_string ) ) {
79 trigger_error( 'hash_equals(): Expected user_string to be a string, ' .
80 gettype( $user_string ) . ' given', E_USER_WARNING );
81
82 return false;
83 }
84
85 $known_string_len = strlen( $known_string );
86 if ( $known_string_len !== strlen( $user_string ) ) {
87 return false;
88 }
89
90 $result = 0;
91 for ( $i = 0; $i < $known_string_len; $i++ ) {
92 $result |= ord( $known_string[$i] ) ^ ord( $user_string[$i] );
93 }
94
95 return ( $result === 0 );
96 }
97 }
98 /// @endcond
99
100 /**
101 * Load an extension
102 *
103 * This queues an extension to be loaded through
104 * the ExtensionRegistry system.
105 *
106 * @param string $ext Name of the extension to load
107 * @param string|null $path Absolute path of where to find the extension.json file
108 * @since 1.25
109 */
110 function wfLoadExtension( $ext, $path = null ) {
111 if ( !$path ) {
112 global $wgExtensionDirectory;
113 $path = "$wgExtensionDirectory/$ext/extension.json";
114 }
115 ExtensionRegistry::getInstance()->queue( $path );
116 }
117
118 /**
119 * Load multiple extensions at once
120 *
121 * Same as wfLoadExtension, but more efficient if you
122 * are loading multiple extensions.
123 *
124 * If you want to specify custom paths, you should interact with
125 * ExtensionRegistry directly.
126 *
127 * @see wfLoadExtension
128 * @param string[] $exts Array of extension names to load
129 * @since 1.25
130 */
131 function wfLoadExtensions( array $exts ) {
132 global $wgExtensionDirectory;
133 $registry = ExtensionRegistry::getInstance();
134 foreach ( $exts as $ext ) {
135 $registry->queue( "$wgExtensionDirectory/$ext/extension.json" );
136 }
137 }
138
139 /**
140 * Load a skin
141 *
142 * @see wfLoadExtension
143 * @param string $skin Name of the extension to load
144 * @param string|null $path Absolute path of where to find the skin.json file
145 * @since 1.25
146 */
147 function wfLoadSkin( $skin, $path = null ) {
148 if ( !$path ) {
149 global $wgStyleDirectory;
150 $path = "$wgStyleDirectory/$skin/skin.json";
151 }
152 ExtensionRegistry::getInstance()->queue( $path );
153 }
154
155 /**
156 * Load multiple skins at once
157 *
158 * @see wfLoadExtensions
159 * @param string[] $skins Array of extension names to load
160 * @since 1.25
161 */
162 function wfLoadSkins( array $skins ) {
163 global $wgStyleDirectory;
164 $registry = ExtensionRegistry::getInstance();
165 foreach ( $skins as $skin ) {
166 $registry->queue( "$wgStyleDirectory/$skin/skin.json" );
167 }
168 }
169
170 /**
171 * Like array_diff( $a, $b ) except that it works with two-dimensional arrays.
172 * @param array $a
173 * @param array $b
174 * @return array
175 */
176 function wfArrayDiff2( $a, $b ) {
177 return array_udiff( $a, $b, 'wfArrayDiff2_cmp' );
178 }
179
180 /**
181 * @param array|string $a
182 * @param array|string $b
183 * @return int
184 */
185 function wfArrayDiff2_cmp( $a, $b ) {
186 if ( is_string( $a ) && is_string( $b ) ) {
187 return strcmp( $a, $b );
188 } elseif ( count( $a ) !== count( $b ) ) {
189 return count( $a ) < count( $b ) ? -1 : 1;
190 } else {
191 reset( $a );
192 reset( $b );
193 while ( ( list( , $valueA ) = each( $a ) ) && ( list( , $valueB ) = each( $b ) ) ) {
194 $cmp = strcmp( $valueA, $valueB );
195 if ( $cmp !== 0 ) {
196 return $cmp;
197 }
198 }
199 return 0;
200 }
201 }
202
203 /**
204 * Appends to second array if $value differs from that in $default
205 *
206 * @param string|int $key
207 * @param mixed $value
208 * @param mixed $default
209 * @param array $changed Array to alter
210 * @throws MWException
211 */
212 function wfAppendToArrayIfNotDefault( $key, $value, $default, &$changed ) {
213 if ( is_null( $changed ) ) {
214 throw new MWException( 'GlobalFunctions::wfAppendToArrayIfNotDefault got null' );
215 }
216 if ( $default[$key] !== $value ) {
217 $changed[$key] = $value;
218 }
219 }
220
221 /**
222 * Merge arrays in the style of getUserPermissionsErrors, with duplicate removal
223 * e.g.
224 * wfMergeErrorArrays(
225 * [ [ 'x' ] ],
226 * [ [ 'x', '2' ] ],
227 * [ [ 'x' ] ],
228 * [ [ 'y' ] ]
229 * );
230 * returns:
231 * [
232 * [ 'x', '2' ],
233 * [ 'x' ],
234 * [ 'y' ]
235 * ]
236 *
237 * @param array $array1,...
238 * @return array
239 */
240 function wfMergeErrorArrays( /*...*/ ) {
241 $args = func_get_args();
242 $out = [];
243 foreach ( $args as $errors ) {
244 foreach ( $errors as $params ) {
245 $originalParams = $params;
246 if ( $params[0] instanceof MessageSpecifier ) {
247 $msg = $params[0];
248 $params = array_merge( [ $msg->getKey() ], $msg->getParams() );
249 }
250 # @todo FIXME: Sometimes get nested arrays for $params,
251 # which leads to E_NOTICEs
252 $spec = implode( "\t", $params );
253 $out[$spec] = $originalParams;
254 }
255 }
256 return array_values( $out );
257 }
258
259 /**
260 * Insert array into another array after the specified *KEY*
261 *
262 * @param array $array The array.
263 * @param array $insert The array to insert.
264 * @param mixed $after The key to insert after
265 * @return array
266 */
267 function wfArrayInsertAfter( array $array, array $insert, $after ) {
268 // Find the offset of the element to insert after.
269 $keys = array_keys( $array );
270 $offsetByKey = array_flip( $keys );
271
272 $offset = $offsetByKey[$after];
273
274 // Insert at the specified offset
275 $before = array_slice( $array, 0, $offset + 1, true );
276 $after = array_slice( $array, $offset + 1, count( $array ) - $offset, true );
277
278 $output = $before + $insert + $after;
279
280 return $output;
281 }
282
283 /**
284 * Recursively converts the parameter (an object) to an array with the same data
285 *
286 * @param object|array $objOrArray
287 * @param bool $recursive
288 * @return array
289 */
290 function wfObjectToArray( $objOrArray, $recursive = true ) {
291 $array = [];
292 if ( is_object( $objOrArray ) ) {
293 $objOrArray = get_object_vars( $objOrArray );
294 }
295 foreach ( $objOrArray as $key => $value ) {
296 if ( $recursive && ( is_object( $value ) || is_array( $value ) ) ) {
297 $value = wfObjectToArray( $value );
298 }
299
300 $array[$key] = $value;
301 }
302
303 return $array;
304 }
305
306 /**
307 * Get a random decimal value between 0 and 1, in a way
308 * not likely to give duplicate values for any realistic
309 * number of articles.
310 *
311 * @note This is designed for use in relation to Special:RandomPage
312 * and the page_random database field.
313 *
314 * @return string
315 */
316 function wfRandom() {
317 // The maximum random value is "only" 2^31-1, so get two random
318 // values to reduce the chance of dupes
319 $max = mt_getrandmax() + 1;
320 $rand = number_format( ( mt_rand() * $max + mt_rand() ) / $max / $max, 12, '.', '' );
321 return $rand;
322 }
323
324 /**
325 * Get a random string containing a number of pseudo-random hex characters.
326 *
327 * @note This is not secure, if you are trying to generate some sort
328 * of token please use MWCryptRand instead.
329 *
330 * @param int $length The length of the string to generate
331 * @return string
332 * @since 1.20
333 */
334 function wfRandomString( $length = 32 ) {
335 $str = '';
336 for ( $n = 0; $n < $length; $n += 7 ) {
337 $str .= sprintf( '%07x', mt_rand() & 0xfffffff );
338 }
339 return substr( $str, 0, $length );
340 }
341
342 /**
343 * We want some things to be included as literal characters in our title URLs
344 * for prettiness, which urlencode encodes by default. According to RFC 1738,
345 * all of the following should be safe:
346 *
347 * ;:@&=$-_.+!*'(),
348 *
349 * RFC 1738 says ~ is unsafe, however RFC 3986 considers it an unreserved
350 * character which should not be encoded. More importantly, google chrome
351 * always converts %7E back to ~, and converting it in this function can
352 * cause a redirect loop (T105265).
353 *
354 * But + is not safe because it's used to indicate a space; &= are only safe in
355 * paths and not in queries (and we don't distinguish here); ' seems kind of
356 * scary; and urlencode() doesn't touch -_. to begin with. Plus, although /
357 * is reserved, we don't care. So the list we unescape is:
358 *
359 * ;:@$!*(),/~
360 *
361 * However, IIS7 redirects fail when the url contains a colon (see T24709),
362 * so no fancy : for IIS7.
363 *
364 * %2F in the page titles seems to fatally break for some reason.
365 *
366 * @param string $s
367 * @return string
368 */
369 function wfUrlencode( $s ) {
370 static $needle;
371
372 if ( is_null( $s ) ) {
373 $needle = null;
374 return '';
375 }
376
377 if ( is_null( $needle ) ) {
378 $needle = [ '%3B', '%40', '%24', '%21', '%2A', '%28', '%29', '%2C', '%2F', '%7E' ];
379 if ( !isset( $_SERVER['SERVER_SOFTWARE'] ) ||
380 ( strpos( $_SERVER['SERVER_SOFTWARE'], 'Microsoft-IIS/7' ) === false )
381 ) {
382 $needle[] = '%3A';
383 }
384 }
385
386 $s = urlencode( $s );
387 $s = str_ireplace(
388 $needle,
389 [ ';', '@', '$', '!', '*', '(', ')', ',', '/', '~', ':' ],
390 $s
391 );
392
393 return $s;
394 }
395
396 /**
397 * This function takes one or two arrays as input, and returns a CGI-style string, e.g.
398 * "days=7&limit=100". Options in the first array override options in the second.
399 * Options set to null or false will not be output.
400 *
401 * @param array $array1 ( String|Array )
402 * @param array|null $array2 ( String|Array )
403 * @param string $prefix
404 * @return string
405 */
406 function wfArrayToCgi( $array1, $array2 = null, $prefix = '' ) {
407 if ( !is_null( $array2 ) ) {
408 $array1 = $array1 + $array2;
409 }
410
411 $cgi = '';
412 foreach ( $array1 as $key => $value ) {
413 if ( !is_null( $value ) && $value !== false ) {
414 if ( $cgi != '' ) {
415 $cgi .= '&';
416 }
417 if ( $prefix !== '' ) {
418 $key = $prefix . "[$key]";
419 }
420 if ( is_array( $value ) ) {
421 $firstTime = true;
422 foreach ( $value as $k => $v ) {
423 $cgi .= $firstTime ? '' : '&';
424 if ( is_array( $v ) ) {
425 $cgi .= wfArrayToCgi( $v, null, $key . "[$k]" );
426 } else {
427 $cgi .= urlencode( $key . "[$k]" ) . '=' . urlencode( $v );
428 }
429 $firstTime = false;
430 }
431 } else {
432 if ( is_object( $value ) ) {
433 $value = $value->__toString();
434 }
435 $cgi .= urlencode( $key ) . '=' . urlencode( $value );
436 }
437 }
438 }
439 return $cgi;
440 }
441
442 /**
443 * This is the logical opposite of wfArrayToCgi(): it accepts a query string as
444 * its argument and returns the same string in array form. This allows compatibility
445 * with legacy functions that accept raw query strings instead of nice
446 * arrays. Of course, keys and values are urldecode()d.
447 *
448 * @param string $query Query string
449 * @return string[] Array version of input
450 */
451 function wfCgiToArray( $query ) {
452 if ( isset( $query[0] ) && $query[0] == '?' ) {
453 $query = substr( $query, 1 );
454 }
455 $bits = explode( '&', $query );
456 $ret = [];
457 foreach ( $bits as $bit ) {
458 if ( $bit === '' ) {
459 continue;
460 }
461 if ( strpos( $bit, '=' ) === false ) {
462 // Pieces like &qwerty become 'qwerty' => '' (at least this is what php does)
463 $key = $bit;
464 $value = '';
465 } else {
466 list( $key, $value ) = explode( '=', $bit );
467 }
468 $key = urldecode( $key );
469 $value = urldecode( $value );
470 if ( strpos( $key, '[' ) !== false ) {
471 $keys = array_reverse( explode( '[', $key ) );
472 $key = array_pop( $keys );
473 $temp = $value;
474 foreach ( $keys as $k ) {
475 $k = substr( $k, 0, -1 );
476 $temp = [ $k => $temp ];
477 }
478 if ( isset( $ret[$key] ) ) {
479 $ret[$key] = array_merge( $ret[$key], $temp );
480 } else {
481 $ret[$key] = $temp;
482 }
483 } else {
484 $ret[$key] = $value;
485 }
486 }
487 return $ret;
488 }
489
490 /**
491 * Append a query string to an existing URL, which may or may not already
492 * have query string parameters already. If so, they will be combined.
493 *
494 * @param string $url
495 * @param string|string[] $query String or associative array
496 * @return string
497 */
498 function wfAppendQuery( $url, $query ) {
499 if ( is_array( $query ) ) {
500 $query = wfArrayToCgi( $query );
501 }
502 if ( $query != '' ) {
503 // Remove the fragment, if there is one
504 $fragment = false;
505 $hashPos = strpos( $url, '#' );
506 if ( $hashPos !== false ) {
507 $fragment = substr( $url, $hashPos );
508 $url = substr( $url, 0, $hashPos );
509 }
510
511 // Add parameter
512 if ( false === strpos( $url, '?' ) ) {
513 $url .= '?';
514 } else {
515 $url .= '&';
516 }
517 $url .= $query;
518
519 // Put the fragment back
520 if ( $fragment !== false ) {
521 $url .= $fragment;
522 }
523 }
524 return $url;
525 }
526
527 /**
528 * Expand a potentially local URL to a fully-qualified URL. Assumes $wgServer
529 * is correct.
530 *
531 * The meaning of the PROTO_* constants is as follows:
532 * PROTO_HTTP: Output a URL starting with http://
533 * PROTO_HTTPS: Output a URL starting with https://
534 * PROTO_RELATIVE: Output a URL starting with // (protocol-relative URL)
535 * PROTO_CURRENT: Output a URL starting with either http:// or https:// , depending
536 * on which protocol was used for the current incoming request
537 * PROTO_CANONICAL: For URLs without a domain, like /w/index.php , use $wgCanonicalServer.
538 * For protocol-relative URLs, use the protocol of $wgCanonicalServer
539 * PROTO_INTERNAL: Like PROTO_CANONICAL, but uses $wgInternalServer instead of $wgCanonicalServer
540 *
541 * @todo this won't work with current-path-relative URLs
542 * like "subdir/foo.html", etc.
543 *
544 * @param string $url Either fully-qualified or a local path + query
545 * @param string $defaultProto One of the PROTO_* constants. Determines the
546 * protocol to use if $url or $wgServer is protocol-relative
547 * @return string Fully-qualified URL, current-path-relative URL or false if
548 * no valid URL can be constructed
549 */
550 function wfExpandUrl( $url, $defaultProto = PROTO_CURRENT ) {
551 global $wgServer, $wgCanonicalServer, $wgInternalServer, $wgRequest,
552 $wgHttpsPort;
553 if ( $defaultProto === PROTO_CANONICAL ) {
554 $serverUrl = $wgCanonicalServer;
555 } elseif ( $defaultProto === PROTO_INTERNAL && $wgInternalServer !== false ) {
556 // Make $wgInternalServer fall back to $wgServer if not set
557 $serverUrl = $wgInternalServer;
558 } else {
559 $serverUrl = $wgServer;
560 if ( $defaultProto === PROTO_CURRENT ) {
561 $defaultProto = $wgRequest->getProtocol() . '://';
562 }
563 }
564
565 // Analyze $serverUrl to obtain its protocol
566 $bits = wfParseUrl( $serverUrl );
567 $serverHasProto = $bits && $bits['scheme'] != '';
568
569 if ( $defaultProto === PROTO_CANONICAL || $defaultProto === PROTO_INTERNAL ) {
570 if ( $serverHasProto ) {
571 $defaultProto = $bits['scheme'] . '://';
572 } else {
573 // $wgCanonicalServer or $wgInternalServer doesn't have a protocol.
574 // This really isn't supposed to happen. Fall back to HTTP in this
575 // ridiculous case.
576 $defaultProto = PROTO_HTTP;
577 }
578 }
579
580 $defaultProtoWithoutSlashes = substr( $defaultProto, 0, -2 );
581
582 if ( substr( $url, 0, 2 ) == '//' ) {
583 $url = $defaultProtoWithoutSlashes . $url;
584 } elseif ( substr( $url, 0, 1 ) == '/' ) {
585 // If $serverUrl is protocol-relative, prepend $defaultProtoWithoutSlashes,
586 // otherwise leave it alone.
587 $url = ( $serverHasProto ? '' : $defaultProtoWithoutSlashes ) . $serverUrl . $url;
588 }
589
590 $bits = wfParseUrl( $url );
591
592 // ensure proper port for HTTPS arrives in URL
593 // https://phabricator.wikimedia.org/T67184
594 if ( $defaultProto === PROTO_HTTPS && $wgHttpsPort != 443 ) {
595 $bits['port'] = $wgHttpsPort;
596 }
597
598 if ( $bits && isset( $bits['path'] ) ) {
599 $bits['path'] = wfRemoveDotSegments( $bits['path'] );
600 return wfAssembleUrl( $bits );
601 } elseif ( $bits ) {
602 # No path to expand
603 return $url;
604 } elseif ( substr( $url, 0, 1 ) != '/' ) {
605 # URL is a relative path
606 return wfRemoveDotSegments( $url );
607 }
608
609 # Expanded URL is not valid.
610 return false;
611 }
612
613 /**
614 * This function will reassemble a URL parsed with wfParseURL. This is useful
615 * if you need to edit part of a URL and put it back together.
616 *
617 * This is the basic structure used (brackets contain keys for $urlParts):
618 * [scheme][delimiter][user]:[pass]@[host]:[port][path]?[query]#[fragment]
619 *
620 * @todo Need to integrate this into wfExpandUrl (see T34168)
621 *
622 * @since 1.19
623 * @param array $urlParts URL parts, as output from wfParseUrl
624 * @return string URL assembled from its component parts
625 */
626 function wfAssembleUrl( $urlParts ) {
627 $result = '';
628
629 if ( isset( $urlParts['delimiter'] ) ) {
630 if ( isset( $urlParts['scheme'] ) ) {
631 $result .= $urlParts['scheme'];
632 }
633
634 $result .= $urlParts['delimiter'];
635 }
636
637 if ( isset( $urlParts['host'] ) ) {
638 if ( isset( $urlParts['user'] ) ) {
639 $result .= $urlParts['user'];
640 if ( isset( $urlParts['pass'] ) ) {
641 $result .= ':' . $urlParts['pass'];
642 }
643 $result .= '@';
644 }
645
646 $result .= $urlParts['host'];
647
648 if ( isset( $urlParts['port'] ) ) {
649 $result .= ':' . $urlParts['port'];
650 }
651 }
652
653 if ( isset( $urlParts['path'] ) ) {
654 $result .= $urlParts['path'];
655 }
656
657 if ( isset( $urlParts['query'] ) ) {
658 $result .= '?' . $urlParts['query'];
659 }
660
661 if ( isset( $urlParts['fragment'] ) ) {
662 $result .= '#' . $urlParts['fragment'];
663 }
664
665 return $result;
666 }
667
668 /**
669 * Remove all dot-segments in the provided URL path. For example,
670 * '/a/./b/../c/' becomes '/a/c/'. For details on the algorithm, please see
671 * RFC3986 section 5.2.4.
672 *
673 * @todo Need to integrate this into wfExpandUrl (see T34168)
674 *
675 * @param string $urlPath URL path, potentially containing dot-segments
676 * @return string URL path with all dot-segments removed
677 */
678 function wfRemoveDotSegments( $urlPath ) {
679 $output = '';
680 $inputOffset = 0;
681 $inputLength = strlen( $urlPath );
682
683 while ( $inputOffset < $inputLength ) {
684 $prefixLengthOne = substr( $urlPath, $inputOffset, 1 );
685 $prefixLengthTwo = substr( $urlPath, $inputOffset, 2 );
686 $prefixLengthThree = substr( $urlPath, $inputOffset, 3 );
687 $prefixLengthFour = substr( $urlPath, $inputOffset, 4 );
688 $trimOutput = false;
689
690 if ( $prefixLengthTwo == './' ) {
691 # Step A, remove leading "./"
692 $inputOffset += 2;
693 } elseif ( $prefixLengthThree == '../' ) {
694 # Step A, remove leading "../"
695 $inputOffset += 3;
696 } elseif ( ( $prefixLengthTwo == '/.' ) && ( $inputOffset + 2 == $inputLength ) ) {
697 # Step B, replace leading "/.$" with "/"
698 $inputOffset += 1;
699 $urlPath[$inputOffset] = '/';
700 } elseif ( $prefixLengthThree == '/./' ) {
701 # Step B, replace leading "/./" with "/"
702 $inputOffset += 2;
703 } elseif ( $prefixLengthThree == '/..' && ( $inputOffset + 3 == $inputLength ) ) {
704 # Step C, replace leading "/..$" with "/" and
705 # remove last path component in output
706 $inputOffset += 2;
707 $urlPath[$inputOffset] = '/';
708 $trimOutput = true;
709 } elseif ( $prefixLengthFour == '/../' ) {
710 # Step C, replace leading "/../" with "/" and
711 # remove last path component in output
712 $inputOffset += 3;
713 $trimOutput = true;
714 } elseif ( ( $prefixLengthOne == '.' ) && ( $inputOffset + 1 == $inputLength ) ) {
715 # Step D, remove "^.$"
716 $inputOffset += 1;
717 } elseif ( ( $prefixLengthTwo == '..' ) && ( $inputOffset + 2 == $inputLength ) ) {
718 # Step D, remove "^..$"
719 $inputOffset += 2;
720 } else {
721 # Step E, move leading path segment to output
722 if ( $prefixLengthOne == '/' ) {
723 $slashPos = strpos( $urlPath, '/', $inputOffset + 1 );
724 } else {
725 $slashPos = strpos( $urlPath, '/', $inputOffset );
726 }
727 if ( $slashPos === false ) {
728 $output .= substr( $urlPath, $inputOffset );
729 $inputOffset = $inputLength;
730 } else {
731 $output .= substr( $urlPath, $inputOffset, $slashPos - $inputOffset );
732 $inputOffset += $slashPos - $inputOffset;
733 }
734 }
735
736 if ( $trimOutput ) {
737 $slashPos = strrpos( $output, '/' );
738 if ( $slashPos === false ) {
739 $output = '';
740 } else {
741 $output = substr( $output, 0, $slashPos );
742 }
743 }
744 }
745
746 return $output;
747 }
748
749 /**
750 * Returns a regular expression of url protocols
751 *
752 * @param bool $includeProtocolRelative If false, remove '//' from the returned protocol list.
753 * DO NOT USE this directly, use wfUrlProtocolsWithoutProtRel() instead
754 * @return string
755 */
756 function wfUrlProtocols( $includeProtocolRelative = true ) {
757 global $wgUrlProtocols;
758
759 // Cache return values separately based on $includeProtocolRelative
760 static $withProtRel = null, $withoutProtRel = null;
761 $cachedValue = $includeProtocolRelative ? $withProtRel : $withoutProtRel;
762 if ( !is_null( $cachedValue ) ) {
763 return $cachedValue;
764 }
765
766 // Support old-style $wgUrlProtocols strings, for backwards compatibility
767 // with LocalSettings files from 1.5
768 if ( is_array( $wgUrlProtocols ) ) {
769 $protocols = [];
770 foreach ( $wgUrlProtocols as $protocol ) {
771 // Filter out '//' if !$includeProtocolRelative
772 if ( $includeProtocolRelative || $protocol !== '//' ) {
773 $protocols[] = preg_quote( $protocol, '/' );
774 }
775 }
776
777 $retval = implode( '|', $protocols );
778 } else {
779 // Ignore $includeProtocolRelative in this case
780 // This case exists for pre-1.6 compatibility, and we can safely assume
781 // that '//' won't appear in a pre-1.6 config because protocol-relative
782 // URLs weren't supported until 1.18
783 $retval = $wgUrlProtocols;
784 }
785
786 // Cache return value
787 if ( $includeProtocolRelative ) {
788 $withProtRel = $retval;
789 } else {
790 $withoutProtRel = $retval;
791 }
792 return $retval;
793 }
794
795 /**
796 * Like wfUrlProtocols(), but excludes '//' from the protocol list. Use this if
797 * you need a regex that matches all URL protocols but does not match protocol-
798 * relative URLs
799 * @return string
800 */
801 function wfUrlProtocolsWithoutProtRel() {
802 return wfUrlProtocols( false );
803 }
804
805 /**
806 * parse_url() work-alike, but non-broken. Differences:
807 *
808 * 1) Does not raise warnings on bad URLs (just returns false).
809 * 2) Handles protocols that don't use :// (e.g., mailto: and news:, as well as
810 * protocol-relative URLs) correctly.
811 * 3) Adds a "delimiter" element to the array, either '://', ':' or '//' (see (2)).
812 *
813 * @param string $url A URL to parse
814 * @return string[]|bool Bits of the URL in an associative array, per PHP docs, false on failure
815 */
816 function wfParseUrl( $url ) {
817 global $wgUrlProtocols; // Allow all protocols defined in DefaultSettings/LocalSettings.php
818
819 // Protocol-relative URLs are handled really badly by parse_url(). It's so
820 // bad that the easiest way to handle them is to just prepend 'http:' and
821 // strip the protocol out later.
822 $wasRelative = substr( $url, 0, 2 ) == '//';
823 if ( $wasRelative ) {
824 $url = "http:$url";
825 }
826 MediaWiki\suppressWarnings();
827 $bits = parse_url( $url );
828 MediaWiki\restoreWarnings();
829 // parse_url() returns an array without scheme for some invalid URLs, e.g.
830 // parse_url("%0Ahttp://example.com") == [ 'host' => '%0Ahttp', 'path' => 'example.com' ]
831 if ( !$bits || !isset( $bits['scheme'] ) ) {
832 return false;
833 }
834
835 // parse_url() incorrectly handles schemes case-sensitively. Convert it to lowercase.
836 $bits['scheme'] = strtolower( $bits['scheme'] );
837
838 // most of the protocols are followed by ://, but mailto: and sometimes news: not, check for it
839 if ( in_array( $bits['scheme'] . '://', $wgUrlProtocols ) ) {
840 $bits['delimiter'] = '://';
841 } elseif ( in_array( $bits['scheme'] . ':', $wgUrlProtocols ) ) {
842 $bits['delimiter'] = ':';
843 // parse_url detects for news: and mailto: the host part of an url as path
844 // We have to correct this wrong detection
845 if ( isset( $bits['path'] ) ) {
846 $bits['host'] = $bits['path'];
847 $bits['path'] = '';
848 }
849 } else {
850 return false;
851 }
852
853 /* Provide an empty host for eg. file:/// urls (see T30627) */
854 if ( !isset( $bits['host'] ) ) {
855 $bits['host'] = '';
856
857 // See T47069
858 if ( isset( $bits['path'] ) ) {
859 /* parse_url loses the third / for file:///c:/ urls (but not on variants) */
860 if ( substr( $bits['path'], 0, 1 ) !== '/' ) {
861 $bits['path'] = '/' . $bits['path'];
862 }
863 } else {
864 $bits['path'] = '';
865 }
866 }
867
868 // If the URL was protocol-relative, fix scheme and delimiter
869 if ( $wasRelative ) {
870 $bits['scheme'] = '';
871 $bits['delimiter'] = '//';
872 }
873 return $bits;
874 }
875
876 /**
877 * Take a URL, make sure it's expanded to fully qualified, and replace any
878 * encoded non-ASCII Unicode characters with their UTF-8 original forms
879 * for more compact display and legibility for local audiences.
880 *
881 * @todo handle punycode domains too
882 *
883 * @param string $url
884 * @return string
885 */
886 function wfExpandIRI( $url ) {
887 return preg_replace_callback(
888 '/((?:%[89A-F][0-9A-F])+)/i',
889 'wfExpandIRI_callback',
890 wfExpandUrl( $url )
891 );
892 }
893
894 /**
895 * Private callback for wfExpandIRI
896 * @param array $matches
897 * @return string
898 */
899 function wfExpandIRI_callback( $matches ) {
900 return urldecode( $matches[1] );
901 }
902
903 /**
904 * Make URL indexes, appropriate for the el_index field of externallinks.
905 *
906 * @param string $url
907 * @return array
908 */
909 function wfMakeUrlIndexes( $url ) {
910 $bits = wfParseUrl( $url );
911
912 // Reverse the labels in the hostname, convert to lower case
913 // For emails reverse domainpart only
914 if ( $bits['scheme'] == 'mailto' ) {
915 $mailparts = explode( '@', $bits['host'], 2 );
916 if ( count( $mailparts ) === 2 ) {
917 $domainpart = strtolower( implode( '.', array_reverse( explode( '.', $mailparts[1] ) ) ) );
918 } else {
919 // No domain specified, don't mangle it
920 $domainpart = '';
921 }
922 $reversedHost = $domainpart . '@' . $mailparts[0];
923 } else {
924 $reversedHost = strtolower( implode( '.', array_reverse( explode( '.', $bits['host'] ) ) ) );
925 }
926 // Add an extra dot to the end
927 // Why? Is it in wrong place in mailto links?
928 if ( substr( $reversedHost, -1, 1 ) !== '.' ) {
929 $reversedHost .= '.';
930 }
931 // Reconstruct the pseudo-URL
932 $prot = $bits['scheme'];
933 $index = $prot . $bits['delimiter'] . $reversedHost;
934 // Leave out user and password. Add the port, path, query and fragment
935 if ( isset( $bits['port'] ) ) {
936 $index .= ':' . $bits['port'];
937 }
938 if ( isset( $bits['path'] ) ) {
939 $index .= $bits['path'];
940 } else {
941 $index .= '/';
942 }
943 if ( isset( $bits['query'] ) ) {
944 $index .= '?' . $bits['query'];
945 }
946 if ( isset( $bits['fragment'] ) ) {
947 $index .= '#' . $bits['fragment'];
948 }
949
950 if ( $prot == '' ) {
951 return [ "http:$index", "https:$index" ];
952 } else {
953 return [ $index ];
954 }
955 }
956
957 /**
958 * Check whether a given URL has a domain that occurs in a given set of domains
959 * @param string $url URL
960 * @param array $domains Array of domains (strings)
961 * @return bool True if the host part of $url ends in one of the strings in $domains
962 */
963 function wfMatchesDomainList( $url, $domains ) {
964 $bits = wfParseUrl( $url );
965 if ( is_array( $bits ) && isset( $bits['host'] ) ) {
966 $host = '.' . $bits['host'];
967 foreach ( (array)$domains as $domain ) {
968 $domain = '.' . $domain;
969 if ( substr( $host, -strlen( $domain ) ) === $domain ) {
970 return true;
971 }
972 }
973 }
974 return false;
975 }
976
977 /**
978 * Sends a line to the debug log if enabled or, optionally, to a comment in output.
979 * In normal operation this is a NOP.
980 *
981 * Controlling globals:
982 * $wgDebugLogFile - points to the log file
983 * $wgDebugRawPage - if false, 'action=raw' hits will not result in debug output.
984 * $wgDebugComments - if on, some debug items may appear in comments in the HTML output.
985 *
986 * @since 1.25 support for additional context data
987 *
988 * @param string $text
989 * @param string|bool $dest Destination of the message:
990 * - 'all': both to the log and HTML (debug toolbar or HTML comments)
991 * - 'private': excluded from HTML output
992 * For backward compatibility, it can also take a boolean:
993 * - true: same as 'all'
994 * - false: same as 'private'
995 * @param array $context Additional logging context data
996 */
997 function wfDebug( $text, $dest = 'all', array $context = [] ) {
998 global $wgDebugRawPage, $wgDebugLogPrefix;
999 global $wgDebugTimestamps, $wgRequestTime;
1000
1001 if ( !$wgDebugRawPage && wfIsDebugRawPage() ) {
1002 return;
1003 }
1004
1005 $text = trim( $text );
1006
1007 if ( $wgDebugTimestamps ) {
1008 $context['seconds_elapsed'] = sprintf(
1009 '%6.4f',
1010 microtime( true ) - $wgRequestTime
1011 );
1012 $context['memory_used'] = sprintf(
1013 '%5.1fM',
1014 ( memory_get_usage( true ) / ( 1024 * 1024 ) )
1015 );
1016 }
1017
1018 if ( $wgDebugLogPrefix !== '' ) {
1019 $context['prefix'] = $wgDebugLogPrefix;
1020 }
1021 $context['private'] = ( $dest === false || $dest === 'private' );
1022
1023 $logger = LoggerFactory::getInstance( 'wfDebug' );
1024 $logger->debug( $text, $context );
1025 }
1026
1027 /**
1028 * Returns true if debug logging should be suppressed if $wgDebugRawPage = false
1029 * @return bool
1030 */
1031 function wfIsDebugRawPage() {
1032 static $cache;
1033 if ( $cache !== null ) {
1034 return $cache;
1035 }
1036 # Check for raw action using $_GET not $wgRequest, since the latter might not be initialised yet
1037 if ( ( isset( $_GET['action'] ) && $_GET['action'] == 'raw' )
1038 || (
1039 isset( $_SERVER['SCRIPT_NAME'] )
1040 && substr( $_SERVER['SCRIPT_NAME'], -8 ) == 'load.php'
1041 )
1042 ) {
1043 $cache = true;
1044 } else {
1045 $cache = false;
1046 }
1047 return $cache;
1048 }
1049
1050 /**
1051 * Send a line giving PHP memory usage.
1052 *
1053 * @param bool $exact Print exact byte values instead of kibibytes (default: false)
1054 */
1055 function wfDebugMem( $exact = false ) {
1056 $mem = memory_get_usage();
1057 if ( !$exact ) {
1058 $mem = floor( $mem / 1024 ) . ' KiB';
1059 } else {
1060 $mem .= ' B';
1061 }
1062 wfDebug( "Memory usage: $mem\n" );
1063 }
1064
1065 /**
1066 * Send a line to a supplementary debug log file, if configured, or main debug
1067 * log if not.
1068 *
1069 * To configure a supplementary log file, set $wgDebugLogGroups[$logGroup] to
1070 * a string filename or an associative array mapping 'destination' to the
1071 * desired filename. The associative array may also contain a 'sample' key
1072 * with an integer value, specifying a sampling factor. Sampled log events
1073 * will be emitted with a 1 in N random chance.
1074 *
1075 * @since 1.23 support for sampling log messages via $wgDebugLogGroups.
1076 * @since 1.25 support for additional context data
1077 * @since 1.25 sample behavior dependent on configured $wgMWLoggerDefaultSpi
1078 *
1079 * @param string $logGroup
1080 * @param string $text
1081 * @param string|bool $dest Destination of the message:
1082 * - 'all': both to the log and HTML (debug toolbar or HTML comments)
1083 * - 'private': only to the specific log if set in $wgDebugLogGroups and
1084 * discarded otherwise
1085 * For backward compatibility, it can also take a boolean:
1086 * - true: same as 'all'
1087 * - false: same as 'private'
1088 * @param array $context Additional logging context data
1089 */
1090 function wfDebugLog(
1091 $logGroup, $text, $dest = 'all', array $context = []
1092 ) {
1093 $text = trim( $text );
1094
1095 $logger = LoggerFactory::getInstance( $logGroup );
1096 $context['private'] = ( $dest === false || $dest === 'private' );
1097 $logger->info( $text, $context );
1098 }
1099
1100 /**
1101 * Log for database errors
1102 *
1103 * @since 1.25 support for additional context data
1104 *
1105 * @param string $text Database error message.
1106 * @param array $context Additional logging context data
1107 */
1108 function wfLogDBError( $text, array $context = [] ) {
1109 $logger = LoggerFactory::getInstance( 'wfLogDBError' );
1110 $logger->error( trim( $text ), $context );
1111 }
1112
1113 /**
1114 * Throws a warning that $function is deprecated
1115 *
1116 * @param string $function
1117 * @param string|bool $version Version of MediaWiki that the function
1118 * was deprecated in (Added in 1.19).
1119 * @param string|bool $component Added in 1.19.
1120 * @param int $callerOffset How far up the call stack is the original
1121 * caller. 2 = function that called the function that called
1122 * wfDeprecated (Added in 1.20)
1123 *
1124 * @return null
1125 */
1126 function wfDeprecated( $function, $version = false, $component = false, $callerOffset = 2 ) {
1127 MWDebug::deprecated( $function, $version, $component, $callerOffset + 1 );
1128 }
1129
1130 /**
1131 * Send a warning either to the debug log or in a PHP error depending on
1132 * $wgDevelopmentWarnings. To log warnings in production, use wfLogWarning() instead.
1133 *
1134 * @param string $msg Message to send
1135 * @param int $callerOffset Number of items to go back in the backtrace to
1136 * find the correct caller (1 = function calling wfWarn, ...)
1137 * @param int $level PHP error level; defaults to E_USER_NOTICE;
1138 * only used when $wgDevelopmentWarnings is true
1139 */
1140 function wfWarn( $msg, $callerOffset = 1, $level = E_USER_NOTICE ) {
1141 MWDebug::warning( $msg, $callerOffset + 1, $level, 'auto' );
1142 }
1143
1144 /**
1145 * Send a warning as a PHP error and the debug log. This is intended for logging
1146 * warnings in production. For logging development warnings, use WfWarn instead.
1147 *
1148 * @param string $msg Message to send
1149 * @param int $callerOffset Number of items to go back in the backtrace to
1150 * find the correct caller (1 = function calling wfLogWarning, ...)
1151 * @param int $level PHP error level; defaults to E_USER_WARNING
1152 */
1153 function wfLogWarning( $msg, $callerOffset = 1, $level = E_USER_WARNING ) {
1154 MWDebug::warning( $msg, $callerOffset + 1, $level, 'production' );
1155 }
1156
1157 /**
1158 * Log to a file without getting "file size exceeded" signals.
1159 *
1160 * Can also log to TCP or UDP with the syntax udp://host:port/prefix. This will
1161 * send lines to the specified port, prefixed by the specified prefix and a space.
1162 * @since 1.25 support for additional context data
1163 *
1164 * @param string $text
1165 * @param string $file Filename
1166 * @param array $context Additional logging context data
1167 * @throws MWException
1168 * @deprecated since 1.25 Use \MediaWiki\Logger\LegacyLogger::emit or UDPTransport
1169 */
1170 function wfErrorLog( $text, $file, array $context = [] ) {
1171 wfDeprecated( __METHOD__, '1.25' );
1172 $logger = LoggerFactory::getInstance( 'wfErrorLog' );
1173 $context['destination'] = $file;
1174 $logger->info( trim( $text ), $context );
1175 }
1176
1177 /**
1178 * @todo document
1179 */
1180 function wfLogProfilingData() {
1181 global $wgDebugLogGroups, $wgDebugRawPage;
1182
1183 $context = RequestContext::getMain();
1184 $request = $context->getRequest();
1185
1186 $profiler = Profiler::instance();
1187 $profiler->setContext( $context );
1188 $profiler->logData();
1189
1190 $config = $context->getConfig();
1191 if ( $config->get( 'StatsdServer' ) ) {
1192 try {
1193 $statsdServer = explode( ':', $config->get( 'StatsdServer' ) );
1194 $statsdHost = $statsdServer[0];
1195 $statsdPort = isset( $statsdServer[1] ) ? $statsdServer[1] : 8125;
1196 $statsdSender = new SocketSender( $statsdHost, $statsdPort );
1197 $statsdClient = new SamplingStatsdClient( $statsdSender, true, false );
1198 $statsdClient->setSamplingRates( $config->get( 'StatsdSamplingRates' ) );
1199 $statsdClient->send( $context->getStats()->getBuffer() );
1200 } catch ( Exception $ex ) {
1201 MWExceptionHandler::logException( $ex );
1202 }
1203 }
1204
1205 # Profiling must actually be enabled...
1206 if ( $profiler instanceof ProfilerStub ) {
1207 return;
1208 }
1209
1210 if ( isset( $wgDebugLogGroups['profileoutput'] )
1211 && $wgDebugLogGroups['profileoutput'] === false
1212 ) {
1213 // Explicitly disabled
1214 return;
1215 }
1216 if ( !$wgDebugRawPage && wfIsDebugRawPage() ) {
1217 return;
1218 }
1219
1220 $ctx = [ 'elapsed' => $request->getElapsedTime() ];
1221 if ( !empty( $_SERVER['HTTP_X_FORWARDED_FOR'] ) ) {
1222 $ctx['forwarded_for'] = $_SERVER['HTTP_X_FORWARDED_FOR'];
1223 }
1224 if ( !empty( $_SERVER['HTTP_CLIENT_IP'] ) ) {
1225 $ctx['client_ip'] = $_SERVER['HTTP_CLIENT_IP'];
1226 }
1227 if ( !empty( $_SERVER['HTTP_FROM'] ) ) {
1228 $ctx['from'] = $_SERVER['HTTP_FROM'];
1229 }
1230 if ( isset( $ctx['forwarded_for'] ) ||
1231 isset( $ctx['client_ip'] ) ||
1232 isset( $ctx['from'] ) ) {
1233 $ctx['proxy'] = $_SERVER['REMOTE_ADDR'];
1234 }
1235
1236 // Don't load $wgUser at this late stage just for statistics purposes
1237 // @todo FIXME: We can detect some anons even if it is not loaded.
1238 // See User::getId()
1239 $user = $context->getUser();
1240 $ctx['anon'] = $user->isItemLoaded( 'id' ) && $user->isAnon();
1241
1242 // Command line script uses a FauxRequest object which does not have
1243 // any knowledge about an URL and throw an exception instead.
1244 try {
1245 $ctx['url'] = urldecode( $request->getRequestURL() );
1246 } catch ( Exception $ignored ) {
1247 // no-op
1248 }
1249
1250 $ctx['output'] = $profiler->getOutput();
1251
1252 $log = LoggerFactory::getInstance( 'profileoutput' );
1253 $log->info( "Elapsed: {elapsed}; URL: <{url}>\n{output}", $ctx );
1254 }
1255
1256 /**
1257 * Increment a statistics counter
1258 *
1259 * @param string $key
1260 * @param int $count
1261 * @return void
1262 */
1263 function wfIncrStats( $key, $count = 1 ) {
1264 $stats = RequestContext::getMain()->getStats();
1265 $stats->updateCount( $key, $count );
1266 }
1267
1268 /**
1269 * Check whether the wiki is in read-only mode.
1270 *
1271 * @return bool
1272 */
1273 function wfReadOnly() {
1274 return wfReadOnlyReason() !== false;
1275 }
1276
1277 /**
1278 * Check if the site is in read-only mode and return the message if so
1279 *
1280 * This checks wfConfiguredReadOnlyReason() and the main load balancer
1281 * for replica DB lag. This may result in DB connection being made.
1282 *
1283 * @return string|bool String when in read-only mode; false otherwise
1284 */
1285 function wfReadOnlyReason() {
1286 $readOnly = wfConfiguredReadOnlyReason();
1287 if ( $readOnly !== false ) {
1288 return $readOnly;
1289 }
1290
1291 static $lbReadOnly = null;
1292 if ( $lbReadOnly === null ) {
1293 // Callers use this method to be aware that data presented to a user
1294 // may be very stale and thus allowing submissions can be problematic.
1295 $lbReadOnly = wfGetLB()->getReadOnlyReason();
1296 }
1297
1298 return $lbReadOnly;
1299 }
1300
1301 /**
1302 * Get the value of $wgReadOnly or the contents of $wgReadOnlyFile.
1303 *
1304 * @return string|bool String when in read-only mode; false otherwise
1305 * @since 1.27
1306 */
1307 function wfConfiguredReadOnlyReason() {
1308 global $wgReadOnly, $wgReadOnlyFile;
1309
1310 if ( $wgReadOnly === null ) {
1311 // Set $wgReadOnly for faster access next time
1312 if ( is_file( $wgReadOnlyFile ) && filesize( $wgReadOnlyFile ) > 0 ) {
1313 $wgReadOnly = file_get_contents( $wgReadOnlyFile );
1314 } else {
1315 $wgReadOnly = false;
1316 }
1317 }
1318
1319 return $wgReadOnly;
1320 }
1321
1322 /**
1323 * Return a Language object from $langcode
1324 *
1325 * @param Language|string|bool $langcode Either:
1326 * - a Language object
1327 * - code of the language to get the message for, if it is
1328 * a valid code create a language for that language, if
1329 * it is a string but not a valid code then make a basic
1330 * language object
1331 * - a boolean: if it's false then use the global object for
1332 * the current user's language (as a fallback for the old parameter
1333 * functionality), or if it is true then use global object
1334 * for the wiki's content language.
1335 * @return Language
1336 */
1337 function wfGetLangObj( $langcode = false ) {
1338 # Identify which language to get or create a language object for.
1339 # Using is_object here due to Stub objects.
1340 if ( is_object( $langcode ) ) {
1341 # Great, we already have the object (hopefully)!
1342 return $langcode;
1343 }
1344
1345 global $wgContLang, $wgLanguageCode;
1346 if ( $langcode === true || $langcode === $wgLanguageCode ) {
1347 # $langcode is the language code of the wikis content language object.
1348 # or it is a boolean and value is true
1349 return $wgContLang;
1350 }
1351
1352 global $wgLang;
1353 if ( $langcode === false || $langcode === $wgLang->getCode() ) {
1354 # $langcode is the language code of user language object.
1355 # or it was a boolean and value is false
1356 return $wgLang;
1357 }
1358
1359 $validCodes = array_keys( Language::fetchLanguageNames() );
1360 if ( in_array( $langcode, $validCodes ) ) {
1361 # $langcode corresponds to a valid language.
1362 return Language::factory( $langcode );
1363 }
1364
1365 # $langcode is a string, but not a valid language code; use content language.
1366 wfDebug( "Invalid language code passed to wfGetLangObj, falling back to content language.\n" );
1367 return $wgContLang;
1368 }
1369
1370 /**
1371 * This is the function for getting translated interface messages.
1372 *
1373 * @see Message class for documentation how to use them.
1374 * @see https://www.mediawiki.org/wiki/Manual:Messages_API
1375 *
1376 * This function replaces all old wfMsg* functions.
1377 *
1378 * @param string|string[]|MessageSpecifier $key Message key, or array of keys, or a MessageSpecifier
1379 * @param mixed $params,... Normal message parameters
1380 * @return Message
1381 *
1382 * @since 1.17
1383 *
1384 * @see Message::__construct
1385 */
1386 function wfMessage( $key /*...*/ ) {
1387 $params = func_get_args();
1388 array_shift( $params );
1389 if ( isset( $params[0] ) && is_array( $params[0] ) ) {
1390 $params = $params[0];
1391 }
1392 return new Message( $key, $params );
1393 }
1394
1395 /**
1396 * This function accepts multiple message keys and returns a message instance
1397 * for the first message which is non-empty. If all messages are empty then an
1398 * instance of the first message key is returned.
1399 *
1400 * @param string|string[] $keys,... Message keys
1401 * @return Message
1402 *
1403 * @since 1.18
1404 *
1405 * @see Message::newFallbackSequence
1406 */
1407 function wfMessageFallback( /*...*/ ) {
1408 $args = func_get_args();
1409 return call_user_func_array( 'Message::newFallbackSequence', $args );
1410 }
1411
1412 /**
1413 * Replace message parameter keys on the given formatted output.
1414 *
1415 * @param string $message
1416 * @param array $args
1417 * @return string
1418 * @private
1419 */
1420 function wfMsgReplaceArgs( $message, $args ) {
1421 # Fix windows line-endings
1422 # Some messages are split with explode("\n", $msg)
1423 $message = str_replace( "\r", '', $message );
1424
1425 // Replace arguments
1426 if ( is_array( $args ) && $args ) {
1427 if ( is_array( $args[0] ) ) {
1428 $args = array_values( $args[0] );
1429 }
1430 $replacementKeys = [];
1431 foreach ( $args as $n => $param ) {
1432 $replacementKeys['$' . ( $n + 1 )] = $param;
1433 }
1434 $message = strtr( $message, $replacementKeys );
1435 }
1436
1437 return $message;
1438 }
1439
1440 /**
1441 * Fetch server name for use in error reporting etc.
1442 * Use real server name if available, so we know which machine
1443 * in a server farm generated the current page.
1444 *
1445 * @return string
1446 */
1447 function wfHostname() {
1448 static $host;
1449 if ( is_null( $host ) ) {
1450
1451 # Hostname overriding
1452 global $wgOverrideHostname;
1453 if ( $wgOverrideHostname !== false ) {
1454 # Set static and skip any detection
1455 $host = $wgOverrideHostname;
1456 return $host;
1457 }
1458
1459 if ( function_exists( 'posix_uname' ) ) {
1460 // This function not present on Windows
1461 $uname = posix_uname();
1462 } else {
1463 $uname = false;
1464 }
1465 if ( is_array( $uname ) && isset( $uname['nodename'] ) ) {
1466 $host = $uname['nodename'];
1467 } elseif ( getenv( 'COMPUTERNAME' ) ) {
1468 # Windows computer name
1469 $host = getenv( 'COMPUTERNAME' );
1470 } else {
1471 # This may be a virtual server.
1472 $host = $_SERVER['SERVER_NAME'];
1473 }
1474 }
1475 return $host;
1476 }
1477
1478 /**
1479 * Returns a script tag that stores the amount of time it took MediaWiki to
1480 * handle the request in milliseconds as 'wgBackendResponseTime'.
1481 *
1482 * If $wgShowHostnames is true, the script will also set 'wgHostname' to the
1483 * hostname of the server handling the request.
1484 *
1485 * @return string
1486 */
1487 function wfReportTime() {
1488 global $wgRequestTime, $wgShowHostnames;
1489
1490 $responseTime = round( ( microtime( true ) - $wgRequestTime ) * 1000 );
1491 $reportVars = [ 'wgBackendResponseTime' => $responseTime ];
1492 if ( $wgShowHostnames ) {
1493 $reportVars['wgHostname'] = wfHostname();
1494 }
1495 return Skin::makeVariablesScript( $reportVars );
1496 }
1497
1498 /**
1499 * Safety wrapper for debug_backtrace().
1500 *
1501 * Will return an empty array if debug_backtrace is disabled, otherwise
1502 * the output from debug_backtrace() (trimmed).
1503 *
1504 * @param int $limit This parameter can be used to limit the number of stack frames returned
1505 *
1506 * @return array Array of backtrace information
1507 */
1508 function wfDebugBacktrace( $limit = 0 ) {
1509 static $disabled = null;
1510
1511 if ( is_null( $disabled ) ) {
1512 $disabled = !function_exists( 'debug_backtrace' );
1513 if ( $disabled ) {
1514 wfDebug( "debug_backtrace() is disabled\n" );
1515 }
1516 }
1517 if ( $disabled ) {
1518 return [];
1519 }
1520
1521 if ( $limit ) {
1522 return array_slice( debug_backtrace( DEBUG_BACKTRACE_PROVIDE_OBJECT, $limit + 1 ), 1 );
1523 } else {
1524 return array_slice( debug_backtrace(), 1 );
1525 }
1526 }
1527
1528 /**
1529 * Get a debug backtrace as a string
1530 *
1531 * @param bool|null $raw If true, the return value is plain text. If false, HTML.
1532 * Defaults to $wgCommandLineMode if unset.
1533 * @return string
1534 * @since 1.25 Supports $raw parameter.
1535 */
1536 function wfBacktrace( $raw = null ) {
1537 global $wgCommandLineMode;
1538
1539 if ( $raw === null ) {
1540 $raw = $wgCommandLineMode;
1541 }
1542
1543 if ( $raw ) {
1544 $frameFormat = "%s line %s calls %s()\n";
1545 $traceFormat = "%s";
1546 } else {
1547 $frameFormat = "<li>%s line %s calls %s()</li>\n";
1548 $traceFormat = "<ul>\n%s</ul>\n";
1549 }
1550
1551 $frames = array_map( function ( $frame ) use ( $frameFormat ) {
1552 $file = !empty( $frame['file'] ) ? basename( $frame['file'] ) : '-';
1553 $line = isset( $frame['line'] ) ? $frame['line'] : '-';
1554 $call = $frame['function'];
1555 if ( !empty( $frame['class'] ) ) {
1556 $call = $frame['class'] . $frame['type'] . $call;
1557 }
1558 return sprintf( $frameFormat, $file, $line, $call );
1559 }, wfDebugBacktrace() );
1560
1561 return sprintf( $traceFormat, implode( '', $frames ) );
1562 }
1563
1564 /**
1565 * Get the name of the function which called this function
1566 * wfGetCaller( 1 ) is the function with the wfGetCaller() call (ie. __FUNCTION__)
1567 * wfGetCaller( 2 ) [default] is the caller of the function running wfGetCaller()
1568 * wfGetCaller( 3 ) is the parent of that.
1569 *
1570 * @param int $level
1571 * @return string
1572 */
1573 function wfGetCaller( $level = 2 ) {
1574 $backtrace = wfDebugBacktrace( $level + 1 );
1575 if ( isset( $backtrace[$level] ) ) {
1576 return wfFormatStackFrame( $backtrace[$level] );
1577 } else {
1578 return 'unknown';
1579 }
1580 }
1581
1582 /**
1583 * Return a string consisting of callers in the stack. Useful sometimes
1584 * for profiling specific points.
1585 *
1586 * @param int $limit The maximum depth of the stack frame to return, or false for the entire stack.
1587 * @return string
1588 */
1589 function wfGetAllCallers( $limit = 3 ) {
1590 $trace = array_reverse( wfDebugBacktrace() );
1591 if ( !$limit || $limit > count( $trace ) - 1 ) {
1592 $limit = count( $trace ) - 1;
1593 }
1594 $trace = array_slice( $trace, -$limit - 1, $limit );
1595 return implode( '/', array_map( 'wfFormatStackFrame', $trace ) );
1596 }
1597
1598 /**
1599 * Return a string representation of frame
1600 *
1601 * @param array $frame
1602 * @return string
1603 */
1604 function wfFormatStackFrame( $frame ) {
1605 if ( !isset( $frame['function'] ) ) {
1606 return 'NO_FUNCTION_GIVEN';
1607 }
1608 return isset( $frame['class'] ) && isset( $frame['type'] ) ?
1609 $frame['class'] . $frame['type'] . $frame['function'] :
1610 $frame['function'];
1611 }
1612
1613 /* Some generic result counters, pulled out of SearchEngine */
1614
1615 /**
1616 * @todo document
1617 *
1618 * @param int $offset
1619 * @param int $limit
1620 * @return string
1621 */
1622 function wfShowingResults( $offset, $limit ) {
1623 return wfMessage( 'showingresults' )->numParams( $limit, $offset + 1 )->parse();
1624 }
1625
1626 /**
1627 * @todo document
1628 * @todo FIXME: We may want to blacklist some broken browsers
1629 *
1630 * @param bool $force
1631 * @return bool Whereas client accept gzip compression
1632 */
1633 function wfClientAcceptsGzip( $force = false ) {
1634 static $result = null;
1635 if ( $result === null || $force ) {
1636 $result = false;
1637 if ( isset( $_SERVER['HTTP_ACCEPT_ENCODING'] ) ) {
1638 # @todo FIXME: We may want to blacklist some broken browsers
1639 $m = [];
1640 if ( preg_match(
1641 '/\bgzip(?:;(q)=([0-9]+(?:\.[0-9]+)))?\b/',
1642 $_SERVER['HTTP_ACCEPT_ENCODING'],
1643 $m
1644 )
1645 ) {
1646 if ( isset( $m[2] ) && ( $m[1] == 'q' ) && ( $m[2] == 0 ) ) {
1647 $result = false;
1648 return $result;
1649 }
1650 wfDebug( "wfClientAcceptsGzip: client accepts gzip.\n" );
1651 $result = true;
1652 }
1653 }
1654 }
1655 return $result;
1656 }
1657
1658 /**
1659 * @deprecated since 1.28, use Parser::escapeWikitext() directly
1660 *
1661 * Escapes the given text so that it may be output using addWikiText()
1662 * without any linking, formatting, etc. making its way through. This
1663 * is achieved by substituting certain characters with HTML entities.
1664 * As required by the callers, "<nowiki>" is not used.
1665 *
1666 * @param string $text Text to be escaped
1667 * @return string
1668 */
1669 function wfEscapeWikiText( $text ) {
1670 global $wgParser;
1671 return $wgParser->escapeWikitext( $text );
1672 }
1673
1674 /**
1675 * Sets dest to source and returns the original value of dest
1676 * If source is NULL, it just returns the value, it doesn't set the variable
1677 * If force is true, it will set the value even if source is NULL
1678 *
1679 * @param mixed $dest
1680 * @param mixed $source
1681 * @param bool $force
1682 * @return mixed
1683 */
1684 function wfSetVar( &$dest, $source, $force = false ) {
1685 $temp = $dest;
1686 if ( !is_null( $source ) || $force ) {
1687 $dest = $source;
1688 }
1689 return $temp;
1690 }
1691
1692 /**
1693 * As for wfSetVar except setting a bit
1694 *
1695 * @param int $dest
1696 * @param int $bit
1697 * @param bool $state
1698 *
1699 * @return bool
1700 */
1701 function wfSetBit( &$dest, $bit, $state = true ) {
1702 $temp = (bool)( $dest & $bit );
1703 if ( !is_null( $state ) ) {
1704 if ( $state ) {
1705 $dest |= $bit;
1706 } else {
1707 $dest &= ~$bit;
1708 }
1709 }
1710 return $temp;
1711 }
1712
1713 /**
1714 * A wrapper around the PHP function var_export().
1715 * Either print it or add it to the regular output ($wgOut).
1716 *
1717 * @param mixed $var A PHP variable to dump.
1718 */
1719 function wfVarDump( $var ) {
1720 global $wgOut;
1721 $s = str_replace( "\n", "<br />\n", var_export( $var, true ) . "\n" );
1722 if ( headers_sent() || !isset( $wgOut ) || !is_object( $wgOut ) ) {
1723 print $s;
1724 } else {
1725 $wgOut->addHTML( $s );
1726 }
1727 }
1728
1729 /**
1730 * Provide a simple HTTP error.
1731 *
1732 * @param int|string $code
1733 * @param string $label
1734 * @param string $desc
1735 */
1736 function wfHttpError( $code, $label, $desc ) {
1737 global $wgOut;
1738 HttpStatus::header( $code );
1739 if ( $wgOut ) {
1740 $wgOut->disable();
1741 $wgOut->sendCacheControl();
1742 }
1743
1744 header( 'Content-type: text/html; charset=utf-8' );
1745 print '<!DOCTYPE html>' .
1746 '<html><head><title>' .
1747 htmlspecialchars( $label ) .
1748 '</title></head><body><h1>' .
1749 htmlspecialchars( $label ) .
1750 '</h1><p>' .
1751 nl2br( htmlspecialchars( $desc ) ) .
1752 "</p></body></html>\n";
1753 }
1754
1755 /**
1756 * Clear away any user-level output buffers, discarding contents.
1757 *
1758 * Suitable for 'starting afresh', for instance when streaming
1759 * relatively large amounts of data without buffering, or wanting to
1760 * output image files without ob_gzhandler's compression.
1761 *
1762 * The optional $resetGzipEncoding parameter controls suppression of
1763 * the Content-Encoding header sent by ob_gzhandler; by default it
1764 * is left. See comments for wfClearOutputBuffers() for why it would
1765 * be used.
1766 *
1767 * Note that some PHP configuration options may add output buffer
1768 * layers which cannot be removed; these are left in place.
1769 *
1770 * @param bool $resetGzipEncoding
1771 */
1772 function wfResetOutputBuffers( $resetGzipEncoding = true ) {
1773 if ( $resetGzipEncoding ) {
1774 // Suppress Content-Encoding and Content-Length
1775 // headers from 1.10+s wfOutputHandler
1776 global $wgDisableOutputCompression;
1777 $wgDisableOutputCompression = true;
1778 }
1779 while ( $status = ob_get_status() ) {
1780 if ( isset( $status['flags'] ) ) {
1781 $flags = PHP_OUTPUT_HANDLER_CLEANABLE | PHP_OUTPUT_HANDLER_REMOVABLE;
1782 $deleteable = ( $status['flags'] & $flags ) === $flags;
1783 } elseif ( isset( $status['del'] ) ) {
1784 $deleteable = $status['del'];
1785 } else {
1786 // Guess that any PHP-internal setting can't be removed.
1787 $deleteable = $status['type'] !== 0; /* PHP_OUTPUT_HANDLER_INTERNAL */
1788 }
1789 if ( !$deleteable ) {
1790 // Give up, and hope the result doesn't break
1791 // output behavior.
1792 break;
1793 }
1794 if ( $status['name'] === 'MediaWikiTestCase::wfResetOutputBuffersBarrier' ) {
1795 // Unit testing barrier to prevent this function from breaking PHPUnit.
1796 break;
1797 }
1798 if ( !ob_end_clean() ) {
1799 // Could not remove output buffer handler; abort now
1800 // to avoid getting in some kind of infinite loop.
1801 break;
1802 }
1803 if ( $resetGzipEncoding ) {
1804 if ( $status['name'] == 'ob_gzhandler' ) {
1805 // Reset the 'Content-Encoding' field set by this handler
1806 // so we can start fresh.
1807 header_remove( 'Content-Encoding' );
1808 break;
1809 }
1810 }
1811 }
1812 }
1813
1814 /**
1815 * More legible than passing a 'false' parameter to wfResetOutputBuffers():
1816 *
1817 * Clear away output buffers, but keep the Content-Encoding header
1818 * produced by ob_gzhandler, if any.
1819 *
1820 * This should be used for HTTP 304 responses, where you need to
1821 * preserve the Content-Encoding header of the real result, but
1822 * also need to suppress the output of ob_gzhandler to keep to spec
1823 * and avoid breaking Firefox in rare cases where the headers and
1824 * body are broken over two packets.
1825 */
1826 function wfClearOutputBuffers() {
1827 wfResetOutputBuffers( false );
1828 }
1829
1830 /**
1831 * Converts an Accept-* header into an array mapping string values to quality
1832 * factors
1833 *
1834 * @param string $accept
1835 * @param string $def Default
1836 * @return float[] Associative array of string => float pairs
1837 */
1838 function wfAcceptToPrefs( $accept, $def = '*/*' ) {
1839 # No arg means accept anything (per HTTP spec)
1840 if ( !$accept ) {
1841 return [ $def => 1.0 ];
1842 }
1843
1844 $prefs = [];
1845
1846 $parts = explode( ',', $accept );
1847
1848 foreach ( $parts as $part ) {
1849 # @todo FIXME: Doesn't deal with params like 'text/html; level=1'
1850 $values = explode( ';', trim( $part ) );
1851 $match = [];
1852 if ( count( $values ) == 1 ) {
1853 $prefs[$values[0]] = 1.0;
1854 } elseif ( preg_match( '/q\s*=\s*(\d*\.\d+)/', $values[1], $match ) ) {
1855 $prefs[$values[0]] = floatval( $match[1] );
1856 }
1857 }
1858
1859 return $prefs;
1860 }
1861
1862 /**
1863 * Checks if a given MIME type matches any of the keys in the given
1864 * array. Basic wildcards are accepted in the array keys.
1865 *
1866 * Returns the matching MIME type (or wildcard) if a match, otherwise
1867 * NULL if no match.
1868 *
1869 * @param string $type
1870 * @param array $avail
1871 * @return string
1872 * @private
1873 */
1874 function mimeTypeMatch( $type, $avail ) {
1875 if ( array_key_exists( $type, $avail ) ) {
1876 return $type;
1877 } else {
1878 $mainType = explode( '/', $type )[0];
1879 if ( array_key_exists( "$mainType/*", $avail ) ) {
1880 return "$mainType/*";
1881 } elseif ( array_key_exists( '*/*', $avail ) ) {
1882 return '*/*';
1883 } else {
1884 return null;
1885 }
1886 }
1887 }
1888
1889 /**
1890 * Returns the 'best' match between a client's requested internet media types
1891 * and the server's list of available types. Each list should be an associative
1892 * array of type to preference (preference is a float between 0.0 and 1.0).
1893 * Wildcards in the types are acceptable.
1894 *
1895 * @param array $cprefs Client's acceptable type list
1896 * @param array $sprefs Server's offered types
1897 * @return string
1898 *
1899 * @todo FIXME: Doesn't handle params like 'text/plain; charset=UTF-8'
1900 * XXX: generalize to negotiate other stuff
1901 */
1902 function wfNegotiateType( $cprefs, $sprefs ) {
1903 $combine = [];
1904
1905 foreach ( array_keys( $sprefs ) as $type ) {
1906 $subType = explode( '/', $type )[1];
1907 if ( $subType != '*' ) {
1908 $ckey = mimeTypeMatch( $type, $cprefs );
1909 if ( $ckey ) {
1910 $combine[$type] = $sprefs[$type] * $cprefs[$ckey];
1911 }
1912 }
1913 }
1914
1915 foreach ( array_keys( $cprefs ) as $type ) {
1916 $subType = explode( '/', $type )[1];
1917 if ( $subType != '*' && !array_key_exists( $type, $sprefs ) ) {
1918 $skey = mimeTypeMatch( $type, $sprefs );
1919 if ( $skey ) {
1920 $combine[$type] = $sprefs[$skey] * $cprefs[$type];
1921 }
1922 }
1923 }
1924
1925 $bestq = 0;
1926 $besttype = null;
1927
1928 foreach ( array_keys( $combine ) as $type ) {
1929 if ( $combine[$type] > $bestq ) {
1930 $besttype = $type;
1931 $bestq = $combine[$type];
1932 }
1933 }
1934
1935 return $besttype;
1936 }
1937
1938 /**
1939 * Reference-counted warning suppression
1940 *
1941 * @deprecated since 1.26, use MediaWiki\suppressWarnings() directly
1942 * @param bool $end
1943 */
1944 function wfSuppressWarnings( $end = false ) {
1945 MediaWiki\suppressWarnings( $end );
1946 }
1947
1948 /**
1949 * @deprecated since 1.26, use MediaWiki\restoreWarnings() directly
1950 * Restore error level to previous value
1951 */
1952 function wfRestoreWarnings() {
1953 MediaWiki\suppressWarnings( true );
1954 }
1955
1956 # Autodetect, convert and provide timestamps of various types
1957
1958 require_once __DIR__ . '/libs/time/defines.php';
1959
1960 /**
1961 * Get a timestamp string in one of various formats
1962 *
1963 * @param mixed $outputtype A timestamp in one of the supported formats, the
1964 * function will autodetect which format is supplied and act accordingly.
1965 * @param mixed $ts Optional timestamp to convert, default 0 for the current time
1966 * @return string|bool String / false The same date in the format specified in $outputtype or false
1967 */
1968 function wfTimestamp( $outputtype = TS_UNIX, $ts = 0 ) {
1969 $ret = MWTimestamp::convert( $outputtype, $ts );
1970 if ( $ret === false ) {
1971 wfDebug( "wfTimestamp() fed bogus time value: TYPE=$outputtype; VALUE=$ts\n" );
1972 }
1973 return $ret;
1974 }
1975
1976 /**
1977 * Return a formatted timestamp, or null if input is null.
1978 * For dealing with nullable timestamp columns in the database.
1979 *
1980 * @param int $outputtype
1981 * @param string $ts
1982 * @return string
1983 */
1984 function wfTimestampOrNull( $outputtype = TS_UNIX, $ts = null ) {
1985 if ( is_null( $ts ) ) {
1986 return null;
1987 } else {
1988 return wfTimestamp( $outputtype, $ts );
1989 }
1990 }
1991
1992 /**
1993 * Convenience function; returns MediaWiki timestamp for the present time.
1994 *
1995 * @return string
1996 */
1997 function wfTimestampNow() {
1998 # return NOW
1999 return MWTimestamp::now( TS_MW );
2000 }
2001
2002 /**
2003 * Check if the operating system is Windows
2004 *
2005 * @return bool True if it's Windows, false otherwise.
2006 */
2007 function wfIsWindows() {
2008 static $isWindows = null;
2009 if ( $isWindows === null ) {
2010 $isWindows = strtoupper( substr( PHP_OS, 0, 3 ) ) === 'WIN';
2011 }
2012 return $isWindows;
2013 }
2014
2015 /**
2016 * Check if we are running under HHVM
2017 *
2018 * @return bool
2019 */
2020 function wfIsHHVM() {
2021 return defined( 'HHVM_VERSION' );
2022 }
2023
2024 /**
2025 * Tries to get the system directory for temporary files. First
2026 * $wgTmpDirectory is checked, and then the TMPDIR, TMP, and TEMP
2027 * environment variables are then checked in sequence, then
2028 * sys_get_temp_dir(), then upload_tmp_dir from php.ini.
2029 *
2030 * NOTE: When possible, use instead the tmpfile() function to create
2031 * temporary files to avoid race conditions on file creation, etc.
2032 *
2033 * @return string
2034 */
2035 function wfTempDir() {
2036 global $wgTmpDirectory;
2037
2038 if ( $wgTmpDirectory !== false ) {
2039 return $wgTmpDirectory;
2040 }
2041
2042 return TempFSFile::getUsableTempDirectory();
2043 }
2044
2045 /**
2046 * Make directory, and make all parent directories if they don't exist
2047 *
2048 * @param string $dir Full path to directory to create
2049 * @param int $mode Chmod value to use, default is $wgDirectoryMode
2050 * @param string $caller Optional caller param for debugging.
2051 * @throws MWException
2052 * @return bool
2053 */
2054 function wfMkdirParents( $dir, $mode = null, $caller = null ) {
2055 global $wgDirectoryMode;
2056
2057 if ( FileBackend::isStoragePath( $dir ) ) { // sanity
2058 throw new MWException( __FUNCTION__ . " given storage path '$dir'." );
2059 }
2060
2061 if ( !is_null( $caller ) ) {
2062 wfDebug( "$caller: called wfMkdirParents($dir)\n" );
2063 }
2064
2065 if ( strval( $dir ) === '' || is_dir( $dir ) ) {
2066 return true;
2067 }
2068
2069 $dir = str_replace( [ '\\', '/' ], DIRECTORY_SEPARATOR, $dir );
2070
2071 if ( is_null( $mode ) ) {
2072 $mode = $wgDirectoryMode;
2073 }
2074
2075 // Turn off the normal warning, we're doing our own below
2076 MediaWiki\suppressWarnings();
2077 $ok = mkdir( $dir, $mode, true ); // PHP5 <3
2078 MediaWiki\restoreWarnings();
2079
2080 if ( !$ok ) {
2081 // directory may have been created on another request since we last checked
2082 if ( is_dir( $dir ) ) {
2083 return true;
2084 }
2085
2086 // PHP doesn't report the path in its warning message, so add our own to aid in diagnosis.
2087 wfLogWarning( sprintf( "failed to mkdir \"%s\" mode 0%o", $dir, $mode ) );
2088 }
2089 return $ok;
2090 }
2091
2092 /**
2093 * Remove a directory and all its content.
2094 * Does not hide error.
2095 * @param string $dir
2096 */
2097 function wfRecursiveRemoveDir( $dir ) {
2098 wfDebug( __FUNCTION__ . "( $dir )\n" );
2099 // taken from http://de3.php.net/manual/en/function.rmdir.php#98622
2100 if ( is_dir( $dir ) ) {
2101 $objects = scandir( $dir );
2102 foreach ( $objects as $object ) {
2103 if ( $object != "." && $object != ".." ) {
2104 if ( filetype( $dir . '/' . $object ) == "dir" ) {
2105 wfRecursiveRemoveDir( $dir . '/' . $object );
2106 } else {
2107 unlink( $dir . '/' . $object );
2108 }
2109 }
2110 }
2111 reset( $objects );
2112 rmdir( $dir );
2113 }
2114 }
2115
2116 /**
2117 * @param int $nr The number to format
2118 * @param int $acc The number of digits after the decimal point, default 2
2119 * @param bool $round Whether or not to round the value, default true
2120 * @return string
2121 */
2122 function wfPercent( $nr, $acc = 2, $round = true ) {
2123 $ret = sprintf( "%.${acc}f", $nr );
2124 return $round ? round( $ret, $acc ) . '%' : "$ret%";
2125 }
2126
2127 /**
2128 * Safety wrapper around ini_get() for boolean settings.
2129 * The values returned from ini_get() are pre-normalized for settings
2130 * set via php.ini or php_flag/php_admin_flag... but *not*
2131 * for those set via php_value/php_admin_value.
2132 *
2133 * It's fairly common for people to use php_value instead of php_flag,
2134 * which can leave you with an 'off' setting giving a false positive
2135 * for code that just takes the ini_get() return value as a boolean.
2136 *
2137 * To make things extra interesting, setting via php_value accepts
2138 * "true" and "yes" as true, but php.ini and php_flag consider them false. :)
2139 * Unrecognized values go false... again opposite PHP's own coercion
2140 * from string to bool.
2141 *
2142 * Luckily, 'properly' set settings will always come back as '0' or '1',
2143 * so we only have to worry about them and the 'improper' settings.
2144 *
2145 * I frickin' hate PHP... :P
2146 *
2147 * @param string $setting
2148 * @return bool
2149 */
2150 function wfIniGetBool( $setting ) {
2151 $val = strtolower( ini_get( $setting ) );
2152 // 'on' and 'true' can't have whitespace around them, but '1' can.
2153 return $val == 'on'
2154 || $val == 'true'
2155 || $val == 'yes'
2156 || preg_match( "/^\s*[+-]?0*[1-9]/", $val ); // approx C atoi() function
2157 }
2158
2159 /**
2160 * Windows-compatible version of escapeshellarg()
2161 * Windows doesn't recognise single-quotes in the shell, but the escapeshellarg()
2162 * function puts single quotes in regardless of OS.
2163 *
2164 * Also fixes the locale problems on Linux in PHP 5.2.6+ (bug backported to
2165 * earlier distro releases of PHP)
2166 *
2167 * @param string ... strings to escape and glue together, or a single array of strings parameter
2168 * @return string
2169 */
2170 function wfEscapeShellArg( /*...*/ ) {
2171 wfInitShellLocale();
2172
2173 $args = func_get_args();
2174 if ( count( $args ) === 1 && is_array( reset( $args ) ) ) {
2175 // If only one argument has been passed, and that argument is an array,
2176 // treat it as a list of arguments
2177 $args = reset( $args );
2178 }
2179
2180 $first = true;
2181 $retVal = '';
2182 foreach ( $args as $arg ) {
2183 if ( !$first ) {
2184 $retVal .= ' ';
2185 } else {
2186 $first = false;
2187 }
2188
2189 if ( wfIsWindows() ) {
2190 // Escaping for an MSVC-style command line parser and CMD.EXE
2191 // @codingStandardsIgnoreStart For long URLs
2192 // Refs:
2193 // * http://web.archive.org/web/20020708081031/http://mailman.lyra.org/pipermail/scite-interest/2002-March/000436.html
2194 // * http://technet.microsoft.com/en-us/library/cc723564.aspx
2195 // * T15518
2196 // * CR r63214
2197 // Double the backslashes before any double quotes. Escape the double quotes.
2198 // @codingStandardsIgnoreEnd
2199 $tokens = preg_split( '/(\\\\*")/', $arg, -1, PREG_SPLIT_DELIM_CAPTURE );
2200 $arg = '';
2201 $iteration = 0;
2202 foreach ( $tokens as $token ) {
2203 if ( $iteration % 2 == 1 ) {
2204 // Delimiter, a double quote preceded by zero or more slashes
2205 $arg .= str_replace( '\\', '\\\\', substr( $token, 0, -1 ) ) . '\\"';
2206 } elseif ( $iteration % 4 == 2 ) {
2207 // ^ in $token will be outside quotes, need to be escaped
2208 $arg .= str_replace( '^', '^^', $token );
2209 } else { // $iteration % 4 == 0
2210 // ^ in $token will appear inside double quotes, so leave as is
2211 $arg .= $token;
2212 }
2213 $iteration++;
2214 }
2215 // Double the backslashes before the end of the string, because
2216 // we will soon add a quote
2217 $m = [];
2218 if ( preg_match( '/^(.*?)(\\\\+)$/', $arg, $m ) ) {
2219 $arg = $m[1] . str_replace( '\\', '\\\\', $m[2] );
2220 }
2221
2222 // Add surrounding quotes
2223 $retVal .= '"' . $arg . '"';
2224 } else {
2225 $retVal .= escapeshellarg( $arg );
2226 }
2227 }
2228 return $retVal;
2229 }
2230
2231 /**
2232 * Check if wfShellExec() is effectively disabled via php.ini config
2233 *
2234 * @return bool|string False or 'disabled'
2235 * @since 1.22
2236 */
2237 function wfShellExecDisabled() {
2238 static $disabled = null;
2239 if ( is_null( $disabled ) ) {
2240 if ( !function_exists( 'proc_open' ) ) {
2241 wfDebug( "proc_open() is disabled\n" );
2242 $disabled = 'disabled';
2243 } else {
2244 $disabled = false;
2245 }
2246 }
2247 return $disabled;
2248 }
2249
2250 /**
2251 * Execute a shell command, with time and memory limits mirrored from the PHP
2252 * configuration if supported.
2253 *
2254 * @param string|string[] $cmd If string, a properly shell-escaped command line,
2255 * or an array of unescaped arguments, in which case each value will be escaped
2256 * Example: [ 'convert', '-font', 'font name' ] would produce "'convert' '-font' 'font name'"
2257 * @param null|mixed &$retval Optional, will receive the program's exit code.
2258 * (non-zero is usually failure). If there is an error from
2259 * read, select, or proc_open(), this will be set to -1.
2260 * @param array $environ Optional environment variables which should be
2261 * added to the executed command environment.
2262 * @param array $limits Optional array with limits(filesize, memory, time, walltime)
2263 * this overwrites the global wgMaxShell* limits.
2264 * @param array $options Array of options:
2265 * - duplicateStderr: Set this to true to duplicate stderr to stdout,
2266 * including errors from limit.sh
2267 * - profileMethod: By default this function will profile based on the calling
2268 * method. Set this to a string for an alternative method to profile from
2269 *
2270 * @return string Collected stdout as a string
2271 */
2272 function wfShellExec( $cmd, &$retval = null, $environ = [],
2273 $limits = [], $options = []
2274 ) {
2275 global $IP, $wgMaxShellMemory, $wgMaxShellFileSize, $wgMaxShellTime,
2276 $wgMaxShellWallClockTime, $wgShellCgroup;
2277
2278 $disabled = wfShellExecDisabled();
2279 if ( $disabled ) {
2280 $retval = 1;
2281 return 'Unable to run external programs, proc_open() is disabled.';
2282 }
2283
2284 $includeStderr = isset( $options['duplicateStderr'] ) && $options['duplicateStderr'];
2285 $profileMethod = isset( $options['profileMethod'] ) ? $options['profileMethod'] : wfGetCaller();
2286
2287 wfInitShellLocale();
2288
2289 $envcmd = '';
2290 foreach ( $environ as $k => $v ) {
2291 if ( wfIsWindows() ) {
2292 /* Surrounding a set in quotes (method used by wfEscapeShellArg) makes the quotes themselves
2293 * appear in the environment variable, so we must use carat escaping as documented in
2294 * http://technet.microsoft.com/en-us/library/cc723564.aspx
2295 * Note however that the quote isn't listed there, but is needed, and the parentheses
2296 * are listed there but doesn't appear to need it.
2297 */
2298 $envcmd .= "set $k=" . preg_replace( '/([&|()<>^"])/', '^\\1', $v ) . '&& ';
2299 } else {
2300 /* Assume this is a POSIX shell, thus required to accept variable assignments before the command
2301 * http://www.opengroup.org/onlinepubs/009695399/utilities/xcu_chap02.html#tag_02_09_01
2302 */
2303 $envcmd .= "$k=" . escapeshellarg( $v ) . ' ';
2304 }
2305 }
2306 if ( is_array( $cmd ) ) {
2307 $cmd = wfEscapeShellArg( $cmd );
2308 }
2309
2310 $cmd = $envcmd . $cmd;
2311
2312 $useLogPipe = false;
2313 if ( is_executable( '/bin/bash' ) ) {
2314 $time = intval( isset( $limits['time'] ) ? $limits['time'] : $wgMaxShellTime );
2315 if ( isset( $limits['walltime'] ) ) {
2316 $wallTime = intval( $limits['walltime'] );
2317 } elseif ( isset( $limits['time'] ) ) {
2318 $wallTime = $time;
2319 } else {
2320 $wallTime = intval( $wgMaxShellWallClockTime );
2321 }
2322 $mem = intval( isset( $limits['memory'] ) ? $limits['memory'] : $wgMaxShellMemory );
2323 $filesize = intval( isset( $limits['filesize'] ) ? $limits['filesize'] : $wgMaxShellFileSize );
2324
2325 if ( $time > 0 || $mem > 0 || $filesize > 0 || $wallTime > 0 ) {
2326 $cmd = '/bin/bash ' . escapeshellarg( "$IP/includes/limit.sh" ) . ' ' .
2327 escapeshellarg( $cmd ) . ' ' .
2328 escapeshellarg(
2329 "MW_INCLUDE_STDERR=" . ( $includeStderr ? '1' : '' ) . ';' .
2330 "MW_CPU_LIMIT=$time; " .
2331 'MW_CGROUP=' . escapeshellarg( $wgShellCgroup ) . '; ' .
2332 "MW_MEM_LIMIT=$mem; " .
2333 "MW_FILE_SIZE_LIMIT=$filesize; " .
2334 "MW_WALL_CLOCK_LIMIT=$wallTime; " .
2335 "MW_USE_LOG_PIPE=yes"
2336 );
2337 $useLogPipe = true;
2338 } elseif ( $includeStderr ) {
2339 $cmd .= ' 2>&1';
2340 }
2341 } elseif ( $includeStderr ) {
2342 $cmd .= ' 2>&1';
2343 }
2344 wfDebug( "wfShellExec: $cmd\n" );
2345
2346 // Don't try to execute commands that exceed Linux's MAX_ARG_STRLEN.
2347 // Other platforms may be more accomodating, but we don't want to be
2348 // accomodating, because very long commands probably include user
2349 // input. See T129506.
2350 if ( strlen( $cmd ) > SHELL_MAX_ARG_STRLEN ) {
2351 throw new Exception( __METHOD__ .
2352 '(): total length of $cmd must not exceed SHELL_MAX_ARG_STRLEN' );
2353 }
2354
2355 $desc = [
2356 0 => [ 'file', 'php://stdin', 'r' ],
2357 1 => [ 'pipe', 'w' ],
2358 2 => [ 'file', 'php://stderr', 'w' ] ];
2359 if ( $useLogPipe ) {
2360 $desc[3] = [ 'pipe', 'w' ];
2361 }
2362 $pipes = null;
2363 $scoped = Profiler::instance()->scopedProfileIn( __FUNCTION__ . '-' . $profileMethod );
2364 $proc = proc_open( $cmd, $desc, $pipes );
2365 if ( !$proc ) {
2366 wfDebugLog( 'exec', "proc_open() failed: $cmd" );
2367 $retval = -1;
2368 return '';
2369 }
2370 $outBuffer = $logBuffer = '';
2371 $emptyArray = [];
2372 $status = false;
2373 $logMsg = false;
2374
2375 /* According to the documentation, it is possible for stream_select()
2376 * to fail due to EINTR. I haven't managed to induce this in testing
2377 * despite sending various signals. If it did happen, the error
2378 * message would take the form:
2379 *
2380 * stream_select(): unable to select [4]: Interrupted system call (max_fd=5)
2381 *
2382 * where [4] is the value of the macro EINTR and "Interrupted system
2383 * call" is string which according to the Linux manual is "possibly"
2384 * localised according to LC_MESSAGES.
2385 */
2386 $eintr = defined( 'SOCKET_EINTR' ) ? SOCKET_EINTR : 4;
2387 $eintrMessage = "stream_select(): unable to select [$eintr]";
2388
2389 $running = true;
2390 $timeout = null;
2391 $numReadyPipes = 0;
2392
2393 while ( $running === true || $numReadyPipes !== 0 ) {
2394 if ( $running ) {
2395 $status = proc_get_status( $proc );
2396 // If the process has terminated, switch to nonblocking selects
2397 // for getting any data still waiting to be read.
2398 if ( !$status['running'] ) {
2399 $running = false;
2400 $timeout = 0;
2401 }
2402 }
2403
2404 $readyPipes = $pipes;
2405
2406 // Clear last error
2407 // @codingStandardsIgnoreStart Generic.PHP.NoSilencedErrors.Discouraged
2408 @trigger_error( '' );
2409 $numReadyPipes = @stream_select( $readyPipes, $emptyArray, $emptyArray, $timeout );
2410 if ( $numReadyPipes === false ) {
2411 // @codingStandardsIgnoreEnd
2412 $error = error_get_last();
2413 if ( strncmp( $error['message'], $eintrMessage, strlen( $eintrMessage ) ) == 0 ) {
2414 continue;
2415 } else {
2416 trigger_error( $error['message'], E_USER_WARNING );
2417 $logMsg = $error['message'];
2418 break;
2419 }
2420 }
2421 foreach ( $readyPipes as $fd => $pipe ) {
2422 $block = fread( $pipe, 65536 );
2423 if ( $block === '' ) {
2424 // End of file
2425 fclose( $pipes[$fd] );
2426 unset( $pipes[$fd] );
2427 if ( !$pipes ) {
2428 break 2;
2429 }
2430 } elseif ( $block === false ) {
2431 // Read error
2432 $logMsg = "Error reading from pipe";
2433 break 2;
2434 } elseif ( $fd == 1 ) {
2435 // From stdout
2436 $outBuffer .= $block;
2437 } elseif ( $fd == 3 ) {
2438 // From log FD
2439 $logBuffer .= $block;
2440 if ( strpos( $block, "\n" ) !== false ) {
2441 $lines = explode( "\n", $logBuffer );
2442 $logBuffer = array_pop( $lines );
2443 foreach ( $lines as $line ) {
2444 wfDebugLog( 'exec', $line );
2445 }
2446 }
2447 }
2448 }
2449 }
2450
2451 foreach ( $pipes as $pipe ) {
2452 fclose( $pipe );
2453 }
2454
2455 // Use the status previously collected if possible, since proc_get_status()
2456 // just calls waitpid() which will not return anything useful the second time.
2457 if ( $running ) {
2458 $status = proc_get_status( $proc );
2459 }
2460
2461 if ( $logMsg !== false ) {
2462 // Read/select error
2463 $retval = -1;
2464 proc_close( $proc );
2465 } elseif ( $status['signaled'] ) {
2466 $logMsg = "Exited with signal {$status['termsig']}";
2467 $retval = 128 + $status['termsig'];
2468 proc_close( $proc );
2469 } else {
2470 if ( $status['running'] ) {
2471 $retval = proc_close( $proc );
2472 } else {
2473 $retval = $status['exitcode'];
2474 proc_close( $proc );
2475 }
2476 if ( $retval == 127 ) {
2477 $logMsg = "Possibly missing executable file";
2478 } elseif ( $retval >= 129 && $retval <= 192 ) {
2479 $logMsg = "Probably exited with signal " . ( $retval - 128 );
2480 }
2481 }
2482
2483 if ( $logMsg !== false ) {
2484 wfDebugLog( 'exec', "$logMsg: $cmd" );
2485 }
2486
2487 return $outBuffer;
2488 }
2489
2490 /**
2491 * Execute a shell command, returning both stdout and stderr. Convenience
2492 * function, as all the arguments to wfShellExec can become unwieldy.
2493 *
2494 * @note This also includes errors from limit.sh, e.g. if $wgMaxShellFileSize is exceeded.
2495 * @param string|string[] $cmd If string, a properly shell-escaped command line,
2496 * or an array of unescaped arguments, in which case each value will be escaped
2497 * Example: [ 'convert', '-font', 'font name' ] would produce "'convert' '-font' 'font name'"
2498 * @param null|mixed &$retval Optional, will receive the program's exit code.
2499 * (non-zero is usually failure)
2500 * @param array $environ Optional environment variables which should be
2501 * added to the executed command environment.
2502 * @param array $limits Optional array with limits(filesize, memory, time, walltime)
2503 * this overwrites the global wgMaxShell* limits.
2504 * @return string Collected stdout and stderr as a string
2505 */
2506 function wfShellExecWithStderr( $cmd, &$retval = null, $environ = [], $limits = [] ) {
2507 return wfShellExec( $cmd, $retval, $environ, $limits,
2508 [ 'duplicateStderr' => true, 'profileMethod' => wfGetCaller() ] );
2509 }
2510
2511 /**
2512 * Workaround for http://bugs.php.net/bug.php?id=45132
2513 * escapeshellarg() destroys non-ASCII characters if LANG is not a UTF-8 locale
2514 */
2515 function wfInitShellLocale() {
2516 static $done = false;
2517 if ( $done ) {
2518 return;
2519 }
2520 $done = true;
2521 global $wgShellLocale;
2522 putenv( "LC_CTYPE=$wgShellLocale" );
2523 setlocale( LC_CTYPE, $wgShellLocale );
2524 }
2525
2526 /**
2527 * Generate a shell-escaped command line string to run a MediaWiki cli script.
2528 * Note that $parameters should be a flat array and an option with an argument
2529 * should consist of two consecutive items in the array (do not use "--option value").
2530 *
2531 * @param string $script MediaWiki cli script path
2532 * @param array $parameters Arguments and options to the script
2533 * @param array $options Associative array of options:
2534 * 'php': The path to the php executable
2535 * 'wrapper': Path to a PHP wrapper to handle the maintenance script
2536 * @return string
2537 */
2538 function wfShellWikiCmd( $script, array $parameters = [], array $options = [] ) {
2539 global $wgPhpCli;
2540 // Give site config file a chance to run the script in a wrapper.
2541 // The caller may likely want to call wfBasename() on $script.
2542 Hooks::run( 'wfShellWikiCmd', [ &$script, &$parameters, &$options ] );
2543 $cmd = isset( $options['php'] ) ? [ $options['php'] ] : [ $wgPhpCli ];
2544 if ( isset( $options['wrapper'] ) ) {
2545 $cmd[] = $options['wrapper'];
2546 }
2547 $cmd[] = $script;
2548 // Escape each parameter for shell
2549 return wfEscapeShellArg( array_merge( $cmd, $parameters ) );
2550 }
2551
2552 /**
2553 * wfMerge attempts to merge differences between three texts.
2554 * Returns true for a clean merge and false for failure or a conflict.
2555 *
2556 * @param string $old
2557 * @param string $mine
2558 * @param string $yours
2559 * @param string $result
2560 * @return bool
2561 */
2562 function wfMerge( $old, $mine, $yours, &$result ) {
2563 global $wgDiff3;
2564
2565 # This check may also protect against code injection in
2566 # case of broken installations.
2567 MediaWiki\suppressWarnings();
2568 $haveDiff3 = $wgDiff3 && file_exists( $wgDiff3 );
2569 MediaWiki\restoreWarnings();
2570
2571 if ( !$haveDiff3 ) {
2572 wfDebug( "diff3 not found\n" );
2573 return false;
2574 }
2575
2576 # Make temporary files
2577 $td = wfTempDir();
2578 $oldtextFile = fopen( $oldtextName = tempnam( $td, 'merge-old-' ), 'w' );
2579 $mytextFile = fopen( $mytextName = tempnam( $td, 'merge-mine-' ), 'w' );
2580 $yourtextFile = fopen( $yourtextName = tempnam( $td, 'merge-your-' ), 'w' );
2581
2582 # NOTE: diff3 issues a warning to stderr if any of the files does not end with
2583 # a newline character. To avoid this, we normalize the trailing whitespace before
2584 # creating the diff.
2585
2586 fwrite( $oldtextFile, rtrim( $old ) . "\n" );
2587 fclose( $oldtextFile );
2588 fwrite( $mytextFile, rtrim( $mine ) . "\n" );
2589 fclose( $mytextFile );
2590 fwrite( $yourtextFile, rtrim( $yours ) . "\n" );
2591 fclose( $yourtextFile );
2592
2593 # Check for a conflict
2594 $cmd = wfEscapeShellArg( $wgDiff3, '-a', '--overlap-only', $mytextName,
2595 $oldtextName, $yourtextName );
2596 $handle = popen( $cmd, 'r' );
2597
2598 if ( fgets( $handle, 1024 ) ) {
2599 $conflict = true;
2600 } else {
2601 $conflict = false;
2602 }
2603 pclose( $handle );
2604
2605 # Merge differences
2606 $cmd = wfEscapeShellArg( $wgDiff3, '-a', '-e', '--merge', $mytextName,
2607 $oldtextName, $yourtextName );
2608 $handle = popen( $cmd, 'r' );
2609 $result = '';
2610 do {
2611 $data = fread( $handle, 8192 );
2612 if ( strlen( $data ) == 0 ) {
2613 break;
2614 }
2615 $result .= $data;
2616 } while ( true );
2617 pclose( $handle );
2618 unlink( $mytextName );
2619 unlink( $oldtextName );
2620 unlink( $yourtextName );
2621
2622 if ( $result === '' && $old !== '' && !$conflict ) {
2623 wfDebug( "Unexpected null result from diff3. Command: $cmd\n" );
2624 $conflict = true;
2625 }
2626 return !$conflict;
2627 }
2628
2629 /**
2630 * Returns unified plain-text diff of two texts.
2631 * "Useful" for machine processing of diffs.
2632 *
2633 * @deprecated since 1.25, use DiffEngine/UnifiedDiffFormatter directly
2634 *
2635 * @param string $before The text before the changes.
2636 * @param string $after The text after the changes.
2637 * @param string $params Command-line options for the diff command.
2638 * @return string Unified diff of $before and $after
2639 */
2640 function wfDiff( $before, $after, $params = '-u' ) {
2641 if ( $before == $after ) {
2642 return '';
2643 }
2644
2645 global $wgDiff;
2646 MediaWiki\suppressWarnings();
2647 $haveDiff = $wgDiff && file_exists( $wgDiff );
2648 MediaWiki\restoreWarnings();
2649
2650 # This check may also protect against code injection in
2651 # case of broken installations.
2652 if ( !$haveDiff ) {
2653 wfDebug( "diff executable not found\n" );
2654 $diffs = new Diff( explode( "\n", $before ), explode( "\n", $after ) );
2655 $format = new UnifiedDiffFormatter();
2656 return $format->format( $diffs );
2657 }
2658
2659 # Make temporary files
2660 $td = wfTempDir();
2661 $oldtextFile = fopen( $oldtextName = tempnam( $td, 'merge-old-' ), 'w' );
2662 $newtextFile = fopen( $newtextName = tempnam( $td, 'merge-your-' ), 'w' );
2663
2664 fwrite( $oldtextFile, $before );
2665 fclose( $oldtextFile );
2666 fwrite( $newtextFile, $after );
2667 fclose( $newtextFile );
2668
2669 // Get the diff of the two files
2670 $cmd = "$wgDiff " . $params . ' ' . wfEscapeShellArg( $oldtextName, $newtextName );
2671
2672 $h = popen( $cmd, 'r' );
2673 if ( !$h ) {
2674 unlink( $oldtextName );
2675 unlink( $newtextName );
2676 throw new Exception( __METHOD__ . '(): popen() failed' );
2677 }
2678
2679 $diff = '';
2680
2681 do {
2682 $data = fread( $h, 8192 );
2683 if ( strlen( $data ) == 0 ) {
2684 break;
2685 }
2686 $diff .= $data;
2687 } while ( true );
2688
2689 // Clean up
2690 pclose( $h );
2691 unlink( $oldtextName );
2692 unlink( $newtextName );
2693
2694 // Kill the --- and +++ lines. They're not useful.
2695 $diff_lines = explode( "\n", $diff );
2696 if ( isset( $diff_lines[0] ) && strpos( $diff_lines[0], '---' ) === 0 ) {
2697 unset( $diff_lines[0] );
2698 }
2699 if ( isset( $diff_lines[1] ) && strpos( $diff_lines[1], '+++' ) === 0 ) {
2700 unset( $diff_lines[1] );
2701 }
2702
2703 $diff = implode( "\n", $diff_lines );
2704
2705 return $diff;
2706 }
2707
2708 /**
2709 * This function works like "use VERSION" in Perl, the program will die with a
2710 * backtrace if the current version of PHP is less than the version provided
2711 *
2712 * This is useful for extensions which due to their nature are not kept in sync
2713 * with releases, and might depend on other versions of PHP than the main code
2714 *
2715 * Note: PHP might die due to parsing errors in some cases before it ever
2716 * manages to call this function, such is life
2717 *
2718 * @see perldoc -f use
2719 *
2720 * @param string|int|float $req_ver The version to check, can be a string, an integer, or a float
2721 * @throws MWException
2722 */
2723 function wfUsePHP( $req_ver ) {
2724 $php_ver = PHP_VERSION;
2725
2726 if ( version_compare( $php_ver, (string)$req_ver, '<' ) ) {
2727 throw new MWException( "PHP $req_ver required--this is only $php_ver" );
2728 }
2729 }
2730
2731 /**
2732 * This function works like "use VERSION" in Perl except it checks the version
2733 * of MediaWiki, the program will die with a backtrace if the current version
2734 * of MediaWiki is less than the version provided.
2735 *
2736 * This is useful for extensions which due to their nature are not kept in sync
2737 * with releases
2738 *
2739 * Note: Due to the behavior of PHP's version_compare() which is used in this
2740 * function, if you want to allow the 'wmf' development versions add a 'c' (or
2741 * any single letter other than 'a', 'b' or 'p') as a post-fix to your
2742 * targeted version number. For example if you wanted to allow any variation
2743 * of 1.22 use `wfUseMW( '1.22c' )`. Using an 'a' or 'b' instead of 'c' will
2744 * not result in the same comparison due to the internal logic of
2745 * version_compare().
2746 *
2747 * @see perldoc -f use
2748 *
2749 * @deprecated since 1.26, use the "requires' property of extension.json
2750 * @param string|int|float $req_ver The version to check, can be a string, an integer, or a float
2751 * @throws MWException
2752 */
2753 function wfUseMW( $req_ver ) {
2754 global $wgVersion;
2755
2756 if ( version_compare( $wgVersion, (string)$req_ver, '<' ) ) {
2757 throw new MWException( "MediaWiki $req_ver required--this is only $wgVersion" );
2758 }
2759 }
2760
2761 /**
2762 * Return the final portion of a pathname.
2763 * Reimplemented because PHP5's "basename()" is buggy with multibyte text.
2764 * http://bugs.php.net/bug.php?id=33898
2765 *
2766 * PHP's basename() only considers '\' a pathchar on Windows and Netware.
2767 * We'll consider it so always, as we don't want '\s' in our Unix paths either.
2768 *
2769 * @param string $path
2770 * @param string $suffix String to remove if present
2771 * @return string
2772 */
2773 function wfBaseName( $path, $suffix = '' ) {
2774 if ( $suffix == '' ) {
2775 $encSuffix = '';
2776 } else {
2777 $encSuffix = '(?:' . preg_quote( $suffix, '#' ) . ')?';
2778 }
2779
2780 $matches = [];
2781 if ( preg_match( "#([^/\\\\]*?){$encSuffix}[/\\\\]*$#", $path, $matches ) ) {
2782 return $matches[1];
2783 } else {
2784 return '';
2785 }
2786 }
2787
2788 /**
2789 * Generate a relative path name to the given file.
2790 * May explode on non-matching case-insensitive paths,
2791 * funky symlinks, etc.
2792 *
2793 * @param string $path Absolute destination path including target filename
2794 * @param string $from Absolute source path, directory only
2795 * @return string
2796 */
2797 function wfRelativePath( $path, $from ) {
2798 // Normalize mixed input on Windows...
2799 $path = str_replace( '/', DIRECTORY_SEPARATOR, $path );
2800 $from = str_replace( '/', DIRECTORY_SEPARATOR, $from );
2801
2802 // Trim trailing slashes -- fix for drive root
2803 $path = rtrim( $path, DIRECTORY_SEPARATOR );
2804 $from = rtrim( $from, DIRECTORY_SEPARATOR );
2805
2806 $pieces = explode( DIRECTORY_SEPARATOR, dirname( $path ) );
2807 $against = explode( DIRECTORY_SEPARATOR, $from );
2808
2809 if ( $pieces[0] !== $against[0] ) {
2810 // Non-matching Windows drive letters?
2811 // Return a full path.
2812 return $path;
2813 }
2814
2815 // Trim off common prefix
2816 while ( count( $pieces ) && count( $against )
2817 && $pieces[0] == $against[0] ) {
2818 array_shift( $pieces );
2819 array_shift( $against );
2820 }
2821
2822 // relative dots to bump us to the parent
2823 while ( count( $against ) ) {
2824 array_unshift( $pieces, '..' );
2825 array_shift( $against );
2826 }
2827
2828 array_push( $pieces, wfBaseName( $path ) );
2829
2830 return implode( DIRECTORY_SEPARATOR, $pieces );
2831 }
2832
2833 /**
2834 * Convert an arbitrarily-long digit string from one numeric base
2835 * to another, optionally zero-padding to a minimum column width.
2836 *
2837 * Supports base 2 through 36; digit values 10-36 are represented
2838 * as lowercase letters a-z. Input is case-insensitive.
2839 *
2840 * @deprecated since 1.27 Use Wikimedia\base_convert() directly
2841 *
2842 * @param string $input Input number
2843 * @param int $sourceBase Base of the input number
2844 * @param int $destBase Desired base of the output
2845 * @param int $pad Minimum number of digits in the output (pad with zeroes)
2846 * @param bool $lowercase Whether to output in lowercase or uppercase
2847 * @param string $engine Either "gmp", "bcmath", or "php"
2848 * @return string|bool The output number as a string, or false on error
2849 */
2850 function wfBaseConvert( $input, $sourceBase, $destBase, $pad = 1,
2851 $lowercase = true, $engine = 'auto'
2852 ) {
2853 return Wikimedia\base_convert( $input, $sourceBase, $destBase, $pad, $lowercase, $engine );
2854 }
2855
2856 /**
2857 * @deprecated since 1.27, PHP's session generation isn't used with
2858 * MediaWiki\Session\SessionManager
2859 */
2860 function wfFixSessionID() {
2861 wfDeprecated( __FUNCTION__, '1.27' );
2862 }
2863
2864 /**
2865 * Reset the session id
2866 *
2867 * @deprecated since 1.27, use MediaWiki\Session\SessionManager instead
2868 * @since 1.22
2869 */
2870 function wfResetSessionID() {
2871 wfDeprecated( __FUNCTION__, '1.27' );
2872 $session = SessionManager::getGlobalSession();
2873 $delay = $session->delaySave();
2874
2875 $session->resetId();
2876
2877 // Make sure a session is started, since that's what the old
2878 // wfResetSessionID() did.
2879 if ( session_id() !== $session->getId() ) {
2880 wfSetupSession( $session->getId() );
2881 }
2882
2883 ScopedCallback::consume( $delay );
2884 }
2885
2886 /**
2887 * Initialise php session
2888 *
2889 * @deprecated since 1.27, use MediaWiki\Session\SessionManager instead.
2890 * Generally, "using" SessionManager will be calling ->getSessionById() or
2891 * ::getGlobalSession() (depending on whether you were passing $sessionId
2892 * here), then calling $session->persist().
2893 * @param bool|string $sessionId
2894 */
2895 function wfSetupSession( $sessionId = false ) {
2896 wfDeprecated( __FUNCTION__, '1.27' );
2897
2898 if ( $sessionId ) {
2899 session_id( $sessionId );
2900 }
2901
2902 $session = SessionManager::getGlobalSession();
2903 $session->persist();
2904
2905 if ( session_id() !== $session->getId() ) {
2906 session_id( $session->getId() );
2907 }
2908 MediaWiki\quietCall( 'session_start' );
2909 }
2910
2911 /**
2912 * Get an object from the precompiled serialized directory
2913 *
2914 * @param string $name
2915 * @return mixed The variable on success, false on failure
2916 */
2917 function wfGetPrecompiledData( $name ) {
2918 global $IP;
2919
2920 $file = "$IP/serialized/$name";
2921 if ( file_exists( $file ) ) {
2922 $blob = file_get_contents( $file );
2923 if ( $blob ) {
2924 return unserialize( $blob );
2925 }
2926 }
2927 return false;
2928 }
2929
2930 /**
2931 * Make a cache key for the local wiki.
2932 *
2933 * @param string $args,...
2934 * @return string
2935 */
2936 function wfMemcKey( /*...*/ ) {
2937 return call_user_func_array(
2938 [ ObjectCache::getLocalClusterInstance(), 'makeKey' ],
2939 func_get_args()
2940 );
2941 }
2942
2943 /**
2944 * Make a cache key for a foreign DB.
2945 *
2946 * Must match what wfMemcKey() would produce in context of the foreign wiki.
2947 *
2948 * @param string $db
2949 * @param string $prefix
2950 * @param string $args,...
2951 * @return string
2952 */
2953 function wfForeignMemcKey( $db, $prefix /*...*/ ) {
2954 $args = array_slice( func_get_args(), 2 );
2955 $keyspace = $prefix ? "$db-$prefix" : $db;
2956 return call_user_func_array(
2957 [ ObjectCache::getLocalClusterInstance(), 'makeKeyInternal' ],
2958 [ $keyspace, $args ]
2959 );
2960 }
2961
2962 /**
2963 * Make a cache key with database-agnostic prefix.
2964 *
2965 * Doesn't have a wiki-specific namespace. Uses a generic 'global' prefix
2966 * instead. Must have a prefix as otherwise keys that use a database name
2967 * in the first segment will clash with wfMemcKey/wfForeignMemcKey.
2968 *
2969 * @since 1.26
2970 * @param string $args,...
2971 * @return string
2972 */
2973 function wfGlobalCacheKey( /*...*/ ) {
2974 return call_user_func_array(
2975 [ ObjectCache::getLocalClusterInstance(), 'makeGlobalKey' ],
2976 func_get_args()
2977 );
2978 }
2979
2980 /**
2981 * Get an ASCII string identifying this wiki
2982 * This is used as a prefix in memcached keys
2983 *
2984 * @return string
2985 */
2986 function wfWikiID() {
2987 global $wgDBprefix, $wgDBname;
2988 if ( $wgDBprefix ) {
2989 return "$wgDBname-$wgDBprefix";
2990 } else {
2991 return $wgDBname;
2992 }
2993 }
2994
2995 /**
2996 * Split a wiki ID into DB name and table prefix
2997 *
2998 * @param string $wiki
2999 *
3000 * @return array
3001 */
3002 function wfSplitWikiID( $wiki ) {
3003 $bits = explode( '-', $wiki, 2 );
3004 if ( count( $bits ) < 2 ) {
3005 $bits[] = '';
3006 }
3007 return $bits;
3008 }
3009
3010 /**
3011 * Get a Database object.
3012 *
3013 * @param int $db Index of the connection to get. May be DB_MASTER for the
3014 * master (for write queries), DB_REPLICA for potentially lagged read
3015 * queries, or an integer >= 0 for a particular server.
3016 *
3017 * @param string|string[] $groups Query groups. An array of group names that this query
3018 * belongs to. May contain a single string if the query is only
3019 * in one group.
3020 *
3021 * @param string|bool $wiki The wiki ID, or false for the current wiki
3022 *
3023 * Note: multiple calls to wfGetDB(DB_REPLICA) during the course of one request
3024 * will always return the same object, unless the underlying connection or load
3025 * balancer is manually destroyed.
3026 *
3027 * Note 2: use $this->getDB() in maintenance scripts that may be invoked by
3028 * updater to ensure that a proper database is being updated.
3029 *
3030 * @todo Replace calls to wfGetDB with calls to LoadBalancer::getConnection()
3031 * on an injected instance of LoadBalancer.
3032 *
3033 * @return DatabaseBase
3034 */
3035 function wfGetDB( $db, $groups = [], $wiki = false ) {
3036 return wfGetLB( $wiki )->getConnection( $db, $groups, $wiki );
3037 }
3038
3039 /**
3040 * Get a load balancer object.
3041 *
3042 * @deprecated since 1.27, use MediaWikiServices::getDBLoadBalancer()
3043 * or MediaWikiServices::getDBLoadBalancerFactory() instead.
3044 *
3045 * @param string|bool $wiki Wiki ID, or false for the current wiki
3046 * @return LoadBalancer
3047 */
3048 function wfGetLB( $wiki = false ) {
3049 if ( $wiki === false ) {
3050 return \MediaWiki\MediaWikiServices::getInstance()->getDBLoadBalancer();
3051 } else {
3052 $factory = \MediaWiki\MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
3053 return $factory->getMainLB( $wiki );
3054 }
3055 }
3056
3057 /**
3058 * Get the load balancer factory object
3059 *
3060 * @deprecated since 1.27, use MediaWikiServices::getDBLoadBalancerFactory() instead.
3061 *
3062 * @return LBFactory
3063 */
3064 function wfGetLBFactory() {
3065 return \MediaWiki\MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
3066 }
3067
3068 /**
3069 * Find a file.
3070 * Shortcut for RepoGroup::singleton()->findFile()
3071 *
3072 * @param string $title String or Title object
3073 * @param array $options Associative array of options (see RepoGroup::findFile)
3074 * @return File|bool File, or false if the file does not exist
3075 */
3076 function wfFindFile( $title, $options = [] ) {
3077 return RepoGroup::singleton()->findFile( $title, $options );
3078 }
3079
3080 /**
3081 * Get an object referring to a locally registered file.
3082 * Returns a valid placeholder object if the file does not exist.
3083 *
3084 * @param Title|string $title
3085 * @return LocalFile|null A File, or null if passed an invalid Title
3086 */
3087 function wfLocalFile( $title ) {
3088 return RepoGroup::singleton()->getLocalRepo()->newFile( $title );
3089 }
3090
3091 /**
3092 * Should low-performance queries be disabled?
3093 *
3094 * @return bool
3095 * @codeCoverageIgnore
3096 */
3097 function wfQueriesMustScale() {
3098 global $wgMiserMode;
3099 return $wgMiserMode
3100 || ( SiteStats::pages() > 100000
3101 && SiteStats::edits() > 1000000
3102 && SiteStats::users() > 10000 );
3103 }
3104
3105 /**
3106 * Get the path to a specified script file, respecting file
3107 * extensions; this is a wrapper around $wgScriptPath etc.
3108 * except for 'index' and 'load' which use $wgScript/$wgLoadScript
3109 *
3110 * @param string $script Script filename, sans extension
3111 * @return string
3112 */
3113 function wfScript( $script = 'index' ) {
3114 global $wgScriptPath, $wgScript, $wgLoadScript;
3115 if ( $script === 'index' ) {
3116 return $wgScript;
3117 } elseif ( $script === 'load' ) {
3118 return $wgLoadScript;
3119 } else {
3120 return "{$wgScriptPath}/{$script}.php";
3121 }
3122 }
3123
3124 /**
3125 * Get the script URL.
3126 *
3127 * @return string Script URL
3128 */
3129 function wfGetScriptUrl() {
3130 if ( isset( $_SERVER['SCRIPT_NAME'] ) ) {
3131 /* as it was called, minus the query string.
3132 *
3133 * Some sites use Apache rewrite rules to handle subdomains,
3134 * and have PHP set up in a weird way that causes PHP_SELF
3135 * to contain the rewritten URL instead of the one that the
3136 * outside world sees.
3137 *
3138 * If in this mode, use SCRIPT_URL instead, which mod_rewrite
3139 * provides containing the "before" URL.
3140 */
3141 return $_SERVER['SCRIPT_NAME'];
3142 } else {
3143 return $_SERVER['URL'];
3144 }
3145 }
3146
3147 /**
3148 * Convenience function converts boolean values into "true"
3149 * or "false" (string) values
3150 *
3151 * @param bool $value
3152 * @return string
3153 */
3154 function wfBoolToStr( $value ) {
3155 return $value ? 'true' : 'false';
3156 }
3157
3158 /**
3159 * Get a platform-independent path to the null file, e.g. /dev/null
3160 *
3161 * @return string
3162 */
3163 function wfGetNull() {
3164 return wfIsWindows() ? 'NUL' : '/dev/null';
3165 }
3166
3167 /**
3168 * Waits for the replica DBs to catch up to the master position
3169 *
3170 * Use this when updating very large numbers of rows, as in maintenance scripts,
3171 * to avoid causing too much lag. Of course, this is a no-op if there are no replica DBs.
3172 *
3173 * By default this waits on the main DB cluster of the current wiki.
3174 * If $cluster is set to "*" it will wait on all DB clusters, including
3175 * external ones. If the lag being waiting on is caused by the code that
3176 * does this check, it makes since to use $ifWritesSince, particularly if
3177 * cluster is "*", to avoid excess overhead.
3178 *
3179 * Never call this function after a big DB write that is still in a transaction.
3180 * This only makes sense after the possible lag inducing changes were committed.
3181 *
3182 * @param float|null $ifWritesSince Only wait if writes were done since this UNIX timestamp
3183 * @param string|bool $wiki Wiki identifier accepted by wfGetLB
3184 * @param string|bool $cluster Cluster name accepted by LBFactory. Default: false.
3185 * @param int|null $timeout Max wait time. Default: 1 day (cli), ~10 seconds (web)
3186 * @return bool Success (able to connect and no timeouts reached)
3187 * @deprecated since 1.27 Use LBFactory::waitForReplication
3188 */
3189 function wfWaitForSlaves(
3190 $ifWritesSince = null, $wiki = false, $cluster = false, $timeout = null
3191 ) {
3192 if ( $timeout === null ) {
3193 $timeout = ( PHP_SAPI === 'cli' ) ? 86400 : 10;
3194 }
3195
3196 if ( $cluster === '*' ) {
3197 $cluster = false;
3198 $wiki = false;
3199 } elseif ( $wiki === false ) {
3200 $wiki = wfWikiID();
3201 }
3202
3203 try {
3204 wfGetLBFactory()->waitForReplication( [
3205 'wiki' => $wiki,
3206 'cluster' => $cluster,
3207 'timeout' => $timeout,
3208 // B/C: first argument used to be "max seconds of lag"; ignore such values
3209 'ifWritesSince' => ( $ifWritesSince > 1e9 ) ? $ifWritesSince : null
3210 ] );
3211 } catch ( DBReplicationWaitError $e ) {
3212 return false;
3213 }
3214
3215 return true;
3216 }
3217
3218 /**
3219 * Count down from $seconds to zero on the terminal, with a one-second pause
3220 * between showing each number. For use in command-line scripts.
3221 *
3222 * @codeCoverageIgnore
3223 * @param int $seconds
3224 */
3225 function wfCountDown( $seconds ) {
3226 for ( $i = $seconds; $i >= 0; $i-- ) {
3227 if ( $i != $seconds ) {
3228 echo str_repeat( "\x08", strlen( $i + 1 ) );
3229 }
3230 echo $i;
3231 flush();
3232 if ( $i ) {
3233 sleep( 1 );
3234 }
3235 }
3236 echo "\n";
3237 }
3238
3239 /**
3240 * Replace all invalid characters with '-'.
3241 * Additional characters can be defined in $wgIllegalFileChars (see T22489).
3242 * By default, $wgIllegalFileChars includes ':', '/', '\'.
3243 *
3244 * @param string $name Filename to process
3245 * @return string
3246 */
3247 function wfStripIllegalFilenameChars( $name ) {
3248 global $wgIllegalFileChars;
3249 $illegalFileChars = $wgIllegalFileChars ? "|[" . $wgIllegalFileChars . "]" : '';
3250 $name = preg_replace(
3251 "/[^" . Title::legalChars() . "]" . $illegalFileChars . "/",
3252 '-',
3253 $name
3254 );
3255 // $wgIllegalFileChars may not include '/' and '\', so we still need to do this
3256 $name = wfBaseName( $name );
3257 return $name;
3258 }
3259
3260 /**
3261 * Set PHP's memory limit to the larger of php.ini or $wgMemoryLimit
3262 *
3263 * @return int Resulting value of the memory limit.
3264 */
3265 function wfMemoryLimit() {
3266 global $wgMemoryLimit;
3267 $memlimit = wfShorthandToInteger( ini_get( 'memory_limit' ) );
3268 if ( $memlimit != -1 ) {
3269 $conflimit = wfShorthandToInteger( $wgMemoryLimit );
3270 if ( $conflimit == -1 ) {
3271 wfDebug( "Removing PHP's memory limit\n" );
3272 MediaWiki\suppressWarnings();
3273 ini_set( 'memory_limit', $conflimit );
3274 MediaWiki\restoreWarnings();
3275 return $conflimit;
3276 } elseif ( $conflimit > $memlimit ) {
3277 wfDebug( "Raising PHP's memory limit to $conflimit bytes\n" );
3278 MediaWiki\suppressWarnings();
3279 ini_set( 'memory_limit', $conflimit );
3280 MediaWiki\restoreWarnings();
3281 return $conflimit;
3282 }
3283 }
3284 return $memlimit;
3285 }
3286
3287 /**
3288 * Set PHP's time limit to the larger of php.ini or $wgTransactionalTimeLimit
3289 *
3290 * @return int Prior time limit
3291 * @since 1.26
3292 */
3293 function wfTransactionalTimeLimit() {
3294 global $wgTransactionalTimeLimit;
3295
3296 $timeLimit = ini_get( 'max_execution_time' );
3297 // Note that CLI scripts use 0
3298 if ( $timeLimit > 0 && $wgTransactionalTimeLimit > $timeLimit ) {
3299 set_time_limit( $wgTransactionalTimeLimit );
3300 }
3301
3302 ignore_user_abort( true ); // ignore client disconnects
3303
3304 return $timeLimit;
3305 }
3306
3307 /**
3308 * Converts shorthand byte notation to integer form
3309 *
3310 * @param string $string
3311 * @param int $default Returned if $string is empty
3312 * @return int
3313 */
3314 function wfShorthandToInteger( $string = '', $default = -1 ) {
3315 $string = trim( $string );
3316 if ( $string === '' ) {
3317 return $default;
3318 }
3319 $last = $string[strlen( $string ) - 1];
3320 $val = intval( $string );
3321 switch ( $last ) {
3322 case 'g':
3323 case 'G':
3324 $val *= 1024;
3325 // break intentionally missing
3326 case 'm':
3327 case 'M':
3328 $val *= 1024;
3329 // break intentionally missing
3330 case 'k':
3331 case 'K':
3332 $val *= 1024;
3333 }
3334
3335 return $val;
3336 }
3337
3338 /**
3339 * Get the normalised IETF language tag
3340 * See unit test for examples.
3341 *
3342 * @param string $code The language code.
3343 * @return string The language code which complying with BCP 47 standards.
3344 */
3345 function wfBCP47( $code ) {
3346 $codeSegment = explode( '-', $code );
3347 $codeBCP = [];
3348 foreach ( $codeSegment as $segNo => $seg ) {
3349 // when previous segment is x, it is a private segment and should be lc
3350 if ( $segNo > 0 && strtolower( $codeSegment[( $segNo - 1 )] ) == 'x' ) {
3351 $codeBCP[$segNo] = strtolower( $seg );
3352 // ISO 3166 country code
3353 } elseif ( ( strlen( $seg ) == 2 ) && ( $segNo > 0 ) ) {
3354 $codeBCP[$segNo] = strtoupper( $seg );
3355 // ISO 15924 script code
3356 } elseif ( ( strlen( $seg ) == 4 ) && ( $segNo > 0 ) ) {
3357 $codeBCP[$segNo] = ucfirst( strtolower( $seg ) );
3358 // Use lowercase for other cases
3359 } else {
3360 $codeBCP[$segNo] = strtolower( $seg );
3361 }
3362 }
3363 $langCode = implode( '-', $codeBCP );
3364 return $langCode;
3365 }
3366
3367 /**
3368 * Get a specific cache object.
3369 *
3370 * @param int|string $cacheType A CACHE_* constants, or other key in $wgObjectCaches
3371 * @return BagOStuff
3372 */
3373 function wfGetCache( $cacheType ) {
3374 return ObjectCache::getInstance( $cacheType );
3375 }
3376
3377 /**
3378 * Get the main cache object
3379 *
3380 * @return BagOStuff
3381 */
3382 function wfGetMainCache() {
3383 global $wgMainCacheType;
3384 return ObjectCache::getInstance( $wgMainCacheType );
3385 }
3386
3387 /**
3388 * Get the cache object used by the message cache
3389 *
3390 * @return BagOStuff
3391 */
3392 function wfGetMessageCacheStorage() {
3393 global $wgMessageCacheType;
3394 return ObjectCache::getInstance( $wgMessageCacheType );
3395 }
3396
3397 /**
3398 * Get the cache object used by the parser cache
3399 *
3400 * @return BagOStuff
3401 */
3402 function wfGetParserCacheStorage() {
3403 global $wgParserCacheType;
3404 return ObjectCache::getInstance( $wgParserCacheType );
3405 }
3406
3407 /**
3408 * Call hook functions defined in $wgHooks
3409 *
3410 * @param string $event Event name
3411 * @param array $args Parameters passed to hook functions
3412 * @param string|null $deprecatedVersion Optionally mark hook as deprecated with version number
3413 *
3414 * @return bool True if no handler aborted the hook
3415 * @deprecated since 1.25 - use Hooks::run
3416 */
3417 function wfRunHooks( $event, array $args = [], $deprecatedVersion = null ) {
3418 return Hooks::run( $event, $args, $deprecatedVersion );
3419 }
3420
3421 /**
3422 * Wrapper around php's unpack.
3423 *
3424 * @param string $format The format string (See php's docs)
3425 * @param string $data A binary string of binary data
3426 * @param int|bool $length The minimum length of $data or false. This is to
3427 * prevent reading beyond the end of $data. false to disable the check.
3428 *
3429 * Also be careful when using this function to read unsigned 32 bit integer
3430 * because php might make it negative.
3431 *
3432 * @throws MWException If $data not long enough, or if unpack fails
3433 * @return array Associative array of the extracted data
3434 */
3435 function wfUnpack( $format, $data, $length = false ) {
3436 if ( $length !== false ) {
3437 $realLen = strlen( $data );
3438 if ( $realLen < $length ) {
3439 throw new MWException( "Tried to use wfUnpack on a "
3440 . "string of length $realLen, but needed one "
3441 . "of at least length $length."
3442 );
3443 }
3444 }
3445
3446 MediaWiki\suppressWarnings();
3447 $result = unpack( $format, $data );
3448 MediaWiki\restoreWarnings();
3449
3450 if ( $result === false ) {
3451 // If it cannot extract the packed data.
3452 throw new MWException( "unpack could not unpack binary data" );
3453 }
3454 return $result;
3455 }
3456
3457 /**
3458 * Determine if an image exists on the 'bad image list'.
3459 *
3460 * The format of MediaWiki:Bad_image_list is as follows:
3461 * * Only list items (lines starting with "*") are considered
3462 * * The first link on a line must be a link to a bad image
3463 * * Any subsequent links on the same line are considered to be exceptions,
3464 * i.e. articles where the image may occur inline.
3465 *
3466 * @param string $name The image name to check
3467 * @param Title|bool $contextTitle The page on which the image occurs, if known
3468 * @param string $blacklist Wikitext of a file blacklist
3469 * @return bool
3470 */
3471 function wfIsBadImage( $name, $contextTitle = false, $blacklist = null ) {
3472 # Handle redirects; callers almost always hit wfFindFile() anyway,
3473 # so just use that method because it has a fast process cache.
3474 $file = wfFindFile( $name ); // get the final name
3475 $name = $file ? $file->getTitle()->getDBkey() : $name;
3476
3477 # Run the extension hook
3478 $bad = false;
3479 if ( !Hooks::run( 'BadImage', [ $name, &$bad ] ) ) {
3480 return $bad;
3481 }
3482
3483 $cache = ObjectCache::getLocalServerInstance( 'hash' );
3484 $key = wfMemcKey( 'bad-image-list', ( $blacklist === null ) ? 'default' : md5( $blacklist ) );
3485 $badImages = $cache->get( $key );
3486
3487 if ( $badImages === false ) { // cache miss
3488 if ( $blacklist === null ) {
3489 $blacklist = wfMessage( 'bad_image_list' )->inContentLanguage()->plain(); // site list
3490 }
3491 # Build the list now
3492 $badImages = [];
3493 $lines = explode( "\n", $blacklist );
3494 foreach ( $lines as $line ) {
3495 # List items only
3496 if ( substr( $line, 0, 1 ) !== '*' ) {
3497 continue;
3498 }
3499
3500 # Find all links
3501 $m = [];
3502 if ( !preg_match_all( '/\[\[:?(.*?)\]\]/', $line, $m ) ) {
3503 continue;
3504 }
3505
3506 $exceptions = [];
3507 $imageDBkey = false;
3508 foreach ( $m[1] as $i => $titleText ) {
3509 $title = Title::newFromText( $titleText );
3510 if ( !is_null( $title ) ) {
3511 if ( $i == 0 ) {
3512 $imageDBkey = $title->getDBkey();
3513 } else {
3514 $exceptions[$title->getPrefixedDBkey()] = true;
3515 }
3516 }
3517 }
3518
3519 if ( $imageDBkey !== false ) {
3520 $badImages[$imageDBkey] = $exceptions;
3521 }
3522 }
3523 $cache->set( $key, $badImages, 60 );
3524 }
3525
3526 $contextKey = $contextTitle ? $contextTitle->getPrefixedDBkey() : false;
3527 $bad = isset( $badImages[$name] ) && !isset( $badImages[$name][$contextKey] );
3528
3529 return $bad;
3530 }
3531
3532 /**
3533 * Determine whether the client at a given source IP is likely to be able to
3534 * access the wiki via HTTPS.
3535 *
3536 * @param string $ip The IPv4/6 address in the normal human-readable form
3537 * @return bool
3538 */
3539 function wfCanIPUseHTTPS( $ip ) {
3540 $canDo = true;
3541 Hooks::run( 'CanIPUseHTTPS', [ $ip, &$canDo ] );
3542 return !!$canDo;
3543 }
3544
3545 /**
3546 * Determine input string is represents as infinity
3547 *
3548 * @param string $str The string to determine
3549 * @return bool
3550 * @since 1.25
3551 */
3552 function wfIsInfinity( $str ) {
3553 $infinityValues = [ 'infinite', 'indefinite', 'infinity', 'never' ];
3554 return in_array( $str, $infinityValues );
3555 }
3556
3557 /**
3558 * Returns true if these thumbnail parameters match one that MediaWiki
3559 * requests from file description pages and/or parser output.
3560 *
3561 * $params is considered non-standard if they involve a non-standard
3562 * width or any non-default parameters aside from width and page number.
3563 * The number of possible files with standard parameters is far less than
3564 * that of all combinations; rate-limiting for them can thus be more generious.
3565 *
3566 * @param File $file
3567 * @param array $params
3568 * @return bool
3569 * @since 1.24 Moved from thumb.php to GlobalFunctions in 1.25
3570 */
3571 function wfThumbIsStandard( File $file, array $params ) {
3572 global $wgThumbLimits, $wgImageLimits, $wgResponsiveImages;
3573
3574 $multipliers = [ 1 ];
3575 if ( $wgResponsiveImages ) {
3576 // These available sizes are hardcoded currently elsewhere in MediaWiki.
3577 // @see Linker::processResponsiveImages
3578 $multipliers[] = 1.5;
3579 $multipliers[] = 2;
3580 }
3581
3582 $handler = $file->getHandler();
3583 if ( !$handler || !isset( $params['width'] ) ) {
3584 return false;
3585 }
3586
3587 $basicParams = [];
3588 if ( isset( $params['page'] ) ) {
3589 $basicParams['page'] = $params['page'];
3590 }
3591
3592 $thumbLimits = [];
3593 $imageLimits = [];
3594 // Expand limits to account for multipliers
3595 foreach ( $multipliers as $multiplier ) {
3596 $thumbLimits = array_merge( $thumbLimits, array_map(
3597 function ( $width ) use ( $multiplier ) {
3598 return round( $width * $multiplier );
3599 }, $wgThumbLimits )
3600 );
3601 $imageLimits = array_merge( $imageLimits, array_map(
3602 function ( $pair ) use ( $multiplier ) {
3603 return [
3604 round( $pair[0] * $multiplier ),
3605 round( $pair[1] * $multiplier ),
3606 ];
3607 }, $wgImageLimits )
3608 );
3609 }
3610
3611 // Check if the width matches one of $wgThumbLimits
3612 if ( in_array( $params['width'], $thumbLimits ) ) {
3613 $normalParams = $basicParams + [ 'width' => $params['width'] ];
3614 // Append any default values to the map (e.g. "lossy", "lossless", ...)
3615 $handler->normaliseParams( $file, $normalParams );
3616 } else {
3617 // If not, then check if the width matchs one of $wgImageLimits
3618 $match = false;
3619 foreach ( $imageLimits as $pair ) {
3620 $normalParams = $basicParams + [ 'width' => $pair[0], 'height' => $pair[1] ];
3621 // Decide whether the thumbnail should be scaled on width or height.
3622 // Also append any default values to the map (e.g. "lossy", "lossless", ...)
3623 $handler->normaliseParams( $file, $normalParams );
3624 // Check if this standard thumbnail size maps to the given width
3625 if ( $normalParams['width'] == $params['width'] ) {
3626 $match = true;
3627 break;
3628 }
3629 }
3630 if ( !$match ) {
3631 return false; // not standard for description pages
3632 }
3633 }
3634
3635 // Check that the given values for non-page, non-width, params are just defaults
3636 foreach ( $params as $key => $value ) {
3637 if ( !isset( $normalParams[$key] ) || $normalParams[$key] != $value ) {
3638 return false;
3639 }
3640 }
3641
3642 return true;
3643 }
3644
3645 /**
3646 * Merges two (possibly) 2 dimensional arrays into the target array ($baseArray).
3647 *
3648 * Values that exist in both values will be combined with += (all values of the array
3649 * of $newValues will be added to the values of the array of $baseArray, while values,
3650 * that exists in both, the value of $baseArray will be used).
3651 *
3652 * @param array $baseArray The array where you want to add the values of $newValues to
3653 * @param array $newValues An array with new values
3654 * @return array The combined array
3655 * @since 1.26
3656 */
3657 function wfArrayPlus2d( array $baseArray, array $newValues ) {
3658 // First merge items that are in both arrays
3659 foreach ( $baseArray as $name => &$groupVal ) {
3660 if ( isset( $newValues[$name] ) ) {
3661 $groupVal += $newValues[$name];
3662 }
3663 }
3664 // Now add items that didn't exist yet
3665 $baseArray += $newValues;
3666
3667 return $baseArray;
3668 }