Merge "maintenance: Script to rename titles for Unicode uppercasing changes"
[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 MediaWiki\Linker\LinkTarget;
28 use MediaWiki\Logger\LoggerFactory;
29 use MediaWiki\MediaWikiServices;
30 use MediaWiki\ProcOpenError;
31 use MediaWiki\Session\SessionManager;
32 use MediaWiki\Shell\Shell;
33 use Wikimedia\WrappedString;
34 use Wikimedia\AtEase\AtEase;
35
36 /**
37 * Load an extension
38 *
39 * This queues an extension to be loaded through
40 * the ExtensionRegistry system.
41 *
42 * @param string $ext Name of the extension to load
43 * @param string|null $path Absolute path of where to find the extension.json file
44 * @since 1.25
45 */
46 function wfLoadExtension( $ext, $path = null ) {
47 if ( !$path ) {
48 global $wgExtensionDirectory;
49 $path = "$wgExtensionDirectory/$ext/extension.json";
50 }
51 ExtensionRegistry::getInstance()->queue( $path );
52 }
53
54 /**
55 * Load multiple extensions at once
56 *
57 * Same as wfLoadExtension, but more efficient if you
58 * are loading multiple extensions.
59 *
60 * If you want to specify custom paths, you should interact with
61 * ExtensionRegistry directly.
62 *
63 * @see wfLoadExtension
64 * @param string[] $exts Array of extension names to load
65 * @since 1.25
66 */
67 function wfLoadExtensions( array $exts ) {
68 global $wgExtensionDirectory;
69 $registry = ExtensionRegistry::getInstance();
70 foreach ( $exts as $ext ) {
71 $registry->queue( "$wgExtensionDirectory/$ext/extension.json" );
72 }
73 }
74
75 /**
76 * Load a skin
77 *
78 * @see wfLoadExtension
79 * @param string $skin Name of the extension to load
80 * @param string|null $path Absolute path of where to find the skin.json file
81 * @since 1.25
82 */
83 function wfLoadSkin( $skin, $path = null ) {
84 if ( !$path ) {
85 global $wgStyleDirectory;
86 $path = "$wgStyleDirectory/$skin/skin.json";
87 }
88 ExtensionRegistry::getInstance()->queue( $path );
89 }
90
91 /**
92 * Load multiple skins at once
93 *
94 * @see wfLoadExtensions
95 * @param string[] $skins Array of extension names to load
96 * @since 1.25
97 */
98 function wfLoadSkins( array $skins ) {
99 global $wgStyleDirectory;
100 $registry = ExtensionRegistry::getInstance();
101 foreach ( $skins as $skin ) {
102 $registry->queue( "$wgStyleDirectory/$skin/skin.json" );
103 }
104 }
105
106 /**
107 * Like array_diff( $a, $b ) except that it works with two-dimensional arrays.
108 * @param array $a
109 * @param array $b
110 * @return array
111 */
112 function wfArrayDiff2( $a, $b ) {
113 return array_udiff( $a, $b, 'wfArrayDiff2_cmp' );
114 }
115
116 /**
117 * @param array|string $a
118 * @param array|string $b
119 * @return int
120 */
121 function wfArrayDiff2_cmp( $a, $b ) {
122 if ( is_string( $a ) && is_string( $b ) ) {
123 return strcmp( $a, $b );
124 } elseif ( count( $a ) !== count( $b ) ) {
125 return count( $a ) <=> count( $b );
126 } else {
127 reset( $a );
128 reset( $b );
129 while ( key( $a ) !== null && key( $b ) !== null ) {
130 $valueA = current( $a );
131 $valueB = current( $b );
132 $cmp = strcmp( $valueA, $valueB );
133 if ( $cmp !== 0 ) {
134 return $cmp;
135 }
136 next( $a );
137 next( $b );
138 }
139 return 0;
140 }
141 }
142
143 /**
144 * Appends to second array if $value differs from that in $default
145 *
146 * @param string|int $key
147 * @param mixed $value
148 * @param mixed $default
149 * @param array &$changed Array to alter
150 * @throws MWException
151 */
152 function wfAppendToArrayIfNotDefault( $key, $value, $default, &$changed ) {
153 if ( is_null( $changed ) ) {
154 throw new MWException( 'GlobalFunctions::wfAppendToArrayIfNotDefault got null' );
155 }
156 if ( $default[$key] !== $value ) {
157 $changed[$key] = $value;
158 }
159 }
160
161 /**
162 * Merge arrays in the style of getUserPermissionsErrors, with duplicate removal
163 * e.g.
164 * wfMergeErrorArrays(
165 * [ [ 'x' ] ],
166 * [ [ 'x', '2' ] ],
167 * [ [ 'x' ] ],
168 * [ [ 'y' ] ]
169 * );
170 * returns:
171 * [
172 * [ 'x', '2' ],
173 * [ 'x' ],
174 * [ 'y' ]
175 * ]
176 *
177 * @param array ...$args
178 * @return array
179 */
180 function wfMergeErrorArrays( ...$args ) {
181 $out = [];
182 foreach ( $args as $errors ) {
183 foreach ( $errors as $params ) {
184 $originalParams = $params;
185 if ( $params[0] instanceof MessageSpecifier ) {
186 $msg = $params[0];
187 $params = array_merge( [ $msg->getKey() ], $msg->getParams() );
188 }
189 # @todo FIXME: Sometimes get nested arrays for $params,
190 # which leads to E_NOTICEs
191 $spec = implode( "\t", $params );
192 $out[$spec] = $originalParams;
193 }
194 }
195 return array_values( $out );
196 }
197
198 /**
199 * Insert array into another array after the specified *KEY*
200 *
201 * @param array $array The array.
202 * @param array $insert The array to insert.
203 * @param mixed $after The key to insert after. Callers need to make sure the key is set.
204 * @return array
205 */
206 function wfArrayInsertAfter( array $array, array $insert, $after ) {
207 // Find the offset of the element to insert after.
208 $keys = array_keys( $array );
209 $offsetByKey = array_flip( $keys );
210
211 $offset = $offsetByKey[$after];
212
213 // Insert at the specified offset
214 $before = array_slice( $array, 0, $offset + 1, true );
215 $after = array_slice( $array, $offset + 1, count( $array ) - $offset, true );
216
217 $output = $before + $insert + $after;
218
219 return $output;
220 }
221
222 /**
223 * Recursively converts the parameter (an object) to an array with the same data
224 *
225 * @param object|array $objOrArray
226 * @param bool $recursive
227 * @return array
228 */
229 function wfObjectToArray( $objOrArray, $recursive = true ) {
230 $array = [];
231 if ( is_object( $objOrArray ) ) {
232 $objOrArray = get_object_vars( $objOrArray );
233 }
234 foreach ( $objOrArray as $key => $value ) {
235 if ( $recursive && ( is_object( $value ) || is_array( $value ) ) ) {
236 $value = wfObjectToArray( $value );
237 }
238
239 $array[$key] = $value;
240 }
241
242 return $array;
243 }
244
245 /**
246 * Get a random decimal value in the domain of [0, 1), in a way
247 * not likely to give duplicate values for any realistic
248 * number of articles.
249 *
250 * @note This is designed for use in relation to Special:RandomPage
251 * and the page_random database field.
252 *
253 * @return string
254 */
255 function wfRandom() {
256 // The maximum random value is "only" 2^31-1, so get two random
257 // values to reduce the chance of dupes
258 $max = mt_getrandmax() + 1;
259 $rand = number_format( ( mt_rand() * $max + mt_rand() ) / $max / $max, 12, '.', '' );
260 return $rand;
261 }
262
263 /**
264 * Get a random string containing a number of pseudo-random hex characters.
265 *
266 * @note This is not secure, if you are trying to generate some sort
267 * of token please use MWCryptRand instead.
268 *
269 * @param int $length The length of the string to generate
270 * @return string
271 * @since 1.20
272 */
273 function wfRandomString( $length = 32 ) {
274 $str = '';
275 for ( $n = 0; $n < $length; $n += 7 ) {
276 $str .= sprintf( '%07x', mt_rand() & 0xfffffff );
277 }
278 return substr( $str, 0, $length );
279 }
280
281 /**
282 * We want some things to be included as literal characters in our title URLs
283 * for prettiness, which urlencode encodes by default. According to RFC 1738,
284 * all of the following should be safe:
285 *
286 * ;:@&=$-_.+!*'(),
287 *
288 * RFC 1738 says ~ is unsafe, however RFC 3986 considers it an unreserved
289 * character which should not be encoded. More importantly, google chrome
290 * always converts %7E back to ~, and converting it in this function can
291 * cause a redirect loop (T105265).
292 *
293 * But + is not safe because it's used to indicate a space; &= are only safe in
294 * paths and not in queries (and we don't distinguish here); ' seems kind of
295 * scary; and urlencode() doesn't touch -_. to begin with. Plus, although /
296 * is reserved, we don't care. So the list we unescape is:
297 *
298 * ;:@$!*(),/~
299 *
300 * However, IIS7 redirects fail when the url contains a colon (see T24709),
301 * so no fancy : for IIS7.
302 *
303 * %2F in the page titles seems to fatally break for some reason.
304 *
305 * @param string $s
306 * @return string
307 */
308 function wfUrlencode( $s ) {
309 static $needle;
310
311 if ( is_null( $s ) ) {
312 // Reset $needle for testing.
313 $needle = null;
314 return '';
315 }
316
317 if ( is_null( $needle ) ) {
318 $needle = [ '%3B', '%40', '%24', '%21', '%2A', '%28', '%29', '%2C', '%2F', '%7E' ];
319 if ( !isset( $_SERVER['SERVER_SOFTWARE'] ) ||
320 ( strpos( $_SERVER['SERVER_SOFTWARE'], 'Microsoft-IIS/7' ) === false )
321 ) {
322 $needle[] = '%3A';
323 }
324 }
325
326 $s = urlencode( $s );
327 $s = str_ireplace(
328 $needle,
329 [ ';', '@', '$', '!', '*', '(', ')', ',', '/', '~', ':' ],
330 $s
331 );
332
333 return $s;
334 }
335
336 /**
337 * This function takes one or two arrays as input, and returns a CGI-style string, e.g.
338 * "days=7&limit=100". Options in the first array override options in the second.
339 * Options set to null or false will not be output.
340 *
341 * @param array $array1 ( String|Array )
342 * @param array|null $array2 ( String|Array )
343 * @param string $prefix
344 * @return string
345 */
346 function wfArrayToCgi( $array1, $array2 = null, $prefix = '' ) {
347 if ( !is_null( $array2 ) ) {
348 $array1 = $array1 + $array2;
349 }
350
351 $cgi = '';
352 foreach ( $array1 as $key => $value ) {
353 if ( !is_null( $value ) && $value !== false ) {
354 if ( $cgi != '' ) {
355 $cgi .= '&';
356 }
357 if ( $prefix !== '' ) {
358 $key = $prefix . "[$key]";
359 }
360 if ( is_array( $value ) ) {
361 $firstTime = true;
362 foreach ( $value as $k => $v ) {
363 $cgi .= $firstTime ? '' : '&';
364 if ( is_array( $v ) ) {
365 $cgi .= wfArrayToCgi( $v, null, $key . "[$k]" );
366 } else {
367 $cgi .= urlencode( $key . "[$k]" ) . '=' . urlencode( $v );
368 }
369 $firstTime = false;
370 }
371 } else {
372 if ( is_object( $value ) ) {
373 $value = $value->__toString();
374 }
375 $cgi .= urlencode( $key ) . '=' . urlencode( $value );
376 }
377 }
378 }
379 return $cgi;
380 }
381
382 /**
383 * This is the logical opposite of wfArrayToCgi(): it accepts a query string as
384 * its argument and returns the same string in array form. This allows compatibility
385 * with legacy functions that accept raw query strings instead of nice
386 * arrays. Of course, keys and values are urldecode()d.
387 *
388 * @param string $query Query string
389 * @return string[] Array version of input
390 */
391 function wfCgiToArray( $query ) {
392 if ( isset( $query[0] ) && $query[0] == '?' ) {
393 $query = substr( $query, 1 );
394 }
395 $bits = explode( '&', $query );
396 $ret = [];
397 foreach ( $bits as $bit ) {
398 if ( $bit === '' ) {
399 continue;
400 }
401 if ( strpos( $bit, '=' ) === false ) {
402 // Pieces like &qwerty become 'qwerty' => '' (at least this is what php does)
403 $key = $bit;
404 $value = '';
405 } else {
406 list( $key, $value ) = explode( '=', $bit );
407 }
408 $key = urldecode( $key );
409 $value = urldecode( $value );
410 if ( strpos( $key, '[' ) !== false ) {
411 $keys = array_reverse( explode( '[', $key ) );
412 $key = array_pop( $keys );
413 $temp = $value;
414 foreach ( $keys as $k ) {
415 $k = substr( $k, 0, -1 );
416 $temp = [ $k => $temp ];
417 }
418 if ( isset( $ret[$key] ) ) {
419 $ret[$key] = array_merge( $ret[$key], $temp );
420 } else {
421 $ret[$key] = $temp;
422 }
423 } else {
424 $ret[$key] = $value;
425 }
426 }
427 return $ret;
428 }
429
430 /**
431 * Append a query string to an existing URL, which may or may not already
432 * have query string parameters already. If so, they will be combined.
433 *
434 * @param string $url
435 * @param string|string[] $query String or associative array
436 * @return string
437 */
438 function wfAppendQuery( $url, $query ) {
439 if ( is_array( $query ) ) {
440 $query = wfArrayToCgi( $query );
441 }
442 if ( $query != '' ) {
443 // Remove the fragment, if there is one
444 $fragment = false;
445 $hashPos = strpos( $url, '#' );
446 if ( $hashPos !== false ) {
447 $fragment = substr( $url, $hashPos );
448 $url = substr( $url, 0, $hashPos );
449 }
450
451 // Add parameter
452 if ( strpos( $url, '?' ) === false ) {
453 $url .= '?';
454 } else {
455 $url .= '&';
456 }
457 $url .= $query;
458
459 // Put the fragment back
460 if ( $fragment !== false ) {
461 $url .= $fragment;
462 }
463 }
464 return $url;
465 }
466
467 /**
468 * Expand a potentially local URL to a fully-qualified URL. Assumes $wgServer
469 * is correct.
470 *
471 * The meaning of the PROTO_* constants is as follows:
472 * PROTO_HTTP: Output a URL starting with http://
473 * PROTO_HTTPS: Output a URL starting with https://
474 * PROTO_RELATIVE: Output a URL starting with // (protocol-relative URL)
475 * PROTO_CURRENT: Output a URL starting with either http:// or https:// , depending
476 * on which protocol was used for the current incoming request
477 * PROTO_CANONICAL: For URLs without a domain, like /w/index.php , use $wgCanonicalServer.
478 * For protocol-relative URLs, use the protocol of $wgCanonicalServer
479 * PROTO_INTERNAL: Like PROTO_CANONICAL, but uses $wgInternalServer instead of $wgCanonicalServer
480 *
481 * @todo this won't work with current-path-relative URLs
482 * like "subdir/foo.html", etc.
483 *
484 * @param string $url Either fully-qualified or a local path + query
485 * @param string|int|null $defaultProto One of the PROTO_* constants. Determines the
486 * protocol to use if $url or $wgServer is protocol-relative
487 * @return string|false Fully-qualified URL, current-path-relative URL or false if
488 * no valid URL can be constructed
489 */
490 function wfExpandUrl( $url, $defaultProto = PROTO_CURRENT ) {
491 global $wgServer, $wgCanonicalServer, $wgInternalServer, $wgRequest,
492 $wgHttpsPort;
493 if ( $defaultProto === PROTO_CANONICAL ) {
494 $serverUrl = $wgCanonicalServer;
495 } elseif ( $defaultProto === PROTO_INTERNAL && $wgInternalServer !== false ) {
496 // Make $wgInternalServer fall back to $wgServer if not set
497 $serverUrl = $wgInternalServer;
498 } else {
499 $serverUrl = $wgServer;
500 if ( $defaultProto === PROTO_CURRENT ) {
501 $defaultProto = $wgRequest->getProtocol() . '://';
502 }
503 }
504
505 // Analyze $serverUrl to obtain its protocol
506 $bits = wfParseUrl( $serverUrl );
507 $serverHasProto = $bits && $bits['scheme'] != '';
508
509 if ( $defaultProto === PROTO_CANONICAL || $defaultProto === PROTO_INTERNAL ) {
510 if ( $serverHasProto ) {
511 $defaultProto = $bits['scheme'] . '://';
512 } else {
513 // $wgCanonicalServer or $wgInternalServer doesn't have a protocol.
514 // This really isn't supposed to happen. Fall back to HTTP in this
515 // ridiculous case.
516 $defaultProto = PROTO_HTTP;
517 }
518 }
519
520 $defaultProtoWithoutSlashes = substr( $defaultProto, 0, -2 );
521
522 if ( substr( $url, 0, 2 ) == '//' ) {
523 $url = $defaultProtoWithoutSlashes . $url;
524 } elseif ( substr( $url, 0, 1 ) == '/' ) {
525 // If $serverUrl is protocol-relative, prepend $defaultProtoWithoutSlashes,
526 // otherwise leave it alone.
527 if ( $serverHasProto ) {
528 $url = $serverUrl . $url;
529 } else {
530 // If an HTTPS URL is synthesized from a protocol-relative $wgServer, allow the
531 // user to override the port number (T67184)
532 if ( $defaultProto === PROTO_HTTPS && $wgHttpsPort != 443 ) {
533 if ( isset( $bits['port'] ) ) {
534 throw new Exception( 'A protocol-relative $wgServer may not contain a port number' );
535 }
536 $url = $defaultProtoWithoutSlashes . $serverUrl . ':' . $wgHttpsPort . $url;
537 } else {
538 $url = $defaultProtoWithoutSlashes . $serverUrl . $url;
539 }
540 }
541 }
542
543 $bits = wfParseUrl( $url );
544
545 if ( $bits && isset( $bits['path'] ) ) {
546 $bits['path'] = wfRemoveDotSegments( $bits['path'] );
547 return wfAssembleUrl( $bits );
548 } elseif ( $bits ) {
549 # No path to expand
550 return $url;
551 } elseif ( substr( $url, 0, 1 ) != '/' ) {
552 # URL is a relative path
553 return wfRemoveDotSegments( $url );
554 }
555
556 # Expanded URL is not valid.
557 return false;
558 }
559
560 /**
561 * Get the wiki's "server", i.e. the protocol and host part of the URL, with a
562 * protocol specified using a PROTO_* constant as in wfExpandUrl()
563 *
564 * @since 1.32
565 * @param string|int|null $proto One of the PROTO_* constants.
566 * @return string The URL
567 */
568 function wfGetServerUrl( $proto ) {
569 $url = wfExpandUrl( '/', $proto );
570 return substr( $url, 0, -1 );
571 }
572
573 /**
574 * This function will reassemble a URL parsed with wfParseURL. This is useful
575 * if you need to edit part of a URL and put it back together.
576 *
577 * This is the basic structure used (brackets contain keys for $urlParts):
578 * [scheme][delimiter][user]:[pass]@[host]:[port][path]?[query]#[fragment]
579 *
580 * @todo Need to integrate this into wfExpandUrl (see T34168)
581 *
582 * @since 1.19
583 * @param array $urlParts URL parts, as output from wfParseUrl
584 * @return string URL assembled from its component parts
585 */
586 function wfAssembleUrl( $urlParts ) {
587 $result = '';
588
589 if ( isset( $urlParts['delimiter'] ) ) {
590 if ( isset( $urlParts['scheme'] ) ) {
591 $result .= $urlParts['scheme'];
592 }
593
594 $result .= $urlParts['delimiter'];
595 }
596
597 if ( isset( $urlParts['host'] ) ) {
598 if ( isset( $urlParts['user'] ) ) {
599 $result .= $urlParts['user'];
600 if ( isset( $urlParts['pass'] ) ) {
601 $result .= ':' . $urlParts['pass'];
602 }
603 $result .= '@';
604 }
605
606 $result .= $urlParts['host'];
607
608 if ( isset( $urlParts['port'] ) ) {
609 $result .= ':' . $urlParts['port'];
610 }
611 }
612
613 if ( isset( $urlParts['path'] ) ) {
614 $result .= $urlParts['path'];
615 }
616
617 if ( isset( $urlParts['query'] ) ) {
618 $result .= '?' . $urlParts['query'];
619 }
620
621 if ( isset( $urlParts['fragment'] ) ) {
622 $result .= '#' . $urlParts['fragment'];
623 }
624
625 return $result;
626 }
627
628 /**
629 * Remove all dot-segments in the provided URL path. For example,
630 * '/a/./b/../c/' becomes '/a/c/'. For details on the algorithm, please see
631 * RFC3986 section 5.2.4.
632 *
633 * @todo Need to integrate this into wfExpandUrl (see T34168)
634 *
635 * @since 1.19
636 *
637 * @param string $urlPath URL path, potentially containing dot-segments
638 * @return string URL path with all dot-segments removed
639 */
640 function wfRemoveDotSegments( $urlPath ) {
641 $output = '';
642 $inputOffset = 0;
643 $inputLength = strlen( $urlPath );
644
645 while ( $inputOffset < $inputLength ) {
646 $prefixLengthOne = substr( $urlPath, $inputOffset, 1 );
647 $prefixLengthTwo = substr( $urlPath, $inputOffset, 2 );
648 $prefixLengthThree = substr( $urlPath, $inputOffset, 3 );
649 $prefixLengthFour = substr( $urlPath, $inputOffset, 4 );
650 $trimOutput = false;
651
652 if ( $prefixLengthTwo == './' ) {
653 # Step A, remove leading "./"
654 $inputOffset += 2;
655 } elseif ( $prefixLengthThree == '../' ) {
656 # Step A, remove leading "../"
657 $inputOffset += 3;
658 } elseif ( ( $prefixLengthTwo == '/.' ) && ( $inputOffset + 2 == $inputLength ) ) {
659 # Step B, replace leading "/.$" with "/"
660 $inputOffset += 1;
661 $urlPath[$inputOffset] = '/';
662 } elseif ( $prefixLengthThree == '/./' ) {
663 # Step B, replace leading "/./" with "/"
664 $inputOffset += 2;
665 } elseif ( $prefixLengthThree == '/..' && ( $inputOffset + 3 == $inputLength ) ) {
666 # Step C, replace leading "/..$" with "/" and
667 # remove last path component in output
668 $inputOffset += 2;
669 $urlPath[$inputOffset] = '/';
670 $trimOutput = true;
671 } elseif ( $prefixLengthFour == '/../' ) {
672 # Step C, replace leading "/../" with "/" and
673 # remove last path component in output
674 $inputOffset += 3;
675 $trimOutput = true;
676 } elseif ( ( $prefixLengthOne == '.' ) && ( $inputOffset + 1 == $inputLength ) ) {
677 # Step D, remove "^.$"
678 $inputOffset += 1;
679 } elseif ( ( $prefixLengthTwo == '..' ) && ( $inputOffset + 2 == $inputLength ) ) {
680 # Step D, remove "^..$"
681 $inputOffset += 2;
682 } else {
683 # Step E, move leading path segment to output
684 if ( $prefixLengthOne == '/' ) {
685 $slashPos = strpos( $urlPath, '/', $inputOffset + 1 );
686 } else {
687 $slashPos = strpos( $urlPath, '/', $inputOffset );
688 }
689 if ( $slashPos === false ) {
690 $output .= substr( $urlPath, $inputOffset );
691 $inputOffset = $inputLength;
692 } else {
693 $output .= substr( $urlPath, $inputOffset, $slashPos - $inputOffset );
694 $inputOffset += $slashPos - $inputOffset;
695 }
696 }
697
698 if ( $trimOutput ) {
699 $slashPos = strrpos( $output, '/' );
700 if ( $slashPos === false ) {
701 $output = '';
702 } else {
703 $output = substr( $output, 0, $slashPos );
704 }
705 }
706 }
707
708 return $output;
709 }
710
711 /**
712 * Returns a regular expression of url protocols
713 *
714 * @param bool $includeProtocolRelative If false, remove '//' from the returned protocol list.
715 * DO NOT USE this directly, use wfUrlProtocolsWithoutProtRel() instead
716 * @return string
717 */
718 function wfUrlProtocols( $includeProtocolRelative = true ) {
719 global $wgUrlProtocols;
720
721 // Cache return values separately based on $includeProtocolRelative
722 static $withProtRel = null, $withoutProtRel = null;
723 $cachedValue = $includeProtocolRelative ? $withProtRel : $withoutProtRel;
724 if ( !is_null( $cachedValue ) ) {
725 return $cachedValue;
726 }
727
728 // Support old-style $wgUrlProtocols strings, for backwards compatibility
729 // with LocalSettings files from 1.5
730 if ( is_array( $wgUrlProtocols ) ) {
731 $protocols = [];
732 foreach ( $wgUrlProtocols as $protocol ) {
733 // Filter out '//' if !$includeProtocolRelative
734 if ( $includeProtocolRelative || $protocol !== '//' ) {
735 $protocols[] = preg_quote( $protocol, '/' );
736 }
737 }
738
739 $retval = implode( '|', $protocols );
740 } else {
741 // Ignore $includeProtocolRelative in this case
742 // This case exists for pre-1.6 compatibility, and we can safely assume
743 // that '//' won't appear in a pre-1.6 config because protocol-relative
744 // URLs weren't supported until 1.18
745 $retval = $wgUrlProtocols;
746 }
747
748 // Cache return value
749 if ( $includeProtocolRelative ) {
750 $withProtRel = $retval;
751 } else {
752 $withoutProtRel = $retval;
753 }
754 return $retval;
755 }
756
757 /**
758 * Like wfUrlProtocols(), but excludes '//' from the protocol list. Use this if
759 * you need a regex that matches all URL protocols but does not match protocol-
760 * relative URLs
761 * @return string
762 */
763 function wfUrlProtocolsWithoutProtRel() {
764 return wfUrlProtocols( false );
765 }
766
767 /**
768 * parse_url() work-alike, but non-broken. Differences:
769 *
770 * 1) Does not raise warnings on bad URLs (just returns false).
771 * 2) Handles protocols that don't use :// (e.g., mailto: and news:, as well as
772 * protocol-relative URLs) correctly.
773 * 3) Adds a "delimiter" element to the array (see (2)).
774 * 4) Verifies that the protocol is on the $wgUrlProtocols whitelist.
775 * 5) Rejects some invalid URLs that parse_url doesn't, e.g. the empty string or URLs starting with
776 * a line feed character.
777 *
778 * @param string $url A URL to parse
779 * @return string[]|bool Bits of the URL in an associative array, or false on failure.
780 * Possible fields:
781 * - scheme: URI scheme (protocol), e.g. 'http', 'mailto'. Lowercase, always present, but can
782 * be an empty string for protocol-relative URLs.
783 * - delimiter: either '://', ':' or '//'. Always present.
784 * - host: domain name / IP. Always present, but could be an empty string, e.g. for file: URLs.
785 * - user: user name, e.g. for HTTP Basic auth URLs such as http://user:pass@example.com/
786 * Missing when there is no username.
787 * - pass: password, same as above.
788 * - path: path including the leading /. Will be missing when empty (e.g. 'http://example.com')
789 * - query: query string (as a string; see wfCgiToArray() for parsing it), can be missing.
790 * - fragment: the part after #, can be missing.
791 */
792 function wfParseUrl( $url ) {
793 global $wgUrlProtocols; // Allow all protocols defined in DefaultSettings/LocalSettings.php
794
795 // Protocol-relative URLs are handled really badly by parse_url(). It's so
796 // bad that the easiest way to handle them is to just prepend 'http:' and
797 // strip the protocol out later.
798 $wasRelative = substr( $url, 0, 2 ) == '//';
799 if ( $wasRelative ) {
800 $url = "http:$url";
801 }
802 AtEase::suppressWarnings();
803 $bits = parse_url( $url );
804 AtEase::restoreWarnings();
805 // parse_url() returns an array without scheme for some invalid URLs, e.g.
806 // parse_url("%0Ahttp://example.com") == [ 'host' => '%0Ahttp', 'path' => 'example.com' ]
807 if ( !$bits || !isset( $bits['scheme'] ) ) {
808 return false;
809 }
810
811 // parse_url() incorrectly handles schemes case-sensitively. Convert it to lowercase.
812 $bits['scheme'] = strtolower( $bits['scheme'] );
813
814 // most of the protocols are followed by ://, but mailto: and sometimes news: not, check for it
815 if ( in_array( $bits['scheme'] . '://', $wgUrlProtocols ) ) {
816 $bits['delimiter'] = '://';
817 } elseif ( in_array( $bits['scheme'] . ':', $wgUrlProtocols ) ) {
818 $bits['delimiter'] = ':';
819 // parse_url detects for news: and mailto: the host part of an url as path
820 // We have to correct this wrong detection
821 if ( isset( $bits['path'] ) ) {
822 $bits['host'] = $bits['path'];
823 $bits['path'] = '';
824 }
825 } else {
826 return false;
827 }
828
829 /* Provide an empty host for eg. file:/// urls (see T30627) */
830 if ( !isset( $bits['host'] ) ) {
831 $bits['host'] = '';
832
833 // See T47069
834 if ( isset( $bits['path'] ) ) {
835 /* parse_url loses the third / for file:///c:/ urls (but not on variants) */
836 if ( substr( $bits['path'], 0, 1 ) !== '/' ) {
837 $bits['path'] = '/' . $bits['path'];
838 }
839 } else {
840 $bits['path'] = '';
841 }
842 }
843
844 // If the URL was protocol-relative, fix scheme and delimiter
845 if ( $wasRelative ) {
846 $bits['scheme'] = '';
847 $bits['delimiter'] = '//';
848 }
849 return $bits;
850 }
851
852 /**
853 * Take a URL, make sure it's expanded to fully qualified, and replace any
854 * encoded non-ASCII Unicode characters with their UTF-8 original forms
855 * for more compact display and legibility for local audiences.
856 *
857 * @todo handle punycode domains too
858 *
859 * @param string $url
860 * @return string
861 */
862 function wfExpandIRI( $url ) {
863 return preg_replace_callback(
864 '/((?:%[89A-F][0-9A-F])+)/i',
865 function ( array $matches ) {
866 return urldecode( $matches[1] );
867 },
868 wfExpandUrl( $url )
869 );
870 }
871
872 /**
873 * Check whether a given URL has a domain that occurs in a given set of domains
874 * @param string $url
875 * @param array $domains Array of domains (strings)
876 * @return bool True if the host part of $url ends in one of the strings in $domains
877 */
878 function wfMatchesDomainList( $url, $domains ) {
879 $bits = wfParseUrl( $url );
880 if ( is_array( $bits ) && isset( $bits['host'] ) ) {
881 $host = '.' . $bits['host'];
882 foreach ( (array)$domains as $domain ) {
883 $domain = '.' . $domain;
884 if ( substr( $host, -strlen( $domain ) ) === $domain ) {
885 return true;
886 }
887 }
888 }
889 return false;
890 }
891
892 /**
893 * Sends a line to the debug log if enabled or, optionally, to a comment in output.
894 * In normal operation this is a NOP.
895 *
896 * Controlling globals:
897 * $wgDebugLogFile - points to the log file
898 * $wgDebugRawPage - if false, 'action=raw' hits will not result in debug output.
899 * $wgDebugComments - if on, some debug items may appear in comments in the HTML output.
900 *
901 * @since 1.25 support for additional context data
902 *
903 * @param string $text
904 * @param string|bool $dest Destination of the message:
905 * - 'all': both to the log and HTML (debug toolbar or HTML comments)
906 * - 'private': excluded from HTML output
907 * For backward compatibility, it can also take a boolean:
908 * - true: same as 'all'
909 * - false: same as 'private'
910 * @param array $context Additional logging context data
911 */
912 function wfDebug( $text, $dest = 'all', array $context = [] ) {
913 global $wgDebugRawPage, $wgDebugLogPrefix;
914 global $wgDebugTimestamps;
915
916 if ( !$wgDebugRawPage && wfIsDebugRawPage() ) {
917 return;
918 }
919
920 $text = trim( $text );
921
922 if ( $wgDebugTimestamps ) {
923 $context['seconds_elapsed'] = sprintf(
924 '%6.4f',
925 microtime( true ) - $_SERVER['REQUEST_TIME_FLOAT']
926 );
927 $context['memory_used'] = sprintf(
928 '%5.1fM',
929 ( memory_get_usage( true ) / ( 1024 * 1024 ) )
930 );
931 }
932
933 if ( $wgDebugLogPrefix !== '' ) {
934 $context['prefix'] = $wgDebugLogPrefix;
935 }
936 $context['private'] = ( $dest === false || $dest === 'private' );
937
938 $logger = LoggerFactory::getInstance( 'wfDebug' );
939 $logger->debug( $text, $context );
940 }
941
942 /**
943 * Returns true if debug logging should be suppressed if $wgDebugRawPage = false
944 * @return bool
945 */
946 function wfIsDebugRawPage() {
947 static $cache;
948 if ( $cache !== null ) {
949 return $cache;
950 }
951 // Check for raw action using $_GET not $wgRequest, since the latter might not be initialised yet
952 // phpcs:ignore MediaWiki.Usage.SuperGlobalsUsage.SuperGlobals
953 if ( ( isset( $_GET['action'] ) && $_GET['action'] == 'raw' )
954 || (
955 isset( $_SERVER['SCRIPT_NAME'] )
956 && substr( $_SERVER['SCRIPT_NAME'], -8 ) == 'load.php'
957 )
958 ) {
959 $cache = true;
960 } else {
961 $cache = false;
962 }
963 return $cache;
964 }
965
966 /**
967 * Send a line giving PHP memory usage.
968 *
969 * @param bool $exact Print exact byte values instead of kibibytes (default: false)
970 */
971 function wfDebugMem( $exact = false ) {
972 $mem = memory_get_usage();
973 if ( !$exact ) {
974 $mem = floor( $mem / 1024 ) . ' KiB';
975 } else {
976 $mem .= ' B';
977 }
978 wfDebug( "Memory usage: $mem\n" );
979 }
980
981 /**
982 * Send a line to a supplementary debug log file, if configured, or main debug
983 * log if not.
984 *
985 * To configure a supplementary log file, set $wgDebugLogGroups[$logGroup] to
986 * a string filename or an associative array mapping 'destination' to the
987 * desired filename. The associative array may also contain a 'sample' key
988 * with an integer value, specifying a sampling factor. Sampled log events
989 * will be emitted with a 1 in N random chance.
990 *
991 * @since 1.23 support for sampling log messages via $wgDebugLogGroups.
992 * @since 1.25 support for additional context data
993 * @since 1.25 sample behavior dependent on configured $wgMWLoggerDefaultSpi
994 *
995 * @param string $logGroup
996 * @param string $text
997 * @param string|bool $dest Destination of the message:
998 * - 'all': both to the log and HTML (debug toolbar or HTML comments)
999 * - 'private': only to the specific log if set in $wgDebugLogGroups and
1000 * discarded otherwise
1001 * For backward compatibility, it can also take a boolean:
1002 * - true: same as 'all'
1003 * - false: same as 'private'
1004 * @param array $context Additional logging context data
1005 */
1006 function wfDebugLog(
1007 $logGroup, $text, $dest = 'all', array $context = []
1008 ) {
1009 $text = trim( $text );
1010
1011 $logger = LoggerFactory::getInstance( $logGroup );
1012 $context['private'] = ( $dest === false || $dest === 'private' );
1013 $logger->info( $text, $context );
1014 }
1015
1016 /**
1017 * Log for database errors
1018 *
1019 * @since 1.25 support for additional context data
1020 *
1021 * @param string $text Database error message.
1022 * @param array $context Additional logging context data
1023 */
1024 function wfLogDBError( $text, array $context = [] ) {
1025 $logger = LoggerFactory::getInstance( 'wfLogDBError' );
1026 $logger->error( trim( $text ), $context );
1027 }
1028
1029 /**
1030 * Throws a warning that $function is deprecated
1031 *
1032 * @param string $function Function that is deprecated.
1033 * @param string|bool $version Version of MediaWiki that the function
1034 * was deprecated in (Added in 1.19).
1035 * @param string|bool $component Component to which the function belongs.
1036 * If false, it is assumed the function is in MediaWiki core (Added in 1.19).
1037 * @param int $callerOffset How far up the call stack is the original
1038 * caller. 2 = function that called the function that called
1039 * wfDeprecated (Added in 1.20).
1040 *
1041 * @throws Exception If the MediaWiki version number is not a string or boolean.
1042 */
1043 function wfDeprecated( $function, $version = false, $component = false, $callerOffset = 2 ) {
1044 if ( is_string( $version ) || is_bool( $version ) ) {
1045 MWDebug::deprecated( $function, $version, $component, $callerOffset + 1 );
1046 } else {
1047 throw new Exception(
1048 "MediaWiki version must either be a string or a boolean. " .
1049 "Example valid version: '1.33'"
1050 );
1051 }
1052 }
1053
1054 /**
1055 * Send a warning either to the debug log or in a PHP error depending on
1056 * $wgDevelopmentWarnings. To log warnings in production, use wfLogWarning() instead.
1057 *
1058 * @param string $msg Message to send
1059 * @param int $callerOffset Number of items to go back in the backtrace to
1060 * find the correct caller (1 = function calling wfWarn, ...)
1061 * @param int $level PHP error level; defaults to E_USER_NOTICE;
1062 * only used when $wgDevelopmentWarnings is true
1063 */
1064 function wfWarn( $msg, $callerOffset = 1, $level = E_USER_NOTICE ) {
1065 MWDebug::warning( $msg, $callerOffset + 1, $level, 'auto' );
1066 }
1067
1068 /**
1069 * Send a warning as a PHP error and the debug log. This is intended for logging
1070 * warnings in production. For logging development warnings, use WfWarn instead.
1071 *
1072 * @param string $msg Message to send
1073 * @param int $callerOffset Number of items to go back in the backtrace to
1074 * find the correct caller (1 = function calling wfLogWarning, ...)
1075 * @param int $level PHP error level; defaults to E_USER_WARNING
1076 */
1077 function wfLogWarning( $msg, $callerOffset = 1, $level = E_USER_WARNING ) {
1078 MWDebug::warning( $msg, $callerOffset + 1, $level, 'production' );
1079 }
1080
1081 /**
1082 * @todo document
1083 * @todo Move logic to MediaWiki.php
1084 */
1085 function wfLogProfilingData() {
1086 global $wgDebugLogGroups, $wgDebugRawPage;
1087
1088 $context = RequestContext::getMain();
1089 $request = $context->getRequest();
1090
1091 $profiler = Profiler::instance();
1092 $profiler->setContext( $context );
1093 $profiler->logData();
1094
1095 // Send out any buffered statsd metrics as needed
1096 MediaWiki::emitBufferedStatsdData(
1097 MediaWikiServices::getInstance()->getStatsdDataFactory(),
1098 $context->getConfig()
1099 );
1100
1101 // Profiling must actually be enabled...
1102 if ( $profiler instanceof ProfilerStub ) {
1103 return;
1104 }
1105
1106 if ( isset( $wgDebugLogGroups['profileoutput'] )
1107 && $wgDebugLogGroups['profileoutput'] === false
1108 ) {
1109 // Explicitly disabled
1110 return;
1111 }
1112 if ( !$wgDebugRawPage && wfIsDebugRawPage() ) {
1113 return;
1114 }
1115
1116 $ctx = [ 'elapsed' => $request->getElapsedTime() ];
1117 if ( !empty( $_SERVER['HTTP_X_FORWARDED_FOR'] ) ) {
1118 $ctx['forwarded_for'] = $_SERVER['HTTP_X_FORWARDED_FOR'];
1119 }
1120 if ( !empty( $_SERVER['HTTP_CLIENT_IP'] ) ) {
1121 $ctx['client_ip'] = $_SERVER['HTTP_CLIENT_IP'];
1122 }
1123 if ( !empty( $_SERVER['HTTP_FROM'] ) ) {
1124 $ctx['from'] = $_SERVER['HTTP_FROM'];
1125 }
1126 if ( isset( $ctx['forwarded_for'] ) ||
1127 isset( $ctx['client_ip'] ) ||
1128 isset( $ctx['from'] ) ) {
1129 $ctx['proxy'] = $_SERVER['REMOTE_ADDR'];
1130 }
1131
1132 // Don't load $wgUser at this late stage just for statistics purposes
1133 // @todo FIXME: We can detect some anons even if it is not loaded.
1134 // See User::getId()
1135 $user = $context->getUser();
1136 $ctx['anon'] = $user->isItemLoaded( 'id' ) && $user->isAnon();
1137
1138 // Command line script uses a FauxRequest object which does not have
1139 // any knowledge about an URL and throw an exception instead.
1140 try {
1141 $ctx['url'] = urldecode( $request->getRequestURL() );
1142 } catch ( Exception $ignored ) {
1143 // no-op
1144 }
1145
1146 $ctx['output'] = $profiler->getOutput();
1147
1148 $log = LoggerFactory::getInstance( 'profileoutput' );
1149 $log->info( "Elapsed: {elapsed}; URL: <{url}>\n{output}", $ctx );
1150 }
1151
1152 /**
1153 * Increment a statistics counter
1154 *
1155 * @param string $key
1156 * @param int $count
1157 * @return void
1158 */
1159 function wfIncrStats( $key, $count = 1 ) {
1160 $stats = MediaWikiServices::getInstance()->getStatsdDataFactory();
1161 $stats->updateCount( $key, $count );
1162 }
1163
1164 /**
1165 * Check whether the wiki is in read-only mode.
1166 *
1167 * @return bool
1168 */
1169 function wfReadOnly() {
1170 return MediaWikiServices::getInstance()->getReadOnlyMode()
1171 ->isReadOnly();
1172 }
1173
1174 /**
1175 * Check if the site is in read-only mode and return the message if so
1176 *
1177 * This checks wfConfiguredReadOnlyReason() and the main load balancer
1178 * for replica DB lag. This may result in DB connection being made.
1179 *
1180 * @return string|bool String when in read-only mode; false otherwise
1181 */
1182 function wfReadOnlyReason() {
1183 return MediaWikiServices::getInstance()->getReadOnlyMode()
1184 ->getReason();
1185 }
1186
1187 /**
1188 * Get the value of $wgReadOnly or the contents of $wgReadOnlyFile.
1189 *
1190 * @return string|bool String when in read-only mode; false otherwise
1191 * @since 1.27
1192 */
1193 function wfConfiguredReadOnlyReason() {
1194 return MediaWikiServices::getInstance()->getConfiguredReadOnlyMode()
1195 ->getReason();
1196 }
1197
1198 /**
1199 * Return a Language object from $langcode
1200 *
1201 * @param Language|string|bool $langcode Either:
1202 * - a Language object
1203 * - code of the language to get the message for, if it is
1204 * a valid code create a language for that language, if
1205 * it is a string but not a valid code then make a basic
1206 * language object
1207 * - a boolean: if it's false then use the global object for
1208 * the current user's language (as a fallback for the old parameter
1209 * functionality), or if it is true then use global object
1210 * for the wiki's content language.
1211 * @return Language
1212 */
1213 function wfGetLangObj( $langcode = false ) {
1214 # Identify which language to get or create a language object for.
1215 # Using is_object here due to Stub objects.
1216 if ( is_object( $langcode ) ) {
1217 # Great, we already have the object (hopefully)!
1218 return $langcode;
1219 }
1220
1221 global $wgLanguageCode;
1222 if ( $langcode === true || $langcode === $wgLanguageCode ) {
1223 # $langcode is the language code of the wikis content language object.
1224 # or it is a boolean and value is true
1225 return MediaWikiServices::getInstance()->getContentLanguage();
1226 }
1227
1228 global $wgLang;
1229 if ( $langcode === false || $langcode === $wgLang->getCode() ) {
1230 # $langcode is the language code of user language object.
1231 # or it was a boolean and value is false
1232 return $wgLang;
1233 }
1234
1235 $validCodes = array_keys( Language::fetchLanguageNames() );
1236 if ( in_array( $langcode, $validCodes ) ) {
1237 # $langcode corresponds to a valid language.
1238 return Language::factory( $langcode );
1239 }
1240
1241 # $langcode is a string, but not a valid language code; use content language.
1242 wfDebug( "Invalid language code passed to wfGetLangObj, falling back to content language.\n" );
1243 return MediaWikiServices::getInstance()->getContentLanguage();
1244 }
1245
1246 /**
1247 * This is the function for getting translated interface messages.
1248 *
1249 * @see Message class for documentation how to use them.
1250 * @see https://www.mediawiki.org/wiki/Manual:Messages_API
1251 *
1252 * This function replaces all old wfMsg* functions.
1253 *
1254 * @param string|string[]|MessageSpecifier $key Message key, or array of keys, or a MessageSpecifier
1255 * @param string|string[] ...$params Normal message parameters
1256 * @return Message
1257 *
1258 * @since 1.17
1259 *
1260 * @see Message::__construct
1261 */
1262 function wfMessage( $key, ...$params ) {
1263 $message = new Message( $key );
1264
1265 // We call Message::params() to reduce code duplication
1266 if ( $params ) {
1267 $message->params( ...$params );
1268 }
1269
1270 return $message;
1271 }
1272
1273 /**
1274 * This function accepts multiple message keys and returns a message instance
1275 * for the first message which is non-empty. If all messages are empty then an
1276 * instance of the first message key is returned.
1277 *
1278 * @param string ...$keys Message keys
1279 * @return Message
1280 *
1281 * @since 1.18
1282 *
1283 * @see Message::newFallbackSequence
1284 */
1285 function wfMessageFallback( ...$keys ) {
1286 return Message::newFallbackSequence( ...$keys );
1287 }
1288
1289 /**
1290 * Replace message parameter keys on the given formatted output.
1291 *
1292 * @param string $message
1293 * @param array $args
1294 * @return string
1295 * @private
1296 */
1297 function wfMsgReplaceArgs( $message, $args ) {
1298 # Fix windows line-endings
1299 # Some messages are split with explode("\n", $msg)
1300 $message = str_replace( "\r", '', $message );
1301
1302 // Replace arguments
1303 if ( is_array( $args ) && $args ) {
1304 if ( is_array( $args[0] ) ) {
1305 $args = array_values( $args[0] );
1306 }
1307 $replacementKeys = [];
1308 foreach ( $args as $n => $param ) {
1309 $replacementKeys['$' . ( $n + 1 )] = $param;
1310 }
1311 $message = strtr( $message, $replacementKeys );
1312 }
1313
1314 return $message;
1315 }
1316
1317 /**
1318 * Fetch server name for use in error reporting etc.
1319 * Use real server name if available, so we know which machine
1320 * in a server farm generated the current page.
1321 *
1322 * @return string
1323 */
1324 function wfHostname() {
1325 static $host;
1326 if ( is_null( $host ) ) {
1327 # Hostname overriding
1328 global $wgOverrideHostname;
1329 if ( $wgOverrideHostname !== false ) {
1330 # Set static and skip any detection
1331 $host = $wgOverrideHostname;
1332 return $host;
1333 }
1334
1335 if ( function_exists( 'posix_uname' ) ) {
1336 // This function not present on Windows
1337 $uname = posix_uname();
1338 } else {
1339 $uname = false;
1340 }
1341 if ( is_array( $uname ) && isset( $uname['nodename'] ) ) {
1342 $host = $uname['nodename'];
1343 } elseif ( getenv( 'COMPUTERNAME' ) ) {
1344 # Windows computer name
1345 $host = getenv( 'COMPUTERNAME' );
1346 } else {
1347 # This may be a virtual server.
1348 $host = $_SERVER['SERVER_NAME'];
1349 }
1350 }
1351 return $host;
1352 }
1353
1354 /**
1355 * Returns a script tag that stores the amount of time it took MediaWiki to
1356 * handle the request in milliseconds as 'wgBackendResponseTime'.
1357 *
1358 * If $wgShowHostnames is true, the script will also set 'wgHostname' to the
1359 * hostname of the server handling the request.
1360 *
1361 * @param string|null $nonce Value from OutputPage::getCSPNonce
1362 * @return string|WrappedString HTML
1363 */
1364 function wfReportTime( $nonce = null ) {
1365 global $wgShowHostnames;
1366
1367 $elapsed = ( microtime( true ) - $_SERVER['REQUEST_TIME_FLOAT'] );
1368 // seconds to milliseconds
1369 $responseTime = round( $elapsed * 1000 );
1370 $reportVars = [ 'wgBackendResponseTime' => $responseTime ];
1371 if ( $wgShowHostnames ) {
1372 $reportVars['wgHostname'] = wfHostname();
1373 }
1374 return Skin::makeVariablesScript( $reportVars, $nonce );
1375 }
1376
1377 /**
1378 * Safety wrapper for debug_backtrace().
1379 *
1380 * Will return an empty array if debug_backtrace is disabled, otherwise
1381 * the output from debug_backtrace() (trimmed).
1382 *
1383 * @param int $limit This parameter can be used to limit the number of stack frames returned
1384 *
1385 * @return array Array of backtrace information
1386 */
1387 function wfDebugBacktrace( $limit = 0 ) {
1388 static $disabled = null;
1389
1390 if ( is_null( $disabled ) ) {
1391 $disabled = !function_exists( 'debug_backtrace' );
1392 if ( $disabled ) {
1393 wfDebug( "debug_backtrace() is disabled\n" );
1394 }
1395 }
1396 if ( $disabled ) {
1397 return [];
1398 }
1399
1400 if ( $limit ) {
1401 return array_slice( debug_backtrace( DEBUG_BACKTRACE_PROVIDE_OBJECT, $limit + 1 ), 1 );
1402 } else {
1403 return array_slice( debug_backtrace(), 1 );
1404 }
1405 }
1406
1407 /**
1408 * Get a debug backtrace as a string
1409 *
1410 * @param bool|null $raw If true, the return value is plain text. If false, HTML.
1411 * Defaults to $wgCommandLineMode if unset.
1412 * @return string
1413 * @since 1.25 Supports $raw parameter.
1414 */
1415 function wfBacktrace( $raw = null ) {
1416 global $wgCommandLineMode;
1417
1418 if ( $raw === null ) {
1419 $raw = $wgCommandLineMode;
1420 }
1421
1422 if ( $raw ) {
1423 $frameFormat = "%s line %s calls %s()\n";
1424 $traceFormat = "%s";
1425 } else {
1426 $frameFormat = "<li>%s line %s calls %s()</li>\n";
1427 $traceFormat = "<ul>\n%s</ul>\n";
1428 }
1429
1430 $frames = array_map( function ( $frame ) use ( $frameFormat ) {
1431 $file = !empty( $frame['file'] ) ? basename( $frame['file'] ) : '-';
1432 $line = $frame['line'] ?? '-';
1433 $call = $frame['function'];
1434 if ( !empty( $frame['class'] ) ) {
1435 $call = $frame['class'] . $frame['type'] . $call;
1436 }
1437 return sprintf( $frameFormat, $file, $line, $call );
1438 }, wfDebugBacktrace() );
1439
1440 return sprintf( $traceFormat, implode( '', $frames ) );
1441 }
1442
1443 /**
1444 * Get the name of the function which called this function
1445 * wfGetCaller( 1 ) is the function with the wfGetCaller() call (ie. __FUNCTION__)
1446 * wfGetCaller( 2 ) [default] is the caller of the function running wfGetCaller()
1447 * wfGetCaller( 3 ) is the parent of that.
1448 *
1449 * @param int $level
1450 * @return string
1451 */
1452 function wfGetCaller( $level = 2 ) {
1453 $backtrace = wfDebugBacktrace( $level + 1 );
1454 if ( isset( $backtrace[$level] ) ) {
1455 return wfFormatStackFrame( $backtrace[$level] );
1456 } else {
1457 return 'unknown';
1458 }
1459 }
1460
1461 /**
1462 * Return a string consisting of callers in the stack. Useful sometimes
1463 * for profiling specific points.
1464 *
1465 * @param int $limit The maximum depth of the stack frame to return, or false for the entire stack.
1466 * @return string
1467 */
1468 function wfGetAllCallers( $limit = 3 ) {
1469 $trace = array_reverse( wfDebugBacktrace() );
1470 if ( !$limit || $limit > count( $trace ) - 1 ) {
1471 $limit = count( $trace ) - 1;
1472 }
1473 $trace = array_slice( $trace, -$limit - 1, $limit );
1474 return implode( '/', array_map( 'wfFormatStackFrame', $trace ) );
1475 }
1476
1477 /**
1478 * Return a string representation of frame
1479 *
1480 * @param array $frame
1481 * @return string
1482 */
1483 function wfFormatStackFrame( $frame ) {
1484 if ( !isset( $frame['function'] ) ) {
1485 return 'NO_FUNCTION_GIVEN';
1486 }
1487 return isset( $frame['class'] ) && isset( $frame['type'] ) ?
1488 $frame['class'] . $frame['type'] . $frame['function'] :
1489 $frame['function'];
1490 }
1491
1492 /* Some generic result counters, pulled out of SearchEngine */
1493
1494 /**
1495 * @todo document
1496 *
1497 * @param int $offset
1498 * @param int $limit
1499 * @return string
1500 */
1501 function wfShowingResults( $offset, $limit ) {
1502 return wfMessage( 'showingresults' )->numParams( $limit, $offset + 1 )->parse();
1503 }
1504
1505 /**
1506 * Whether the client accept gzip encoding
1507 *
1508 * Uses the Accept-Encoding header to check if the client supports gzip encoding.
1509 * Use this when considering to send a gzip-encoded response to the client.
1510 *
1511 * @param bool $force Forces another check even if we already have a cached result.
1512 * @return bool
1513 */
1514 function wfClientAcceptsGzip( $force = false ) {
1515 static $result = null;
1516 if ( $result === null || $force ) {
1517 $result = false;
1518 if ( isset( $_SERVER['HTTP_ACCEPT_ENCODING'] ) ) {
1519 # @todo FIXME: We may want to blacklist some broken browsers
1520 $m = [];
1521 if ( preg_match(
1522 '/\bgzip(?:;(q)=([0-9]+(?:\.[0-9]+)))?\b/',
1523 $_SERVER['HTTP_ACCEPT_ENCODING'],
1524 $m
1525 )
1526 ) {
1527 if ( isset( $m[2] ) && ( $m[1] == 'q' ) && ( $m[2] == 0 ) ) {
1528 $result = false;
1529 return $result;
1530 }
1531 wfDebug( "wfClientAcceptsGzip: client accepts gzip.\n" );
1532 $result = true;
1533 }
1534 }
1535 }
1536 return $result;
1537 }
1538
1539 /**
1540 * Escapes the given text so that it may be output using addWikiText()
1541 * without any linking, formatting, etc. making its way through. This
1542 * is achieved by substituting certain characters with HTML entities.
1543 * As required by the callers, "<nowiki>" is not used.
1544 *
1545 * @param string $text Text to be escaped
1546 * @param-taint $text escapes_html
1547 * @return string
1548 */
1549 function wfEscapeWikiText( $text ) {
1550 global $wgEnableMagicLinks;
1551 static $repl = null, $repl2 = null;
1552 if ( $repl === null || defined( 'MW_PARSER_TEST' ) || defined( 'MW_PHPUNIT_TEST' ) ) {
1553 // Tests depend upon being able to change $wgEnableMagicLinks, so don't cache
1554 // in those situations
1555 $repl = [
1556 '"' => '&#34;', '&' => '&#38;', "'" => '&#39;', '<' => '&#60;',
1557 '=' => '&#61;', '>' => '&#62;', '[' => '&#91;', ']' => '&#93;',
1558 '{' => '&#123;', '|' => '&#124;', '}' => '&#125;', ';' => '&#59;',
1559 "\n#" => "\n&#35;", "\r#" => "\r&#35;",
1560 "\n*" => "\n&#42;", "\r*" => "\r&#42;",
1561 "\n:" => "\n&#58;", "\r:" => "\r&#58;",
1562 "\n " => "\n&#32;", "\r " => "\r&#32;",
1563 "\n\n" => "\n&#10;", "\r\n" => "&#13;\n",
1564 "\n\r" => "\n&#13;", "\r\r" => "\r&#13;",
1565 "\n\t" => "\n&#9;", "\r\t" => "\r&#9;", // "\n\t\n" is treated like "\n\n"
1566 "\n----" => "\n&#45;---", "\r----" => "\r&#45;---",
1567 '__' => '_&#95;', '://' => '&#58;//',
1568 ];
1569
1570 $magicLinks = array_keys( array_filter( $wgEnableMagicLinks ) );
1571 // We have to catch everything "\s" matches in PCRE
1572 foreach ( $magicLinks as $magic ) {
1573 $repl["$magic "] = "$magic&#32;";
1574 $repl["$magic\t"] = "$magic&#9;";
1575 $repl["$magic\r"] = "$magic&#13;";
1576 $repl["$magic\n"] = "$magic&#10;";
1577 $repl["$magic\f"] = "$magic&#12;";
1578 }
1579
1580 // And handle protocols that don't use "://"
1581 global $wgUrlProtocols;
1582 $repl2 = [];
1583 foreach ( $wgUrlProtocols as $prot ) {
1584 if ( substr( $prot, -1 ) === ':' ) {
1585 $repl2[] = preg_quote( substr( $prot, 0, -1 ), '/' );
1586 }
1587 }
1588 $repl2 = $repl2 ? '/\b(' . implode( '|', $repl2 ) . '):/i' : '/^(?!)/';
1589 }
1590 $text = substr( strtr( "\n$text", $repl ), 1 );
1591 $text = preg_replace( $repl2, '$1&#58;', $text );
1592 return $text;
1593 }
1594
1595 /**
1596 * Sets dest to source and returns the original value of dest
1597 * If source is NULL, it just returns the value, it doesn't set the variable
1598 * If force is true, it will set the value even if source is NULL
1599 *
1600 * @param mixed &$dest
1601 * @param mixed $source
1602 * @param bool $force
1603 * @return mixed
1604 */
1605 function wfSetVar( &$dest, $source, $force = false ) {
1606 $temp = $dest;
1607 if ( !is_null( $source ) || $force ) {
1608 $dest = $source;
1609 }
1610 return $temp;
1611 }
1612
1613 /**
1614 * As for wfSetVar except setting a bit
1615 *
1616 * @param int &$dest
1617 * @param int $bit
1618 * @param bool $state
1619 *
1620 * @return bool
1621 */
1622 function wfSetBit( &$dest, $bit, $state = true ) {
1623 $temp = (bool)( $dest & $bit );
1624 if ( !is_null( $state ) ) {
1625 if ( $state ) {
1626 $dest |= $bit;
1627 } else {
1628 $dest &= ~$bit;
1629 }
1630 }
1631 return $temp;
1632 }
1633
1634 /**
1635 * A wrapper around the PHP function var_export().
1636 * Either print it or add it to the regular output ($wgOut).
1637 *
1638 * @param mixed $var A PHP variable to dump.
1639 */
1640 function wfVarDump( $var ) {
1641 global $wgOut;
1642 $s = str_replace( "\n", "<br />\n", var_export( $var, true ) . "\n" );
1643 if ( headers_sent() || !isset( $wgOut ) || !is_object( $wgOut ) ) {
1644 print $s;
1645 } else {
1646 $wgOut->addHTML( $s );
1647 }
1648 }
1649
1650 /**
1651 * Provide a simple HTTP error.
1652 *
1653 * @param int|string $code
1654 * @param string $label
1655 * @param string $desc
1656 */
1657 function wfHttpError( $code, $label, $desc ) {
1658 global $wgOut;
1659 HttpStatus::header( $code );
1660 if ( $wgOut ) {
1661 $wgOut->disable();
1662 $wgOut->sendCacheControl();
1663 }
1664
1665 MediaWiki\HeaderCallback::warnIfHeadersSent();
1666 header( 'Content-type: text/html; charset=utf-8' );
1667 print '<!DOCTYPE html>' .
1668 '<html><head><title>' .
1669 htmlspecialchars( $label ) .
1670 '</title></head><body><h1>' .
1671 htmlspecialchars( $label ) .
1672 '</h1><p>' .
1673 nl2br( htmlspecialchars( $desc ) ) .
1674 "</p></body></html>\n";
1675 }
1676
1677 /**
1678 * Clear away any user-level output buffers, discarding contents.
1679 *
1680 * Suitable for 'starting afresh', for instance when streaming
1681 * relatively large amounts of data without buffering, or wanting to
1682 * output image files without ob_gzhandler's compression.
1683 *
1684 * The optional $resetGzipEncoding parameter controls suppression of
1685 * the Content-Encoding header sent by ob_gzhandler; by default it
1686 * is left. See comments for wfClearOutputBuffers() for why it would
1687 * be used.
1688 *
1689 * Note that some PHP configuration options may add output buffer
1690 * layers which cannot be removed; these are left in place.
1691 *
1692 * @param bool $resetGzipEncoding
1693 */
1694 function wfResetOutputBuffers( $resetGzipEncoding = true ) {
1695 if ( $resetGzipEncoding ) {
1696 // Suppress Content-Encoding and Content-Length
1697 // headers from OutputHandler::handle.
1698 global $wgDisableOutputCompression;
1699 $wgDisableOutputCompression = true;
1700 }
1701 while ( $status = ob_get_status() ) {
1702 if ( isset( $status['flags'] ) ) {
1703 $flags = PHP_OUTPUT_HANDLER_CLEANABLE | PHP_OUTPUT_HANDLER_REMOVABLE;
1704 $deleteable = ( $status['flags'] & $flags ) === $flags;
1705 } elseif ( isset( $status['del'] ) ) {
1706 $deleteable = $status['del'];
1707 } else {
1708 // Guess that any PHP-internal setting can't be removed.
1709 $deleteable = $status['type'] !== 0; /* PHP_OUTPUT_HANDLER_INTERNAL */
1710 }
1711 if ( !$deleteable ) {
1712 // Give up, and hope the result doesn't break
1713 // output behavior.
1714 break;
1715 }
1716 if ( $status['name'] === 'MediaWikiTestCase::wfResetOutputBuffersBarrier' ) {
1717 // Unit testing barrier to prevent this function from breaking PHPUnit.
1718 break;
1719 }
1720 if ( !ob_end_clean() ) {
1721 // Could not remove output buffer handler; abort now
1722 // to avoid getting in some kind of infinite loop.
1723 break;
1724 }
1725 if ( $resetGzipEncoding && $status['name'] == 'ob_gzhandler' ) {
1726 // Reset the 'Content-Encoding' field set by this handler
1727 // so we can start fresh.
1728 header_remove( 'Content-Encoding' );
1729 break;
1730 }
1731 }
1732 }
1733
1734 /**
1735 * More legible than passing a 'false' parameter to wfResetOutputBuffers():
1736 *
1737 * Clear away output buffers, but keep the Content-Encoding header
1738 * produced by ob_gzhandler, if any.
1739 *
1740 * This should be used for HTTP 304 responses, where you need to
1741 * preserve the Content-Encoding header of the real result, but
1742 * also need to suppress the output of ob_gzhandler to keep to spec
1743 * and avoid breaking Firefox in rare cases where the headers and
1744 * body are broken over two packets.
1745 */
1746 function wfClearOutputBuffers() {
1747 wfResetOutputBuffers( false );
1748 }
1749
1750 /**
1751 * Converts an Accept-* header into an array mapping string values to quality
1752 * factors
1753 *
1754 * @param string $accept
1755 * @param string $def Default
1756 * @return float[] Associative array of string => float pairs
1757 */
1758 function wfAcceptToPrefs( $accept, $def = '*/*' ) {
1759 # No arg means accept anything (per HTTP spec)
1760 if ( !$accept ) {
1761 return [ $def => 1.0 ];
1762 }
1763
1764 $prefs = [];
1765
1766 $parts = explode( ',', $accept );
1767
1768 foreach ( $parts as $part ) {
1769 # @todo FIXME: Doesn't deal with params like 'text/html; level=1'
1770 $values = explode( ';', trim( $part ) );
1771 $match = [];
1772 if ( count( $values ) == 1 ) {
1773 $prefs[$values[0]] = 1.0;
1774 } elseif ( preg_match( '/q\s*=\s*(\d*\.\d+)/', $values[1], $match ) ) {
1775 $prefs[$values[0]] = floatval( $match[1] );
1776 }
1777 }
1778
1779 return $prefs;
1780 }
1781
1782 /**
1783 * Checks if a given MIME type matches any of the keys in the given
1784 * array. Basic wildcards are accepted in the array keys.
1785 *
1786 * Returns the matching MIME type (or wildcard) if a match, otherwise
1787 * NULL if no match.
1788 *
1789 * @param string $type
1790 * @param array $avail
1791 * @return string
1792 * @private
1793 */
1794 function mimeTypeMatch( $type, $avail ) {
1795 if ( array_key_exists( $type, $avail ) ) {
1796 return $type;
1797 } else {
1798 $mainType = explode( '/', $type )[0];
1799 if ( array_key_exists( "$mainType/*", $avail ) ) {
1800 return "$mainType/*";
1801 } elseif ( array_key_exists( '*/*', $avail ) ) {
1802 return '*/*';
1803 } else {
1804 return null;
1805 }
1806 }
1807 }
1808
1809 /**
1810 * Returns the 'best' match between a client's requested internet media types
1811 * and the server's list of available types. Each list should be an associative
1812 * array of type to preference (preference is a float between 0.0 and 1.0).
1813 * Wildcards in the types are acceptable.
1814 *
1815 * @param array $cprefs Client's acceptable type list
1816 * @param array $sprefs Server's offered types
1817 * @return string
1818 *
1819 * @todo FIXME: Doesn't handle params like 'text/plain; charset=UTF-8'
1820 * XXX: generalize to negotiate other stuff
1821 */
1822 function wfNegotiateType( $cprefs, $sprefs ) {
1823 $combine = [];
1824
1825 foreach ( array_keys( $sprefs ) as $type ) {
1826 $subType = explode( '/', $type )[1];
1827 if ( $subType != '*' ) {
1828 $ckey = mimeTypeMatch( $type, $cprefs );
1829 if ( $ckey ) {
1830 $combine[$type] = $sprefs[$type] * $cprefs[$ckey];
1831 }
1832 }
1833 }
1834
1835 foreach ( array_keys( $cprefs ) as $type ) {
1836 $subType = explode( '/', $type )[1];
1837 if ( $subType != '*' && !array_key_exists( $type, $sprefs ) ) {
1838 $skey = mimeTypeMatch( $type, $sprefs );
1839 if ( $skey ) {
1840 $combine[$type] = $sprefs[$skey] * $cprefs[$type];
1841 }
1842 }
1843 }
1844
1845 $bestq = 0;
1846 $besttype = null;
1847
1848 foreach ( array_keys( $combine ) as $type ) {
1849 if ( $combine[$type] > $bestq ) {
1850 $besttype = $type;
1851 $bestq = $combine[$type];
1852 }
1853 }
1854
1855 return $besttype;
1856 }
1857
1858 /**
1859 * Get a timestamp string in one of various formats
1860 *
1861 * @param mixed $outputtype A timestamp in one of the supported formats, the
1862 * function will autodetect which format is supplied and act accordingly.
1863 * @param mixed $ts Optional timestamp to convert, default 0 for the current time
1864 * @return string|bool String / false The same date in the format specified in $outputtype or false
1865 */
1866 function wfTimestamp( $outputtype = TS_UNIX, $ts = 0 ) {
1867 $ret = MWTimestamp::convert( $outputtype, $ts );
1868 if ( $ret === false ) {
1869 wfDebug( "wfTimestamp() fed bogus time value: TYPE=$outputtype; VALUE=$ts\n" );
1870 }
1871 return $ret;
1872 }
1873
1874 /**
1875 * Return a formatted timestamp, or null if input is null.
1876 * For dealing with nullable timestamp columns in the database.
1877 *
1878 * @param int $outputtype
1879 * @param string|null $ts
1880 * @return string
1881 */
1882 function wfTimestampOrNull( $outputtype = TS_UNIX, $ts = null ) {
1883 if ( is_null( $ts ) ) {
1884 return null;
1885 } else {
1886 return wfTimestamp( $outputtype, $ts );
1887 }
1888 }
1889
1890 /**
1891 * Convenience function; returns MediaWiki timestamp for the present time.
1892 *
1893 * @return string TS_MW timestamp
1894 */
1895 function wfTimestampNow() {
1896 return MWTimestamp::now( TS_MW );
1897 }
1898
1899 /**
1900 * Check if the operating system is Windows
1901 *
1902 * @return bool True if it's Windows, false otherwise.
1903 */
1904 function wfIsWindows() {
1905 static $isWindows = null;
1906 if ( $isWindows === null ) {
1907 $isWindows = strtoupper( substr( PHP_OS, 0, 3 ) ) === 'WIN';
1908 }
1909 return $isWindows;
1910 }
1911
1912 /**
1913 * Check if we are running under HHVM
1914 *
1915 * @return bool
1916 */
1917 function wfIsHHVM() {
1918 return defined( 'HHVM_VERSION' );
1919 }
1920
1921 /**
1922 * Check if we are running from the commandline
1923 *
1924 * @since 1.31
1925 * @return bool
1926 */
1927 function wfIsCLI() {
1928 return PHP_SAPI === 'cli' || PHP_SAPI === 'phpdbg';
1929 }
1930
1931 /**
1932 * Tries to get the system directory for temporary files. First
1933 * $wgTmpDirectory is checked, and then the TMPDIR, TMP, and TEMP
1934 * environment variables are then checked in sequence, then
1935 * sys_get_temp_dir(), then upload_tmp_dir from php.ini.
1936 *
1937 * NOTE: When possible, use instead the tmpfile() function to create
1938 * temporary files to avoid race conditions on file creation, etc.
1939 *
1940 * @return string
1941 */
1942 function wfTempDir() {
1943 global $wgTmpDirectory;
1944
1945 if ( $wgTmpDirectory !== false ) {
1946 return $wgTmpDirectory;
1947 }
1948
1949 return TempFSFile::getUsableTempDirectory();
1950 }
1951
1952 /**
1953 * Make directory, and make all parent directories if they don't exist
1954 *
1955 * @param string $dir Full path to directory to create
1956 * @param int|null $mode Chmod value to use, default is $wgDirectoryMode
1957 * @param string|null $caller Optional caller param for debugging.
1958 * @throws MWException
1959 * @return bool
1960 */
1961 function wfMkdirParents( $dir, $mode = null, $caller = null ) {
1962 global $wgDirectoryMode;
1963
1964 if ( FileBackend::isStoragePath( $dir ) ) { // sanity
1965 throw new MWException( __FUNCTION__ . " given storage path '$dir'." );
1966 }
1967
1968 if ( !is_null( $caller ) ) {
1969 wfDebug( "$caller: called wfMkdirParents($dir)\n" );
1970 }
1971
1972 if ( strval( $dir ) === '' || is_dir( $dir ) ) {
1973 return true;
1974 }
1975
1976 $dir = str_replace( [ '\\', '/' ], DIRECTORY_SEPARATOR, $dir );
1977
1978 if ( is_null( $mode ) ) {
1979 $mode = $wgDirectoryMode;
1980 }
1981
1982 // Turn off the normal warning, we're doing our own below
1983 AtEase::suppressWarnings();
1984 $ok = mkdir( $dir, $mode, true ); // PHP5 <3
1985 AtEase::restoreWarnings();
1986
1987 if ( !$ok ) {
1988 // directory may have been created on another request since we last checked
1989 if ( is_dir( $dir ) ) {
1990 return true;
1991 }
1992
1993 // PHP doesn't report the path in its warning message, so add our own to aid in diagnosis.
1994 wfLogWarning( sprintf( "failed to mkdir \"%s\" mode 0%o", $dir, $mode ) );
1995 }
1996 return $ok;
1997 }
1998
1999 /**
2000 * Remove a directory and all its content.
2001 * Does not hide error.
2002 * @param string $dir
2003 */
2004 function wfRecursiveRemoveDir( $dir ) {
2005 wfDebug( __FUNCTION__ . "( $dir )\n" );
2006 // taken from https://www.php.net/manual/en/function.rmdir.php#98622
2007 if ( is_dir( $dir ) ) {
2008 $objects = scandir( $dir );
2009 foreach ( $objects as $object ) {
2010 if ( $object != "." && $object != ".." ) {
2011 if ( filetype( $dir . '/' . $object ) == "dir" ) {
2012 wfRecursiveRemoveDir( $dir . '/' . $object );
2013 } else {
2014 unlink( $dir . '/' . $object );
2015 }
2016 }
2017 }
2018 reset( $objects );
2019 rmdir( $dir );
2020 }
2021 }
2022
2023 /**
2024 * @param int $nr The number to format
2025 * @param int $acc The number of digits after the decimal point, default 2
2026 * @param bool $round Whether or not to round the value, default true
2027 * @return string
2028 */
2029 function wfPercent( $nr, $acc = 2, $round = true ) {
2030 $ret = sprintf( "%.${acc}f", $nr );
2031 return $round ? round( $ret, $acc ) . '%' : "$ret%";
2032 }
2033
2034 /**
2035 * Safety wrapper around ini_get() for boolean settings.
2036 * The values returned from ini_get() are pre-normalized for settings
2037 * set via php.ini or php_flag/php_admin_flag... but *not*
2038 * for those set via php_value/php_admin_value.
2039 *
2040 * It's fairly common for people to use php_value instead of php_flag,
2041 * which can leave you with an 'off' setting giving a false positive
2042 * for code that just takes the ini_get() return value as a boolean.
2043 *
2044 * To make things extra interesting, setting via php_value accepts
2045 * "true" and "yes" as true, but php.ini and php_flag consider them false. :)
2046 * Unrecognized values go false... again opposite PHP's own coercion
2047 * from string to bool.
2048 *
2049 * Luckily, 'properly' set settings will always come back as '0' or '1',
2050 * so we only have to worry about them and the 'improper' settings.
2051 *
2052 * I frickin' hate PHP... :P
2053 *
2054 * @param string $setting
2055 * @return bool
2056 */
2057 function wfIniGetBool( $setting ) {
2058 return wfStringToBool( ini_get( $setting ) );
2059 }
2060
2061 /**
2062 * Convert string value to boolean, when the following are interpreted as true:
2063 * - on
2064 * - true
2065 * - yes
2066 * - Any number, except 0
2067 * All other strings are interpreted as false.
2068 *
2069 * @param string $val
2070 * @return bool
2071 * @since 1.31
2072 */
2073 function wfStringToBool( $val ) {
2074 $val = strtolower( $val );
2075 // 'on' and 'true' can't have whitespace around them, but '1' can.
2076 return $val == 'on'
2077 || $val == 'true'
2078 || $val == 'yes'
2079 || preg_match( "/^\s*[+-]?0*[1-9]/", $val ); // approx C atoi() function
2080 }
2081
2082 /**
2083 * Version of escapeshellarg() that works better on Windows.
2084 *
2085 * Originally, this fixed the incorrect use of single quotes on Windows
2086 * (https://bugs.php.net/bug.php?id=26285) and the locale problems on Linux in
2087 * PHP 5.2.6+ (bug backported to earlier distro releases of PHP).
2088 *
2089 * @param string|string[] ...$args strings to escape and glue together,
2090 * or a single array of strings parameter
2091 * @return string
2092 * @deprecated since 1.30 use MediaWiki\Shell\Shell::escape()
2093 */
2094 function wfEscapeShellArg( ...$args ) {
2095 return Shell::escape( ...$args );
2096 }
2097
2098 /**
2099 * Execute a shell command, with time and memory limits mirrored from the PHP
2100 * configuration if supported.
2101 *
2102 * @param string|string[] $cmd If string, a properly shell-escaped command line,
2103 * or an array of unescaped arguments, in which case each value will be escaped
2104 * Example: [ 'convert', '-font', 'font name' ] would produce "'convert' '-font' 'font name'"
2105 * @param null|mixed &$retval Optional, will receive the program's exit code.
2106 * (non-zero is usually failure). If there is an error from
2107 * read, select, or proc_open(), this will be set to -1.
2108 * @param array $environ Optional environment variables which should be
2109 * added to the executed command environment.
2110 * @param array $limits Optional array with limits(filesize, memory, time, walltime)
2111 * this overwrites the global wgMaxShell* limits.
2112 * @param array $options Array of options:
2113 * - duplicateStderr: Set this to true to duplicate stderr to stdout,
2114 * including errors from limit.sh
2115 * - profileMethod: By default this function will profile based on the calling
2116 * method. Set this to a string for an alternative method to profile from
2117 *
2118 * @return string Collected stdout as a string
2119 * @deprecated since 1.30 use class MediaWiki\Shell\Shell
2120 */
2121 function wfShellExec( $cmd, &$retval = null, $environ = [],
2122 $limits = [], $options = []
2123 ) {
2124 if ( Shell::isDisabled() ) {
2125 $retval = 1;
2126 // Backwards compatibility be upon us...
2127 return 'Unable to run external programs, proc_open() is disabled.';
2128 }
2129
2130 if ( is_array( $cmd ) ) {
2131 $cmd = Shell::escape( $cmd );
2132 }
2133
2134 $includeStderr = isset( $options['duplicateStderr'] ) && $options['duplicateStderr'];
2135 $profileMethod = $options['profileMethod'] ?? wfGetCaller();
2136
2137 try {
2138 $result = Shell::command( [] )
2139 ->unsafeParams( (array)$cmd )
2140 ->environment( $environ )
2141 ->limits( $limits )
2142 ->includeStderr( $includeStderr )
2143 ->profileMethod( $profileMethod )
2144 // For b/c
2145 ->restrict( Shell::RESTRICT_NONE )
2146 ->execute();
2147 } catch ( ProcOpenError $ex ) {
2148 $retval = -1;
2149 return '';
2150 }
2151
2152 $retval = $result->getExitCode();
2153
2154 return $result->getStdout();
2155 }
2156
2157 /**
2158 * Execute a shell command, returning both stdout and stderr. Convenience
2159 * function, as all the arguments to wfShellExec can become unwieldy.
2160 *
2161 * @note This also includes errors from limit.sh, e.g. if $wgMaxShellFileSize is exceeded.
2162 * @param string|string[] $cmd If string, a properly shell-escaped command line,
2163 * or an array of unescaped arguments, in which case each value will be escaped
2164 * Example: [ 'convert', '-font', 'font name' ] would produce "'convert' '-font' 'font name'"
2165 * @param null|mixed &$retval Optional, will receive the program's exit code.
2166 * (non-zero is usually failure)
2167 * @param array $environ Optional environment variables which should be
2168 * added to the executed command environment.
2169 * @param array $limits Optional array with limits(filesize, memory, time, walltime)
2170 * this overwrites the global wgMaxShell* limits.
2171 * @return string Collected stdout and stderr as a string
2172 * @deprecated since 1.30 use class MediaWiki\Shell\Shell
2173 */
2174 function wfShellExecWithStderr( $cmd, &$retval = null, $environ = [], $limits = [] ) {
2175 return wfShellExec( $cmd, $retval, $environ, $limits,
2176 [ 'duplicateStderr' => true, 'profileMethod' => wfGetCaller() ] );
2177 }
2178
2179 /**
2180 * Generate a shell-escaped command line string to run a MediaWiki cli script.
2181 * Note that $parameters should be a flat array and an option with an argument
2182 * should consist of two consecutive items in the array (do not use "--option value").
2183 *
2184 * @deprecated since 1.31, use Shell::makeScriptCommand()
2185 *
2186 * @param string $script MediaWiki cli script path
2187 * @param array $parameters Arguments and options to the script
2188 * @param array $options Associative array of options:
2189 * 'php': The path to the php executable
2190 * 'wrapper': Path to a PHP wrapper to handle the maintenance script
2191 * @return string
2192 */
2193 function wfShellWikiCmd( $script, array $parameters = [], array $options = [] ) {
2194 global $wgPhpCli;
2195 // Give site config file a chance to run the script in a wrapper.
2196 // The caller may likely want to call wfBasename() on $script.
2197 Hooks::run( 'wfShellWikiCmd', [ &$script, &$parameters, &$options ] );
2198 $cmd = [ $options['php'] ?? $wgPhpCli ];
2199 if ( isset( $options['wrapper'] ) ) {
2200 $cmd[] = $options['wrapper'];
2201 }
2202 $cmd[] = $script;
2203 // Escape each parameter for shell
2204 return Shell::escape( array_merge( $cmd, $parameters ) );
2205 }
2206
2207 /**
2208 * wfMerge attempts to merge differences between three texts.
2209 * Returns true for a clean merge and false for failure or a conflict.
2210 *
2211 * @param string $old
2212 * @param string $mine
2213 * @param string $yours
2214 * @param string &$result
2215 * @param string|null &$mergeAttemptResult
2216 * @return bool
2217 */
2218 function wfMerge( $old, $mine, $yours, &$result, &$mergeAttemptResult = null ) {
2219 global $wgDiff3;
2220
2221 # This check may also protect against code injection in
2222 # case of broken installations.
2223 AtEase::suppressWarnings();
2224 $haveDiff3 = $wgDiff3 && file_exists( $wgDiff3 );
2225 AtEase::restoreWarnings();
2226
2227 if ( !$haveDiff3 ) {
2228 wfDebug( "diff3 not found\n" );
2229 return false;
2230 }
2231
2232 # Make temporary files
2233 $td = wfTempDir();
2234 $oldtextFile = fopen( $oldtextName = tempnam( $td, 'merge-old-' ), 'w' );
2235 $mytextFile = fopen( $mytextName = tempnam( $td, 'merge-mine-' ), 'w' );
2236 $yourtextFile = fopen( $yourtextName = tempnam( $td, 'merge-your-' ), 'w' );
2237
2238 # NOTE: diff3 issues a warning to stderr if any of the files does not end with
2239 # a newline character. To avoid this, we normalize the trailing whitespace before
2240 # creating the diff.
2241
2242 fwrite( $oldtextFile, rtrim( $old ) . "\n" );
2243 fclose( $oldtextFile );
2244 fwrite( $mytextFile, rtrim( $mine ) . "\n" );
2245 fclose( $mytextFile );
2246 fwrite( $yourtextFile, rtrim( $yours ) . "\n" );
2247 fclose( $yourtextFile );
2248
2249 # Check for a conflict
2250 $cmd = Shell::escape( $wgDiff3, '-a', '--overlap-only', $mytextName,
2251 $oldtextName, $yourtextName );
2252 $handle = popen( $cmd, 'r' );
2253
2254 $mergeAttemptResult = '';
2255 do {
2256 $data = fread( $handle, 8192 );
2257 if ( strlen( $data ) == 0 ) {
2258 break;
2259 }
2260 $mergeAttemptResult .= $data;
2261 } while ( true );
2262 pclose( $handle );
2263
2264 $conflict = $mergeAttemptResult !== '';
2265
2266 # Merge differences
2267 $cmd = Shell::escape( $wgDiff3, '-a', '-e', '--merge', $mytextName,
2268 $oldtextName, $yourtextName );
2269 $handle = popen( $cmd, 'r' );
2270 $result = '';
2271 do {
2272 $data = fread( $handle, 8192 );
2273 if ( strlen( $data ) == 0 ) {
2274 break;
2275 }
2276 $result .= $data;
2277 } while ( true );
2278 pclose( $handle );
2279 unlink( $mytextName );
2280 unlink( $oldtextName );
2281 unlink( $yourtextName );
2282
2283 if ( $result === '' && $old !== '' && !$conflict ) {
2284 wfDebug( "Unexpected null result from diff3. Command: $cmd\n" );
2285 $conflict = true;
2286 }
2287 return !$conflict;
2288 }
2289
2290 /**
2291 * Returns unified plain-text diff of two texts.
2292 * "Useful" for machine processing of diffs.
2293 *
2294 * @deprecated since 1.25, use DiffEngine/UnifiedDiffFormatter directly
2295 *
2296 * @param string $before The text before the changes.
2297 * @param string $after The text after the changes.
2298 * @param string $params Command-line options for the diff command.
2299 * @return string Unified diff of $before and $after
2300 */
2301 function wfDiff( $before, $after, $params = '-u' ) {
2302 if ( $before == $after ) {
2303 return '';
2304 }
2305
2306 global $wgDiff;
2307 AtEase::suppressWarnings();
2308 $haveDiff = $wgDiff && file_exists( $wgDiff );
2309 AtEase::restoreWarnings();
2310
2311 # This check may also protect against code injection in
2312 # case of broken installations.
2313 if ( !$haveDiff ) {
2314 wfDebug( "diff executable not found\n" );
2315 $diffs = new Diff( explode( "\n", $before ), explode( "\n", $after ) );
2316 $format = new UnifiedDiffFormatter();
2317 return $format->format( $diffs );
2318 }
2319
2320 # Make temporary files
2321 $td = wfTempDir();
2322 $oldtextFile = fopen( $oldtextName = tempnam( $td, 'merge-old-' ), 'w' );
2323 $newtextFile = fopen( $newtextName = tempnam( $td, 'merge-your-' ), 'w' );
2324
2325 fwrite( $oldtextFile, $before );
2326 fclose( $oldtextFile );
2327 fwrite( $newtextFile, $after );
2328 fclose( $newtextFile );
2329
2330 // Get the diff of the two files
2331 $cmd = "$wgDiff " . $params . ' ' . Shell::escape( $oldtextName, $newtextName );
2332
2333 $h = popen( $cmd, 'r' );
2334 if ( !$h ) {
2335 unlink( $oldtextName );
2336 unlink( $newtextName );
2337 throw new Exception( __METHOD__ . '(): popen() failed' );
2338 }
2339
2340 $diff = '';
2341
2342 do {
2343 $data = fread( $h, 8192 );
2344 if ( strlen( $data ) == 0 ) {
2345 break;
2346 }
2347 $diff .= $data;
2348 } while ( true );
2349
2350 // Clean up
2351 pclose( $h );
2352 unlink( $oldtextName );
2353 unlink( $newtextName );
2354
2355 // Kill the --- and +++ lines. They're not useful.
2356 $diff_lines = explode( "\n", $diff );
2357 if ( isset( $diff_lines[0] ) && strpos( $diff_lines[0], '---' ) === 0 ) {
2358 unset( $diff_lines[0] );
2359 }
2360 if ( isset( $diff_lines[1] ) && strpos( $diff_lines[1], '+++' ) === 0 ) {
2361 unset( $diff_lines[1] );
2362 }
2363
2364 $diff = implode( "\n", $diff_lines );
2365
2366 return $diff;
2367 }
2368
2369 /**
2370 * Return the final portion of a pathname.
2371 * Reimplemented because PHP5's "basename()" is buggy with multibyte text.
2372 * https://bugs.php.net/bug.php?id=33898
2373 *
2374 * PHP's basename() only considers '\' a pathchar on Windows and Netware.
2375 * We'll consider it so always, as we don't want '\s' in our Unix paths either.
2376 *
2377 * @param string $path
2378 * @param string $suffix String to remove if present
2379 * @return string
2380 */
2381 function wfBaseName( $path, $suffix = '' ) {
2382 if ( $suffix == '' ) {
2383 $encSuffix = '';
2384 } else {
2385 $encSuffix = '(?:' . preg_quote( $suffix, '#' ) . ')?';
2386 }
2387
2388 $matches = [];
2389 if ( preg_match( "#([^/\\\\]*?){$encSuffix}[/\\\\]*$#", $path, $matches ) ) {
2390 return $matches[1];
2391 } else {
2392 return '';
2393 }
2394 }
2395
2396 /**
2397 * Generate a relative path name to the given file.
2398 * May explode on non-matching case-insensitive paths,
2399 * funky symlinks, etc.
2400 *
2401 * @param string $path Absolute destination path including target filename
2402 * @param string $from Absolute source path, directory only
2403 * @return string
2404 */
2405 function wfRelativePath( $path, $from ) {
2406 // Normalize mixed input on Windows...
2407 $path = str_replace( '/', DIRECTORY_SEPARATOR, $path );
2408 $from = str_replace( '/', DIRECTORY_SEPARATOR, $from );
2409
2410 // Trim trailing slashes -- fix for drive root
2411 $path = rtrim( $path, DIRECTORY_SEPARATOR );
2412 $from = rtrim( $from, DIRECTORY_SEPARATOR );
2413
2414 $pieces = explode( DIRECTORY_SEPARATOR, dirname( $path ) );
2415 $against = explode( DIRECTORY_SEPARATOR, $from );
2416
2417 if ( $pieces[0] !== $against[0] ) {
2418 // Non-matching Windows drive letters?
2419 // Return a full path.
2420 return $path;
2421 }
2422
2423 // Trim off common prefix
2424 while ( count( $pieces ) && count( $against )
2425 && $pieces[0] == $against[0] ) {
2426 array_shift( $pieces );
2427 array_shift( $against );
2428 }
2429
2430 // relative dots to bump us to the parent
2431 while ( count( $against ) ) {
2432 array_unshift( $pieces, '..' );
2433 array_shift( $against );
2434 }
2435
2436 array_push( $pieces, wfBaseName( $path ) );
2437
2438 return implode( DIRECTORY_SEPARATOR, $pieces );
2439 }
2440
2441 /**
2442 * Initialise php session
2443 *
2444 * @deprecated since 1.27, use MediaWiki\Session\SessionManager instead.
2445 * Generally, "using" SessionManager will be calling ->getSessionById() or
2446 * ::getGlobalSession() (depending on whether you were passing $sessionId
2447 * here), then calling $session->persist().
2448 * @param bool|string $sessionId
2449 */
2450 function wfSetupSession( $sessionId = false ) {
2451 wfDeprecated( __FUNCTION__, '1.27' );
2452
2453 if ( $sessionId ) {
2454 session_id( $sessionId );
2455 }
2456
2457 $session = SessionManager::getGlobalSession();
2458 $session->persist();
2459
2460 if ( session_id() !== $session->getId() ) {
2461 session_id( $session->getId() );
2462 }
2463 AtEase::quietCall( 'session_start' );
2464 }
2465
2466 /**
2467 * Get an object from the precompiled serialized directory
2468 *
2469 * @param string $name
2470 * @return mixed The variable on success, false on failure
2471 */
2472 function wfGetPrecompiledData( $name ) {
2473 global $IP;
2474
2475 $file = "$IP/serialized/$name";
2476 if ( file_exists( $file ) ) {
2477 $blob = file_get_contents( $file );
2478 if ( $blob ) {
2479 return unserialize( $blob );
2480 }
2481 }
2482 return false;
2483 }
2484
2485 /**
2486 * Make a cache key for the local wiki.
2487 *
2488 * @deprecated since 1.30 Call makeKey on a BagOStuff instance
2489 * @param string ...$args
2490 * @return string
2491 */
2492 function wfMemcKey( ...$args ) {
2493 return ObjectCache::getLocalClusterInstance()->makeKey( ...$args );
2494 }
2495
2496 /**
2497 * Make a cache key for a foreign DB.
2498 *
2499 * Must match what wfMemcKey() would produce in context of the foreign wiki.
2500 *
2501 * @param string $db
2502 * @param string $prefix
2503 * @param string ...$args
2504 * @return string
2505 */
2506 function wfForeignMemcKey( $db, $prefix, ...$args ) {
2507 $keyspace = $prefix ? "$db-$prefix" : $db;
2508 return ObjectCache::getLocalClusterInstance()->makeKeyInternal( $keyspace, $args );
2509 }
2510
2511 /**
2512 * Make a cache key with database-agnostic prefix.
2513 *
2514 * Doesn't have a wiki-specific namespace. Uses a generic 'global' prefix
2515 * instead. Must have a prefix as otherwise keys that use a database name
2516 * in the first segment will clash with wfMemcKey/wfForeignMemcKey.
2517 *
2518 * @deprecated since 1.30 Call makeGlobalKey on a BagOStuff instance
2519 * @since 1.26
2520 * @param string ...$args
2521 * @return string
2522 */
2523 function wfGlobalCacheKey( ...$args ) {
2524 return ObjectCache::getLocalClusterInstance()->makeGlobalKey( ...$args );
2525 }
2526
2527 /**
2528 * Get an ASCII string identifying this wiki
2529 * This is used as a prefix in memcached keys
2530 *
2531 * @return string
2532 */
2533 function wfWikiID() {
2534 global $wgDBprefix, $wgDBname;
2535 if ( $wgDBprefix ) {
2536 return "$wgDBname-$wgDBprefix";
2537 } else {
2538 return $wgDBname;
2539 }
2540 }
2541
2542 /**
2543 * Get a Database object.
2544 *
2545 * @param int $db Index of the connection to get. May be DB_MASTER for the
2546 * master (for write queries), DB_REPLICA for potentially lagged read
2547 * queries, or an integer >= 0 for a particular server.
2548 *
2549 * @param string|string[] $groups Query groups. An array of group names that this query
2550 * belongs to. May contain a single string if the query is only
2551 * in one group.
2552 *
2553 * @param string|bool $wiki The wiki ID, or false for the current wiki
2554 *
2555 * Note: multiple calls to wfGetDB(DB_REPLICA) during the course of one request
2556 * will always return the same object, unless the underlying connection or load
2557 * balancer is manually destroyed.
2558 *
2559 * Note 2: use $this->getDB() in maintenance scripts that may be invoked by
2560 * updater to ensure that a proper database is being updated.
2561 *
2562 * @todo Replace calls to wfGetDB with calls to LoadBalancer::getConnection()
2563 * on an injected instance of LoadBalancer.
2564 *
2565 * @return \Wikimedia\Rdbms\Database
2566 */
2567 function wfGetDB( $db, $groups = [], $wiki = false ) {
2568 return wfGetLB( $wiki )->getConnection( $db, $groups, $wiki );
2569 }
2570
2571 /**
2572 * Get a load balancer object.
2573 *
2574 * @deprecated since 1.27, use MediaWikiServices::getInstance()->getDBLoadBalancer()
2575 * or MediaWikiServices::getInstance()->getDBLoadBalancerFactory() instead.
2576 *
2577 * @param string|bool $wiki Wiki ID, or false for the current wiki
2578 * @return \Wikimedia\Rdbms\LoadBalancer
2579 */
2580 function wfGetLB( $wiki = false ) {
2581 if ( $wiki === false ) {
2582 return MediaWikiServices::getInstance()->getDBLoadBalancer();
2583 } else {
2584 $factory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
2585 return $factory->getMainLB( $wiki );
2586 }
2587 }
2588
2589 /**
2590 * Find a file.
2591 * @deprecated since 1.34, use MediaWikiServices
2592 * @param string|LinkTarget $title String or LinkTarget object
2593 * @param array $options Associative array of options (see RepoGroup::findFile)
2594 * @return File|bool File, or false if the file does not exist
2595 */
2596 function wfFindFile( $title, $options = [] ) {
2597 return MediaWikiServices::getInstance()->getRepoGroup()->findFile( $title, $options );
2598 }
2599
2600 /**
2601 * Get an object referring to a locally registered file.
2602 * Returns a valid placeholder object if the file does not exist.
2603 *
2604 * @deprecated since 1.34, use MediaWikiServices
2605 * @param Title|string $title
2606 * @return LocalFile|null A File, or null if passed an invalid Title
2607 */
2608 function wfLocalFile( $title ) {
2609 return MediaWikiServices::getInstance()->getRepoGroup()->getLocalRepo()->newFile( $title );
2610 }
2611
2612 /**
2613 * Should low-performance queries be disabled?
2614 *
2615 * @return bool
2616 * @codeCoverageIgnore
2617 */
2618 function wfQueriesMustScale() {
2619 global $wgMiserMode;
2620 return $wgMiserMode
2621 || ( SiteStats::pages() > 100000
2622 && SiteStats::edits() > 1000000
2623 && SiteStats::users() > 10000 );
2624 }
2625
2626 /**
2627 * Get the path to a specified script file, respecting file
2628 * extensions; this is a wrapper around $wgScriptPath etc.
2629 * except for 'index' and 'load' which use $wgScript/$wgLoadScript
2630 *
2631 * @param string $script Script filename, sans extension
2632 * @return string
2633 */
2634 function wfScript( $script = 'index' ) {
2635 global $wgScriptPath, $wgScript, $wgLoadScript;
2636 if ( $script === 'index' ) {
2637 return $wgScript;
2638 } elseif ( $script === 'load' ) {
2639 return $wgLoadScript;
2640 } else {
2641 return "{$wgScriptPath}/{$script}.php";
2642 }
2643 }
2644
2645 /**
2646 * Get the script URL.
2647 *
2648 * @return string Script URL
2649 */
2650 function wfGetScriptUrl() {
2651 if ( isset( $_SERVER['SCRIPT_NAME'] ) ) {
2652 /* as it was called, minus the query string.
2653 *
2654 * Some sites use Apache rewrite rules to handle subdomains,
2655 * and have PHP set up in a weird way that causes PHP_SELF
2656 * to contain the rewritten URL instead of the one that the
2657 * outside world sees.
2658 *
2659 * If in this mode, use SCRIPT_URL instead, which mod_rewrite
2660 * provides containing the "before" URL.
2661 */
2662 return $_SERVER['SCRIPT_NAME'];
2663 } else {
2664 return $_SERVER['URL'];
2665 }
2666 }
2667
2668 /**
2669 * Convenience function converts boolean values into "true"
2670 * or "false" (string) values
2671 *
2672 * @param bool $value
2673 * @return string
2674 */
2675 function wfBoolToStr( $value ) {
2676 return $value ? 'true' : 'false';
2677 }
2678
2679 /**
2680 * Get a platform-independent path to the null file, e.g. /dev/null
2681 *
2682 * @return string
2683 */
2684 function wfGetNull() {
2685 return wfIsWindows() ? 'NUL' : '/dev/null';
2686 }
2687
2688 /**
2689 * Waits for the replica DBs to catch up to the master position
2690 *
2691 * Use this when updating very large numbers of rows, as in maintenance scripts,
2692 * to avoid causing too much lag. Of course, this is a no-op if there are no replica DBs.
2693 *
2694 * By default this waits on the main DB cluster of the current wiki.
2695 * If $cluster is set to "*" it will wait on all DB clusters, including
2696 * external ones. If the lag being waiting on is caused by the code that
2697 * does this check, it makes since to use $ifWritesSince, particularly if
2698 * cluster is "*", to avoid excess overhead.
2699 *
2700 * Never call this function after a big DB write that is still in a transaction.
2701 * This only makes sense after the possible lag inducing changes were committed.
2702 *
2703 * @param float|null $ifWritesSince Only wait if writes were done since this UNIX timestamp
2704 * @param string|bool $wiki Wiki identifier accepted by wfGetLB
2705 * @param string|bool $cluster Cluster name accepted by LBFactory. Default: false.
2706 * @param int|null $timeout Max wait time. Default: 60 seconds (cli), 1 second (web)
2707 * @return bool Success (able to connect and no timeouts reached)
2708 * @deprecated since 1.27 Use LBFactory::waitForReplication
2709 */
2710 function wfWaitForSlaves(
2711 $ifWritesSince = null, $wiki = false, $cluster = false, $timeout = null
2712 ) {
2713 $lbFactory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
2714
2715 if ( $cluster === '*' ) {
2716 $cluster = false;
2717 $domain = false;
2718 } elseif ( $wiki === false ) {
2719 $domain = $lbFactory->getLocalDomainID();
2720 } else {
2721 $domain = $wiki;
2722 }
2723
2724 $opts = [
2725 'domain' => $domain,
2726 'cluster' => $cluster,
2727 // B/C: first argument used to be "max seconds of lag"; ignore such values
2728 'ifWritesSince' => ( $ifWritesSince > 1e9 ) ? $ifWritesSince : null
2729 ];
2730 if ( $timeout !== null ) {
2731 $opts['timeout'] = $timeout;
2732 }
2733
2734 return $lbFactory->waitForReplication( $opts );
2735 }
2736
2737 /**
2738 * Replace all invalid characters with '-'.
2739 * Additional characters can be defined in $wgIllegalFileChars (see T22489).
2740 * By default, $wgIllegalFileChars includes ':', '/', '\'.
2741 *
2742 * @param string $name Filename to process
2743 * @return string
2744 */
2745 function wfStripIllegalFilenameChars( $name ) {
2746 global $wgIllegalFileChars;
2747 $illegalFileChars = $wgIllegalFileChars ? "|[" . $wgIllegalFileChars . "]" : '';
2748 $name = preg_replace(
2749 "/[^" . Title::legalChars() . "]" . $illegalFileChars . "/",
2750 '-',
2751 $name
2752 );
2753 // $wgIllegalFileChars may not include '/' and '\', so we still need to do this
2754 $name = wfBaseName( $name );
2755 return $name;
2756 }
2757
2758 /**
2759 * Set PHP's memory limit to the larger of php.ini or $wgMemoryLimit
2760 *
2761 * @return int Resulting value of the memory limit.
2762 */
2763 function wfMemoryLimit() {
2764 global $wgMemoryLimit;
2765 $memlimit = wfShorthandToInteger( ini_get( 'memory_limit' ) );
2766 if ( $memlimit != -1 ) {
2767 $conflimit = wfShorthandToInteger( $wgMemoryLimit );
2768 if ( $conflimit == -1 ) {
2769 wfDebug( "Removing PHP's memory limit\n" );
2770 Wikimedia\suppressWarnings();
2771 ini_set( 'memory_limit', $conflimit );
2772 Wikimedia\restoreWarnings();
2773 return $conflimit;
2774 } elseif ( $conflimit > $memlimit ) {
2775 wfDebug( "Raising PHP's memory limit to $conflimit bytes\n" );
2776 Wikimedia\suppressWarnings();
2777 ini_set( 'memory_limit', $conflimit );
2778 Wikimedia\restoreWarnings();
2779 return $conflimit;
2780 }
2781 }
2782 return $memlimit;
2783 }
2784
2785 /**
2786 * Set PHP's time limit to the larger of php.ini or $wgTransactionalTimeLimit
2787 *
2788 * @return int Prior time limit
2789 * @since 1.26
2790 */
2791 function wfTransactionalTimeLimit() {
2792 global $wgTransactionalTimeLimit;
2793
2794 $timeLimit = ini_get( 'max_execution_time' );
2795 // Note that CLI scripts use 0
2796 if ( $timeLimit > 0 && $wgTransactionalTimeLimit > $timeLimit ) {
2797 set_time_limit( $wgTransactionalTimeLimit );
2798 }
2799
2800 ignore_user_abort( true ); // ignore client disconnects
2801
2802 return $timeLimit;
2803 }
2804
2805 /**
2806 * Converts shorthand byte notation to integer form
2807 *
2808 * @param string $string
2809 * @param int $default Returned if $string is empty
2810 * @return int
2811 */
2812 function wfShorthandToInteger( $string = '', $default = -1 ) {
2813 $string = trim( $string );
2814 if ( $string === '' ) {
2815 return $default;
2816 }
2817 $last = $string[strlen( $string ) - 1];
2818 $val = intval( $string );
2819 switch ( $last ) {
2820 case 'g':
2821 case 'G':
2822 $val *= 1024;
2823 // break intentionally missing
2824 case 'm':
2825 case 'M':
2826 $val *= 1024;
2827 // break intentionally missing
2828 case 'k':
2829 case 'K':
2830 $val *= 1024;
2831 }
2832
2833 return $val;
2834 }
2835
2836 /**
2837 * Get a specific cache object.
2838 *
2839 * @deprecated since 1.32, use ObjectCache::getInstance() instead
2840 * @param int|string $cacheType A CACHE_* constants, or other key in $wgObjectCaches
2841 * @return BagOStuff
2842 */
2843 function wfGetCache( $cacheType ) {
2844 return ObjectCache::getInstance( $cacheType );
2845 }
2846
2847 /**
2848 * Get the main cache object
2849 *
2850 * @deprecated since 1.32, use ObjectCache::getLocalClusterInstance() instead
2851 * @return BagOStuff
2852 */
2853 function wfGetMainCache() {
2854 return ObjectCache::getLocalClusterInstance();
2855 }
2856
2857 /**
2858 * Get the cache object used by the message cache
2859 *
2860 * @return BagOStuff
2861 */
2862 function wfGetMessageCacheStorage() {
2863 global $wgMessageCacheType;
2864 return ObjectCache::getInstance( $wgMessageCacheType );
2865 }
2866
2867 /**
2868 * Wrapper around php's unpack.
2869 *
2870 * @param string $format The format string (See php's docs)
2871 * @param string $data A binary string of binary data
2872 * @param int|bool $length The minimum length of $data or false. This is to
2873 * prevent reading beyond the end of $data. false to disable the check.
2874 *
2875 * Also be careful when using this function to read unsigned 32 bit integer
2876 * because php might make it negative.
2877 *
2878 * @throws MWException If $data not long enough, or if unpack fails
2879 * @return array Associative array of the extracted data
2880 */
2881 function wfUnpack( $format, $data, $length = false ) {
2882 if ( $length !== false ) {
2883 $realLen = strlen( $data );
2884 if ( $realLen < $length ) {
2885 throw new MWException( "Tried to use wfUnpack on a "
2886 . "string of length $realLen, but needed one "
2887 . "of at least length $length."
2888 );
2889 }
2890 }
2891
2892 Wikimedia\suppressWarnings();
2893 $result = unpack( $format, $data );
2894 Wikimedia\restoreWarnings();
2895
2896 if ( $result === false ) {
2897 // If it cannot extract the packed data.
2898 throw new MWException( "unpack could not unpack binary data" );
2899 }
2900 return $result;
2901 }
2902
2903 /**
2904 * Determine if an image exists on the 'bad image list'.
2905 *
2906 * The format of MediaWiki:Bad_image_list is as follows:
2907 * * Only list items (lines starting with "*") are considered
2908 * * The first link on a line must be a link to a bad image
2909 * * Any subsequent links on the same line are considered to be exceptions,
2910 * i.e. articles where the image may occur inline.
2911 *
2912 * @param string $name The image name to check
2913 * @param Title|bool $contextTitle The page on which the image occurs, if known
2914 * @param string|null $blacklist Wikitext of a file blacklist
2915 * @return bool
2916 */
2917 function wfIsBadImage( $name, $contextTitle = false, $blacklist = null ) {
2918 # Handle redirects; callers almost always hit wfFindFile() anyway,
2919 # so just use that method because it has a fast process cache.
2920 $file = MediaWikiServices::getInstance()->getRepoGroup()->findFile( $name ); // get the final name
2921 $name = $file ? $file->getTitle()->getDBkey() : $name;
2922
2923 # Run the extension hook
2924 $bad = false;
2925 if ( !Hooks::run( 'BadImage', [ $name, &$bad ] ) ) {
2926 return (bool)$bad;
2927 }
2928
2929 $cache = ObjectCache::getLocalServerInstance( 'hash' );
2930 $key = $cache->makeKey(
2931 'bad-image-list', ( $blacklist === null ) ? 'default' : md5( $blacklist )
2932 );
2933 $badImages = $cache->get( $key );
2934
2935 if ( $badImages === false ) { // cache miss
2936 if ( $blacklist === null ) {
2937 $blacklist = wfMessage( 'bad_image_list' )->inContentLanguage()->plain(); // site list
2938 }
2939 # Build the list now
2940 $badImages = [];
2941 $lines = explode( "\n", $blacklist );
2942 foreach ( $lines as $line ) {
2943 # List items only
2944 if ( substr( $line, 0, 1 ) !== '*' ) {
2945 continue;
2946 }
2947
2948 # Find all links
2949 $m = [];
2950 if ( !preg_match_all( '/\[\[:?(.*?)\]\]/', $line, $m ) ) {
2951 continue;
2952 }
2953
2954 $exceptions = [];
2955 $imageDBkey = false;
2956 foreach ( $m[1] as $i => $titleText ) {
2957 $title = Title::newFromText( $titleText );
2958 if ( !is_null( $title ) ) {
2959 if ( $i == 0 ) {
2960 $imageDBkey = $title->getDBkey();
2961 } else {
2962 $exceptions[$title->getPrefixedDBkey()] = true;
2963 }
2964 }
2965 }
2966
2967 if ( $imageDBkey !== false ) {
2968 $badImages[$imageDBkey] = $exceptions;
2969 }
2970 }
2971 $cache->set( $key, $badImages, 60 );
2972 }
2973
2974 $contextKey = $contextTitle ? $contextTitle->getPrefixedDBkey() : false;
2975 $bad = isset( $badImages[$name] ) && !isset( $badImages[$name][$contextKey] );
2976
2977 return $bad;
2978 }
2979
2980 /**
2981 * Determine whether the client at a given source IP is likely to be able to
2982 * access the wiki via HTTPS.
2983 *
2984 * @param string $ip The IPv4/6 address in the normal human-readable form
2985 * @return bool
2986 */
2987 function wfCanIPUseHTTPS( $ip ) {
2988 $canDo = true;
2989 Hooks::run( 'CanIPUseHTTPS', [ $ip, &$canDo ] );
2990 return (bool)$canDo;
2991 }
2992
2993 /**
2994 * Determine input string is represents as infinity
2995 *
2996 * @param string $str The string to determine
2997 * @return bool
2998 * @since 1.25
2999 */
3000 function wfIsInfinity( $str ) {
3001 // These are hardcoded elsewhere in MediaWiki (e.g. mediawiki.special.block.js).
3002 $infinityValues = [ 'infinite', 'indefinite', 'infinity', 'never' ];
3003 return in_array( $str, $infinityValues );
3004 }
3005
3006 /**
3007 * Returns true if these thumbnail parameters match one that MediaWiki
3008 * requests from file description pages and/or parser output.
3009 *
3010 * $params is considered non-standard if they involve a non-standard
3011 * width or any non-default parameters aside from width and page number.
3012 * The number of possible files with standard parameters is far less than
3013 * that of all combinations; rate-limiting for them can thus be more generious.
3014 *
3015 * @param File $file
3016 * @param array $params
3017 * @return bool
3018 * @since 1.24 Moved from thumb.php to GlobalFunctions in 1.25
3019 */
3020 function wfThumbIsStandard( File $file, array $params ) {
3021 global $wgThumbLimits, $wgImageLimits, $wgResponsiveImages;
3022
3023 $multipliers = [ 1 ];
3024 if ( $wgResponsiveImages ) {
3025 // These available sizes are hardcoded currently elsewhere in MediaWiki.
3026 // @see Linker::processResponsiveImages
3027 $multipliers[] = 1.5;
3028 $multipliers[] = 2;
3029 }
3030
3031 $handler = $file->getHandler();
3032 if ( !$handler || !isset( $params['width'] ) ) {
3033 return false;
3034 }
3035
3036 $basicParams = [];
3037 if ( isset( $params['page'] ) ) {
3038 $basicParams['page'] = $params['page'];
3039 }
3040
3041 $thumbLimits = [];
3042 $imageLimits = [];
3043 // Expand limits to account for multipliers
3044 foreach ( $multipliers as $multiplier ) {
3045 $thumbLimits = array_merge( $thumbLimits, array_map(
3046 function ( $width ) use ( $multiplier ) {
3047 return round( $width * $multiplier );
3048 }, $wgThumbLimits )
3049 );
3050 $imageLimits = array_merge( $imageLimits, array_map(
3051 function ( $pair ) use ( $multiplier ) {
3052 return [
3053 round( $pair[0] * $multiplier ),
3054 round( $pair[1] * $multiplier ),
3055 ];
3056 }, $wgImageLimits )
3057 );
3058 }
3059
3060 // Check if the width matches one of $wgThumbLimits
3061 if ( in_array( $params['width'], $thumbLimits ) ) {
3062 $normalParams = $basicParams + [ 'width' => $params['width'] ];
3063 // Append any default values to the map (e.g. "lossy", "lossless", ...)
3064 $handler->normaliseParams( $file, $normalParams );
3065 } else {
3066 // If not, then check if the width matchs one of $wgImageLimits
3067 $match = false;
3068 foreach ( $imageLimits as $pair ) {
3069 $normalParams = $basicParams + [ 'width' => $pair[0], 'height' => $pair[1] ];
3070 // Decide whether the thumbnail should be scaled on width or height.
3071 // Also append any default values to the map (e.g. "lossy", "lossless", ...)
3072 $handler->normaliseParams( $file, $normalParams );
3073 // Check if this standard thumbnail size maps to the given width
3074 if ( $normalParams['width'] == $params['width'] ) {
3075 $match = true;
3076 break;
3077 }
3078 }
3079 if ( !$match ) {
3080 return false; // not standard for description pages
3081 }
3082 }
3083
3084 // Check that the given values for non-page, non-width, params are just defaults
3085 foreach ( $params as $key => $value ) {
3086 if ( !isset( $normalParams[$key] ) || $normalParams[$key] != $value ) {
3087 return false;
3088 }
3089 }
3090
3091 return true;
3092 }
3093
3094 /**
3095 * Merges two (possibly) 2 dimensional arrays into the target array ($baseArray).
3096 *
3097 * Values that exist in both values will be combined with += (all values of the array
3098 * of $newValues will be added to the values of the array of $baseArray, while values,
3099 * that exists in both, the value of $baseArray will be used).
3100 *
3101 * @param array $baseArray The array where you want to add the values of $newValues to
3102 * @param array $newValues An array with new values
3103 * @return array The combined array
3104 * @since 1.26
3105 */
3106 function wfArrayPlus2d( array $baseArray, array $newValues ) {
3107 // First merge items that are in both arrays
3108 foreach ( $baseArray as $name => &$groupVal ) {
3109 if ( isset( $newValues[$name] ) ) {
3110 $groupVal += $newValues[$name];
3111 }
3112 }
3113 // Now add items that didn't exist yet
3114 $baseArray += $newValues;
3115
3116 return $baseArray;
3117 }
3118
3119 /**
3120 * Get system resource usage of current request context.
3121 * Invokes the getrusage(2) system call, requesting RUSAGE_SELF if on PHP5
3122 * or RUSAGE_THREAD if on HHVM. Returns false if getrusage is not available.
3123 *
3124 * @since 1.24
3125 * @return array|bool Resource usage data or false if no data available.
3126 */
3127 function wfGetRusage() {
3128 if ( !function_exists( 'getrusage' ) ) {
3129 return false;
3130 } elseif ( defined( 'HHVM_VERSION' ) && PHP_OS === 'Linux' ) {
3131 return getrusage( 2 /* RUSAGE_THREAD */ );
3132 } else {
3133 return getrusage( 0 /* RUSAGE_SELF */ );
3134 }
3135 }