3 * Global functions used everywhere.
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.
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.
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
23 if ( !defined( 'MEDIAWIKI' ) ) {
24 die( "This file is part of MediaWiki, it is not a valid entry point" );
27 use Liuggio\StatsdClient\Sender\SocketSender
;
28 use MediaWiki\Logger\LoggerFactory
;
29 use MediaWiki\Session\SessionManager
;
31 // Hide compatibility functions from Doxygen
34 * Compatibility functions
36 * We support PHP 5.5.9 and up.
37 * Re-implementations of newer functions or functions in non-standard
38 * PHP extensions may be included here.
41 // hash_equals function only exists in PHP >= 5.6.0
42 // http://php.net/hash_equals
43 if ( !function_exists( 'hash_equals' ) ) {
45 * Check whether a user-provided string is equal to a fixed-length secret string
46 * without revealing bytes of the secret string through timing differences.
48 * The usual way to compare strings (PHP's === operator or the underlying memcmp()
49 * function in C) is to compare corresponding bytes and stop at the first difference,
50 * which would take longer for a partial match than for a complete mismatch. This
51 * is not secure when one of the strings (e.g. an HMAC or token) must remain secret
52 * and the other may come from an attacker. Statistical analysis of timing measurements
53 * over many requests may allow the attacker to guess the string's bytes one at a time
54 * (and check his guesses) even if the timing differences are extremely small.
56 * When making such a security-sensitive comparison, it is essential that the sequence
57 * in which instructions are executed and memory locations are accessed not depend on
58 * the secret string's value. HOWEVER, for simplicity, we do not attempt to minimize
59 * the inevitable leakage of the string's length. That is generally known anyway as
60 * a chararacteristic of the hash function used to compute the secret value.
62 * Longer explanation: http://www.emerose.com/timing-attacks-explained
65 * @param string $known_string Fixed-length secret string to compare against
66 * @param string $user_string User-provided string
67 * @return bool True if the strings are the same, false otherwise
69 function hash_equals( $known_string, $user_string ) {
70 // Strict type checking as in PHP's native implementation
71 if ( !is_string( $known_string ) ) {
72 trigger_error( 'hash_equals(): Expected known_string to be a string, ' .
73 gettype( $known_string ) . ' given', E_USER_WARNING
);
78 if ( !is_string( $user_string ) ) {
79 trigger_error( 'hash_equals(): Expected user_string to be a string, ' .
80 gettype( $user_string ) . ' given', E_USER_WARNING
);
85 $known_string_len = strlen( $known_string );
86 if ( $known_string_len !== strlen( $user_string ) ) {
91 for ( $i = 0; $i < $known_string_len; $i++
) {
92 $result |
= ord( $known_string[$i] ) ^
ord( $user_string[$i] );
95 return ( $result === 0 );
103 * This queues an extension to be loaded through
104 * the ExtensionRegistry system.
106 * @param string $ext Name of the extension to load
107 * @param string|null $path Absolute path of where to find the extension.json file
110 function wfLoadExtension( $ext, $path = null ) {
112 global $wgExtensionDirectory;
113 $path = "$wgExtensionDirectory/$ext/extension.json";
115 ExtensionRegistry
::getInstance()->queue( $path );
119 * Load multiple extensions at once
121 * Same as wfLoadExtension, but more efficient if you
122 * are loading multiple extensions.
124 * If you want to specify custom paths, you should interact with
125 * ExtensionRegistry directly.
127 * @see wfLoadExtension
128 * @param string[] $exts Array of extension names to load
131 function wfLoadExtensions( array $exts ) {
132 global $wgExtensionDirectory;
133 $registry = ExtensionRegistry
::getInstance();
134 foreach ( $exts as $ext ) {
135 $registry->queue( "$wgExtensionDirectory/$ext/extension.json" );
142 * @see wfLoadExtension
143 * @param string $skin Name of the extension to load
144 * @param string|null $path Absolute path of where to find the skin.json file
147 function wfLoadSkin( $skin, $path = null ) {
149 global $wgStyleDirectory;
150 $path = "$wgStyleDirectory/$skin/skin.json";
152 ExtensionRegistry
::getInstance()->queue( $path );
156 * Load multiple skins at once
158 * @see wfLoadExtensions
159 * @param string[] $skins Array of extension names to load
162 function wfLoadSkins( array $skins ) {
163 global $wgStyleDirectory;
164 $registry = ExtensionRegistry
::getInstance();
165 foreach ( $skins as $skin ) {
166 $registry->queue( "$wgStyleDirectory/$skin/skin.json" );
171 * Like array_diff( $a, $b ) except that it works with two-dimensional arrays.
176 function wfArrayDiff2( $a, $b ) {
177 return array_udiff( $a, $b, 'wfArrayDiff2_cmp' );
181 * @param array|string $a
182 * @param array|string $b
185 function wfArrayDiff2_cmp( $a, $b ) {
186 if ( is_string( $a ) && is_string( $b ) ) {
187 return strcmp( $a, $b );
188 } elseif ( count( $a ) !== count( $b ) ) {
189 return count( $a ) < count( $b ) ?
-1 : 1;
193 while ( ( list( , $valueA ) = each( $a ) ) && ( list( , $valueB ) = each( $b ) ) ) {
194 $cmp = strcmp( $valueA, $valueB );
204 * Appends to second array if $value differs from that in $default
206 * @param string|int $key
207 * @param mixed $value
208 * @param mixed $default
209 * @param array $changed Array to alter
210 * @throws MWException
212 function wfAppendToArrayIfNotDefault( $key, $value, $default, &$changed ) {
213 if ( is_null( $changed ) ) {
214 throw new MWException( 'GlobalFunctions::wfAppendToArrayIfNotDefault got null' );
216 if ( $default[$key] !== $value ) {
217 $changed[$key] = $value;
222 * Merge arrays in the style of getUserPermissionsErrors, with duplicate removal
224 * wfMergeErrorArrays(
237 * @param array $array1,...
240 function wfMergeErrorArrays( /*...*/ ) {
241 $args = func_get_args();
243 foreach ( $args as $errors ) {
244 foreach ( $errors as $params ) {
245 $originalParams = $params;
246 if ( $params[0] instanceof MessageSpecifier
) {
248 $params = array_merge( [ $msg->getKey() ], $msg->getParams() );
250 # @todo FIXME: Sometimes get nested arrays for $params,
251 # which leads to E_NOTICEs
252 $spec = implode( "\t", $params );
253 $out[$spec] = $originalParams;
256 return array_values( $out );
260 * Insert array into another array after the specified *KEY*
262 * @param array $array The array.
263 * @param array $insert The array to insert.
264 * @param mixed $after The key to insert after
267 function wfArrayInsertAfter( array $array, array $insert, $after ) {
268 // Find the offset of the element to insert after.
269 $keys = array_keys( $array );
270 $offsetByKey = array_flip( $keys );
272 $offset = $offsetByKey[$after];
274 // Insert at the specified offset
275 $before = array_slice( $array, 0, $offset +
1, true );
276 $after = array_slice( $array, $offset +
1, count( $array ) - $offset, true );
278 $output = $before +
$insert +
$after;
284 * Recursively converts the parameter (an object) to an array with the same data
286 * @param object|array $objOrArray
287 * @param bool $recursive
290 function wfObjectToArray( $objOrArray, $recursive = true ) {
292 if ( is_object( $objOrArray ) ) {
293 $objOrArray = get_object_vars( $objOrArray );
295 foreach ( $objOrArray as $key => $value ) {
296 if ( $recursive && ( is_object( $value ) ||
is_array( $value ) ) ) {
297 $value = wfObjectToArray( $value );
300 $array[$key] = $value;
307 * Get a random decimal value between 0 and 1, in a way
308 * not likely to give duplicate values for any realistic
309 * number of articles.
311 * @note This is designed for use in relation to Special:RandomPage
312 * and the page_random database field.
316 function wfRandom() {
317 // The maximum random value is "only" 2^31-1, so get two random
318 // values to reduce the chance of dupes
319 $max = mt_getrandmax() +
1;
320 $rand = number_format( ( mt_rand() * $max +
mt_rand() ) / $max / $max, 12, '.', '' );
325 * Get a random string containing a number of pseudo-random hex characters.
327 * @note This is not secure, if you are trying to generate some sort
328 * of token please use MWCryptRand instead.
330 * @param int $length The length of the string to generate
334 function wfRandomString( $length = 32 ) {
336 for ( $n = 0; $n < $length; $n +
= 7 ) {
337 $str .= sprintf( '%07x', mt_rand() & 0xfffffff );
339 return substr( $str, 0, $length );
343 * We want some things to be included as literal characters in our title URLs
344 * for prettiness, which urlencode encodes by default. According to RFC 1738,
345 * all of the following should be safe:
349 * RFC 1738 says ~ is unsafe, however RFC 3986 considers it an unreserved
350 * character which should not be encoded. More importantly, google chrome
351 * always converts %7E back to ~, and converting it in this function can
352 * cause a redirect loop (T105265).
354 * But + is not safe because it's used to indicate a space; &= are only safe in
355 * paths and not in queries (and we don't distinguish here); ' seems kind of
356 * scary; and urlencode() doesn't touch -_. to begin with. Plus, although /
357 * is reserved, we don't care. So the list we unescape is:
361 * However, IIS7 redirects fail when the url contains a colon (Bug 22709),
362 * so no fancy : for IIS7.
364 * %2F in the page titles seems to fatally break for some reason.
369 function wfUrlencode( $s ) {
372 if ( is_null( $s ) ) {
377 if ( is_null( $needle ) ) {
378 $needle = [ '%3B', '%40', '%24', '%21', '%2A', '%28', '%29', '%2C', '%2F', '%7E' ];
379 if ( !isset( $_SERVER['SERVER_SOFTWARE'] ) ||
380 ( strpos( $_SERVER['SERVER_SOFTWARE'], 'Microsoft-IIS/7' ) === false )
386 $s = urlencode( $s );
389 [ ';', '@', '$', '!', '*', '(', ')', ',', '/', '~', ':' ],
397 * This function takes one or two arrays as input, and returns a CGI-style string, e.g.
398 * "days=7&limit=100". Options in the first array override options in the second.
399 * Options set to null or false will not be output.
401 * @param array $array1 ( String|Array )
402 * @param array|null $array2 ( String|Array )
403 * @param string $prefix
406 function wfArrayToCgi( $array1, $array2 = null, $prefix = '' ) {
407 if ( !is_null( $array2 ) ) {
408 $array1 = $array1 +
$array2;
412 foreach ( $array1 as $key => $value ) {
413 if ( !is_null( $value ) && $value !== false ) {
417 if ( $prefix !== '' ) {
418 $key = $prefix . "[$key]";
420 if ( is_array( $value ) ) {
422 foreach ( $value as $k => $v ) {
423 $cgi .= $firstTime ?
'' : '&';
424 if ( is_array( $v ) ) {
425 $cgi .= wfArrayToCgi( $v, null, $key . "[$k]" );
427 $cgi .= urlencode( $key . "[$k]" ) . '=' . urlencode( $v );
432 if ( is_object( $value ) ) {
433 $value = $value->__toString();
435 $cgi .= urlencode( $key ) . '=' . urlencode( $value );
443 * This is the logical opposite of wfArrayToCgi(): it accepts a query string as
444 * its argument and returns the same string in array form. This allows compatibility
445 * with legacy functions that accept raw query strings instead of nice
446 * arrays. Of course, keys and values are urldecode()d.
448 * @param string $query Query string
449 * @return string[] Array version of input
451 function wfCgiToArray( $query ) {
452 if ( isset( $query[0] ) && $query[0] == '?' ) {
453 $query = substr( $query, 1 );
455 $bits = explode( '&', $query );
457 foreach ( $bits as $bit ) {
461 if ( strpos( $bit, '=' ) === false ) {
462 // Pieces like &qwerty become 'qwerty' => '' (at least this is what php does)
466 list( $key, $value ) = explode( '=', $bit );
468 $key = urldecode( $key );
469 $value = urldecode( $value );
470 if ( strpos( $key, '[' ) !== false ) {
471 $keys = array_reverse( explode( '[', $key ) );
472 $key = array_pop( $keys );
474 foreach ( $keys as $k ) {
475 $k = substr( $k, 0, -1 );
476 $temp = [ $k => $temp ];
478 if ( isset( $ret[$key] ) ) {
479 $ret[$key] = array_merge( $ret[$key], $temp );
491 * Append a query string to an existing URL, which may or may not already
492 * have query string parameters already. If so, they will be combined.
495 * @param string|string[] $query String or associative array
498 function wfAppendQuery( $url, $query ) {
499 if ( is_array( $query ) ) {
500 $query = wfArrayToCgi( $query );
502 if ( $query != '' ) {
503 // Remove the fragment, if there is one
505 $hashPos = strpos( $url, '#' );
506 if ( $hashPos !== false ) {
507 $fragment = substr( $url, $hashPos );
508 $url = substr( $url, 0, $hashPos );
512 if ( false === strpos( $url, '?' ) ) {
519 // Put the fragment back
520 if ( $fragment !== false ) {
528 * Expand a potentially local URL to a fully-qualified URL. Assumes $wgServer
531 * The meaning of the PROTO_* constants is as follows:
532 * PROTO_HTTP: Output a URL starting with http://
533 * PROTO_HTTPS: Output a URL starting with https://
534 * PROTO_RELATIVE: Output a URL starting with // (protocol-relative URL)
535 * PROTO_CURRENT: Output a URL starting with either http:// or https:// , depending
536 * on which protocol was used for the current incoming request
537 * PROTO_CANONICAL: For URLs without a domain, like /w/index.php , use $wgCanonicalServer.
538 * For protocol-relative URLs, use the protocol of $wgCanonicalServer
539 * PROTO_INTERNAL: Like PROTO_CANONICAL, but uses $wgInternalServer instead of $wgCanonicalServer
541 * @todo this won't work with current-path-relative URLs
542 * like "subdir/foo.html", etc.
544 * @param string $url Either fully-qualified or a local path + query
545 * @param string $defaultProto One of the PROTO_* constants. Determines the
546 * protocol to use if $url or $wgServer is protocol-relative
547 * @return string Fully-qualified URL, current-path-relative URL or false if
548 * no valid URL can be constructed
550 function wfExpandUrl( $url, $defaultProto = PROTO_CURRENT
) {
551 global $wgServer, $wgCanonicalServer, $wgInternalServer, $wgRequest,
553 if ( $defaultProto === PROTO_CANONICAL
) {
554 $serverUrl = $wgCanonicalServer;
555 } elseif ( $defaultProto === PROTO_INTERNAL
&& $wgInternalServer !== false ) {
556 // Make $wgInternalServer fall back to $wgServer if not set
557 $serverUrl = $wgInternalServer;
559 $serverUrl = $wgServer;
560 if ( $defaultProto === PROTO_CURRENT
) {
561 $defaultProto = $wgRequest->getProtocol() . '://';
565 // Analyze $serverUrl to obtain its protocol
566 $bits = wfParseUrl( $serverUrl );
567 $serverHasProto = $bits && $bits['scheme'] != '';
569 if ( $defaultProto === PROTO_CANONICAL ||
$defaultProto === PROTO_INTERNAL
) {
570 if ( $serverHasProto ) {
571 $defaultProto = $bits['scheme'] . '://';
573 // $wgCanonicalServer or $wgInternalServer doesn't have a protocol.
574 // This really isn't supposed to happen. Fall back to HTTP in this
576 $defaultProto = PROTO_HTTP
;
580 $defaultProtoWithoutSlashes = substr( $defaultProto, 0, -2 );
582 if ( substr( $url, 0, 2 ) == '//' ) {
583 $url = $defaultProtoWithoutSlashes . $url;
584 } elseif ( substr( $url, 0, 1 ) == '/' ) {
585 // If $serverUrl is protocol-relative, prepend $defaultProtoWithoutSlashes,
586 // otherwise leave it alone.
587 $url = ( $serverHasProto ?
'' : $defaultProtoWithoutSlashes ) . $serverUrl . $url;
590 $bits = wfParseUrl( $url );
592 // ensure proper port for HTTPS arrives in URL
593 // https://phabricator.wikimedia.org/T67184
594 if ( $defaultProto === PROTO_HTTPS
&& $wgHttpsPort != 443 ) {
595 $bits['port'] = $wgHttpsPort;
598 if ( $bits && isset( $bits['path'] ) ) {
599 $bits['path'] = wfRemoveDotSegments( $bits['path'] );
600 return wfAssembleUrl( $bits );
604 } elseif ( substr( $url, 0, 1 ) != '/' ) {
605 # URL is a relative path
606 return wfRemoveDotSegments( $url );
609 # Expanded URL is not valid.
614 * This function will reassemble a URL parsed with wfParseURL. This is useful
615 * if you need to edit part of a URL and put it back together.
617 * This is the basic structure used (brackets contain keys for $urlParts):
618 * [scheme][delimiter][user]:[pass]@[host]:[port][path]?[query]#[fragment]
620 * @todo Need to integrate this into wfExpandUrl (bug 32168)
623 * @param array $urlParts URL parts, as output from wfParseUrl
624 * @return string URL assembled from its component parts
626 function wfAssembleUrl( $urlParts ) {
629 if ( isset( $urlParts['delimiter'] ) ) {
630 if ( isset( $urlParts['scheme'] ) ) {
631 $result .= $urlParts['scheme'];
634 $result .= $urlParts['delimiter'];
637 if ( isset( $urlParts['host'] ) ) {
638 if ( isset( $urlParts['user'] ) ) {
639 $result .= $urlParts['user'];
640 if ( isset( $urlParts['pass'] ) ) {
641 $result .= ':' . $urlParts['pass'];
646 $result .= $urlParts['host'];
648 if ( isset( $urlParts['port'] ) ) {
649 $result .= ':' . $urlParts['port'];
653 if ( isset( $urlParts['path'] ) ) {
654 $result .= $urlParts['path'];
657 if ( isset( $urlParts['query'] ) ) {
658 $result .= '?' . $urlParts['query'];
661 if ( isset( $urlParts['fragment'] ) ) {
662 $result .= '#' . $urlParts['fragment'];
669 * Remove all dot-segments in the provided URL path. For example,
670 * '/a/./b/../c/' becomes '/a/c/'. For details on the algorithm, please see
671 * RFC3986 section 5.2.4.
673 * @todo Need to integrate this into wfExpandUrl (bug 32168)
675 * @param string $urlPath URL path, potentially containing dot-segments
676 * @return string URL path with all dot-segments removed
678 function wfRemoveDotSegments( $urlPath ) {
681 $inputLength = strlen( $urlPath );
683 while ( $inputOffset < $inputLength ) {
684 $prefixLengthOne = substr( $urlPath, $inputOffset, 1 );
685 $prefixLengthTwo = substr( $urlPath, $inputOffset, 2 );
686 $prefixLengthThree = substr( $urlPath, $inputOffset, 3 );
687 $prefixLengthFour = substr( $urlPath, $inputOffset, 4 );
690 if ( $prefixLengthTwo == './' ) {
691 # Step A, remove leading "./"
693 } elseif ( $prefixLengthThree == '../' ) {
694 # Step A, remove leading "../"
696 } elseif ( ( $prefixLengthTwo == '/.' ) && ( $inputOffset +
2 == $inputLength ) ) {
697 # Step B, replace leading "/.$" with "/"
699 $urlPath[$inputOffset] = '/';
700 } elseif ( $prefixLengthThree == '/./' ) {
701 # Step B, replace leading "/./" with "/"
703 } elseif ( $prefixLengthThree == '/..' && ( $inputOffset +
3 == $inputLength ) ) {
704 # Step C, replace leading "/..$" with "/" and
705 # remove last path component in output
707 $urlPath[$inputOffset] = '/';
709 } elseif ( $prefixLengthFour == '/../' ) {
710 # Step C, replace leading "/../" with "/" and
711 # remove last path component in output
714 } elseif ( ( $prefixLengthOne == '.' ) && ( $inputOffset +
1 == $inputLength ) ) {
715 # Step D, remove "^.$"
717 } elseif ( ( $prefixLengthTwo == '..' ) && ( $inputOffset +
2 == $inputLength ) ) {
718 # Step D, remove "^..$"
721 # Step E, move leading path segment to output
722 if ( $prefixLengthOne == '/' ) {
723 $slashPos = strpos( $urlPath, '/', $inputOffset +
1 );
725 $slashPos = strpos( $urlPath, '/', $inputOffset );
727 if ( $slashPos === false ) {
728 $output .= substr( $urlPath, $inputOffset );
729 $inputOffset = $inputLength;
731 $output .= substr( $urlPath, $inputOffset, $slashPos - $inputOffset );
732 $inputOffset +
= $slashPos - $inputOffset;
737 $slashPos = strrpos( $output, '/' );
738 if ( $slashPos === false ) {
741 $output = substr( $output, 0, $slashPos );
750 * Returns a regular expression of url protocols
752 * @param bool $includeProtocolRelative If false, remove '//' from the returned protocol list.
753 * DO NOT USE this directly, use wfUrlProtocolsWithoutProtRel() instead
756 function wfUrlProtocols( $includeProtocolRelative = true ) {
757 global $wgUrlProtocols;
759 // Cache return values separately based on $includeProtocolRelative
760 static $withProtRel = null, $withoutProtRel = null;
761 $cachedValue = $includeProtocolRelative ?
$withProtRel : $withoutProtRel;
762 if ( !is_null( $cachedValue ) ) {
766 // Support old-style $wgUrlProtocols strings, for backwards compatibility
767 // with LocalSettings files from 1.5
768 if ( is_array( $wgUrlProtocols ) ) {
770 foreach ( $wgUrlProtocols as $protocol ) {
771 // Filter out '//' if !$includeProtocolRelative
772 if ( $includeProtocolRelative ||
$protocol !== '//' ) {
773 $protocols[] = preg_quote( $protocol, '/' );
777 $retval = implode( '|', $protocols );
779 // Ignore $includeProtocolRelative in this case
780 // This case exists for pre-1.6 compatibility, and we can safely assume
781 // that '//' won't appear in a pre-1.6 config because protocol-relative
782 // URLs weren't supported until 1.18
783 $retval = $wgUrlProtocols;
786 // Cache return value
787 if ( $includeProtocolRelative ) {
788 $withProtRel = $retval;
790 $withoutProtRel = $retval;
796 * Like wfUrlProtocols(), but excludes '//' from the protocol list. Use this if
797 * you need a regex that matches all URL protocols but does not match protocol-
801 function wfUrlProtocolsWithoutProtRel() {
802 return wfUrlProtocols( false );
806 * parse_url() work-alike, but non-broken. Differences:
808 * 1) Does not raise warnings on bad URLs (just returns false).
809 * 2) Handles protocols that don't use :// (e.g., mailto: and news:, as well as
810 * protocol-relative URLs) correctly.
811 * 3) Adds a "delimiter" element to the array, either '://', ':' or '//' (see (2)).
813 * @param string $url A URL to parse
814 * @return string[]|bool Bits of the URL in an associative array, per PHP docs, false on failure
816 function wfParseUrl( $url ) {
817 global $wgUrlProtocols; // Allow all protocols defined in DefaultSettings/LocalSettings.php
819 // Protocol-relative URLs are handled really badly by parse_url(). It's so
820 // bad that the easiest way to handle them is to just prepend 'http:' and
821 // strip the protocol out later.
822 $wasRelative = substr( $url, 0, 2 ) == '//';
823 if ( $wasRelative ) {
826 MediaWiki\
suppressWarnings();
827 $bits = parse_url( $url );
828 MediaWiki\restoreWarnings
();
829 // parse_url() returns an array without scheme for some invalid URLs, e.g.
830 // parse_url("%0Ahttp://example.com") == [ 'host' => '%0Ahttp', 'path' => 'example.com' ]
831 if ( !$bits ||
!isset( $bits['scheme'] ) ) {
835 // parse_url() incorrectly handles schemes case-sensitively. Convert it to lowercase.
836 $bits['scheme'] = strtolower( $bits['scheme'] );
838 // most of the protocols are followed by ://, but mailto: and sometimes news: not, check for it
839 if ( in_array( $bits['scheme'] . '://', $wgUrlProtocols ) ) {
840 $bits['delimiter'] = '://';
841 } elseif ( in_array( $bits['scheme'] . ':', $wgUrlProtocols ) ) {
842 $bits['delimiter'] = ':';
843 // parse_url detects for news: and mailto: the host part of an url as path
844 // We have to correct this wrong detection
845 if ( isset( $bits['path'] ) ) {
846 $bits['host'] = $bits['path'];
853 /* Provide an empty host for eg. file:/// urls (see bug 28627) */
854 if ( !isset( $bits['host'] ) ) {
858 if ( isset( $bits['path'] ) ) {
859 /* parse_url loses the third / for file:///c:/ urls (but not on variants) */
860 if ( substr( $bits['path'], 0, 1 ) !== '/' ) {
861 $bits['path'] = '/' . $bits['path'];
868 // If the URL was protocol-relative, fix scheme and delimiter
869 if ( $wasRelative ) {
870 $bits['scheme'] = '';
871 $bits['delimiter'] = '//';
877 * Take a URL, make sure it's expanded to fully qualified, and replace any
878 * encoded non-ASCII Unicode characters with their UTF-8 original forms
879 * for more compact display and legibility for local audiences.
881 * @todo handle punycode domains too
886 function wfExpandIRI( $url ) {
887 return preg_replace_callback(
888 '/((?:%[89A-F][0-9A-F])+)/i',
889 'wfExpandIRI_callback',
895 * Private callback for wfExpandIRI
896 * @param array $matches
899 function wfExpandIRI_callback( $matches ) {
900 return urldecode( $matches[1] );
904 * Make URL indexes, appropriate for the el_index field of externallinks.
909 function wfMakeUrlIndexes( $url ) {
910 $bits = wfParseUrl( $url );
912 // Reverse the labels in the hostname, convert to lower case
913 // For emails reverse domainpart only
914 if ( $bits['scheme'] == 'mailto' ) {
915 $mailparts = explode( '@', $bits['host'], 2 );
916 if ( count( $mailparts ) === 2 ) {
917 $domainpart = strtolower( implode( '.', array_reverse( explode( '.', $mailparts[1] ) ) ) );
919 // No domain specified, don't mangle it
922 $reversedHost = $domainpart . '@' . $mailparts[0];
924 $reversedHost = strtolower( implode( '.', array_reverse( explode( '.', $bits['host'] ) ) ) );
926 // Add an extra dot to the end
927 // Why? Is it in wrong place in mailto links?
928 if ( substr( $reversedHost, -1, 1 ) !== '.' ) {
929 $reversedHost .= '.';
931 // Reconstruct the pseudo-URL
932 $prot = $bits['scheme'];
933 $index = $prot . $bits['delimiter'] . $reversedHost;
934 // Leave out user and password. Add the port, path, query and fragment
935 if ( isset( $bits['port'] ) ) {
936 $index .= ':' . $bits['port'];
938 if ( isset( $bits['path'] ) ) {
939 $index .= $bits['path'];
943 if ( isset( $bits['query'] ) ) {
944 $index .= '?' . $bits['query'];
946 if ( isset( $bits['fragment'] ) ) {
947 $index .= '#' . $bits['fragment'];
951 return [ "http:$index", "https:$index" ];
958 * Check whether a given URL has a domain that occurs in a given set of domains
959 * @param string $url URL
960 * @param array $domains Array of domains (strings)
961 * @return bool True if the host part of $url ends in one of the strings in $domains
963 function wfMatchesDomainList( $url, $domains ) {
964 $bits = wfParseUrl( $url );
965 if ( is_array( $bits ) && isset( $bits['host'] ) ) {
966 $host = '.' . $bits['host'];
967 foreach ( (array)$domains as $domain ) {
968 $domain = '.' . $domain;
969 if ( substr( $host, -strlen( $domain ) ) === $domain ) {
978 * Sends a line to the debug log if enabled or, optionally, to a comment in output.
979 * In normal operation this is a NOP.
981 * Controlling globals:
982 * $wgDebugLogFile - points to the log file
983 * $wgDebugRawPage - if false, 'action=raw' hits will not result in debug output.
984 * $wgDebugComments - if on, some debug items may appear in comments in the HTML output.
986 * @since 1.25 support for additional context data
988 * @param string $text
989 * @param string|bool $dest Destination of the message:
990 * - 'all': both to the log and HTML (debug toolbar or HTML comments)
991 * - 'private': excluded from HTML output
992 * For backward compatibility, it can also take a boolean:
993 * - true: same as 'all'
994 * - false: same as 'private'
995 * @param array $context Additional logging context data
997 function wfDebug( $text, $dest = 'all', array $context = [] ) {
998 global $wgDebugRawPage, $wgDebugLogPrefix;
999 global $wgDebugTimestamps, $wgRequestTime;
1001 if ( !$wgDebugRawPage && wfIsDebugRawPage() ) {
1005 $text = trim( $text );
1007 if ( $wgDebugTimestamps ) {
1008 $context['seconds_elapsed'] = sprintf(
1010 microtime( true ) - $wgRequestTime
1012 $context['memory_used'] = sprintf(
1014 ( memory_get_usage( true ) / ( 1024 * 1024 ) )
1018 if ( $wgDebugLogPrefix !== '' ) {
1019 $context['prefix'] = $wgDebugLogPrefix;
1021 $context['private'] = ( $dest === false ||
$dest === 'private' );
1023 $logger = LoggerFactory
::getInstance( 'wfDebug' );
1024 $logger->debug( $text, $context );
1028 * Returns true if debug logging should be suppressed if $wgDebugRawPage = false
1031 function wfIsDebugRawPage() {
1033 if ( $cache !== null ) {
1036 # Check for raw action using $_GET not $wgRequest, since the latter might not be initialised yet
1037 if ( ( isset( $_GET['action'] ) && $_GET['action'] == 'raw' )
1039 isset( $_SERVER['SCRIPT_NAME'] )
1040 && substr( $_SERVER['SCRIPT_NAME'], -8 ) == 'load.php'
1051 * Send a line giving PHP memory usage.
1053 * @param bool $exact Print exact byte values instead of kibibytes (default: false)
1055 function wfDebugMem( $exact = false ) {
1056 $mem = memory_get_usage();
1058 $mem = floor( $mem / 1024 ) . ' KiB';
1062 wfDebug( "Memory usage: $mem\n" );
1066 * Send a line to a supplementary debug log file, if configured, or main debug
1069 * To configure a supplementary log file, set $wgDebugLogGroups[$logGroup] to
1070 * a string filename or an associative array mapping 'destination' to the
1071 * desired filename. The associative array may also contain a 'sample' key
1072 * with an integer value, specifying a sampling factor. Sampled log events
1073 * will be emitted with a 1 in N random chance.
1075 * @since 1.23 support for sampling log messages via $wgDebugLogGroups.
1076 * @since 1.25 support for additional context data
1077 * @since 1.25 sample behavior dependent on configured $wgMWLoggerDefaultSpi
1079 * @param string $logGroup
1080 * @param string $text
1081 * @param string|bool $dest Destination of the message:
1082 * - 'all': both to the log and HTML (debug toolbar or HTML comments)
1083 * - 'private': only to the specific log if set in $wgDebugLogGroups and
1084 * discarded otherwise
1085 * For backward compatibility, it can also take a boolean:
1086 * - true: same as 'all'
1087 * - false: same as 'private'
1088 * @param array $context Additional logging context data
1090 function wfDebugLog(
1091 $logGroup, $text, $dest = 'all', array $context = []
1093 $text = trim( $text );
1095 $logger = LoggerFactory
::getInstance( $logGroup );
1096 $context['private'] = ( $dest === false ||
$dest === 'private' );
1097 $logger->info( $text, $context );
1101 * Log for database errors
1103 * @since 1.25 support for additional context data
1105 * @param string $text Database error message.
1106 * @param array $context Additional logging context data
1108 function wfLogDBError( $text, array $context = [] ) {
1109 $logger = LoggerFactory
::getInstance( 'wfLogDBError' );
1110 $logger->error( trim( $text ), $context );
1114 * Throws a warning that $function is deprecated
1116 * @param string $function
1117 * @param string|bool $version Version of MediaWiki that the function
1118 * was deprecated in (Added in 1.19).
1119 * @param string|bool $component Added in 1.19.
1120 * @param int $callerOffset How far up the call stack is the original
1121 * caller. 2 = function that called the function that called
1122 * wfDeprecated (Added in 1.20)
1126 function wfDeprecated( $function, $version = false, $component = false, $callerOffset = 2 ) {
1127 MWDebug
::deprecated( $function, $version, $component, $callerOffset +
1 );
1131 * Send a warning either to the debug log or in a PHP error depending on
1132 * $wgDevelopmentWarnings. To log warnings in production, use wfLogWarning() instead.
1134 * @param string $msg Message to send
1135 * @param int $callerOffset Number of items to go back in the backtrace to
1136 * find the correct caller (1 = function calling wfWarn, ...)
1137 * @param int $level PHP error level; defaults to E_USER_NOTICE;
1138 * only used when $wgDevelopmentWarnings is true
1140 function wfWarn( $msg, $callerOffset = 1, $level = E_USER_NOTICE
) {
1141 MWDebug
::warning( $msg, $callerOffset +
1, $level, 'auto' );
1145 * Send a warning as a PHP error and the debug log. This is intended for logging
1146 * warnings in production. For logging development warnings, use WfWarn instead.
1148 * @param string $msg Message to send
1149 * @param int $callerOffset Number of items to go back in the backtrace to
1150 * find the correct caller (1 = function calling wfLogWarning, ...)
1151 * @param int $level PHP error level; defaults to E_USER_WARNING
1153 function wfLogWarning( $msg, $callerOffset = 1, $level = E_USER_WARNING
) {
1154 MWDebug
::warning( $msg, $callerOffset +
1, $level, 'production' );
1158 * Log to a file without getting "file size exceeded" signals.
1160 * Can also log to TCP or UDP with the syntax udp://host:port/prefix. This will
1161 * send lines to the specified port, prefixed by the specified prefix and a space.
1162 * @since 1.25 support for additional context data
1164 * @param string $text
1165 * @param string $file Filename
1166 * @param array $context Additional logging context data
1167 * @throws MWException
1168 * @deprecated since 1.25 Use \MediaWiki\Logger\LegacyLogger::emit or UDPTransport
1170 function wfErrorLog( $text, $file, array $context = [] ) {
1171 wfDeprecated( __METHOD__
, '1.25' );
1172 $logger = LoggerFactory
::getInstance( 'wfErrorLog' );
1173 $context['destination'] = $file;
1174 $logger->info( trim( $text ), $context );
1180 function wfLogProfilingData() {
1181 global $wgDebugLogGroups, $wgDebugRawPage;
1183 $context = RequestContext
::getMain();
1184 $request = $context->getRequest();
1186 $profiler = Profiler
::instance();
1187 $profiler->setContext( $context );
1188 $profiler->logData();
1190 $config = $context->getConfig();
1191 if ( $config->get( 'StatsdServer' ) ) {
1193 $statsdServer = explode( ':', $config->get( 'StatsdServer' ) );
1194 $statsdHost = $statsdServer[0];
1195 $statsdPort = isset( $statsdServer[1] ) ?
$statsdServer[1] : 8125;
1196 $statsdSender = new SocketSender( $statsdHost, $statsdPort );
1197 $statsdClient = new SamplingStatsdClient( $statsdSender, true, false );
1198 $statsdClient->setSamplingRates( $config->get( 'StatsdSamplingRates' ) );
1199 $statsdClient->send( $context->getStats()->getBuffer() );
1200 } catch ( Exception
$ex ) {
1201 MWExceptionHandler
::logException( $ex );
1205 # Profiling must actually be enabled...
1206 if ( $profiler instanceof ProfilerStub
) {
1210 if ( isset( $wgDebugLogGroups['profileoutput'] )
1211 && $wgDebugLogGroups['profileoutput'] === false
1213 // Explicitly disabled
1216 if ( !$wgDebugRawPage && wfIsDebugRawPage() ) {
1220 $ctx = [ 'elapsed' => $request->getElapsedTime() ];
1221 if ( !empty( $_SERVER['HTTP_X_FORWARDED_FOR'] ) ) {
1222 $ctx['forwarded_for'] = $_SERVER['HTTP_X_FORWARDED_FOR'];
1224 if ( !empty( $_SERVER['HTTP_CLIENT_IP'] ) ) {
1225 $ctx['client_ip'] = $_SERVER['HTTP_CLIENT_IP'];
1227 if ( !empty( $_SERVER['HTTP_FROM'] ) ) {
1228 $ctx['from'] = $_SERVER['HTTP_FROM'];
1230 if ( isset( $ctx['forwarded_for'] ) ||
1231 isset( $ctx['client_ip'] ) ||
1232 isset( $ctx['from'] ) ) {
1233 $ctx['proxy'] = $_SERVER['REMOTE_ADDR'];
1236 // Don't load $wgUser at this late stage just for statistics purposes
1237 // @todo FIXME: We can detect some anons even if it is not loaded.
1238 // See User::getId()
1239 $user = $context->getUser();
1240 $ctx['anon'] = $user->isItemLoaded( 'id' ) && $user->isAnon();
1242 // Command line script uses a FauxRequest object which does not have
1243 // any knowledge about an URL and throw an exception instead.
1245 $ctx['url'] = urldecode( $request->getRequestURL() );
1246 } catch ( Exception
$ignored ) {
1250 $ctx['output'] = $profiler->getOutput();
1252 $log = LoggerFactory
::getInstance( 'profileoutput' );
1253 $log->info( "Elapsed: {elapsed}; URL: <{url}>\n{output}", $ctx );
1257 * Increment a statistics counter
1259 * @param string $key
1263 function wfIncrStats( $key, $count = 1 ) {
1264 $stats = RequestContext
::getMain()->getStats();
1265 $stats->updateCount( $key, $count );
1269 * Check whether the wiki is in read-only mode.
1273 function wfReadOnly() {
1274 return wfReadOnlyReason() !== false;
1278 * Check if the site is in read-only mode and return the message if so
1280 * This checks wfConfiguredReadOnlyReason() and the main load balancer
1281 * for replica DB lag. This may result in DB connection being made.
1283 * @return string|bool String when in read-only mode; false otherwise
1285 function wfReadOnlyReason() {
1286 $readOnly = wfConfiguredReadOnlyReason();
1287 if ( $readOnly !== false ) {
1291 static $lbReadOnly = null;
1292 if ( $lbReadOnly === null ) {
1293 // Callers use this method to be aware that data presented to a user
1294 // may be very stale and thus allowing submissions can be problematic.
1295 $lbReadOnly = wfGetLB()->getReadOnlyReason();
1302 * Get the value of $wgReadOnly or the contents of $wgReadOnlyFile.
1304 * @return string|bool String when in read-only mode; false otherwise
1307 function wfConfiguredReadOnlyReason() {
1308 global $wgReadOnly, $wgReadOnlyFile;
1310 if ( $wgReadOnly === null ) {
1311 // Set $wgReadOnly for faster access next time
1312 if ( is_file( $wgReadOnlyFile ) && filesize( $wgReadOnlyFile ) > 0 ) {
1313 $wgReadOnly = file_get_contents( $wgReadOnlyFile );
1315 $wgReadOnly = false;
1323 * Return a Language object from $langcode
1325 * @param Language|string|bool $langcode Either:
1326 * - a Language object
1327 * - code of the language to get the message for, if it is
1328 * a valid code create a language for that language, if
1329 * it is a string but not a valid code then make a basic
1331 * - a boolean: if it's false then use the global object for
1332 * the current user's language (as a fallback for the old parameter
1333 * functionality), or if it is true then use global object
1334 * for the wiki's content language.
1337 function wfGetLangObj( $langcode = false ) {
1338 # Identify which language to get or create a language object for.
1339 # Using is_object here due to Stub objects.
1340 if ( is_object( $langcode ) ) {
1341 # Great, we already have the object (hopefully)!
1345 global $wgContLang, $wgLanguageCode;
1346 if ( $langcode === true ||
$langcode === $wgLanguageCode ) {
1347 # $langcode is the language code of the wikis content language object.
1348 # or it is a boolean and value is true
1353 if ( $langcode === false ||
$langcode === $wgLang->getCode() ) {
1354 # $langcode is the language code of user language object.
1355 # or it was a boolean and value is false
1359 $validCodes = array_keys( Language
::fetchLanguageNames() );
1360 if ( in_array( $langcode, $validCodes ) ) {
1361 # $langcode corresponds to a valid language.
1362 return Language
::factory( $langcode );
1365 # $langcode is a string, but not a valid language code; use content language.
1366 wfDebug( "Invalid language code passed to wfGetLangObj, falling back to content language.\n" );
1371 * This is the function for getting translated interface messages.
1373 * @see Message class for documentation how to use them.
1374 * @see https://www.mediawiki.org/wiki/Manual:Messages_API
1376 * This function replaces all old wfMsg* functions.
1378 * @param string|string[]|MessageSpecifier $key Message key, or array of keys, or a MessageSpecifier
1379 * @param mixed $params,... Normal message parameters
1384 * @see Message::__construct
1386 function wfMessage( $key /*...*/ ) {
1387 $params = func_get_args();
1388 array_shift( $params );
1389 if ( isset( $params[0] ) && is_array( $params[0] ) ) {
1390 $params = $params[0];
1392 return new Message( $key, $params );
1396 * This function accepts multiple message keys and returns a message instance
1397 * for the first message which is non-empty. If all messages are empty then an
1398 * instance of the first message key is returned.
1400 * @param string|string[] $keys,... Message keys
1405 * @see Message::newFallbackSequence
1407 function wfMessageFallback( /*...*/ ) {
1408 $args = func_get_args();
1409 return call_user_func_array( 'Message::newFallbackSequence', $args );
1413 * Replace message parameter keys on the given formatted output.
1415 * @param string $message
1416 * @param array $args
1420 function wfMsgReplaceArgs( $message, $args ) {
1421 # Fix windows line-endings
1422 # Some messages are split with explode("\n", $msg)
1423 $message = str_replace( "\r", '', $message );
1425 // Replace arguments
1426 if ( is_array( $args ) && $args ) {
1427 if ( is_array( $args[0] ) ) {
1428 $args = array_values( $args[0] );
1430 $replacementKeys = [];
1431 foreach ( $args as $n => $param ) {
1432 $replacementKeys['$' . ( $n +
1 )] = $param;
1434 $message = strtr( $message, $replacementKeys );
1441 * Fetch server name for use in error reporting etc.
1442 * Use real server name if available, so we know which machine
1443 * in a server farm generated the current page.
1447 function wfHostname() {
1449 if ( is_null( $host ) ) {
1451 # Hostname overriding
1452 global $wgOverrideHostname;
1453 if ( $wgOverrideHostname !== false ) {
1454 # Set static and skip any detection
1455 $host = $wgOverrideHostname;
1459 if ( function_exists( 'posix_uname' ) ) {
1460 // This function not present on Windows
1461 $uname = posix_uname();
1465 if ( is_array( $uname ) && isset( $uname['nodename'] ) ) {
1466 $host = $uname['nodename'];
1467 } elseif ( getenv( 'COMPUTERNAME' ) ) {
1468 # Windows computer name
1469 $host = getenv( 'COMPUTERNAME' );
1471 # This may be a virtual server.
1472 $host = $_SERVER['SERVER_NAME'];
1479 * Returns a script tag that stores the amount of time it took MediaWiki to
1480 * handle the request in milliseconds as 'wgBackendResponseTime'.
1482 * If $wgShowHostnames is true, the script will also set 'wgHostname' to the
1483 * hostname of the server handling the request.
1487 function wfReportTime() {
1488 global $wgRequestTime, $wgShowHostnames;
1490 $responseTime = round( ( microtime( true ) - $wgRequestTime ) * 1000 );
1491 $reportVars = [ 'wgBackendResponseTime' => $responseTime ];
1492 if ( $wgShowHostnames ) {
1493 $reportVars['wgHostname'] = wfHostname();
1495 return Skin
::makeVariablesScript( $reportVars );
1499 * Safety wrapper for debug_backtrace().
1501 * Will return an empty array if debug_backtrace is disabled, otherwise
1502 * the output from debug_backtrace() (trimmed).
1504 * @param int $limit This parameter can be used to limit the number of stack frames returned
1506 * @return array Array of backtrace information
1508 function wfDebugBacktrace( $limit = 0 ) {
1509 static $disabled = null;
1511 if ( is_null( $disabled ) ) {
1512 $disabled = !function_exists( 'debug_backtrace' );
1514 wfDebug( "debug_backtrace() is disabled\n" );
1522 return array_slice( debug_backtrace( DEBUG_BACKTRACE_PROVIDE_OBJECT
, $limit +
1 ), 1 );
1524 return array_slice( debug_backtrace(), 1 );
1529 * Get a debug backtrace as a string
1531 * @param bool|null $raw If true, the return value is plain text. If false, HTML.
1532 * Defaults to $wgCommandLineMode if unset.
1534 * @since 1.25 Supports $raw parameter.
1536 function wfBacktrace( $raw = null ) {
1537 global $wgCommandLineMode;
1539 if ( $raw === null ) {
1540 $raw = $wgCommandLineMode;
1544 $frameFormat = "%s line %s calls %s()\n";
1545 $traceFormat = "%s";
1547 $frameFormat = "<li>%s line %s calls %s()</li>\n";
1548 $traceFormat = "<ul>\n%s</ul>\n";
1551 $frames = array_map( function ( $frame ) use ( $frameFormat ) {
1552 $file = !empty( $frame['file'] ) ?
basename( $frame['file'] ) : '-';
1553 $line = isset( $frame['line'] ) ?
$frame['line'] : '-';
1554 $call = $frame['function'];
1555 if ( !empty( $frame['class'] ) ) {
1556 $call = $frame['class'] . $frame['type'] . $call;
1558 return sprintf( $frameFormat, $file, $line, $call );
1559 }, wfDebugBacktrace() );
1561 return sprintf( $traceFormat, implode( '', $frames ) );
1565 * Get the name of the function which called this function
1566 * wfGetCaller( 1 ) is the function with the wfGetCaller() call (ie. __FUNCTION__)
1567 * wfGetCaller( 2 ) [default] is the caller of the function running wfGetCaller()
1568 * wfGetCaller( 3 ) is the parent of that.
1573 function wfGetCaller( $level = 2 ) {
1574 $backtrace = wfDebugBacktrace( $level +
1 );
1575 if ( isset( $backtrace[$level] ) ) {
1576 return wfFormatStackFrame( $backtrace[$level] );
1583 * Return a string consisting of callers in the stack. Useful sometimes
1584 * for profiling specific points.
1586 * @param int $limit The maximum depth of the stack frame to return, or false for the entire stack.
1589 function wfGetAllCallers( $limit = 3 ) {
1590 $trace = array_reverse( wfDebugBacktrace() );
1591 if ( !$limit ||
$limit > count( $trace ) - 1 ) {
1592 $limit = count( $trace ) - 1;
1594 $trace = array_slice( $trace, -$limit - 1, $limit );
1595 return implode( '/', array_map( 'wfFormatStackFrame', $trace ) );
1599 * Return a string representation of frame
1601 * @param array $frame
1604 function wfFormatStackFrame( $frame ) {
1605 if ( !isset( $frame['function'] ) ) {
1606 return 'NO_FUNCTION_GIVEN';
1608 return isset( $frame['class'] ) && isset( $frame['type'] ) ?
1609 $frame['class'] . $frame['type'] . $frame['function'] :
1613 /* Some generic result counters, pulled out of SearchEngine */
1618 * @param int $offset
1622 function wfShowingResults( $offset, $limit ) {
1623 return wfMessage( 'showingresults' )->numParams( $limit, $offset +
1 )->parse();
1628 * @todo FIXME: We may want to blacklist some broken browsers
1630 * @param bool $force
1631 * @return bool Whereas client accept gzip compression
1633 function wfClientAcceptsGzip( $force = false ) {
1634 static $result = null;
1635 if ( $result === null ||
$force ) {
1637 if ( isset( $_SERVER['HTTP_ACCEPT_ENCODING'] ) ) {
1638 # @todo FIXME: We may want to blacklist some broken browsers
1641 '/\bgzip(?:;(q)=([0-9]+(?:\.[0-9]+)))?\b/',
1642 $_SERVER['HTTP_ACCEPT_ENCODING'],
1646 if ( isset( $m[2] ) && ( $m[1] == 'q' ) && ( $m[2] == 0 ) ) {
1650 wfDebug( "wfClientAcceptsGzip: client accepts gzip.\n" );
1659 * Escapes the given text so that it may be output using addWikiText()
1660 * without any linking, formatting, etc. making its way through. This
1661 * is achieved by substituting certain characters with HTML entities.
1662 * As required by the callers, "<nowiki>" is not used.
1664 * @param string $text Text to be escaped
1667 function wfEscapeWikiText( $text ) {
1668 static $repl = null, $repl2 = null;
1669 if ( $repl === null ) {
1671 '"' => '"', '&' => '&', "'" => ''', '<' => '<',
1672 '=' => '=', '>' => '>', '[' => '[', ']' => ']',
1673 '{' => '{', '|' => '|', '}' => '}', ';' => ';',
1674 "\n#" => "\n#", "\r#" => "\r#",
1675 "\n*" => "\n*", "\r*" => "\r*",
1676 "\n:" => "\n:", "\r:" => "\r:",
1677 "\n " => "\n ", "\r " => "\r ",
1678 "\n\n" => "\n ", "\r\n" => " \n",
1679 "\n\r" => "\n ", "\r\r" => "\r ",
1680 "\n\t" => "\n	", "\r\t" => "\r	", // "\n\t\n" is treated like "\n\n"
1681 "\n----" => "\n----", "\r----" => "\r----",
1682 '__' => '__', '://' => '://',
1685 // We have to catch everything "\s" matches in PCRE
1686 foreach ( [ 'ISBN', 'RFC', 'PMID' ] as $magic ) {
1687 $repl["$magic "] = "$magic ";
1688 $repl["$magic\t"] = "$magic	";
1689 $repl["$magic\r"] = "$magic ";
1690 $repl["$magic\n"] = "$magic ";
1691 $repl["$magic\f"] = "$magic";
1694 // And handle protocols that don't use "://"
1695 global $wgUrlProtocols;
1697 foreach ( $wgUrlProtocols as $prot ) {
1698 if ( substr( $prot, -1 ) === ':' ) {
1699 $repl2[] = preg_quote( substr( $prot, 0, -1 ), '/' );
1702 $repl2 = $repl2 ?
'/\b(' . implode( '|', $repl2 ) . '):/i' : '/^(?!)/';
1704 $text = substr( strtr( "\n$text", $repl ), 1 );
1705 $text = preg_replace( $repl2, '$1:', $text );
1710 * Sets dest to source and returns the original value of dest
1711 * If source is NULL, it just returns the value, it doesn't set the variable
1712 * If force is true, it will set the value even if source is NULL
1714 * @param mixed $dest
1715 * @param mixed $source
1716 * @param bool $force
1719 function wfSetVar( &$dest, $source, $force = false ) {
1721 if ( !is_null( $source ) ||
$force ) {
1728 * As for wfSetVar except setting a bit
1732 * @param bool $state
1736 function wfSetBit( &$dest, $bit, $state = true ) {
1737 $temp = (bool)( $dest & $bit );
1738 if ( !is_null( $state ) ) {
1749 * A wrapper around the PHP function var_export().
1750 * Either print it or add it to the regular output ($wgOut).
1752 * @param mixed $var A PHP variable to dump.
1754 function wfVarDump( $var ) {
1756 $s = str_replace( "\n", "<br />\n", var_export( $var, true ) . "\n" );
1757 if ( headers_sent() ||
!isset( $wgOut ) ||
!is_object( $wgOut ) ) {
1760 $wgOut->addHTML( $s );
1765 * Provide a simple HTTP error.
1767 * @param int|string $code
1768 * @param string $label
1769 * @param string $desc
1771 function wfHttpError( $code, $label, $desc ) {
1773 HttpStatus
::header( $code );
1776 $wgOut->sendCacheControl();
1779 header( 'Content-type: text/html; charset=utf-8' );
1780 print '<!DOCTYPE html>' .
1781 '<html><head><title>' .
1782 htmlspecialchars( $label ) .
1783 '</title></head><body><h1>' .
1784 htmlspecialchars( $label ) .
1786 nl2br( htmlspecialchars( $desc ) ) .
1787 "</p></body></html>\n";
1791 * Clear away any user-level output buffers, discarding contents.
1793 * Suitable for 'starting afresh', for instance when streaming
1794 * relatively large amounts of data without buffering, or wanting to
1795 * output image files without ob_gzhandler's compression.
1797 * The optional $resetGzipEncoding parameter controls suppression of
1798 * the Content-Encoding header sent by ob_gzhandler; by default it
1799 * is left. See comments for wfClearOutputBuffers() for why it would
1802 * Note that some PHP configuration options may add output buffer
1803 * layers which cannot be removed; these are left in place.
1805 * @param bool $resetGzipEncoding
1807 function wfResetOutputBuffers( $resetGzipEncoding = true ) {
1808 if ( $resetGzipEncoding ) {
1809 // Suppress Content-Encoding and Content-Length
1810 // headers from 1.10+s wfOutputHandler
1811 global $wgDisableOutputCompression;
1812 $wgDisableOutputCompression = true;
1814 while ( $status = ob_get_status() ) {
1815 if ( isset( $status['flags'] ) ) {
1816 $flags = PHP_OUTPUT_HANDLER_CLEANABLE | PHP_OUTPUT_HANDLER_REMOVABLE
;
1817 $deleteable = ( $status['flags'] & $flags ) === $flags;
1818 } elseif ( isset( $status['del'] ) ) {
1819 $deleteable = $status['del'];
1821 // Guess that any PHP-internal setting can't be removed.
1822 $deleteable = $status['type'] !== 0; /* PHP_OUTPUT_HANDLER_INTERNAL */
1824 if ( !$deleteable ) {
1825 // Give up, and hope the result doesn't break
1829 if ( $status['name'] === 'MediaWikiTestCase::wfResetOutputBuffersBarrier' ) {
1830 // Unit testing barrier to prevent this function from breaking PHPUnit.
1833 if ( !ob_end_clean() ) {
1834 // Could not remove output buffer handler; abort now
1835 // to avoid getting in some kind of infinite loop.
1838 if ( $resetGzipEncoding ) {
1839 if ( $status['name'] == 'ob_gzhandler' ) {
1840 // Reset the 'Content-Encoding' field set by this handler
1841 // so we can start fresh.
1842 header_remove( 'Content-Encoding' );
1850 * More legible than passing a 'false' parameter to wfResetOutputBuffers():
1852 * Clear away output buffers, but keep the Content-Encoding header
1853 * produced by ob_gzhandler, if any.
1855 * This should be used for HTTP 304 responses, where you need to
1856 * preserve the Content-Encoding header of the real result, but
1857 * also need to suppress the output of ob_gzhandler to keep to spec
1858 * and avoid breaking Firefox in rare cases where the headers and
1859 * body are broken over two packets.
1861 function wfClearOutputBuffers() {
1862 wfResetOutputBuffers( false );
1866 * Converts an Accept-* header into an array mapping string values to quality
1869 * @param string $accept
1870 * @param string $def Default
1871 * @return float[] Associative array of string => float pairs
1873 function wfAcceptToPrefs( $accept, $def = '*/*' ) {
1874 # No arg means accept anything (per HTTP spec)
1876 return [ $def => 1.0 ];
1881 $parts = explode( ',', $accept );
1883 foreach ( $parts as $part ) {
1884 # @todo FIXME: Doesn't deal with params like 'text/html; level=1'
1885 $values = explode( ';', trim( $part ) );
1887 if ( count( $values ) == 1 ) {
1888 $prefs[$values[0]] = 1.0;
1889 } elseif ( preg_match( '/q\s*=\s*(\d*\.\d+)/', $values[1], $match ) ) {
1890 $prefs[$values[0]] = floatval( $match[1] );
1898 * Checks if a given MIME type matches any of the keys in the given
1899 * array. Basic wildcards are accepted in the array keys.
1901 * Returns the matching MIME type (or wildcard) if a match, otherwise
1904 * @param string $type
1905 * @param array $avail
1909 function mimeTypeMatch( $type, $avail ) {
1910 if ( array_key_exists( $type, $avail ) ) {
1913 $mainType = explode( '/', $type )[0];
1914 if ( array_key_exists( "$mainType/*", $avail ) ) {
1915 return "$mainType/*";
1916 } elseif ( array_key_exists( '*/*', $avail ) ) {
1925 * Returns the 'best' match between a client's requested internet media types
1926 * and the server's list of available types. Each list should be an associative
1927 * array of type to preference (preference is a float between 0.0 and 1.0).
1928 * Wildcards in the types are acceptable.
1930 * @param array $cprefs Client's acceptable type list
1931 * @param array $sprefs Server's offered types
1934 * @todo FIXME: Doesn't handle params like 'text/plain; charset=UTF-8'
1935 * XXX: generalize to negotiate other stuff
1937 function wfNegotiateType( $cprefs, $sprefs ) {
1940 foreach ( array_keys( $sprefs ) as $type ) {
1941 $subType = explode( '/', $type )[1];
1942 if ( $subType != '*' ) {
1943 $ckey = mimeTypeMatch( $type, $cprefs );
1945 $combine[$type] = $sprefs[$type] * $cprefs[$ckey];
1950 foreach ( array_keys( $cprefs ) as $type ) {
1951 $subType = explode( '/', $type )[1];
1952 if ( $subType != '*' && !array_key_exists( $type, $sprefs ) ) {
1953 $skey = mimeTypeMatch( $type, $sprefs );
1955 $combine[$type] = $sprefs[$skey] * $cprefs[$type];
1963 foreach ( array_keys( $combine ) as $type ) {
1964 if ( $combine[$type] > $bestq ) {
1966 $bestq = $combine[$type];
1974 * Reference-counted warning suppression
1976 * @deprecated since 1.26, use MediaWiki\suppressWarnings() directly
1979 function wfSuppressWarnings( $end = false ) {
1980 MediaWiki\
suppressWarnings( $end );
1984 * @deprecated since 1.26, use MediaWiki\restoreWarnings() directly
1985 * Restore error level to previous value
1987 function wfRestoreWarnings() {
1988 MediaWiki\
suppressWarnings( true );
1991 # Autodetect, convert and provide timestamps of various types
1994 * Unix time - the number of seconds since 1970-01-01 00:00:00 UTC
1996 define( 'TS_UNIX', 0 );
1999 * MediaWiki concatenated string timestamp (YYYYMMDDHHMMSS)
2001 define( 'TS_MW', 1 );
2004 * MySQL DATETIME (YYYY-MM-DD HH:MM:SS)
2006 define( 'TS_DB', 2 );
2009 * RFC 2822 format, for E-mail and HTTP headers
2011 define( 'TS_RFC2822', 3 );
2014 * ISO 8601 format with no timezone: 1986-02-09T20:00:00Z
2016 * This is used by Special:Export
2018 define( 'TS_ISO_8601', 4 );
2021 * An Exif timestamp (YYYY:MM:DD HH:MM:SS)
2023 * @see http://exif.org/Exif2-2.PDF The Exif 2.2 spec, see page 28 for the
2024 * DateTime tag and page 36 for the DateTimeOriginal and
2025 * DateTimeDigitized tags.
2027 define( 'TS_EXIF', 5 );
2030 * Oracle format time.
2032 define( 'TS_ORACLE', 6 );
2035 * Postgres format time.
2037 define( 'TS_POSTGRES', 7 );
2040 * ISO 8601 basic format with no timezone: 19860209T200000Z. This is used by ResourceLoader
2042 define( 'TS_ISO_8601_BASIC', 9 );
2045 * Get a timestamp string in one of various formats
2047 * @param mixed $outputtype A timestamp in one of the supported formats, the
2048 * function will autodetect which format is supplied and act accordingly.
2049 * @param mixed $ts Optional timestamp to convert, default 0 for the current time
2050 * @return string|bool String / false The same date in the format specified in $outputtype or false
2052 function wfTimestamp( $outputtype = TS_UNIX
, $ts = 0 ) {
2054 $timestamp = new MWTimestamp( $ts );
2055 return $timestamp->getTimestamp( $outputtype );
2056 } catch ( TimestampException
$e ) {
2057 wfDebug( "wfTimestamp() fed bogus time value: TYPE=$outputtype; VALUE=$ts\n" );
2063 * Return a formatted timestamp, or null if input is null.
2064 * For dealing with nullable timestamp columns in the database.
2066 * @param int $outputtype
2070 function wfTimestampOrNull( $outputtype = TS_UNIX
, $ts = null ) {
2071 if ( is_null( $ts ) ) {
2074 return wfTimestamp( $outputtype, $ts );
2079 * Convenience function; returns MediaWiki timestamp for the present time.
2083 function wfTimestampNow() {
2085 return wfTimestamp( TS_MW
, time() );
2089 * Check if the operating system is Windows
2091 * @return bool True if it's Windows, false otherwise.
2093 function wfIsWindows() {
2094 static $isWindows = null;
2095 if ( $isWindows === null ) {
2096 $isWindows = strtoupper( substr( PHP_OS
, 0, 3 ) ) === 'WIN';
2102 * Check if we are running under HHVM
2106 function wfIsHHVM() {
2107 return defined( 'HHVM_VERSION' );
2111 * Tries to get the system directory for temporary files. First
2112 * $wgTmpDirectory is checked, and then the TMPDIR, TMP, and TEMP
2113 * environment variables are then checked in sequence, then
2114 * sys_get_temp_dir(), then upload_tmp_dir from php.ini.
2116 * NOTE: When possible, use instead the tmpfile() function to create
2117 * temporary files to avoid race conditions on file creation, etc.
2121 function wfTempDir() {
2122 global $wgTmpDirectory;
2124 if ( $wgTmpDirectory !== false ) {
2125 return $wgTmpDirectory;
2128 $tmpDir = array_map( "getenv", [ 'TMPDIR', 'TMP', 'TEMP' ] );
2129 $tmpDir[] = sys_get_temp_dir();
2130 $tmpDir[] = ini_get( 'upload_tmp_dir' );
2132 foreach ( $tmpDir as $tmp ) {
2133 if ( $tmp && file_exists( $tmp ) && is_dir( $tmp ) && is_writable( $tmp ) ) {
2139 * PHP on Windows will detect C:\Windows\Temp as not writable even though PHP can write to it
2140 * so create a directory within that called 'mwtmp' with a suffix of the user running the
2142 * The user is included as if various scripts are run by different users they will likely
2143 * not be able to access each others temporary files.
2145 if ( wfIsWindows() ) {
2146 $tmp = sys_get_temp_dir() . DIRECTORY_SEPARATOR
. 'mwtmp' . '-' . get_current_user();
2147 if ( !file_exists( $tmp ) ) {
2150 if ( file_exists( $tmp ) && is_dir( $tmp ) && is_writable( $tmp ) ) {
2155 throw new MWException( 'No writable temporary directory could be found. ' .
2156 'Please set $wgTmpDirectory to a writable directory.' );
2160 * Make directory, and make all parent directories if they don't exist
2162 * @param string $dir Full path to directory to create
2163 * @param int $mode Chmod value to use, default is $wgDirectoryMode
2164 * @param string $caller Optional caller param for debugging.
2165 * @throws MWException
2168 function wfMkdirParents( $dir, $mode = null, $caller = null ) {
2169 global $wgDirectoryMode;
2171 if ( FileBackend
::isStoragePath( $dir ) ) { // sanity
2172 throw new MWException( __FUNCTION__
. " given storage path '$dir'." );
2175 if ( !is_null( $caller ) ) {
2176 wfDebug( "$caller: called wfMkdirParents($dir)\n" );
2179 if ( strval( $dir ) === '' ||
is_dir( $dir ) ) {
2183 $dir = str_replace( [ '\\', '/' ], DIRECTORY_SEPARATOR
, $dir );
2185 if ( is_null( $mode ) ) {
2186 $mode = $wgDirectoryMode;
2189 // Turn off the normal warning, we're doing our own below
2190 MediaWiki\
suppressWarnings();
2191 $ok = mkdir( $dir, $mode, true ); // PHP5 <3
2192 MediaWiki\restoreWarnings
();
2195 // directory may have been created on another request since we last checked
2196 if ( is_dir( $dir ) ) {
2200 // PHP doesn't report the path in its warning message, so add our own to aid in diagnosis.
2201 wfLogWarning( sprintf( "failed to mkdir \"%s\" mode 0%o", $dir, $mode ) );
2207 * Remove a directory and all its content.
2208 * Does not hide error.
2209 * @param string $dir
2211 function wfRecursiveRemoveDir( $dir ) {
2212 wfDebug( __FUNCTION__
. "( $dir )\n" );
2213 // taken from http://de3.php.net/manual/en/function.rmdir.php#98622
2214 if ( is_dir( $dir ) ) {
2215 $objects = scandir( $dir );
2216 foreach ( $objects as $object ) {
2217 if ( $object != "." && $object != ".." ) {
2218 if ( filetype( $dir . '/' . $object ) == "dir" ) {
2219 wfRecursiveRemoveDir( $dir . '/' . $object );
2221 unlink( $dir . '/' . $object );
2231 * @param int $nr The number to format
2232 * @param int $acc The number of digits after the decimal point, default 2
2233 * @param bool $round Whether or not to round the value, default true
2236 function wfPercent( $nr, $acc = 2, $round = true ) {
2237 $ret = sprintf( "%.${acc}f", $nr );
2238 return $round ?
round( $ret, $acc ) . '%' : "$ret%";
2242 * Safety wrapper around ini_get() for boolean settings.
2243 * The values returned from ini_get() are pre-normalized for settings
2244 * set via php.ini or php_flag/php_admin_flag... but *not*
2245 * for those set via php_value/php_admin_value.
2247 * It's fairly common for people to use php_value instead of php_flag,
2248 * which can leave you with an 'off' setting giving a false positive
2249 * for code that just takes the ini_get() return value as a boolean.
2251 * To make things extra interesting, setting via php_value accepts
2252 * "true" and "yes" as true, but php.ini and php_flag consider them false. :)
2253 * Unrecognized values go false... again opposite PHP's own coercion
2254 * from string to bool.
2256 * Luckily, 'properly' set settings will always come back as '0' or '1',
2257 * so we only have to worry about them and the 'improper' settings.
2259 * I frickin' hate PHP... :P
2261 * @param string $setting
2264 function wfIniGetBool( $setting ) {
2265 $val = strtolower( ini_get( $setting ) );
2266 // 'on' and 'true' can't have whitespace around them, but '1' can.
2270 ||
preg_match( "/^\s*[+-]?0*[1-9]/", $val ); // approx C atoi() function
2274 * Windows-compatible version of escapeshellarg()
2275 * Windows doesn't recognise single-quotes in the shell, but the escapeshellarg()
2276 * function puts single quotes in regardless of OS.
2278 * Also fixes the locale problems on Linux in PHP 5.2.6+ (bug backported to
2279 * earlier distro releases of PHP)
2281 * @param string ... strings to escape and glue together, or a single array of strings parameter
2284 function wfEscapeShellArg( /*...*/ ) {
2285 wfInitShellLocale();
2287 $args = func_get_args();
2288 if ( count( $args ) === 1 && is_array( reset( $args ) ) ) {
2289 // If only one argument has been passed, and that argument is an array,
2290 // treat it as a list of arguments
2291 $args = reset( $args );
2296 foreach ( $args as $arg ) {
2303 if ( wfIsWindows() ) {
2304 // Escaping for an MSVC-style command line parser and CMD.EXE
2305 // @codingStandardsIgnoreStart For long URLs
2307 // * http://web.archive.org/web/20020708081031/http://mailman.lyra.org/pipermail/scite-interest/2002-March/000436.html
2308 // * http://technet.microsoft.com/en-us/library/cc723564.aspx
2311 // Double the backslashes before any double quotes. Escape the double quotes.
2312 // @codingStandardsIgnoreEnd
2313 $tokens = preg_split( '/(\\\\*")/', $arg, -1, PREG_SPLIT_DELIM_CAPTURE
);
2316 foreach ( $tokens as $token ) {
2317 if ( $iteration %
2 == 1 ) {
2318 // Delimiter, a double quote preceded by zero or more slashes
2319 $arg .= str_replace( '\\', '\\\\', substr( $token, 0, -1 ) ) . '\\"';
2320 } elseif ( $iteration %
4 == 2 ) {
2321 // ^ in $token will be outside quotes, need to be escaped
2322 $arg .= str_replace( '^', '^^', $token );
2323 } else { // $iteration % 4 == 0
2324 // ^ in $token will appear inside double quotes, so leave as is
2329 // Double the backslashes before the end of the string, because
2330 // we will soon add a quote
2332 if ( preg_match( '/^(.*?)(\\\\+)$/', $arg, $m ) ) {
2333 $arg = $m[1] . str_replace( '\\', '\\\\', $m[2] );
2336 // Add surrounding quotes
2337 $retVal .= '"' . $arg . '"';
2339 $retVal .= escapeshellarg( $arg );
2346 * Check if wfShellExec() is effectively disabled via php.ini config
2348 * @return bool|string False or 'disabled'
2351 function wfShellExecDisabled() {
2352 static $disabled = null;
2353 if ( is_null( $disabled ) ) {
2354 if ( !function_exists( 'proc_open' ) ) {
2355 wfDebug( "proc_open() is disabled\n" );
2356 $disabled = 'disabled';
2365 * Execute a shell command, with time and memory limits mirrored from the PHP
2366 * configuration if supported.
2368 * @param string|string[] $cmd If string, a properly shell-escaped command line,
2369 * or an array of unescaped arguments, in which case each value will be escaped
2370 * Example: [ 'convert', '-font', 'font name' ] would produce "'convert' '-font' 'font name'"
2371 * @param null|mixed &$retval Optional, will receive the program's exit code.
2372 * (non-zero is usually failure). If there is an error from
2373 * read, select, or proc_open(), this will be set to -1.
2374 * @param array $environ Optional environment variables which should be
2375 * added to the executed command environment.
2376 * @param array $limits Optional array with limits(filesize, memory, time, walltime)
2377 * this overwrites the global wgMaxShell* limits.
2378 * @param array $options Array of options:
2379 * - duplicateStderr: Set this to true to duplicate stderr to stdout,
2380 * including errors from limit.sh
2381 * - profileMethod: By default this function will profile based on the calling
2382 * method. Set this to a string for an alternative method to profile from
2384 * @return string Collected stdout as a string
2386 function wfShellExec( $cmd, &$retval = null, $environ = [],
2387 $limits = [], $options = []
2389 global $IP, $wgMaxShellMemory, $wgMaxShellFileSize, $wgMaxShellTime,
2390 $wgMaxShellWallClockTime, $wgShellCgroup;
2392 $disabled = wfShellExecDisabled();
2395 return 'Unable to run external programs, proc_open() is disabled.';
2398 $includeStderr = isset( $options['duplicateStderr'] ) && $options['duplicateStderr'];
2399 $profileMethod = isset( $options['profileMethod'] ) ?
$options['profileMethod'] : wfGetCaller();
2401 wfInitShellLocale();
2404 foreach ( $environ as $k => $v ) {
2405 if ( wfIsWindows() ) {
2406 /* Surrounding a set in quotes (method used by wfEscapeShellArg) makes the quotes themselves
2407 * appear in the environment variable, so we must use carat escaping as documented in
2408 * http://technet.microsoft.com/en-us/library/cc723564.aspx
2409 * Note however that the quote isn't listed there, but is needed, and the parentheses
2410 * are listed there but doesn't appear to need it.
2412 $envcmd .= "set $k=" . preg_replace( '/([&|()<>^"])/', '^\\1', $v ) . '&& ';
2414 /* Assume this is a POSIX shell, thus required to accept variable assignments before the command
2415 * http://www.opengroup.org/onlinepubs/009695399/utilities/xcu_chap02.html#tag_02_09_01
2417 $envcmd .= "$k=" . escapeshellarg( $v ) . ' ';
2420 if ( is_array( $cmd ) ) {
2421 $cmd = wfEscapeShellArg( $cmd );
2424 $cmd = $envcmd . $cmd;
2426 $useLogPipe = false;
2427 if ( is_executable( '/bin/bash' ) ) {
2428 $time = intval( isset( $limits['time'] ) ?
$limits['time'] : $wgMaxShellTime );
2429 if ( isset( $limits['walltime'] ) ) {
2430 $wallTime = intval( $limits['walltime'] );
2431 } elseif ( isset( $limits['time'] ) ) {
2434 $wallTime = intval( $wgMaxShellWallClockTime );
2436 $mem = intval( isset( $limits['memory'] ) ?
$limits['memory'] : $wgMaxShellMemory );
2437 $filesize = intval( isset( $limits['filesize'] ) ?
$limits['filesize'] : $wgMaxShellFileSize );
2439 if ( $time > 0 ||
$mem > 0 ||
$filesize > 0 ||
$wallTime > 0 ) {
2440 $cmd = '/bin/bash ' . escapeshellarg( "$IP/includes/limit.sh" ) . ' ' .
2441 escapeshellarg( $cmd ) . ' ' .
2443 "MW_INCLUDE_STDERR=" . ( $includeStderr ?
'1' : '' ) . ';' .
2444 "MW_CPU_LIMIT=$time; " .
2445 'MW_CGROUP=' . escapeshellarg( $wgShellCgroup ) . '; ' .
2446 "MW_MEM_LIMIT=$mem; " .
2447 "MW_FILE_SIZE_LIMIT=$filesize; " .
2448 "MW_WALL_CLOCK_LIMIT=$wallTime; " .
2449 "MW_USE_LOG_PIPE=yes"
2452 } elseif ( $includeStderr ) {
2455 } elseif ( $includeStderr ) {
2458 wfDebug( "wfShellExec: $cmd\n" );
2460 // Don't try to execute commands that exceed Linux's MAX_ARG_STRLEN.
2461 // Other platforms may be more accomodating, but we don't want to be
2462 // accomodating, because very long commands probably include user
2463 // input. See T129506.
2464 if ( strlen( $cmd ) > SHELL_MAX_ARG_STRLEN
) {
2465 throw new Exception( __METHOD__
.
2466 '(): total length of $cmd must not exceed SHELL_MAX_ARG_STRLEN' );
2470 0 => [ 'file', 'php://stdin', 'r' ],
2471 1 => [ 'pipe', 'w' ],
2472 2 => [ 'file', 'php://stderr', 'w' ] ];
2473 if ( $useLogPipe ) {
2474 $desc[3] = [ 'pipe', 'w' ];
2477 $scoped = Profiler
::instance()->scopedProfileIn( __FUNCTION__
. '-' . $profileMethod );
2478 $proc = proc_open( $cmd, $desc, $pipes );
2480 wfDebugLog( 'exec', "proc_open() failed: $cmd" );
2484 $outBuffer = $logBuffer = '';
2489 /* According to the documentation, it is possible for stream_select()
2490 * to fail due to EINTR. I haven't managed to induce this in testing
2491 * despite sending various signals. If it did happen, the error
2492 * message would take the form:
2494 * stream_select(): unable to select [4]: Interrupted system call (max_fd=5)
2496 * where [4] is the value of the macro EINTR and "Interrupted system
2497 * call" is string which according to the Linux manual is "possibly"
2498 * localised according to LC_MESSAGES.
2500 $eintr = defined( 'SOCKET_EINTR' ) ? SOCKET_EINTR
: 4;
2501 $eintrMessage = "stream_select(): unable to select [$eintr]";
2507 while ( $running === true ||
$numReadyPipes !== 0 ) {
2509 $status = proc_get_status( $proc );
2510 // If the process has terminated, switch to nonblocking selects
2511 // for getting any data still waiting to be read.
2512 if ( !$status['running'] ) {
2518 $readyPipes = $pipes;
2521 // @codingStandardsIgnoreStart Generic.PHP.NoSilencedErrors.Discouraged
2522 @trigger_error
( '' );
2523 $numReadyPipes = @stream_select
( $readyPipes, $emptyArray, $emptyArray, $timeout );
2524 if ( $numReadyPipes === false ) {
2525 // @codingStandardsIgnoreEnd
2526 $error = error_get_last();
2527 if ( strncmp( $error['message'], $eintrMessage, strlen( $eintrMessage ) ) == 0 ) {
2530 trigger_error( $error['message'], E_USER_WARNING
);
2531 $logMsg = $error['message'];
2535 foreach ( $readyPipes as $fd => $pipe ) {
2536 $block = fread( $pipe, 65536 );
2537 if ( $block === '' ) {
2539 fclose( $pipes[$fd] );
2540 unset( $pipes[$fd] );
2544 } elseif ( $block === false ) {
2546 $logMsg = "Error reading from pipe";
2548 } elseif ( $fd == 1 ) {
2550 $outBuffer .= $block;
2551 } elseif ( $fd == 3 ) {
2553 $logBuffer .= $block;
2554 if ( strpos( $block, "\n" ) !== false ) {
2555 $lines = explode( "\n", $logBuffer );
2556 $logBuffer = array_pop( $lines );
2557 foreach ( $lines as $line ) {
2558 wfDebugLog( 'exec', $line );
2565 foreach ( $pipes as $pipe ) {
2569 // Use the status previously collected if possible, since proc_get_status()
2570 // just calls waitpid() which will not return anything useful the second time.
2572 $status = proc_get_status( $proc );
2575 if ( $logMsg !== false ) {
2576 // Read/select error
2578 proc_close( $proc );
2579 } elseif ( $status['signaled'] ) {
2580 $logMsg = "Exited with signal {$status['termsig']}";
2581 $retval = 128 +
$status['termsig'];
2582 proc_close( $proc );
2584 if ( $status['running'] ) {
2585 $retval = proc_close( $proc );
2587 $retval = $status['exitcode'];
2588 proc_close( $proc );
2590 if ( $retval == 127 ) {
2591 $logMsg = "Possibly missing executable file";
2592 } elseif ( $retval >= 129 && $retval <= 192 ) {
2593 $logMsg = "Probably exited with signal " . ( $retval - 128 );
2597 if ( $logMsg !== false ) {
2598 wfDebugLog( 'exec', "$logMsg: $cmd" );
2605 * Execute a shell command, returning both stdout and stderr. Convenience
2606 * function, as all the arguments to wfShellExec can become unwieldy.
2608 * @note This also includes errors from limit.sh, e.g. if $wgMaxShellFileSize is exceeded.
2609 * @param string|string[] $cmd If string, a properly shell-escaped command line,
2610 * or an array of unescaped arguments, in which case each value will be escaped
2611 * Example: [ 'convert', '-font', 'font name' ] would produce "'convert' '-font' 'font name'"
2612 * @param null|mixed &$retval Optional, will receive the program's exit code.
2613 * (non-zero is usually failure)
2614 * @param array $environ Optional environment variables which should be
2615 * added to the executed command environment.
2616 * @param array $limits Optional array with limits(filesize, memory, time, walltime)
2617 * this overwrites the global wgMaxShell* limits.
2618 * @return string Collected stdout and stderr as a string
2620 function wfShellExecWithStderr( $cmd, &$retval = null, $environ = [], $limits = [] ) {
2621 return wfShellExec( $cmd, $retval, $environ, $limits,
2622 [ 'duplicateStderr' => true, 'profileMethod' => wfGetCaller() ] );
2626 * Workaround for http://bugs.php.net/bug.php?id=45132
2627 * escapeshellarg() destroys non-ASCII characters if LANG is not a UTF-8 locale
2629 function wfInitShellLocale() {
2630 static $done = false;
2635 global $wgShellLocale;
2636 putenv( "LC_CTYPE=$wgShellLocale" );
2637 setlocale( LC_CTYPE
, $wgShellLocale );
2641 * Generate a shell-escaped command line string to run a MediaWiki cli script.
2642 * Note that $parameters should be a flat array and an option with an argument
2643 * should consist of two consecutive items in the array (do not use "--option value").
2645 * @param string $script MediaWiki cli script path
2646 * @param array $parameters Arguments and options to the script
2647 * @param array $options Associative array of options:
2648 * 'php': The path to the php executable
2649 * 'wrapper': Path to a PHP wrapper to handle the maintenance script
2652 function wfShellWikiCmd( $script, array $parameters = [], array $options = [] ) {
2654 // Give site config file a chance to run the script in a wrapper.
2655 // The caller may likely want to call wfBasename() on $script.
2656 Hooks
::run( 'wfShellWikiCmd', [ &$script, &$parameters, &$options ] );
2657 $cmd = isset( $options['php'] ) ?
[ $options['php'] ] : [ $wgPhpCli ];
2658 if ( isset( $options['wrapper'] ) ) {
2659 $cmd[] = $options['wrapper'];
2662 // Escape each parameter for shell
2663 return wfEscapeShellArg( array_merge( $cmd, $parameters ) );
2667 * wfMerge attempts to merge differences between three texts.
2668 * Returns true for a clean merge and false for failure or a conflict.
2670 * @param string $old
2671 * @param string $mine
2672 * @param string $yours
2673 * @param string $result
2676 function wfMerge( $old, $mine, $yours, &$result ) {
2679 # This check may also protect against code injection in
2680 # case of broken installations.
2681 MediaWiki\
suppressWarnings();
2682 $haveDiff3 = $wgDiff3 && file_exists( $wgDiff3 );
2683 MediaWiki\restoreWarnings
();
2685 if ( !$haveDiff3 ) {
2686 wfDebug( "diff3 not found\n" );
2690 # Make temporary files
2692 $oldtextFile = fopen( $oldtextName = tempnam( $td, 'merge-old-' ), 'w' );
2693 $mytextFile = fopen( $mytextName = tempnam( $td, 'merge-mine-' ), 'w' );
2694 $yourtextFile = fopen( $yourtextName = tempnam( $td, 'merge-your-' ), 'w' );
2696 # NOTE: diff3 issues a warning to stderr if any of the files does not end with
2697 # a newline character. To avoid this, we normalize the trailing whitespace before
2698 # creating the diff.
2700 fwrite( $oldtextFile, rtrim( $old ) . "\n" );
2701 fclose( $oldtextFile );
2702 fwrite( $mytextFile, rtrim( $mine ) . "\n" );
2703 fclose( $mytextFile );
2704 fwrite( $yourtextFile, rtrim( $yours ) . "\n" );
2705 fclose( $yourtextFile );
2707 # Check for a conflict
2708 $cmd = wfEscapeShellArg( $wgDiff3, '-a', '--overlap-only', $mytextName,
2709 $oldtextName, $yourtextName );
2710 $handle = popen( $cmd, 'r' );
2712 if ( fgets( $handle, 1024 ) ) {
2720 $cmd = wfEscapeShellArg( $wgDiff3, '-a', '-e', '--merge', $mytextName,
2721 $oldtextName, $yourtextName );
2722 $handle = popen( $cmd, 'r' );
2725 $data = fread( $handle, 8192 );
2726 if ( strlen( $data ) == 0 ) {
2732 unlink( $mytextName );
2733 unlink( $oldtextName );
2734 unlink( $yourtextName );
2736 if ( $result === '' && $old !== '' && !$conflict ) {
2737 wfDebug( "Unexpected null result from diff3. Command: $cmd\n" );
2744 * Returns unified plain-text diff of two texts.
2745 * "Useful" for machine processing of diffs.
2747 * @deprecated since 1.25, use DiffEngine/UnifiedDiffFormatter directly
2749 * @param string $before The text before the changes.
2750 * @param string $after The text after the changes.
2751 * @param string $params Command-line options for the diff command.
2752 * @return string Unified diff of $before and $after
2754 function wfDiff( $before, $after, $params = '-u' ) {
2755 if ( $before == $after ) {
2760 MediaWiki\
suppressWarnings();
2761 $haveDiff = $wgDiff && file_exists( $wgDiff );
2762 MediaWiki\restoreWarnings
();
2764 # This check may also protect against code injection in
2765 # case of broken installations.
2767 wfDebug( "diff executable not found\n" );
2768 $diffs = new Diff( explode( "\n", $before ), explode( "\n", $after ) );
2769 $format = new UnifiedDiffFormatter();
2770 return $format->format( $diffs );
2773 # Make temporary files
2775 $oldtextFile = fopen( $oldtextName = tempnam( $td, 'merge-old-' ), 'w' );
2776 $newtextFile = fopen( $newtextName = tempnam( $td, 'merge-your-' ), 'w' );
2778 fwrite( $oldtextFile, $before );
2779 fclose( $oldtextFile );
2780 fwrite( $newtextFile, $after );
2781 fclose( $newtextFile );
2783 // Get the diff of the two files
2784 $cmd = "$wgDiff " . $params . ' ' . wfEscapeShellArg( $oldtextName, $newtextName );
2786 $h = popen( $cmd, 'r' );
2788 unlink( $oldtextName );
2789 unlink( $newtextName );
2790 throw new Exception( __METHOD__
. '(): popen() failed' );
2796 $data = fread( $h, 8192 );
2797 if ( strlen( $data ) == 0 ) {
2805 unlink( $oldtextName );
2806 unlink( $newtextName );
2808 // Kill the --- and +++ lines. They're not useful.
2809 $diff_lines = explode( "\n", $diff );
2810 if ( isset( $diff_lines[0] ) && strpos( $diff_lines[0], '---' ) === 0 ) {
2811 unset( $diff_lines[0] );
2813 if ( isset( $diff_lines[1] ) && strpos( $diff_lines[1], '+++' ) === 0 ) {
2814 unset( $diff_lines[1] );
2817 $diff = implode( "\n", $diff_lines );
2823 * This function works like "use VERSION" in Perl, the program will die with a
2824 * backtrace if the current version of PHP is less than the version provided
2826 * This is useful for extensions which due to their nature are not kept in sync
2827 * with releases, and might depend on other versions of PHP than the main code
2829 * Note: PHP might die due to parsing errors in some cases before it ever
2830 * manages to call this function, such is life
2832 * @see perldoc -f use
2834 * @param string|int|float $req_ver The version to check, can be a string, an integer, or a float
2835 * @throws MWException
2837 function wfUsePHP( $req_ver ) {
2838 $php_ver = PHP_VERSION
;
2840 if ( version_compare( $php_ver, (string)$req_ver, '<' ) ) {
2841 throw new MWException( "PHP $req_ver required--this is only $php_ver" );
2846 * This function works like "use VERSION" in Perl except it checks the version
2847 * of MediaWiki, the program will die with a backtrace if the current version
2848 * of MediaWiki is less than the version provided.
2850 * This is useful for extensions which due to their nature are not kept in sync
2853 * Note: Due to the behavior of PHP's version_compare() which is used in this
2854 * function, if you want to allow the 'wmf' development versions add a 'c' (or
2855 * any single letter other than 'a', 'b' or 'p') as a post-fix to your
2856 * targeted version number. For example if you wanted to allow any variation
2857 * of 1.22 use `wfUseMW( '1.22c' )`. Using an 'a' or 'b' instead of 'c' will
2858 * not result in the same comparison due to the internal logic of
2859 * version_compare().
2861 * @see perldoc -f use
2863 * @deprecated since 1.26, use the "requires' property of extension.json
2864 * @param string|int|float $req_ver The version to check, can be a string, an integer, or a float
2865 * @throws MWException
2867 function wfUseMW( $req_ver ) {
2870 if ( version_compare( $wgVersion, (string)$req_ver, '<' ) ) {
2871 throw new MWException( "MediaWiki $req_ver required--this is only $wgVersion" );
2876 * Return the final portion of a pathname.
2877 * Reimplemented because PHP5's "basename()" is buggy with multibyte text.
2878 * http://bugs.php.net/bug.php?id=33898
2880 * PHP's basename() only considers '\' a pathchar on Windows and Netware.
2881 * We'll consider it so always, as we don't want '\s' in our Unix paths either.
2883 * @param string $path
2884 * @param string $suffix String to remove if present
2887 function wfBaseName( $path, $suffix = '' ) {
2888 if ( $suffix == '' ) {
2891 $encSuffix = '(?:' . preg_quote( $suffix, '#' ) . ')?';
2895 if ( preg_match( "#([^/\\\\]*?){$encSuffix}[/\\\\]*$#", $path, $matches ) ) {
2903 * Generate a relative path name to the given file.
2904 * May explode on non-matching case-insensitive paths,
2905 * funky symlinks, etc.
2907 * @param string $path Absolute destination path including target filename
2908 * @param string $from Absolute source path, directory only
2911 function wfRelativePath( $path, $from ) {
2912 // Normalize mixed input on Windows...
2913 $path = str_replace( '/', DIRECTORY_SEPARATOR
, $path );
2914 $from = str_replace( '/', DIRECTORY_SEPARATOR
, $from );
2916 // Trim trailing slashes -- fix for drive root
2917 $path = rtrim( $path, DIRECTORY_SEPARATOR
);
2918 $from = rtrim( $from, DIRECTORY_SEPARATOR
);
2920 $pieces = explode( DIRECTORY_SEPARATOR
, dirname( $path ) );
2921 $against = explode( DIRECTORY_SEPARATOR
, $from );
2923 if ( $pieces[0] !== $against[0] ) {
2924 // Non-matching Windows drive letters?
2925 // Return a full path.
2929 // Trim off common prefix
2930 while ( count( $pieces ) && count( $against )
2931 && $pieces[0] == $against[0] ) {
2932 array_shift( $pieces );
2933 array_shift( $against );
2936 // relative dots to bump us to the parent
2937 while ( count( $against ) ) {
2938 array_unshift( $pieces, '..' );
2939 array_shift( $against );
2942 array_push( $pieces, wfBaseName( $path ) );
2944 return implode( DIRECTORY_SEPARATOR
, $pieces );
2948 * Convert an arbitrarily-long digit string from one numeric base
2949 * to another, optionally zero-padding to a minimum column width.
2951 * Supports base 2 through 36; digit values 10-36 are represented
2952 * as lowercase letters a-z. Input is case-insensitive.
2954 * @deprecated since 1.27 Use Wikimedia\base_convert() directly
2956 * @param string $input Input number
2957 * @param int $sourceBase Base of the input number
2958 * @param int $destBase Desired base of the output
2959 * @param int $pad Minimum number of digits in the output (pad with zeroes)
2960 * @param bool $lowercase Whether to output in lowercase or uppercase
2961 * @param string $engine Either "gmp", "bcmath", or "php"
2962 * @return string|bool The output number as a string, or false on error
2964 function wfBaseConvert( $input, $sourceBase, $destBase, $pad = 1,
2965 $lowercase = true, $engine = 'auto'
2967 return Wikimedia\base_convert
( $input, $sourceBase, $destBase, $pad, $lowercase, $engine );
2971 * @deprecated since 1.27, PHP's session generation isn't used with
2972 * MediaWiki\Session\SessionManager
2974 function wfFixSessionID() {
2975 wfDeprecated( __FUNCTION__
, '1.27' );
2979 * Reset the session id
2981 * @deprecated since 1.27, use MediaWiki\Session\SessionManager instead
2984 function wfResetSessionID() {
2985 wfDeprecated( __FUNCTION__
, '1.27' );
2986 $session = SessionManager
::getGlobalSession();
2987 $delay = $session->delaySave();
2989 $session->resetId();
2991 // Make sure a session is started, since that's what the old
2992 // wfResetSessionID() did.
2993 if ( session_id() !== $session->getId() ) {
2994 wfSetupSession( $session->getId() );
2997 ScopedCallback
::consume( $delay );
3001 * Initialise php session
3003 * @deprecated since 1.27, use MediaWiki\Session\SessionManager instead.
3004 * Generally, "using" SessionManager will be calling ->getSessionById() or
3005 * ::getGlobalSession() (depending on whether you were passing $sessionId
3006 * here), then calling $session->persist().
3007 * @param bool|string $sessionId
3009 function wfSetupSession( $sessionId = false ) {
3010 wfDeprecated( __FUNCTION__
, '1.27' );
3013 session_id( $sessionId );
3016 $session = SessionManager
::getGlobalSession();
3017 $session->persist();
3019 if ( session_id() !== $session->getId() ) {
3020 session_id( $session->getId() );
3022 MediaWiki\
quietCall( 'session_start' );
3026 * Get an object from the precompiled serialized directory
3028 * @param string $name
3029 * @return mixed The variable on success, false on failure
3031 function wfGetPrecompiledData( $name ) {
3034 $file = "$IP/serialized/$name";
3035 if ( file_exists( $file ) ) {
3036 $blob = file_get_contents( $file );
3038 return unserialize( $blob );
3045 * Make a cache key for the local wiki.
3047 * @param string $args,...
3050 function wfMemcKey( /*...*/ ) {
3051 return call_user_func_array(
3052 [ ObjectCache
::getLocalClusterInstance(), 'makeKey' ],
3058 * Make a cache key for a foreign DB.
3060 * Must match what wfMemcKey() would produce in context of the foreign wiki.
3063 * @param string $prefix
3064 * @param string $args,...
3067 function wfForeignMemcKey( $db, $prefix /*...*/ ) {
3068 $args = array_slice( func_get_args(), 2 );
3069 $keyspace = $prefix ?
"$db-$prefix" : $db;
3070 return call_user_func_array(
3071 [ ObjectCache
::getLocalClusterInstance(), 'makeKeyInternal' ],
3072 [ $keyspace, $args ]
3077 * Make a cache key with database-agnostic prefix.
3079 * Doesn't have a wiki-specific namespace. Uses a generic 'global' prefix
3080 * instead. Must have a prefix as otherwise keys that use a database name
3081 * in the first segment will clash with wfMemcKey/wfForeignMemcKey.
3084 * @param string $args,...
3087 function wfGlobalCacheKey( /*...*/ ) {
3088 return call_user_func_array(
3089 [ ObjectCache
::getLocalClusterInstance(), 'makeGlobalKey' ],
3095 * Get an ASCII string identifying this wiki
3096 * This is used as a prefix in memcached keys
3100 function wfWikiID() {
3101 global $wgDBprefix, $wgDBname;
3102 if ( $wgDBprefix ) {
3103 return "$wgDBname-$wgDBprefix";
3110 * Split a wiki ID into DB name and table prefix
3112 * @param string $wiki
3116 function wfSplitWikiID( $wiki ) {
3117 $bits = explode( '-', $wiki, 2 );
3118 if ( count( $bits ) < 2 ) {
3125 * Get a Database object.
3127 * @param int $db Index of the connection to get. May be DB_MASTER for the
3128 * master (for write queries), DB_REPLICA for potentially lagged read
3129 * queries, or an integer >= 0 for a particular server.
3131 * @param string|string[] $groups Query groups. An array of group names that this query
3132 * belongs to. May contain a single string if the query is only
3135 * @param string|bool $wiki The wiki ID, or false for the current wiki
3137 * Note: multiple calls to wfGetDB(DB_REPLICA) during the course of one request
3138 * will always return the same object, unless the underlying connection or load
3139 * balancer is manually destroyed.
3141 * Note 2: use $this->getDB() in maintenance scripts that may be invoked by
3142 * updater to ensure that a proper database is being updated.
3144 * @todo Replace calls to wfGetDB with calls to LoadBalancer::getConnection()
3145 * on an injected instance of LoadBalancer.
3147 * @return DatabaseBase
3149 function wfGetDB( $db, $groups = [], $wiki = false ) {
3150 return wfGetLB( $wiki )->getConnection( $db, $groups, $wiki );
3154 * Get a load balancer object.
3156 * @deprecated since 1.27, use MediaWikiServices::getDBLoadBalancer()
3157 * or MediaWikiServices::getDBLoadBalancerFactory() instead.
3159 * @param string|bool $wiki Wiki ID, or false for the current wiki
3160 * @return LoadBalancer
3162 function wfGetLB( $wiki = false ) {
3163 if ( $wiki === false ) {
3164 return \MediaWiki\MediaWikiServices
::getInstance()->getDBLoadBalancer();
3166 $factory = \MediaWiki\MediaWikiServices
::getInstance()->getDBLoadBalancerFactory();
3167 return $factory->getMainLB( $wiki );
3172 * Get the load balancer factory object
3174 * @deprecated since 1.27, use MediaWikiServices::getDBLoadBalancerFactory() instead.
3178 function wfGetLBFactory() {
3179 return \MediaWiki\MediaWikiServices
::getInstance()->getDBLoadBalancerFactory();
3184 * Shortcut for RepoGroup::singleton()->findFile()
3186 * @param string $title String or Title object
3187 * @param array $options Associative array of options (see RepoGroup::findFile)
3188 * @return File|bool File, or false if the file does not exist
3190 function wfFindFile( $title, $options = [] ) {
3191 return RepoGroup
::singleton()->findFile( $title, $options );
3195 * Get an object referring to a locally registered file.
3196 * Returns a valid placeholder object if the file does not exist.
3198 * @param Title|string $title
3199 * @return LocalFile|null A File, or null if passed an invalid Title
3201 function wfLocalFile( $title ) {
3202 return RepoGroup
::singleton()->getLocalRepo()->newFile( $title );
3206 * Should low-performance queries be disabled?
3209 * @codeCoverageIgnore
3211 function wfQueriesMustScale() {
3212 global $wgMiserMode;
3214 ||
( SiteStats
::pages() > 100000
3215 && SiteStats
::edits() > 1000000
3216 && SiteStats
::users() > 10000 );
3220 * Get the path to a specified script file, respecting file
3221 * extensions; this is a wrapper around $wgScriptPath etc.
3222 * except for 'index' and 'load' which use $wgScript/$wgLoadScript
3224 * @param string $script Script filename, sans extension
3227 function wfScript( $script = 'index' ) {
3228 global $wgScriptPath, $wgScript, $wgLoadScript;
3229 if ( $script === 'index' ) {
3231 } elseif ( $script === 'load' ) {
3232 return $wgLoadScript;
3234 return "{$wgScriptPath}/{$script}.php";
3239 * Get the script URL.
3241 * @return string Script URL
3243 function wfGetScriptUrl() {
3244 if ( isset( $_SERVER['SCRIPT_NAME'] ) ) {
3245 /* as it was called, minus the query string.
3247 * Some sites use Apache rewrite rules to handle subdomains,
3248 * and have PHP set up in a weird way that causes PHP_SELF
3249 * to contain the rewritten URL instead of the one that the
3250 * outside world sees.
3252 * If in this mode, use SCRIPT_URL instead, which mod_rewrite
3253 * provides containing the "before" URL.
3255 return $_SERVER['SCRIPT_NAME'];
3257 return $_SERVER['URL'];
3262 * Convenience function converts boolean values into "true"
3263 * or "false" (string) values
3265 * @param bool $value
3268 function wfBoolToStr( $value ) {
3269 return $value ?
'true' : 'false';
3273 * Get a platform-independent path to the null file, e.g. /dev/null
3277 function wfGetNull() {
3278 return wfIsWindows() ?
'NUL' : '/dev/null';
3282 * Waits for the replica DBs to catch up to the master position
3284 * Use this when updating very large numbers of rows, as in maintenance scripts,
3285 * to avoid causing too much lag. Of course, this is a no-op if there are no replica DBs.
3287 * By default this waits on the main DB cluster of the current wiki.
3288 * If $cluster is set to "*" it will wait on all DB clusters, including
3289 * external ones. If the lag being waiting on is caused by the code that
3290 * does this check, it makes since to use $ifWritesSince, particularly if
3291 * cluster is "*", to avoid excess overhead.
3293 * Never call this function after a big DB write that is still in a transaction.
3294 * This only makes sense after the possible lag inducing changes were committed.
3296 * @param float|null $ifWritesSince Only wait if writes were done since this UNIX timestamp
3297 * @param string|bool $wiki Wiki identifier accepted by wfGetLB
3298 * @param string|bool $cluster Cluster name accepted by LBFactory. Default: false.
3299 * @param int|null $timeout Max wait time. Default: 1 day (cli), ~10 seconds (web)
3300 * @return bool Success (able to connect and no timeouts reached)
3301 * @deprecated since 1.27 Use LBFactory::waitForReplication
3303 function wfWaitForSlaves(
3304 $ifWritesSince = null, $wiki = false, $cluster = false, $timeout = null
3306 if ( $timeout === null ) {
3307 $timeout = ( PHP_SAPI
=== 'cli' ) ?
86400 : 10;
3310 if ( $cluster === '*' ) {
3313 } elseif ( $wiki === false ) {
3318 wfGetLBFactory()->waitForReplication( [
3320 'cluster' => $cluster,
3321 'timeout' => $timeout,
3322 // B/C: first argument used to be "max seconds of lag"; ignore such values
3323 'ifWritesSince' => ( $ifWritesSince > 1e9
) ?
$ifWritesSince : null
3325 } catch ( DBReplicationWaitError
$e ) {
3333 * Count down from $seconds to zero on the terminal, with a one-second pause
3334 * between showing each number. For use in command-line scripts.
3336 * @codeCoverageIgnore
3337 * @param int $seconds
3339 function wfCountDown( $seconds ) {
3340 for ( $i = $seconds; $i >= 0; $i-- ) {
3341 if ( $i != $seconds ) {
3342 echo str_repeat( "\x08", strlen( $i +
1 ) );
3354 * Replace all invalid characters with '-'.
3355 * Additional characters can be defined in $wgIllegalFileChars (see T22489).
3356 * By default, $wgIllegalFileChars includes ':', '/', '\'.
3358 * @param string $name Filename to process
3361 function wfStripIllegalFilenameChars( $name ) {
3362 global $wgIllegalFileChars;
3363 $illegalFileChars = $wgIllegalFileChars ?
"|[" . $wgIllegalFileChars . "]" : '';
3364 $name = preg_replace(
3365 "/[^" . Title
::legalChars() . "]" . $illegalFileChars . "/",
3369 // $wgIllegalFileChars may not include '/' and '\', so we still need to do this
3370 $name = wfBaseName( $name );
3375 * Set PHP's memory limit to the larger of php.ini or $wgMemoryLimit
3377 * @return int Resulting value of the memory limit.
3379 function wfMemoryLimit() {
3380 global $wgMemoryLimit;
3381 $memlimit = wfShorthandToInteger( ini_get( 'memory_limit' ) );
3382 if ( $memlimit != -1 ) {
3383 $conflimit = wfShorthandToInteger( $wgMemoryLimit );
3384 if ( $conflimit == -1 ) {
3385 wfDebug( "Removing PHP's memory limit\n" );
3386 MediaWiki\
suppressWarnings();
3387 ini_set( 'memory_limit', $conflimit );
3388 MediaWiki\restoreWarnings
();
3390 } elseif ( $conflimit > $memlimit ) {
3391 wfDebug( "Raising PHP's memory limit to $conflimit bytes\n" );
3392 MediaWiki\
suppressWarnings();
3393 ini_set( 'memory_limit', $conflimit );
3394 MediaWiki\restoreWarnings
();
3402 * Set PHP's time limit to the larger of php.ini or $wgTransactionalTimeLimit
3404 * @return int Prior time limit
3407 function wfTransactionalTimeLimit() {
3408 global $wgTransactionalTimeLimit;
3410 $timeLimit = ini_get( 'max_execution_time' );
3411 // Note that CLI scripts use 0
3412 if ( $timeLimit > 0 && $wgTransactionalTimeLimit > $timeLimit ) {
3413 set_time_limit( $wgTransactionalTimeLimit );
3416 ignore_user_abort( true ); // ignore client disconnects
3422 * Converts shorthand byte notation to integer form
3424 * @param string $string
3425 * @param int $default Returned if $string is empty
3428 function wfShorthandToInteger( $string = '', $default = -1 ) {
3429 $string = trim( $string );
3430 if ( $string === '' ) {
3433 $last = $string[strlen( $string ) - 1];
3434 $val = intval( $string );
3439 // break intentionally missing
3443 // break intentionally missing
3453 * Get the normalised IETF language tag
3454 * See unit test for examples.
3456 * @param string $code The language code.
3457 * @return string The language code which complying with BCP 47 standards.
3459 function wfBCP47( $code ) {
3460 $codeSegment = explode( '-', $code );
3462 foreach ( $codeSegment as $segNo => $seg ) {
3463 // when previous segment is x, it is a private segment and should be lc
3464 if ( $segNo > 0 && strtolower( $codeSegment[( $segNo - 1 )] ) == 'x' ) {
3465 $codeBCP[$segNo] = strtolower( $seg );
3466 // ISO 3166 country code
3467 } elseif ( ( strlen( $seg ) == 2 ) && ( $segNo > 0 ) ) {
3468 $codeBCP[$segNo] = strtoupper( $seg );
3469 // ISO 15924 script code
3470 } elseif ( ( strlen( $seg ) == 4 ) && ( $segNo > 0 ) ) {
3471 $codeBCP[$segNo] = ucfirst( strtolower( $seg ) );
3472 // Use lowercase for other cases
3474 $codeBCP[$segNo] = strtolower( $seg );
3477 $langCode = implode( '-', $codeBCP );
3482 * Get a specific cache object.
3484 * @param int|string $cacheType A CACHE_* constants, or other key in $wgObjectCaches
3487 function wfGetCache( $cacheType ) {
3488 return ObjectCache
::getInstance( $cacheType );
3492 * Get the main cache object
3496 function wfGetMainCache() {
3497 global $wgMainCacheType;
3498 return ObjectCache
::getInstance( $wgMainCacheType );
3502 * Get the cache object used by the message cache
3506 function wfGetMessageCacheStorage() {
3507 global $wgMessageCacheType;
3508 return ObjectCache
::getInstance( $wgMessageCacheType );
3512 * Get the cache object used by the parser cache
3516 function wfGetParserCacheStorage() {
3517 global $wgParserCacheType;
3518 return ObjectCache
::getInstance( $wgParserCacheType );
3522 * Call hook functions defined in $wgHooks
3524 * @param string $event Event name
3525 * @param array $args Parameters passed to hook functions
3526 * @param string|null $deprecatedVersion Optionally mark hook as deprecated with version number
3528 * @return bool True if no handler aborted the hook
3529 * @deprecated since 1.25 - use Hooks::run
3531 function wfRunHooks( $event, array $args = [], $deprecatedVersion = null ) {
3532 return Hooks
::run( $event, $args, $deprecatedVersion );
3536 * Wrapper around php's unpack.
3538 * @param string $format The format string (See php's docs)
3539 * @param string $data A binary string of binary data
3540 * @param int|bool $length The minimum length of $data or false. This is to
3541 * prevent reading beyond the end of $data. false to disable the check.
3543 * Also be careful when using this function to read unsigned 32 bit integer
3544 * because php might make it negative.
3546 * @throws MWException If $data not long enough, or if unpack fails
3547 * @return array Associative array of the extracted data
3549 function wfUnpack( $format, $data, $length = false ) {
3550 if ( $length !== false ) {
3551 $realLen = strlen( $data );
3552 if ( $realLen < $length ) {
3553 throw new MWException( "Tried to use wfUnpack on a "
3554 . "string of length $realLen, but needed one "
3555 . "of at least length $length."
3560 MediaWiki\
suppressWarnings();
3561 $result = unpack( $format, $data );
3562 MediaWiki\restoreWarnings
();
3564 if ( $result === false ) {
3565 // If it cannot extract the packed data.
3566 throw new MWException( "unpack could not unpack binary data" );
3572 * Determine if an image exists on the 'bad image list'.
3574 * The format of MediaWiki:Bad_image_list is as follows:
3575 * * Only list items (lines starting with "*") are considered
3576 * * The first link on a line must be a link to a bad image
3577 * * Any subsequent links on the same line are considered to be exceptions,
3578 * i.e. articles where the image may occur inline.
3580 * @param string $name The image name to check
3581 * @param Title|bool $contextTitle The page on which the image occurs, if known
3582 * @param string $blacklist Wikitext of a file blacklist
3585 function wfIsBadImage( $name, $contextTitle = false, $blacklist = null ) {
3586 # Handle redirects; callers almost always hit wfFindFile() anyway,
3587 # so just use that method because it has a fast process cache.
3588 $file = wfFindFile( $name ); // get the final name
3589 $name = $file ?
$file->getTitle()->getDBkey() : $name;
3591 # Run the extension hook
3593 if ( !Hooks
::run( 'BadImage', [ $name, &$bad ] ) ) {
3597 $cache = ObjectCache
::getLocalServerInstance( 'hash' );
3598 $key = wfMemcKey( 'bad-image-list', ( $blacklist === null ) ?
'default' : md5( $blacklist ) );
3599 $badImages = $cache->get( $key );
3601 if ( $badImages === false ) { // cache miss
3602 if ( $blacklist === null ) {
3603 $blacklist = wfMessage( 'bad_image_list' )->inContentLanguage()->plain(); // site list
3605 # Build the list now
3607 $lines = explode( "\n", $blacklist );
3608 foreach ( $lines as $line ) {
3610 if ( substr( $line, 0, 1 ) !== '*' ) {
3616 if ( !preg_match_all( '/\[\[:?(.*?)\]\]/', $line, $m ) ) {
3621 $imageDBkey = false;
3622 foreach ( $m[1] as $i => $titleText ) {
3623 $title = Title
::newFromText( $titleText );
3624 if ( !is_null( $title ) ) {
3626 $imageDBkey = $title->getDBkey();
3628 $exceptions[$title->getPrefixedDBkey()] = true;
3633 if ( $imageDBkey !== false ) {
3634 $badImages[$imageDBkey] = $exceptions;
3637 $cache->set( $key, $badImages, 60 );
3640 $contextKey = $contextTitle ?
$contextTitle->getPrefixedDBkey() : false;
3641 $bad = isset( $badImages[$name] ) && !isset( $badImages[$name][$contextKey] );
3647 * Determine whether the client at a given source IP is likely to be able to
3648 * access the wiki via HTTPS.
3650 * @param string $ip The IPv4/6 address in the normal human-readable form
3653 function wfCanIPUseHTTPS( $ip ) {
3655 Hooks
::run( 'CanIPUseHTTPS', [ $ip, &$canDo ] );
3660 * Determine input string is represents as infinity
3662 * @param string $str The string to determine
3666 function wfIsInfinity( $str ) {
3667 $infinityValues = [ 'infinite', 'indefinite', 'infinity', 'never' ];
3668 return in_array( $str, $infinityValues );
3672 * Returns true if these thumbnail parameters match one that MediaWiki
3673 * requests from file description pages and/or parser output.
3675 * $params is considered non-standard if they involve a non-standard
3676 * width or any non-default parameters aside from width and page number.
3677 * The number of possible files with standard parameters is far less than
3678 * that of all combinations; rate-limiting for them can thus be more generious.
3681 * @param array $params
3683 * @since 1.24 Moved from thumb.php to GlobalFunctions in 1.25
3685 function wfThumbIsStandard( File
$file, array $params ) {
3686 global $wgThumbLimits, $wgImageLimits, $wgResponsiveImages;
3688 $multipliers = [ 1 ];
3689 if ( $wgResponsiveImages ) {
3690 // These available sizes are hardcoded currently elsewhere in MediaWiki.
3691 // @see Linker::processResponsiveImages
3692 $multipliers[] = 1.5;
3696 $handler = $file->getHandler();
3697 if ( !$handler ||
!isset( $params['width'] ) ) {
3702 if ( isset( $params['page'] ) ) {
3703 $basicParams['page'] = $params['page'];
3708 // Expand limits to account for multipliers
3709 foreach ( $multipliers as $multiplier ) {
3710 $thumbLimits = array_merge( $thumbLimits, array_map(
3711 function ( $width ) use ( $multiplier ) {
3712 return round( $width * $multiplier );
3715 $imageLimits = array_merge( $imageLimits, array_map(
3716 function ( $pair ) use ( $multiplier ) {
3718 round( $pair[0] * $multiplier ),
3719 round( $pair[1] * $multiplier ),
3725 // Check if the width matches one of $wgThumbLimits
3726 if ( in_array( $params['width'], $thumbLimits ) ) {
3727 $normalParams = $basicParams +
[ 'width' => $params['width'] ];
3728 // Append any default values to the map (e.g. "lossy", "lossless", ...)
3729 $handler->normaliseParams( $file, $normalParams );
3731 // If not, then check if the width matchs one of $wgImageLimits
3733 foreach ( $imageLimits as $pair ) {
3734 $normalParams = $basicParams +
[ 'width' => $pair[0], 'height' => $pair[1] ];
3735 // Decide whether the thumbnail should be scaled on width or height.
3736 // Also append any default values to the map (e.g. "lossy", "lossless", ...)
3737 $handler->normaliseParams( $file, $normalParams );
3738 // Check if this standard thumbnail size maps to the given width
3739 if ( $normalParams['width'] == $params['width'] ) {
3745 return false; // not standard for description pages
3749 // Check that the given values for non-page, non-width, params are just defaults
3750 foreach ( $params as $key => $value ) {
3751 if ( !isset( $normalParams[$key] ) ||
$normalParams[$key] != $value ) {
3760 * Merges two (possibly) 2 dimensional arrays into the target array ($baseArray).
3762 * Values that exist in both values will be combined with += (all values of the array
3763 * of $newValues will be added to the values of the array of $baseArray, while values,
3764 * that exists in both, the value of $baseArray will be used).
3766 * @param array $baseArray The array where you want to add the values of $newValues to
3767 * @param array $newValues An array with new values
3768 * @return array The combined array
3771 function wfArrayPlus2d( array $baseArray, array $newValues ) {
3772 // First merge items that are in both arrays
3773 foreach ( $baseArray as $name => &$groupVal ) {
3774 if ( isset( $newValues[$name] ) ) {
3775 $groupVal +
= $newValues[$name];
3778 // Now add items that didn't exist yet
3779 $baseArray +
= $newValues;