includes/GlobalFunctions.php

   1 <?php
   2 /**
   3  * Global functions used everywhere.
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  */
  22
  23 if ( !defined( 'MEDIAWIKI' ) ) {
  24         die( "This file is part of MediaWiki, it is not a valid entry point" );
  25 }
  26
  27 use Liuggio\StatsdClient\Sender\SocketSender;
  28 use MediaWiki\Logger\LoggerFactory;
  29 use MediaWiki\Session\SessionManager;
  30 use Wikimedia\ScopedCallback;
  31 use Wikimedia\Rdbms\DBReplicationWaitError;
  32
  33 // Hide compatibility functions from Doxygen
  34 /// @cond
  35 /**
  36  * Compatibility functions
  37  *
  38  * We support PHP 5.5.9 and up.
  39  * Re-implementations of newer functions or functions in non-standard
  40  * PHP extensions may be included here.
  41  */
  42
  43 // hash_equals function only exists in PHP >= 5.6.0
  44 // https://secure.php.net/hash_equals
  45 if ( !function_exists( 'hash_equals' ) ) {
  46         /**
  47          * Check whether a user-provided string is equal to a fixed-length secret string
  48          * without revealing bytes of the secret string through timing differences.
  49          *
  50          * The usual way to compare strings (PHP's === operator or the underlying memcmp()
  51          * function in C) is to compare corresponding bytes and stop at the first difference,
  52          * which would take longer for a partial match than for a complete mismatch. This
  53          * is not secure when one of the strings (e.g. an HMAC or token) must remain secret
  54          * and the other may come from an attacker. Statistical analysis of timing measurements
  55          * over many requests may allow the attacker to guess the string's bytes one at a time
  56          * (and check his guesses) even if the timing differences are extremely small.
  57          *
  58          * When making such a security-sensitive comparison, it is essential that the sequence
  59          * in which instructions are executed and memory locations are accessed not depend on
  60          * the secret string's value. HOWEVER, for simplicity, we do not attempt to minimize
  61          * the inevitable leakage of the string's length. That is generally known anyway as
  62          * a chararacteristic of the hash function used to compute the secret value.
  63          *
  64          * Longer explanation: http://www.emerose.com/timing-attacks-explained
  65          *
  66          * @codeCoverageIgnore
  67          * @param string $known_string Fixed-length secret string to compare against
  68          * @param string $user_string User-provided string
  69          * @return bool True if the strings are the same, false otherwise
  70          */
  71         function hash_equals( $known_string, $user_string ) {
  72                 // Strict type checking as in PHP's native implementation
  73                 if ( !is_string( $known_string ) ) {
  74                         trigger_error( 'hash_equals(): Expected known_string to be a string, ' .
  75                                 gettype( $known_string ) . ' given', E_USER_WARNING );
  76
  77                         return false;
  78                 }
  79
  80                 if ( !is_string( $user_string ) ) {
  81                         trigger_error( 'hash_equals(): Expected user_string to be a string, ' .
  82                                 gettype( $user_string ) . ' given', E_USER_WARNING );
  83
  84                         return false;
  85                 }
  86
  87                 $known_string_len = strlen( $known_string );
  88                 if ( $known_string_len !== strlen( $user_string ) ) {
  89                         return false;
  90                 }
  91
  92                 $result = 0;
  93                 for ( $i = 0; $i < $known_string_len; $i++ ) {
  94                         $result |= ord( $known_string[$i] ) ^ ord( $user_string[$i] );
  95                 }
  96
  97                 return ( $result === 0 );
  98         }
  99 }
 100 /// @endcond
 101
 102 /**
 103  * Load an extension
 104  *
 105  * This queues an extension to be loaded through
 106  * the ExtensionRegistry system.
 107  *
 108  * @param string $ext Name of the extension to load
 109  * @param string|null $path Absolute path of where to find the extension.json file
 110  * @since 1.25
 111  */
 112 function wfLoadExtension( $ext, $path = null ) {
 113         if ( !$path ) {
 114                 global $wgExtensionDirectory;
 115                 $path = "$wgExtensionDirectory/$ext/extension.json";
 116         }
 117         ExtensionRegistry::getInstance()->queue( $path );
 118 }
 119
 120 /**
 121  * Load multiple extensions at once
 122  *
 123  * Same as wfLoadExtension, but more efficient if you
 124  * are loading multiple extensions.
 125  *
 126  * If you want to specify custom paths, you should interact with
 127  * ExtensionRegistry directly.
 128  *
 129  * @see wfLoadExtension
 130  * @param string[] $exts Array of extension names to load
 131  * @since 1.25
 132  */
 133 function wfLoadExtensions( array $exts ) {
 134         global $wgExtensionDirectory;
 135         $registry = ExtensionRegistry::getInstance();
 136         foreach ( $exts as $ext ) {
 137                 $registry->queue( "$wgExtensionDirectory/$ext/extension.json" );
 138         }
 139 }
 140
 141 /**
 142  * Load a skin
 143  *
 144  * @see wfLoadExtension
 145  * @param string $skin Name of the extension to load
 146  * @param string|null $path Absolute path of where to find the skin.json file
 147  * @since 1.25
 148  */
 149 function wfLoadSkin( $skin, $path = null ) {
 150         if ( !$path ) {
 151                 global $wgStyleDirectory;
 152                 $path = "$wgStyleDirectory/$skin/skin.json";
 153         }
 154         ExtensionRegistry::getInstance()->queue( $path );
 155 }
 156
 157 /**
 158  * Load multiple skins at once
 159  *
 160  * @see wfLoadExtensions
 161  * @param string[] $skins Array of extension names to load
 162  * @since 1.25
 163  */
 164 function wfLoadSkins( array $skins ) {
 165         global $wgStyleDirectory;
 166         $registry = ExtensionRegistry::getInstance();
 167         foreach ( $skins as $skin ) {
 168                 $registry->queue( "$wgStyleDirectory/$skin/skin.json" );
 169         }
 170 }
 171
 172 /**
 173  * Like array_diff( $a, $b ) except that it works with two-dimensional arrays.
 174  * @param array $a
 175  * @param array $b
 176  * @return array
 177  */
 178 function wfArrayDiff2( $a, $b ) {
 179         return array_udiff( $a, $b, 'wfArrayDiff2_cmp' );
 180 }
 181
 182 /**
 183  * @param array|string $a
 184  * @param array|string $b
 185  * @return int
 186  */
 187 function wfArrayDiff2_cmp( $a, $b ) {
 188         if ( is_string( $a ) && is_string( $b ) ) {
 189                 return strcmp( $a, $b );
 190         } elseif ( count( $a ) !== count( $b ) ) {
 191                 return count( $a ) < count( $b ) ? -1 : 1;
 192         } else {
 193                 reset( $a );
 194                 reset( $b );
 195                 while ( ( list( , $valueA ) = each( $a ) ) && ( list( , $valueB ) = each( $b ) ) ) {
 196                         $cmp = strcmp( $valueA, $valueB );
 197                         if ( $cmp !== 0 ) {
 198                                 return $cmp;
 199                         }
 200                 }
 201                 return 0;
 202         }
 203 }
 204
 205 /**
 206  * Appends to second array if $value differs from that in $default
 207  *
 208  * @param string|int $key
 209  * @param mixed $value
 210  * @param mixed $default
 211  * @param array $changed Array to alter
 212  * @throws MWException
 213  */
 214 function wfAppendToArrayIfNotDefault( $key, $value, $default, &$changed ) {
 215         if ( is_null( $changed ) ) {
 216                 throw new MWException( 'GlobalFunctions::wfAppendToArrayIfNotDefault got null' );
 217         }
 218         if ( $default[$key] !== $value ) {
 219                 $changed[$key] = $value;
 220         }
 221 }
 222
 223 /**
 224  * Merge arrays in the style of getUserPermissionsErrors, with duplicate removal
 225  * e.g.
 226  *     wfMergeErrorArrays(
 227  *       [ [ 'x' ] ],
 228  *       [ [ 'x', '2' ] ],
 229  *       [ [ 'x' ] ],
 230  *       [ [ 'y' ] ]
 231  *     );
 232  * returns:
 233  *     [
 234  *       [ 'x', '2' ],
 235  *       [ 'x' ],
 236  *       [ 'y' ]
 237  *     ]
 238  *
 239  * @param array $array1,...
 240  * @return array
 241  */
 242 function wfMergeErrorArrays( /*...*/ ) {
 243         $args = func_get_args();
 244         $out = [];
 245         foreach ( $args as $errors ) {
 246                 foreach ( $errors as $params ) {
 247                         $originalParams = $params;
 248                         if ( $params[0] instanceof MessageSpecifier ) {
 249                                 $msg = $params[0];
 250                                 $params = array_merge( [ $msg->getKey() ], $msg->getParams() );
 251                         }
 252                         # @todo FIXME: Sometimes get nested arrays for $params,
 253                         # which leads to E_NOTICEs
 254                         $spec = implode( "\t", $params );
 255                         $out[$spec] = $originalParams;
 256                 }
 257         }
 258         return array_values( $out );
 259 }
 260
 261 /**
 262  * Insert array into another array after the specified *KEY*
 263  *
 264  * @param array $array The array.
 265  * @param array $insert The array to insert.
 266  * @param mixed $after The key to insert after
 267  * @return array
 268  */
 269 function wfArrayInsertAfter( array $array, array $insert, $after ) {
 270         // Find the offset of the element to insert after.
 271         $keys = array_keys( $array );
 272         $offsetByKey = array_flip( $keys );
 273
 274         $offset = $offsetByKey[$after];
 275
 276         // Insert at the specified offset
 277         $before = array_slice( $array, 0, $offset + 1, true );
 278         $after = array_slice( $array, $offset + 1, count( $array ) - $offset, true );
 279
 280         $output = $before + $insert + $after;
 281
 282         return $output;
 283 }
 284
 285 /**
 286  * Recursively converts the parameter (an object) to an array with the same data
 287  *
 288  * @param object|array $objOrArray
 289  * @param bool $recursive
 290  * @return array
 291  */
 292 function wfObjectToArray( $objOrArray, $recursive = true ) {
 293         $array = [];
 294         if ( is_object( $objOrArray ) ) {
 295                 $objOrArray = get_object_vars( $objOrArray );
 296         }
 297         foreach ( $objOrArray as $key => $value ) {
 298                 if ( $recursive && ( is_object( $value ) || is_array( $value ) ) ) {
 299                         $value = wfObjectToArray( $value );
 300                 }
 301
 302                 $array[$key] = $value;
 303         }
 304
 305         return $array;
 306 }
 307
 308 /**
 309  * Get a random decimal value between 0 and 1, in a way
 310  * not likely to give duplicate values for any realistic
 311  * number of articles.
 312  *
 313  * @note This is designed for use in relation to Special:RandomPage
 314  *       and the page_random database field.
 315  *
 316  * @return string
 317  */
 318 function wfRandom() {
 319         // The maximum random value is "only" 2^31-1, so get two random
 320         // values to reduce the chance of dupes
 321         $max = mt_getrandmax() + 1;
 322         $rand = number_format( ( mt_rand() * $max + mt_rand() ) / $max / $max, 12, '.', '' );
 323         return $rand;
 324 }
 325
 326 /**
 327  * Get a random string containing a number of pseudo-random hex characters.
 328  *
 329  * @note This is not secure, if you are trying to generate some sort
 330  *       of token please use MWCryptRand instead.
 331  *
 332  * @param int $length The length of the string to generate
 333  * @return string
 334  * @since 1.20
 335  */
 336 function wfRandomString( $length = 32 ) {
 337         $str = '';
 338         for ( $n = 0; $n < $length; $n += 7 ) {
 339                 $str .= sprintf( '%07x', mt_rand() & 0xfffffff );
 340         }
 341         return substr( $str, 0, $length );
 342 }
 343
 344 /**
 345  * We want some things to be included as literal characters in our title URLs
 346  * for prettiness, which urlencode encodes by default.  According to RFC 1738,
 347  * all of the following should be safe:
 348  *
 349  * ;:@&=$-_.+!*'(),
 350  *
 351  * RFC 1738 says ~ is unsafe, however RFC 3986 considers it an unreserved
 352  * character which should not be encoded. More importantly, google chrome
 353  * always converts %7E back to ~, and converting it in this function can
 354  * cause a redirect loop (T105265).
 355  *
 356  * But + is not safe because it's used to indicate a space; &= are only safe in
 357  * paths and not in queries (and we don't distinguish here); ' seems kind of
 358  * scary; and urlencode() doesn't touch -_. to begin with.  Plus, although /
 359  * is reserved, we don't care.  So the list we unescape is:
 360  *
 361  * ;:@$!*(),/~
 362  *
 363  * However, IIS7 redirects fail when the url contains a colon (see T24709),
 364  * so no fancy : for IIS7.
 365  *
 366  * %2F in the page titles seems to fatally break for some reason.
 367  *
 368  * @param string $s
 369  * @return string
 370  */
 371 function wfUrlencode( $s ) {
 372         static $needle;
 373
 374         if ( is_null( $s ) ) {
 375                 $needle = null;
 376                 return '';
 377         }
 378
 379         if ( is_null( $needle ) ) {
 380                 $needle = [ '%3B', '%40', '%24', '%21', '%2A', '%28', '%29', '%2C', '%2F', '%7E' ];
 381                 if ( !isset( $_SERVER['SERVER_SOFTWARE'] ) ||
 382                         ( strpos( $_SERVER['SERVER_SOFTWARE'], 'Microsoft-IIS/7' ) === false )
 383                 ) {
 384                         $needle[] = '%3A';
 385                 }
 386         }
 387
 388         $s = urlencode( $s );
 389         $s = str_ireplace(
 390                 $needle,
 391                 [ ';', '@', '$', '!', '*', '(', ')', ',', '/', '~', ':' ],
 392                 $s
 393         );
 394
 395         return $s;
 396 }
 397
 398 /**
 399  * This function takes one or two arrays as input, and returns a CGI-style string, e.g.
 400  * "days=7&limit=100". Options in the first array override options in the second.
 401  * Options set to null or false will not be output.
 402  *
 403  * @param array $array1 ( String|Array )
 404  * @param array|null $array2 ( String|Array )
 405  * @param string $prefix
 406  * @return string
 407  */
 408 function wfArrayToCgi( $array1, $array2 = null, $prefix = '' ) {
 409         if ( !is_null( $array2 ) ) {
 410                 $array1 = $array1 + $array2;
 411         }
 412
 413         $cgi = '';
 414         foreach ( $array1 as $key => $value ) {
 415                 if ( !is_null( $value ) && $value !== false ) {
 416                         if ( $cgi != '' ) {
 417                                 $cgi .= '&';
 418                         }
 419                         if ( $prefix !== '' ) {
 420                                 $key = $prefix . "[$key]";
 421                         }
 422                         if ( is_array( $value ) ) {
 423                                 $firstTime = true;
 424                                 foreach ( $value as $k => $v ) {
 425                                         $cgi .= $firstTime ? '' : '&';
 426                                         if ( is_array( $v ) ) {
 427                                                 $cgi .= wfArrayToCgi( $v, null, $key . "[$k]" );
 428                                         } else {
 429                                                 $cgi .= urlencode( $key . "[$k]" ) . '=' . urlencode( $v );
 430                                         }
 431                                         $firstTime = false;
 432                                 }
 433                         } else {
 434                                 if ( is_object( $value ) ) {
 435                                         $value = $value->__toString();
 436                                 }
 437                                 $cgi .= urlencode( $key ) . '=' . urlencode( $value );
 438                         }
 439                 }
 440         }
 441         return $cgi;
 442 }
 443
 444 /**
 445  * This is the logical opposite of wfArrayToCgi(): it accepts a query string as
 446  * its argument and returns the same string in array form.  This allows compatibility
 447  * with legacy functions that accept raw query strings instead of nice
 448  * arrays.  Of course, keys and values are urldecode()d.
 449  *
 450  * @param string $query Query string
 451  * @return string[] Array version of input
 452  */
 453 function wfCgiToArray( $query ) {
 454         if ( isset( $query[0] ) && $query[0] == '?' ) {
 455                 $query = substr( $query, 1 );
 456         }
 457         $bits = explode( '&', $query );
 458         $ret = [];
 459         foreach ( $bits as $bit ) {
 460                 if ( $bit === '' ) {
 461                         continue;
 462                 }
 463                 if ( strpos( $bit, '=' ) === false ) {
 464                         // Pieces like &qwerty become 'qwerty' => '' (at least this is what php does)
 465                         $key = $bit;
 466                         $value = '';
 467                 } else {
 468                         list( $key, $value ) = explode( '=', $bit );
 469                 }
 470                 $key = urldecode( $key );
 471                 $value = urldecode( $value );
 472                 if ( strpos( $key, '[' ) !== false ) {
 473                         $keys = array_reverse( explode( '[', $key ) );
 474                         $key = array_pop( $keys );
 475                         $temp = $value;
 476                         foreach ( $keys as $k ) {
 477                                 $k = substr( $k, 0, -1 );
 478                                 $temp = [ $k => $temp ];
 479                         }
 480                         if ( isset( $ret[$key] ) ) {
 481                                 $ret[$key] = array_merge( $ret[$key], $temp );
 482                         } else {
 483                                 $ret[$key] = $temp;
 484                         }
 485                 } else {
 486                         $ret[$key] = $value;
 487                 }
 488         }
 489         return $ret;
 490 }
 491
 492 /**
 493  * Append a query string to an existing URL, which may or may not already
 494  * have query string parameters already. If so, they will be combined.
 495  *
 496  * @param string $url
 497  * @param string|string[] $query String or associative array
 498  * @return string
 499  */
 500 function wfAppendQuery( $url, $query ) {
 501         if ( is_array( $query ) ) {
 502                 $query = wfArrayToCgi( $query );
 503         }
 504         if ( $query != '' ) {
 505                 // Remove the fragment, if there is one
 506                 $fragment = false;
 507                 $hashPos = strpos( $url, '#' );
 508                 if ( $hashPos !== false ) {
 509                         $fragment = substr( $url, $hashPos );
 510                         $url = substr( $url, 0, $hashPos );
 511                 }
 512
 513                 // Add parameter
 514                 if ( false === strpos( $url, '?' ) ) {
 515                         $url .= '?';
 516                 } else {
 517                         $url .= '&';
 518                 }
 519                 $url .= $query;
 520
 521                 // Put the fragment back
 522                 if ( $fragment !== false ) {
 523                         $url .= $fragment;
 524                 }
 525         }
 526         return $url;
 527 }
 528
 529 /**
 530  * Expand a potentially local URL to a fully-qualified URL. Assumes $wgServer
 531  * is correct.
 532  *
 533  * The meaning of the PROTO_* constants is as follows:
 534  * PROTO_HTTP: Output a URL starting with http://
 535  * PROTO_HTTPS: Output a URL starting with https://
 536  * PROTO_RELATIVE: Output a URL starting with // (protocol-relative URL)
 537  * PROTO_CURRENT: Output a URL starting with either http:// or https:// , depending
 538  *    on which protocol was used for the current incoming request
 539  * PROTO_CANONICAL: For URLs without a domain, like /w/index.php , use $wgCanonicalServer.
 540  *    For protocol-relative URLs, use the protocol of $wgCanonicalServer
 541  * PROTO_INTERNAL: Like PROTO_CANONICAL, but uses $wgInternalServer instead of $wgCanonicalServer
 542  *
 543  * @todo this won't work with current-path-relative URLs
 544  * like "subdir/foo.html", etc.
 545  *
 546  * @param string $url Either fully-qualified or a local path + query
 547  * @param string $defaultProto One of the PROTO_* constants. Determines the
 548  *    protocol to use if $url or $wgServer is protocol-relative
 549  * @return string|false Fully-qualified URL, current-path-relative URL or false if
 550  *    no valid URL can be constructed
 551  */
 552 function wfExpandUrl( $url, $defaultProto = PROTO_CURRENT ) {
 553         global $wgServer, $wgCanonicalServer, $wgInternalServer, $wgRequest,
 554                 $wgHttpsPort;
 555         if ( $defaultProto === PROTO_CANONICAL ) {
 556                 $serverUrl = $wgCanonicalServer;
 557         } elseif ( $defaultProto === PROTO_INTERNAL && $wgInternalServer !== false ) {
 558                 // Make $wgInternalServer fall back to $wgServer if not set
 559                 $serverUrl = $wgInternalServer;
 560         } else {
 561                 $serverUrl = $wgServer;
 562                 if ( $defaultProto === PROTO_CURRENT ) {
 563                         $defaultProto = $wgRequest->getProtocol() . '://';
 564                 }
 565         }
 566
 567         // Analyze $serverUrl to obtain its protocol
 568         $bits = wfParseUrl( $serverUrl );
 569         $serverHasProto = $bits && $bits['scheme'] != '';
 570
 571         if ( $defaultProto === PROTO_CANONICAL || $defaultProto === PROTO_INTERNAL ) {
 572                 if ( $serverHasProto ) {
 573                         $defaultProto = $bits['scheme'] . '://';
 574                 } else {
 575                         // $wgCanonicalServer or $wgInternalServer doesn't have a protocol.
 576                         // This really isn't supposed to happen. Fall back to HTTP in this
 577                         // ridiculous case.
 578                         $defaultProto = PROTO_HTTP;
 579                 }
 580         }
 581
 582         $defaultProtoWithoutSlashes = substr( $defaultProto, 0, -2 );
 583
 584         if ( substr( $url, 0, 2 ) == '//' ) {
 585                 $url = $defaultProtoWithoutSlashes . $url;
 586         } elseif ( substr( $url, 0, 1 ) == '/' ) {
 587                 // If $serverUrl is protocol-relative, prepend $defaultProtoWithoutSlashes,
 588                 // otherwise leave it alone.
 589                 $url = ( $serverHasProto ? '' : $defaultProtoWithoutSlashes ) . $serverUrl . $url;
 590         }
 591
 592         $bits = wfParseUrl( $url );
 593
 594         // ensure proper port for HTTPS arrives in URL
 595         // https://phabricator.wikimedia.org/T67184
 596         if ( $defaultProto === PROTO_HTTPS && $wgHttpsPort != 443 ) {
 597                 $bits['port'] = $wgHttpsPort;
 598         }
 599
 600         if ( $bits && isset( $bits['path'] ) ) {
 601                 $bits['path'] = wfRemoveDotSegments( $bits['path'] );
 602                 return wfAssembleUrl( $bits );
 603         } elseif ( $bits ) {
 604                 # No path to expand
 605                 return $url;
 606         } elseif ( substr( $url, 0, 1 ) != '/' ) {
 607                 # URL is a relative path
 608                 return wfRemoveDotSegments( $url );
 609         }
 610
 611         # Expanded URL is not valid.
 612         return false;
 613 }
 614
 615 /**
 616  * This function will reassemble a URL parsed with wfParseURL.  This is useful
 617  * if you need to edit part of a URL and put it back together.
 618  *
 619  * This is the basic structure used (brackets contain keys for $urlParts):
 620  * [scheme][delimiter][user]:[pass]@[host]:[port][path]?[query]#[fragment]
 621  *
 622  * @todo Need to integrate this into wfExpandUrl (see T34168)
 623  *
 624  * @since 1.19
 625  * @param array $urlParts URL parts, as output from wfParseUrl
 626  * @return string URL assembled from its component parts
 627  */
 628 function wfAssembleUrl( $urlParts ) {
 629         $result = '';
 630
 631         if ( isset( $urlParts['delimiter'] ) ) {
 632                 if ( isset( $urlParts['scheme'] ) ) {
 633                         $result .= $urlParts['scheme'];
 634                 }
 635
 636                 $result .= $urlParts['delimiter'];
 637         }
 638
 639         if ( isset( $urlParts['host'] ) ) {
 640                 if ( isset( $urlParts['user'] ) ) {
 641                         $result .= $urlParts['user'];
 642                         if ( isset( $urlParts['pass'] ) ) {
 643                                 $result .= ':' . $urlParts['pass'];
 644                         }
 645                         $result .= '@';
 646                 }
 647
 648                 $result .= $urlParts['host'];
 649
 650                 if ( isset( $urlParts['port'] ) ) {
 651                         $result .= ':' . $urlParts['port'];
 652                 }
 653         }
 654
 655         if ( isset( $urlParts['path'] ) ) {
 656                 $result .= $urlParts['path'];
 657         }
 658
 659         if ( isset( $urlParts['query'] ) ) {
 660                 $result .= '?' . $urlParts['query'];
 661         }
 662
 663         if ( isset( $urlParts['fragment'] ) ) {
 664                 $result .= '#' . $urlParts['fragment'];
 665         }
 666
 667         return $result;
 668 }
 669
 670 /**
 671  * Remove all dot-segments in the provided URL path.  For example,
 672  * '/a/./b/../c/' becomes '/a/c/'.  For details on the algorithm, please see
 673  * RFC3986 section 5.2.4.
 674  *
 675  * @todo Need to integrate this into wfExpandUrl (see T34168)
 676  *
 677  * @param string $urlPath URL path, potentially containing dot-segments
 678  * @return string URL path with all dot-segments removed
 679  */
 680 function wfRemoveDotSegments( $urlPath ) {
 681         $output = '';
 682         $inputOffset = 0;
 683         $inputLength = strlen( $urlPath );
 684
 685         while ( $inputOffset < $inputLength ) {
 686                 $prefixLengthOne = substr( $urlPath, $inputOffset, 1 );
 687                 $prefixLengthTwo = substr( $urlPath, $inputOffset, 2 );
 688                 $prefixLengthThree = substr( $urlPath, $inputOffset, 3 );
 689                 $prefixLengthFour = substr( $urlPath, $inputOffset, 4 );
 690                 $trimOutput = false;
 691
 692                 if ( $prefixLengthTwo == './' ) {
 693                         # Step A, remove leading "./"
 694                         $inputOffset += 2;
 695                 } elseif ( $prefixLengthThree == '../' ) {
 696                         # Step A, remove leading "../"
 697                         $inputOffset += 3;
 698                 } elseif ( ( $prefixLengthTwo == '/.' ) && ( $inputOffset + 2 == $inputLength ) ) {
 699                         # Step B, replace leading "/.$" with "/"
 700                         $inputOffset += 1;
 701                         $urlPath[$inputOffset] = '/';
 702                 } elseif ( $prefixLengthThree == '/./' ) {
 703                         # Step B, replace leading "/./" with "/"
 704                         $inputOffset += 2;
 705                 } elseif ( $prefixLengthThree == '/..' && ( $inputOffset + 3 == $inputLength ) ) {
 706                         # Step C, replace leading "/..$" with "/" and
 707                         # remove last path component in output
 708                         $inputOffset += 2;
 709                         $urlPath[$inputOffset] = '/';
 710                         $trimOutput = true;
 711                 } elseif ( $prefixLengthFour == '/../' ) {
 712                         # Step C, replace leading "/../" with "/" and
 713                         # remove last path component in output
 714                         $inputOffset += 3;
 715                         $trimOutput = true;
 716                 } elseif ( ( $prefixLengthOne == '.' ) && ( $inputOffset + 1 == $inputLength ) ) {
 717                         # Step D, remove "^.$"
 718                         $inputOffset += 1;
 719                 } elseif ( ( $prefixLengthTwo == '..' ) && ( $inputOffset + 2 == $inputLength ) ) {
 720                         # Step D, remove "^..$"
 721                         $inputOffset += 2;
 722                 } else {
 723                         # Step E, move leading path segment to output
 724                         if ( $prefixLengthOne == '/' ) {
 725                                 $slashPos = strpos( $urlPath, '/', $inputOffset + 1 );
 726                         } else {
 727                                 $slashPos = strpos( $urlPath, '/', $inputOffset );
 728                         }
 729                         if ( $slashPos === false ) {
 730                                 $output .= substr( $urlPath, $inputOffset );
 731                                 $inputOffset = $inputLength;
 732                         } else {
 733                                 $output .= substr( $urlPath, $inputOffset, $slashPos - $inputOffset );
 734                                 $inputOffset += $slashPos - $inputOffset;
 735                         }
 736                 }
 737
 738                 if ( $trimOutput ) {
 739                         $slashPos = strrpos( $output, '/' );
 740                         if ( $slashPos === false ) {
 741                                 $output = '';
 742                         } else {
 743                                 $output = substr( $output, 0, $slashPos );
 744                         }
 745                 }
 746         }
 747
 748         return $output;
 749 }
 750
 751 /**
 752  * Returns a regular expression of url protocols
 753  *
 754  * @param bool $includeProtocolRelative If false, remove '//' from the returned protocol list.
 755  *        DO NOT USE this directly, use wfUrlProtocolsWithoutProtRel() instead
 756  * @return string
 757  */
 758 function wfUrlProtocols( $includeProtocolRelative = true ) {
 759         global $wgUrlProtocols;
 760
 761         // Cache return values separately based on $includeProtocolRelative
 762         static $withProtRel = null, $withoutProtRel = null;
 763         $cachedValue = $includeProtocolRelative ? $withProtRel : $withoutProtRel;
 764         if ( !is_null( $cachedValue ) ) {
 765                 return $cachedValue;
 766         }
 767
 768         // Support old-style $wgUrlProtocols strings, for backwards compatibility
 769         // with LocalSettings files from 1.5
 770         if ( is_array( $wgUrlProtocols ) ) {
 771                 $protocols = [];
 772                 foreach ( $wgUrlProtocols as $protocol ) {
 773                         // Filter out '//' if !$includeProtocolRelative
 774                         if ( $includeProtocolRelative || $protocol !== '//' ) {
 775                                 $protocols[] = preg_quote( $protocol, '/' );
 776                         }
 777                 }
 778
 779                 $retval = implode( '|', $protocols );
 780         } else {
 781                 // Ignore $includeProtocolRelative in this case
 782                 // This case exists for pre-1.6 compatibility, and we can safely assume
 783                 // that '//' won't appear in a pre-1.6 config because protocol-relative
 784                 // URLs weren't supported until 1.18
 785                 $retval = $wgUrlProtocols;
 786         }
 787
 788         // Cache return value
 789         if ( $includeProtocolRelative ) {
 790                 $withProtRel = $retval;
 791         } else {
 792                 $withoutProtRel = $retval;
 793         }
 794         return $retval;
 795 }
 796
 797 /**
 798  * Like wfUrlProtocols(), but excludes '//' from the protocol list. Use this if
 799  * you need a regex that matches all URL protocols but does not match protocol-
 800  * relative URLs
 801  * @return string
 802  */
 803 function wfUrlProtocolsWithoutProtRel() {
 804         return wfUrlProtocols( false );
 805 }
 806
 807 /**
 808  * parse_url() work-alike, but non-broken.  Differences:
 809  *
 810  * 1) Does not raise warnings on bad URLs (just returns false).
 811  * 2) Handles protocols that don't use :// (e.g., mailto: and news:, as well as
 812  *    protocol-relative URLs) correctly.
 813  * 3) Adds a "delimiter" element to the array, either '://', ':' or '//' (see (2)).
 814  *
 815  * @param string $url A URL to parse
 816  * @return string[]|bool Bits of the URL in an associative array, per PHP docs, false on failure
 817  */
 818 function wfParseUrl( $url ) {
 819         global $wgUrlProtocols; // Allow all protocols defined in DefaultSettings/LocalSettings.php
 820
 821         // Protocol-relative URLs are handled really badly by parse_url(). It's so
 822         // bad that the easiest way to handle them is to just prepend 'http:' and
 823         // strip the protocol out later.
 824         $wasRelative = substr( $url, 0, 2 ) == '//';
 825         if ( $wasRelative ) {
 826                 $url = "http:$url";
 827         }
 828         MediaWiki\suppressWarnings();
 829         $bits = parse_url( $url );
 830         MediaWiki\restoreWarnings();
 831         // parse_url() returns an array without scheme for some invalid URLs, e.g.
 832         // parse_url("%0Ahttp://example.com") == [ 'host' => '%0Ahttp', 'path' => 'example.com' ]
 833         if ( !$bits || !isset( $bits['scheme'] ) ) {
 834                 return false;
 835         }
 836
 837         // parse_url() incorrectly handles schemes case-sensitively. Convert it to lowercase.
 838         $bits['scheme'] = strtolower( $bits['scheme'] );
 839
 840         // most of the protocols are followed by ://, but mailto: and sometimes news: not, check for it
 841         if ( in_array( $bits['scheme'] . '://', $wgUrlProtocols ) ) {
 842                 $bits['delimiter'] = '://';
 843         } elseif ( in_array( $bits['scheme'] . ':', $wgUrlProtocols ) ) {
 844                 $bits['delimiter'] = ':';
 845                 // parse_url detects for news: and mailto: the host part of an url as path
 846                 // We have to correct this wrong detection
 847                 if ( isset( $bits['path'] ) ) {
 848                         $bits['host'] = $bits['path'];
 849                         $bits['path'] = '';
 850                 }
 851         } else {
 852                 return false;
 853         }
 854
 855         /* Provide an empty host for eg. file:/// urls (see T30627) */
 856         if ( !isset( $bits['host'] ) ) {
 857                 $bits['host'] = '';
 858
 859                 // See T47069
 860                 if ( isset( $bits['path'] ) ) {
 861                         /* parse_url loses the third / for file:///c:/ urls (but not on variants) */
 862                         if ( substr( $bits['path'], 0, 1 ) !== '/' ) {
 863                                 $bits['path'] = '/' . $bits['path'];
 864                         }
 865                 } else {
 866                         $bits['path'] = '';
 867                 }
 868         }
 869
 870         // If the URL was protocol-relative, fix scheme and delimiter
 871         if ( $wasRelative ) {
 872                 $bits['scheme'] = '';
 873                 $bits['delimiter'] = '//';
 874         }
 875         return $bits;
 876 }
 877
 878 /**
 879  * Take a URL, make sure it's expanded to fully qualified, and replace any
 880  * encoded non-ASCII Unicode characters with their UTF-8 original forms
 881  * for more compact display and legibility for local audiences.
 882  *
 883  * @todo handle punycode domains too
 884  *
 885  * @param string $url
 886  * @return string
 887  */
 888 function wfExpandIRI( $url ) {
 889         return preg_replace_callback(
 890                 '/((?:%[89A-F][0-9A-F])+)/i',
 891                 'wfExpandIRI_callback',
 892                 wfExpandUrl( $url )
 893         );
 894 }
 895
 896 /**
 897  * Private callback for wfExpandIRI
 898  * @param array $matches
 899  * @return string
 900  */
 901 function wfExpandIRI_callback( $matches ) {
 902         return urldecode( $matches[1] );
 903 }
 904
 905 /**
 906  * Make URL indexes, appropriate for the el_index field of externallinks.
 907  *
 908  * @param string $url
 909  * @return array
 910  */
 911 function wfMakeUrlIndexes( $url ) {
 912         $bits = wfParseUrl( $url );
 913
 914         // Reverse the labels in the hostname, convert to lower case
 915         // For emails reverse domainpart only
 916         if ( $bits['scheme'] == 'mailto' ) {
 917                 $mailparts = explode( '@', $bits['host'], 2 );
 918                 if ( count( $mailparts ) === 2 ) {
 919                         $domainpart = strtolower( implode( '.', array_reverse( explode( '.', $mailparts[1] ) ) ) );
 920                 } else {
 921                         // No domain specified, don't mangle it
 922                         $domainpart = '';
 923                 }
 924                 $reversedHost = $domainpart . '@' . $mailparts[0];
 925         } else {
 926                 $reversedHost = strtolower( implode( '.', array_reverse( explode( '.', $bits['host'] ) ) ) );
 927         }
 928         // Add an extra dot to the end
 929         // Why? Is it in wrong place in mailto links?
 930         if ( substr( $reversedHost, -1, 1 ) !== '.' ) {
 931                 $reversedHost .= '.';
 932         }
 933         // Reconstruct the pseudo-URL
 934         $prot = $bits['scheme'];
 935         $index = $prot . $bits['delimiter'] . $reversedHost;
 936         // Leave out user and password. Add the port, path, query and fragment
 937         if ( isset( $bits['port'] ) ) {
 938                 $index .= ':' . $bits['port'];
 939         }
 940         if ( isset( $bits['path'] ) ) {
 941                 $index .= $bits['path'];
 942         } else {
 943                 $index .= '/';
 944         }
 945         if ( isset( $bits['query'] ) ) {
 946                 $index .= '?' . $bits['query'];
 947         }
 948         if ( isset( $bits['fragment'] ) ) {
 949                 $index .= '#' . $bits['fragment'];
 950         }
 951
 952         if ( $prot == '' ) {
 953                 return [ "http:$index", "https:$index" ];
 954         } else {
 955                 return [ $index ];
 956         }
 957 }
 958
 959 /**
 960  * Check whether a given URL has a domain that occurs in a given set of domains
 961  * @param string $url URL
 962  * @param array $domains Array of domains (strings)
 963  * @return bool True if the host part of $url ends in one of the strings in $domains
 964  */
 965 function wfMatchesDomainList( $url, $domains ) {
 966         $bits = wfParseUrl( $url );
 967         if ( is_array( $bits ) && isset( $bits['host'] ) ) {
 968                 $host = '.' . $bits['host'];
 969                 foreach ( (array)$domains as $domain ) {
 970                         $domain = '.' . $domain;
 971                         if ( substr( $host, -strlen( $domain ) ) === $domain ) {
 972                                 return true;
 973                         }
 974                 }
 975         }
 976         return false;
 977 }
 978
 979 /**
 980  * Sends a line to the debug log if enabled or, optionally, to a comment in output.
 981  * In normal operation this is a NOP.
 982  *
 983  * Controlling globals:
 984  * $wgDebugLogFile - points to the log file
 985  * $wgDebugRawPage - if false, 'action=raw' hits will not result in debug output.
 986  * $wgDebugComments - if on, some debug items may appear in comments in the HTML output.
 987  *
 988  * @since 1.25 support for additional context data
 989  *
 990  * @param string $text
 991  * @param string|bool $dest Destination of the message:
 992  *     - 'all': both to the log and HTML (debug toolbar or HTML comments)
 993  *     - 'private': excluded from HTML output
 994  *   For backward compatibility, it can also take a boolean:
 995  *     - true: same as 'all'
 996  *     - false: same as 'private'
 997  * @param array $context Additional logging context data
 998  */
 999 function wfDebug( $text, $dest = 'all', array $context = [] ) {
1000         global $wgDebugRawPage, $wgDebugLogPrefix;
1001         global $wgDebugTimestamps, $wgRequestTime;
1002
1003         if ( !$wgDebugRawPage && wfIsDebugRawPage() ) {
1004                 return;
1005         }
1006
1007         $text = trim( $text );
1008
1009         if ( $wgDebugTimestamps ) {
1010                 $context['seconds_elapsed'] = sprintf(
1011                         '%6.4f',
1012                         microtime( true ) - $wgRequestTime
1013                 );
1014                 $context['memory_used'] = sprintf(
1015                         '%5.1fM',
1016                         ( memory_get_usage( true ) / ( 1024 * 1024 ) )
1017                 );
1018         }
1019
1020         if ( $wgDebugLogPrefix !== '' ) {
1021                 $context['prefix'] = $wgDebugLogPrefix;
1022         }
1023         $context['private'] = ( $dest === false || $dest === 'private' );
1024
1025         $logger = LoggerFactory::getInstance( 'wfDebug' );
1026         $logger->debug( $text, $context );
1027 }
1028
1029 /**
1030  * Returns true if debug logging should be suppressed if $wgDebugRawPage = false
1031  * @return bool
1032  */
1033 function wfIsDebugRawPage() {
1034         static $cache;
1035         if ( $cache !== null ) {
1036                 return $cache;
1037         }
1038         # Check for raw action using $_GET not $wgRequest, since the latter might not be initialised yet
1039         if ( ( isset( $_GET['action'] ) && $_GET['action'] == 'raw' )
1040                 || (
1041                         isset( $_SERVER['SCRIPT_NAME'] )
1042                         && substr( $_SERVER['SCRIPT_NAME'], -8 ) == 'load.php'
1043                 )
1044         ) {
1045                 $cache = true;
1046         } else {
1047                 $cache = false;
1048         }
1049         return $cache;
1050 }
1051
1052 /**
1053  * Send a line giving PHP memory usage.
1054  *
1055  * @param bool $exact Print exact byte values instead of kibibytes (default: false)
1056  */
1057 function wfDebugMem( $exact = false ) {
1058         $mem = memory_get_usage();
1059         if ( !$exact ) {
1060                 $mem = floor( $mem / 1024 ) . ' KiB';
1061         } else {
1062                 $mem .= ' B';
1063         }
1064         wfDebug( "Memory usage: $mem\n" );
1065 }
1066
1067 /**
1068  * Send a line to a supplementary debug log file, if configured, or main debug
1069  * log if not.
1070  *
1071  * To configure a supplementary log file, set $wgDebugLogGroups[$logGroup] to
1072  * a string filename or an associative array mapping 'destination' to the
1073  * desired filename. The associative array may also contain a 'sample' key
1074  * with an integer value, specifying a sampling factor. Sampled log events
1075  * will be emitted with a 1 in N random chance.
1076  *
1077  * @since 1.23 support for sampling log messages via $wgDebugLogGroups.
1078  * @since 1.25 support for additional context data
1079  * @since 1.25 sample behavior dependent on configured $wgMWLoggerDefaultSpi
1080  *
1081  * @param string $logGroup
1082  * @param string $text
1083  * @param string|bool $dest Destination of the message:
1084  *     - 'all': both to the log and HTML (debug toolbar or HTML comments)
1085  *     - 'private': only to the specific log if set in $wgDebugLogGroups and
1086  *       discarded otherwise
1087  *   For backward compatibility, it can also take a boolean:
1088  *     - true: same as 'all'
1089  *     - false: same as 'private'
1090  * @param array $context Additional logging context data
1091  */
1092 function wfDebugLog(
1093         $logGroup, $text, $dest = 'all', array $context = []
1094 ) {
1095         $text = trim( $text );
1096
1097         $logger = LoggerFactory::getInstance( $logGroup );
1098         $context['private'] = ( $dest === false || $dest === 'private' );
1099         $logger->info( $text, $context );
1100 }
1101
1102 /**
1103  * Log for database errors
1104  *
1105  * @since 1.25 support for additional context data
1106  *
1107  * @param string $text Database error message.
1108  * @param array $context Additional logging context data
1109  */
1110 function wfLogDBError( $text, array $context = [] ) {
1111         $logger = LoggerFactory::getInstance( 'wfLogDBError' );
1112         $logger->error( trim( $text ), $context );
1113 }
1114
1115 /**
1116  * Throws a warning that $function is deprecated
1117  *
1118  * @param string $function
1119  * @param string|bool $version Version of MediaWiki that the function
1120  *    was deprecated in (Added in 1.19).
1121  * @param string|bool $component Added in 1.19.
1122  * @param int $callerOffset How far up the call stack is the original
1123  *    caller. 2 = function that called the function that called
1124  *    wfDeprecated (Added in 1.20)
1125  *
1126  * @return null
1127  */
1128 function wfDeprecated( $function, $version = false, $component = false, $callerOffset = 2 ) {
1129         MWDebug::deprecated( $function, $version, $component, $callerOffset + 1 );
1130 }
1131
1132 /**
1133  * Send a warning either to the debug log or in a PHP error depending on
1134  * $wgDevelopmentWarnings. To log warnings in production, use wfLogWarning() instead.
1135  *
1136  * @param string $msg Message to send
1137  * @param int $callerOffset Number of items to go back in the backtrace to
1138  *        find the correct caller (1 = function calling wfWarn, ...)
1139  * @param int $level PHP error level; defaults to E_USER_NOTICE;
1140  *        only used when $wgDevelopmentWarnings is true
1141  */
1142 function wfWarn( $msg, $callerOffset = 1, $level = E_USER_NOTICE ) {
1143         MWDebug::warning( $msg, $callerOffset + 1, $level, 'auto' );
1144 }
1145
1146 /**
1147  * Send a warning as a PHP error and the debug log. This is intended for logging
1148  * warnings in production. For logging development warnings, use WfWarn instead.
1149  *
1150  * @param string $msg Message to send
1151  * @param int $callerOffset Number of items to go back in the backtrace to
1152  *        find the correct caller (1 = function calling wfLogWarning, ...)
1153  * @param int $level PHP error level; defaults to E_USER_WARNING
1154  */
1155 function wfLogWarning( $msg, $callerOffset = 1, $level = E_USER_WARNING ) {
1156         MWDebug::warning( $msg, $callerOffset + 1, $level, 'production' );
1157 }
1158
1159 /**
1160  * Log to a file without getting "file size exceeded" signals.
1161  *
1162  * Can also log to TCP or UDP with the syntax udp://host:port/prefix. This will
1163  * send lines to the specified port, prefixed by the specified prefix and a space.
1164  * @since 1.25 support for additional context data
1165  *
1166  * @param string $text
1167  * @param string $file Filename
1168  * @param array $context Additional logging context data
1169  * @throws MWException
1170  * @deprecated since 1.25 Use \MediaWiki\Logger\LegacyLogger::emit or UDPTransport
1171  */
1172 function wfErrorLog( $text, $file, array $context = [] ) {
1173         wfDeprecated( __METHOD__, '1.25' );
1174         $logger = LoggerFactory::getInstance( 'wfErrorLog' );
1175         $context['destination'] = $file;
1176         $logger->info( trim( $text ), $context );
1177 }
1178
1179 /**
1180  * @todo document
1181  */
1182 function wfLogProfilingData() {
1183         global $wgDebugLogGroups, $wgDebugRawPage;
1184
1185         $context = RequestContext::getMain();
1186         $request = $context->getRequest();
1187
1188         $profiler = Profiler::instance();
1189         $profiler->setContext( $context );
1190         $profiler->logData();
1191
1192         $config = $context->getConfig();
1193         if ( $config->get( 'StatsdServer' ) ) {
1194                 try {
1195                         $statsdServer = explode( ':', $config->get( 'StatsdServer' ) );
1196                         $statsdHost = $statsdServer[0];
1197                         $statsdPort = isset( $statsdServer[1] ) ? $statsdServer[1] : 8125;
1198                         $statsdSender = new SocketSender( $statsdHost, $statsdPort );
1199                         $statsdClient = new SamplingStatsdClient( $statsdSender, true, false );
1200                         $statsdClient->setSamplingRates( $config->get( 'StatsdSamplingRates' ) );
1201                         $statsdClient->send( $context->getStats()->getBuffer() );
1202                 } catch ( Exception $ex ) {
1203                         MWExceptionHandler::logException( $ex );
1204                 }
1205         }
1206
1207         # Profiling must actually be enabled...
1208         if ( $profiler instanceof ProfilerStub ) {
1209                 return;
1210         }
1211
1212         if ( isset( $wgDebugLogGroups['profileoutput'] )
1213                 && $wgDebugLogGroups['profileoutput'] === false
1214         ) {
1215                 // Explicitly disabled
1216                 return;
1217         }
1218         if ( !$wgDebugRawPage && wfIsDebugRawPage() ) {
1219                 return;
1220         }
1221
1222         $ctx = [ 'elapsed' => $request->getElapsedTime() ];
1223         if ( !empty( $_SERVER['HTTP_X_FORWARDED_FOR'] ) ) {
1224                 $ctx['forwarded_for'] = $_SERVER['HTTP_X_FORWARDED_FOR'];
1225         }
1226         if ( !empty( $_SERVER['HTTP_CLIENT_IP'] ) ) {
1227                 $ctx['client_ip'] = $_SERVER['HTTP_CLIENT_IP'];
1228         }
1229         if ( !empty( $_SERVER['HTTP_FROM'] ) ) {
1230                 $ctx['from'] = $_SERVER['HTTP_FROM'];
1231         }
1232         if ( isset( $ctx['forwarded_for'] ) ||
1233                 isset( $ctx['client_ip'] ) ||
1234                 isset( $ctx['from'] ) ) {
1235                 $ctx['proxy'] = $_SERVER['REMOTE_ADDR'];
1236         }
1237
1238         // Don't load $wgUser at this late stage just for statistics purposes
1239         // @todo FIXME: We can detect some anons even if it is not loaded.
1240         // See User::getId()
1241         $user = $context->getUser();
1242         $ctx['anon'] = $user->isItemLoaded( 'id' ) && $user->isAnon();
1243
1244         // Command line script uses a FauxRequest object which does not have
1245         // any knowledge about an URL and throw an exception instead.
1246         try {
1247                 $ctx['url'] = urldecode( $request->getRequestURL() );
1248         } catch ( Exception $ignored ) {
1249                 // no-op
1250         }
1251
1252         $ctx['output'] = $profiler->getOutput();
1253
1254         $log = LoggerFactory::getInstance( 'profileoutput' );
1255         $log->info( "Elapsed: {elapsed}; URL: <{url}>\n{output}", $ctx );
1256 }
1257
1258 /**
1259  * Increment a statistics counter
1260  *
1261  * @param string $key
1262  * @param int $count
1263  * @return void
1264  */
1265 function wfIncrStats( $key, $count = 1 ) {
1266         $stats = RequestContext::getMain()->getStats();
1267         $stats->updateCount( $key, $count );
1268 }
1269
1270 /**
1271  * Check whether the wiki is in read-only mode.
1272  *
1273  * @return bool
1274  */
1275 function wfReadOnly() {
1276         return wfReadOnlyReason() !== false;
1277 }
1278
1279 /**
1280  * Check if the site is in read-only mode and return the message if so
1281  *
1282  * This checks wfConfiguredReadOnlyReason() and the main load balancer
1283  * for replica DB lag. This may result in DB connection being made.
1284  *
1285  * @return string|bool String when in read-only mode; false otherwise
1286  */
1287 function wfReadOnlyReason() {
1288         $readOnly = wfConfiguredReadOnlyReason();
1289         if ( $readOnly !== false ) {
1290                 return $readOnly;
1291         }
1292
1293         static $lbReadOnly = null;
1294         if ( $lbReadOnly === null ) {
1295                 // Callers use this method to be aware that data presented to a user
1296                 // may be very stale and thus allowing submissions can be problematic.
1297                 $lbReadOnly = wfGetLB()->getReadOnlyReason();
1298         }
1299
1300         return $lbReadOnly;
1301 }
1302
1303 /**
1304  * Get the value of $wgReadOnly or the contents of $wgReadOnlyFile.
1305  *
1306  * @return string|bool String when in read-only mode; false otherwise
1307  * @since 1.27
1308  */
1309 function wfConfiguredReadOnlyReason() {
1310         global $wgReadOnly, $wgReadOnlyFile;
1311
1312         if ( $wgReadOnly === null ) {
1313                 // Set $wgReadOnly for faster access next time
1314                 if ( is_file( $wgReadOnlyFile ) && filesize( $wgReadOnlyFile ) > 0 ) {
1315                         $wgReadOnly = file_get_contents( $wgReadOnlyFile );
1316                 } else {
1317                         $wgReadOnly = false;
1318                 }
1319         }
1320
1321         return $wgReadOnly;
1322 }
1323
1324 /**
1325  * Return a Language object from $langcode
1326  *
1327  * @param Language|string|bool $langcode Either:
1328  *                  - a Language object
1329  *                  - code of the language to get the message for, if it is
1330  *                    a valid code create a language for that language, if
1331  *                    it is a string but not a valid code then make a basic
1332  *                    language object
1333  *                  - a boolean: if it's false then use the global object for
1334  *                    the current user's language (as a fallback for the old parameter
1335  *                    functionality), or if it is true then use global object
1336  *                    for the wiki's content language.
1337  * @return Language
1338  */
1339 function wfGetLangObj( $langcode = false ) {
1340         # Identify which language to get or create a language object for.
1341         # Using is_object here due to Stub objects.
1342         if ( is_object( $langcode ) ) {
1343                 # Great, we already have the object (hopefully)!
1344                 return $langcode;
1345         }
1346
1347         global $wgContLang, $wgLanguageCode;
1348         if ( $langcode === true || $langcode === $wgLanguageCode ) {
1349                 # $langcode is the language code of the wikis content language object.
1350                 # or it is a boolean and value is true
1351                 return $wgContLang;
1352         }
1353
1354         global $wgLang;
1355         if ( $langcode === false || $langcode === $wgLang->getCode() ) {
1356                 # $langcode is the language code of user language object.
1357                 # or it was a boolean and value is false
1358                 return $wgLang;
1359         }
1360
1361         $validCodes = array_keys( Language::fetchLanguageNames() );
1362         if ( in_array( $langcode, $validCodes ) ) {
1363                 # $langcode corresponds to a valid language.
1364                 return Language::factory( $langcode );
1365         }
1366
1367         # $langcode is a string, but not a valid language code; use content language.
1368         wfDebug( "Invalid language code passed to wfGetLangObj, falling back to content language.\n" );
1369         return $wgContLang;
1370 }
1371
1372 /**
1373  * This is the function for getting translated interface messages.
1374  *
1375  * @see Message class for documentation how to use them.
1376  * @see https://www.mediawiki.org/wiki/Manual:Messages_API
1377  *
1378  * This function replaces all old wfMsg* functions.
1379  *
1380  * @param string|string[]|MessageSpecifier $key Message key, or array of keys, or a MessageSpecifier
1381  * @param mixed $params,... Normal message parameters
1382  * @return Message
1383  *
1384  * @since 1.17
1385  *
1386  * @see Message::__construct
1387  */
1388 function wfMessage( $key /*...*/ ) {
1389         $message = new Message( $key );
1390
1391         // We call Message::params() to reduce code duplication
1392         $params = func_get_args();
1393         array_shift( $params );
1394         if ( $params ) {
1395                 call_user_func_array( [ $message, 'params' ], $params );
1396         }
1397
1398         return $message;
1399 }
1400
1401 /**
1402  * This function accepts multiple message keys and returns a message instance
1403  * for the first message which is non-empty. If all messages are empty then an
1404  * instance of the first message key is returned.
1405  *
1406  * @param string|string[] $keys,... Message keys
1407  * @return Message
1408  *
1409  * @since 1.18
1410  *
1411  * @see Message::newFallbackSequence
1412  */
1413 function wfMessageFallback( /*...*/ ) {
1414         $args = func_get_args();
1415         return call_user_func_array( 'Message::newFallbackSequence', $args );
1416 }
1417
1418 /**
1419  * Replace message parameter keys on the given formatted output.
1420  *
1421  * @param string $message
1422  * @param array $args
1423  * @return string
1424  * @private
1425  */
1426 function wfMsgReplaceArgs( $message, $args ) {
1427         # Fix windows line-endings
1428         # Some messages are split with explode("\n", $msg)
1429         $message = str_replace( "\r", '', $message );
1430
1431         // Replace arguments
1432         if ( is_array( $args ) && $args ) {
1433                 if ( is_array( $args[0] ) ) {
1434                         $args = array_values( $args[0] );
1435                 }
1436                 $replacementKeys = [];
1437                 foreach ( $args as $n => $param ) {
1438                         $replacementKeys['$' . ( $n + 1 )] = $param;
1439                 }
1440                 $message = strtr( $message, $replacementKeys );
1441         }
1442
1443         return $message;
1444 }
1445
1446 /**
1447  * Fetch server name for use in error reporting etc.
1448  * Use real server name if available, so we know which machine
1449  * in a server farm generated the current page.
1450  *
1451  * @return string
1452  */
1453 function wfHostname() {
1454         static $host;
1455         if ( is_null( $host ) ) {
1456
1457                 # Hostname overriding
1458                 global $wgOverrideHostname;
1459                 if ( $wgOverrideHostname !== false ) {
1460                         # Set static and skip any detection
1461                         $host = $wgOverrideHostname;
1462                         return $host;
1463                 }
1464
1465                 if ( function_exists( 'posix_uname' ) ) {
1466                         // This function not present on Windows
1467                         $uname = posix_uname();
1468                 } else {
1469                         $uname = false;
1470                 }
1471                 if ( is_array( $uname ) && isset( $uname['nodename'] ) ) {
1472                         $host = $uname['nodename'];
1473                 } elseif ( getenv( 'COMPUTERNAME' ) ) {
1474                         # Windows computer name
1475                         $host = getenv( 'COMPUTERNAME' );
1476                 } else {
1477                         # This may be a virtual server.
1478                         $host = $_SERVER['SERVER_NAME'];
1479                 }
1480         }
1481         return $host;
1482 }
1483
1484 /**
1485  * Returns a script tag that stores the amount of time it took MediaWiki to
1486  * handle the request in milliseconds as 'wgBackendResponseTime'.
1487  *
1488  * If $wgShowHostnames is true, the script will also set 'wgHostname' to the
1489  * hostname of the server handling the request.
1490  *
1491  * @return string
1492  */
1493 function wfReportTime() {
1494         global $wgRequestTime, $wgShowHostnames;
1495
1496         $responseTime = round( ( microtime( true ) - $wgRequestTime ) * 1000 );
1497         $reportVars = [ 'wgBackendResponseTime' => $responseTime ];
1498         if ( $wgShowHostnames ) {
1499                 $reportVars['wgHostname'] = wfHostname();
1500         }
1501         return Skin::makeVariablesScript( $reportVars );
1502 }
1503
1504 /**
1505  * Safety wrapper for debug_backtrace().
1506  *
1507  * Will return an empty array if debug_backtrace is disabled, otherwise
1508  * the output from debug_backtrace() (trimmed).
1509  *
1510  * @param int $limit This parameter can be used to limit the number of stack frames returned
1511  *
1512  * @return array Array of backtrace information
1513  */
1514 function wfDebugBacktrace( $limit = 0 ) {
1515         static $disabled = null;
1516
1517         if ( is_null( $disabled ) ) {
1518                 $disabled = !function_exists( 'debug_backtrace' );
1519                 if ( $disabled ) {
1520                         wfDebug( "debug_backtrace() is disabled\n" );
1521                 }
1522         }
1523         if ( $disabled ) {
1524                 return [];
1525         }
1526
1527         if ( $limit ) {
1528                 return array_slice( debug_backtrace( DEBUG_BACKTRACE_PROVIDE_OBJECT, $limit + 1 ), 1 );
1529         } else {
1530                 return array_slice( debug_backtrace(), 1 );
1531         }
1532 }
1533
1534 /**
1535  * Get a debug backtrace as a string
1536  *
1537  * @param bool|null $raw If true, the return value is plain text. If false, HTML.
1538  *   Defaults to $wgCommandLineMode if unset.
1539  * @return string
1540  * @since 1.25 Supports $raw parameter.
1541  */
1542 function wfBacktrace( $raw = null ) {
1543         global $wgCommandLineMode;
1544
1545         if ( $raw === null ) {
1546                 $raw = $wgCommandLineMode;
1547         }
1548
1549         if ( $raw ) {
1550                 $frameFormat = "%s line %s calls %s()\n";
1551                 $traceFormat = "%s";
1552         } else {
1553                 $frameFormat = "<li>%s line %s calls %s()</li>\n";
1554                 $traceFormat = "<ul>\n%s</ul>\n";
1555         }
1556
1557         $frames = array_map( function ( $frame ) use ( $frameFormat ) {
1558                 $file = !empty( $frame['file'] ) ? basename( $frame['file'] ) : '-';
1559                 $line = isset( $frame['line'] ) ? $frame['line'] : '-';
1560                 $call = $frame['function'];
1561                 if ( !empty( $frame['class'] ) ) {
1562                         $call = $frame['class'] . $frame['type'] . $call;
1563                 }
1564                 return sprintf( $frameFormat, $file, $line, $call );
1565         }, wfDebugBacktrace() );
1566
1567         return sprintf( $traceFormat, implode( '', $frames ) );
1568 }
1569
1570 /**
1571  * Get the name of the function which called this function
1572  * wfGetCaller( 1 ) is the function with the wfGetCaller() call (ie. __FUNCTION__)
1573  * wfGetCaller( 2 ) [default] is the caller of the function running wfGetCaller()
1574  * wfGetCaller( 3 ) is the parent of that.
1575  *
1576  * @param int $level
1577  * @return string
1578  */
1579 function wfGetCaller( $level = 2 ) {
1580         $backtrace = wfDebugBacktrace( $level + 1 );
1581         if ( isset( $backtrace[$level] ) ) {
1582                 return wfFormatStackFrame( $backtrace[$level] );
1583         } else {
1584                 return 'unknown';
1585         }
1586 }
1587
1588 /**
1589  * Return a string consisting of callers in the stack. Useful sometimes
1590  * for profiling specific points.
1591  *
1592  * @param int $limit The maximum depth of the stack frame to return, or false for the entire stack.
1593  * @return string
1594  */
1595 function wfGetAllCallers( $limit = 3 ) {
1596         $trace = array_reverse( wfDebugBacktrace() );
1597         if ( !$limit || $limit > count( $trace ) - 1 ) {
1598                 $limit = count( $trace ) - 1;
1599         }
1600         $trace = array_slice( $trace, -$limit - 1, $limit );
1601         return implode( '/', array_map( 'wfFormatStackFrame', $trace ) );
1602 }
1603
1604 /**
1605  * Return a string representation of frame
1606  *
1607  * @param array $frame
1608  * @return string
1609  */
1610 function wfFormatStackFrame( $frame ) {
1611         if ( !isset( $frame['function'] ) ) {
1612                 return 'NO_FUNCTION_GIVEN';
1613         }
1614         return isset( $frame['class'] ) && isset( $frame['type'] ) ?
1615                 $frame['class'] . $frame['type'] . $frame['function'] :
1616                 $frame['function'];
1617 }
1618
1619 /* Some generic result counters, pulled out of SearchEngine */
1620
1621 /**
1622  * @todo document
1623  *
1624  * @param int $offset
1625  * @param int $limit
1626  * @return string
1627  */
1628 function wfShowingResults( $offset, $limit ) {
1629         return wfMessage( 'showingresults' )->numParams( $limit, $offset + 1 )->parse();
1630 }
1631
1632 /**
1633  * Whether the client accept gzip encoding
1634  *
1635  * Uses the Accept-Encoding header to check if the client supports gzip encoding.
1636  * Use this when considering to send a gzip-encoded response to the client.
1637  *
1638  * @param bool $force Forces another check even if we already have a cached result.
1639  * @return bool
1640  */
1641 function wfClientAcceptsGzip( $force = false ) {
1642         static $result = null;
1643         if ( $result === null || $force ) {
1644                 $result = false;
1645                 if ( isset( $_SERVER['HTTP_ACCEPT_ENCODING'] ) ) {
1646                         # @todo FIXME: We may want to blacklist some broken browsers
1647                         $m = [];
1648                         if ( preg_match(
1649                                         '/\bgzip(?:;(q)=([0-9]+(?:\.[0-9]+)))?\b/',
1650                                         $_SERVER['HTTP_ACCEPT_ENCODING'],
1651                                         $m
1652                                 )
1653                         ) {
1654                                 if ( isset( $m[2] ) && ( $m[1] == 'q' ) && ( $m[2] == 0 ) ) {
1655                                         $result = false;
1656                                         return $result;
1657                                 }
1658                                 wfDebug( "wfClientAcceptsGzip: client accepts gzip.\n" );
1659                                 $result = true;
1660                         }
1661                 }
1662         }
1663         return $result;
1664 }
1665
1666 /**
1667  * Escapes the given text so that it may be output using addWikiText()
1668  * without any linking, formatting, etc. making its way through. This
1669  * is achieved by substituting certain characters with HTML entities.
1670  * As required by the callers, "<nowiki>" is not used.
1671  *
1672  * @param string $text Text to be escaped
1673  * @return string
1674  */
1675 function wfEscapeWikiText( $text ) {
1676         global $wgEnableMagicLinks;
1677         static $repl = null, $repl2 = null;
1678         if ( $repl === null || defined( 'MW_PARSER_TEST' ) || defined( 'MW_PHPUNIT_TEST' ) ) {
1679                 // Tests depend upon being able to change $wgEnableMagicLinks, so don't cache
1680                 // in those situations
1681                 $repl = [
1682                         '"' => '&#34;', '&' => '&#38;', "'" => '&#39;', '<' => '&#60;',
1683                         '=' => '&#61;', '>' => '&#62;', '[' => '&#91;', ']' => '&#93;',
1684                         '{' => '&#123;', '|' => '&#124;', '}' => '&#125;', ';' => '&#59;',
1685                         "\n#" => "\n&#35;", "\r#" => "\r&#35;",
1686                         "\n*" => "\n&#42;", "\r*" => "\r&#42;",
1687                         "\n:" => "\n&#58;", "\r:" => "\r&#58;",
1688                         "\n " => "\n&#32;", "\r " => "\r&#32;",
1689                         "\n\n" => "\n&#10;", "\r\n" => "&#13;\n",
1690                         "\n\r" => "\n&#13;", "\r\r" => "\r&#13;",
1691                         "\n\t" => "\n&#9;", "\r\t" => "\r&#9;", // "\n\t\n" is treated like "\n\n"
1692                         "\n----" => "\n&#45;---", "\r----" => "\r&#45;---",
1693                         '__' => '_&#95;', '://' => '&#58;//',
1694                 ];
1695
1696                 $magicLinks = array_keys( array_filter( $wgEnableMagicLinks ) );
1697                 // We have to catch everything "\s" matches in PCRE
1698                 foreach ( $magicLinks as $magic ) {
1699                         $repl["$magic "] = "$magic&#32;";
1700                         $repl["$magic\t"] = "$magic&#9;";
1701                         $repl["$magic\r"] = "$magic&#13;";
1702                         $repl["$magic\n"] = "$magic&#10;";
1703                         $repl["$magic\f"] = "$magic&#12;";
1704                 }
1705
1706                 // And handle protocols that don't use "://"
1707                 global $wgUrlProtocols;
1708                 $repl2 = [];
1709                 foreach ( $wgUrlProtocols as $prot ) {
1710                         if ( substr( $prot, -1 ) === ':' ) {
1711                                 $repl2[] = preg_quote( substr( $prot, 0, -1 ), '/' );
1712                         }
1713                 }
1714                 $repl2 = $repl2 ? '/\b(' . implode( '|', $repl2 ) . '):/i' : '/^(?!)/';
1715         }
1716         $text = substr( strtr( "\n$text", $repl ), 1 );
1717         $text = preg_replace( $repl2, '$1&#58;', $text );
1718         return $text;
1719 }
1720
1721 /**
1722  * Sets dest to source and returns the original value of dest
1723  * If source is NULL, it just returns the value, it doesn't set the variable
1724  * If force is true, it will set the value even if source is NULL
1725  *
1726  * @param mixed $dest
1727  * @param mixed $source
1728  * @param bool $force
1729  * @return mixed
1730  */
1731 function wfSetVar( &$dest, $source, $force = false ) {
1732         $temp = $dest;
1733         if ( !is_null( $source ) || $force ) {
1734                 $dest = $source;
1735         }
1736         return $temp;
1737 }
1738
1739 /**
1740  * As for wfSetVar except setting a bit
1741  *
1742  * @param int $dest
1743  * @param int $bit
1744  * @param bool $state
1745  *
1746  * @return bool
1747  */
1748 function wfSetBit( &$dest, $bit, $state = true ) {
1749         $temp = (bool)( $dest & $bit );
1750         if ( !is_null( $state ) ) {
1751                 if ( $state ) {
1752                         $dest |= $bit;
1753                 } else {
1754                         $dest &= ~$bit;
1755                 }
1756         }
1757         return $temp;
1758 }
1759
1760 /**
1761  * A wrapper around the PHP function var_export().
1762  * Either print it or add it to the regular output ($wgOut).
1763  *
1764  * @param mixed $var A PHP variable to dump.
1765  */
1766 function wfVarDump( $var ) {
1767         global $wgOut;
1768         $s = str_replace( "\n", "<br />\n", var_export( $var, true ) . "\n" );
1769         if ( headers_sent() || !isset( $wgOut ) || !is_object( $wgOut ) ) {
1770                 print $s;
1771         } else {
1772                 $wgOut->addHTML( $s );
1773         }
1774 }
1775
1776 /**
1777  * Provide a simple HTTP error.
1778  *
1779  * @param int|string $code
1780  * @param string $label
1781  * @param string $desc
1782  */
1783 function wfHttpError( $code, $label, $desc ) {
1784         global $wgOut;
1785         HttpStatus::header( $code );
1786         if ( $wgOut ) {
1787                 $wgOut->disable();
1788                 $wgOut->sendCacheControl();
1789         }
1790
1791         MediaWiki\HeaderCallback::warnIfHeadersSent();
1792         header( 'Content-type: text/html; charset=utf-8' );
1793         print '<!DOCTYPE html>' .
1794                 '<html><head><title>' .
1795                 htmlspecialchars( $label ) .
1796                 '</title></head><body><h1>' .
1797                 htmlspecialchars( $label ) .
1798                 '</h1><p>' .
1799                 nl2br( htmlspecialchars( $desc ) ) .
1800                 "</p></body></html>\n";
1801 }
1802
1803 /**
1804  * Clear away any user-level output buffers, discarding contents.
1805  *
1806  * Suitable for 'starting afresh', for instance when streaming
1807  * relatively large amounts of data without buffering, or wanting to
1808  * output image files without ob_gzhandler's compression.
1809  *
1810  * The optional $resetGzipEncoding parameter controls suppression of
1811  * the Content-Encoding header sent by ob_gzhandler; by default it
1812  * is left. See comments for wfClearOutputBuffers() for why it would
1813  * be used.
1814  *
1815  * Note that some PHP configuration options may add output buffer
1816  * layers which cannot be removed; these are left in place.
1817  *
1818  * @param bool $resetGzipEncoding
1819  */
1820 function wfResetOutputBuffers( $resetGzipEncoding = true ) {
1821         if ( $resetGzipEncoding ) {
1822                 // Suppress Content-Encoding and Content-Length
1823                 // headers from 1.10+s wfOutputHandler
1824                 global $wgDisableOutputCompression;
1825                 $wgDisableOutputCompression = true;
1826         }
1827         while ( $status = ob_get_status() ) {
1828                 if ( isset( $status['flags'] ) ) {
1829                         $flags = PHP_OUTPUT_HANDLER_CLEANABLE | PHP_OUTPUT_HANDLER_REMOVABLE;
1830                         $deleteable = ( $status['flags'] & $flags ) === $flags;
1831                 } elseif ( isset( $status['del'] ) ) {
1832                         $deleteable = $status['del'];
1833                 } else {
1834                         // Guess that any PHP-internal setting can't be removed.
1835                         $deleteable = $status['type'] !== 0; /* PHP_OUTPUT_HANDLER_INTERNAL */
1836                 }
1837                 if ( !$deleteable ) {
1838                         // Give up, and hope the result doesn't break
1839                         // output behavior.
1840                         break;
1841                 }
1842                 if ( $status['name'] === 'MediaWikiTestCase::wfResetOutputBuffersBarrier' ) {
1843                         // Unit testing barrier to prevent this function from breaking PHPUnit.
1844                         break;
1845                 }
1846                 if ( !ob_end_clean() ) {
1847                         // Could not remove output buffer handler; abort now
1848                         // to avoid getting in some kind of infinite loop.
1849                         break;
1850                 }
1851                 if ( $resetGzipEncoding ) {
1852                         if ( $status['name'] == 'ob_gzhandler' ) {
1853                                 // Reset the 'Content-Encoding' field set by this handler
1854                                 // so we can start fresh.
1855                                 header_remove( 'Content-Encoding' );
1856                                 break;
1857                         }
1858                 }
1859         }
1860 }
1861
1862 /**
1863  * More legible than passing a 'false' parameter to wfResetOutputBuffers():
1864  *
1865  * Clear away output buffers, but keep the Content-Encoding header
1866  * produced by ob_gzhandler, if any.
1867  *
1868  * This should be used for HTTP 304 responses, where you need to
1869  * preserve the Content-Encoding header of the real result, but
1870  * also need to suppress the output of ob_gzhandler to keep to spec
1871  * and avoid breaking Firefox in rare cases where the headers and
1872  * body are broken over two packets.
1873  */
1874 function wfClearOutputBuffers() {
1875         wfResetOutputBuffers( false );
1876 }
1877
1878 /**
1879  * Converts an Accept-* header into an array mapping string values to quality
1880  * factors
1881  *
1882  * @param string $accept
1883  * @param string $def Default
1884  * @return float[] Associative array of string => float pairs
1885  */
1886 function wfAcceptToPrefs( $accept, $def = '*/*' ) {
1887         # No arg means accept anything (per HTTP spec)
1888         if ( !$accept ) {
1889                 return [ $def => 1.0 ];
1890         }
1891
1892         $prefs = [];
1893
1894         $parts = explode( ',', $accept );
1895
1896         foreach ( $parts as $part ) {
1897                 # @todo FIXME: Doesn't deal with params like 'text/html; level=1'
1898                 $values = explode( ';', trim( $part ) );
1899                 $match = [];
1900                 if ( count( $values ) == 1 ) {
1901                         $prefs[$values[0]] = 1.0;
1902                 } elseif ( preg_match( '/q\s*=\s*(\d*\.\d+)/', $values[1], $match ) ) {
1903                         $prefs[$values[0]] = floatval( $match[1] );
1904                 }
1905         }
1906
1907         return $prefs;
1908 }
1909
1910 /**
1911  * Checks if a given MIME type matches any of the keys in the given
1912  * array. Basic wildcards are accepted in the array keys.
1913  *
1914  * Returns the matching MIME type (or wildcard) if a match, otherwise
1915  * NULL if no match.
1916  *
1917  * @param string $type
1918  * @param array $avail
1919  * @return string
1920  * @private
1921  */
1922 function mimeTypeMatch( $type, $avail ) {
1923         if ( array_key_exists( $type, $avail ) ) {
1924                 return $type;
1925         } else {
1926                 $mainType = explode( '/', $type )[0];
1927                 if ( array_key_exists( "$mainType/*", $avail ) ) {
1928                         return "$mainType/*";
1929                 } elseif ( array_key_exists( '*/*', $avail ) ) {
1930                         return '*/*';
1931                 } else {
1932                         return null;
1933                 }
1934         }
1935 }
1936
1937 /**
1938  * Returns the 'best' match between a client's requested internet media types
1939  * and the server's list of available types. Each list should be an associative
1940  * array of type to preference (preference is a float between 0.0 and 1.0).
1941  * Wildcards in the types are acceptable.
1942  *
1943  * @param array $cprefs Client's acceptable type list
1944  * @param array $sprefs Server's offered types
1945  * @return string
1946  *
1947  * @todo FIXME: Doesn't handle params like 'text/plain; charset=UTF-8'
1948  * XXX: generalize to negotiate other stuff
1949  */
1950 function wfNegotiateType( $cprefs, $sprefs ) {
1951         $combine = [];
1952
1953         foreach ( array_keys( $sprefs ) as $type ) {
1954                 $subType = explode( '/', $type )[1];
1955                 if ( $subType != '*' ) {
1956                         $ckey = mimeTypeMatch( $type, $cprefs );
1957                         if ( $ckey ) {
1958                                 $combine[$type] = $sprefs[$type] * $cprefs[$ckey];
1959                         }
1960                 }
1961         }
1962
1963         foreach ( array_keys( $cprefs ) as $type ) {
1964                 $subType = explode( '/', $type )[1];
1965                 if ( $subType != '*' && !array_key_exists( $type, $sprefs ) ) {
1966                         $skey = mimeTypeMatch( $type, $sprefs );
1967                         if ( $skey ) {
1968                                 $combine[$type] = $sprefs[$skey] * $cprefs[$type];
1969                         }
1970                 }
1971         }
1972
1973         $bestq = 0;
1974         $besttype = null;
1975
1976         foreach ( array_keys( $combine ) as $type ) {
1977                 if ( $combine[$type] > $bestq ) {
1978                         $besttype = $type;
1979                         $bestq = $combine[$type];
1980                 }
1981         }
1982
1983         return $besttype;
1984 }
1985
1986 /**
1987  * Reference-counted warning suppression
1988  *
1989  * @deprecated since 1.26, use MediaWiki\suppressWarnings() directly
1990  * @param bool $end
1991  */
1992 function wfSuppressWarnings( $end = false ) {
1993         MediaWiki\suppressWarnings( $end );
1994 }
1995
1996 /**
1997  * @deprecated since 1.26, use MediaWiki\restoreWarnings() directly
1998  * Restore error level to previous value
1999  */
2000 function wfRestoreWarnings() {
2001         MediaWiki\suppressWarnings( true );
2002 }
2003
2004 /**
2005  * Get a timestamp string in one of various formats
2006  *
2007  * @param mixed $outputtype A timestamp in one of the supported formats, the
2008  *   function will autodetect which format is supplied and act accordingly.
2009  * @param mixed $ts Optional timestamp to convert, default 0 for the current time
2010  * @return string|bool String / false The same date in the format specified in $outputtype or false
2011  */
2012 function wfTimestamp( $outputtype = TS_UNIX, $ts = 0 ) {
2013         $ret = MWTimestamp::convert( $outputtype, $ts );
2014         if ( $ret === false ) {
2015                 wfDebug( "wfTimestamp() fed bogus time value: TYPE=$outputtype; VALUE=$ts\n" );
2016         }
2017         return $ret;
2018 }
2019
2020 /**
2021  * Return a formatted timestamp, or null if input is null.
2022  * For dealing with nullable timestamp columns in the database.
2023  *
2024  * @param int $outputtype
2025  * @param string $ts
2026  * @return string
2027  */
2028 function wfTimestampOrNull( $outputtype = TS_UNIX, $ts = null ) {
2029         if ( is_null( $ts ) ) {
2030                 return null;
2031         } else {
2032                 return wfTimestamp( $outputtype, $ts );
2033         }
2034 }
2035
2036 /**
2037  * Convenience function; returns MediaWiki timestamp for the present time.
2038  *
2039  * @return string
2040  */
2041 function wfTimestampNow() {
2042         # return NOW
2043         return MWTimestamp::now( TS_MW );
2044 }
2045
2046 /**
2047  * Check if the operating system is Windows
2048  *
2049  * @return bool True if it's Windows, false otherwise.
2050  */
2051 function wfIsWindows() {
2052         static $isWindows = null;
2053         if ( $isWindows === null ) {
2054                 $isWindows = strtoupper( substr( PHP_OS, 0, 3 ) ) === 'WIN';
2055         }
2056         return $isWindows;
2057 }
2058
2059 /**
2060  * Check if we are running under HHVM
2061  *
2062  * @return bool
2063  */
2064 function wfIsHHVM() {
2065         return defined( 'HHVM_VERSION' );
2066 }
2067
2068 /**
2069  * Tries to get the system directory for temporary files. First
2070  * $wgTmpDirectory is checked, and then the TMPDIR, TMP, and TEMP
2071  * environment variables are then checked in sequence, then
2072  * sys_get_temp_dir(), then upload_tmp_dir from php.ini.
2073  *
2074  * NOTE: When possible, use instead the tmpfile() function to create
2075  * temporary files to avoid race conditions on file creation, etc.
2076  *
2077  * @return string
2078  */
2079 function wfTempDir() {
2080         global $wgTmpDirectory;
2081
2082         if ( $wgTmpDirectory !== false ) {
2083                 return $wgTmpDirectory;
2084         }
2085
2086         return TempFSFile::getUsableTempDirectory();
2087 }
2088
2089 /**
2090  * Make directory, and make all parent directories if they don't exist
2091  *
2092  * @param string $dir Full path to directory to create
2093  * @param int $mode Chmod value to use, default is $wgDirectoryMode
2094  * @param string $caller Optional caller param for debugging.
2095  * @throws MWException
2096  * @return bool
2097  */
2098 function wfMkdirParents( $dir, $mode = null, $caller = null ) {
2099         global $wgDirectoryMode;
2100
2101         if ( FileBackend::isStoragePath( $dir ) ) { // sanity
2102                 throw new MWException( __FUNCTION__ . " given storage path '$dir'." );
2103         }
2104
2105         if ( !is_null( $caller ) ) {
2106                 wfDebug( "$caller: called wfMkdirParents($dir)\n" );
2107         }
2108
2109         if ( strval( $dir ) === '' || is_dir( $dir ) ) {
2110                 return true;
2111         }
2112
2113         $dir = str_replace( [ '\\', '/' ], DIRECTORY_SEPARATOR, $dir );
2114
2115         if ( is_null( $mode ) ) {
2116                 $mode = $wgDirectoryMode;
2117         }
2118
2119         // Turn off the normal warning, we're doing our own below
2120         MediaWiki\suppressWarnings();
2121         $ok = mkdir( $dir, $mode, true ); // PHP5 <3
2122         MediaWiki\restoreWarnings();
2123
2124         if ( !$ok ) {
2125                 // directory may have been created on another request since we last checked
2126                 if ( is_dir( $dir ) ) {
2127                         return true;
2128                 }
2129
2130                 // PHP doesn't report the path in its warning message, so add our own to aid in diagnosis.
2131                 wfLogWarning( sprintf( "failed to mkdir \"%s\" mode 0%o", $dir, $mode ) );
2132         }
2133         return $ok;
2134 }
2135
2136 /**
2137  * Remove a directory and all its content.
2138  * Does not hide error.
2139  * @param string $dir
2140  */
2141 function wfRecursiveRemoveDir( $dir ) {
2142         wfDebug( __FUNCTION__ . "( $dir )\n" );
2143         // taken from https://secure.php.net/manual/en/function.rmdir.php#98622
2144         if ( is_dir( $dir ) ) {
2145                 $objects = scandir( $dir );
2146                 foreach ( $objects as $object ) {
2147                         if ( $object != "." && $object != ".." ) {
2148                                 if ( filetype( $dir . '/' . $object ) == "dir" ) {
2149                                         wfRecursiveRemoveDir( $dir . '/' . $object );
2150                                 } else {
2151                                         unlink( $dir . '/' . $object );
2152                                 }
2153                         }
2154                 }
2155                 reset( $objects );
2156                 rmdir( $dir );
2157         }
2158 }
2159
2160 /**
2161  * @param int $nr The number to format
2162  * @param int $acc The number of digits after the decimal point, default 2
2163  * @param bool $round Whether or not to round the value, default true
2164  * @return string
2165  */
2166 function wfPercent( $nr, $acc = 2, $round = true ) {
2167         $ret = sprintf( "%.${acc}f", $nr );
2168         return $round ? round( $ret, $acc ) . '%' : "$ret%";
2169 }
2170
2171 /**
2172  * Safety wrapper around ini_get() for boolean settings.
2173  * The values returned from ini_get() are pre-normalized for settings
2174  * set via php.ini or php_flag/php_admin_flag... but *not*
2175  * for those set via php_value/php_admin_value.
2176  *
2177  * It's fairly common for people to use php_value instead of php_flag,
2178  * which can leave you with an 'off' setting giving a false positive
2179  * for code that just takes the ini_get() return value as a boolean.
2180  *
2181  * To make things extra interesting, setting via php_value accepts
2182  * "true" and "yes" as true, but php.ini and php_flag consider them false. :)
2183  * Unrecognized values go false... again opposite PHP's own coercion
2184  * from string to bool.
2185  *
2186  * Luckily, 'properly' set settings will always come back as '0' or '1',
2187  * so we only have to worry about them and the 'improper' settings.
2188  *
2189  * I frickin' hate PHP... :P
2190  *
2191  * @param string $setting
2192  * @return bool
2193  */
2194 function wfIniGetBool( $setting ) {
2195         $val = strtolower( ini_get( $setting ) );
2196         // 'on' and 'true' can't have whitespace around them, but '1' can.
2197         return $val == 'on'
2198                 || $val == 'true'
2199                 || $val == 'yes'
2200                 || preg_match( "/^\s*[+-]?0*[1-9]/", $val ); // approx C atoi() function
2201 }
2202
2203 /**
2204  * Version of escapeshellarg() that works better on Windows.
2205  *
2206  * Originally, this fixed the incorrect use of single quotes on Windows
2207  * (https://bugs.php.net/bug.php?id=26285) and the locale problems on Linux in
2208  * PHP 5.2.6+ (bug backported to earlier distro releases of PHP).
2209  *
2210  * @param string ... strings to escape and glue together, or a single array of strings parameter
2211  * @return string
2212  */
2213 function wfEscapeShellArg( /*...*/ ) {
2214         wfInitShellLocale();
2215
2216         $args = func_get_args();
2217         if ( count( $args ) === 1 && is_array( reset( $args ) ) ) {
2218                 // If only one argument has been passed, and that argument is an array,
2219                 // treat it as a list of arguments
2220                 $args = reset( $args );
2221         }
2222
2223         $first = true;
2224         $retVal = '';
2225         foreach ( $args as $arg ) {
2226                 if ( !$first ) {
2227                         $retVal .= ' ';
2228                 } else {
2229                         $first = false;
2230                 }
2231
2232                 if ( wfIsWindows() ) {
2233                         // Escaping for an MSVC-style command line parser and CMD.EXE
2234                         // @codingStandardsIgnoreStart For long URLs
2235                         // Refs:
2236                         //  * https://web.archive.org/web/20020708081031/http://mailman.lyra.org/pipermail/scite-interest/2002-March/000436.html
2237                         //  * https://technet.microsoft.com/en-us/library/cc723564.aspx
2238                         //  * T15518
2239                         //  * CR r63214
2240                         // Double the backslashes before any double quotes. Escape the double quotes.
2241                         // @codingStandardsIgnoreEnd
2242                         $tokens = preg_split( '/(\\\\*")/', $arg, -1, PREG_SPLIT_DELIM_CAPTURE );
2243                         $arg = '';
2244                         $iteration = 0;
2245                         foreach ( $tokens as $token ) {
2246                                 if ( $iteration % 2 == 1 ) {
2247                                         // Delimiter, a double quote preceded by zero or more slashes
2248                                         $arg .= str_replace( '\\', '\\\\', substr( $token, 0, -1 ) ) . '\\"';
2249                                 } elseif ( $iteration % 4 == 2 ) {
2250                                         // ^ in $token will be outside quotes, need to be escaped
2251                                         $arg .= str_replace( '^', '^^', $token );
2252                                 } else { // $iteration % 4 == 0
2253                                         // ^ in $token will appear inside double quotes, so leave as is
2254                                         $arg .= $token;
2255                                 }
2256                                 $iteration++;
2257                         }
2258                         // Double the backslashes before the end of the string, because
2259                         // we will soon add a quote
2260                         $m = [];
2261                         if ( preg_match( '/^(.*?)(\\\\+)$/', $arg, $m ) ) {
2262                                 $arg = $m[1] . str_replace( '\\', '\\\\', $m[2] );
2263                         }
2264
2265                         // Add surrounding quotes
2266                         $retVal .= '"' . $arg . '"';
2267                 } else {
2268                         $retVal .= escapeshellarg( $arg );
2269                 }
2270         }
2271         return $retVal;
2272 }
2273
2274 /**
2275  * Check if wfShellExec() is effectively disabled via php.ini config
2276  *
2277  * @return bool|string False or 'disabled'
2278  * @since 1.22
2279  */
2280 function wfShellExecDisabled() {
2281         static $disabled = null;
2282         if ( is_null( $disabled ) ) {
2283                 if ( !function_exists( 'proc_open' ) ) {
2284                         wfDebug( "proc_open() is disabled\n" );
2285                         $disabled = 'disabled';
2286                 } else {
2287                         $disabled = false;
2288                 }
2289         }
2290         return $disabled;
2291 }
2292
2293 /**
2294  * Execute a shell command, with time and memory limits mirrored from the PHP
2295  * configuration if supported.
2296  *
2297  * @param string|string[] $cmd If string, a properly shell-escaped command line,
2298  *   or an array of unescaped arguments, in which case each value will be escaped
2299  *   Example:   [ 'convert', '-font', 'font name' ] would produce "'convert' '-font' 'font name'"
2300  * @param null|mixed &$retval Optional, will receive the program's exit code.
2301  *   (non-zero is usually failure). If there is an error from
2302  *   read, select, or proc_open(), this will be set to -1.
2303  * @param array $environ Optional environment variables which should be
2304  *   added to the executed command environment.
2305  * @param array $limits Optional array with limits(filesize, memory, time, walltime)
2306  *   this overwrites the global wgMaxShell* limits.
2307  * @param array $options Array of options:
2308  *   - duplicateStderr: Set this to true to duplicate stderr to stdout,
2309  *     including errors from limit.sh
2310  *   - profileMethod: By default this function will profile based on the calling
2311  *     method. Set this to a string for an alternative method to profile from
2312  *
2313  * @return string Collected stdout as a string
2314  */
2315 function wfShellExec( $cmd, &$retval = null, $environ = [],
2316         $limits = [], $options = []
2317 ) {
2318         global $IP, $wgMaxShellMemory, $wgMaxShellFileSize, $wgMaxShellTime,
2319                 $wgMaxShellWallClockTime, $wgShellCgroup;
2320
2321         $disabled = wfShellExecDisabled();
2322         if ( $disabled ) {
2323                 $retval = 1;
2324                 return 'Unable to run external programs, proc_open() is disabled.';
2325         }
2326
2327         $includeStderr = isset( $options['duplicateStderr'] ) && $options['duplicateStderr'];
2328         $profileMethod = isset( $options['profileMethod'] ) ? $options['profileMethod'] : wfGetCaller();
2329
2330         wfInitShellLocale();
2331
2332         $envcmd = '';
2333         foreach ( $environ as $k => $v ) {
2334                 if ( wfIsWindows() ) {
2335                         /* Surrounding a set in quotes (method used by wfEscapeShellArg) makes the quotes themselves
2336                          * appear in the environment variable, so we must use carat escaping as documented in
2337                          * https://technet.microsoft.com/en-us/library/cc723564.aspx
2338                          * Note however that the quote isn't listed there, but is needed, and the parentheses
2339                          * are listed there but doesn't appear to need it.
2340                          */
2341                         $envcmd .= "set $k=" . preg_replace( '/([&|()<>^"])/', '^\\1', $v ) . '&& ';
2342                 } else {
2343                         /* Assume this is a POSIX shell, thus required to accept variable assignments before the command
2344                          * http://www.opengroup.org/onlinepubs/009695399/utilities/xcu_chap02.html#tag_02_09_01
2345                          */
2346                         $envcmd .= "$k=" . escapeshellarg( $v ) . ' ';
2347                 }
2348         }
2349         if ( is_array( $cmd ) ) {
2350                 $cmd = wfEscapeShellArg( $cmd );
2351         }
2352
2353         $cmd = $envcmd . $cmd;
2354
2355         $useLogPipe = false;
2356         if ( is_executable( '/bin/bash' ) ) {
2357                 $time = intval( isset( $limits['time'] ) ? $limits['time'] : $wgMaxShellTime );
2358                 if ( isset( $limits['walltime'] ) ) {
2359                         $wallTime = intval( $limits['walltime'] );
2360                 } elseif ( isset( $limits['time'] ) ) {
2361                         $wallTime = $time;
2362                 } else {
2363                         $wallTime = intval( $wgMaxShellWallClockTime );
2364                 }
2365                 $mem = intval( isset( $limits['memory'] ) ? $limits['memory'] : $wgMaxShellMemory );
2366                 $filesize = intval( isset( $limits['filesize'] ) ? $limits['filesize'] : $wgMaxShellFileSize );
2367
2368                 if ( $time > 0 || $mem > 0 || $filesize > 0 || $wallTime > 0 ) {
2369                         $cmd = '/bin/bash ' . escapeshellarg( "$IP/includes/limit.sh" ) . ' ' .
2370                                 escapeshellarg( $cmd ) . ' ' .
2371                                 escapeshellarg(
2372                                         "MW_INCLUDE_STDERR=" . ( $includeStderr ? '1' : '' ) . ';' .
2373                                         "MW_CPU_LIMIT=$time; " .
2374                                         'MW_CGROUP=' . escapeshellarg( $wgShellCgroup ) . '; ' .
2375                                         "MW_MEM_LIMIT=$mem; " .
2376                                         "MW_FILE_SIZE_LIMIT=$filesize; " .
2377                                         "MW_WALL_CLOCK_LIMIT=$wallTime; " .
2378                                         "MW_USE_LOG_PIPE=yes"
2379                                 );
2380                         $useLogPipe = true;
2381                 } elseif ( $includeStderr ) {
2382                         $cmd .= ' 2>&1';
2383                 }
2384         } elseif ( $includeStderr ) {
2385                 $cmd .= ' 2>&1';
2386         }
2387         wfDebug( "wfShellExec: $cmd\n" );
2388
2389         // Don't try to execute commands that exceed Linux's MAX_ARG_STRLEN.
2390         // Other platforms may be more accomodating, but we don't want to be
2391         // accomodating, because very long commands probably include user
2392         // input. See T129506.
2393         if ( strlen( $cmd ) > SHELL_MAX_ARG_STRLEN ) {
2394                 throw new Exception( __METHOD__ .
2395                         '(): total length of $cmd must not exceed SHELL_MAX_ARG_STRLEN' );
2396         }
2397
2398         $desc = [
2399                 0 => [ 'file', 'php://stdin', 'r' ],
2400                 1 => [ 'pipe', 'w' ],
2401                 2 => [ 'file', 'php://stderr', 'w' ] ];
2402         if ( $useLogPipe ) {
2403                 $desc[3] = [ 'pipe', 'w' ];
2404         }
2405         $pipes = null;
2406         $scoped = Profiler::instance()->scopedProfileIn( __FUNCTION__ . '-' . $profileMethod );
2407         $proc = proc_open( $cmd, $desc, $pipes );
2408         if ( !$proc ) {
2409                 wfDebugLog( 'exec', "proc_open() failed: $cmd" );
2410                 $retval = -1;
2411                 return '';
2412         }
2413         $outBuffer = $logBuffer = '';
2414         $emptyArray = [];
2415         $status = false;
2416         $logMsg = false;
2417
2418         /* According to the documentation, it is possible for stream_select()
2419          * to fail due to EINTR. I haven't managed to induce this in testing
2420          * despite sending various signals. If it did happen, the error
2421          * message would take the form:
2422          *
2423          * stream_select(): unable to select [4]: Interrupted system call (max_fd=5)
2424          *
2425          * where [4] is the value of the macro EINTR and "Interrupted system
2426          * call" is string which according to the Linux manual is "possibly"
2427          * localised according to LC_MESSAGES.
2428          */
2429         $eintr = defined( 'SOCKET_EINTR' ) ? SOCKET_EINTR : 4;
2430         $eintrMessage = "stream_select(): unable to select [$eintr]";
2431
2432         $running = true;
2433         $timeout = null;
2434         $numReadyPipes = 0;
2435
2436         while ( $running === true || $numReadyPipes !== 0 ) {
2437                 if ( $running ) {
2438                         $status = proc_get_status( $proc );
2439                         // If the process has terminated, switch to nonblocking selects
2440                         // for getting any data still waiting to be read.
2441                         if ( !$status['running'] ) {
2442                                 $running = false;
2443                                 $timeout = 0;
2444                         }
2445                 }
2446
2447                 $readyPipes = $pipes;
2448
2449                 // Clear last error
2450                 // @codingStandardsIgnoreStart Generic.PHP.NoSilencedErrors.Discouraged
2451                 @trigger_error( '' );
2452                 $numReadyPipes = @stream_select( $readyPipes, $emptyArray, $emptyArray, $timeout );
2453                 if ( $numReadyPipes === false ) {
2454                         // @codingStandardsIgnoreEnd
2455                         $error = error_get_last();
2456                         if ( strncmp( $error['message'], $eintrMessage, strlen( $eintrMessage ) ) == 0 ) {
2457                                 continue;
2458                         } else {
2459                                 trigger_error( $error['message'], E_USER_WARNING );
2460                                 $logMsg = $error['message'];
2461                                 break;
2462                         }
2463                 }
2464                 foreach ( $readyPipes as $fd => $pipe ) {
2465                         $block = fread( $pipe, 65536 );
2466                         if ( $block === '' ) {
2467                                 // End of file
2468                                 fclose( $pipes[$fd] );
2469                                 unset( $pipes[$fd] );
2470                                 if ( !$pipes ) {
2471                                         break 2;
2472                                 }
2473                         } elseif ( $block === false ) {
2474                                 // Read error
2475                                 $logMsg = "Error reading from pipe";
2476                                 break 2;
2477                         } elseif ( $fd == 1 ) {
2478                                 // From stdout
2479                                 $outBuffer .= $block;
2480                         } elseif ( $fd == 3 ) {
2481                                 // From log FD
2482                                 $logBuffer .= $block;
2483                                 if ( strpos( $block, "\n" ) !== false ) {
2484                                         $lines = explode( "\n", $logBuffer );
2485                                         $logBuffer = array_pop( $lines );
2486                                         foreach ( $lines as $line ) {
2487                                                 wfDebugLog( 'exec', $line );
2488                                         }
2489                                 }
2490                         }
2491                 }
2492         }
2493
2494         foreach ( $pipes as $pipe ) {
2495                 fclose( $pipe );
2496         }
2497
2498         // Use the status previously collected if possible, since proc_get_status()
2499         // just calls waitpid() which will not return anything useful the second time.
2500         if ( $running ) {
2501                 $status = proc_get_status( $proc );
2502         }
2503
2504         if ( $logMsg !== false ) {
2505                 // Read/select error
2506                 $retval = -1;
2507                 proc_close( $proc );
2508         } elseif ( $status['signaled'] ) {
2509                 $logMsg = "Exited with signal {$status['termsig']}";
2510                 $retval = 128 + $status['termsig'];
2511                 proc_close( $proc );
2512         } else {
2513                 if ( $status['running'] ) {
2514                         $retval = proc_close( $proc );
2515                 } else {
2516                         $retval = $status['exitcode'];
2517                         proc_close( $proc );
2518                 }
2519                 if ( $retval == 127 ) {
2520                         $logMsg = "Possibly missing executable file";
2521                 } elseif ( $retval >= 129 && $retval <= 192 ) {
2522                         $logMsg = "Probably exited with signal " . ( $retval - 128 );
2523                 }
2524         }
2525
2526         if ( $logMsg !== false ) {
2527                 wfDebugLog( 'exec', "$logMsg: $cmd" );
2528         }
2529
2530         return $outBuffer;
2531 }
2532
2533 /**
2534  * Execute a shell command, returning both stdout and stderr. Convenience
2535  * function, as all the arguments to wfShellExec can become unwieldy.
2536  *
2537  * @note This also includes errors from limit.sh, e.g. if $wgMaxShellFileSize is exceeded.
2538  * @param string|string[] $cmd If string, a properly shell-escaped command line,
2539  *   or an array of unescaped arguments, in which case each value will be escaped
2540  *   Example:   [ 'convert', '-font', 'font name' ] would produce "'convert' '-font' 'font name'"
2541  * @param null|mixed &$retval Optional, will receive the program's exit code.
2542  *   (non-zero is usually failure)
2543  * @param array $environ Optional environment variables which should be
2544  *   added to the executed command environment.
2545  * @param array $limits Optional array with limits(filesize, memory, time, walltime)
2546  *   this overwrites the global wgMaxShell* limits.
2547  * @return string Collected stdout and stderr as a string
2548  */
2549 function wfShellExecWithStderr( $cmd, &$retval = null, $environ = [], $limits = [] ) {
2550         return wfShellExec( $cmd, $retval, $environ, $limits,
2551                 [ 'duplicateStderr' => true, 'profileMethod' => wfGetCaller() ] );
2552 }
2553
2554 /**
2555  * Workaround for https://bugs.php.net/bug.php?id=45132
2556  * escapeshellarg() destroys non-ASCII characters if LANG is not a UTF-8 locale
2557  */
2558 function wfInitShellLocale() {
2559         static $done = false;
2560         if ( $done ) {
2561                 return;
2562         }
2563         $done = true;
2564         global $wgShellLocale;
2565         putenv( "LC_CTYPE=$wgShellLocale" );
2566         setlocale( LC_CTYPE, $wgShellLocale );
2567 }
2568
2569 /**
2570  * Generate a shell-escaped command line string to run a MediaWiki cli script.
2571  * Note that $parameters should be a flat array and an option with an argument
2572  * should consist of two consecutive items in the array (do not use "--option value").
2573  *
2574  * @param string $script MediaWiki cli script path
2575  * @param array $parameters Arguments and options to the script
2576  * @param array $options Associative array of options:
2577  *     'php': The path to the php executable
2578  *     'wrapper': Path to a PHP wrapper to handle the maintenance script
2579  * @return string
2580  */
2581 function wfShellWikiCmd( $script, array $parameters = [], array $options = [] ) {
2582         global $wgPhpCli;
2583         // Give site config file a chance to run the script in a wrapper.
2584         // The caller may likely want to call wfBasename() on $script.
2585         Hooks::run( 'wfShellWikiCmd', [ &$script, &$parameters, &$options ] );
2586         $cmd = isset( $options['php'] ) ? [ $options['php'] ] : [ $wgPhpCli ];
2587         if ( isset( $options['wrapper'] ) ) {
2588                 $cmd[] = $options['wrapper'];
2589         }
2590         $cmd[] = $script;
2591         // Escape each parameter for shell
2592         return wfEscapeShellArg( array_merge( $cmd, $parameters ) );
2593 }
2594
2595 /**
2596  * wfMerge attempts to merge differences between three texts.
2597  * Returns true for a clean merge and false for failure or a conflict.
2598  *
2599  * @param string $old
2600  * @param string $mine
2601  * @param string $yours
2602  * @param string $result
2603  * @return bool
2604  */
2605 function wfMerge( $old, $mine, $yours, &$result ) {
2606         global $wgDiff3;
2607
2608         # This check may also protect against code injection in
2609         # case of broken installations.
2610         MediaWiki\suppressWarnings();
2611         $haveDiff3 = $wgDiff3 && file_exists( $wgDiff3 );
2612         MediaWiki\restoreWarnings();
2613
2614         if ( !$haveDiff3 ) {
2615                 wfDebug( "diff3 not found\n" );
2616                 return false;
2617         }
2618
2619         # Make temporary files
2620         $td = wfTempDir();
2621         $oldtextFile = fopen( $oldtextName = tempnam( $td, 'merge-old-' ), 'w' );
2622         $mytextFile = fopen( $mytextName = tempnam( $td, 'merge-mine-' ), 'w' );
2623         $yourtextFile = fopen( $yourtextName = tempnam( $td, 'merge-your-' ), 'w' );
2624
2625         # NOTE: diff3 issues a warning to stderr if any of the files does not end with
2626         #       a newline character. To avoid this, we normalize the trailing whitespace before
2627         #       creating the diff.
2628
2629         fwrite( $oldtextFile, rtrim( $old ) . "\n" );
2630         fclose( $oldtextFile );
2631         fwrite( $mytextFile, rtrim( $mine ) . "\n" );
2632         fclose( $mytextFile );
2633         fwrite( $yourtextFile, rtrim( $yours ) . "\n" );
2634         fclose( $yourtextFile );
2635
2636         # Check for a conflict
2637         $cmd = wfEscapeShellArg( $wgDiff3, '-a', '--overlap-only', $mytextName,
2638                 $oldtextName, $yourtextName );
2639         $handle = popen( $cmd, 'r' );
2640
2641         if ( fgets( $handle, 1024 ) ) {
2642                 $conflict = true;
2643         } else {
2644                 $conflict = false;
2645         }
2646         pclose( $handle );
2647
2648         # Merge differences
2649         $cmd = wfEscapeShellArg( $wgDiff3, '-a', '-e', '--merge', $mytextName,
2650                 $oldtextName, $yourtextName );
2651         $handle = popen( $cmd, 'r' );
2652         $result = '';
2653         do {
2654                 $data = fread( $handle, 8192 );
2655                 if ( strlen( $data ) == 0 ) {
2656                         break;
2657                 }
2658                 $result .= $data;
2659         } while ( true );
2660         pclose( $handle );
2661         unlink( $mytextName );
2662         unlink( $oldtextName );
2663         unlink( $yourtextName );
2664
2665         if ( $result === '' && $old !== '' && !$conflict ) {
2666                 wfDebug( "Unexpected null result from diff3. Command: $cmd\n" );
2667                 $conflict = true;
2668         }
2669         return !$conflict;
2670 }
2671
2672 /**
2673  * Returns unified plain-text diff of two texts.
2674  * "Useful" for machine processing of diffs.
2675  *
2676  * @deprecated since 1.25, use DiffEngine/UnifiedDiffFormatter directly
2677  *
2678  * @param string $before The text before the changes.
2679  * @param string $after The text after the changes.
2680  * @param string $params Command-line options for the diff command.
2681  * @return string Unified diff of $before and $after
2682  */
2683 function wfDiff( $before, $after, $params = '-u' ) {
2684         if ( $before == $after ) {
2685                 return '';
2686         }
2687
2688         global $wgDiff;
2689         MediaWiki\suppressWarnings();
2690         $haveDiff = $wgDiff && file_exists( $wgDiff );
2691         MediaWiki\restoreWarnings();
2692
2693         # This check may also protect against code injection in
2694         # case of broken installations.
2695         if ( !$haveDiff ) {
2696                 wfDebug( "diff executable not found\n" );
2697                 $diffs = new Diff( explode( "\n", $before ), explode( "\n", $after ) );
2698                 $format = new UnifiedDiffFormatter();
2699                 return $format->format( $diffs );
2700         }
2701
2702         # Make temporary files
2703         $td = wfTempDir();
2704         $oldtextFile = fopen( $oldtextName = tempnam( $td, 'merge-old-' ), 'w' );
2705         $newtextFile = fopen( $newtextName = tempnam( $td, 'merge-your-' ), 'w' );
2706
2707         fwrite( $oldtextFile, $before );
2708         fclose( $oldtextFile );
2709         fwrite( $newtextFile, $after );
2710         fclose( $newtextFile );
2711
2712         // Get the diff of the two files
2713         $cmd = "$wgDiff " . $params . ' ' . wfEscapeShellArg( $oldtextName, $newtextName );
2714
2715         $h = popen( $cmd, 'r' );
2716         if ( !$h ) {
2717                 unlink( $oldtextName );
2718                 unlink( $newtextName );
2719                 throw new Exception( __METHOD__ . '(): popen() failed' );
2720         }
2721
2722         $diff = '';
2723
2724         do {
2725                 $data = fread( $h, 8192 );
2726                 if ( strlen( $data ) == 0 ) {
2727                         break;
2728                 }
2729                 $diff .= $data;
2730         } while ( true );
2731
2732         // Clean up
2733         pclose( $h );
2734         unlink( $oldtextName );
2735         unlink( $newtextName );
2736
2737         // Kill the --- and +++ lines. They're not useful.
2738         $diff_lines = explode( "\n", $diff );
2739         if ( isset( $diff_lines[0] ) && strpos( $diff_lines[0], '---' ) === 0 ) {
2740                 unset( $diff_lines[0] );
2741         }
2742         if ( isset( $diff_lines[1] ) && strpos( $diff_lines[1], '+++' ) === 0 ) {
2743                 unset( $diff_lines[1] );
2744         }
2745
2746         $diff = implode( "\n", $diff_lines );
2747
2748         return $diff;
2749 }
2750
2751 /**
2752  * This function works like "use VERSION" in Perl, the program will die with a
2753  * backtrace if the current version of PHP is less than the version provided
2754  *
2755  * This is useful for extensions which due to their nature are not kept in sync
2756  * with releases, and might depend on other versions of PHP than the main code
2757  *
2758  * Note: PHP might die due to parsing errors in some cases before it ever
2759  *       manages to call this function, such is life
2760  *
2761  * @see perldoc -f use
2762  *
2763  * @param string|int|float $req_ver The version to check, can be a string, an integer, or a float
2764  * @throws MWException
2765  */
2766 function wfUsePHP( $req_ver ) {
2767         $php_ver = PHP_VERSION;
2768
2769         if ( version_compare( $php_ver, (string)$req_ver, '<' ) ) {
2770                 throw new MWException( "PHP $req_ver required--this is only $php_ver" );
2771         }
2772 }
2773
2774 /**
2775  * This function works like "use VERSION" in Perl except it checks the version
2776  * of MediaWiki, the program will die with a backtrace if the current version
2777  * of MediaWiki is less than the version provided.
2778  *
2779  * This is useful for extensions which due to their nature are not kept in sync
2780  * with releases
2781  *
2782  * Note: Due to the behavior of PHP's version_compare() which is used in this
2783  * function, if you want to allow the 'wmf' development versions add a 'c' (or
2784  * any single letter other than 'a', 'b' or 'p') as a post-fix to your
2785  * targeted version number. For example if you wanted to allow any variation
2786  * of 1.22 use `wfUseMW( '1.22c' )`. Using an 'a' or 'b' instead of 'c' will
2787  * not result in the same comparison due to the internal logic of
2788  * version_compare().
2789  *
2790  * @see perldoc -f use
2791  *
2792  * @deprecated since 1.26, use the "requires' property of extension.json
2793  * @param string|int|float $req_ver The version to check, can be a string, an integer, or a float
2794  * @throws MWException
2795  */
2796 function wfUseMW( $req_ver ) {
2797         global $wgVersion;
2798
2799         if ( version_compare( $wgVersion, (string)$req_ver, '<' ) ) {
2800                 throw new MWException( "MediaWiki $req_ver required--this is only $wgVersion" );
2801         }
2802 }
2803
2804 /**
2805  * Return the final portion of a pathname.
2806  * Reimplemented because PHP5's "basename()" is buggy with multibyte text.
2807  * https://bugs.php.net/bug.php?id=33898
2808  *
2809  * PHP's basename() only considers '\' a pathchar on Windows and Netware.
2810  * We'll consider it so always, as we don't want '\s' in our Unix paths either.
2811  *
2812  * @param string $path
2813  * @param string $suffix String to remove if present
2814  * @return string
2815  */
2816 function wfBaseName( $path, $suffix = '' ) {
2817         if ( $suffix == '' ) {
2818                 $encSuffix = '';
2819         } else {
2820                 $encSuffix = '(?:' . preg_quote( $suffix, '#' ) . ')?';
2821         }
2822
2823         $matches = [];
2824         if ( preg_match( "#([^/\\\\]*?){$encSuffix}[/\\\\]*$#", $path, $matches ) ) {
2825                 return $matches[1];
2826         } else {
2827                 return '';
2828         }
2829 }
2830
2831 /**
2832  * Generate a relative path name to the given file.
2833  * May explode on non-matching case-insensitive paths,
2834  * funky symlinks, etc.
2835  *
2836  * @param string $path Absolute destination path including target filename
2837  * @param string $from Absolute source path, directory only
2838  * @return string
2839  */
2840 function wfRelativePath( $path, $from ) {
2841         // Normalize mixed input on Windows...
2842         $path = str_replace( '/', DIRECTORY_SEPARATOR, $path );
2843         $from = str_replace( '/', DIRECTORY_SEPARATOR, $from );
2844
2845         // Trim trailing slashes -- fix for drive root
2846         $path = rtrim( $path, DIRECTORY_SEPARATOR );
2847         $from = rtrim( $from, DIRECTORY_SEPARATOR );
2848
2849         $pieces = explode( DIRECTORY_SEPARATOR, dirname( $path ) );
2850         $against = explode( DIRECTORY_SEPARATOR, $from );
2851
2852         if ( $pieces[0] !== $against[0] ) {
2853                 // Non-matching Windows drive letters?
2854                 // Return a full path.
2855                 return $path;
2856         }
2857
2858         // Trim off common prefix
2859         while ( count( $pieces ) && count( $against )
2860                 && $pieces[0] == $against[0] ) {
2861                 array_shift( $pieces );
2862                 array_shift( $against );
2863         }
2864
2865         // relative dots to bump us to the parent
2866         while ( count( $against ) ) {
2867                 array_unshift( $pieces, '..' );
2868                 array_shift( $against );
2869         }
2870
2871         array_push( $pieces, wfBaseName( $path ) );
2872
2873         return implode( DIRECTORY_SEPARATOR, $pieces );
2874 }
2875
2876 /**
2877  * Convert an arbitrarily-long digit string from one numeric base
2878  * to another, optionally zero-padding to a minimum column width.
2879  *
2880  * Supports base 2 through 36; digit values 10-36 are represented
2881  * as lowercase letters a-z. Input is case-insensitive.
2882  *
2883  * @deprecated since 1.27 Use Wikimedia\base_convert() directly
2884  *
2885  * @param string $input Input number
2886  * @param int $sourceBase Base of the input number
2887  * @param int $destBase Desired base of the output
2888  * @param int $pad Minimum number of digits in the output (pad with zeroes)
2889  * @param bool $lowercase Whether to output in lowercase or uppercase
2890  * @param string $engine Either "gmp", "bcmath", or "php"
2891  * @return string|bool The output number as a string, or false on error
2892  */
2893 function wfBaseConvert( $input, $sourceBase, $destBase, $pad = 1,
2894         $lowercase = true, $engine = 'auto'
2895 ) {
2896         return Wikimedia\base_convert( $input, $sourceBase, $destBase, $pad, $lowercase, $engine );
2897 }
2898
2899 /**
2900  * @deprecated since 1.27, PHP's session generation isn't used with
2901  *  MediaWiki\Session\SessionManager
2902  */
2903 function wfFixSessionID() {
2904         wfDeprecated( __FUNCTION__, '1.27' );
2905 }
2906
2907 /**
2908  * Reset the session id
2909  *
2910  * @deprecated since 1.27, use MediaWiki\Session\SessionManager instead
2911  * @since 1.22
2912  */
2913 function wfResetSessionID() {
2914         wfDeprecated( __FUNCTION__, '1.27' );
2915         $session = SessionManager::getGlobalSession();
2916         $delay = $session->delaySave();
2917
2918         $session->resetId();
2919
2920         // Make sure a session is started, since that's what the old
2921         // wfResetSessionID() did.
2922         if ( session_id() !== $session->getId() ) {
2923                 wfSetupSession( $session->getId() );
2924         }
2925
2926         ScopedCallback::consume( $delay );
2927 }
2928
2929 /**
2930  * Initialise php session
2931  *
2932  * @deprecated since 1.27, use MediaWiki\Session\SessionManager instead.
2933  *  Generally, "using" SessionManager will be calling ->getSessionById() or
2934  *  ::getGlobalSession() (depending on whether you were passing $sessionId
2935  *  here), then calling $session->persist().
2936  * @param bool|string $sessionId
2937  */
2938 function wfSetupSession( $sessionId = false ) {
2939         wfDeprecated( __FUNCTION__, '1.27' );
2940
2941         if ( $sessionId ) {
2942                 session_id( $sessionId );
2943         }
2944
2945         $session = SessionManager::getGlobalSession();
2946         $session->persist();
2947
2948         if ( session_id() !== $session->getId() ) {
2949                 session_id( $session->getId() );
2950         }
2951         MediaWiki\quietCall( 'session_start' );
2952 }
2953
2954 /**
2955  * Get an object from the precompiled serialized directory
2956  *
2957  * @param string $name
2958  * @return mixed The variable on success, false on failure
2959  */
2960 function wfGetPrecompiledData( $name ) {
2961         global $IP;
2962
2963         $file = "$IP/serialized/$name";
2964         if ( file_exists( $file ) ) {
2965                 $blob = file_get_contents( $file );
2966                 if ( $blob ) {
2967                         return unserialize( $blob );
2968                 }
2969         }
2970         return false;
2971 }
2972
2973 /**
2974  * Make a cache key for the local wiki.
2975  *
2976  * @param string $args,...
2977  * @return string
2978  */
2979 function wfMemcKey( /*...*/ ) {
2980         return call_user_func_array(
2981                 [ ObjectCache::getLocalClusterInstance(), 'makeKey' ],
2982                 func_get_args()
2983         );
2984 }
2985
2986 /**
2987  * Make a cache key for a foreign DB.
2988  *
2989  * Must match what wfMemcKey() would produce in context of the foreign wiki.
2990  *
2991  * @param string $db
2992  * @param string $prefix
2993  * @param string $args,...
2994  * @return string
2995  */
2996 function wfForeignMemcKey( $db, $prefix /*...*/ ) {
2997         $args = array_slice( func_get_args(), 2 );
2998         $keyspace = $prefix ? "$db-$prefix" : $db;
2999         return call_user_func_array(
3000                 [ ObjectCache::getLocalClusterInstance(), 'makeKeyInternal' ],
3001                 [ $keyspace, $args ]
3002         );
3003 }
3004
3005 /**
3006  * Make a cache key with database-agnostic prefix.
3007  *
3008  * Doesn't have a wiki-specific namespace. Uses a generic 'global' prefix
3009  * instead. Must have a prefix as otherwise keys that use a database name
3010  * in the first segment will clash with wfMemcKey/wfForeignMemcKey.
3011  *
3012  * @since 1.26
3013  * @param string $args,...
3014  * @return string
3015  */
3016 function wfGlobalCacheKey( /*...*/ ) {
3017         return call_user_func_array(
3018                 [ ObjectCache::getLocalClusterInstance(), 'makeGlobalKey' ],
3019                 func_get_args()
3020         );
3021 }
3022
3023 /**
3024  * Get an ASCII string identifying this wiki
3025  * This is used as a prefix in memcached keys
3026  *
3027  * @return string
3028  */
3029 function wfWikiID() {
3030         global $wgDBprefix, $wgDBname;
3031         if ( $wgDBprefix ) {
3032                 return "$wgDBname-$wgDBprefix";
3033         } else {
3034                 return $wgDBname;
3035         }
3036 }
3037
3038 /**
3039  * Split a wiki ID into DB name and table prefix
3040  *
3041  * @param string $wiki
3042  *
3043  * @return array
3044  */
3045 function wfSplitWikiID( $wiki ) {
3046         $bits = explode( '-', $wiki, 2 );
3047         if ( count( $bits ) < 2 ) {
3048                 $bits[] = '';
3049         }
3050         return $bits;
3051 }
3052
3053 /**
3054  * Get a Database object.
3055  *
3056  * @param int $db Index of the connection to get. May be DB_MASTER for the
3057  *            master (for write queries), DB_REPLICA for potentially lagged read
3058  *            queries, or an integer >= 0 for a particular server.
3059  *
3060  * @param string|string[] $groups Query groups. An array of group names that this query
3061  *                belongs to. May contain a single string if the query is only
3062  *                in one group.
3063  *
3064  * @param string|bool $wiki The wiki ID, or false for the current wiki
3065  *
3066  * Note: multiple calls to wfGetDB(DB_REPLICA) during the course of one request
3067  * will always return the same object, unless the underlying connection or load
3068  * balancer is manually destroyed.
3069  *
3070  * Note 2: use $this->getDB() in maintenance scripts that may be invoked by
3071  * updater to ensure that a proper database is being updated.
3072  *
3073  * @todo Replace calls to wfGetDB with calls to LoadBalancer::getConnection()
3074  *       on an injected instance of LoadBalancer.
3075  *
3076  * @return \Wikimedia\Rdbms\Database
3077  */
3078 function wfGetDB( $db, $groups = [], $wiki = false ) {
3079         return wfGetLB( $wiki )->getConnection( $db, $groups, $wiki );
3080 }
3081
3082 /**
3083  * Get a load balancer object.
3084  *
3085  * @deprecated since 1.27, use MediaWikiServices::getDBLoadBalancer()
3086  *              or MediaWikiServices::getDBLoadBalancerFactory() instead.
3087  *
3088  * @param string|bool $wiki Wiki ID, or false for the current wiki
3089  * @return \Wikimedia\Rdbms\LoadBalancer
3090  */
3091 function wfGetLB( $wiki = false ) {
3092         if ( $wiki === false ) {
3093                 return \MediaWiki\MediaWikiServices::getInstance()->getDBLoadBalancer();
3094         } else {
3095                 $factory = \MediaWiki\MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
3096                 return $factory->getMainLB( $wiki );
3097         }
3098 }
3099
3100 /**
3101  * Get the load balancer factory object
3102  *
3103  * @deprecated since 1.27, use MediaWikiServices::getDBLoadBalancerFactory() instead.
3104  *
3105  * @return \Wikimedia\Rdbms\LBFactory
3106  */
3107 function wfGetLBFactory() {
3108         return \MediaWiki\MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
3109 }
3110
3111 /**
3112  * Find a file.
3113  * Shortcut for RepoGroup::singleton()->findFile()
3114  *
3115  * @param string $title String or Title object
3116  * @param array $options Associative array of options (see RepoGroup::findFile)
3117  * @return File|bool File, or false if the file does not exist
3118  */
3119 function wfFindFile( $title, $options = [] ) {
3120         return RepoGroup::singleton()->findFile( $title, $options );
3121 }
3122
3123 /**
3124  * Get an object referring to a locally registered file.
3125  * Returns a valid placeholder object if the file does not exist.
3126  *
3127  * @param Title|string $title
3128  * @return LocalFile|null A File, or null if passed an invalid Title
3129  */
3130 function wfLocalFile( $title ) {
3131         return RepoGroup::singleton()->getLocalRepo()->newFile( $title );
3132 }
3133
3134 /**
3135  * Should low-performance queries be disabled?
3136  *
3137  * @return bool
3138  * @codeCoverageIgnore
3139  */
3140 function wfQueriesMustScale() {
3141         global $wgMiserMode;
3142         return $wgMiserMode
3143                 || ( SiteStats::pages() > 100000
3144                 && SiteStats::edits() > 1000000
3145                 && SiteStats::users() > 10000 );
3146 }
3147
3148 /**
3149  * Get the path to a specified script file, respecting file
3150  * extensions; this is a wrapper around $wgScriptPath etc.
3151  * except for 'index' and 'load' which use $wgScript/$wgLoadScript
3152  *
3153  * @param string $script Script filename, sans extension
3154  * @return string
3155  */
3156 function wfScript( $script = 'index' ) {
3157         global $wgScriptPath, $wgScript, $wgLoadScript;
3158         if ( $script === 'index' ) {
3159                 return $wgScript;
3160         } elseif ( $script === 'load' ) {
3161                 return $wgLoadScript;
3162         } else {
3163                 return "{$wgScriptPath}/{$script}.php";
3164         }
3165 }
3166
3167 /**
3168  * Get the script URL.
3169  *
3170  * @return string Script URL
3171  */
3172 function wfGetScriptUrl() {
3173         if ( isset( $_SERVER['SCRIPT_NAME'] ) ) {
3174                 /* as it was called, minus the query string.
3175                  *
3176                  * Some sites use Apache rewrite rules to handle subdomains,
3177                  * and have PHP set up in a weird way that causes PHP_SELF
3178                  * to contain the rewritten URL instead of the one that the
3179                  * outside world sees.
3180                  *
3181                  * If in this mode, use SCRIPT_URL instead, which mod_rewrite
3182                  * provides containing the "before" URL.
3183                  */
3184                 return $_SERVER['SCRIPT_NAME'];
3185         } else {
3186                 return $_SERVER['URL'];
3187         }
3188 }
3189
3190 /**
3191  * Convenience function converts boolean values into "true"
3192  * or "false" (string) values
3193  *
3194  * @param bool $value
3195  * @return string
3196  */
3197 function wfBoolToStr( $value ) {
3198         return $value ? 'true' : 'false';
3199 }
3200
3201 /**
3202  * Get a platform-independent path to the null file, e.g. /dev/null
3203  *
3204  * @return string
3205  */
3206 function wfGetNull() {
3207         return wfIsWindows() ? 'NUL' : '/dev/null';
3208 }
3209
3210 /**
3211  * Waits for the replica DBs to catch up to the master position
3212  *
3213  * Use this when updating very large numbers of rows, as in maintenance scripts,
3214  * to avoid causing too much lag. Of course, this is a no-op if there are no replica DBs.
3215  *
3216  * By default this waits on the main DB cluster of the current wiki.
3217  * If $cluster is set to "*" it will wait on all DB clusters, including
3218  * external ones. If the lag being waiting on is caused by the code that
3219  * does this check, it makes since to use $ifWritesSince, particularly if
3220  * cluster is "*", to avoid excess overhead.
3221  *
3222  * Never call this function after a big DB write that is still in a transaction.
3223  * This only makes sense after the possible lag inducing changes were committed.
3224  *
3225  * @param float|null $ifWritesSince Only wait if writes were done since this UNIX timestamp
3226  * @param string|bool $wiki Wiki identifier accepted by wfGetLB
3227  * @param string|bool $cluster Cluster name accepted by LBFactory. Default: false.
3228  * @param int|null $timeout Max wait time. Default: 1 day (cli), ~10 seconds (web)
3229  * @return bool Success (able to connect and no timeouts reached)
3230  * @deprecated since 1.27 Use LBFactory::waitForReplication
3231  */
3232 function wfWaitForSlaves(
3233         $ifWritesSince = null, $wiki = false, $cluster = false, $timeout = null
3234 ) {
3235         if ( $timeout === null ) {
3236                 $timeout = ( PHP_SAPI === 'cli' ) ? 86400 : 10;
3237         }
3238
3239         if ( $cluster === '*' ) {
3240                 $cluster = false;
3241                 $wiki = false;
3242         } elseif ( $wiki === false ) {
3243                 $wiki = wfWikiID();
3244         }
3245
3246         try {
3247                 wfGetLBFactory()->waitForReplication( [
3248                         'wiki' => $wiki,
3249                         'cluster' => $cluster,
3250                         'timeout' => $timeout,
3251                         // B/C: first argument used to be "max seconds of lag"; ignore such values
3252                         'ifWritesSince' => ( $ifWritesSince > 1e9 ) ? $ifWritesSince : null
3253                 ] );
3254         } catch ( DBReplicationWaitError $e ) {
3255                 return false;
3256         }
3257
3258         return true;
3259 }
3260
3261 /**
3262  * Count down from $seconds to zero on the terminal, with a one-second pause
3263  * between showing each number. For use in command-line scripts.
3264  *
3265  * @codeCoverageIgnore
3266  * @param int $seconds
3267  */
3268 function wfCountDown( $seconds ) {
3269         for ( $i = $seconds; $i >= 0; $i-- ) {
3270                 if ( $i != $seconds ) {
3271                         echo str_repeat( "\x08", strlen( $i + 1 ) );
3272                 }
3273                 echo $i;
3274                 flush();
3275                 if ( $i ) {
3276                         sleep( 1 );
3277                 }
3278         }
3279         echo "\n";
3280 }
3281
3282 /**
3283  * Replace all invalid characters with '-'.
3284  * Additional characters can be defined in $wgIllegalFileChars (see T22489).
3285  * By default, $wgIllegalFileChars includes ':', '/', '\'.
3286  *
3287  * @param string $name Filename to process
3288  * @return string
3289  */
3290 function wfStripIllegalFilenameChars( $name ) {
3291         global $wgIllegalFileChars;
3292         $illegalFileChars = $wgIllegalFileChars ? "|[" . $wgIllegalFileChars . "]" : '';
3293         $name = preg_replace(
3294                 "/[^" . Title::legalChars() . "]" . $illegalFileChars . "/",
3295                 '-',
3296                 $name
3297         );
3298         // $wgIllegalFileChars may not include '/' and '\', so we still need to do this
3299         $name = wfBaseName( $name );
3300         return $name;
3301 }
3302
3303 /**
3304  * Set PHP's memory limit to the larger of php.ini or $wgMemoryLimit
3305  *
3306  * @return int Resulting value of the memory limit.
3307  */
3308 function wfMemoryLimit() {
3309         global $wgMemoryLimit;
3310         $memlimit = wfShorthandToInteger( ini_get( 'memory_limit' ) );
3311         if ( $memlimit != -1 ) {
3312                 $conflimit = wfShorthandToInteger( $wgMemoryLimit );
3313                 if ( $conflimit == -1 ) {
3314                         wfDebug( "Removing PHP's memory limit\n" );
3315                         MediaWiki\suppressWarnings();
3316                         ini_set( 'memory_limit', $conflimit );
3317                         MediaWiki\restoreWarnings();
3318                         return $conflimit;
3319                 } elseif ( $conflimit > $memlimit ) {
3320                         wfDebug( "Raising PHP's memory limit to $conflimit bytes\n" );
3321                         MediaWiki\suppressWarnings();
3322                         ini_set( 'memory_limit', $conflimit );
3323                         MediaWiki\restoreWarnings();
3324                         return $conflimit;
3325                 }
3326         }
3327         return $memlimit;
3328 }
3329
3330 /**
3331  * Set PHP's time limit to the larger of php.ini or $wgTransactionalTimeLimit
3332  *
3333  * @return int Prior time limit
3334  * @since 1.26
3335  */
3336 function wfTransactionalTimeLimit() {
3337         global $wgTransactionalTimeLimit;
3338
3339         $timeLimit = ini_get( 'max_execution_time' );
3340         // Note that CLI scripts use 0
3341         if ( $timeLimit > 0 && $wgTransactionalTimeLimit > $timeLimit ) {
3342                 set_time_limit( $wgTransactionalTimeLimit );
3343         }
3344
3345         ignore_user_abort( true ); // ignore client disconnects
3346
3347         return $timeLimit;
3348 }
3349
3350 /**
3351  * Converts shorthand byte notation to integer form
3352  *
3353  * @param string $string
3354  * @param int $default Returned if $string is empty
3355  * @return int
3356  */
3357 function wfShorthandToInteger( $string = '', $default = -1 ) {
3358         $string = trim( $string );
3359         if ( $string === '' ) {
3360                 return $default;
3361         }
3362         $last = $string[strlen( $string ) - 1];
3363         $val = intval( $string );
3364         switch ( $last ) {
3365                 case 'g':
3366                 case 'G':
3367                         $val *= 1024;
3368                         // break intentionally missing
3369                 case 'm':
3370                 case 'M':
3371                         $val *= 1024;
3372                         // break intentionally missing
3373                 case 'k':
3374                 case 'K':
3375                         $val *= 1024;
3376         }
3377
3378         return $val;
3379 }
3380
3381 /**
3382  * Get the normalised IETF language tag
3383  * See unit test for examples.
3384  *
3385  * @param string $code The language code.
3386  * @return string The language code which complying with BCP 47 standards.
3387  */
3388 function wfBCP47( $code ) {
3389         $codeSegment = explode( '-', $code );
3390         $codeBCP = [];
3391         foreach ( $codeSegment as $segNo => $seg ) {
3392                 // when previous segment is x, it is a private segment and should be lc
3393                 if ( $segNo > 0 && strtolower( $codeSegment[( $segNo - 1 )] ) == 'x' ) {
3394                         $codeBCP[$segNo] = strtolower( $seg );
3395                 // ISO 3166 country code
3396                 } elseif ( ( strlen( $seg ) == 2 ) && ( $segNo > 0 ) ) {
3397                         $codeBCP[$segNo] = strtoupper( $seg );
3398                 // ISO 15924 script code
3399                 } elseif ( ( strlen( $seg ) == 4 ) && ( $segNo > 0 ) ) {
3400                         $codeBCP[$segNo] = ucfirst( strtolower( $seg ) );
3401                 // Use lowercase for other cases
3402                 } else {
3403                         $codeBCP[$segNo] = strtolower( $seg );
3404                 }
3405         }
3406         $langCode = implode( '-', $codeBCP );
3407         return $langCode;
3408 }
3409
3410 /**
3411  * Get a specific cache object.
3412  *
3413  * @param int|string $cacheType A CACHE_* constants, or other key in $wgObjectCaches
3414  * @return BagOStuff
3415  */
3416 function wfGetCache( $cacheType ) {
3417         return ObjectCache::getInstance( $cacheType );
3418 }
3419
3420 /**
3421  * Get the main cache object
3422  *
3423  * @return BagOStuff
3424  */
3425 function wfGetMainCache() {
3426         global $wgMainCacheType;
3427         return ObjectCache::getInstance( $wgMainCacheType );
3428 }
3429
3430 /**
3431  * Get the cache object used by the message cache
3432  *
3433  * @return BagOStuff
3434  */
3435 function wfGetMessageCacheStorage() {
3436         global $wgMessageCacheType;
3437         return ObjectCache::getInstance( $wgMessageCacheType );
3438 }
3439
3440 /**
3441  * Get the cache object used by the parser cache
3442  *
3443  * @return BagOStuff
3444  */
3445 function wfGetParserCacheStorage() {
3446         global $wgParserCacheType;
3447         return ObjectCache::getInstance( $wgParserCacheType );
3448 }
3449
3450 /**
3451  * Call hook functions defined in $wgHooks
3452  *
3453  * @param string $event Event name
3454  * @param array $args Parameters passed to hook functions
3455  * @param string|null $deprecatedVersion Optionally mark hook as deprecated with version number
3456  *
3457  * @return bool True if no handler aborted the hook
3458  * @deprecated since 1.25 - use Hooks::run
3459  */
3460 function wfRunHooks( $event, array $args = [], $deprecatedVersion = null ) {
3461         return Hooks::run( $event, $args, $deprecatedVersion );
3462 }
3463
3464 /**
3465  * Wrapper around php's unpack.
3466  *
3467  * @param string $format The format string (See php's docs)
3468  * @param string $data A binary string of binary data
3469  * @param int|bool $length The minimum length of $data or false. This is to
3470  *      prevent reading beyond the end of $data. false to disable the check.
3471  *
3472  * Also be careful when using this function to read unsigned 32 bit integer
3473  * because php might make it negative.
3474  *
3475  * @throws MWException If $data not long enough, or if unpack fails
3476  * @return array Associative array of the extracted data
3477  */
3478 function wfUnpack( $format, $data, $length = false ) {
3479         if ( $length !== false ) {
3480                 $realLen = strlen( $data );
3481                 if ( $realLen < $length ) {
3482                         throw new MWException( "Tried to use wfUnpack on a "
3483                                 . "string of length $realLen, but needed one "
3484                                 . "of at least length $length."
3485                         );
3486                 }
3487         }
3488
3489         MediaWiki\suppressWarnings();
3490         $result = unpack( $format, $data );
3491         MediaWiki\restoreWarnings();
3492
3493         if ( $result === false ) {
3494                 // If it cannot extract the packed data.
3495                 throw new MWException( "unpack could not unpack binary data" );
3496         }
3497         return $result;
3498 }
3499
3500 /**
3501  * Determine if an image exists on the 'bad image list'.
3502  *
3503  * The format of MediaWiki:Bad_image_list is as follows:
3504  *    * Only list items (lines starting with "*") are considered
3505  *    * The first link on a line must be a link to a bad image
3506  *    * Any subsequent links on the same line are considered to be exceptions,
3507  *      i.e. articles where the image may occur inline.
3508  *
3509  * @param string $name The image name to check
3510  * @param Title|bool $contextTitle The page on which the image occurs, if known
3511  * @param string $blacklist Wikitext of a file blacklist
3512  * @return bool
3513  */
3514 function wfIsBadImage( $name, $contextTitle = false, $blacklist = null ) {
3515         # Handle redirects; callers almost always hit wfFindFile() anyway,
3516         # so just use that method because it has a fast process cache.
3517         $file = wfFindFile( $name ); // get the final name
3518         $name = $file ? $file->getTitle()->getDBkey() : $name;
3519
3520         # Run the extension hook
3521         $bad = false;
3522         if ( !Hooks::run( 'BadImage', [ $name, &$bad ] ) ) {
3523                 return (bool)$bad;
3524         }
3525
3526         $cache = ObjectCache::getLocalServerInstance( 'hash' );
3527         $key = wfMemcKey( 'bad-image-list', ( $blacklist === null ) ? 'default' : md5( $blacklist ) );
3528         $badImages = $cache->get( $key );
3529
3530         if ( $badImages === false ) { // cache miss
3531                 if ( $blacklist === null ) {
3532                         $blacklist = wfMessage( 'bad_image_list' )->inContentLanguage()->plain(); // site list
3533                 }
3534                 # Build the list now
3535                 $badImages = [];
3536                 $lines = explode( "\n", $blacklist );
3537                 foreach ( $lines as $line ) {
3538                         # List items only
3539                         if ( substr( $line, 0, 1 ) !== '*' ) {
3540                                 continue;
3541                         }
3542
3543                         # Find all links
3544                         $m = [];
3545                         if ( !preg_match_all( '/\[\[:?(.*?)\]\]/', $line, $m ) ) {
3546                                 continue;
3547                         }
3548
3549                         $exceptions = [];
3550                         $imageDBkey = false;
3551                         foreach ( $m[1] as $i => $titleText ) {
3552                                 $title = Title::newFromText( $titleText );
3553                                 if ( !is_null( $title ) ) {
3554                                         if ( $i == 0 ) {
3555                                                 $imageDBkey = $title->getDBkey();
3556                                         } else {
3557                                                 $exceptions[$title->getPrefixedDBkey()] = true;
3558                                         }
3559                                 }
3560                         }
3561
3562                         if ( $imageDBkey !== false ) {
3563                                 $badImages[$imageDBkey] = $exceptions;
3564                         }
3565                 }
3566                 $cache->set( $key, $badImages, 60 );
3567         }
3568
3569         $contextKey = $contextTitle ? $contextTitle->getPrefixedDBkey() : false;
3570         $bad = isset( $badImages[$name] ) && !isset( $badImages[$name][$contextKey] );
3571
3572         return $bad;
3573 }
3574
3575 /**
3576  * Determine whether the client at a given source IP is likely to be able to
3577  * access the wiki via HTTPS.
3578  *
3579  * @param string $ip The IPv4/6 address in the normal human-readable form
3580  * @return bool
3581  */
3582 function wfCanIPUseHTTPS( $ip ) {
3583         $canDo = true;
3584         Hooks::run( 'CanIPUseHTTPS', [ $ip, &$canDo ] );
3585         return !!$canDo;
3586 }
3587
3588 /**
3589  * Determine input string is represents as infinity
3590  *
3591  * @param string $str The string to determine
3592  * @return bool
3593  * @since 1.25
3594  */
3595 function wfIsInfinity( $str ) {
3596         $infinityValues = [ 'infinite', 'indefinite', 'infinity', 'never' ];
3597         return in_array( $str, $infinityValues );
3598 }
3599
3600 /**
3601  * Returns true if these thumbnail parameters match one that MediaWiki
3602  * requests from file description pages and/or parser output.
3603  *
3604  * $params is considered non-standard if they involve a non-standard
3605  * width or any non-default parameters aside from width and page number.
3606  * The number of possible files with standard parameters is far less than
3607  * that of all combinations; rate-limiting for them can thus be more generious.
3608  *
3609  * @param File $file
3610  * @param array $params
3611  * @return bool
3612  * @since 1.24 Moved from thumb.php to GlobalFunctions in 1.25
3613  */
3614 function wfThumbIsStandard( File $file, array $params ) {
3615         global $wgThumbLimits, $wgImageLimits, $wgResponsiveImages;
3616
3617         $multipliers = [ 1 ];
3618         if ( $wgResponsiveImages ) {
3619                 // These available sizes are hardcoded currently elsewhere in MediaWiki.
3620                 // @see Linker::processResponsiveImages
3621                 $multipliers[] = 1.5;
3622                 $multipliers[] = 2;
3623         }
3624
3625         $handler = $file->getHandler();
3626         if ( !$handler || !isset( $params['width'] ) ) {
3627                 return false;
3628         }
3629
3630         $basicParams = [];
3631         if ( isset( $params['page'] ) ) {
3632                 $basicParams['page'] = $params['page'];
3633         }
3634
3635         $thumbLimits = [];
3636         $imageLimits = [];
3637         // Expand limits to account for multipliers
3638         foreach ( $multipliers as $multiplier ) {
3639                 $thumbLimits = array_merge( $thumbLimits, array_map(
3640                         function ( $width ) use ( $multiplier ) {
3641                                 return round( $width * $multiplier );
3642                         }, $wgThumbLimits )
3643                 );
3644                 $imageLimits = array_merge( $imageLimits, array_map(
3645                         function ( $pair ) use ( $multiplier ) {
3646                                 return [
3647                                         round( $pair[0] * $multiplier ),
3648                                         round( $pair[1] * $multiplier ),
3649                                 ];
3650                         }, $wgImageLimits )
3651                 );
3652         }
3653
3654         // Check if the width matches one of $wgThumbLimits
3655         if ( in_array( $params['width'], $thumbLimits ) ) {
3656                 $normalParams = $basicParams + [ 'width' => $params['width'] ];
3657                 // Append any default values to the map (e.g. "lossy", "lossless", ...)
3658                 $handler->normaliseParams( $file, $normalParams );
3659         } else {
3660                 // If not, then check if the width matchs one of $wgImageLimits
3661                 $match = false;
3662                 foreach ( $imageLimits as $pair ) {
3663                         $normalParams = $basicParams + [ 'width' => $pair[0], 'height' => $pair[1] ];
3664                         // Decide whether the thumbnail should be scaled on width or height.
3665                         // Also append any default values to the map (e.g. "lossy", "lossless", ...)
3666                         $handler->normaliseParams( $file, $normalParams );
3667                         // Check if this standard thumbnail size maps to the given width
3668                         if ( $normalParams['width'] == $params['width'] ) {
3669                                 $match = true;
3670                                 break;
3671                         }
3672                 }
3673                 if ( !$match ) {
3674                         return false; // not standard for description pages
3675                 }
3676         }
3677
3678         // Check that the given values for non-page, non-width, params are just defaults
3679         foreach ( $params as $key => $value ) {
3680                 if ( !isset( $normalParams[$key] ) || $normalParams[$key] != $value ) {
3681                         return false;
3682                 }
3683         }
3684
3685         return true;
3686 }
3687
3688 /**
3689  * Merges two (possibly) 2 dimensional arrays into the target array ($baseArray).
3690  *
3691  * Values that exist in both values will be combined with += (all values of the array
3692  * of $newValues will be added to the values of the array of $baseArray, while values,
3693  * that exists in both, the value of $baseArray will be used).
3694  *
3695  * @param array $baseArray The array where you want to add the values of $newValues to
3696  * @param array $newValues An array with new values
3697  * @return array The combined array
3698  * @since 1.26
3699  */
3700 function wfArrayPlus2d( array $baseArray, array $newValues ) {
3701         // First merge items that are in both arrays
3702         foreach ( $baseArray as $name => &$groupVal ) {
3703                 if ( isset( $newValues[$name] ) ) {
3704                         $groupVal += $newValues[$name];
3705                 }
3706         }
3707         // Now add items that didn't exist yet
3708         $baseArray += $newValues;
3709
3710         return $baseArray;
3711 }