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 MediaWiki\MediaWikiServices;
  31 use Wikimedia\ScopedCallback;
  32 use Wikimedia\Rdbms\DBReplicationWaitError;
  33
  34 // Hide compatibility functions from Doxygen
  35 /// @cond
  36 /**
  37  * Compatibility functions
  38  *
  39  * We support PHP 5.5.9 and up.
  40  * Re-implementations of newer functions or functions in non-standard
  41  * PHP extensions may be included here.
  42  */
  43
  44 // hash_equals function only exists in PHP >= 5.6.0
  45 // https://secure.php.net/hash_equals
  46 if ( !function_exists( 'hash_equals' ) ) {
  47         /**
  48          * Check whether a user-provided string is equal to a fixed-length secret string
  49          * without revealing bytes of the secret string through timing differences.
  50          *
  51          * The usual way to compare strings (PHP's === operator or the underlying memcmp()
  52          * function in C) is to compare corresponding bytes and stop at the first difference,
  53          * which would take longer for a partial match than for a complete mismatch. This
  54          * is not secure when one of the strings (e.g. an HMAC or token) must remain secret
  55          * and the other may come from an attacker. Statistical analysis of timing measurements
  56          * over many requests may allow the attacker to guess the string's bytes one at a time
  57          * (and check his guesses) even if the timing differences are extremely small.
  58          *
  59          * When making such a security-sensitive comparison, it is essential that the sequence
  60          * in which instructions are executed and memory locations are accessed not depend on
  61          * the secret string's value. HOWEVER, for simplicity, we do not attempt to minimize
  62          * the inevitable leakage of the string's length. That is generally known anyway as
  63          * a chararacteristic of the hash function used to compute the secret value.
  64          *
  65          * Longer explanation: http://www.emerose.com/timing-attacks-explained
  66          *
  67          * @codeCoverageIgnore
  68          * @param string $known_string Fixed-length secret string to compare against
  69          * @param string $user_string User-provided string
  70          * @return bool True if the strings are the same, false otherwise
  71          */
  72         function hash_equals( $known_string, $user_string ) {
  73                 // Strict type checking as in PHP's native implementation
  74                 if ( !is_string( $known_string ) ) {
  75                         trigger_error( 'hash_equals(): Expected known_string to be a string, ' .
  76                                 gettype( $known_string ) . ' given', E_USER_WARNING );
  77
  78                         return false;
  79                 }
  80
  81                 if ( !is_string( $user_string ) ) {
  82                         trigger_error( 'hash_equals(): Expected user_string to be a string, ' .
  83                                 gettype( $user_string ) . ' given', E_USER_WARNING );
  84
  85                         return false;
  86                 }
  87
  88                 $known_string_len = strlen( $known_string );
  89                 if ( $known_string_len !== strlen( $user_string ) ) {
  90                         return false;
  91                 }
  92
  93                 $result = 0;
  94                 for ( $i = 0; $i < $known_string_len; $i++ ) {
  95                         $result |= ord( $known_string[$i] ) ^ ord( $user_string[$i] );
  96                 }
  97
  98                 return ( $result === 0 );
  99         }
 100 }
 101 /// @endcond
 102
 103 /**
 104  * Load an extension
 105  *
 106  * This queues an extension to be loaded through
 107  * the ExtensionRegistry system.
 108  *
 109  * @param string $ext Name of the extension to load
 110  * @param string|null $path Absolute path of where to find the extension.json file
 111  * @since 1.25
 112  */
 113 function wfLoadExtension( $ext, $path = null ) {
 114         if ( !$path ) {
 115                 global $wgExtensionDirectory;
 116                 $path = "$wgExtensionDirectory/$ext/extension.json";
 117         }
 118         ExtensionRegistry::getInstance()->queue( $path );
 119 }
 120
 121 /**
 122  * Load multiple extensions at once
 123  *
 124  * Same as wfLoadExtension, but more efficient if you
 125  * are loading multiple extensions.
 126  *
 127  * If you want to specify custom paths, you should interact with
 128  * ExtensionRegistry directly.
 129  *
 130  * @see wfLoadExtension
 131  * @param string[] $exts Array of extension names to load
 132  * @since 1.25
 133  */
 134 function wfLoadExtensions( array $exts ) {
 135         global $wgExtensionDirectory;
 136         $registry = ExtensionRegistry::getInstance();
 137         foreach ( $exts as $ext ) {
 138                 $registry->queue( "$wgExtensionDirectory/$ext/extension.json" );
 139         }
 140 }
 141
 142 /**
 143  * Load a skin
 144  *
 145  * @see wfLoadExtension
 146  * @param string $skin Name of the extension to load
 147  * @param string|null $path Absolute path of where to find the skin.json file
 148  * @since 1.25
 149  */
 150 function wfLoadSkin( $skin, $path = null ) {
 151         if ( !$path ) {
 152                 global $wgStyleDirectory;
 153                 $path = "$wgStyleDirectory/$skin/skin.json";
 154         }
 155         ExtensionRegistry::getInstance()->queue( $path );
 156 }
 157
 158 /**
 159  * Load multiple skins at once
 160  *
 161  * @see wfLoadExtensions
 162  * @param string[] $skins Array of extension names to load
 163  * @since 1.25
 164  */
 165 function wfLoadSkins( array $skins ) {
 166         global $wgStyleDirectory;
 167         $registry = ExtensionRegistry::getInstance();
 168         foreach ( $skins as $skin ) {
 169                 $registry->queue( "$wgStyleDirectory/$skin/skin.json" );
 170         }
 171 }
 172
 173 /**
 174  * Like array_diff( $a, $b ) except that it works with two-dimensional arrays.
 175  * @param array $a
 176  * @param array $b
 177  * @return array
 178  */
 179 function wfArrayDiff2( $a, $b ) {
 180         return array_udiff( $a, $b, 'wfArrayDiff2_cmp' );
 181 }
 182
 183 /**
 184  * @param array|string $a
 185  * @param array|string $b
 186  * @return int
 187  */
 188 function wfArrayDiff2_cmp( $a, $b ) {
 189         if ( is_string( $a ) && is_string( $b ) ) {
 190                 return strcmp( $a, $b );
 191         } elseif ( count( $a ) !== count( $b ) ) {
 192                 return count( $a ) < count( $b ) ? -1 : 1;
 193         } else {
 194                 reset( $a );
 195                 reset( $b );
 196                 while ( ( list( , $valueA ) = each( $a ) ) && ( list( , $valueB ) = each( $b ) ) ) {
 197                         $cmp = strcmp( $valueA, $valueB );
 198                         if ( $cmp !== 0 ) {
 199                                 return $cmp;
 200                         }
 201                 }
 202                 return 0;
 203         }
 204 }
 205
 206 /**
 207  * Like array_filter with ARRAY_FILTER_USE_BOTH, but works pre-5.6.
 208  *
 209  * @param array $arr
 210  * @param callable $callback Will be called with the array value and key (in that order) and
 211  *   should return a bool which will determine whether the array element is kept.
 212  * @return array
 213  */
 214 function wfArrayFilter( array $arr, callable $callback ) {
 215         if ( defined( 'ARRAY_FILTER_USE_BOTH' ) ) {
 216                 return array_filter( $arr, $callback, ARRAY_FILTER_USE_BOTH );
 217         }
 218         $filteredKeys = array_filter( array_keys( $arr ), function ( $key ) use ( $arr, $callback ) {
 219                 return call_user_func( $callback, $arr[$key], $key );
 220         } );
 221         return array_intersect_key( $arr, array_fill_keys( $filteredKeys, true ) );
 222 }
 223
 224 /**
 225  * Like array_filter with ARRAY_FILTER_USE_KEY, but works pre-5.6.
 226  *
 227  * @param array $arr
 228  * @param callable $callback Will be called with the array key and should return a bool which
 229  *   will determine whether the array element is kept.
 230  * @return array
 231  */
 232 function wfArrayFilterByKey( array $arr, callable $callback ) {
 233         return wfArrayFilter( $arr, function ( $val, $key ) use ( $callback ) {
 234                 return call_user_func( $callback, $key );
 235         } );
 236 }
 237
 238 /**
 239  * Appends to second array if $value differs from that in $default
 240  *
 241  * @param string|int $key
 242  * @param mixed $value
 243  * @param mixed $default
 244  * @param array $changed Array to alter
 245  * @throws MWException
 246  */
 247 function wfAppendToArrayIfNotDefault( $key, $value, $default, &$changed ) {
 248         if ( is_null( $changed ) ) {
 249                 throw new MWException( 'GlobalFunctions::wfAppendToArrayIfNotDefault got null' );
 250         }
 251         if ( $default[$key] !== $value ) {
 252                 $changed[$key] = $value;
 253         }
 254 }
 255
 256 /**
 257  * Merge arrays in the style of getUserPermissionsErrors, with duplicate removal
 258  * e.g.
 259  *     wfMergeErrorArrays(
 260  *       [ [ 'x' ] ],
 261  *       [ [ 'x', '2' ] ],
 262  *       [ [ 'x' ] ],
 263  *       [ [ 'y' ] ]
 264  *     );
 265  * returns:
 266  *     [
 267  *       [ 'x', '2' ],
 268  *       [ 'x' ],
 269  *       [ 'y' ]
 270  *     ]
 271  *
 272  * @param array $array1,...
 273  * @return array
 274  */
 275 function wfMergeErrorArrays( /*...*/ ) {
 276         $args = func_get_args();
 277         $out = [];
 278         foreach ( $args as $errors ) {
 279                 foreach ( $errors as $params ) {
 280                         $originalParams = $params;
 281                         if ( $params[0] instanceof MessageSpecifier ) {
 282                                 $msg = $params[0];
 283                                 $params = array_merge( [ $msg->getKey() ], $msg->getParams() );
 284                         }
 285                         # @todo FIXME: Sometimes get nested arrays for $params,
 286                         # which leads to E_NOTICEs
 287                         $spec = implode( "\t", $params );
 288                         $out[$spec] = $originalParams;
 289                 }
 290         }
 291         return array_values( $out );
 292 }
 293
 294 /**
 295  * Insert array into another array after the specified *KEY*
 296  *
 297  * @param array $array The array.
 298  * @param array $insert The array to insert.
 299  * @param mixed $after The key to insert after
 300  * @return array
 301  */
 302 function wfArrayInsertAfter( array $array, array $insert, $after ) {
 303         // Find the offset of the element to insert after.
 304         $keys = array_keys( $array );
 305         $offsetByKey = array_flip( $keys );
 306
 307         $offset = $offsetByKey[$after];
 308
 309         // Insert at the specified offset
 310         $before = array_slice( $array, 0, $offset + 1, true );
 311         $after = array_slice( $array, $offset + 1, count( $array ) - $offset, true );
 312
 313         $output = $before + $insert + $after;
 314
 315         return $output;
 316 }
 317
 318 /**
 319  * Recursively converts the parameter (an object) to an array with the same data
 320  *
 321  * @param object|array $objOrArray
 322  * @param bool $recursive
 323  * @return array
 324  */
 325 function wfObjectToArray( $objOrArray, $recursive = true ) {
 326         $array = [];
 327         if ( is_object( $objOrArray ) ) {
 328                 $objOrArray = get_object_vars( $objOrArray );
 329         }
 330         foreach ( $objOrArray as $key => $value ) {
 331                 if ( $recursive && ( is_object( $value ) || is_array( $value ) ) ) {
 332                         $value = wfObjectToArray( $value );
 333                 }
 334
 335                 $array[$key] = $value;
 336         }
 337
 338         return $array;
 339 }
 340
 341 /**
 342  * Get a random decimal value between 0 and 1, in a way
 343  * not likely to give duplicate values for any realistic
 344  * number of articles.
 345  *
 346  * @note This is designed for use in relation to Special:RandomPage
 347  *       and the page_random database field.
 348  *
 349  * @return string
 350  */
 351 function wfRandom() {
 352         // The maximum random value is "only" 2^31-1, so get two random
 353         // values to reduce the chance of dupes
 354         $max = mt_getrandmax() + 1;
 355         $rand = number_format( ( mt_rand() * $max + mt_rand() ) / $max / $max, 12, '.', '' );
 356         return $rand;
 357 }
 358
 359 /**
 360  * Get a random string containing a number of pseudo-random hex characters.
 361  *
 362  * @note This is not secure, if you are trying to generate some sort
 363  *       of token please use MWCryptRand instead.
 364  *
 365  * @param int $length The length of the string to generate
 366  * @return string
 367  * @since 1.20
 368  */
 369 function wfRandomString( $length = 32 ) {
 370         $str = '';
 371         for ( $n = 0; $n < $length; $n += 7 ) {
 372                 $str .= sprintf( '%07x', mt_rand() & 0xfffffff );
 373         }
 374         return substr( $str, 0, $length );
 375 }
 376
 377 /**
 378  * We want some things to be included as literal characters in our title URLs
 379  * for prettiness, which urlencode encodes by default.  According to RFC 1738,
 380  * all of the following should be safe:
 381  *
 382  * ;:@&=$-_.+!*'(),
 383  *
 384  * RFC 1738 says ~ is unsafe, however RFC 3986 considers it an unreserved
 385  * character which should not be encoded. More importantly, google chrome
 386  * always converts %7E back to ~, and converting it in this function can
 387  * cause a redirect loop (T105265).
 388  *
 389  * But + is not safe because it's used to indicate a space; &= are only safe in
 390  * paths and not in queries (and we don't distinguish here); ' seems kind of
 391  * scary; and urlencode() doesn't touch -_. to begin with.  Plus, although /
 392  * is reserved, we don't care.  So the list we unescape is:
 393  *
 394  * ;:@$!*(),/~
 395  *
 396  * However, IIS7 redirects fail when the url contains a colon (see T24709),
 397  * so no fancy : for IIS7.
 398  *
 399  * %2F in the page titles seems to fatally break for some reason.
 400  *
 401  * @param string $s
 402  * @return string
 403  */
 404 function wfUrlencode( $s ) {
 405         static $needle;
 406
 407         if ( is_null( $s ) ) {
 408                 $needle = null;
 409                 return '';
 410         }
 411
 412         if ( is_null( $needle ) ) {
 413                 $needle = [ '%3B', '%40', '%24', '%21', '%2A', '%28', '%29', '%2C', '%2F', '%7E' ];
 414                 if ( !isset( $_SERVER['SERVER_SOFTWARE'] ) ||
 415                         ( strpos( $_SERVER['SERVER_SOFTWARE'], 'Microsoft-IIS/7' ) === false )
 416                 ) {
 417                         $needle[] = '%3A';
 418                 }
 419         }
 420
 421         $s = urlencode( $s );
 422         $s = str_ireplace(
 423                 $needle,
 424                 [ ';', '@', '$', '!', '*', '(', ')', ',', '/', '~', ':' ],
 425                 $s
 426         );
 427
 428         return $s;
 429 }
 430
 431 /**
 432  * This function takes one or two arrays as input, and returns a CGI-style string, e.g.
 433  * "days=7&limit=100". Options in the first array override options in the second.
 434  * Options set to null or false will not be output.
 435  *
 436  * @param array $array1 ( String|Array )
 437  * @param array|null $array2 ( String|Array )
 438  * @param string $prefix
 439  * @return string
 440  */
 441 function wfArrayToCgi( $array1, $array2 = null, $prefix = '' ) {
 442         if ( !is_null( $array2 ) ) {
 443                 $array1 = $array1 + $array2;
 444         }
 445
 446         $cgi = '';
 447         foreach ( $array1 as $key => $value ) {
 448                 if ( !is_null( $value ) && $value !== false ) {
 449                         if ( $cgi != '' ) {
 450                                 $cgi .= '&';
 451                         }
 452                         if ( $prefix !== '' ) {
 453                                 $key = $prefix . "[$key]";
 454                         }
 455                         if ( is_array( $value ) ) {
 456                                 $firstTime = true;
 457                                 foreach ( $value as $k => $v ) {
 458                                         $cgi .= $firstTime ? '' : '&';
 459                                         if ( is_array( $v ) ) {
 460                                                 $cgi .= wfArrayToCgi( $v, null, $key . "[$k]" );
 461                                         } else {
 462                                                 $cgi .= urlencode( $key . "[$k]" ) . '=' . urlencode( $v );
 463                                         }
 464                                         $firstTime = false;
 465                                 }
 466                         } else {
 467                                 if ( is_object( $value ) ) {
 468                                         $value = $value->__toString();
 469                                 }
 470                                 $cgi .= urlencode( $key ) . '=' . urlencode( $value );
 471                         }
 472                 }
 473         }
 474         return $cgi;
 475 }
 476
 477 /**
 478  * This is the logical opposite of wfArrayToCgi(): it accepts a query string as
 479  * its argument and returns the same string in array form.  This allows compatibility
 480  * with legacy functions that accept raw query strings instead of nice
 481  * arrays.  Of course, keys and values are urldecode()d.
 482  *
 483  * @param string $query Query string
 484  * @return string[] Array version of input
 485  */
 486 function wfCgiToArray( $query ) {
 487         if ( isset( $query[0] ) && $query[0] == '?' ) {
 488                 $query = substr( $query, 1 );
 489         }
 490         $bits = explode( '&', $query );
 491         $ret = [];
 492         foreach ( $bits as $bit ) {
 493                 if ( $bit === '' ) {
 494                         continue;
 495                 }
 496                 if ( strpos( $bit, '=' ) === false ) {
 497                         // Pieces like &qwerty become 'qwerty' => '' (at least this is what php does)
 498                         $key = $bit;
 499                         $value = '';
 500                 } else {
 501                         list( $key, $value ) = explode( '=', $bit );
 502                 }
 503                 $key = urldecode( $key );
 504                 $value = urldecode( $value );
 505                 if ( strpos( $key, '[' ) !== false ) {
 506                         $keys = array_reverse( explode( '[', $key ) );
 507                         $key = array_pop( $keys );
 508                         $temp = $value;
 509                         foreach ( $keys as $k ) {
 510                                 $k = substr( $k, 0, -1 );
 511                                 $temp = [ $k => $temp ];
 512                         }
 513                         if ( isset( $ret[$key] ) ) {
 514                                 $ret[$key] = array_merge( $ret[$key], $temp );
 515                         } else {
 516                                 $ret[$key] = $temp;
 517                         }
 518                 } else {
 519                         $ret[$key] = $value;
 520                 }
 521         }
 522         return $ret;
 523 }
 524
 525 /**
 526  * Append a query string to an existing URL, which may or may not already
 527  * have query string parameters already. If so, they will be combined.
 528  *
 529  * @param string $url
 530  * @param string|string[] $query String or associative array
 531  * @return string
 532  */
 533 function wfAppendQuery( $url, $query ) {
 534         if ( is_array( $query ) ) {
 535                 $query = wfArrayToCgi( $query );
 536         }
 537         if ( $query != '' ) {
 538                 // Remove the fragment, if there is one
 539                 $fragment = false;
 540                 $hashPos = strpos( $url, '#' );
 541                 if ( $hashPos !== false ) {
 542                         $fragment = substr( $url, $hashPos );
 543                         $url = substr( $url, 0, $hashPos );
 544                 }
 545
 546                 // Add parameter
 547                 if ( false === strpos( $url, '?' ) ) {
 548                         $url .= '?';
 549                 } else {
 550                         $url .= '&';
 551                 }
 552                 $url .= $query;
 553
 554                 // Put the fragment back
 555                 if ( $fragment !== false ) {
 556                         $url .= $fragment;
 557                 }
 558         }
 559         return $url;
 560 }
 561
 562 /**
 563  * Expand a potentially local URL to a fully-qualified URL. Assumes $wgServer
 564  * is correct.
 565  *
 566  * The meaning of the PROTO_* constants is as follows:
 567  * PROTO_HTTP: Output a URL starting with http://
 568  * PROTO_HTTPS: Output a URL starting with https://
 569  * PROTO_RELATIVE: Output a URL starting with // (protocol-relative URL)
 570  * PROTO_CURRENT: Output a URL starting with either http:// or https:// , depending
 571  *    on which protocol was used for the current incoming request
 572  * PROTO_CANONICAL: For URLs without a domain, like /w/index.php , use $wgCanonicalServer.
 573  *    For protocol-relative URLs, use the protocol of $wgCanonicalServer
 574  * PROTO_INTERNAL: Like PROTO_CANONICAL, but uses $wgInternalServer instead of $wgCanonicalServer
 575  *
 576  * @todo this won't work with current-path-relative URLs
 577  * like "subdir/foo.html", etc.
 578  *
 579  * @param string $url Either fully-qualified or a local path + query
 580  * @param string $defaultProto One of the PROTO_* constants. Determines the
 581  *    protocol to use if $url or $wgServer is protocol-relative
 582  * @return string|false Fully-qualified URL, current-path-relative URL or false if
 583  *    no valid URL can be constructed
 584  */
 585 function wfExpandUrl( $url, $defaultProto = PROTO_CURRENT ) {
 586         global $wgServer, $wgCanonicalServer, $wgInternalServer, $wgRequest,
 587                 $wgHttpsPort;
 588         if ( $defaultProto === PROTO_CANONICAL ) {
 589                 $serverUrl = $wgCanonicalServer;
 590         } elseif ( $defaultProto === PROTO_INTERNAL && $wgInternalServer !== false ) {
 591                 // Make $wgInternalServer fall back to $wgServer if not set
 592                 $serverUrl = $wgInternalServer;
 593         } else {
 594                 $serverUrl = $wgServer;
 595                 if ( $defaultProto === PROTO_CURRENT ) {
 596                         $defaultProto = $wgRequest->getProtocol() . '://';
 597                 }
 598         }
 599
 600         // Analyze $serverUrl to obtain its protocol
 601         $bits = wfParseUrl( $serverUrl );
 602         $serverHasProto = $bits && $bits['scheme'] != '';
 603
 604         if ( $defaultProto === PROTO_CANONICAL || $defaultProto === PROTO_INTERNAL ) {
 605                 if ( $serverHasProto ) {
 606                         $defaultProto = $bits['scheme'] . '://';
 607                 } else {
 608                         // $wgCanonicalServer or $wgInternalServer doesn't have a protocol.
 609                         // This really isn't supposed to happen. Fall back to HTTP in this
 610                         // ridiculous case.
 611                         $defaultProto = PROTO_HTTP;
 612                 }
 613         }
 614
 615         $defaultProtoWithoutSlashes = substr( $defaultProto, 0, -2 );
 616
 617         if ( substr( $url, 0, 2 ) == '//' ) {
 618                 $url = $defaultProtoWithoutSlashes . $url;
 619         } elseif ( substr( $url, 0, 1 ) == '/' ) {
 620                 // If $serverUrl is protocol-relative, prepend $defaultProtoWithoutSlashes,
 621                 // otherwise leave it alone.
 622                 $url = ( $serverHasProto ? '' : $defaultProtoWithoutSlashes ) . $serverUrl . $url;
 623         }
 624
 625         $bits = wfParseUrl( $url );
 626
 627         // ensure proper port for HTTPS arrives in URL
 628         // https://phabricator.wikimedia.org/T67184
 629         if ( $defaultProto === PROTO_HTTPS && $wgHttpsPort != 443 ) {
 630                 $bits['port'] = $wgHttpsPort;
 631         }
 632
 633         if ( $bits && isset( $bits['path'] ) ) {
 634                 $bits['path'] = wfRemoveDotSegments( $bits['path'] );
 635                 return wfAssembleUrl( $bits );
 636         } elseif ( $bits ) {
 637                 # No path to expand
 638                 return $url;
 639         } elseif ( substr( $url, 0, 1 ) != '/' ) {
 640                 # URL is a relative path
 641                 return wfRemoveDotSegments( $url );
 642         }
 643
 644         # Expanded URL is not valid.
 645         return false;
 646 }
 647
 648 /**
 649  * This function will reassemble a URL parsed with wfParseURL.  This is useful
 650  * if you need to edit part of a URL and put it back together.
 651  *
 652  * This is the basic structure used (brackets contain keys for $urlParts):
 653  * [scheme][delimiter][user]:[pass]@[host]:[port][path]?[query]#[fragment]
 654  *
 655  * @todo Need to integrate this into wfExpandUrl (see T34168)
 656  *
 657  * @since 1.19
 658  * @param array $urlParts URL parts, as output from wfParseUrl
 659  * @return string URL assembled from its component parts
 660  */
 661 function wfAssembleUrl( $urlParts ) {
 662         $result = '';
 663
 664         if ( isset( $urlParts['delimiter'] ) ) {
 665                 if ( isset( $urlParts['scheme'] ) ) {
 666                         $result .= $urlParts['scheme'];
 667                 }
 668
 669                 $result .= $urlParts['delimiter'];
 670         }
 671
 672         if ( isset( $urlParts['host'] ) ) {
 673                 if ( isset( $urlParts['user'] ) ) {
 674                         $result .= $urlParts['user'];
 675                         if ( isset( $urlParts['pass'] ) ) {
 676                                 $result .= ':' . $urlParts['pass'];
 677                         }
 678                         $result .= '@';
 679                 }
 680
 681                 $result .= $urlParts['host'];
 682
 683                 if ( isset( $urlParts['port'] ) ) {
 684                         $result .= ':' . $urlParts['port'];
 685                 }
 686         }
 687
 688         if ( isset( $urlParts['path'] ) ) {
 689                 $result .= $urlParts['path'];
 690         }
 691
 692         if ( isset( $urlParts['query'] ) ) {
 693                 $result .= '?' . $urlParts['query'];
 694         }
 695
 696         if ( isset( $urlParts['fragment'] ) ) {
 697                 $result .= '#' . $urlParts['fragment'];
 698         }
 699
 700         return $result;
 701 }
 702
 703 /**
 704  * Remove all dot-segments in the provided URL path.  For example,
 705  * '/a/./b/../c/' becomes '/a/c/'.  For details on the algorithm, please see
 706  * RFC3986 section 5.2.4.
 707  *
 708  * @todo Need to integrate this into wfExpandUrl (see T34168)
 709  *
 710  * @param string $urlPath URL path, potentially containing dot-segments
 711  * @return string URL path with all dot-segments removed
 712  */
 713 function wfRemoveDotSegments( $urlPath ) {
 714         $output = '';
 715         $inputOffset = 0;
 716         $inputLength = strlen( $urlPath );
 717
 718         while ( $inputOffset < $inputLength ) {
 719                 $prefixLengthOne = substr( $urlPath, $inputOffset, 1 );
 720                 $prefixLengthTwo = substr( $urlPath, $inputOffset, 2 );
 721                 $prefixLengthThree = substr( $urlPath, $inputOffset, 3 );
 722                 $prefixLengthFour = substr( $urlPath, $inputOffset, 4 );
 723                 $trimOutput = false;
 724
 725                 if ( $prefixLengthTwo == './' ) {
 726                         # Step A, remove leading "./"
 727                         $inputOffset += 2;
 728                 } elseif ( $prefixLengthThree == '../' ) {
 729                         # Step A, remove leading "../"
 730                         $inputOffset += 3;
 731                 } elseif ( ( $prefixLengthTwo == '/.' ) && ( $inputOffset + 2 == $inputLength ) ) {
 732                         # Step B, replace leading "/.$" with "/"
 733                         $inputOffset += 1;
 734                         $urlPath[$inputOffset] = '/';
 735                 } elseif ( $prefixLengthThree == '/./' ) {
 736                         # Step B, replace leading "/./" with "/"
 737                         $inputOffset += 2;
 738                 } elseif ( $prefixLengthThree == '/..' && ( $inputOffset + 3 == $inputLength ) ) {
 739                         # Step C, replace leading "/..$" with "/" and
 740                         # remove last path component in output
 741                         $inputOffset += 2;
 742                         $urlPath[$inputOffset] = '/';
 743                         $trimOutput = true;
 744                 } elseif ( $prefixLengthFour == '/../' ) {
 745                         # Step C, replace leading "/../" with "/" and
 746                         # remove last path component in output
 747                         $inputOffset += 3;
 748                         $trimOutput = true;
 749                 } elseif ( ( $prefixLengthOne == '.' ) && ( $inputOffset + 1 == $inputLength ) ) {
 750                         # Step D, remove "^.$"
 751                         $inputOffset += 1;
 752                 } elseif ( ( $prefixLengthTwo == '..' ) && ( $inputOffset + 2 == $inputLength ) ) {
 753                         # Step D, remove "^..$"
 754                         $inputOffset += 2;
 755                 } else {
 756                         # Step E, move leading path segment to output
 757                         if ( $prefixLengthOne == '/' ) {
 758                                 $slashPos = strpos( $urlPath, '/', $inputOffset + 1 );
 759                         } else {
 760                                 $slashPos = strpos( $urlPath, '/', $inputOffset );
 761                         }
 762                         if ( $slashPos === false ) {
 763                                 $output .= substr( $urlPath, $inputOffset );
 764                                 $inputOffset = $inputLength;
 765                         } else {
 766                                 $output .= substr( $urlPath, $inputOffset, $slashPos - $inputOffset );
 767                                 $inputOffset += $slashPos - $inputOffset;
 768                         }
 769                 }
 770
 771                 if ( $trimOutput ) {
 772                         $slashPos = strrpos( $output, '/' );
 773                         if ( $slashPos === false ) {
 774                                 $output = '';
 775                         } else {
 776                                 $output = substr( $output, 0, $slashPos );
 777                         }
 778                 }
 779         }
 780
 781         return $output;
 782 }
 783
 784 /**
 785  * Returns a regular expression of url protocols
 786  *
 787  * @param bool $includeProtocolRelative If false, remove '//' from the returned protocol list.
 788  *        DO NOT USE this directly, use wfUrlProtocolsWithoutProtRel() instead
 789  * @return string
 790  */
 791 function wfUrlProtocols( $includeProtocolRelative = true ) {
 792         global $wgUrlProtocols;
 793
 794         // Cache return values separately based on $includeProtocolRelative
 795         static $withProtRel = null, $withoutProtRel = null;
 796         $cachedValue = $includeProtocolRelative ? $withProtRel : $withoutProtRel;
 797         if ( !is_null( $cachedValue ) ) {
 798                 return $cachedValue;
 799         }
 800
 801         // Support old-style $wgUrlProtocols strings, for backwards compatibility
 802         // with LocalSettings files from 1.5
 803         if ( is_array( $wgUrlProtocols ) ) {
 804                 $protocols = [];
 805                 foreach ( $wgUrlProtocols as $protocol ) {
 806                         // Filter out '//' if !$includeProtocolRelative
 807                         if ( $includeProtocolRelative || $protocol !== '//' ) {
 808                                 $protocols[] = preg_quote( $protocol, '/' );
 809                         }
 810                 }
 811
 812                 $retval = implode( '|', $protocols );
 813         } else {
 814                 // Ignore $includeProtocolRelative in this case
 815                 // This case exists for pre-1.6 compatibility, and we can safely assume
 816                 // that '//' won't appear in a pre-1.6 config because protocol-relative
 817                 // URLs weren't supported until 1.18
 818                 $retval = $wgUrlProtocols;
 819         }
 820
 821         // Cache return value
 822         if ( $includeProtocolRelative ) {
 823                 $withProtRel = $retval;
 824         } else {
 825                 $withoutProtRel = $retval;
 826         }
 827         return $retval;
 828 }
 829
 830 /**
 831  * Like wfUrlProtocols(), but excludes '//' from the protocol list. Use this if
 832  * you need a regex that matches all URL protocols but does not match protocol-
 833  * relative URLs
 834  * @return string
 835  */
 836 function wfUrlProtocolsWithoutProtRel() {
 837         return wfUrlProtocols( false );
 838 }
 839
 840 /**
 841  * parse_url() work-alike, but non-broken.  Differences:
 842  *
 843  * 1) Does not raise warnings on bad URLs (just returns false).
 844  * 2) Handles protocols that don't use :// (e.g., mailto: and news:, as well as
 845  *    protocol-relative URLs) correctly.
 846  * 3) Adds a "delimiter" element to the array (see (2)).
 847  * 4) Verifies that the protocol is on the $wgUrlProtocols whitelist.
 848  * 5) Rejects some invalid URLs that parse_url doesn't, e.g. the empty string or URLs starting with
 849  *    a line feed character.
 850  *
 851  * @param string $url A URL to parse
 852  * @return string[]|bool Bits of the URL in an associative array, or false on failure.
 853  *   Possible fields:
 854  *   - scheme: URI scheme (protocol), e.g. 'http', 'mailto'. Lowercase, always present, but can
 855  *       be an empty string for protocol-relative URLs.
 856  *   - delimiter: either '://', ':' or '//'. Always present.
 857  *   - host: domain name / IP. Always present, but could be an empty string, e.g. for file: URLs.
 858  *   - user: user name, e.g. for HTTP Basic auth URLs such as http://user:pass@example.com/
 859  *       Missing when there is no username.
 860  *   - pass: password, same as above.
 861  *   - path: path including the leading /. Will be missing when empty (e.g. 'http://example.com')
 862  *   - query: query string (as a string; see wfCgiToArray() for parsing it), can be missing.
 863  *   - fragment: the part after #, can be missing.
 864  */
 865 function wfParseUrl( $url ) {
 866         global $wgUrlProtocols; // Allow all protocols defined in DefaultSettings/LocalSettings.php
 867
 868         // Protocol-relative URLs are handled really badly by parse_url(). It's so
 869         // bad that the easiest way to handle them is to just prepend 'http:' and
 870         // strip the protocol out later.
 871         $wasRelative = substr( $url, 0, 2 ) == '//';
 872         if ( $wasRelative ) {
 873                 $url = "http:$url";
 874         }
 875         MediaWiki\suppressWarnings();
 876         $bits = parse_url( $url );
 877         MediaWiki\restoreWarnings();
 878         // parse_url() returns an array without scheme for some invalid URLs, e.g.
 879         // parse_url("%0Ahttp://example.com") == [ 'host' => '%0Ahttp', 'path' => 'example.com' ]
 880         if ( !$bits || !isset( $bits['scheme'] ) ) {
 881                 return false;
 882         }
 883
 884         // parse_url() incorrectly handles schemes case-sensitively. Convert it to lowercase.
 885         $bits['scheme'] = strtolower( $bits['scheme'] );
 886
 887         // most of the protocols are followed by ://, but mailto: and sometimes news: not, check for it
 888         if ( in_array( $bits['scheme'] . '://', $wgUrlProtocols ) ) {
 889                 $bits['delimiter'] = '://';
 890         } elseif ( in_array( $bits['scheme'] . ':', $wgUrlProtocols ) ) {
 891                 $bits['delimiter'] = ':';
 892                 // parse_url detects for news: and mailto: the host part of an url as path
 893                 // We have to correct this wrong detection
 894                 if ( isset( $bits['path'] ) ) {
 895                         $bits['host'] = $bits['path'];
 896                         $bits['path'] = '';
 897                 }
 898         } else {
 899                 return false;
 900         }
 901
 902         /* Provide an empty host for eg. file:/// urls (see T30627) */
 903         if ( !isset( $bits['host'] ) ) {
 904                 $bits['host'] = '';
 905
 906                 // See T47069
 907                 if ( isset( $bits['path'] ) ) {
 908                         /* parse_url loses the third / for file:///c:/ urls (but not on variants) */
 909                         if ( substr( $bits['path'], 0, 1 ) !== '/' ) {
 910                                 $bits['path'] = '/' . $bits['path'];
 911                         }
 912                 } else {
 913                         $bits['path'] = '';
 914                 }
 915         }
 916
 917         // If the URL was protocol-relative, fix scheme and delimiter
 918         if ( $wasRelative ) {
 919                 $bits['scheme'] = '';
 920                 $bits['delimiter'] = '//';
 921         }
 922         return $bits;
 923 }
 924
 925 /**
 926  * Take a URL, make sure it's expanded to fully qualified, and replace any
 927  * encoded non-ASCII Unicode characters with their UTF-8 original forms
 928  * for more compact display and legibility for local audiences.
 929  *
 930  * @todo handle punycode domains too
 931  *
 932  * @param string $url
 933  * @return string
 934  */
 935 function wfExpandIRI( $url ) {
 936         return preg_replace_callback(
 937                 '/((?:%[89A-F][0-9A-F])+)/i',
 938                 'wfExpandIRI_callback',
 939                 wfExpandUrl( $url )
 940         );
 941 }
 942
 943 /**
 944  * Private callback for wfExpandIRI
 945  * @param array $matches
 946  * @return string
 947  */
 948 function wfExpandIRI_callback( $matches ) {
 949         return urldecode( $matches[1] );
 950 }
 951
 952 /**
 953  * Make URL indexes, appropriate for the el_index field of externallinks.
 954  *
 955  * @param string $url
 956  * @return array
 957  */
 958 function wfMakeUrlIndexes( $url ) {
 959         $bits = wfParseUrl( $url );
 960
 961         // Reverse the labels in the hostname, convert to lower case
 962         // For emails reverse domainpart only
 963         if ( $bits['scheme'] == 'mailto' ) {
 964                 $mailparts = explode( '@', $bits['host'], 2 );
 965                 if ( count( $mailparts ) === 2 ) {
 966                         $domainpart = strtolower( implode( '.', array_reverse( explode( '.', $mailparts[1] ) ) ) );
 967                 } else {
 968                         // No domain specified, don't mangle it
 969                         $domainpart = '';
 970                 }
 971                 $reversedHost = $domainpart . '@' . $mailparts[0];
 972         } else {
 973                 $reversedHost = strtolower( implode( '.', array_reverse( explode( '.', $bits['host'] ) ) ) );
 974         }
 975         // Add an extra dot to the end
 976         // Why? Is it in wrong place in mailto links?
 977         if ( substr( $reversedHost, -1, 1 ) !== '.' ) {
 978                 $reversedHost .= '.';
 979         }
 980         // Reconstruct the pseudo-URL
 981         $prot = $bits['scheme'];
 982         $index = $prot . $bits['delimiter'] . $reversedHost;
 983         // Leave out user and password. Add the port, path, query and fragment
 984         if ( isset( $bits['port'] ) ) {
 985                 $index .= ':' . $bits['port'];
 986         }
 987         if ( isset( $bits['path'] ) ) {
 988                 $index .= $bits['path'];
 989         } else {
 990                 $index .= '/';
 991         }
 992         if ( isset( $bits['query'] ) ) {
 993                 $index .= '?' . $bits['query'];
 994         }
 995         if ( isset( $bits['fragment'] ) ) {
 996                 $index .= '#' . $bits['fragment'];
 997         }
 998
 999         if ( $prot == '' ) {
1000                 return [ "http:$index", "https:$index" ];
1001         } else {
1002                 return [ $index ];
1003         }
1004 }
1005
1006 /**
1007  * Check whether a given URL has a domain that occurs in a given set of domains
1008  * @param string $url URL
1009  * @param array $domains Array of domains (strings)
1010  * @return bool True if the host part of $url ends in one of the strings in $domains
1011  */
1012 function wfMatchesDomainList( $url, $domains ) {
1013         $bits = wfParseUrl( $url );
1014         if ( is_array( $bits ) && isset( $bits['host'] ) ) {
1015                 $host = '.' . $bits['host'];
1016                 foreach ( (array)$domains as $domain ) {
1017                         $domain = '.' . $domain;
1018                         if ( substr( $host, -strlen( $domain ) ) === $domain ) {
1019                                 return true;
1020                         }
1021                 }
1022         }
1023         return false;
1024 }
1025
1026 /**
1027  * Sends a line to the debug log if enabled or, optionally, to a comment in output.
1028  * In normal operation this is a NOP.
1029  *
1030  * Controlling globals:
1031  * $wgDebugLogFile - points to the log file
1032  * $wgDebugRawPage - if false, 'action=raw' hits will not result in debug output.
1033  * $wgDebugComments - if on, some debug items may appear in comments in the HTML output.
1034  *
1035  * @since 1.25 support for additional context data
1036  *
1037  * @param string $text
1038  * @param string|bool $dest Destination of the message:
1039  *     - 'all': both to the log and HTML (debug toolbar or HTML comments)
1040  *     - 'private': excluded from HTML output
1041  *   For backward compatibility, it can also take a boolean:
1042  *     - true: same as 'all'
1043  *     - false: same as 'private'
1044  * @param array $context Additional logging context data
1045  */
1046 function wfDebug( $text, $dest = 'all', array $context = [] ) {
1047         global $wgDebugRawPage, $wgDebugLogPrefix;
1048         global $wgDebugTimestamps, $wgRequestTime;
1049
1050         if ( !$wgDebugRawPage && wfIsDebugRawPage() ) {
1051                 return;
1052         }
1053
1054         $text = trim( $text );
1055
1056         if ( $wgDebugTimestamps ) {
1057                 $context['seconds_elapsed'] = sprintf(
1058                         '%6.4f',
1059                         microtime( true ) - $wgRequestTime
1060                 );
1061                 $context['memory_used'] = sprintf(
1062                         '%5.1fM',
1063                         ( memory_get_usage( true ) / ( 1024 * 1024 ) )
1064                 );
1065         }
1066
1067         if ( $wgDebugLogPrefix !== '' ) {
1068                 $context['prefix'] = $wgDebugLogPrefix;
1069         }
1070         $context['private'] = ( $dest === false || $dest === 'private' );
1071
1072         $logger = LoggerFactory::getInstance( 'wfDebug' );
1073         $logger->debug( $text, $context );
1074 }
1075
1076 /**
1077  * Returns true if debug logging should be suppressed if $wgDebugRawPage = false
1078  * @return bool
1079  */
1080 function wfIsDebugRawPage() {
1081         static $cache;
1082         if ( $cache !== null ) {
1083                 return $cache;
1084         }
1085         # Check for raw action using $_GET not $wgRequest, since the latter might not be initialised yet
1086         if ( ( isset( $_GET['action'] ) && $_GET['action'] == 'raw' )
1087                 || (
1088                         isset( $_SERVER['SCRIPT_NAME'] )
1089                         && substr( $_SERVER['SCRIPT_NAME'], -8 ) == 'load.php'
1090                 )
1091         ) {
1092                 $cache = true;
1093         } else {
1094                 $cache = false;
1095         }
1096         return $cache;
1097 }
1098
1099 /**
1100  * Send a line giving PHP memory usage.
1101  *
1102  * @param bool $exact Print exact byte values instead of kibibytes (default: false)
1103  */
1104 function wfDebugMem( $exact = false ) {
1105         $mem = memory_get_usage();
1106         if ( !$exact ) {
1107                 $mem = floor( $mem / 1024 ) . ' KiB';
1108         } else {
1109                 $mem .= ' B';
1110         }
1111         wfDebug( "Memory usage: $mem\n" );
1112 }
1113
1114 /**
1115  * Send a line to a supplementary debug log file, if configured, or main debug
1116  * log if not.
1117  *
1118  * To configure a supplementary log file, set $wgDebugLogGroups[$logGroup] to
1119  * a string filename or an associative array mapping 'destination' to the
1120  * desired filename. The associative array may also contain a 'sample' key
1121  * with an integer value, specifying a sampling factor. Sampled log events
1122  * will be emitted with a 1 in N random chance.
1123  *
1124  * @since 1.23 support for sampling log messages via $wgDebugLogGroups.
1125  * @since 1.25 support for additional context data
1126  * @since 1.25 sample behavior dependent on configured $wgMWLoggerDefaultSpi
1127  *
1128  * @param string $logGroup
1129  * @param string $text
1130  * @param string|bool $dest Destination of the message:
1131  *     - 'all': both to the log and HTML (debug toolbar or HTML comments)
1132  *     - 'private': only to the specific log if set in $wgDebugLogGroups and
1133  *       discarded otherwise
1134  *   For backward compatibility, it can also take a boolean:
1135  *     - true: same as 'all'
1136  *     - false: same as 'private'
1137  * @param array $context Additional logging context data
1138  */
1139 function wfDebugLog(
1140         $logGroup, $text, $dest = 'all', array $context = []
1141 ) {
1142         $text = trim( $text );
1143
1144         $logger = LoggerFactory::getInstance( $logGroup );
1145         $context['private'] = ( $dest === false || $dest === 'private' );
1146         $logger->info( $text, $context );
1147 }
1148
1149 /**
1150  * Log for database errors
1151  *
1152  * @since 1.25 support for additional context data
1153  *
1154  * @param string $text Database error message.
1155  * @param array $context Additional logging context data
1156  */
1157 function wfLogDBError( $text, array $context = [] ) {
1158         $logger = LoggerFactory::getInstance( 'wfLogDBError' );
1159         $logger->error( trim( $text ), $context );
1160 }
1161
1162 /**
1163  * Throws a warning that $function is deprecated
1164  *
1165  * @param string $function
1166  * @param string|bool $version Version of MediaWiki that the function
1167  *    was deprecated in (Added in 1.19).
1168  * @param string|bool $component Added in 1.19.
1169  * @param int $callerOffset How far up the call stack is the original
1170  *    caller. 2 = function that called the function that called
1171  *    wfDeprecated (Added in 1.20)
1172  *
1173  * @return null
1174  */
1175 function wfDeprecated( $function, $version = false, $component = false, $callerOffset = 2 ) {
1176         MWDebug::deprecated( $function, $version, $component, $callerOffset + 1 );
1177 }
1178
1179 /**
1180  * Send a warning either to the debug log or in a PHP error depending on
1181  * $wgDevelopmentWarnings. To log warnings in production, use wfLogWarning() instead.
1182  *
1183  * @param string $msg Message to send
1184  * @param int $callerOffset Number of items to go back in the backtrace to
1185  *        find the correct caller (1 = function calling wfWarn, ...)
1186  * @param int $level PHP error level; defaults to E_USER_NOTICE;
1187  *        only used when $wgDevelopmentWarnings is true
1188  */
1189 function wfWarn( $msg, $callerOffset = 1, $level = E_USER_NOTICE ) {
1190         MWDebug::warning( $msg, $callerOffset + 1, $level, 'auto' );
1191 }
1192
1193 /**
1194  * Send a warning as a PHP error and the debug log. This is intended for logging
1195  * warnings in production. For logging development warnings, use WfWarn instead.
1196  *
1197  * @param string $msg Message to send
1198  * @param int $callerOffset Number of items to go back in the backtrace to
1199  *        find the correct caller (1 = function calling wfLogWarning, ...)
1200  * @param int $level PHP error level; defaults to E_USER_WARNING
1201  */
1202 function wfLogWarning( $msg, $callerOffset = 1, $level = E_USER_WARNING ) {
1203         MWDebug::warning( $msg, $callerOffset + 1, $level, 'production' );
1204 }
1205
1206 /**
1207  * Log to a file without getting "file size exceeded" signals.
1208  *
1209  * Can also log to TCP or UDP with the syntax udp://host:port/prefix. This will
1210  * send lines to the specified port, prefixed by the specified prefix and a space.
1211  * @since 1.25 support for additional context data
1212  *
1213  * @param string $text
1214  * @param string $file Filename
1215  * @param array $context Additional logging context data
1216  * @throws MWException
1217  * @deprecated since 1.25 Use \MediaWiki\Logger\LegacyLogger::emit or UDPTransport
1218  */
1219 function wfErrorLog( $text, $file, array $context = [] ) {
1220         wfDeprecated( __METHOD__, '1.25' );
1221         $logger = LoggerFactory::getInstance( 'wfErrorLog' );
1222         $context['destination'] = $file;
1223         $logger->info( trim( $text ), $context );
1224 }
1225
1226 /**
1227  * @todo document
1228  */
1229 function wfLogProfilingData() {
1230         global $wgDebugLogGroups, $wgDebugRawPage;
1231
1232         $context = RequestContext::getMain();
1233         $request = $context->getRequest();
1234
1235         $profiler = Profiler::instance();
1236         $profiler->setContext( $context );
1237         $profiler->logData();
1238
1239         $config = $context->getConfig();
1240         $stats = MediaWikiServices::getInstance()->getStatsdDataFactory();
1241         if ( $config->get( 'StatsdServer' ) && $stats->hasData() ) {
1242                 try {
1243                         $statsdServer = explode( ':', $config->get( 'StatsdServer' ) );
1244                         $statsdHost = $statsdServer[0];
1245                         $statsdPort = isset( $statsdServer[1] ) ? $statsdServer[1] : 8125;
1246                         $statsdSender = new SocketSender( $statsdHost, $statsdPort );
1247                         $statsdClient = new SamplingStatsdClient( $statsdSender, true, false );
1248                         $statsdClient->setSamplingRates( $config->get( 'StatsdSamplingRates' ) );
1249                         $statsdClient->send( $stats->getData() );
1250                 } catch ( Exception $ex ) {
1251                         MWExceptionHandler::logException( $ex );
1252                 }
1253         }
1254
1255         # Profiling must actually be enabled...
1256         if ( $profiler instanceof ProfilerStub ) {
1257                 return;
1258         }
1259
1260         if ( isset( $wgDebugLogGroups['profileoutput'] )
1261                 && $wgDebugLogGroups['profileoutput'] === false
1262         ) {
1263                 // Explicitly disabled
1264                 return;
1265         }
1266         if ( !$wgDebugRawPage && wfIsDebugRawPage() ) {
1267                 return;
1268         }
1269
1270         $ctx = [ 'elapsed' => $request->getElapsedTime() ];
1271         if ( !empty( $_SERVER['HTTP_X_FORWARDED_FOR'] ) ) {
1272                 $ctx['forwarded_for'] = $_SERVER['HTTP_X_FORWARDED_FOR'];
1273         }
1274         if ( !empty( $_SERVER['HTTP_CLIENT_IP'] ) ) {
1275                 $ctx['client_ip'] = $_SERVER['HTTP_CLIENT_IP'];
1276         }
1277         if ( !empty( $_SERVER['HTTP_FROM'] ) ) {
1278                 $ctx['from'] = $_SERVER['HTTP_FROM'];
1279         }
1280         if ( isset( $ctx['forwarded_for'] ) ||
1281                 isset( $ctx['client_ip'] ) ||
1282                 isset( $ctx['from'] ) ) {
1283                 $ctx['proxy'] = $_SERVER['REMOTE_ADDR'];
1284         }
1285
1286         // Don't load $wgUser at this late stage just for statistics purposes
1287         // @todo FIXME: We can detect some anons even if it is not loaded.
1288         // See User::getId()
1289         $user = $context->getUser();
1290         $ctx['anon'] = $user->isItemLoaded( 'id' ) && $user->isAnon();
1291
1292         // Command line script uses a FauxRequest object which does not have
1293         // any knowledge about an URL and throw an exception instead.
1294         try {
1295                 $ctx['url'] = urldecode( $request->getRequestURL() );
1296         } catch ( Exception $ignored ) {
1297                 // no-op
1298         }
1299
1300         $ctx['output'] = $profiler->getOutput();
1301
1302         $log = LoggerFactory::getInstance( 'profileoutput' );
1303         $log->info( "Elapsed: {elapsed}; URL: <{url}>\n{output}", $ctx );
1304 }
1305
1306 /**
1307  * Increment a statistics counter
1308  *
1309  * @param string $key
1310  * @param int $count
1311  * @return void
1312  */
1313 function wfIncrStats( $key, $count = 1 ) {
1314         $stats = MediaWikiServices::getInstance()->getStatsdDataFactory();
1315         $stats->updateCount( $key, $count );
1316 }
1317
1318 /**
1319  * Check whether the wiki is in read-only mode.
1320  *
1321  * @return bool
1322  */
1323 function wfReadOnly() {
1324         return MediaWikiServices::getInstance()->getReadOnlyMode()
1325                 ->isReadOnly();
1326 }
1327
1328 /**
1329  * Check if the site is in read-only mode and return the message if so
1330  *
1331  * This checks wfConfiguredReadOnlyReason() and the main load balancer
1332  * for replica DB lag. This may result in DB connection being made.
1333  *
1334  * @return string|bool String when in read-only mode; false otherwise
1335  */
1336 function wfReadOnlyReason() {
1337         return MediaWikiServices::getInstance()->getReadOnlyMode()
1338                 ->getReason();
1339 }
1340
1341 /**
1342  * Get the value of $wgReadOnly or the contents of $wgReadOnlyFile.
1343  *
1344  * @return string|bool String when in read-only mode; false otherwise
1345  * @since 1.27
1346  */
1347 function wfConfiguredReadOnlyReason() {
1348         return MediaWikiServices::getInstance()->getConfiguredReadOnlyMode()
1349                 ->getReason();
1350 }
1351
1352 /**
1353  * Return a Language object from $langcode
1354  *
1355  * @param Language|string|bool $langcode Either:
1356  *                  - a Language object
1357  *                  - code of the language to get the message for, if it is
1358  *                    a valid code create a language for that language, if
1359  *                    it is a string but not a valid code then make a basic
1360  *                    language object
1361  *                  - a boolean: if it's false then use the global object for
1362  *                    the current user's language (as a fallback for the old parameter
1363  *                    functionality), or if it is true then use global object
1364  *                    for the wiki's content language.
1365  * @return Language
1366  */
1367 function wfGetLangObj( $langcode = false ) {
1368         # Identify which language to get or create a language object for.
1369         # Using is_object here due to Stub objects.
1370         if ( is_object( $langcode ) ) {
1371                 # Great, we already have the object (hopefully)!
1372                 return $langcode;
1373         }
1374
1375         global $wgContLang, $wgLanguageCode;
1376         if ( $langcode === true || $langcode === $wgLanguageCode ) {
1377                 # $langcode is the language code of the wikis content language object.
1378                 # or it is a boolean and value is true
1379                 return $wgContLang;
1380         }
1381
1382         global $wgLang;
1383         if ( $langcode === false || $langcode === $wgLang->getCode() ) {
1384                 # $langcode is the language code of user language object.
1385                 # or it was a boolean and value is false
1386                 return $wgLang;
1387         }
1388
1389         $validCodes = array_keys( Language::fetchLanguageNames() );
1390         if ( in_array( $langcode, $validCodes ) ) {
1391                 # $langcode corresponds to a valid language.
1392                 return Language::factory( $langcode );
1393         }
1394
1395         # $langcode is a string, but not a valid language code; use content language.
1396         wfDebug( "Invalid language code passed to wfGetLangObj, falling back to content language.\n" );
1397         return $wgContLang;
1398 }
1399
1400 /**
1401  * This is the function for getting translated interface messages.
1402  *
1403  * @see Message class for documentation how to use them.
1404  * @see https://www.mediawiki.org/wiki/Manual:Messages_API
1405  *
1406  * This function replaces all old wfMsg* functions.
1407  *
1408  * @param string|string[]|MessageSpecifier $key Message key, or array of keys, or a MessageSpecifier
1409  * @param mixed $params,... Normal message parameters
1410  * @return Message
1411  *
1412  * @since 1.17
1413  *
1414  * @see Message::__construct
1415  */
1416 function wfMessage( $key /*...*/ ) {
1417         $message = new Message( $key );
1418
1419         // We call Message::params() to reduce code duplication
1420         $params = func_get_args();
1421         array_shift( $params );
1422         if ( $params ) {
1423                 call_user_func_array( [ $message, 'params' ], $params );
1424         }
1425
1426         return $message;
1427 }
1428
1429 /**
1430  * This function accepts multiple message keys and returns a message instance
1431  * for the first message which is non-empty. If all messages are empty then an
1432  * instance of the first message key is returned.
1433  *
1434  * @param string|string[] $keys,... Message keys
1435  * @return Message
1436  *
1437  * @since 1.18
1438  *
1439  * @see Message::newFallbackSequence
1440  */
1441 function wfMessageFallback( /*...*/ ) {
1442         $args = func_get_args();
1443         return call_user_func_array( 'Message::newFallbackSequence', $args );
1444 }
1445
1446 /**
1447  * Replace message parameter keys on the given formatted output.
1448  *
1449  * @param string $message
1450  * @param array $args
1451  * @return string
1452  * @private
1453  */
1454 function wfMsgReplaceArgs( $message, $args ) {
1455         # Fix windows line-endings
1456         # Some messages are split with explode("\n", $msg)
1457         $message = str_replace( "\r", '', $message );
1458
1459         // Replace arguments
1460         if ( is_array( $args ) && $args ) {
1461                 if ( is_array( $args[0] ) ) {
1462                         $args = array_values( $args[0] );
1463                 }
1464                 $replacementKeys = [];
1465                 foreach ( $args as $n => $param ) {
1466                         $replacementKeys['$' . ( $n + 1 )] = $param;
1467                 }
1468                 $message = strtr( $message, $replacementKeys );
1469         }
1470
1471         return $message;
1472 }
1473
1474 /**
1475  * Fetch server name for use in error reporting etc.
1476  * Use real server name if available, so we know which machine
1477  * in a server farm generated the current page.
1478  *
1479  * @return string
1480  */
1481 function wfHostname() {
1482         static $host;
1483         if ( is_null( $host ) ) {
1484                 # Hostname overriding
1485                 global $wgOverrideHostname;
1486                 if ( $wgOverrideHostname !== false ) {
1487                         # Set static and skip any detection
1488                         $host = $wgOverrideHostname;
1489                         return $host;
1490                 }
1491
1492                 if ( function_exists( 'posix_uname' ) ) {
1493                         // This function not present on Windows
1494                         $uname = posix_uname();
1495                 } else {
1496                         $uname = false;
1497                 }
1498                 if ( is_array( $uname ) && isset( $uname['nodename'] ) ) {
1499                         $host = $uname['nodename'];
1500                 } elseif ( getenv( 'COMPUTERNAME' ) ) {
1501                         # Windows computer name
1502                         $host = getenv( 'COMPUTERNAME' );
1503                 } else {
1504                         # This may be a virtual server.
1505                         $host = $_SERVER['SERVER_NAME'];
1506                 }
1507         }
1508         return $host;
1509 }
1510
1511 /**
1512  * Returns a script tag that stores the amount of time it took MediaWiki to
1513  * handle the request in milliseconds as 'wgBackendResponseTime'.
1514  *
1515  * If $wgShowHostnames is true, the script will also set 'wgHostname' to the
1516  * hostname of the server handling the request.
1517  *
1518  * @return string
1519  */
1520 function wfReportTime() {
1521         global $wgRequestTime, $wgShowHostnames;
1522
1523         $responseTime = round( ( microtime( true ) - $wgRequestTime ) * 1000 );
1524         $reportVars = [ 'wgBackendResponseTime' => $responseTime ];
1525         if ( $wgShowHostnames ) {
1526                 $reportVars['wgHostname'] = wfHostname();
1527         }
1528         return Skin::makeVariablesScript( $reportVars );
1529 }
1530
1531 /**
1532  * Safety wrapper for debug_backtrace().
1533  *
1534  * Will return an empty array if debug_backtrace is disabled, otherwise
1535  * the output from debug_backtrace() (trimmed).
1536  *
1537  * @param int $limit This parameter can be used to limit the number of stack frames returned
1538  *
1539  * @return array Array of backtrace information
1540  */
1541 function wfDebugBacktrace( $limit = 0 ) {
1542         static $disabled = null;
1543
1544         if ( is_null( $disabled ) ) {
1545                 $disabled = !function_exists( 'debug_backtrace' );
1546                 if ( $disabled ) {
1547                         wfDebug( "debug_backtrace() is disabled\n" );
1548                 }
1549         }
1550         if ( $disabled ) {
1551                 return [];
1552         }
1553
1554         if ( $limit ) {
1555                 return array_slice( debug_backtrace( DEBUG_BACKTRACE_PROVIDE_OBJECT, $limit + 1 ), 1 );
1556         } else {
1557                 return array_slice( debug_backtrace(), 1 );
1558         }
1559 }
1560
1561 /**
1562  * Get a debug backtrace as a string
1563  *
1564  * @param bool|null $raw If true, the return value is plain text. If false, HTML.
1565  *   Defaults to $wgCommandLineMode if unset.
1566  * @return string
1567  * @since 1.25 Supports $raw parameter.
1568  */
1569 function wfBacktrace( $raw = null ) {
1570         global $wgCommandLineMode;
1571
1572         if ( $raw === null ) {
1573                 $raw = $wgCommandLineMode;
1574         }
1575
1576         if ( $raw ) {
1577                 $frameFormat = "%s line %s calls %s()\n";
1578                 $traceFormat = "%s";
1579         } else {
1580                 $frameFormat = "<li>%s line %s calls %s()</li>\n";
1581                 $traceFormat = "<ul>\n%s</ul>\n";
1582         }
1583
1584         $frames = array_map( function ( $frame ) use ( $frameFormat ) {
1585                 $file = !empty( $frame['file'] ) ? basename( $frame['file'] ) : '-';
1586                 $line = isset( $frame['line'] ) ? $frame['line'] : '-';
1587                 $call = $frame['function'];
1588                 if ( !empty( $frame['class'] ) ) {
1589                         $call = $frame['class'] . $frame['type'] . $call;
1590                 }
1591                 return sprintf( $frameFormat, $file, $line, $call );
1592         }, wfDebugBacktrace() );
1593
1594         return sprintf( $traceFormat, implode( '', $frames ) );
1595 }
1596
1597 /**
1598  * Get the name of the function which called this function
1599  * wfGetCaller( 1 ) is the function with the wfGetCaller() call (ie. __FUNCTION__)
1600  * wfGetCaller( 2 ) [default] is the caller of the function running wfGetCaller()
1601  * wfGetCaller( 3 ) is the parent of that.
1602  *
1603  * @param int $level
1604  * @return string
1605  */
1606 function wfGetCaller( $level = 2 ) {
1607         $backtrace = wfDebugBacktrace( $level + 1 );
1608         if ( isset( $backtrace[$level] ) ) {
1609                 return wfFormatStackFrame( $backtrace[$level] );
1610         } else {
1611                 return 'unknown';
1612         }
1613 }
1614
1615 /**
1616  * Return a string consisting of callers in the stack. Useful sometimes
1617  * for profiling specific points.
1618  *
1619  * @param int $limit The maximum depth of the stack frame to return, or false for the entire stack.
1620  * @return string
1621  */
1622 function wfGetAllCallers( $limit = 3 ) {
1623         $trace = array_reverse( wfDebugBacktrace() );
1624         if ( !$limit || $limit > count( $trace ) - 1 ) {
1625                 $limit = count( $trace ) - 1;
1626         }
1627         $trace = array_slice( $trace, -$limit - 1, $limit );
1628         return implode( '/', array_map( 'wfFormatStackFrame', $trace ) );
1629 }
1630
1631 /**
1632  * Return a string representation of frame
1633  *
1634  * @param array $frame
1635  * @return string
1636  */
1637 function wfFormatStackFrame( $frame ) {
1638         if ( !isset( $frame['function'] ) ) {
1639                 return 'NO_FUNCTION_GIVEN';
1640         }
1641         return isset( $frame['class'] ) && isset( $frame['type'] ) ?
1642                 $frame['class'] . $frame['type'] . $frame['function'] :
1643                 $frame['function'];
1644 }
1645
1646 /* Some generic result counters, pulled out of SearchEngine */
1647
1648 /**
1649  * @todo document
1650  *
1651  * @param int $offset
1652  * @param int $limit
1653  * @return string
1654  */
1655 function wfShowingResults( $offset, $limit ) {
1656         return wfMessage( 'showingresults' )->numParams( $limit, $offset + 1 )->parse();
1657 }
1658
1659 /**
1660  * Whether the client accept gzip encoding
1661  *
1662  * Uses the Accept-Encoding header to check if the client supports gzip encoding.
1663  * Use this when considering to send a gzip-encoded response to the client.
1664  *
1665  * @param bool $force Forces another check even if we already have a cached result.
1666  * @return bool
1667  */
1668 function wfClientAcceptsGzip( $force = false ) {
1669         static $result = null;
1670         if ( $result === null || $force ) {
1671                 $result = false;
1672                 if ( isset( $_SERVER['HTTP_ACCEPT_ENCODING'] ) ) {
1673                         # @todo FIXME: We may want to blacklist some broken browsers
1674                         $m = [];
1675                         if ( preg_match(
1676                                         '/\bgzip(?:;(q)=([0-9]+(?:\.[0-9]+)))?\b/',
1677                                         $_SERVER['HTTP_ACCEPT_ENCODING'],
1678                                         $m
1679                                 )
1680                         ) {
1681                                 if ( isset( $m[2] ) && ( $m[1] == 'q' ) && ( $m[2] == 0 ) ) {
1682                                         $result = false;
1683                                         return $result;
1684                                 }
1685                                 wfDebug( "wfClientAcceptsGzip: client accepts gzip.\n" );
1686                                 $result = true;
1687                         }
1688                 }
1689         }
1690         return $result;
1691 }
1692
1693 /**
1694  * Escapes the given text so that it may be output using addWikiText()
1695  * without any linking, formatting, etc. making its way through. This
1696  * is achieved by substituting certain characters with HTML entities.
1697  * As required by the callers, "<nowiki>" is not used.
1698  *
1699  * @param string $text Text to be escaped
1700  * @return string
1701  */
1702 function wfEscapeWikiText( $text ) {
1703         global $wgEnableMagicLinks;
1704         static $repl = null, $repl2 = null;
1705         if ( $repl === null || defined( 'MW_PARSER_TEST' ) || defined( 'MW_PHPUNIT_TEST' ) ) {
1706                 // Tests depend upon being able to change $wgEnableMagicLinks, so don't cache
1707                 // in those situations
1708                 $repl = [
1709                         '"' => '&#34;', '&' => '&#38;', "'" => '&#39;', '<' => '&#60;',
1710                         '=' => '&#61;', '>' => '&#62;', '[' => '&#91;', ']' => '&#93;',
1711                         '{' => '&#123;', '|' => '&#124;', '}' => '&#125;', ';' => '&#59;',
1712                         "\n#" => "\n&#35;", "\r#" => "\r&#35;",
1713                         "\n*" => "\n&#42;", "\r*" => "\r&#42;",
1714                         "\n:" => "\n&#58;", "\r:" => "\r&#58;",
1715                         "\n " => "\n&#32;", "\r " => "\r&#32;",
1716                         "\n\n" => "\n&#10;", "\r\n" => "&#13;\n",
1717                         "\n\r" => "\n&#13;", "\r\r" => "\r&#13;",
1718                         "\n\t" => "\n&#9;", "\r\t" => "\r&#9;", // "\n\t\n" is treated like "\n\n"
1719                         "\n----" => "\n&#45;---", "\r----" => "\r&#45;---",
1720                         '__' => '_&#95;', '://' => '&#58;//',
1721                 ];
1722
1723                 $magicLinks = array_keys( array_filter( $wgEnableMagicLinks ) );
1724                 // We have to catch everything "\s" matches in PCRE
1725                 foreach ( $magicLinks as $magic ) {
1726                         $repl["$magic "] = "$magic&#32;";
1727                         $repl["$magic\t"] = "$magic&#9;";
1728                         $repl["$magic\r"] = "$magic&#13;";
1729                         $repl["$magic\n"] = "$magic&#10;";
1730                         $repl["$magic\f"] = "$magic&#12;";
1731                 }
1732
1733                 // And handle protocols that don't use "://"
1734                 global $wgUrlProtocols;
1735                 $repl2 = [];
1736                 foreach ( $wgUrlProtocols as $prot ) {
1737                         if ( substr( $prot, -1 ) === ':' ) {
1738                                 $repl2[] = preg_quote( substr( $prot, 0, -1 ), '/' );
1739                         }
1740                 }
1741                 $repl2 = $repl2 ? '/\b(' . implode( '|', $repl2 ) . '):/i' : '/^(?!)/';
1742         }
1743         $text = substr( strtr( "\n$text", $repl ), 1 );
1744         $text = preg_replace( $repl2, '$1&#58;', $text );
1745         return $text;
1746 }
1747
1748 /**
1749  * Sets dest to source and returns the original value of dest
1750  * If source is NULL, it just returns the value, it doesn't set the variable
1751  * If force is true, it will set the value even if source is NULL
1752  *
1753  * @param mixed $dest
1754  * @param mixed $source
1755  * @param bool $force
1756  * @return mixed
1757  */
1758 function wfSetVar( &$dest, $source, $force = false ) {
1759         $temp = $dest;
1760         if ( !is_null( $source ) || $force ) {
1761                 $dest = $source;
1762         }
1763         return $temp;
1764 }
1765
1766 /**
1767  * As for wfSetVar except setting a bit
1768  *
1769  * @param int $dest
1770  * @param int $bit
1771  * @param bool $state
1772  *
1773  * @return bool
1774  */
1775 function wfSetBit( &$dest, $bit, $state = true ) {
1776         $temp = (bool)( $dest & $bit );
1777         if ( !is_null( $state ) ) {
1778                 if ( $state ) {
1779                         $dest |= $bit;
1780                 } else {
1781                         $dest &= ~$bit;
1782                 }
1783         }
1784         return $temp;
1785 }
1786
1787 /**
1788  * A wrapper around the PHP function var_export().
1789  * Either print it or add it to the regular output ($wgOut).
1790  *
1791  * @param mixed $var A PHP variable to dump.
1792  */
1793 function wfVarDump( $var ) {
1794         global $wgOut;
1795         $s = str_replace( "\n", "<br />\n", var_export( $var, true ) . "\n" );
1796         if ( headers_sent() || !isset( $wgOut ) || !is_object( $wgOut ) ) {
1797                 print $s;
1798         } else {
1799                 $wgOut->addHTML( $s );
1800         }
1801 }
1802
1803 /**
1804  * Provide a simple HTTP error.
1805  *
1806  * @param int|string $code
1807  * @param string $label
1808  * @param string $desc
1809  */
1810 function wfHttpError( $code, $label, $desc ) {
1811         global $wgOut;
1812         HttpStatus::header( $code );
1813         if ( $wgOut ) {
1814                 $wgOut->disable();
1815                 $wgOut->sendCacheControl();
1816         }
1817
1818         MediaWiki\HeaderCallback::warnIfHeadersSent();
1819         header( 'Content-type: text/html; charset=utf-8' );
1820         print '<!DOCTYPE html>' .
1821                 '<html><head><title>' .
1822                 htmlspecialchars( $label ) .
1823                 '</title></head><body><h1>' .
1824                 htmlspecialchars( $label ) .
1825                 '</h1><p>' .
1826                 nl2br( htmlspecialchars( $desc ) ) .
1827                 "</p></body></html>\n";
1828 }
1829
1830 /**
1831  * Clear away any user-level output buffers, discarding contents.
1832  *
1833  * Suitable for 'starting afresh', for instance when streaming
1834  * relatively large amounts of data without buffering, or wanting to
1835  * output image files without ob_gzhandler's compression.
1836  *
1837  * The optional $resetGzipEncoding parameter controls suppression of
1838  * the Content-Encoding header sent by ob_gzhandler; by default it
1839  * is left. See comments for wfClearOutputBuffers() for why it would
1840  * be used.
1841  *
1842  * Note that some PHP configuration options may add output buffer
1843  * layers which cannot be removed; these are left in place.
1844  *
1845  * @param bool $resetGzipEncoding
1846  */
1847 function wfResetOutputBuffers( $resetGzipEncoding = true ) {
1848         if ( $resetGzipEncoding ) {
1849                 // Suppress Content-Encoding and Content-Length
1850                 // headers from 1.10+s wfOutputHandler
1851                 global $wgDisableOutputCompression;
1852                 $wgDisableOutputCompression = true;
1853         }
1854         while ( $status = ob_get_status() ) {
1855                 if ( isset( $status['flags'] ) ) {
1856                         $flags = PHP_OUTPUT_HANDLER_CLEANABLE | PHP_OUTPUT_HANDLER_REMOVABLE;
1857                         $deleteable = ( $status['flags'] & $flags ) === $flags;
1858                 } elseif ( isset( $status['del'] ) ) {
1859                         $deleteable = $status['del'];
1860                 } else {
1861                         // Guess that any PHP-internal setting can't be removed.
1862                         $deleteable = $status['type'] !== 0; /* PHP_OUTPUT_HANDLER_INTERNAL */
1863                 }
1864                 if ( !$deleteable ) {
1865                         // Give up, and hope the result doesn't break
1866                         // output behavior.
1867                         break;
1868                 }
1869                 if ( $status['name'] === 'MediaWikiTestCase::wfResetOutputBuffersBarrier' ) {
1870                         // Unit testing barrier to prevent this function from breaking PHPUnit.
1871                         break;
1872                 }
1873                 if ( !ob_end_clean() ) {
1874                         // Could not remove output buffer handler; abort now
1875                         // to avoid getting in some kind of infinite loop.
1876                         break;
1877                 }
1878                 if ( $resetGzipEncoding ) {
1879                         if ( $status['name'] == 'ob_gzhandler' ) {
1880                                 // Reset the 'Content-Encoding' field set by this handler
1881                                 // so we can start fresh.
1882                                 header_remove( 'Content-Encoding' );
1883                                 break;
1884                         }
1885                 }
1886         }
1887 }
1888
1889 /**
1890  * More legible than passing a 'false' parameter to wfResetOutputBuffers():
1891  *
1892  * Clear away output buffers, but keep the Content-Encoding header
1893  * produced by ob_gzhandler, if any.
1894  *
1895  * This should be used for HTTP 304 responses, where you need to
1896  * preserve the Content-Encoding header of the real result, but
1897  * also need to suppress the output of ob_gzhandler to keep to spec
1898  * and avoid breaking Firefox in rare cases where the headers and
1899  * body are broken over two packets.
1900  */
1901 function wfClearOutputBuffers() {
1902         wfResetOutputBuffers( false );
1903 }
1904
1905 /**
1906  * Converts an Accept-* header into an array mapping string values to quality
1907  * factors
1908  *
1909  * @param string $accept
1910  * @param string $def Default
1911  * @return float[] Associative array of string => float pairs
1912  */
1913 function wfAcceptToPrefs( $accept, $def = '*/*' ) {
1914         # No arg means accept anything (per HTTP spec)
1915         if ( !$accept ) {
1916                 return [ $def => 1.0 ];
1917         }
1918
1919         $prefs = [];
1920
1921         $parts = explode( ',', $accept );
1922
1923         foreach ( $parts as $part ) {
1924                 # @todo FIXME: Doesn't deal with params like 'text/html; level=1'
1925                 $values = explode( ';', trim( $part ) );
1926                 $match = [];
1927                 if ( count( $values ) == 1 ) {
1928                         $prefs[$values[0]] = 1.0;
1929                 } elseif ( preg_match( '/q\s*=\s*(\d*\.\d+)/', $values[1], $match ) ) {
1930                         $prefs[$values[0]] = floatval( $match[1] );
1931                 }
1932         }
1933
1934         return $prefs;
1935 }
1936
1937 /**
1938  * Checks if a given MIME type matches any of the keys in the given
1939  * array. Basic wildcards are accepted in the array keys.
1940  *
1941  * Returns the matching MIME type (or wildcard) if a match, otherwise
1942  * NULL if no match.
1943  *
1944  * @param string $type
1945  * @param array $avail
1946  * @return string
1947  * @private
1948  */
1949 function mimeTypeMatch( $type, $avail ) {
1950         if ( array_key_exists( $type, $avail ) ) {
1951                 return $type;
1952         } else {
1953                 $mainType = explode( '/', $type )[0];
1954                 if ( array_key_exists( "$mainType/*", $avail ) ) {
1955                         return "$mainType/*";
1956                 } elseif ( array_key_exists( '*/*', $avail ) ) {
1957                         return '*/*';
1958                 } else {
1959                         return null;
1960                 }
1961         }
1962 }
1963
1964 /**
1965  * Returns the 'best' match between a client's requested internet media types
1966  * and the server's list of available types. Each list should be an associative
1967  * array of type to preference (preference is a float between 0.0 and 1.0).
1968  * Wildcards in the types are acceptable.
1969  *
1970  * @param array $cprefs Client's acceptable type list
1971  * @param array $sprefs Server's offered types
1972  * @return string
1973  *
1974  * @todo FIXME: Doesn't handle params like 'text/plain; charset=UTF-8'
1975  * XXX: generalize to negotiate other stuff
1976  */
1977 function wfNegotiateType( $cprefs, $sprefs ) {
1978         $combine = [];
1979
1980         foreach ( array_keys( $sprefs ) as $type ) {
1981                 $subType = explode( '/', $type )[1];
1982                 if ( $subType != '*' ) {
1983                         $ckey = mimeTypeMatch( $type, $cprefs );
1984                         if ( $ckey ) {
1985                                 $combine[$type] = $sprefs[$type] * $cprefs[$ckey];
1986                         }
1987                 }
1988         }
1989
1990         foreach ( array_keys( $cprefs ) as $type ) {
1991                 $subType = explode( '/', $type )[1];
1992                 if ( $subType != '*' && !array_key_exists( $type, $sprefs ) ) {
1993                         $skey = mimeTypeMatch( $type, $sprefs );
1994                         if ( $skey ) {
1995                                 $combine[$type] = $sprefs[$skey] * $cprefs[$type];
1996                         }
1997                 }
1998         }
1999
2000         $bestq = 0;
2001         $besttype = null;
2002
2003         foreach ( array_keys( $combine ) as $type ) {
2004                 if ( $combine[$type] > $bestq ) {
2005                         $besttype = $type;
2006                         $bestq = $combine[$type];
2007                 }
2008         }
2009
2010         return $besttype;
2011 }
2012
2013 /**
2014  * Reference-counted warning suppression
2015  *
2016  * @deprecated since 1.26, use MediaWiki\suppressWarnings() directly
2017  * @param bool $end
2018  */
2019 function wfSuppressWarnings( $end = false ) {
2020         MediaWiki\suppressWarnings( $end );
2021 }
2022
2023 /**
2024  * @deprecated since 1.26, use MediaWiki\restoreWarnings() directly
2025  * Restore error level to previous value
2026  */
2027 function wfRestoreWarnings() {
2028         MediaWiki\suppressWarnings( true );
2029 }
2030
2031 /**
2032  * Get a timestamp string in one of various formats
2033  *
2034  * @param mixed $outputtype A timestamp in one of the supported formats, the
2035  *   function will autodetect which format is supplied and act accordingly.
2036  * @param mixed $ts Optional timestamp to convert, default 0 for the current time
2037  * @return string|bool String / false The same date in the format specified in $outputtype or false
2038  */
2039 function wfTimestamp( $outputtype = TS_UNIX, $ts = 0 ) {
2040         $ret = MWTimestamp::convert( $outputtype, $ts );
2041         if ( $ret === false ) {
2042                 wfDebug( "wfTimestamp() fed bogus time value: TYPE=$outputtype; VALUE=$ts\n" );
2043         }
2044         return $ret;
2045 }
2046
2047 /**
2048  * Return a formatted timestamp, or null if input is null.
2049  * For dealing with nullable timestamp columns in the database.
2050  *
2051  * @param int $outputtype
2052  * @param string $ts
2053  * @return string
2054  */
2055 function wfTimestampOrNull( $outputtype = TS_UNIX, $ts = null ) {
2056         if ( is_null( $ts ) ) {
2057                 return null;
2058         } else {
2059                 return wfTimestamp( $outputtype, $ts );
2060         }
2061 }
2062
2063 /**
2064  * Convenience function; returns MediaWiki timestamp for the present time.
2065  *
2066  * @return string
2067  */
2068 function wfTimestampNow() {
2069         # return NOW
2070         return MWTimestamp::now( TS_MW );
2071 }
2072
2073 /**
2074  * Check if the operating system is Windows
2075  *
2076  * @return bool True if it's Windows, false otherwise.
2077  */
2078 function wfIsWindows() {
2079         static $isWindows = null;
2080         if ( $isWindows === null ) {
2081                 $isWindows = strtoupper( substr( PHP_OS, 0, 3 ) ) === 'WIN';
2082         }
2083         return $isWindows;
2084 }
2085
2086 /**
2087  * Check if we are running under HHVM
2088  *
2089  * @return bool
2090  */
2091 function wfIsHHVM() {
2092         return defined( 'HHVM_VERSION' );
2093 }
2094
2095 /**
2096  * Tries to get the system directory for temporary files. First
2097  * $wgTmpDirectory is checked, and then the TMPDIR, TMP, and TEMP
2098  * environment variables are then checked in sequence, then
2099  * sys_get_temp_dir(), then upload_tmp_dir from php.ini.
2100  *
2101  * NOTE: When possible, use instead the tmpfile() function to create
2102  * temporary files to avoid race conditions on file creation, etc.
2103  *
2104  * @return string
2105  */
2106 function wfTempDir() {
2107         global $wgTmpDirectory;
2108
2109         if ( $wgTmpDirectory !== false ) {
2110                 return $wgTmpDirectory;
2111         }
2112
2113         return TempFSFile::getUsableTempDirectory();
2114 }
2115
2116 /**
2117  * Make directory, and make all parent directories if they don't exist
2118  *
2119  * @param string $dir Full path to directory to create
2120  * @param int $mode Chmod value to use, default is $wgDirectoryMode
2121  * @param string $caller Optional caller param for debugging.
2122  * @throws MWException
2123  * @return bool
2124  */
2125 function wfMkdirParents( $dir, $mode = null, $caller = null ) {
2126         global $wgDirectoryMode;
2127
2128         if ( FileBackend::isStoragePath( $dir ) ) { // sanity
2129                 throw new MWException( __FUNCTION__ . " given storage path '$dir'." );
2130         }
2131
2132         if ( !is_null( $caller ) ) {
2133                 wfDebug( "$caller: called wfMkdirParents($dir)\n" );
2134         }
2135
2136         if ( strval( $dir ) === '' || is_dir( $dir ) ) {
2137                 return true;
2138         }
2139
2140         $dir = str_replace( [ '\\', '/' ], DIRECTORY_SEPARATOR, $dir );
2141
2142         if ( is_null( $mode ) ) {
2143                 $mode = $wgDirectoryMode;
2144         }
2145
2146         // Turn off the normal warning, we're doing our own below
2147         MediaWiki\suppressWarnings();
2148         $ok = mkdir( $dir, $mode, true ); // PHP5 <3
2149         MediaWiki\restoreWarnings();
2150
2151         if ( !$ok ) {
2152                 // directory may have been created on another request since we last checked
2153                 if ( is_dir( $dir ) ) {
2154                         return true;
2155                 }
2156
2157                 // PHP doesn't report the path in its warning message, so add our own to aid in diagnosis.
2158                 wfLogWarning( sprintf( "failed to mkdir \"%s\" mode 0%o", $dir, $mode ) );
2159         }
2160         return $ok;
2161 }
2162
2163 /**
2164  * Remove a directory and all its content.
2165  * Does not hide error.
2166  * @param string $dir
2167  */
2168 function wfRecursiveRemoveDir( $dir ) {
2169         wfDebug( __FUNCTION__ . "( $dir )\n" );
2170         // taken from https://secure.php.net/manual/en/function.rmdir.php#98622
2171         if ( is_dir( $dir ) ) {
2172                 $objects = scandir( $dir );
2173                 foreach ( $objects as $object ) {
2174                         if ( $object != "." && $object != ".." ) {
2175                                 if ( filetype( $dir . '/' . $object ) == "dir" ) {
2176                                         wfRecursiveRemoveDir( $dir . '/' . $object );
2177                                 } else {
2178                                         unlink( $dir . '/' . $object );
2179                                 }
2180                         }
2181                 }
2182                 reset( $objects );
2183                 rmdir( $dir );
2184         }
2185 }
2186
2187 /**
2188  * @param int $nr The number to format
2189  * @param int $acc The number of digits after the decimal point, default 2
2190  * @param bool $round Whether or not to round the value, default true
2191  * @return string
2192  */
2193 function wfPercent( $nr, $acc = 2, $round = true ) {
2194         $ret = sprintf( "%.${acc}f", $nr );
2195         return $round ? round( $ret, $acc ) . '%' : "$ret%";
2196 }
2197
2198 /**
2199  * Safety wrapper around ini_get() for boolean settings.
2200  * The values returned from ini_get() are pre-normalized for settings
2201  * set via php.ini or php_flag/php_admin_flag... but *not*
2202  * for those set via php_value/php_admin_value.
2203  *
2204  * It's fairly common for people to use php_value instead of php_flag,
2205  * which can leave you with an 'off' setting giving a false positive
2206  * for code that just takes the ini_get() return value as a boolean.
2207  *
2208  * To make things extra interesting, setting via php_value accepts
2209  * "true" and "yes" as true, but php.ini and php_flag consider them false. :)
2210  * Unrecognized values go false... again opposite PHP's own coercion
2211  * from string to bool.
2212  *
2213  * Luckily, 'properly' set settings will always come back as '0' or '1',
2214  * so we only have to worry about them and the 'improper' settings.
2215  *
2216  * I frickin' hate PHP... :P
2217  *
2218  * @param string $setting
2219  * @return bool
2220  */
2221 function wfIniGetBool( $setting ) {
2222         $val = strtolower( ini_get( $setting ) );
2223         // 'on' and 'true' can't have whitespace around them, but '1' can.
2224         return $val == 'on'
2225                 || $val == 'true'
2226                 || $val == 'yes'
2227                 || preg_match( "/^\s*[+-]?0*[1-9]/", $val ); // approx C atoi() function
2228 }
2229
2230 /**
2231  * Version of escapeshellarg() that works better on Windows.
2232  *
2233  * Originally, this fixed the incorrect use of single quotes on Windows
2234  * (https://bugs.php.net/bug.php?id=26285) and the locale problems on Linux in
2235  * PHP 5.2.6+ (bug backported to earlier distro releases of PHP).
2236  *
2237  * @param string ... strings to escape and glue together, or a single array of strings parameter
2238  * @return string
2239  */
2240 function wfEscapeShellArg( /*...*/ ) {
2241         $args = func_get_args();
2242         if ( count( $args ) === 1 && is_array( reset( $args ) ) ) {
2243                 // If only one argument has been passed, and that argument is an array,
2244                 // treat it as a list of arguments
2245                 $args = reset( $args );
2246         }
2247
2248         $first = true;
2249         $retVal = '';
2250         foreach ( $args as $arg ) {
2251                 if ( !$first ) {
2252                         $retVal .= ' ';
2253                 } else {
2254                         $first = false;
2255                 }
2256
2257                 if ( wfIsWindows() ) {
2258                         // Escaping for an MSVC-style command line parser and CMD.EXE
2259                         // @codingStandardsIgnoreStart For long URLs
2260                         // Refs:
2261                         //  * https://web.archive.org/web/20020708081031/http://mailman.lyra.org/pipermail/scite-interest/2002-March/000436.html
2262                         //  * https://technet.microsoft.com/en-us/library/cc723564.aspx
2263                         //  * T15518
2264                         //  * CR r63214
2265                         // Double the backslashes before any double quotes. Escape the double quotes.
2266                         // @codingStandardsIgnoreEnd
2267                         $tokens = preg_split( '/(\\\\*")/', $arg, -1, PREG_SPLIT_DELIM_CAPTURE );
2268                         $arg = '';
2269                         $iteration = 0;
2270                         foreach ( $tokens as $token ) {
2271                                 if ( $iteration % 2 == 1 ) {
2272                                         // Delimiter, a double quote preceded by zero or more slashes
2273                                         $arg .= str_replace( '\\', '\\\\', substr( $token, 0, -1 ) ) . '\\"';
2274                                 } elseif ( $iteration % 4 == 2 ) {
2275                                         // ^ in $token will be outside quotes, need to be escaped
2276                                         $arg .= str_replace( '^', '^^', $token );
2277                                 } else { // $iteration % 4 == 0
2278                                         // ^ in $token will appear inside double quotes, so leave as is
2279                                         $arg .= $token;
2280                                 }
2281                                 $iteration++;
2282                         }
2283                         // Double the backslashes before the end of the string, because
2284                         // we will soon add a quote
2285                         $m = [];
2286                         if ( preg_match( '/^(.*?)(\\\\+)$/', $arg, $m ) ) {
2287                                 $arg = $m[1] . str_replace( '\\', '\\\\', $m[2] );
2288                         }
2289
2290                         // Add surrounding quotes
2291                         $retVal .= '"' . $arg . '"';
2292                 } else {
2293                         $retVal .= escapeshellarg( $arg );
2294                 }
2295         }
2296         return $retVal;
2297 }
2298
2299 /**
2300  * Check if wfShellExec() is effectively disabled via php.ini config
2301  *
2302  * @return bool|string False or 'disabled'
2303  * @since 1.22
2304  */
2305 function wfShellExecDisabled() {
2306         static $disabled = null;
2307         if ( is_null( $disabled ) ) {
2308                 if ( !function_exists( 'proc_open' ) ) {
2309                         wfDebug( "proc_open() is disabled\n" );
2310                         $disabled = 'disabled';
2311                 } else {
2312                         $disabled = false;
2313                 }
2314         }
2315         return $disabled;
2316 }
2317
2318 /**
2319  * Execute a shell command, with time and memory limits mirrored from the PHP
2320  * configuration if supported.
2321  *
2322  * @param string|string[] $cmd If string, a properly shell-escaped command line,
2323  *   or an array of unescaped arguments, in which case each value will be escaped
2324  *   Example:   [ 'convert', '-font', 'font name' ] would produce "'convert' '-font' 'font name'"
2325  * @param null|mixed &$retval Optional, will receive the program's exit code.
2326  *   (non-zero is usually failure). If there is an error from
2327  *   read, select, or proc_open(), this will be set to -1.
2328  * @param array $environ Optional environment variables which should be
2329  *   added to the executed command environment.
2330  * @param array $limits Optional array with limits(filesize, memory, time, walltime)
2331  *   this overwrites the global wgMaxShell* limits.
2332  * @param array $options Array of options:
2333  *   - duplicateStderr: Set this to true to duplicate stderr to stdout,
2334  *     including errors from limit.sh
2335  *   - profileMethod: By default this function will profile based on the calling
2336  *     method. Set this to a string for an alternative method to profile from
2337  *
2338  * @return string Collected stdout as a string
2339  */
2340 function wfShellExec( $cmd, &$retval = null, $environ = [],
2341         $limits = [], $options = []
2342 ) {
2343         global $IP, $wgMaxShellMemory, $wgMaxShellFileSize, $wgMaxShellTime,
2344                 $wgMaxShellWallClockTime, $wgShellCgroup;
2345
2346         $disabled = wfShellExecDisabled();
2347         if ( $disabled ) {
2348                 $retval = 1;
2349                 return 'Unable to run external programs, proc_open() is disabled.';
2350         }
2351
2352         $includeStderr = isset( $options['duplicateStderr'] ) && $options['duplicateStderr'];
2353         $profileMethod = isset( $options['profileMethod'] ) ? $options['profileMethod'] : wfGetCaller();
2354
2355         $envcmd = '';
2356         foreach ( $environ as $k => $v ) {
2357                 if ( wfIsWindows() ) {
2358                         /* Surrounding a set in quotes (method used by wfEscapeShellArg) makes the quotes themselves
2359                          * appear in the environment variable, so we must use carat escaping as documented in
2360                          * https://technet.microsoft.com/en-us/library/cc723564.aspx
2361                          * Note however that the quote isn't listed there, but is needed, and the parentheses
2362                          * are listed there but doesn't appear to need it.
2363                          */
2364                         $envcmd .= "set $k=" . preg_replace( '/([&|()<>^"])/', '^\\1', $v ) . '&& ';
2365                 } else {
2366                         /* Assume this is a POSIX shell, thus required to accept variable assignments before the command
2367                          * http://www.opengroup.org/onlinepubs/009695399/utilities/xcu_chap02.html#tag_02_09_01
2368                          */
2369                         $envcmd .= "$k=" . escapeshellarg( $v ) . ' ';
2370                 }
2371         }
2372         if ( is_array( $cmd ) ) {
2373                 $cmd = wfEscapeShellArg( $cmd );
2374         }
2375
2376         $cmd = $envcmd . $cmd;
2377
2378         $useLogPipe = false;
2379         if ( is_executable( '/bin/bash' ) ) {
2380                 $time = intval( isset( $limits['time'] ) ? $limits['time'] : $wgMaxShellTime );
2381                 if ( isset( $limits['walltime'] ) ) {
2382                         $wallTime = intval( $limits['walltime'] );
2383                 } elseif ( isset( $limits['time'] ) ) {
2384                         $wallTime = $time;
2385                 } else {
2386                         $wallTime = intval( $wgMaxShellWallClockTime );
2387                 }
2388                 $mem = intval( isset( $limits['memory'] ) ? $limits['memory'] : $wgMaxShellMemory );
2389                 $filesize = intval( isset( $limits['filesize'] ) ? $limits['filesize'] : $wgMaxShellFileSize );
2390
2391                 if ( $time > 0 || $mem > 0 || $filesize > 0 || $wallTime > 0 ) {
2392                         $cmd = '/bin/bash ' . escapeshellarg( "$IP/includes/limit.sh" ) . ' ' .
2393                                 escapeshellarg( $cmd ) . ' ' .
2394                                 escapeshellarg(
2395                                         "MW_INCLUDE_STDERR=" . ( $includeStderr ? '1' : '' ) . ';' .
2396                                         "MW_CPU_LIMIT=$time; " .
2397                                         'MW_CGROUP=' . escapeshellarg( $wgShellCgroup ) . '; ' .
2398                                         "MW_MEM_LIMIT=$mem; " .
2399                                         "MW_FILE_SIZE_LIMIT=$filesize; " .
2400                                         "MW_WALL_CLOCK_LIMIT=$wallTime; " .
2401                                         "MW_USE_LOG_PIPE=yes"
2402                                 );
2403                         $useLogPipe = true;
2404                 } elseif ( $includeStderr ) {
2405                         $cmd .= ' 2>&1';
2406                 }
2407         } elseif ( $includeStderr ) {
2408                 $cmd .= ' 2>&1';
2409         }
2410         wfDebug( "wfShellExec: $cmd\n" );
2411
2412         // Don't try to execute commands that exceed Linux's MAX_ARG_STRLEN.
2413         // Other platforms may be more accomodating, but we don't want to be
2414         // accomodating, because very long commands probably include user
2415         // input. See T129506.
2416         if ( strlen( $cmd ) > SHELL_MAX_ARG_STRLEN ) {
2417                 throw new Exception( __METHOD__ .
2418                         '(): total length of $cmd must not exceed SHELL_MAX_ARG_STRLEN' );
2419         }
2420
2421         $desc = [
2422                 0 => [ 'file', 'php://stdin', 'r' ],
2423                 1 => [ 'pipe', 'w' ],
2424                 2 => [ 'file', 'php://stderr', 'w' ] ];
2425         if ( $useLogPipe ) {
2426                 $desc[3] = [ 'pipe', 'w' ];
2427         }
2428         $pipes = null;
2429         $scoped = Profiler::instance()->scopedProfileIn( __FUNCTION__ . '-' . $profileMethod );
2430         $proc = proc_open( $cmd, $desc, $pipes );
2431         if ( !$proc ) {
2432                 wfDebugLog( 'exec', "proc_open() failed: $cmd" );
2433                 $retval = -1;
2434                 return '';
2435         }
2436         $outBuffer = $logBuffer = '';
2437         $emptyArray = [];
2438         $status = false;
2439         $logMsg = false;
2440
2441         /* According to the documentation, it is possible for stream_select()
2442          * to fail due to EINTR. I haven't managed to induce this in testing
2443          * despite sending various signals. If it did happen, the error
2444          * message would take the form:
2445          *
2446          * stream_select(): unable to select [4]: Interrupted system call (max_fd=5)
2447          *
2448          * where [4] is the value of the macro EINTR and "Interrupted system
2449          * call" is string which according to the Linux manual is "possibly"
2450          * localised according to LC_MESSAGES.
2451          */
2452         $eintr = defined( 'SOCKET_EINTR' ) ? SOCKET_EINTR : 4;
2453         $eintrMessage = "stream_select(): unable to select [$eintr]";
2454
2455         $running = true;
2456         $timeout = null;
2457         $numReadyPipes = 0;
2458
2459         while ( $running === true || $numReadyPipes !== 0 ) {
2460                 if ( $running ) {
2461                         $status = proc_get_status( $proc );
2462                         // If the process has terminated, switch to nonblocking selects
2463                         // for getting any data still waiting to be read.
2464                         if ( !$status['running'] ) {
2465                                 $running = false;
2466                                 $timeout = 0;
2467                         }
2468                 }
2469
2470                 $readyPipes = $pipes;
2471
2472                 // Clear last error
2473                 // @codingStandardsIgnoreStart Generic.PHP.NoSilencedErrors.Discouraged
2474                 @trigger_error( '' );
2475                 $numReadyPipes = @stream_select( $readyPipes, $emptyArray, $emptyArray, $timeout );
2476                 if ( $numReadyPipes === false ) {
2477                         // @codingStandardsIgnoreEnd
2478                         $error = error_get_last();
2479                         if ( strncmp( $error['message'], $eintrMessage, strlen( $eintrMessage ) ) == 0 ) {
2480                                 continue;
2481                         } else {
2482                                 trigger_error( $error['message'], E_USER_WARNING );
2483                                 $logMsg = $error['message'];
2484                                 break;
2485                         }
2486                 }
2487                 foreach ( $readyPipes as $fd => $pipe ) {
2488                         $block = fread( $pipe, 65536 );
2489                         if ( $block === '' ) {
2490                                 // End of file
2491                                 fclose( $pipes[$fd] );
2492                                 unset( $pipes[$fd] );
2493                                 if ( !$pipes ) {
2494                                         break 2;
2495                                 }
2496                         } elseif ( $block === false ) {
2497                                 // Read error
2498                                 $logMsg = "Error reading from pipe";
2499                                 break 2;
2500                         } elseif ( $fd == 1 ) {
2501                                 // From stdout
2502                                 $outBuffer .= $block;
2503                         } elseif ( $fd == 3 ) {
2504                                 // From log FD
2505                                 $logBuffer .= $block;
2506                                 if ( strpos( $block, "\n" ) !== false ) {
2507                                         $lines = explode( "\n", $logBuffer );
2508                                         $logBuffer = array_pop( $lines );
2509                                         foreach ( $lines as $line ) {
2510                                                 wfDebugLog( 'exec', $line );
2511                                         }
2512                                 }
2513                         }
2514                 }
2515         }
2516
2517         foreach ( $pipes as $pipe ) {
2518                 fclose( $pipe );
2519         }
2520
2521         // Use the status previously collected if possible, since proc_get_status()
2522         // just calls waitpid() which will not return anything useful the second time.
2523         if ( $running ) {
2524                 $status = proc_get_status( $proc );
2525         }
2526
2527         if ( $logMsg !== false ) {
2528                 // Read/select error
2529                 $retval = -1;
2530                 proc_close( $proc );
2531         } elseif ( $status['signaled'] ) {
2532                 $logMsg = "Exited with signal {$status['termsig']}";
2533                 $retval = 128 + $status['termsig'];
2534                 proc_close( $proc );
2535         } else {
2536                 if ( $status['running'] ) {
2537                         $retval = proc_close( $proc );
2538                 } else {
2539                         $retval = $status['exitcode'];
2540                         proc_close( $proc );
2541                 }
2542                 if ( $retval == 127 ) {
2543                         $logMsg = "Possibly missing executable file";
2544                 } elseif ( $retval >= 129 && $retval <= 192 ) {
2545                         $logMsg = "Probably exited with signal " . ( $retval - 128 );
2546                 }
2547         }
2548
2549         if ( $logMsg !== false ) {
2550                 wfDebugLog( 'exec', "$logMsg: $cmd" );
2551         }
2552
2553         return $outBuffer;
2554 }
2555
2556 /**
2557  * Execute a shell command, returning both stdout and stderr. Convenience
2558  * function, as all the arguments to wfShellExec can become unwieldy.
2559  *
2560  * @note This also includes errors from limit.sh, e.g. if $wgMaxShellFileSize is exceeded.
2561  * @param string|string[] $cmd If string, a properly shell-escaped command line,
2562  *   or an array of unescaped arguments, in which case each value will be escaped
2563  *   Example:   [ 'convert', '-font', 'font name' ] would produce "'convert' '-font' 'font name'"
2564  * @param null|mixed &$retval Optional, will receive the program's exit code.
2565  *   (non-zero is usually failure)
2566  * @param array $environ Optional environment variables which should be
2567  *   added to the executed command environment.
2568  * @param array $limits Optional array with limits(filesize, memory, time, walltime)
2569  *   this overwrites the global wgMaxShell* limits.
2570  * @return string Collected stdout and stderr as a string
2571  */
2572 function wfShellExecWithStderr( $cmd, &$retval = null, $environ = [], $limits = [] ) {
2573         return wfShellExec( $cmd, $retval, $environ, $limits,
2574                 [ 'duplicateStderr' => true, 'profileMethod' => wfGetCaller() ] );
2575 }
2576
2577 /**
2578  * Formerly set the locale for locale-sensitive operations
2579  *
2580  * This is now done in Setup.php.
2581  *
2582  * @deprecated since 1.30, no longer needed
2583  * @see $wgShellLocale
2584  */
2585 function wfInitShellLocale() {
2586 }
2587
2588 /**
2589  * Generate a shell-escaped command line string to run a MediaWiki cli script.
2590  * Note that $parameters should be a flat array and an option with an argument
2591  * should consist of two consecutive items in the array (do not use "--option value").
2592  *
2593  * @param string $script MediaWiki cli script path
2594  * @param array $parameters Arguments and options to the script
2595  * @param array $options Associative array of options:
2596  *     'php': The path to the php executable
2597  *     'wrapper': Path to a PHP wrapper to handle the maintenance script
2598  * @return string
2599  */
2600 function wfShellWikiCmd( $script, array $parameters = [], array $options = [] ) {
2601         global $wgPhpCli;
2602         // Give site config file a chance to run the script in a wrapper.
2603         // The caller may likely want to call wfBasename() on $script.
2604         Hooks::run( 'wfShellWikiCmd', [ &$script, &$parameters, &$options ] );
2605         $cmd = isset( $options['php'] ) ? [ $options['php'] ] : [ $wgPhpCli ];
2606         if ( isset( $options['wrapper'] ) ) {
2607                 $cmd[] = $options['wrapper'];
2608         }
2609         $cmd[] = $script;
2610         // Escape each parameter for shell
2611         return wfEscapeShellArg( array_merge( $cmd, $parameters ) );
2612 }
2613
2614 /**
2615  * wfMerge attempts to merge differences between three texts.
2616  * Returns true for a clean merge and false for failure or a conflict.
2617  *
2618  * @param string $old
2619  * @param string $mine
2620  * @param string $yours
2621  * @param string $result
2622  * @return bool
2623  */
2624 function wfMerge( $old, $mine, $yours, &$result ) {
2625         global $wgDiff3;
2626
2627         # This check may also protect against code injection in
2628         # case of broken installations.
2629         MediaWiki\suppressWarnings();
2630         $haveDiff3 = $wgDiff3 && file_exists( $wgDiff3 );
2631         MediaWiki\restoreWarnings();
2632
2633         if ( !$haveDiff3 ) {
2634                 wfDebug( "diff3 not found\n" );
2635                 return false;
2636         }
2637
2638         # Make temporary files
2639         $td = wfTempDir();
2640         $oldtextFile = fopen( $oldtextName = tempnam( $td, 'merge-old-' ), 'w' );
2641         $mytextFile = fopen( $mytextName = tempnam( $td, 'merge-mine-' ), 'w' );
2642         $yourtextFile = fopen( $yourtextName = tempnam( $td, 'merge-your-' ), 'w' );
2643
2644         # NOTE: diff3 issues a warning to stderr if any of the files does not end with
2645         #       a newline character. To avoid this, we normalize the trailing whitespace before
2646         #       creating the diff.
2647
2648         fwrite( $oldtextFile, rtrim( $old ) . "\n" );
2649         fclose( $oldtextFile );
2650         fwrite( $mytextFile, rtrim( $mine ) . "\n" );
2651         fclose( $mytextFile );
2652         fwrite( $yourtextFile, rtrim( $yours ) . "\n" );
2653         fclose( $yourtextFile );
2654
2655         # Check for a conflict
2656         $cmd = wfEscapeShellArg( $wgDiff3, '-a', '--overlap-only', $mytextName,
2657                 $oldtextName, $yourtextName );
2658         $handle = popen( $cmd, 'r' );
2659
2660         if ( fgets( $handle, 1024 ) ) {
2661                 $conflict = true;
2662         } else {
2663                 $conflict = false;
2664         }
2665         pclose( $handle );
2666
2667         # Merge differences
2668         $cmd = wfEscapeShellArg( $wgDiff3, '-a', '-e', '--merge', $mytextName,
2669                 $oldtextName, $yourtextName );
2670         $handle = popen( $cmd, 'r' );
2671         $result = '';
2672         do {
2673                 $data = fread( $handle, 8192 );
2674                 if ( strlen( $data ) == 0 ) {
2675                         break;
2676                 }
2677                 $result .= $data;
2678         } while ( true );
2679         pclose( $handle );
2680         unlink( $mytextName );
2681         unlink( $oldtextName );
2682         unlink( $yourtextName );
2683
2684         if ( $result === '' && $old !== '' && !$conflict ) {
2685                 wfDebug( "Unexpected null result from diff3. Command: $cmd\n" );
2686                 $conflict = true;
2687         }
2688         return !$conflict;
2689 }
2690
2691 /**
2692  * Returns unified plain-text diff of two texts.
2693  * "Useful" for machine processing of diffs.
2694  *
2695  * @deprecated since 1.25, use DiffEngine/UnifiedDiffFormatter directly
2696  *
2697  * @param string $before The text before the changes.
2698  * @param string $after The text after the changes.
2699  * @param string $params Command-line options for the diff command.
2700  * @return string Unified diff of $before and $after
2701  */
2702 function wfDiff( $before, $after, $params = '-u' ) {
2703         if ( $before == $after ) {
2704                 return '';
2705         }
2706
2707         global $wgDiff;
2708         MediaWiki\suppressWarnings();
2709         $haveDiff = $wgDiff && file_exists( $wgDiff );
2710         MediaWiki\restoreWarnings();
2711
2712         # This check may also protect against code injection in
2713         # case of broken installations.
2714         if ( !$haveDiff ) {
2715                 wfDebug( "diff executable not found\n" );
2716                 $diffs = new Diff( explode( "\n", $before ), explode( "\n", $after ) );
2717                 $format = new UnifiedDiffFormatter();
2718                 return $format->format( $diffs );
2719         }
2720
2721         # Make temporary files
2722         $td = wfTempDir();
2723         $oldtextFile = fopen( $oldtextName = tempnam( $td, 'merge-old-' ), 'w' );
2724         $newtextFile = fopen( $newtextName = tempnam( $td, 'merge-your-' ), 'w' );
2725
2726         fwrite( $oldtextFile, $before );
2727         fclose( $oldtextFile );
2728         fwrite( $newtextFile, $after );
2729         fclose( $newtextFile );
2730
2731         // Get the diff of the two files
2732         $cmd = "$wgDiff " . $params . ' ' . wfEscapeShellArg( $oldtextName, $newtextName );
2733
2734         $h = popen( $cmd, 'r' );
2735         if ( !$h ) {
2736                 unlink( $oldtextName );
2737                 unlink( $newtextName );
2738                 throw new Exception( __METHOD__ . '(): popen() failed' );
2739         }
2740
2741         $diff = '';
2742
2743         do {
2744                 $data = fread( $h, 8192 );
2745                 if ( strlen( $data ) == 0 ) {
2746                         break;
2747                 }
2748                 $diff .= $data;
2749         } while ( true );
2750
2751         // Clean up
2752         pclose( $h );
2753         unlink( $oldtextName );
2754         unlink( $newtextName );
2755
2756         // Kill the --- and +++ lines. They're not useful.
2757         $diff_lines = explode( "\n", $diff );
2758         if ( isset( $diff_lines[0] ) && strpos( $diff_lines[0], '---' ) === 0 ) {
2759                 unset( $diff_lines[0] );
2760         }
2761         if ( isset( $diff_lines[1] ) && strpos( $diff_lines[1], '+++' ) === 0 ) {
2762                 unset( $diff_lines[1] );
2763         }
2764
2765         $diff = implode( "\n", $diff_lines );
2766
2767         return $diff;
2768 }
2769
2770 /**
2771  * This function works like "use VERSION" in Perl, the program will die with a
2772  * backtrace if the current version of PHP is less than the version provided
2773  *
2774  * This is useful for extensions which due to their nature are not kept in sync
2775  * with releases, and might depend on other versions of PHP than the main code
2776  *
2777  * Note: PHP might die due to parsing errors in some cases before it ever
2778  *       manages to call this function, such is life
2779  *
2780  * @see perldoc -f use
2781  *
2782  * @param string|int|float $req_ver The version to check, can be a string, an integer, or a float
2783  * @throws MWException
2784  */
2785 function wfUsePHP( $req_ver ) {
2786         $php_ver = PHP_VERSION;
2787
2788         if ( version_compare( $php_ver, (string)$req_ver, '<' ) ) {
2789                 throw new MWException( "PHP $req_ver required--this is only $php_ver" );
2790         }
2791 }
2792
2793 /**
2794  * This function works like "use VERSION" in Perl except it checks the version
2795  * of MediaWiki, the program will die with a backtrace if the current version
2796  * of MediaWiki is less than the version provided.
2797  *
2798  * This is useful for extensions which due to their nature are not kept in sync
2799  * with releases
2800  *
2801  * Note: Due to the behavior of PHP's version_compare() which is used in this
2802  * function, if you want to allow the 'wmf' development versions add a 'c' (or
2803  * any single letter other than 'a', 'b' or 'p') as a post-fix to your
2804  * targeted version number. For example if you wanted to allow any variation
2805  * of 1.22 use `wfUseMW( '1.22c' )`. Using an 'a' or 'b' instead of 'c' will
2806  * not result in the same comparison due to the internal logic of
2807  * version_compare().
2808  *
2809  * @see perldoc -f use
2810  *
2811  * @deprecated since 1.26, use the "requires' property of extension.json
2812  * @param string|int|float $req_ver The version to check, can be a string, an integer, or a float
2813  * @throws MWException
2814  */
2815 function wfUseMW( $req_ver ) {
2816         global $wgVersion;
2817
2818         if ( version_compare( $wgVersion, (string)$req_ver, '<' ) ) {
2819                 throw new MWException( "MediaWiki $req_ver required--this is only $wgVersion" );
2820         }
2821 }
2822
2823 /**
2824  * Return the final portion of a pathname.
2825  * Reimplemented because PHP5's "basename()" is buggy with multibyte text.
2826  * https://bugs.php.net/bug.php?id=33898
2827  *
2828  * PHP's basename() only considers '\' a pathchar on Windows and Netware.
2829  * We'll consider it so always, as we don't want '\s' in our Unix paths either.
2830  *
2831  * @param string $path
2832  * @param string $suffix String to remove if present
2833  * @return string
2834  */
2835 function wfBaseName( $path, $suffix = '' ) {
2836         if ( $suffix == '' ) {
2837                 $encSuffix = '';
2838         } else {
2839                 $encSuffix = '(?:' . preg_quote( $suffix, '#' ) . ')?';
2840         }
2841
2842         $matches = [];
2843         if ( preg_match( "#([^/\\\\]*?){$encSuffix}[/\\\\]*$#", $path, $matches ) ) {
2844                 return $matches[1];
2845         } else {
2846                 return '';
2847         }
2848 }
2849
2850 /**
2851  * Generate a relative path name to the given file.
2852  * May explode on non-matching case-insensitive paths,
2853  * funky symlinks, etc.
2854  *
2855  * @param string $path Absolute destination path including target filename
2856  * @param string $from Absolute source path, directory only
2857  * @return string
2858  */
2859 function wfRelativePath( $path, $from ) {
2860         // Normalize mixed input on Windows...
2861         $path = str_replace( '/', DIRECTORY_SEPARATOR, $path );
2862         $from = str_replace( '/', DIRECTORY_SEPARATOR, $from );
2863
2864         // Trim trailing slashes -- fix for drive root
2865         $path = rtrim( $path, DIRECTORY_SEPARATOR );
2866         $from = rtrim( $from, DIRECTORY_SEPARATOR );
2867
2868         $pieces = explode( DIRECTORY_SEPARATOR, dirname( $path ) );
2869         $against = explode( DIRECTORY_SEPARATOR, $from );
2870
2871         if ( $pieces[0] !== $against[0] ) {
2872                 // Non-matching Windows drive letters?
2873                 // Return a full path.
2874                 return $path;
2875         }
2876
2877         // Trim off common prefix
2878         while ( count( $pieces ) && count( $against )
2879                 && $pieces[0] == $against[0] ) {
2880                 array_shift( $pieces );
2881                 array_shift( $against );
2882         }
2883
2884         // relative dots to bump us to the parent
2885         while ( count( $against ) ) {
2886                 array_unshift( $pieces, '..' );
2887                 array_shift( $against );
2888         }
2889
2890         array_push( $pieces, wfBaseName( $path ) );
2891
2892         return implode( DIRECTORY_SEPARATOR, $pieces );
2893 }
2894
2895 /**
2896  * Convert an arbitrarily-long digit string from one numeric base
2897  * to another, optionally zero-padding to a minimum column width.
2898  *
2899  * Supports base 2 through 36; digit values 10-36 are represented
2900  * as lowercase letters a-z. Input is case-insensitive.
2901  *
2902  * @deprecated since 1.27 Use Wikimedia\base_convert() directly
2903  *
2904  * @param string $input Input number
2905  * @param int $sourceBase Base of the input number
2906  * @param int $destBase Desired base of the output
2907  * @param int $pad Minimum number of digits in the output (pad with zeroes)
2908  * @param bool $lowercase Whether to output in lowercase or uppercase
2909  * @param string $engine Either "gmp", "bcmath", or "php"
2910  * @return string|bool The output number as a string, or false on error
2911  */
2912 function wfBaseConvert( $input, $sourceBase, $destBase, $pad = 1,
2913         $lowercase = true, $engine = 'auto'
2914 ) {
2915         return Wikimedia\base_convert( $input, $sourceBase, $destBase, $pad, $lowercase, $engine );
2916 }
2917
2918 /**
2919  * @deprecated since 1.27, PHP's session generation isn't used with
2920  *  MediaWiki\Session\SessionManager
2921  */
2922 function wfFixSessionID() {
2923         wfDeprecated( __FUNCTION__, '1.27' );
2924 }
2925
2926 /**
2927  * Reset the session id
2928  *
2929  * @deprecated since 1.27, use MediaWiki\Session\SessionManager instead
2930  * @since 1.22
2931  */
2932 function wfResetSessionID() {
2933         wfDeprecated( __FUNCTION__, '1.27' );
2934         $session = SessionManager::getGlobalSession();
2935         $delay = $session->delaySave();
2936
2937         $session->resetId();
2938
2939         // Make sure a session is started, since that's what the old
2940         // wfResetSessionID() did.
2941         if ( session_id() !== $session->getId() ) {
2942                 wfSetupSession( $session->getId() );
2943         }
2944
2945         ScopedCallback::consume( $delay );
2946 }
2947
2948 /**
2949  * Initialise php session
2950  *
2951  * @deprecated since 1.27, use MediaWiki\Session\SessionManager instead.
2952  *  Generally, "using" SessionManager will be calling ->getSessionById() or
2953  *  ::getGlobalSession() (depending on whether you were passing $sessionId
2954  *  here), then calling $session->persist().
2955  * @param bool|string $sessionId
2956  */
2957 function wfSetupSession( $sessionId = false ) {
2958         wfDeprecated( __FUNCTION__, '1.27' );
2959
2960         if ( $sessionId ) {
2961                 session_id( $sessionId );
2962         }
2963
2964         $session = SessionManager::getGlobalSession();
2965         $session->persist();
2966
2967         if ( session_id() !== $session->getId() ) {
2968                 session_id( $session->getId() );
2969         }
2970         MediaWiki\quietCall( 'session_start' );
2971 }
2972
2973 /**
2974  * Get an object from the precompiled serialized directory
2975  *
2976  * @param string $name
2977  * @return mixed The variable on success, false on failure
2978  */
2979 function wfGetPrecompiledData( $name ) {
2980         global $IP;
2981
2982         $file = "$IP/serialized/$name";
2983         if ( file_exists( $file ) ) {
2984                 $blob = file_get_contents( $file );
2985                 if ( $blob ) {
2986                         return unserialize( $blob );
2987                 }
2988         }
2989         return false;
2990 }
2991
2992 /**
2993  * Make a cache key for the local wiki.
2994  *
2995  * @deprecated since 1.30 Call makeKey on a BagOStuff instance
2996  * @param string $args,...
2997  * @return string
2998  */
2999 function wfMemcKey( /*...*/ ) {
3000         return call_user_func_array(
3001                 [ ObjectCache::getLocalClusterInstance(), 'makeKey' ],
3002                 func_get_args()
3003         );
3004 }
3005
3006 /**
3007  * Make a cache key for a foreign DB.
3008  *
3009  * Must match what wfMemcKey() would produce in context of the foreign wiki.
3010  *
3011  * @param string $db
3012  * @param string $prefix
3013  * @param string $args,...
3014  * @return string
3015  */
3016 function wfForeignMemcKey( $db, $prefix /*...*/ ) {
3017         $args = array_slice( func_get_args(), 2 );
3018         $keyspace = $prefix ? "$db-$prefix" : $db;
3019         return call_user_func_array(
3020                 [ ObjectCache::getLocalClusterInstance(), 'makeKeyInternal' ],
3021                 [ $keyspace, $args ]
3022         );
3023 }
3024
3025 /**
3026  * Make a cache key with database-agnostic prefix.
3027  *
3028  * Doesn't have a wiki-specific namespace. Uses a generic 'global' prefix
3029  * instead. Must have a prefix as otherwise keys that use a database name
3030  * in the first segment will clash with wfMemcKey/wfForeignMemcKey.
3031  *
3032  * @deprecated since 1.30 Call makeGlobalKey on a BagOStuff instance
3033  * @since 1.26
3034  * @param string $args,...
3035  * @return string
3036  */
3037 function wfGlobalCacheKey( /*...*/ ) {
3038         return call_user_func_array(
3039                 [ ObjectCache::getLocalClusterInstance(), 'makeGlobalKey' ],
3040                 func_get_args()
3041         );
3042 }
3043
3044 /**
3045  * Get an ASCII string identifying this wiki
3046  * This is used as a prefix in memcached keys
3047  *
3048  * @return string
3049  */
3050 function wfWikiID() {
3051         global $wgDBprefix, $wgDBname;
3052         if ( $wgDBprefix ) {
3053                 return "$wgDBname-$wgDBprefix";
3054         } else {
3055                 return $wgDBname;
3056         }
3057 }
3058
3059 /**
3060  * Split a wiki ID into DB name and table prefix
3061  *
3062  * @param string $wiki
3063  *
3064  * @return array
3065  */
3066 function wfSplitWikiID( $wiki ) {
3067         $bits = explode( '-', $wiki, 2 );
3068         if ( count( $bits ) < 2 ) {
3069                 $bits[] = '';
3070         }
3071         return $bits;
3072 }
3073
3074 /**
3075  * Get a Database object.
3076  *
3077  * @param int $db Index of the connection to get. May be DB_MASTER for the
3078  *            master (for write queries), DB_REPLICA for potentially lagged read
3079  *            queries, or an integer >= 0 for a particular server.
3080  *
3081  * @param string|string[] $groups Query groups. An array of group names that this query
3082  *                belongs to. May contain a single string if the query is only
3083  *                in one group.
3084  *
3085  * @param string|bool $wiki The wiki ID, or false for the current wiki
3086  *
3087  * Note: multiple calls to wfGetDB(DB_REPLICA) during the course of one request
3088  * will always return the same object, unless the underlying connection or load
3089  * balancer is manually destroyed.
3090  *
3091  * Note 2: use $this->getDB() in maintenance scripts that may be invoked by
3092  * updater to ensure that a proper database is being updated.
3093  *
3094  * @todo Replace calls to wfGetDB with calls to LoadBalancer::getConnection()
3095  *       on an injected instance of LoadBalancer.
3096  *
3097  * @return \Wikimedia\Rdbms\Database
3098  */
3099 function wfGetDB( $db, $groups = [], $wiki = false ) {
3100         return wfGetLB( $wiki )->getConnection( $db, $groups, $wiki );
3101 }
3102
3103 /**
3104  * Get a load balancer object.
3105  *
3106  * @deprecated since 1.27, use MediaWikiServices::getDBLoadBalancer()
3107  *              or MediaWikiServices::getDBLoadBalancerFactory() instead.
3108  *
3109  * @param string|bool $wiki Wiki ID, or false for the current wiki
3110  * @return \Wikimedia\Rdbms\LoadBalancer
3111  */
3112 function wfGetLB( $wiki = false ) {
3113         if ( $wiki === false ) {
3114                 return MediaWikiServices::getInstance()->getDBLoadBalancer();
3115         } else {
3116                 $factory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
3117                 return $factory->getMainLB( $wiki );
3118         }
3119 }
3120
3121 /**
3122  * Get the load balancer factory object
3123  *
3124  * @deprecated since 1.27, use MediaWikiServices::getDBLoadBalancerFactory() instead.
3125  *
3126  * @return \Wikimedia\Rdbms\LBFactory
3127  */
3128 function wfGetLBFactory() {
3129         return MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
3130 }
3131
3132 /**
3133  * Find a file.
3134  * Shortcut for RepoGroup::singleton()->findFile()
3135  *
3136  * @param string $title String or Title object
3137  * @param array $options Associative array of options (see RepoGroup::findFile)
3138  * @return File|bool File, or false if the file does not exist
3139  */
3140 function wfFindFile( $title, $options = [] ) {
3141         return RepoGroup::singleton()->findFile( $title, $options );
3142 }
3143
3144 /**
3145  * Get an object referring to a locally registered file.
3146  * Returns a valid placeholder object if the file does not exist.
3147  *
3148  * @param Title|string $title
3149  * @return LocalFile|null A File, or null if passed an invalid Title
3150  */
3151 function wfLocalFile( $title ) {
3152         return RepoGroup::singleton()->getLocalRepo()->newFile( $title );
3153 }
3154
3155 /**
3156  * Should low-performance queries be disabled?
3157  *
3158  * @return bool
3159  * @codeCoverageIgnore
3160  */
3161 function wfQueriesMustScale() {
3162         global $wgMiserMode;
3163         return $wgMiserMode
3164                 || ( SiteStats::pages() > 100000
3165                 && SiteStats::edits() > 1000000
3166                 && SiteStats::users() > 10000 );
3167 }
3168
3169 /**
3170  * Get the path to a specified script file, respecting file
3171  * extensions; this is a wrapper around $wgScriptPath etc.
3172  * except for 'index' and 'load' which use $wgScript/$wgLoadScript
3173  *
3174  * @param string $script Script filename, sans extension
3175  * @return string
3176  */
3177 function wfScript( $script = 'index' ) {
3178         global $wgScriptPath, $wgScript, $wgLoadScript;
3179         if ( $script === 'index' ) {
3180                 return $wgScript;
3181         } elseif ( $script === 'load' ) {
3182                 return $wgLoadScript;
3183         } else {
3184                 return "{$wgScriptPath}/{$script}.php";
3185         }
3186 }
3187
3188 /**
3189  * Get the script URL.
3190  *
3191  * @return string Script URL
3192  */
3193 function wfGetScriptUrl() {
3194         if ( isset( $_SERVER['SCRIPT_NAME'] ) ) {
3195                 /* as it was called, minus the query string.
3196                  *
3197                  * Some sites use Apache rewrite rules to handle subdomains,
3198                  * and have PHP set up in a weird way that causes PHP_SELF
3199                  * to contain the rewritten URL instead of the one that the
3200                  * outside world sees.
3201                  *
3202                  * If in this mode, use SCRIPT_URL instead, which mod_rewrite
3203                  * provides containing the "before" URL.
3204                  */
3205                 return $_SERVER['SCRIPT_NAME'];
3206         } else {
3207                 return $_SERVER['URL'];
3208         }
3209 }
3210
3211 /**
3212  * Convenience function converts boolean values into "true"
3213  * or "false" (string) values
3214  *
3215  * @param bool $value
3216  * @return string
3217  */
3218 function wfBoolToStr( $value ) {
3219         return $value ? 'true' : 'false';
3220 }
3221
3222 /**
3223  * Get a platform-independent path to the null file, e.g. /dev/null
3224  *
3225  * @return string
3226  */
3227 function wfGetNull() {
3228         return wfIsWindows() ? 'NUL' : '/dev/null';
3229 }
3230
3231 /**
3232  * Waits for the replica DBs to catch up to the master position
3233  *
3234  * Use this when updating very large numbers of rows, as in maintenance scripts,
3235  * to avoid causing too much lag. Of course, this is a no-op if there are no replica DBs.
3236  *
3237  * By default this waits on the main DB cluster of the current wiki.
3238  * If $cluster is set to "*" it will wait on all DB clusters, including
3239  * external ones. If the lag being waiting on is caused by the code that
3240  * does this check, it makes since to use $ifWritesSince, particularly if
3241  * cluster is "*", to avoid excess overhead.
3242  *
3243  * Never call this function after a big DB write that is still in a transaction.
3244  * This only makes sense after the possible lag inducing changes were committed.
3245  *
3246  * @param float|null $ifWritesSince Only wait if writes were done since this UNIX timestamp
3247  * @param string|bool $wiki Wiki identifier accepted by wfGetLB
3248  * @param string|bool $cluster Cluster name accepted by LBFactory. Default: false.
3249  * @param int|null $timeout Max wait time. Default: 1 day (cli), ~10 seconds (web)
3250  * @return bool Success (able to connect and no timeouts reached)
3251  * @deprecated since 1.27 Use LBFactory::waitForReplication
3252  */
3253 function wfWaitForSlaves(
3254         $ifWritesSince = null, $wiki = false, $cluster = false, $timeout = null
3255 ) {
3256         if ( $timeout === null ) {
3257                 $timeout = ( PHP_SAPI === 'cli' ) ? 86400 : 10;
3258         }
3259
3260         if ( $cluster === '*' ) {
3261                 $cluster = false;
3262                 $wiki = false;
3263         } elseif ( $wiki === false ) {
3264                 $wiki = wfWikiID();
3265         }
3266
3267         try {
3268                 wfGetLBFactory()->waitForReplication( [
3269                         'wiki' => $wiki,
3270                         'cluster' => $cluster,
3271                         'timeout' => $timeout,
3272                         // B/C: first argument used to be "max seconds of lag"; ignore such values
3273                         'ifWritesSince' => ( $ifWritesSince > 1e9 ) ? $ifWritesSince : null
3274                 ] );
3275         } catch ( DBReplicationWaitError $e ) {
3276                 return false;
3277         }
3278
3279         return true;
3280 }
3281
3282 /**
3283  * Count down from $seconds to zero on the terminal, with a one-second pause
3284  * between showing each number. For use in command-line scripts.
3285  *
3286  * @codeCoverageIgnore
3287  * @param int $seconds
3288  */
3289 function wfCountDown( $seconds ) {
3290         for ( $i = $seconds; $i >= 0; $i-- ) {
3291                 if ( $i != $seconds ) {
3292                         echo str_repeat( "\x08", strlen( $i + 1 ) );
3293                 }
3294                 echo $i;
3295                 flush();
3296                 if ( $i ) {
3297                         sleep( 1 );
3298                 }
3299         }
3300         echo "\n";
3301 }
3302
3303 /**
3304  * Replace all invalid characters with '-'.
3305  * Additional characters can be defined in $wgIllegalFileChars (see T22489).
3306  * By default, $wgIllegalFileChars includes ':', '/', '\'.
3307  *
3308  * @param string $name Filename to process
3309  * @return string
3310  */
3311 function wfStripIllegalFilenameChars( $name ) {
3312         global $wgIllegalFileChars;
3313         $illegalFileChars = $wgIllegalFileChars ? "|[" . $wgIllegalFileChars . "]" : '';
3314         $name = preg_replace(
3315                 "/[^" . Title::legalChars() . "]" . $illegalFileChars . "/",
3316                 '-',
3317                 $name
3318         );
3319         // $wgIllegalFileChars may not include '/' and '\', so we still need to do this
3320         $name = wfBaseName( $name );
3321         return $name;
3322 }
3323
3324 /**
3325  * Set PHP's memory limit to the larger of php.ini or $wgMemoryLimit
3326  *
3327  * @return int Resulting value of the memory limit.
3328  */
3329 function wfMemoryLimit() {
3330         global $wgMemoryLimit;
3331         $memlimit = wfShorthandToInteger( ini_get( 'memory_limit' ) );
3332         if ( $memlimit != -1 ) {
3333                 $conflimit = wfShorthandToInteger( $wgMemoryLimit );
3334                 if ( $conflimit == -1 ) {
3335                         wfDebug( "Removing PHP's memory limit\n" );
3336                         MediaWiki\suppressWarnings();
3337                         ini_set( 'memory_limit', $conflimit );
3338                         MediaWiki\restoreWarnings();
3339                         return $conflimit;
3340                 } elseif ( $conflimit > $memlimit ) {
3341                         wfDebug( "Raising PHP's memory limit to $conflimit bytes\n" );
3342                         MediaWiki\suppressWarnings();
3343                         ini_set( 'memory_limit', $conflimit );
3344                         MediaWiki\restoreWarnings();
3345                         return $conflimit;
3346                 }
3347         }
3348         return $memlimit;
3349 }
3350
3351 /**
3352  * Set PHP's time limit to the larger of php.ini or $wgTransactionalTimeLimit
3353  *
3354  * @return int Prior time limit
3355  * @since 1.26
3356  */
3357 function wfTransactionalTimeLimit() {
3358         global $wgTransactionalTimeLimit;
3359
3360         $timeLimit = ini_get( 'max_execution_time' );
3361         // Note that CLI scripts use 0
3362         if ( $timeLimit > 0 && $wgTransactionalTimeLimit > $timeLimit ) {
3363                 set_time_limit( $wgTransactionalTimeLimit );
3364         }
3365
3366         ignore_user_abort( true ); // ignore client disconnects
3367
3368         return $timeLimit;
3369 }
3370
3371 /**
3372  * Converts shorthand byte notation to integer form
3373  *
3374  * @param string $string
3375  * @param int $default Returned if $string is empty
3376  * @return int
3377  */
3378 function wfShorthandToInteger( $string = '', $default = -1 ) {
3379         $string = trim( $string );
3380         if ( $string === '' ) {
3381                 return $default;
3382         }
3383         $last = $string[strlen( $string ) - 1];
3384         $val = intval( $string );
3385         switch ( $last ) {
3386                 case 'g':
3387                 case 'G':
3388                         $val *= 1024;
3389                         // break intentionally missing
3390                 case 'm':
3391                 case 'M':
3392                         $val *= 1024;
3393                         // break intentionally missing
3394                 case 'k':
3395                 case 'K':
3396                         $val *= 1024;
3397         }
3398
3399         return $val;
3400 }
3401
3402 /**
3403  * Get the normalised IETF language tag
3404  * See unit test for examples.
3405  *
3406  * @param string $code The language code.
3407  * @return string The language code which complying with BCP 47 standards.
3408  */
3409 function wfBCP47( $code ) {
3410         $codeSegment = explode( '-', $code );
3411         $codeBCP = [];
3412         foreach ( $codeSegment as $segNo => $seg ) {
3413                 // when previous segment is x, it is a private segment and should be lc
3414                 if ( $segNo > 0 && strtolower( $codeSegment[( $segNo - 1 )] ) == 'x' ) {
3415                         $codeBCP[$segNo] = strtolower( $seg );
3416                 // ISO 3166 country code
3417                 } elseif ( ( strlen( $seg ) == 2 ) && ( $segNo > 0 ) ) {
3418                         $codeBCP[$segNo] = strtoupper( $seg );
3419                 // ISO 15924 script code
3420                 } elseif ( ( strlen( $seg ) == 4 ) && ( $segNo > 0 ) ) {
3421                         $codeBCP[$segNo] = ucfirst( strtolower( $seg ) );
3422                 // Use lowercase for other cases
3423                 } else {
3424                         $codeBCP[$segNo] = strtolower( $seg );
3425                 }
3426         }
3427         $langCode = implode( '-', $codeBCP );
3428         return $langCode;
3429 }
3430
3431 /**
3432  * Get a specific cache object.
3433  *
3434  * @param int|string $cacheType A CACHE_* constants, or other key in $wgObjectCaches
3435  * @return BagOStuff
3436  */
3437 function wfGetCache( $cacheType ) {
3438         return ObjectCache::getInstance( $cacheType );
3439 }
3440
3441 /**
3442  * Get the main cache object
3443  *
3444  * @return BagOStuff
3445  */
3446 function wfGetMainCache() {
3447         global $wgMainCacheType;
3448         return ObjectCache::getInstance( $wgMainCacheType );
3449 }
3450
3451 /**
3452  * Get the cache object used by the message cache
3453  *
3454  * @return BagOStuff
3455  */
3456 function wfGetMessageCacheStorage() {
3457         global $wgMessageCacheType;
3458         return ObjectCache::getInstance( $wgMessageCacheType );
3459 }
3460
3461 /**
3462  * Get the cache object used by the parser cache
3463  *
3464  * @deprecated since 1.30, use MediaWikiServices::getParserCache()->getCacheStorage()
3465  * @return BagOStuff
3466  */
3467 function wfGetParserCacheStorage() {
3468         global $wgParserCacheType;
3469         return ObjectCache::getInstance( $wgParserCacheType );
3470 }
3471
3472 /**
3473  * Call hook functions defined in $wgHooks
3474  *
3475  * @param string $event Event name
3476  * @param array $args Parameters passed to hook functions
3477  * @param string|null $deprecatedVersion Optionally mark hook as deprecated with version number
3478  *
3479  * @return bool True if no handler aborted the hook
3480  * @deprecated since 1.25 - use Hooks::run
3481  */
3482 function wfRunHooks( $event, array $args = [], $deprecatedVersion = null ) {
3483         return Hooks::run( $event, $args, $deprecatedVersion );
3484 }
3485
3486 /**
3487  * Wrapper around php's unpack.
3488  *
3489  * @param string $format The format string (See php's docs)
3490  * @param string $data A binary string of binary data
3491  * @param int|bool $length The minimum length of $data or false. This is to
3492  *      prevent reading beyond the end of $data. false to disable the check.
3493  *
3494  * Also be careful when using this function to read unsigned 32 bit integer
3495  * because php might make it negative.
3496  *
3497  * @throws MWException If $data not long enough, or if unpack fails
3498  * @return array Associative array of the extracted data
3499  */
3500 function wfUnpack( $format, $data, $length = false ) {
3501         if ( $length !== false ) {
3502                 $realLen = strlen( $data );
3503                 if ( $realLen < $length ) {
3504                         throw new MWException( "Tried to use wfUnpack on a "
3505                                 . "string of length $realLen, but needed one "
3506                                 . "of at least length $length."
3507                         );
3508                 }
3509         }
3510
3511         MediaWiki\suppressWarnings();
3512         $result = unpack( $format, $data );
3513         MediaWiki\restoreWarnings();
3514
3515         if ( $result === false ) {
3516                 // If it cannot extract the packed data.
3517                 throw new MWException( "unpack could not unpack binary data" );
3518         }
3519         return $result;
3520 }
3521
3522 /**
3523  * Determine if an image exists on the 'bad image list'.
3524  *
3525  * The format of MediaWiki:Bad_image_list is as follows:
3526  *    * Only list items (lines starting with "*") are considered
3527  *    * The first link on a line must be a link to a bad image
3528  *    * Any subsequent links on the same line are considered to be exceptions,
3529  *      i.e. articles where the image may occur inline.
3530  *
3531  * @param string $name The image name to check
3532  * @param Title|bool $contextTitle The page on which the image occurs, if known
3533  * @param string $blacklist Wikitext of a file blacklist
3534  * @return bool
3535  */
3536 function wfIsBadImage( $name, $contextTitle = false, $blacklist = null ) {
3537         # Handle redirects; callers almost always hit wfFindFile() anyway,
3538         # so just use that method because it has a fast process cache.
3539         $file = wfFindFile( $name ); // get the final name
3540         $name = $file ? $file->getTitle()->getDBkey() : $name;
3541
3542         # Run the extension hook
3543         $bad = false;
3544         if ( !Hooks::run( 'BadImage', [ $name, &$bad ] ) ) {
3545                 return (bool)$bad;
3546         }
3547
3548         $cache = ObjectCache::getLocalServerInstance( 'hash' );
3549         $key = $cache->makeKey(
3550                 'bad-image-list', ( $blacklist === null ) ? 'default' : md5( $blacklist )
3551         );
3552         $badImages = $cache->get( $key );
3553
3554         if ( $badImages === false ) { // cache miss
3555                 if ( $blacklist === null ) {
3556                         $blacklist = wfMessage( 'bad_image_list' )->inContentLanguage()->plain(); // site list
3557                 }
3558                 # Build the list now
3559                 $badImages = [];
3560                 $lines = explode( "\n", $blacklist );
3561                 foreach ( $lines as $line ) {
3562                         # List items only
3563                         if ( substr( $line, 0, 1 ) !== '*' ) {
3564                                 continue;
3565                         }
3566
3567                         # Find all links
3568                         $m = [];
3569                         if ( !preg_match_all( '/\[\[:?(.*?)\]\]/', $line, $m ) ) {
3570                                 continue;
3571                         }
3572
3573                         $exceptions = [];
3574                         $imageDBkey = false;
3575                         foreach ( $m[1] as $i => $titleText ) {
3576                                 $title = Title::newFromText( $titleText );
3577                                 if ( !is_null( $title ) ) {
3578                                         if ( $i == 0 ) {
3579                                                 $imageDBkey = $title->getDBkey();
3580                                         } else {
3581                                                 $exceptions[$title->getPrefixedDBkey()] = true;
3582                                         }
3583                                 }
3584                         }
3585
3586                         if ( $imageDBkey !== false ) {
3587                                 $badImages[$imageDBkey] = $exceptions;
3588                         }
3589                 }
3590                 $cache->set( $key, $badImages, 60 );
3591         }
3592
3593         $contextKey = $contextTitle ? $contextTitle->getPrefixedDBkey() : false;
3594         $bad = isset( $badImages[$name] ) && !isset( $badImages[$name][$contextKey] );
3595
3596         return $bad;
3597 }
3598
3599 /**
3600  * Determine whether the client at a given source IP is likely to be able to
3601  * access the wiki via HTTPS.
3602  *
3603  * @param string $ip The IPv4/6 address in the normal human-readable form
3604  * @return bool
3605  */
3606 function wfCanIPUseHTTPS( $ip ) {
3607         $canDo = true;
3608         Hooks::run( 'CanIPUseHTTPS', [ $ip, &$canDo ] );
3609         return !!$canDo;
3610 }
3611
3612 /**
3613  * Determine input string is represents as infinity
3614  *
3615  * @param string $str The string to determine
3616  * @return bool
3617  * @since 1.25
3618  */
3619 function wfIsInfinity( $str ) {
3620         // These are hardcoded elsewhere in MediaWiki (e.g. mediawiki.special.block.js).
3621         $infinityValues = [ 'infinite', 'indefinite', 'infinity', 'never' ];
3622         return in_array( $str, $infinityValues );
3623 }
3624
3625 /**
3626  * Returns true if these thumbnail parameters match one that MediaWiki
3627  * requests from file description pages and/or parser output.
3628  *
3629  * $params is considered non-standard if they involve a non-standard
3630  * width or any non-default parameters aside from width and page number.
3631  * The number of possible files with standard parameters is far less than
3632  * that of all combinations; rate-limiting for them can thus be more generious.
3633  *
3634  * @param File $file
3635  * @param array $params
3636  * @return bool
3637  * @since 1.24 Moved from thumb.php to GlobalFunctions in 1.25
3638  */
3639 function wfThumbIsStandard( File $file, array $params ) {
3640         global $wgThumbLimits, $wgImageLimits, $wgResponsiveImages;
3641
3642         $multipliers = [ 1 ];
3643         if ( $wgResponsiveImages ) {
3644                 // These available sizes are hardcoded currently elsewhere in MediaWiki.
3645                 // @see Linker::processResponsiveImages
3646                 $multipliers[] = 1.5;
3647                 $multipliers[] = 2;
3648         }
3649
3650         $handler = $file->getHandler();
3651         if ( !$handler || !isset( $params['width'] ) ) {
3652                 return false;
3653         }
3654
3655         $basicParams = [];
3656         if ( isset( $params['page'] ) ) {
3657                 $basicParams['page'] = $params['page'];
3658         }
3659
3660         $thumbLimits = [];
3661         $imageLimits = [];
3662         // Expand limits to account for multipliers
3663         foreach ( $multipliers as $multiplier ) {
3664                 $thumbLimits = array_merge( $thumbLimits, array_map(
3665                         function ( $width ) use ( $multiplier ) {
3666                                 return round( $width * $multiplier );
3667                         }, $wgThumbLimits )
3668                 );
3669                 $imageLimits = array_merge( $imageLimits, array_map(
3670                         function ( $pair ) use ( $multiplier ) {
3671                                 return [
3672                                         round( $pair[0] * $multiplier ),
3673                                         round( $pair[1] * $multiplier ),
3674                                 ];
3675                         }, $wgImageLimits )
3676                 );
3677         }
3678
3679         // Check if the width matches one of $wgThumbLimits
3680         if ( in_array( $params['width'], $thumbLimits ) ) {
3681                 $normalParams = $basicParams + [ 'width' => $params['width'] ];
3682                 // Append any default values to the map (e.g. "lossy", "lossless", ...)
3683                 $handler->normaliseParams( $file, $normalParams );
3684         } else {
3685                 // If not, then check if the width matchs one of $wgImageLimits
3686                 $match = false;
3687                 foreach ( $imageLimits as $pair ) {
3688                         $normalParams = $basicParams + [ 'width' => $pair[0], 'height' => $pair[1] ];
3689                         // Decide whether the thumbnail should be scaled on width or height.
3690                         // Also append any default values to the map (e.g. "lossy", "lossless", ...)
3691                         $handler->normaliseParams( $file, $normalParams );
3692                         // Check if this standard thumbnail size maps to the given width
3693                         if ( $normalParams['width'] == $params['width'] ) {
3694                                 $match = true;
3695                                 break;
3696                         }
3697                 }
3698                 if ( !$match ) {
3699                         return false; // not standard for description pages
3700                 }
3701         }
3702
3703         // Check that the given values for non-page, non-width, params are just defaults
3704         foreach ( $params as $key => $value ) {
3705                 if ( !isset( $normalParams[$key] ) || $normalParams[$key] != $value ) {
3706                         return false;
3707                 }
3708         }
3709
3710         return true;
3711 }
3712
3713 /**
3714  * Merges two (possibly) 2 dimensional arrays into the target array ($baseArray).
3715  *
3716  * Values that exist in both values will be combined with += (all values of the array
3717  * of $newValues will be added to the values of the array of $baseArray, while values,
3718  * that exists in both, the value of $baseArray will be used).
3719  *
3720  * @param array $baseArray The array where you want to add the values of $newValues to
3721  * @param array $newValues An array with new values
3722  * @return array The combined array
3723  * @since 1.26
3724  */
3725 function wfArrayPlus2d( array $baseArray, array $newValues ) {
3726         // First merge items that are in both arrays
3727         foreach ( $baseArray as $name => &$groupVal ) {
3728                 if ( isset( $newValues[$name] ) ) {
3729                         $groupVal += $newValues[$name];
3730                 }
3731         }
3732         // Now add items that didn't exist yet
3733         $baseArray += $newValues;
3734
3735         return $baseArray;
3736 }