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
  33 // Hide compatibility functions from Doxygen
  34 /// @cond
  35 /**
  36  * Compatibility functions
  37  *
  38  * We support PHP 5.5.9 and up.
  39  * Re-implementations of newer functions or functions in non-standard
  40  * PHP extensions may be included here.
  41  */
  42
  43 // hash_equals function only exists in PHP >= 5.6.0
  44 // https://secure.php.net/hash_equals
  45 if ( !function_exists( 'hash_equals' ) ) {
  46         /**
  47          * Check whether a user-provided string is equal to a fixed-length secret string
  48          * without revealing bytes of the secret string through timing differences.
  49          *
  50          * The usual way to compare strings (PHP's === operator or the underlying memcmp()
  51          * function in C) is to compare corresponding bytes and stop at the first difference,
  52          * which would take longer for a partial match than for a complete mismatch. This
  53          * is not secure when one of the strings (e.g. an HMAC or token) must remain secret
  54          * and the other may come from an attacker. Statistical analysis of timing measurements
  55          * over many requests may allow the attacker to guess the string's bytes one at a time
  56          * (and check his guesses) even if the timing differences are extremely small.
  57          *
  58          * When making such a security-sensitive comparison, it is essential that the sequence
  59          * in which instructions are executed and memory locations are accessed not depend on
  60          * the secret string's value. HOWEVER, for simplicity, we do not attempt to minimize
  61          * the inevitable leakage of the string's length. That is generally known anyway as
  62          * a chararacteristic of the hash function used to compute the secret value.
  63          *
  64          * Longer explanation: http://www.emerose.com/timing-attacks-explained
  65          *
  66          * @codeCoverageIgnore
  67          * @param string $known_string Fixed-length secret string to compare against
  68          * @param string $user_string User-provided string
  69          * @return bool True if the strings are the same, false otherwise
  70          */
  71         function hash_equals( $known_string, $user_string ) {
  72                 // Strict type checking as in PHP's native implementation
  73                 if ( !is_string( $known_string ) ) {
  74                         trigger_error( 'hash_equals(): Expected known_string to be a string, ' .
  75                                 gettype( $known_string ) . ' given', E_USER_WARNING );
  76
  77                         return false;
  78                 }
  79
  80                 if ( !is_string( $user_string ) ) {
  81                         trigger_error( 'hash_equals(): Expected user_string to be a string, ' .
  82                                 gettype( $user_string ) . ' given', E_USER_WARNING );
  83
  84                         return false;
  85                 }
  86
  87                 $known_string_len = strlen( $known_string );
  88                 if ( $known_string_len !== strlen( $user_string ) ) {
  89                         return false;
  90                 }
  91
  92                 $result = 0;
  93                 for ( $i = 0; $i < $known_string_len; $i++ ) {
  94                         $result |= ord( $known_string[$i] ) ^ ord( $user_string[$i] );
  95                 }
  96
  97                 return ( $result === 0 );
  98         }
  99 }
 100 /// @endcond
 101
 102 /**
 103  * Load an extension
 104  *
 105  * This queues an extension to be loaded through
 106  * the ExtensionRegistry system.
 107  *
 108  * @param string $ext Name of the extension to load
 109  * @param string|null $path Absolute path of where to find the extension.json file
 110  * @since 1.25
 111  */
 112 function wfLoadExtension( $ext, $path = null ) {
 113         if ( !$path ) {
 114                 global $wgExtensionDirectory;
 115                 $path = "$wgExtensionDirectory/$ext/extension.json";
 116         }
 117         ExtensionRegistry::getInstance()->queue( $path );
 118 }
 119
 120 /**
 121  * Load multiple extensions at once
 122  *
 123  * Same as wfLoadExtension, but more efficient if you
 124  * are loading multiple extensions.
 125  *
 126  * If you want to specify custom paths, you should interact with
 127  * ExtensionRegistry directly.
 128  *
 129  * @see wfLoadExtension
 130  * @param string[] $exts Array of extension names to load
 131  * @since 1.25
 132  */
 133 function wfLoadExtensions( array $exts ) {
 134         global $wgExtensionDirectory;
 135         $registry = ExtensionRegistry::getInstance();
 136         foreach ( $exts as $ext ) {
 137                 $registry->queue( "$wgExtensionDirectory/$ext/extension.json" );
 138         }
 139 }
 140
 141 /**
 142  * Load a skin
 143  *
 144  * @see wfLoadExtension
 145  * @param string $skin Name of the extension to load
 146  * @param string|null $path Absolute path of where to find the skin.json file
 147  * @since 1.25
 148  */
 149 function wfLoadSkin( $skin, $path = null ) {
 150         if ( !$path ) {
 151                 global $wgStyleDirectory;
 152                 $path = "$wgStyleDirectory/$skin/skin.json";
 153         }
 154         ExtensionRegistry::getInstance()->queue( $path );
 155 }
 156
 157 /**
 158  * Load multiple skins at once
 159  *
 160  * @see wfLoadExtensions
 161  * @param string[] $skins Array of extension names to load
 162  * @since 1.25
 163  */
 164 function wfLoadSkins( array $skins ) {
 165         global $wgStyleDirectory;
 166         $registry = ExtensionRegistry::getInstance();
 167         foreach ( $skins as $skin ) {
 168                 $registry->queue( "$wgStyleDirectory/$skin/skin.json" );
 169         }
 170 }
 171
 172 /**
 173  * Like array_diff( $a, $b ) except that it works with two-dimensional arrays.
 174  * @param array $a
 175  * @param array $b
 176  * @return array
 177  */
 178 function wfArrayDiff2( $a, $b ) {
 179         return array_udiff( $a, $b, 'wfArrayDiff2_cmp' );
 180 }
 181
 182 /**
 183  * @param array|string $a
 184  * @param array|string $b
 185  * @return int
 186  */
 187 function wfArrayDiff2_cmp( $a, $b ) {
 188         if ( is_string( $a ) && is_string( $b ) ) {
 189                 return strcmp( $a, $b );
 190         } elseif ( count( $a ) !== count( $b ) ) {
 191                 return count( $a ) < count( $b ) ? -1 : 1;
 192         } else {
 193                 reset( $a );
 194                 reset( $b );
 195                 while ( ( list( , $valueA ) = each( $a ) ) && ( list( , $valueB ) = each( $b ) ) ) {
 196                         $cmp = strcmp( $valueA, $valueB );
 197                         if ( $cmp !== 0 ) {
 198                                 return $cmp;
 199                         }
 200                 }
 201                 return 0;
 202         }
 203 }
 204
 205 /**
 206  * Appends to second array if $value differs from that in $default
 207  *
 208  * @param string|int $key
 209  * @param mixed $value
 210  * @param mixed $default
 211  * @param array $changed Array to alter
 212  * @throws MWException
 213  */
 214 function wfAppendToArrayIfNotDefault( $key, $value, $default, &$changed ) {
 215         if ( is_null( $changed ) ) {
 216                 throw new MWException( 'GlobalFunctions::wfAppendToArrayIfNotDefault got null' );
 217         }
 218         if ( $default[$key] !== $value ) {
 219                 $changed[$key] = $value;
 220         }
 221 }
 222
 223 /**
 224  * Merge arrays in the style of getUserPermissionsErrors, with duplicate removal
 225  * e.g.
 226  *     wfMergeErrorArrays(
 227  *       [ [ 'x' ] ],
 228  *       [ [ 'x', '2' ] ],
 229  *       [ [ 'x' ] ],
 230  *       [ [ 'y' ] ]
 231  *     );
 232  * returns:
 233  *     [
 234  *       [ 'x', '2' ],
 235  *       [ 'x' ],
 236  *       [ 'y' ]
 237  *     ]
 238  *
 239  * @param array $array1,...
 240  * @return array
 241  */
 242 function wfMergeErrorArrays( /*...*/ ) {
 243         $args = func_get_args();
 244         $out = [];
 245         foreach ( $args as $errors ) {
 246                 foreach ( $errors as $params ) {
 247                         $originalParams = $params;
 248                         if ( $params[0] instanceof MessageSpecifier ) {
 249                                 $msg = $params[0];
 250                                 $params = array_merge( [ $msg->getKey() ], $msg->getParams() );
 251                         }
 252                         # @todo FIXME: Sometimes get nested arrays for $params,
 253                         # which leads to E_NOTICEs
 254                         $spec = implode( "\t", $params );
 255                         $out[$spec] = $originalParams;
 256                 }
 257         }
 258         return array_values( $out );
 259 }
 260
 261 /**
 262  * Insert array into another array after the specified *KEY*
 263  *
 264  * @param array $array The array.
 265  * @param array $insert The array to insert.
 266  * @param mixed $after The key to insert after
 267  * @return array
 268  */
 269 function wfArrayInsertAfter( array $array, array $insert, $after ) {
 270         // Find the offset of the element to insert after.
 271         $keys = array_keys( $array );
 272         $offsetByKey = array_flip( $keys );
 273
 274         $offset = $offsetByKey[$after];
 275
 276         // Insert at the specified offset
 277         $before = array_slice( $array, 0, $offset + 1, true );
 278         $after = array_slice( $array, $offset + 1, count( $array ) - $offset, true );
 279
 280         $output = $before + $insert + $after;
 281
 282         return $output;
 283 }
 284
 285 /**
 286  * Recursively converts the parameter (an object) to an array with the same data
 287  *
 288  * @param object|array $objOrArray
 289  * @param bool $recursive
 290  * @return array
 291  */
 292 function wfObjectToArray( $objOrArray, $recursive = true ) {
 293         $array = [];
 294         if ( is_object( $objOrArray ) ) {
 295                 $objOrArray = get_object_vars( $objOrArray );
 296         }
 297         foreach ( $objOrArray as $key => $value ) {
 298                 if ( $recursive && ( is_object( $value ) || is_array( $value ) ) ) {
 299                         $value = wfObjectToArray( $value );
 300                 }
 301
 302                 $array[$key] = $value;
 303         }
 304
 305         return $array;
 306 }
 307
 308 /**
 309  * Get a random decimal value between 0 and 1, in a way
 310  * not likely to give duplicate values for any realistic
 311  * number of articles.
 312  *
 313  * @note This is designed for use in relation to Special:RandomPage
 314  *       and the page_random database field.
 315  *
 316  * @return string
 317  */
 318 function wfRandom() {
 319         // The maximum random value is "only" 2^31-1, so get two random
 320         // values to reduce the chance of dupes
 321         $max = mt_getrandmax() + 1;
 322         $rand = number_format( ( mt_rand() * $max + mt_rand() ) / $max / $max, 12, '.', '' );
 323         return $rand;
 324 }
 325
 326 /**
 327  * Get a random string containing a number of pseudo-random hex characters.
 328  *
 329  * @note This is not secure, if you are trying to generate some sort
 330  *       of token please use MWCryptRand instead.
 331  *
 332  * @param int $length The length of the string to generate
 333  * @return string
 334  * @since 1.20
 335  */
 336 function wfRandomString( $length = 32 ) {
 337         $str = '';
 338         for ( $n = 0; $n < $length; $n += 7 ) {
 339                 $str .= sprintf( '%07x', mt_rand() & 0xfffffff );
 340         }
 341         return substr( $str, 0, $length );
 342 }
 343
 344 /**
 345  * We want some things to be included as literal characters in our title URLs
 346  * for prettiness, which urlencode encodes by default.  According to RFC 1738,
 347  * all of the following should be safe:
 348  *
 349  * ;:@&=$-_.+!*'(),
 350  *
 351  * RFC 1738 says ~ is unsafe, however RFC 3986 considers it an unreserved
 352  * character which should not be encoded. More importantly, google chrome
 353  * always converts %7E back to ~, and converting it in this function can
 354  * cause a redirect loop (T105265).
 355  *
 356  * But + is not safe because it's used to indicate a space; &= are only safe in
 357  * paths and not in queries (and we don't distinguish here); ' seems kind of
 358  * scary; and urlencode() doesn't touch -_. to begin with.  Plus, although /
 359  * is reserved, we don't care.  So the list we unescape is:
 360  *
 361  * ;:@$!*(),/~
 362  *
 363  * However, IIS7 redirects fail when the url contains a colon (see T24709),
 364  * so no fancy : for IIS7.
 365  *
 366  * %2F in the page titles seems to fatally break for some reason.
 367  *
 368  * @param string $s
 369  * @return string
 370  */
 371 function wfUrlencode( $s ) {
 372         static $needle;
 373
 374         if ( is_null( $s ) ) {
 375                 $needle = null;
 376                 return '';
 377         }
 378
 379         if ( is_null( $needle ) ) {
 380                 $needle = [ '%3B', '%40', '%24', '%21', '%2A', '%28', '%29', '%2C', '%2F', '%7E' ];
 381                 if ( !isset( $_SERVER['SERVER_SOFTWARE'] ) ||
 382                         ( strpos( $_SERVER['SERVER_SOFTWARE'], 'Microsoft-IIS/7' ) === false )
 383                 ) {
 384                         $needle[] = '%3A';
 385                 }
 386         }
 387
 388         $s = urlencode( $s );
 389         $s = str_ireplace(
 390                 $needle,
 391                 [ ';', '@', '$', '!', '*', '(', ')', ',', '/', '~', ':' ],
 392                 $s
 393         );
 394
 395         return $s;
 396 }
 397
 398 /**
 399  * This function takes one or two arrays as input, and returns a CGI-style string, e.g.
 400  * "days=7&limit=100". Options in the first array override options in the second.
 401  * Options set to null or false will not be output.
 402  *
 403  * @param array $array1 ( String|Array )
 404  * @param array|null $array2 ( String|Array )
 405  * @param string $prefix
 406  * @return string
 407  */
 408 function wfArrayToCgi( $array1, $array2 = null, $prefix = '' ) {
 409         if ( !is_null( $array2 ) ) {
 410                 $array1 = $array1 + $array2;
 411         }
 412
 413         $cgi = '';
 414         foreach ( $array1 as $key => $value ) {
 415                 if ( !is_null( $value ) && $value !== false ) {
 416                         if ( $cgi != '' ) {
 417                                 $cgi .= '&';
 418                         }
 419                         if ( $prefix !== '' ) {
 420                                 $key = $prefix . "[$key]";
 421                         }
 422                         if ( is_array( $value ) ) {
 423                                 $firstTime = true;
 424                                 foreach ( $value as $k => $v ) {
 425                                         $cgi .= $firstTime ? '' : '&';
 426                                         if ( is_array( $v ) ) {
 427                                                 $cgi .= wfArrayToCgi( $v, null, $key . "[$k]" );
 428                                         } else {
 429                                                 $cgi .= urlencode( $key . "[$k]" ) . '=' . urlencode( $v );
 430                                         }
 431                                         $firstTime = false;
 432                                 }
 433                         } else {
 434                                 if ( is_object( $value ) ) {
 435                                         $value = $value->__toString();
 436                                 }
 437                                 $cgi .= urlencode( $key ) . '=' . urlencode( $value );
 438                         }
 439                 }
 440         }
 441         return $cgi;
 442 }
 443
 444 /**
 445  * This is the logical opposite of wfArrayToCgi(): it accepts a query string as
 446  * its argument and returns the same string in array form.  This allows compatibility
 447  * with legacy functions that accept raw query strings instead of nice
 448  * arrays.  Of course, keys and values are urldecode()d.
 449  *
 450  * @param string $query Query string
 451  * @return string[] Array version of input
 452  */
 453 function wfCgiToArray( $query ) {
 454         if ( isset( $query[0] ) && $query[0] == '?' ) {
 455                 $query = substr( $query, 1 );
 456         }
 457         $bits = explode( '&', $query );
 458         $ret = [];
 459         foreach ( $bits as $bit ) {
 460                 if ( $bit === '' ) {
 461                         continue;
 462                 }
 463                 if ( strpos( $bit, '=' ) === false ) {
 464                         // Pieces like &qwerty become 'qwerty' => '' (at least this is what php does)
 465                         $key = $bit;
 466                         $value = '';
 467                 } else {
 468                         list( $key, $value ) = explode( '=', $bit );
 469                 }
 470                 $key = urldecode( $key );
 471                 $value = urldecode( $value );
 472                 if ( strpos( $key, '[' ) !== false ) {
 473                         $keys = array_reverse( explode( '[', $key ) );
 474                         $key = array_pop( $keys );
 475                         $temp = $value;
 476                         foreach ( $keys as $k ) {
 477                                 $k = substr( $k, 0, -1 );
 478                                 $temp = [ $k => $temp ];
 479                         }
 480                         if ( isset( $ret[$key] ) ) {
 481                                 $ret[$key] = array_merge( $ret[$key], $temp );
 482                         } else {
 483                                 $ret[$key] = $temp;
 484                         }
 485                 } else {
 486                         $ret[$key] = $value;
 487                 }
 488         }
 489         return $ret;
 490 }
 491
 492 /**
 493  * Append a query string to an existing URL, which may or may not already
 494  * have query string parameters already. If so, they will be combined.
 495  *
 496  * @param string $url
 497  * @param string|string[] $query String or associative array
 498  * @return string
 499  */
 500 function wfAppendQuery( $url, $query ) {
 501         if ( is_array( $query ) ) {
 502                 $query = wfArrayToCgi( $query );
 503         }
 504         if ( $query != '' ) {
 505                 // Remove the fragment, if there is one
 506                 $fragment = false;
 507                 $hashPos = strpos( $url, '#' );
 508                 if ( $hashPos !== false ) {
 509                         $fragment = substr( $url, $hashPos );
 510                         $url = substr( $url, 0, $hashPos );
 511                 }
 512
 513                 // Add parameter
 514                 if ( false === strpos( $url, '?' ) ) {
 515                         $url .= '?';
 516                 } else {
 517                         $url .= '&';
 518                 }
 519                 $url .= $query;
 520
 521                 // Put the fragment back
 522                 if ( $fragment !== false ) {
 523                         $url .= $fragment;
 524                 }
 525         }
 526         return $url;
 527 }
 528
 529 /**
 530  * Expand a potentially local URL to a fully-qualified URL. Assumes $wgServer
 531  * is correct.
 532  *
 533  * The meaning of the PROTO_* constants is as follows:
 534  * PROTO_HTTP: Output a URL starting with http://
 535  * PROTO_HTTPS: Output a URL starting with https://
 536  * PROTO_RELATIVE: Output a URL starting with // (protocol-relative URL)
 537  * PROTO_CURRENT: Output a URL starting with either http:// or https:// , depending
 538  *    on which protocol was used for the current incoming request
 539  * PROTO_CANONICAL: For URLs without a domain, like /w/index.php , use $wgCanonicalServer.
 540  *    For protocol-relative URLs, use the protocol of $wgCanonicalServer
 541  * PROTO_INTERNAL: Like PROTO_CANONICAL, but uses $wgInternalServer instead of $wgCanonicalServer
 542  *
 543  * @todo this won't work with current-path-relative URLs
 544  * like "subdir/foo.html", etc.
 545  *
 546  * @param string $url Either fully-qualified or a local path + query
 547  * @param string $defaultProto One of the PROTO_* constants. Determines the
 548  *    protocol to use if $url or $wgServer is protocol-relative
 549  * @return string|false Fully-qualified URL, current-path-relative URL or false if
 550  *    no valid URL can be constructed
 551  */
 552 function wfExpandUrl( $url, $defaultProto = PROTO_CURRENT ) {
 553         global $wgServer, $wgCanonicalServer, $wgInternalServer, $wgRequest,
 554                 $wgHttpsPort;
 555         if ( $defaultProto === PROTO_CANONICAL ) {
 556                 $serverUrl = $wgCanonicalServer;
 557         } elseif ( $defaultProto === PROTO_INTERNAL && $wgInternalServer !== false ) {
 558                 // Make $wgInternalServer fall back to $wgServer if not set
 559                 $serverUrl = $wgInternalServer;
 560         } else {
 561                 $serverUrl = $wgServer;
 562                 if ( $defaultProto === PROTO_CURRENT ) {
 563                         $defaultProto = $wgRequest->getProtocol() . '://';
 564                 }
 565         }
 566
 567         // Analyze $serverUrl to obtain its protocol
 568         $bits = wfParseUrl( $serverUrl );
 569         $serverHasProto = $bits && $bits['scheme'] != '';
 570
 571         if ( $defaultProto === PROTO_CANONICAL || $defaultProto === PROTO_INTERNAL ) {
 572                 if ( $serverHasProto ) {
 573                         $defaultProto = $bits['scheme'] . '://';
 574                 } else {
 575                         // $wgCanonicalServer or $wgInternalServer doesn't have a protocol.
 576                         // This really isn't supposed to happen. Fall back to HTTP in this
 577                         // ridiculous case.
 578                         $defaultProto = PROTO_HTTP;
 579                 }
 580         }
 581
 582         $defaultProtoWithoutSlashes = substr( $defaultProto, 0, -2 );
 583
 584         if ( substr( $url, 0, 2 ) == '//' ) {
 585                 $url = $defaultProtoWithoutSlashes . $url;
 586         } elseif ( substr( $url, 0, 1 ) == '/' ) {
 587                 // If $serverUrl is protocol-relative, prepend $defaultProtoWithoutSlashes,
 588                 // otherwise leave it alone.
 589                 $url = ( $serverHasProto ? '' : $defaultProtoWithoutSlashes ) . $serverUrl . $url;
 590         }
 591
 592         $bits = wfParseUrl( $url );
 593
 594         // ensure proper port for HTTPS arrives in URL
 595         // https://phabricator.wikimedia.org/T67184
 596         if ( $defaultProto === PROTO_HTTPS && $wgHttpsPort != 443 ) {
 597                 $bits['port'] = $wgHttpsPort;
 598         }
 599
 600         if ( $bits && isset( $bits['path'] ) ) {
 601                 $bits['path'] = wfRemoveDotSegments( $bits['path'] );
 602                 return wfAssembleUrl( $bits );
 603         } elseif ( $bits ) {
 604                 # No path to expand
 605                 return $url;
 606         } elseif ( substr( $url, 0, 1 ) != '/' ) {
 607                 # URL is a relative path
 608                 return wfRemoveDotSegments( $url );
 609         }
 610
 611         # Expanded URL is not valid.
 612         return false;
 613 }
 614
 615 /**
 616  * This function will reassemble a URL parsed with wfParseURL.  This is useful
 617  * if you need to edit part of a URL and put it back together.
 618  *
 619  * This is the basic structure used (brackets contain keys for $urlParts):
 620  * [scheme][delimiter][user]:[pass]@[host]:[port][path]?[query]#[fragment]
 621  *
 622  * @todo Need to integrate this into wfExpandUrl (see T34168)
 623  *
 624  * @since 1.19
 625  * @param array $urlParts URL parts, as output from wfParseUrl
 626  * @return string URL assembled from its component parts
 627  */
 628 function wfAssembleUrl( $urlParts ) {
 629         $result = '';
 630
 631         if ( isset( $urlParts['delimiter'] ) ) {
 632                 if ( isset( $urlParts['scheme'] ) ) {
 633                         $result .= $urlParts['scheme'];
 634                 }
 635
 636                 $result .= $urlParts['delimiter'];
 637         }
 638
 639         if ( isset( $urlParts['host'] ) ) {
 640                 if ( isset( $urlParts['user'] ) ) {
 641                         $result .= $urlParts['user'];
 642                         if ( isset( $urlParts['pass'] ) ) {
 643                                 $result .= ':' . $urlParts['pass'];
 644                         }
 645                         $result .= '@';
 646                 }
 647
 648                 $result .= $urlParts['host'];
 649
 650                 if ( isset( $urlParts['port'] ) ) {
 651                         $result .= ':' . $urlParts['port'];
 652                 }
 653         }
 654
 655         if ( isset( $urlParts['path'] ) ) {
 656                 $result .= $urlParts['path'];
 657         }
 658
 659         if ( isset( $urlParts['query'] ) ) {
 660                 $result .= '?' . $urlParts['query'];
 661         }
 662
 663         if ( isset( $urlParts['fragment'] ) ) {
 664                 $result .= '#' . $urlParts['fragment'];
 665         }
 666
 667         return $result;
 668 }
 669
 670 /**
 671  * Remove all dot-segments in the provided URL path.  For example,
 672  * '/a/./b/../c/' becomes '/a/c/'.  For details on the algorithm, please see
 673  * RFC3986 section 5.2.4.
 674  *
 675  * @todo Need to integrate this into wfExpandUrl (see T34168)
 676  *
 677  * @param string $urlPath URL path, potentially containing dot-segments
 678  * @return string URL path with all dot-segments removed
 679  */
 680 function wfRemoveDotSegments( $urlPath ) {
 681         $output = '';
 682         $inputOffset = 0;
 683         $inputLength = strlen( $urlPath );
 684
 685         while ( $inputOffset < $inputLength ) {
 686                 $prefixLengthOne = substr( $urlPath, $inputOffset, 1 );
 687                 $prefixLengthTwo = substr( $urlPath, $inputOffset, 2 );
 688                 $prefixLengthThree = substr( $urlPath, $inputOffset, 3 );
 689                 $prefixLengthFour = substr( $urlPath, $inputOffset, 4 );
 690                 $trimOutput = false;
 691
 692                 if ( $prefixLengthTwo == './' ) {
 693                         # Step A, remove leading "./"
 694                         $inputOffset += 2;
 695                 } elseif ( $prefixLengthThree == '../' ) {
 696                         # Step A, remove leading "../"
 697                         $inputOffset += 3;
 698                 } elseif ( ( $prefixLengthTwo == '/.' ) && ( $inputOffset + 2 == $inputLength ) ) {
 699                         # Step B, replace leading "/.$" with "/"
 700                         $inputOffset += 1;
 701                         $urlPath[$inputOffset] = '/';
 702                 } elseif ( $prefixLengthThree == '/./' ) {
 703                         # Step B, replace leading "/./" with "/"
 704                         $inputOffset += 2;
 705                 } elseif ( $prefixLengthThree == '/..' && ( $inputOffset + 3 == $inputLength ) ) {
 706                         # Step C, replace leading "/..$" with "/" and
 707                         # remove last path component in output
 708                         $inputOffset += 2;
 709                         $urlPath[$inputOffset] = '/';
 710                         $trimOutput = true;
 711                 } elseif ( $prefixLengthFour == '/../' ) {
 712                         # Step C, replace leading "/../" with "/" and
 713                         # remove last path component in output
 714                         $inputOffset += 3;
 715                         $trimOutput = true;
 716                 } elseif ( ( $prefixLengthOne == '.' ) && ( $inputOffset + 1 == $inputLength ) ) {
 717                         # Step D, remove "^.$"
 718                         $inputOffset += 1;
 719                 } elseif ( ( $prefixLengthTwo == '..' ) && ( $inputOffset + 2 == $inputLength ) ) {
 720                         # Step D, remove "^..$"
 721                         $inputOffset += 2;
 722                 } else {
 723                         # Step E, move leading path segment to output
 724                         if ( $prefixLengthOne == '/' ) {
 725                                 $slashPos = strpos( $urlPath, '/', $inputOffset + 1 );
 726                         } else {
 727                                 $slashPos = strpos( $urlPath, '/', $inputOffset );
 728                         }
 729                         if ( $slashPos === false ) {
 730                                 $output .= substr( $urlPath, $inputOffset );
 731                                 $inputOffset = $inputLength;
 732                         } else {
 733                                 $output .= substr( $urlPath, $inputOffset, $slashPos - $inputOffset );
 734                                 $inputOffset += $slashPos - $inputOffset;
 735                         }
 736                 }
 737
 738                 if ( $trimOutput ) {
 739                         $slashPos = strrpos( $output, '/' );
 740                         if ( $slashPos === false ) {
 741                                 $output = '';
 742                         } else {
 743                                 $output = substr( $output, 0, $slashPos );
 744                         }
 745                 }
 746         }
 747
 748         return $output;
 749 }
 750
 751 /**
 752  * Returns a regular expression of url protocols
 753  *
 754  * @param bool $includeProtocolRelative If false, remove '//' from the returned protocol list.
 755  *        DO NOT USE this directly, use wfUrlProtocolsWithoutProtRel() instead
 756  * @return string
 757  */
 758 function wfUrlProtocols( $includeProtocolRelative = true ) {
 759         global $wgUrlProtocols;
 760
 761         // Cache return values separately based on $includeProtocolRelative
 762         static $withProtRel = null, $withoutProtRel = null;
 763         $cachedValue = $includeProtocolRelative ? $withProtRel : $withoutProtRel;
 764         if ( !is_null( $cachedValue ) ) {
 765                 return $cachedValue;
 766         }
 767
 768         // Support old-style $wgUrlProtocols strings, for backwards compatibility
 769         // with LocalSettings files from 1.5
 770         if ( is_array( $wgUrlProtocols ) ) {
 771                 $protocols = [];
 772                 foreach ( $wgUrlProtocols as $protocol ) {
 773                         // Filter out '//' if !$includeProtocolRelative
 774                         if ( $includeProtocolRelative || $protocol !== '//' ) {
 775                                 $protocols[] = preg_quote( $protocol, '/' );
 776                         }
 777                 }
 778
 779                 $retval = implode( '|', $protocols );
 780         } else {
 781                 // Ignore $includeProtocolRelative in this case
 782                 // This case exists for pre-1.6 compatibility, and we can safely assume
 783                 // that '//' won't appear in a pre-1.6 config because protocol-relative
 784                 // URLs weren't supported until 1.18
 785                 $retval = $wgUrlProtocols;
 786         }
 787
 788         // Cache return value
 789         if ( $includeProtocolRelative ) {
 790                 $withProtRel = $retval;
 791         } else {
 792                 $withoutProtRel = $retval;
 793         }
 794         return $retval;
 795 }
 796
 797 /**
 798  * Like wfUrlProtocols(), but excludes '//' from the protocol list. Use this if
 799  * you need a regex that matches all URL protocols but does not match protocol-
 800  * relative URLs
 801  * @return string
 802  */
 803 function wfUrlProtocolsWithoutProtRel() {
 804         return wfUrlProtocols( false );
 805 }
 806
 807 /**
 808  * parse_url() work-alike, but non-broken.  Differences:
 809  *
 810  * 1) Does not raise warnings on bad URLs (just returns false).
 811  * 2) Handles protocols that don't use :// (e.g., mailto: and news:, as well as
 812  *    protocol-relative URLs) correctly.
 813  * 3) Adds a "delimiter" element to the array, either '://', ':' or '//' (see (2)).
 814  *
 815  * @param string $url A URL to parse
 816  * @return string[]|bool Bits of the URL in an associative array, per PHP docs, false on failure
 817  */
 818 function wfParseUrl( $url ) {
 819         global $wgUrlProtocols; // Allow all protocols defined in DefaultSettings/LocalSettings.php
 820
 821         // Protocol-relative URLs are handled really badly by parse_url(). It's so
 822         // bad that the easiest way to handle them is to just prepend 'http:' and
 823         // strip the protocol out later.
 824         $wasRelative = substr( $url, 0, 2 ) == '//';
 825         if ( $wasRelative ) {
 826                 $url = "http:$url";
 827         }
 828         MediaWiki\suppressWarnings();
 829         $bits = parse_url( $url );
 830         MediaWiki\restoreWarnings();
 831         // parse_url() returns an array without scheme for some invalid URLs, e.g.
 832         // parse_url("%0Ahttp://example.com") == [ 'host' => '%0Ahttp', 'path' => 'example.com' ]
 833         if ( !$bits || !isset( $bits['scheme'] ) ) {
 834                 return false;
 835         }
 836
 837         // parse_url() incorrectly handles schemes case-sensitively. Convert it to lowercase.
 838         $bits['scheme'] = strtolower( $bits['scheme'] );
 839
 840         // most of the protocols are followed by ://, but mailto: and sometimes news: not, check for it
 841         if ( in_array( $bits['scheme'] . '://', $wgUrlProtocols ) ) {
 842                 $bits['delimiter'] = '://';
 843         } elseif ( in_array( $bits['scheme'] . ':', $wgUrlProtocols ) ) {
 844                 $bits['delimiter'] = ':';
 845                 // parse_url detects for news: and mailto: the host part of an url as path
 846                 // We have to correct this wrong detection
 847                 if ( isset( $bits['path'] ) ) {
 848                         $bits['host'] = $bits['path'];
 849                         $bits['path'] = '';
 850                 }
 851         } else {
 852                 return false;
 853         }
 854
 855         /* Provide an empty host for eg. file:/// urls (see T30627) */
 856         if ( !isset( $bits['host'] ) ) {
 857                 $bits['host'] = '';
 858
 859                 // See T47069
 860                 if ( isset( $bits['path'] ) ) {
 861                         /* parse_url loses the third / for file:///c:/ urls (but not on variants) */
 862                         if ( substr( $bits['path'], 0, 1 ) !== '/' ) {
 863                                 $bits['path'] = '/' . $bits['path'];
 864                         }
 865                 } else {
 866                         $bits['path'] = '';
 867                 }
 868         }
 869
 870         // If the URL was protocol-relative, fix scheme and delimiter
 871         if ( $wasRelative ) {
 872                 $bits['scheme'] = '';
 873                 $bits['delimiter'] = '//';
 874         }
 875         return $bits;
 876 }
 877
 878 /**
 879  * Take a URL, make sure it's expanded to fully qualified, and replace any
 880  * encoded non-ASCII Unicode characters with their UTF-8 original forms
 881  * for more compact display and legibility for local audiences.
 882  *
 883  * @todo handle punycode domains too
 884  *
 885  * @param string $url
 886  * @return string
 887  */
 888 function wfExpandIRI( $url ) {
 889         return preg_replace_callback(
 890                 '/((?:%[89A-F][0-9A-F])+)/i',
 891                 'wfExpandIRI_callback',
 892                 wfExpandUrl( $url )
 893         );
 894 }
 895
 896 /**
 897  * Private callback for wfExpandIRI
 898  * @param array $matches
 899  * @return string
 900  */
 901 function wfExpandIRI_callback( $matches ) {
 902         return urldecode( $matches[1] );
 903 }
 904
 905 /**
 906  * Make URL indexes, appropriate for the el_index field of externallinks.
 907  *
 908  * @param string $url
 909  * @return array
 910  */
 911 function wfMakeUrlIndexes( $url ) {
 912         $bits = wfParseUrl( $url );
 913
 914         // Reverse the labels in the hostname, convert to lower case
 915         // For emails reverse domainpart only
 916         if ( $bits['scheme'] == 'mailto' ) {
 917                 $mailparts = explode( '@', $bits['host'], 2 );
 918                 if ( count( $mailparts ) === 2 ) {
 919                         $domainpart = strtolower( implode( '.', array_reverse( explode( '.', $mailparts[1] ) ) ) );
 920                 } else {
 921                         // No domain specified, don't mangle it
 922                         $domainpart = '';
 923                 }
 924                 $reversedHost = $domainpart . '@' . $mailparts[0];
 925         } else {
 926                 $reversedHost = strtolower( implode( '.', array_reverse( explode( '.', $bits['host'] ) ) ) );
 927         }
 928         // Add an extra dot to the end
 929         // Why? Is it in wrong place in mailto links?
 930         if ( substr( $reversedHost, -1, 1 ) !== '.' ) {
 931                 $reversedHost .= '.';
 932         }
 933         // Reconstruct the pseudo-URL
 934         $prot = $bits['scheme'];
 935         $index = $prot . $bits['delimiter'] . $reversedHost;
 936         // Leave out user and password. Add the port, path, query and fragment
 937         if ( isset( $bits['port'] ) ) {
 938                 $index .= ':' . $bits['port'];
 939         }
 940         if ( isset( $bits['path'] ) ) {
 941                 $index .= $bits['path'];
 942         } else {
 943                 $index .= '/';
 944         }
 945         if ( isset( $bits['query'] ) ) {
 946                 $index .= '?' . $bits['query'];
 947         }
 948         if ( isset( $bits['fragment'] ) ) {
 949                 $index .= '#' . $bits['fragment'];
 950         }
 951
 952         if ( $prot == '' ) {
 953                 return [ "http:$index", "https:$index" ];
 954         } else {
 955                 return [ $index ];
 956         }
 957 }
 958
 959 /**
 960  * Check whether a given URL has a domain that occurs in a given set of domains
 961  * @param string $url URL
 962  * @param array $domains Array of domains (strings)
 963  * @return bool True if the host part of $url ends in one of the strings in $domains
 964  */
 965 function wfMatchesDomainList( $url, $domains ) {
 966         $bits = wfParseUrl( $url );
 967         if ( is_array( $bits ) && isset( $bits['host'] ) ) {
 968                 $host = '.' . $bits['host'];
 969                 foreach ( (array)$domains as $domain ) {
 970                         $domain = '.' . $domain;
 971                         if ( substr( $host, -strlen( $domain ) ) === $domain ) {
 972                                 return true;
 973                         }
 974                 }
 975         }
 976         return false;
 977 }
 978
 979 /**
 980  * Sends a line to the debug log if enabled or, optionally, to a comment in output.
 981  * In normal operation this is a NOP.
 982  *
 983  * Controlling globals:
 984  * $wgDebugLogFile - points to the log file
 985  * $wgDebugRawPage - if false, 'action=raw' hits will not result in debug output.
 986  * $wgDebugComments - if on, some debug items may appear in comments in the HTML output.
 987  *
 988  * @since 1.25 support for additional context data
 989  *
 990  * @param string $text
 991  * @param string|bool $dest Destination of the message:
 992  *     - 'all': both to the log and HTML (debug toolbar or HTML comments)
 993  *     - 'private': excluded from HTML output
 994  *   For backward compatibility, it can also take a boolean:
 995  *     - true: same as 'all'
 996  *     - false: same as 'private'
 997  * @param array $context Additional logging context data
 998  */
 999 function wfDebug( $text, $dest = 'all', array $context = [] ) {
1000         global $wgDebugRawPage, $wgDebugLogPrefix;
1001         global $wgDebugTimestamps, $wgRequestTime;
1002
1003         if ( !$wgDebugRawPage && wfIsDebugRawPage() ) {
1004                 return;
1005         }
1006
1007         $text = trim( $text );
1008
1009         if ( $wgDebugTimestamps ) {
1010                 $context['seconds_elapsed'] = sprintf(
1011                         '%6.4f',
1012                         microtime( true ) - $wgRequestTime
1013                 );
1014                 $context['memory_used'] = sprintf(
1015                         '%5.1fM',
1016                         ( memory_get_usage( true ) / ( 1024 * 1024 ) )
1017                 );
1018         }
1019
1020         if ( $wgDebugLogPrefix !== '' ) {
1021                 $context['prefix'] = $wgDebugLogPrefix;
1022         }
1023         $context['private'] = ( $dest === false || $dest === 'private' );
1024
1025         $logger = LoggerFactory::getInstance( 'wfDebug' );
1026         $logger->debug( $text, $context );
1027 }
1028
1029 /**
1030  * Returns true if debug logging should be suppressed if $wgDebugRawPage = false
1031  * @return bool
1032  */
1033 function wfIsDebugRawPage() {
1034         static $cache;
1035         if ( $cache !== null ) {
1036                 return $cache;
1037         }
1038         # Check for raw action using $_GET not $wgRequest, since the latter might not be initialised yet
1039         if ( ( isset( $_GET['action'] ) && $_GET['action'] == 'raw' )
1040                 || (
1041                         isset( $_SERVER['SCRIPT_NAME'] )
1042                         && substr( $_SERVER['SCRIPT_NAME'], -8 ) == 'load.php'
1043                 )
1044         ) {
1045                 $cache = true;
1046         } else {
1047                 $cache = false;
1048         }
1049         return $cache;
1050 }
1051
1052 /**
1053  * Send a line giving PHP memory usage.
1054  *
1055  * @param bool $exact Print exact byte values instead of kibibytes (default: false)
1056  */
1057 function wfDebugMem( $exact = false ) {
1058         $mem = memory_get_usage();
1059         if ( !$exact ) {
1060                 $mem = floor( $mem / 1024 ) . ' KiB';
1061         } else {
1062                 $mem .= ' B';
1063         }
1064         wfDebug( "Memory usage: $mem\n" );
1065 }
1066
1067 /**
1068  * Send a line to a supplementary debug log file, if configured, or main debug
1069  * log if not.
1070  *
1071  * To configure a supplementary log file, set $wgDebugLogGroups[$logGroup] to
1072  * a string filename or an associative array mapping 'destination' to the
1073  * desired filename. The associative array may also contain a 'sample' key
1074  * with an integer value, specifying a sampling factor. Sampled log events
1075  * will be emitted with a 1 in N random chance.
1076  *
1077  * @since 1.23 support for sampling log messages via $wgDebugLogGroups.
1078  * @since 1.25 support for additional context data
1079  * @since 1.25 sample behavior dependent on configured $wgMWLoggerDefaultSpi
1080  *
1081  * @param string $logGroup
1082  * @param string $text
1083  * @param string|bool $dest Destination of the message:
1084  *     - 'all': both to the log and HTML (debug toolbar or HTML comments)
1085  *     - 'private': only to the specific log if set in $wgDebugLogGroups and
1086  *       discarded otherwise
1087  *   For backward compatibility, it can also take a boolean:
1088  *     - true: same as 'all'
1089  *     - false: same as 'private'
1090  * @param array $context Additional logging context data
1091  */
1092 function wfDebugLog(
1093         $logGroup, $text, $dest = 'all', array $context = []
1094 ) {
1095         $text = trim( $text );
1096
1097         $logger = LoggerFactory::getInstance( $logGroup );
1098         $context['private'] = ( $dest === false || $dest === 'private' );
1099         $logger->info( $text, $context );
1100 }
1101
1102 /**
1103  * Log for database errors
1104  *
1105  * @since 1.25 support for additional context data
1106  *
1107  * @param string $text Database error message.
1108  * @param array $context Additional logging context data
1109  */
1110 function wfLogDBError( $text, array $context = [] ) {
1111         $logger = LoggerFactory::getInstance( 'wfLogDBError' );
1112         $logger->error( trim( $text ), $context );
1113 }
1114
1115 /**
1116  * Throws a warning that $function is deprecated
1117  *
1118  * @param string $function
1119  * @param string|bool $version Version of MediaWiki that the function
1120  *    was deprecated in (Added in 1.19).
1121  * @param string|bool $component Added in 1.19.
1122  * @param int $callerOffset How far up the call stack is the original
1123  *    caller. 2 = function that called the function that called
1124  *    wfDeprecated (Added in 1.20)
1125  *
1126  * @return null
1127  */
1128 function wfDeprecated( $function, $version = false, $component = false, $callerOffset = 2 ) {
1129         MWDebug::deprecated( $function, $version, $component, $callerOffset + 1 );
1130 }
1131
1132 /**
1133  * Send a warning either to the debug log or in a PHP error depending on
1134  * $wgDevelopmentWarnings. To log warnings in production, use wfLogWarning() instead.
1135  *
1136  * @param string $msg Message to send
1137  * @param int $callerOffset Number of items to go back in the backtrace to
1138  *        find the correct caller (1 = function calling wfWarn, ...)
1139  * @param int $level PHP error level; defaults to E_USER_NOTICE;
1140  *        only used when $wgDevelopmentWarnings is true
1141  */
1142 function wfWarn( $msg, $callerOffset = 1, $level = E_USER_NOTICE ) {
1143         MWDebug::warning( $msg, $callerOffset + 1, $level, 'auto' );
1144 }
1145
1146 /**
1147  * Send a warning as a PHP error and the debug log. This is intended for logging
1148  * warnings in production. For logging development warnings, use WfWarn instead.
1149  *
1150  * @param string $msg Message to send
1151  * @param int $callerOffset Number of items to go back in the backtrace to
1152  *        find the correct caller (1 = function calling wfLogWarning, ...)
1153  * @param int $level PHP error level; defaults to E_USER_WARNING
1154  */
1155 function wfLogWarning( $msg, $callerOffset = 1, $level = E_USER_WARNING ) {
1156         MWDebug::warning( $msg, $callerOffset + 1, $level, 'production' );
1157 }
1158
1159 /**
1160  * Log to a file without getting "file size exceeded" signals.
1161  *
1162  * Can also log to TCP or UDP with the syntax udp://host:port/prefix. This will
1163  * send lines to the specified port, prefixed by the specified prefix and a space.
1164  * @since 1.25 support for additional context data
1165  *
1166  * @param string $text
1167  * @param string $file Filename
1168  * @param array $context Additional logging context data
1169  * @throws MWException
1170  * @deprecated since 1.25 Use \MediaWiki\Logger\LegacyLogger::emit or UDPTransport
1171  */
1172 function wfErrorLog( $text, $file, array $context = [] ) {
1173         wfDeprecated( __METHOD__, '1.25' );
1174         $logger = LoggerFactory::getInstance( 'wfErrorLog' );
1175         $context['destination'] = $file;
1176         $logger->info( trim( $text ), $context );
1177 }
1178
1179 /**
1180  * @todo document
1181  */
1182 function wfLogProfilingData() {
1183         global $wgDebugLogGroups, $wgDebugRawPage;
1184
1185         $context = RequestContext::getMain();
1186         $request = $context->getRequest();
1187
1188         $profiler = Profiler::instance();
1189         $profiler->setContext( $context );
1190         $profiler->logData();
1191
1192         $config = $context->getConfig();
1193         if ( $config->get( 'StatsdServer' ) ) {
1194                 try {
1195                         $statsdServer = explode( ':', $config->get( 'StatsdServer' ) );
1196                         $statsdHost = $statsdServer[0];
1197                         $statsdPort = isset( $statsdServer[1] ) ? $statsdServer[1] : 8125;
1198                         $statsdSender = new SocketSender( $statsdHost, $statsdPort );
1199                         $statsdClient = new SamplingStatsdClient( $statsdSender, true, false );
1200                         $statsdClient->setSamplingRates( $config->get( 'StatsdSamplingRates' ) );
1201                         $statsdClient->send(
1202                                 MediaWikiServices::getInstance()->getStatsdDataFactory()->getBuffer()
1203                         );
1204                 } catch ( Exception $ex ) {
1205                         MWExceptionHandler::logException( $ex );
1206                 }
1207         }
1208
1209         # Profiling must actually be enabled...
1210         if ( $profiler instanceof ProfilerStub ) {
1211                 return;
1212         }
1213
1214         if ( isset( $wgDebugLogGroups['profileoutput'] )
1215                 && $wgDebugLogGroups['profileoutput'] === false
1216         ) {
1217                 // Explicitly disabled
1218                 return;
1219         }
1220         if ( !$wgDebugRawPage && wfIsDebugRawPage() ) {
1221                 return;
1222         }
1223
1224         $ctx = [ 'elapsed' => $request->getElapsedTime() ];
1225         if ( !empty( $_SERVER['HTTP_X_FORWARDED_FOR'] ) ) {
1226                 $ctx['forwarded_for'] = $_SERVER['HTTP_X_FORWARDED_FOR'];
1227         }
1228         if ( !empty( $_SERVER['HTTP_CLIENT_IP'] ) ) {
1229                 $ctx['client_ip'] = $_SERVER['HTTP_CLIENT_IP'];
1230         }
1231         if ( !empty( $_SERVER['HTTP_FROM'] ) ) {
1232                 $ctx['from'] = $_SERVER['HTTP_FROM'];
1233         }
1234         if ( isset( $ctx['forwarded_for'] ) ||
1235                 isset( $ctx['client_ip'] ) ||
1236                 isset( $ctx['from'] ) ) {
1237                 $ctx['proxy'] = $_SERVER['REMOTE_ADDR'];
1238         }
1239
1240         // Don't load $wgUser at this late stage just for statistics purposes
1241         // @todo FIXME: We can detect some anons even if it is not loaded.
1242         // See User::getId()
1243         $user = $context->getUser();
1244         $ctx['anon'] = $user->isItemLoaded( 'id' ) && $user->isAnon();
1245
1246         // Command line script uses a FauxRequest object which does not have
1247         // any knowledge about an URL and throw an exception instead.
1248         try {
1249                 $ctx['url'] = urldecode( $request->getRequestURL() );
1250         } catch ( Exception $ignored ) {
1251                 // no-op
1252         }
1253
1254         $ctx['output'] = $profiler->getOutput();
1255
1256         $log = LoggerFactory::getInstance( 'profileoutput' );
1257         $log->info( "Elapsed: {elapsed}; URL: <{url}>\n{output}", $ctx );
1258 }
1259
1260 /**
1261  * Increment a statistics counter
1262  *
1263  * @param string $key
1264  * @param int $count
1265  * @return void
1266  */
1267 function wfIncrStats( $key, $count = 1 ) {
1268         $stats = MediaWikiServices::getInstance()->getStatsdDataFactory();
1269         $stats->updateCount( $key, $count );
1270 }
1271
1272 /**
1273  * Check whether the wiki is in read-only mode.
1274  *
1275  * @return bool
1276  */
1277 function wfReadOnly() {
1278         return wfReadOnlyReason() !== false;
1279 }
1280
1281 /**
1282  * Check if the site is in read-only mode and return the message if so
1283  *
1284  * This checks wfConfiguredReadOnlyReason() and the main load balancer
1285  * for replica DB lag. This may result in DB connection being made.
1286  *
1287  * @return string|bool String when in read-only mode; false otherwise
1288  */
1289 function wfReadOnlyReason() {
1290         $readOnly = wfConfiguredReadOnlyReason();
1291         if ( $readOnly !== false ) {
1292                 return $readOnly;
1293         }
1294
1295         static $lbReadOnly = null;
1296         if ( $lbReadOnly === null ) {
1297                 // Callers use this method to be aware that data presented to a user
1298                 // may be very stale and thus allowing submissions can be problematic.
1299                 $lbReadOnly = wfGetLB()->getReadOnlyReason();
1300         }
1301
1302         return $lbReadOnly;
1303 }
1304
1305 /**
1306  * Get the value of $wgReadOnly or the contents of $wgReadOnlyFile.
1307  *
1308  * @return string|bool String when in read-only mode; false otherwise
1309  * @since 1.27
1310  */
1311 function wfConfiguredReadOnlyReason() {
1312         global $wgReadOnly, $wgReadOnlyFile;
1313
1314         if ( $wgReadOnly === null ) {
1315                 // Set $wgReadOnly for faster access next time
1316                 if ( is_file( $wgReadOnlyFile ) && filesize( $wgReadOnlyFile ) > 0 ) {
1317                         $wgReadOnly = file_get_contents( $wgReadOnlyFile );
1318                 } else {
1319                         $wgReadOnly = false;
1320                 }
1321         }
1322
1323         return $wgReadOnly;
1324 }
1325
1326 /**
1327  * Return a Language object from $langcode
1328  *
1329  * @param Language|string|bool $langcode Either:
1330  *                  - a Language object
1331  *                  - code of the language to get the message for, if it is
1332  *                    a valid code create a language for that language, if
1333  *                    it is a string but not a valid code then make a basic
1334  *                    language object
1335  *                  - a boolean: if it's false then use the global object for
1336  *                    the current user's language (as a fallback for the old parameter
1337  *                    functionality), or if it is true then use global object
1338  *                    for the wiki's content language.
1339  * @return Language
1340  */
1341 function wfGetLangObj( $langcode = false ) {
1342         # Identify which language to get or create a language object for.
1343         # Using is_object here due to Stub objects.
1344         if ( is_object( $langcode ) ) {
1345                 # Great, we already have the object (hopefully)!
1346                 return $langcode;
1347         }
1348
1349         global $wgContLang, $wgLanguageCode;
1350         if ( $langcode === true || $langcode === $wgLanguageCode ) {
1351                 # $langcode is the language code of the wikis content language object.
1352                 # or it is a boolean and value is true
1353                 return $wgContLang;
1354         }
1355
1356         global $wgLang;
1357         if ( $langcode === false || $langcode === $wgLang->getCode() ) {
1358                 # $langcode is the language code of user language object.
1359                 # or it was a boolean and value is false
1360                 return $wgLang;
1361         }
1362
1363         $validCodes = array_keys( Language::fetchLanguageNames() );
1364         if ( in_array( $langcode, $validCodes ) ) {
1365                 # $langcode corresponds to a valid language.
1366                 return Language::factory( $langcode );
1367         }
1368
1369         # $langcode is a string, but not a valid language code; use content language.
1370         wfDebug( "Invalid language code passed to wfGetLangObj, falling back to content language.\n" );
1371         return $wgContLang;
1372 }
1373
1374 /**
1375  * This is the function for getting translated interface messages.
1376  *
1377  * @see Message class for documentation how to use them.
1378  * @see https://www.mediawiki.org/wiki/Manual:Messages_API
1379  *
1380  * This function replaces all old wfMsg* functions.
1381  *
1382  * @param string|string[]|MessageSpecifier $key Message key, or array of keys, or a MessageSpecifier
1383  * @param mixed $params,... Normal message parameters
1384  * @return Message
1385  *
1386  * @since 1.17
1387  *
1388  * @see Message::__construct
1389  */
1390 function wfMessage( $key /*...*/ ) {
1391         $message = new Message( $key );
1392
1393         // We call Message::params() to reduce code duplication
1394         $params = func_get_args();
1395         array_shift( $params );
1396         if ( $params ) {
1397                 call_user_func_array( [ $message, 'params' ], $params );
1398         }
1399
1400         return $message;
1401 }
1402
1403 /**
1404  * This function accepts multiple message keys and returns a message instance
1405  * for the first message which is non-empty. If all messages are empty then an
1406  * instance of the first message key is returned.
1407  *
1408  * @param string|string[] $keys,... Message keys
1409  * @return Message
1410  *
1411  * @since 1.18
1412  *
1413  * @see Message::newFallbackSequence
1414  */
1415 function wfMessageFallback( /*...*/ ) {
1416         $args = func_get_args();
1417         return call_user_func_array( 'Message::newFallbackSequence', $args );
1418 }
1419
1420 /**
1421  * Replace message parameter keys on the given formatted output.
1422  *
1423  * @param string $message
1424  * @param array $args
1425  * @return string
1426  * @private
1427  */
1428 function wfMsgReplaceArgs( $message, $args ) {
1429         # Fix windows line-endings
1430         # Some messages are split with explode("\n", $msg)
1431         $message = str_replace( "\r", '', $message );
1432
1433         // Replace arguments
1434         if ( is_array( $args ) && $args ) {
1435                 if ( is_array( $args[0] ) ) {
1436                         $args = array_values( $args[0] );
1437                 }
1438                 $replacementKeys = [];
1439                 foreach ( $args as $n => $param ) {
1440                         $replacementKeys['$' . ( $n + 1 )] = $param;
1441                 }
1442                 $message = strtr( $message, $replacementKeys );
1443         }
1444
1445         return $message;
1446 }
1447
1448 /**
1449  * Fetch server name for use in error reporting etc.
1450  * Use real server name if available, so we know which machine
1451  * in a server farm generated the current page.
1452  *
1453  * @return string
1454  */
1455 function wfHostname() {
1456         static $host;
1457         if ( is_null( $host ) ) {
1458
1459                 # Hostname overriding
1460                 global $wgOverrideHostname;
1461                 if ( $wgOverrideHostname !== false ) {
1462                         # Set static and skip any detection
1463                         $host = $wgOverrideHostname;
1464                         return $host;
1465                 }
1466
1467                 if ( function_exists( 'posix_uname' ) ) {
1468                         // This function not present on Windows
1469                         $uname = posix_uname();
1470                 } else {
1471                         $uname = false;
1472                 }
1473                 if ( is_array( $uname ) && isset( $uname['nodename'] ) ) {
1474                         $host = $uname['nodename'];
1475                 } elseif ( getenv( 'COMPUTERNAME' ) ) {
1476                         # Windows computer name
1477                         $host = getenv( 'COMPUTERNAME' );
1478                 } else {
1479                         # This may be a virtual server.
1480                         $host = $_SERVER['SERVER_NAME'];
1481                 }
1482         }
1483         return $host;
1484 }
1485
1486 /**
1487  * Returns a script tag that stores the amount of time it took MediaWiki to
1488  * handle the request in milliseconds as 'wgBackendResponseTime'.
1489  *
1490  * If $wgShowHostnames is true, the script will also set 'wgHostname' to the
1491  * hostname of the server handling the request.
1492  *
1493  * @return string
1494  */
1495 function wfReportTime() {
1496         global $wgRequestTime, $wgShowHostnames;
1497
1498         $responseTime = round( ( microtime( true ) - $wgRequestTime ) * 1000 );
1499         $reportVars = [ 'wgBackendResponseTime' => $responseTime ];
1500         if ( $wgShowHostnames ) {
1501                 $reportVars['wgHostname'] = wfHostname();
1502         }
1503         return Skin::makeVariablesScript( $reportVars );
1504 }
1505
1506 /**
1507  * Safety wrapper for debug_backtrace().
1508  *
1509  * Will return an empty array if debug_backtrace is disabled, otherwise
1510  * the output from debug_backtrace() (trimmed).
1511  *
1512  * @param int $limit This parameter can be used to limit the number of stack frames returned
1513  *
1514  * @return array Array of backtrace information
1515  */
1516 function wfDebugBacktrace( $limit = 0 ) {
1517         static $disabled = null;
1518
1519         if ( is_null( $disabled ) ) {
1520                 $disabled = !function_exists( 'debug_backtrace' );
1521                 if ( $disabled ) {
1522                         wfDebug( "debug_backtrace() is disabled\n" );
1523                 }
1524         }
1525         if ( $disabled ) {
1526                 return [];
1527         }
1528
1529         if ( $limit ) {
1530                 return array_slice( debug_backtrace( DEBUG_BACKTRACE_PROVIDE_OBJECT, $limit + 1 ), 1 );
1531         } else {
1532                 return array_slice( debug_backtrace(), 1 );
1533         }
1534 }
1535
1536 /**
1537  * Get a debug backtrace as a string
1538  *
1539  * @param bool|null $raw If true, the return value is plain text. If false, HTML.
1540  *   Defaults to $wgCommandLineMode if unset.
1541  * @return string
1542  * @since 1.25 Supports $raw parameter.
1543  */
1544 function wfBacktrace( $raw = null ) {
1545         global $wgCommandLineMode;
1546
1547         if ( $raw === null ) {
1548                 $raw = $wgCommandLineMode;
1549         }
1550
1551         if ( $raw ) {
1552                 $frameFormat = "%s line %s calls %s()\n";
1553                 $traceFormat = "%s";
1554         } else {
1555                 $frameFormat = "<li>%s line %s calls %s()</li>\n";
1556                 $traceFormat = "<ul>\n%s</ul>\n";
1557         }
1558
1559         $frames = array_map( function ( $frame ) use ( $frameFormat ) {
1560                 $file = !empty( $frame['file'] ) ? basename( $frame['file'] ) : '-';
1561                 $line = isset( $frame['line'] ) ? $frame['line'] : '-';
1562                 $call = $frame['function'];
1563                 if ( !empty( $frame['class'] ) ) {
1564                         $call = $frame['class'] . $frame['type'] . $call;
1565                 }
1566                 return sprintf( $frameFormat, $file, $line, $call );
1567         }, wfDebugBacktrace() );
1568
1569         return sprintf( $traceFormat, implode( '', $frames ) );
1570 }
1571
1572 /**
1573  * Get the name of the function which called this function
1574  * wfGetCaller( 1 ) is the function with the wfGetCaller() call (ie. __FUNCTION__)
1575  * wfGetCaller( 2 ) [default] is the caller of the function running wfGetCaller()
1576  * wfGetCaller( 3 ) is the parent of that.
1577  *
1578  * @param int $level
1579  * @return string
1580  */
1581 function wfGetCaller( $level = 2 ) {
1582         $backtrace = wfDebugBacktrace( $level + 1 );
1583         if ( isset( $backtrace[$level] ) ) {
1584                 return wfFormatStackFrame( $backtrace[$level] );
1585         } else {
1586                 return 'unknown';
1587         }
1588 }
1589
1590 /**
1591  * Return a string consisting of callers in the stack. Useful sometimes
1592  * for profiling specific points.
1593  *
1594  * @param int $limit The maximum depth of the stack frame to return, or false for the entire stack.
1595  * @return string
1596  */
1597 function wfGetAllCallers( $limit = 3 ) {
1598         $trace = array_reverse( wfDebugBacktrace() );
1599         if ( !$limit || $limit > count( $trace ) - 1 ) {
1600                 $limit = count( $trace ) - 1;
1601         }
1602         $trace = array_slice( $trace, -$limit - 1, $limit );
1603         return implode( '/', array_map( 'wfFormatStackFrame', $trace ) );
1604 }
1605
1606 /**
1607  * Return a string representation of frame
1608  *
1609  * @param array $frame
1610  * @return string
1611  */
1612 function wfFormatStackFrame( $frame ) {
1613         if ( !isset( $frame['function'] ) ) {
1614                 return 'NO_FUNCTION_GIVEN';
1615         }
1616         return isset( $frame['class'] ) && isset( $frame['type'] ) ?
1617                 $frame['class'] . $frame['type'] . $frame['function'] :
1618                 $frame['function'];
1619 }
1620
1621 /* Some generic result counters, pulled out of SearchEngine */
1622
1623 /**
1624  * @todo document
1625  *
1626  * @param int $offset
1627  * @param int $limit
1628  * @return string
1629  */
1630 function wfShowingResults( $offset, $limit ) {
1631         return wfMessage( 'showingresults' )->numParams( $limit, $offset + 1 )->parse();
1632 }
1633
1634 /**
1635  * Whether the client accept gzip encoding
1636  *
1637  * Uses the Accept-Encoding header to check if the client supports gzip encoding.
1638  * Use this when considering to send a gzip-encoded response to the client.
1639  *
1640  * @param bool $force Forces another check even if we already have a cached result.
1641  * @return bool
1642  */
1643 function wfClientAcceptsGzip( $force = false ) {
1644         static $result = null;
1645         if ( $result === null || $force ) {
1646                 $result = false;
1647                 if ( isset( $_SERVER['HTTP_ACCEPT_ENCODING'] ) ) {
1648                         # @todo FIXME: We may want to blacklist some broken browsers
1649                         $m = [];
1650                         if ( preg_match(
1651                                         '/\bgzip(?:;(q)=([0-9]+(?:\.[0-9]+)))?\b/',
1652                                         $_SERVER['HTTP_ACCEPT_ENCODING'],
1653                                         $m
1654                                 )
1655                         ) {
1656                                 if ( isset( $m[2] ) && ( $m[1] == 'q' ) && ( $m[2] == 0 ) ) {
1657                                         $result = false;
1658                                         return $result;
1659                                 }
1660                                 wfDebug( "wfClientAcceptsGzip: client accepts gzip.\n" );
1661                                 $result = true;
1662                         }
1663                 }
1664         }
1665         return $result;
1666 }
1667
1668 /**
1669  * Escapes the given text so that it may be output using addWikiText()
1670  * without any linking, formatting, etc. making its way through. This
1671  * is achieved by substituting certain characters with HTML entities.
1672  * As required by the callers, "<nowiki>" is not used.
1673  *
1674  * @param string $text Text to be escaped
1675  * @return string
1676  */
1677 function wfEscapeWikiText( $text ) {
1678         global $wgEnableMagicLinks;
1679         static $repl = null, $repl2 = null;
1680         if ( $repl === null || defined( 'MW_PARSER_TEST' ) || defined( 'MW_PHPUNIT_TEST' ) ) {
1681                 // Tests depend upon being able to change $wgEnableMagicLinks, so don't cache
1682                 // in those situations
1683                 $repl = [
1684                         '"' => '&#34;', '&' => '&#38;', "'" => '&#39;', '<' => '&#60;',
1685                         '=' => '&#61;', '>' => '&#62;', '[' => '&#91;', ']' => '&#93;',
1686                         '{' => '&#123;', '|' => '&#124;', '}' => '&#125;', ';' => '&#59;',
1687                         "\n#" => "\n&#35;", "\r#" => "\r&#35;",
1688                         "\n*" => "\n&#42;", "\r*" => "\r&#42;",
1689                         "\n:" => "\n&#58;", "\r:" => "\r&#58;",
1690                         "\n " => "\n&#32;", "\r " => "\r&#32;",
1691                         "\n\n" => "\n&#10;", "\r\n" => "&#13;\n",
1692                         "\n\r" => "\n&#13;", "\r\r" => "\r&#13;",
1693                         "\n\t" => "\n&#9;", "\r\t" => "\r&#9;", // "\n\t\n" is treated like "\n\n"
1694                         "\n----" => "\n&#45;---", "\r----" => "\r&#45;---",
1695                         '__' => '_&#95;', '://' => '&#58;//',
1696                 ];
1697
1698                 $magicLinks = array_keys( array_filter( $wgEnableMagicLinks ) );
1699                 // We have to catch everything "\s" matches in PCRE
1700                 foreach ( $magicLinks as $magic ) {
1701                         $repl["$magic "] = "$magic&#32;";
1702                         $repl["$magic\t"] = "$magic&#9;";
1703                         $repl["$magic\r"] = "$magic&#13;";
1704                         $repl["$magic\n"] = "$magic&#10;";
1705                         $repl["$magic\f"] = "$magic&#12;";
1706                 }
1707
1708                 // And handle protocols that don't use "://"
1709                 global $wgUrlProtocols;
1710                 $repl2 = [];
1711                 foreach ( $wgUrlProtocols as $prot ) {
1712                         if ( substr( $prot, -1 ) === ':' ) {
1713                                 $repl2[] = preg_quote( substr( $prot, 0, -1 ), '/' );
1714                         }
1715                 }
1716                 $repl2 = $repl2 ? '/\b(' . implode( '|', $repl2 ) . '):/i' : '/^(?!)/';
1717         }
1718         $text = substr( strtr( "\n$text", $repl ), 1 );
1719         $text = preg_replace( $repl2, '$1&#58;', $text );
1720         return $text;
1721 }
1722
1723 /**
1724  * Sets dest to source and returns the original value of dest
1725  * If source is NULL, it just returns the value, it doesn't set the variable
1726  * If force is true, it will set the value even if source is NULL
1727  *
1728  * @param mixed $dest
1729  * @param mixed $source
1730  * @param bool $force
1731  * @return mixed
1732  */
1733 function wfSetVar( &$dest, $source, $force = false ) {
1734         $temp = $dest;
1735         if ( !is_null( $source ) || $force ) {
1736                 $dest = $source;
1737         }
1738         return $temp;
1739 }
1740
1741 /**
1742  * As for wfSetVar except setting a bit
1743  *
1744  * @param int $dest
1745  * @param int $bit
1746  * @param bool $state
1747  *
1748  * @return bool
1749  */
1750 function wfSetBit( &$dest, $bit, $state = true ) {
1751         $temp = (bool)( $dest & $bit );
1752         if ( !is_null( $state ) ) {
1753                 if ( $state ) {
1754                         $dest |= $bit;
1755                 } else {
1756                         $dest &= ~$bit;
1757                 }
1758         }
1759         return $temp;
1760 }
1761
1762 /**
1763  * A wrapper around the PHP function var_export().
1764  * Either print it or add it to the regular output ($wgOut).
1765  *
1766  * @param mixed $var A PHP variable to dump.
1767  */
1768 function wfVarDump( $var ) {
1769         global $wgOut;
1770         $s = str_replace( "\n", "<br />\n", var_export( $var, true ) . "\n" );
1771         if ( headers_sent() || !isset( $wgOut ) || !is_object( $wgOut ) ) {
1772                 print $s;
1773         } else {
1774                 $wgOut->addHTML( $s );
1775         }
1776 }
1777
1778 /**
1779  * Provide a simple HTTP error.
1780  *
1781  * @param int|string $code
1782  * @param string $label
1783  * @param string $desc
1784  */
1785 function wfHttpError( $code, $label, $desc ) {
1786         global $wgOut;
1787         HttpStatus::header( $code );
1788         if ( $wgOut ) {
1789                 $wgOut->disable();
1790                 $wgOut->sendCacheControl();
1791         }
1792
1793         MediaWiki\HeaderCallback::warnIfHeadersSent();
1794         header( 'Content-type: text/html; charset=utf-8' );
1795         print '<!DOCTYPE html>' .
1796                 '<html><head><title>' .
1797                 htmlspecialchars( $label ) .
1798                 '</title></head><body><h1>' .
1799                 htmlspecialchars( $label ) .
1800                 '</h1><p>' .
1801                 nl2br( htmlspecialchars( $desc ) ) .
1802                 "</p></body></html>\n";
1803 }
1804
1805 /**
1806  * Clear away any user-level output buffers, discarding contents.
1807  *
1808  * Suitable for 'starting afresh', for instance when streaming
1809  * relatively large amounts of data without buffering, or wanting to
1810  * output image files without ob_gzhandler's compression.
1811  *
1812  * The optional $resetGzipEncoding parameter controls suppression of
1813  * the Content-Encoding header sent by ob_gzhandler; by default it
1814  * is left. See comments for wfClearOutputBuffers() for why it would
1815  * be used.
1816  *
1817  * Note that some PHP configuration options may add output buffer
1818  * layers which cannot be removed; these are left in place.
1819  *
1820  * @param bool $resetGzipEncoding
1821  */
1822 function wfResetOutputBuffers( $resetGzipEncoding = true ) {
1823         if ( $resetGzipEncoding ) {
1824                 // Suppress Content-Encoding and Content-Length
1825                 // headers from 1.10+s wfOutputHandler
1826                 global $wgDisableOutputCompression;
1827                 $wgDisableOutputCompression = true;
1828         }
1829         while ( $status = ob_get_status() ) {
1830                 if ( isset( $status['flags'] ) ) {
1831                         $flags = PHP_OUTPUT_HANDLER_CLEANABLE | PHP_OUTPUT_HANDLER_REMOVABLE;
1832                         $deleteable = ( $status['flags'] & $flags ) === $flags;
1833                 } elseif ( isset( $status['del'] ) ) {
1834                         $deleteable = $status['del'];
1835                 } else {
1836                         // Guess that any PHP-internal setting can't be removed.
1837                         $deleteable = $status['type'] !== 0; /* PHP_OUTPUT_HANDLER_INTERNAL */
1838                 }
1839                 if ( !$deleteable ) {
1840                         // Give up, and hope the result doesn't break
1841                         // output behavior.
1842                         break;
1843                 }
1844                 if ( $status['name'] === 'MediaWikiTestCase::wfResetOutputBuffersBarrier' ) {
1845                         // Unit testing barrier to prevent this function from breaking PHPUnit.
1846                         break;
1847                 }
1848                 if ( !ob_end_clean() ) {
1849                         // Could not remove output buffer handler; abort now
1850                         // to avoid getting in some kind of infinite loop.
1851                         break;
1852                 }
1853                 if ( $resetGzipEncoding ) {
1854                         if ( $status['name'] == 'ob_gzhandler' ) {
1855                                 // Reset the 'Content-Encoding' field set by this handler
1856                                 // so we can start fresh.
1857                                 header_remove( 'Content-Encoding' );
1858                                 break;
1859                         }
1860                 }
1861         }
1862 }
1863
1864 /**
1865  * More legible than passing a 'false' parameter to wfResetOutputBuffers():
1866  *
1867  * Clear away output buffers, but keep the Content-Encoding header
1868  * produced by ob_gzhandler, if any.
1869  *
1870  * This should be used for HTTP 304 responses, where you need to
1871  * preserve the Content-Encoding header of the real result, but
1872  * also need to suppress the output of ob_gzhandler to keep to spec
1873  * and avoid breaking Firefox in rare cases where the headers and
1874  * body are broken over two packets.
1875  */
1876 function wfClearOutputBuffers() {
1877         wfResetOutputBuffers( false );
1878 }
1879
1880 /**
1881  * Converts an Accept-* header into an array mapping string values to quality
1882  * factors
1883  *
1884  * @param string $accept
1885  * @param string $def Default
1886  * @return float[] Associative array of string => float pairs
1887  */
1888 function wfAcceptToPrefs( $accept, $def = '*/*' ) {
1889         # No arg means accept anything (per HTTP spec)
1890         if ( !$accept ) {
1891                 return [ $def => 1.0 ];
1892         }
1893
1894         $prefs = [];
1895
1896         $parts = explode( ',', $accept );
1897
1898         foreach ( $parts as $part ) {
1899                 # @todo FIXME: Doesn't deal with params like 'text/html; level=1'
1900                 $values = explode( ';', trim( $part ) );
1901                 $match = [];
1902                 if ( count( $values ) == 1 ) {
1903                         $prefs[$values[0]] = 1.0;
1904                 } elseif ( preg_match( '/q\s*=\s*(\d*\.\d+)/', $values[1], $match ) ) {
1905                         $prefs[$values[0]] = floatval( $match[1] );
1906                 }
1907         }
1908
1909         return $prefs;
1910 }
1911
1912 /**
1913  * Checks if a given MIME type matches any of the keys in the given
1914  * array. Basic wildcards are accepted in the array keys.
1915  *
1916  * Returns the matching MIME type (or wildcard) if a match, otherwise
1917  * NULL if no match.
1918  *
1919  * @param string $type
1920  * @param array $avail
1921  * @return string
1922  * @private
1923  */
1924 function mimeTypeMatch( $type, $avail ) {
1925         if ( array_key_exists( $type, $avail ) ) {
1926                 return $type;
1927         } else {
1928                 $mainType = explode( '/', $type )[0];
1929                 if ( array_key_exists( "$mainType/*", $avail ) ) {
1930                         return "$mainType/*";
1931                 } elseif ( array_key_exists( '*/*', $avail ) ) {
1932                         return '*/*';
1933                 } else {
1934                         return null;
1935                 }
1936         }
1937 }
1938
1939 /**
1940  * Returns the 'best' match between a client's requested internet media types
1941  * and the server's list of available types. Each list should be an associative
1942  * array of type to preference (preference is a float between 0.0 and 1.0).
1943  * Wildcards in the types are acceptable.
1944  *
1945  * @param array $cprefs Client's acceptable type list
1946  * @param array $sprefs Server's offered types
1947  * @return string
1948  *
1949  * @todo FIXME: Doesn't handle params like 'text/plain; charset=UTF-8'
1950  * XXX: generalize to negotiate other stuff
1951  */
1952 function wfNegotiateType( $cprefs, $sprefs ) {
1953         $combine = [];
1954
1955         foreach ( array_keys( $sprefs ) as $type ) {
1956                 $subType = explode( '/', $type )[1];
1957                 if ( $subType != '*' ) {
1958                         $ckey = mimeTypeMatch( $type, $cprefs );
1959                         if ( $ckey ) {
1960                                 $combine[$type] = $sprefs[$type] * $cprefs[$ckey];
1961                         }
1962                 }
1963         }
1964
1965         foreach ( array_keys( $cprefs ) as $type ) {
1966                 $subType = explode( '/', $type )[1];
1967                 if ( $subType != '*' && !array_key_exists( $type, $sprefs ) ) {
1968                         $skey = mimeTypeMatch( $type, $sprefs );
1969                         if ( $skey ) {
1970                                 $combine[$type] = $sprefs[$skey] * $cprefs[$type];
1971                         }
1972                 }
1973         }
1974
1975         $bestq = 0;
1976         $besttype = null;
1977
1978         foreach ( array_keys( $combine ) as $type ) {
1979                 if ( $combine[$type] > $bestq ) {
1980                         $besttype = $type;
1981                         $bestq = $combine[$type];
1982                 }
1983         }
1984
1985         return $besttype;
1986 }
1987
1988 /**
1989  * Reference-counted warning suppression
1990  *
1991  * @deprecated since 1.26, use MediaWiki\suppressWarnings() directly
1992  * @param bool $end
1993  */
1994 function wfSuppressWarnings( $end = false ) {
1995         MediaWiki\suppressWarnings( $end );
1996 }
1997
1998 /**
1999  * @deprecated since 1.26, use MediaWiki\restoreWarnings() directly
2000  * Restore error level to previous value
2001  */
2002 function wfRestoreWarnings() {
2003         MediaWiki\suppressWarnings( true );
2004 }
2005
2006 /**
2007  * Get a timestamp string in one of various formats
2008  *
2009  * @param mixed $outputtype A timestamp in one of the supported formats, the
2010  *   function will autodetect which format is supplied and act accordingly.
2011  * @param mixed $ts Optional timestamp to convert, default 0 for the current time
2012  * @return string|bool String / false The same date in the format specified in $outputtype or false
2013  */
2014 function wfTimestamp( $outputtype = TS_UNIX, $ts = 0 ) {
2015         $ret = MWTimestamp::convert( $outputtype, $ts );
2016         if ( $ret === false ) {
2017                 wfDebug( "wfTimestamp() fed bogus time value: TYPE=$outputtype; VALUE=$ts\n" );
2018         }
2019         return $ret;
2020 }
2021
2022 /**
2023  * Return a formatted timestamp, or null if input is null.
2024  * For dealing with nullable timestamp columns in the database.
2025  *
2026  * @param int $outputtype
2027  * @param string $ts
2028  * @return string
2029  */
2030 function wfTimestampOrNull( $outputtype = TS_UNIX, $ts = null ) {
2031         if ( is_null( $ts ) ) {
2032                 return null;
2033         } else {
2034                 return wfTimestamp( $outputtype, $ts );
2035         }
2036 }
2037
2038 /**
2039  * Convenience function; returns MediaWiki timestamp for the present time.
2040  *
2041  * @return string
2042  */
2043 function wfTimestampNow() {
2044         # return NOW
2045         return MWTimestamp::now( TS_MW );
2046 }
2047
2048 /**
2049  * Check if the operating system is Windows
2050  *
2051  * @return bool True if it's Windows, false otherwise.
2052  */
2053 function wfIsWindows() {
2054         static $isWindows = null;
2055         if ( $isWindows === null ) {
2056                 $isWindows = strtoupper( substr( PHP_OS, 0, 3 ) ) === 'WIN';
2057         }
2058         return $isWindows;
2059 }
2060
2061 /**
2062  * Check if we are running under HHVM
2063  *
2064  * @return bool
2065  */
2066 function wfIsHHVM() {
2067         return defined( 'HHVM_VERSION' );
2068 }
2069
2070 /**
2071  * Tries to get the system directory for temporary files. First
2072  * $wgTmpDirectory is checked, and then the TMPDIR, TMP, and TEMP
2073  * environment variables are then checked in sequence, then
2074  * sys_get_temp_dir(), then upload_tmp_dir from php.ini.
2075  *
2076  * NOTE: When possible, use instead the tmpfile() function to create
2077  * temporary files to avoid race conditions on file creation, etc.
2078  *
2079  * @return string
2080  */
2081 function wfTempDir() {
2082         global $wgTmpDirectory;
2083
2084         if ( $wgTmpDirectory !== false ) {
2085                 return $wgTmpDirectory;
2086         }
2087
2088         return TempFSFile::getUsableTempDirectory();
2089 }
2090
2091 /**
2092  * Make directory, and make all parent directories if they don't exist
2093  *
2094  * @param string $dir Full path to directory to create
2095  * @param int $mode Chmod value to use, default is $wgDirectoryMode
2096  * @param string $caller Optional caller param for debugging.
2097  * @throws MWException
2098  * @return bool
2099  */
2100 function wfMkdirParents( $dir, $mode = null, $caller = null ) {
2101         global $wgDirectoryMode;
2102
2103         if ( FileBackend::isStoragePath( $dir ) ) { // sanity
2104                 throw new MWException( __FUNCTION__ . " given storage path '$dir'." );
2105         }
2106
2107         if ( !is_null( $caller ) ) {
2108                 wfDebug( "$caller: called wfMkdirParents($dir)\n" );
2109         }
2110
2111         if ( strval( $dir ) === '' || is_dir( $dir ) ) {
2112                 return true;
2113         }
2114
2115         $dir = str_replace( [ '\\', '/' ], DIRECTORY_SEPARATOR, $dir );
2116
2117         if ( is_null( $mode ) ) {
2118                 $mode = $wgDirectoryMode;
2119         }
2120
2121         // Turn off the normal warning, we're doing our own below
2122         MediaWiki\suppressWarnings();
2123         $ok = mkdir( $dir, $mode, true ); // PHP5 <3
2124         MediaWiki\restoreWarnings();
2125
2126         if ( !$ok ) {
2127                 // directory may have been created on another request since we last checked
2128                 if ( is_dir( $dir ) ) {
2129                         return true;
2130                 }
2131
2132                 // PHP doesn't report the path in its warning message, so add our own to aid in diagnosis.
2133                 wfLogWarning( sprintf( "failed to mkdir \"%s\" mode 0%o", $dir, $mode ) );
2134         }
2135         return $ok;
2136 }
2137
2138 /**
2139  * Remove a directory and all its content.
2140  * Does not hide error.
2141  * @param string $dir
2142  */
2143 function wfRecursiveRemoveDir( $dir ) {
2144         wfDebug( __FUNCTION__ . "( $dir )\n" );
2145         // taken from https://secure.php.net/manual/en/function.rmdir.php#98622
2146         if ( is_dir( $dir ) ) {
2147                 $objects = scandir( $dir );
2148                 foreach ( $objects as $object ) {
2149                         if ( $object != "." && $object != ".." ) {
2150                                 if ( filetype( $dir . '/' . $object ) == "dir" ) {
2151                                         wfRecursiveRemoveDir( $dir . '/' . $object );
2152                                 } else {
2153                                         unlink( $dir . '/' . $object );
2154                                 }
2155                         }
2156                 }
2157                 reset( $objects );
2158                 rmdir( $dir );
2159         }
2160 }
2161
2162 /**
2163  * @param int $nr The number to format
2164  * @param int $acc The number of digits after the decimal point, default 2
2165  * @param bool $round Whether or not to round the value, default true
2166  * @return string
2167  */
2168 function wfPercent( $nr, $acc = 2, $round = true ) {
2169         $ret = sprintf( "%.${acc}f", $nr );
2170         return $round ? round( $ret, $acc ) . '%' : "$ret%";
2171 }
2172
2173 /**
2174  * Safety wrapper around ini_get() for boolean settings.
2175  * The values returned from ini_get() are pre-normalized for settings
2176  * set via php.ini or php_flag/php_admin_flag... but *not*
2177  * for those set via php_value/php_admin_value.
2178  *
2179  * It's fairly common for people to use php_value instead of php_flag,
2180  * which can leave you with an 'off' setting giving a false positive
2181  * for code that just takes the ini_get() return value as a boolean.
2182  *
2183  * To make things extra interesting, setting via php_value accepts
2184  * "true" and "yes" as true, but php.ini and php_flag consider them false. :)
2185  * Unrecognized values go false... again opposite PHP's own coercion
2186  * from string to bool.
2187  *
2188  * Luckily, 'properly' set settings will always come back as '0' or '1',
2189  * so we only have to worry about them and the 'improper' settings.
2190  *
2191  * I frickin' hate PHP... :P
2192  *
2193  * @param string $setting
2194  * @return bool
2195  */
2196 function wfIniGetBool( $setting ) {
2197         $val = strtolower( ini_get( $setting ) );
2198         // 'on' and 'true' can't have whitespace around them, but '1' can.
2199         return $val == 'on'
2200                 || $val == 'true'
2201                 || $val == 'yes'
2202                 || preg_match( "/^\s*[+-]?0*[1-9]/", $val ); // approx C atoi() function
2203 }
2204
2205 /**
2206  * Version of escapeshellarg() that works better on Windows.
2207  *
2208  * Originally, this fixed the incorrect use of single quotes on Windows
2209  * (https://bugs.php.net/bug.php?id=26285) and the locale problems on Linux in
2210  * PHP 5.2.6+ (bug backported to earlier distro releases of PHP).
2211  *
2212  * @param string ... strings to escape and glue together, or a single array of strings parameter
2213  * @return string
2214  */
2215 function wfEscapeShellArg( /*...*/ ) {
2216         wfInitShellLocale();
2217
2218         $args = func_get_args();
2219         if ( count( $args ) === 1 && is_array( reset( $args ) ) ) {
2220                 // If only one argument has been passed, and that argument is an array,
2221                 // treat it as a list of arguments
2222                 $args = reset( $args );
2223         }
2224
2225         $first = true;
2226         $retVal = '';
2227         foreach ( $args as $arg ) {
2228                 if ( !$first ) {
2229                         $retVal .= ' ';
2230                 } else {
2231                         $first = false;
2232                 }
2233
2234                 if ( wfIsWindows() ) {
2235                         // Escaping for an MSVC-style command line parser and CMD.EXE
2236                         // @codingStandardsIgnoreStart For long URLs
2237                         // Refs:
2238                         //  * https://web.archive.org/web/20020708081031/http://mailman.lyra.org/pipermail/scite-interest/2002-March/000436.html
2239                         //  * https://technet.microsoft.com/en-us/library/cc723564.aspx
2240                         //  * T15518
2241                         //  * CR r63214
2242                         // Double the backslashes before any double quotes. Escape the double quotes.
2243                         // @codingStandardsIgnoreEnd
2244                         $tokens = preg_split( '/(\\\\*")/', $arg, -1, PREG_SPLIT_DELIM_CAPTURE );
2245                         $arg = '';
2246                         $iteration = 0;
2247                         foreach ( $tokens as $token ) {
2248                                 if ( $iteration % 2 == 1 ) {
2249                                         // Delimiter, a double quote preceded by zero or more slashes
2250                                         $arg .= str_replace( '\\', '\\\\', substr( $token, 0, -1 ) ) . '\\"';
2251                                 } elseif ( $iteration % 4 == 2 ) {
2252                                         // ^ in $token will be outside quotes, need to be escaped
2253                                         $arg .= str_replace( '^', '^^', $token );
2254                                 } else { // $iteration % 4 == 0
2255                                         // ^ in $token will appear inside double quotes, so leave as is
2256                                         $arg .= $token;
2257                                 }
2258                                 $iteration++;
2259                         }
2260                         // Double the backslashes before the end of the string, because
2261                         // we will soon add a quote
2262                         $m = [];
2263                         if ( preg_match( '/^(.*?)(\\\\+)$/', $arg, $m ) ) {
2264                                 $arg = $m[1] . str_replace( '\\', '\\\\', $m[2] );
2265                         }
2266
2267                         // Add surrounding quotes
2268                         $retVal .= '"' . $arg . '"';
2269                 } else {
2270                         $retVal .= escapeshellarg( $arg );
2271                 }
2272         }
2273         return $retVal;
2274 }
2275
2276 /**
2277  * Check if wfShellExec() is effectively disabled via php.ini config
2278  *
2279  * @return bool|string False or 'disabled'
2280  * @since 1.22
2281  */
2282 function wfShellExecDisabled() {
2283         static $disabled = null;
2284         if ( is_null( $disabled ) ) {
2285                 if ( !function_exists( 'proc_open' ) ) {
2286                         wfDebug( "proc_open() is disabled\n" );
2287                         $disabled = 'disabled';
2288                 } else {
2289                         $disabled = false;
2290                 }
2291         }
2292         return $disabled;
2293 }
2294
2295 /**
2296  * Execute a shell command, with time and memory limits mirrored from the PHP
2297  * configuration if supported.
2298  *
2299  * @param string|string[] $cmd If string, a properly shell-escaped command line,
2300  *   or an array of unescaped arguments, in which case each value will be escaped
2301  *   Example:   [ 'convert', '-font', 'font name' ] would produce "'convert' '-font' 'font name'"
2302  * @param null|mixed &$retval Optional, will receive the program's exit code.
2303  *   (non-zero is usually failure). If there is an error from
2304  *   read, select, or proc_open(), this will be set to -1.
2305  * @param array $environ Optional environment variables which should be
2306  *   added to the executed command environment.
2307  * @param array $limits Optional array with limits(filesize, memory, time, walltime)
2308  *   this overwrites the global wgMaxShell* limits.
2309  * @param array $options Array of options:
2310  *   - duplicateStderr: Set this to true to duplicate stderr to stdout,
2311  *     including errors from limit.sh
2312  *   - profileMethod: By default this function will profile based on the calling
2313  *     method. Set this to a string for an alternative method to profile from
2314  *
2315  * @return string Collected stdout as a string
2316  */
2317 function wfShellExec( $cmd, &$retval = null, $environ = [],
2318         $limits = [], $options = []
2319 ) {
2320         global $IP, $wgMaxShellMemory, $wgMaxShellFileSize, $wgMaxShellTime,
2321                 $wgMaxShellWallClockTime, $wgShellCgroup;
2322
2323         $disabled = wfShellExecDisabled();
2324         if ( $disabled ) {
2325                 $retval = 1;
2326                 return 'Unable to run external programs, proc_open() is disabled.';
2327         }
2328
2329         $includeStderr = isset( $options['duplicateStderr'] ) && $options['duplicateStderr'];
2330         $profileMethod = isset( $options['profileMethod'] ) ? $options['profileMethod'] : wfGetCaller();
2331
2332         wfInitShellLocale();
2333
2334         $envcmd = '';
2335         foreach ( $environ as $k => $v ) {
2336                 if ( wfIsWindows() ) {
2337                         /* Surrounding a set in quotes (method used by wfEscapeShellArg) makes the quotes themselves
2338                          * appear in the environment variable, so we must use carat escaping as documented in
2339                          * https://technet.microsoft.com/en-us/library/cc723564.aspx
2340                          * Note however that the quote isn't listed there, but is needed, and the parentheses
2341                          * are listed there but doesn't appear to need it.
2342                          */
2343                         $envcmd .= "set $k=" . preg_replace( '/([&|()<>^"])/', '^\\1', $v ) . '&& ';
2344                 } else {
2345                         /* Assume this is a POSIX shell, thus required to accept variable assignments before the command
2346                          * http://www.opengroup.org/onlinepubs/009695399/utilities/xcu_chap02.html#tag_02_09_01
2347                          */
2348                         $envcmd .= "$k=" . escapeshellarg( $v ) . ' ';
2349                 }
2350         }
2351         if ( is_array( $cmd ) ) {
2352                 $cmd = wfEscapeShellArg( $cmd );
2353         }
2354
2355         $cmd = $envcmd . $cmd;
2356
2357         $useLogPipe = false;
2358         if ( is_executable( '/bin/bash' ) ) {
2359                 $time = intval( isset( $limits['time'] ) ? $limits['time'] : $wgMaxShellTime );
2360                 if ( isset( $limits['walltime'] ) ) {
2361                         $wallTime = intval( $limits['walltime'] );
2362                 } elseif ( isset( $limits['time'] ) ) {
2363                         $wallTime = $time;
2364                 } else {
2365                         $wallTime = intval( $wgMaxShellWallClockTime );
2366                 }
2367                 $mem = intval( isset( $limits['memory'] ) ? $limits['memory'] : $wgMaxShellMemory );
2368                 $filesize = intval( isset( $limits['filesize'] ) ? $limits['filesize'] : $wgMaxShellFileSize );
2369
2370                 if ( $time > 0 || $mem > 0 || $filesize > 0 || $wallTime > 0 ) {
2371                         $cmd = '/bin/bash ' . escapeshellarg( "$IP/includes/limit.sh" ) . ' ' .
2372                                 escapeshellarg( $cmd ) . ' ' .
2373                                 escapeshellarg(
2374                                         "MW_INCLUDE_STDERR=" . ( $includeStderr ? '1' : '' ) . ';' .
2375                                         "MW_CPU_LIMIT=$time; " .
2376                                         'MW_CGROUP=' . escapeshellarg( $wgShellCgroup ) . '; ' .
2377                                         "MW_MEM_LIMIT=$mem; " .
2378                                         "MW_FILE_SIZE_LIMIT=$filesize; " .
2379                                         "MW_WALL_CLOCK_LIMIT=$wallTime; " .
2380                                         "MW_USE_LOG_PIPE=yes"
2381                                 );
2382                         $useLogPipe = true;
2383                 } elseif ( $includeStderr ) {
2384                         $cmd .= ' 2>&1';
2385                 }
2386         } elseif ( $includeStderr ) {
2387                 $cmd .= ' 2>&1';
2388         }
2389         wfDebug( "wfShellExec: $cmd\n" );
2390
2391         // Don't try to execute commands that exceed Linux's MAX_ARG_STRLEN.
2392         // Other platforms may be more accomodating, but we don't want to be
2393         // accomodating, because very long commands probably include user
2394         // input. See T129506.
2395         if ( strlen( $cmd ) > SHELL_MAX_ARG_STRLEN ) {
2396                 throw new Exception( __METHOD__ .
2397                         '(): total length of $cmd must not exceed SHELL_MAX_ARG_STRLEN' );
2398         }
2399
2400         $desc = [
2401                 0 => [ 'file', 'php://stdin', 'r' ],
2402                 1 => [ 'pipe', 'w' ],
2403                 2 => [ 'file', 'php://stderr', 'w' ] ];
2404         if ( $useLogPipe ) {
2405                 $desc[3] = [ 'pipe', 'w' ];
2406         }
2407         $pipes = null;
2408         $scoped = Profiler::instance()->scopedProfileIn( __FUNCTION__ . '-' . $profileMethod );
2409         $proc = proc_open( $cmd, $desc, $pipes );
2410         if ( !$proc ) {
2411                 wfDebugLog( 'exec', "proc_open() failed: $cmd" );
2412                 $retval = -1;
2413                 return '';
2414         }
2415         $outBuffer = $logBuffer = '';
2416         $emptyArray = [];
2417         $status = false;
2418         $logMsg = false;
2419
2420         /* According to the documentation, it is possible for stream_select()
2421          * to fail due to EINTR. I haven't managed to induce this in testing
2422          * despite sending various signals. If it did happen, the error
2423          * message would take the form:
2424          *
2425          * stream_select(): unable to select [4]: Interrupted system call (max_fd=5)
2426          *
2427          * where [4] is the value of the macro EINTR and "Interrupted system
2428          * call" is string which according to the Linux manual is "possibly"
2429          * localised according to LC_MESSAGES.
2430          */
2431         $eintr = defined( 'SOCKET_EINTR' ) ? SOCKET_EINTR : 4;
2432         $eintrMessage = "stream_select(): unable to select [$eintr]";
2433
2434         $running = true;
2435         $timeout = null;
2436         $numReadyPipes = 0;
2437
2438         while ( $running === true || $numReadyPipes !== 0 ) {
2439                 if ( $running ) {
2440                         $status = proc_get_status( $proc );
2441                         // If the process has terminated, switch to nonblocking selects
2442                         // for getting any data still waiting to be read.
2443                         if ( !$status['running'] ) {
2444                                 $running = false;
2445                                 $timeout = 0;
2446                         }
2447                 }
2448
2449                 $readyPipes = $pipes;
2450
2451                 // Clear last error
2452                 // @codingStandardsIgnoreStart Generic.PHP.NoSilencedErrors.Discouraged
2453                 @trigger_error( '' );
2454                 $numReadyPipes = @stream_select( $readyPipes, $emptyArray, $emptyArray, $timeout );
2455                 if ( $numReadyPipes === false ) {
2456                         // @codingStandardsIgnoreEnd
2457                         $error = error_get_last();
2458                         if ( strncmp( $error['message'], $eintrMessage, strlen( $eintrMessage ) ) == 0 ) {
2459                                 continue;
2460                         } else {
2461                                 trigger_error( $error['message'], E_USER_WARNING );
2462                                 $logMsg = $error['message'];
2463                                 break;
2464                         }
2465                 }
2466                 foreach ( $readyPipes as $fd => $pipe ) {
2467                         $block = fread( $pipe, 65536 );
2468                         if ( $block === '' ) {
2469                                 // End of file
2470                                 fclose( $pipes[$fd] );
2471                                 unset( $pipes[$fd] );
2472                                 if ( !$pipes ) {
2473                                         break 2;
2474                                 }
2475                         } elseif ( $block === false ) {
2476                                 // Read error
2477                                 $logMsg = "Error reading from pipe";
2478                                 break 2;
2479                         } elseif ( $fd == 1 ) {
2480                                 // From stdout
2481                                 $outBuffer .= $block;
2482                         } elseif ( $fd == 3 ) {
2483                                 // From log FD
2484                                 $logBuffer .= $block;
2485                                 if ( strpos( $block, "\n" ) !== false ) {
2486                                         $lines = explode( "\n", $logBuffer );
2487                                         $logBuffer = array_pop( $lines );
2488                                         foreach ( $lines as $line ) {
2489                                                 wfDebugLog( 'exec', $line );
2490                                         }
2491                                 }
2492                         }
2493                 }
2494         }
2495
2496         foreach ( $pipes as $pipe ) {
2497                 fclose( $pipe );
2498         }
2499
2500         // Use the status previously collected if possible, since proc_get_status()
2501         // just calls waitpid() which will not return anything useful the second time.
2502         if ( $running ) {
2503                 $status = proc_get_status( $proc );
2504         }
2505
2506         if ( $logMsg !== false ) {
2507                 // Read/select error
2508                 $retval = -1;
2509                 proc_close( $proc );
2510         } elseif ( $status['signaled'] ) {
2511                 $logMsg = "Exited with signal {$status['termsig']}";
2512                 $retval = 128 + $status['termsig'];
2513                 proc_close( $proc );
2514         } else {
2515                 if ( $status['running'] ) {
2516                         $retval = proc_close( $proc );
2517                 } else {
2518                         $retval = $status['exitcode'];
2519                         proc_close( $proc );
2520                 }
2521                 if ( $retval == 127 ) {
2522                         $logMsg = "Possibly missing executable file";
2523                 } elseif ( $retval >= 129 && $retval <= 192 ) {
2524                         $logMsg = "Probably exited with signal " . ( $retval - 128 );
2525                 }
2526         }
2527
2528         if ( $logMsg !== false ) {
2529                 wfDebugLog( 'exec', "$logMsg: $cmd" );
2530         }
2531
2532         return $outBuffer;
2533 }
2534
2535 /**
2536  * Execute a shell command, returning both stdout and stderr. Convenience
2537  * function, as all the arguments to wfShellExec can become unwieldy.
2538  *
2539  * @note This also includes errors from limit.sh, e.g. if $wgMaxShellFileSize is exceeded.
2540  * @param string|string[] $cmd If string, a properly shell-escaped command line,
2541  *   or an array of unescaped arguments, in which case each value will be escaped
2542  *   Example:   [ 'convert', '-font', 'font name' ] would produce "'convert' '-font' 'font name'"
2543  * @param null|mixed &$retval Optional, will receive the program's exit code.
2544  *   (non-zero is usually failure)
2545  * @param array $environ Optional environment variables which should be
2546  *   added to the executed command environment.
2547  * @param array $limits Optional array with limits(filesize, memory, time, walltime)
2548  *   this overwrites the global wgMaxShell* limits.
2549  * @return string Collected stdout and stderr as a string
2550  */
2551 function wfShellExecWithStderr( $cmd, &$retval = null, $environ = [], $limits = [] ) {
2552         return wfShellExec( $cmd, $retval, $environ, $limits,
2553                 [ 'duplicateStderr' => true, 'profileMethod' => wfGetCaller() ] );
2554 }
2555
2556 /**
2557  * Workaround for https://bugs.php.net/bug.php?id=45132
2558  * escapeshellarg() destroys non-ASCII characters if LANG is not a UTF-8 locale
2559  */
2560 function wfInitShellLocale() {
2561         static $done = false;
2562         if ( $done ) {
2563                 return;
2564         }
2565         $done = true;
2566         global $wgShellLocale;
2567         putenv( "LC_CTYPE=$wgShellLocale" );
2568         setlocale( LC_CTYPE, $wgShellLocale );
2569 }
2570
2571 /**
2572  * Generate a shell-escaped command line string to run a MediaWiki cli script.
2573  * Note that $parameters should be a flat array and an option with an argument
2574  * should consist of two consecutive items in the array (do not use "--option value").
2575  *
2576  * @param string $script MediaWiki cli script path
2577  * @param array $parameters Arguments and options to the script
2578  * @param array $options Associative array of options:
2579  *     'php': The path to the php executable
2580  *     'wrapper': Path to a PHP wrapper to handle the maintenance script
2581  * @return string
2582  */
2583 function wfShellWikiCmd( $script, array $parameters = [], array $options = [] ) {
2584         global $wgPhpCli;
2585         // Give site config file a chance to run the script in a wrapper.
2586         // The caller may likely want to call wfBasename() on $script.
2587         Hooks::run( 'wfShellWikiCmd', [ &$script, &$parameters, &$options ] );
2588         $cmd = isset( $options['php'] ) ? [ $options['php'] ] : [ $wgPhpCli ];
2589         if ( isset( $options['wrapper'] ) ) {
2590                 $cmd[] = $options['wrapper'];
2591         }
2592         $cmd[] = $script;
2593         // Escape each parameter for shell
2594         return wfEscapeShellArg( array_merge( $cmd, $parameters ) );
2595 }
2596
2597 /**
2598  * wfMerge attempts to merge differences between three texts.
2599  * Returns true for a clean merge and false for failure or a conflict.
2600  *
2601  * @param string $old
2602  * @param string $mine
2603  * @param string $yours
2604  * @param string $result
2605  * @return bool
2606  */
2607 function wfMerge( $old, $mine, $yours, &$result ) {
2608         global $wgDiff3;
2609
2610         # This check may also protect against code injection in
2611         # case of broken installations.
2612         MediaWiki\suppressWarnings();
2613         $haveDiff3 = $wgDiff3 && file_exists( $wgDiff3 );
2614         MediaWiki\restoreWarnings();
2615
2616         if ( !$haveDiff3 ) {
2617                 wfDebug( "diff3 not found\n" );
2618                 return false;
2619         }
2620
2621         # Make temporary files
2622         $td = wfTempDir();
2623         $oldtextFile = fopen( $oldtextName = tempnam( $td, 'merge-old-' ), 'w' );
2624         $mytextFile = fopen( $mytextName = tempnam( $td, 'merge-mine-' ), 'w' );
2625         $yourtextFile = fopen( $yourtextName = tempnam( $td, 'merge-your-' ), 'w' );
2626
2627         # NOTE: diff3 issues a warning to stderr if any of the files does not end with
2628         #       a newline character. To avoid this, we normalize the trailing whitespace before
2629         #       creating the diff.
2630
2631         fwrite( $oldtextFile, rtrim( $old ) . "\n" );
2632         fclose( $oldtextFile );
2633         fwrite( $mytextFile, rtrim( $mine ) . "\n" );
2634         fclose( $mytextFile );
2635         fwrite( $yourtextFile, rtrim( $yours ) . "\n" );
2636         fclose( $yourtextFile );
2637
2638         # Check for a conflict
2639         $cmd = wfEscapeShellArg( $wgDiff3, '-a', '--overlap-only', $mytextName,
2640                 $oldtextName, $yourtextName );
2641         $handle = popen( $cmd, 'r' );
2642
2643         if ( fgets( $handle, 1024 ) ) {
2644                 $conflict = true;
2645         } else {
2646                 $conflict = false;
2647         }
2648         pclose( $handle );
2649
2650         # Merge differences
2651         $cmd = wfEscapeShellArg( $wgDiff3, '-a', '-e', '--merge', $mytextName,
2652                 $oldtextName, $yourtextName );
2653         $handle = popen( $cmd, 'r' );
2654         $result = '';
2655         do {
2656                 $data = fread( $handle, 8192 );
2657                 if ( strlen( $data ) == 0 ) {
2658                         break;
2659                 }
2660                 $result .= $data;
2661         } while ( true );
2662         pclose( $handle );
2663         unlink( $mytextName );
2664         unlink( $oldtextName );
2665         unlink( $yourtextName );
2666
2667         if ( $result === '' && $old !== '' && !$conflict ) {
2668                 wfDebug( "Unexpected null result from diff3. Command: $cmd\n" );
2669                 $conflict = true;
2670         }
2671         return !$conflict;
2672 }
2673
2674 /**
2675  * Returns unified plain-text diff of two texts.
2676  * "Useful" for machine processing of diffs.
2677  *
2678  * @deprecated since 1.25, use DiffEngine/UnifiedDiffFormatter directly
2679  *
2680  * @param string $before The text before the changes.
2681  * @param string $after The text after the changes.
2682  * @param string $params Command-line options for the diff command.
2683  * @return string Unified diff of $before and $after
2684  */
2685 function wfDiff( $before, $after, $params = '-u' ) {
2686         if ( $before == $after ) {
2687                 return '';
2688         }
2689
2690         global $wgDiff;
2691         MediaWiki\suppressWarnings();
2692         $haveDiff = $wgDiff && file_exists( $wgDiff );
2693         MediaWiki\restoreWarnings();
2694
2695         # This check may also protect against code injection in
2696         # case of broken installations.
2697         if ( !$haveDiff ) {
2698                 wfDebug( "diff executable not found\n" );
2699                 $diffs = new Diff( explode( "\n", $before ), explode( "\n", $after ) );
2700                 $format = new UnifiedDiffFormatter();
2701                 return $format->format( $diffs );
2702         }
2703
2704         # Make temporary files
2705         $td = wfTempDir();
2706         $oldtextFile = fopen( $oldtextName = tempnam( $td, 'merge-old-' ), 'w' );
2707         $newtextFile = fopen( $newtextName = tempnam( $td, 'merge-your-' ), 'w' );
2708
2709         fwrite( $oldtextFile, $before );
2710         fclose( $oldtextFile );
2711         fwrite( $newtextFile, $after );
2712         fclose( $newtextFile );
2713
2714         // Get the diff of the two files
2715         $cmd = "$wgDiff " . $params . ' ' . wfEscapeShellArg( $oldtextName, $newtextName );
2716
2717         $h = popen( $cmd, 'r' );
2718         if ( !$h ) {
2719                 unlink( $oldtextName );
2720                 unlink( $newtextName );
2721                 throw new Exception( __METHOD__ . '(): popen() failed' );
2722         }
2723
2724         $diff = '';
2725
2726         do {
2727                 $data = fread( $h, 8192 );
2728                 if ( strlen( $data ) == 0 ) {
2729                         break;
2730                 }
2731                 $diff .= $data;
2732         } while ( true );
2733
2734         // Clean up
2735         pclose( $h );
2736         unlink( $oldtextName );
2737         unlink( $newtextName );
2738
2739         // Kill the --- and +++ lines. They're not useful.
2740         $diff_lines = explode( "\n", $diff );
2741         if ( isset( $diff_lines[0] ) && strpos( $diff_lines[0], '---' ) === 0 ) {
2742                 unset( $diff_lines[0] );
2743         }
2744         if ( isset( $diff_lines[1] ) && strpos( $diff_lines[1], '+++' ) === 0 ) {
2745                 unset( $diff_lines[1] );
2746         }
2747
2748         $diff = implode( "\n", $diff_lines );
2749
2750         return $diff;
2751 }
2752
2753 /**
2754  * This function works like "use VERSION" in Perl, the program will die with a
2755  * backtrace if the current version of PHP is less than the version provided
2756  *
2757  * This is useful for extensions which due to their nature are not kept in sync
2758  * with releases, and might depend on other versions of PHP than the main code
2759  *
2760  * Note: PHP might die due to parsing errors in some cases before it ever
2761  *       manages to call this function, such is life
2762  *
2763  * @see perldoc -f use
2764  *
2765  * @param string|int|float $req_ver The version to check, can be a string, an integer, or a float
2766  * @throws MWException
2767  */
2768 function wfUsePHP( $req_ver ) {
2769         $php_ver = PHP_VERSION;
2770
2771         if ( version_compare( $php_ver, (string)$req_ver, '<' ) ) {
2772                 throw new MWException( "PHP $req_ver required--this is only $php_ver" );
2773         }
2774 }
2775
2776 /**
2777  * This function works like "use VERSION" in Perl except it checks the version
2778  * of MediaWiki, the program will die with a backtrace if the current version
2779  * of MediaWiki is less than the version provided.
2780  *
2781  * This is useful for extensions which due to their nature are not kept in sync
2782  * with releases
2783  *
2784  * Note: Due to the behavior of PHP's version_compare() which is used in this
2785  * function, if you want to allow the 'wmf' development versions add a 'c' (or
2786  * any single letter other than 'a', 'b' or 'p') as a post-fix to your
2787  * targeted version number. For example if you wanted to allow any variation
2788  * of 1.22 use `wfUseMW( '1.22c' )`. Using an 'a' or 'b' instead of 'c' will
2789  * not result in the same comparison due to the internal logic of
2790  * version_compare().
2791  *
2792  * @see perldoc -f use
2793  *
2794  * @deprecated since 1.26, use the "requires' property of extension.json
2795  * @param string|int|float $req_ver The version to check, can be a string, an integer, or a float
2796  * @throws MWException
2797  */
2798 function wfUseMW( $req_ver ) {
2799         global $wgVersion;
2800
2801         if ( version_compare( $wgVersion, (string)$req_ver, '<' ) ) {
2802                 throw new MWException( "MediaWiki $req_ver required--this is only $wgVersion" );
2803         }
2804 }
2805
2806 /**
2807  * Return the final portion of a pathname.
2808  * Reimplemented because PHP5's "basename()" is buggy with multibyte text.
2809  * https://bugs.php.net/bug.php?id=33898
2810  *
2811  * PHP's basename() only considers '\' a pathchar on Windows and Netware.
2812  * We'll consider it so always, as we don't want '\s' in our Unix paths either.
2813  *
2814  * @param string $path
2815  * @param string $suffix String to remove if present
2816  * @return string
2817  */
2818 function wfBaseName( $path, $suffix = '' ) {
2819         if ( $suffix == '' ) {
2820                 $encSuffix = '';
2821         } else {
2822                 $encSuffix = '(?:' . preg_quote( $suffix, '#' ) . ')?';
2823         }
2824
2825         $matches = [];
2826         if ( preg_match( "#([^/\\\\]*?){$encSuffix}[/\\\\]*$#", $path, $matches ) ) {
2827                 return $matches[1];
2828         } else {
2829                 return '';
2830         }
2831 }
2832
2833 /**
2834  * Generate a relative path name to the given file.
2835  * May explode on non-matching case-insensitive paths,
2836  * funky symlinks, etc.
2837  *
2838  * @param string $path Absolute destination path including target filename
2839  * @param string $from Absolute source path, directory only
2840  * @return string
2841  */
2842 function wfRelativePath( $path, $from ) {
2843         // Normalize mixed input on Windows...
2844         $path = str_replace( '/', DIRECTORY_SEPARATOR, $path );
2845         $from = str_replace( '/', DIRECTORY_SEPARATOR, $from );
2846
2847         // Trim trailing slashes -- fix for drive root
2848         $path = rtrim( $path, DIRECTORY_SEPARATOR );
2849         $from = rtrim( $from, DIRECTORY_SEPARATOR );
2850
2851         $pieces = explode( DIRECTORY_SEPARATOR, dirname( $path ) );
2852         $against = explode( DIRECTORY_SEPARATOR, $from );
2853
2854         if ( $pieces[0] !== $against[0] ) {
2855                 // Non-matching Windows drive letters?
2856                 // Return a full path.
2857                 return $path;
2858         }
2859
2860         // Trim off common prefix
2861         while ( count( $pieces ) && count( $against )
2862                 && $pieces[0] == $against[0] ) {
2863                 array_shift( $pieces );
2864                 array_shift( $against );
2865         }
2866
2867         // relative dots to bump us to the parent
2868         while ( count( $against ) ) {
2869                 array_unshift( $pieces, '..' );
2870                 array_shift( $against );
2871         }
2872
2873         array_push( $pieces, wfBaseName( $path ) );
2874
2875         return implode( DIRECTORY_SEPARATOR, $pieces );
2876 }
2877
2878 /**
2879  * Convert an arbitrarily-long digit string from one numeric base
2880  * to another, optionally zero-padding to a minimum column width.
2881  *
2882  * Supports base 2 through 36; digit values 10-36 are represented
2883  * as lowercase letters a-z. Input is case-insensitive.
2884  *
2885  * @deprecated since 1.27 Use Wikimedia\base_convert() directly
2886  *
2887  * @param string $input Input number
2888  * @param int $sourceBase Base of the input number
2889  * @param int $destBase Desired base of the output
2890  * @param int $pad Minimum number of digits in the output (pad with zeroes)
2891  * @param bool $lowercase Whether to output in lowercase or uppercase
2892  * @param string $engine Either "gmp", "bcmath", or "php"
2893  * @return string|bool The output number as a string, or false on error
2894  */
2895 function wfBaseConvert( $input, $sourceBase, $destBase, $pad = 1,
2896         $lowercase = true, $engine = 'auto'
2897 ) {
2898         return Wikimedia\base_convert( $input, $sourceBase, $destBase, $pad, $lowercase, $engine );
2899 }
2900
2901 /**
2902  * @deprecated since 1.27, PHP's session generation isn't used with
2903  *  MediaWiki\Session\SessionManager
2904  */
2905 function wfFixSessionID() {
2906         wfDeprecated( __FUNCTION__, '1.27' );
2907 }
2908
2909 /**
2910  * Reset the session id
2911  *
2912  * @deprecated since 1.27, use MediaWiki\Session\SessionManager instead
2913  * @since 1.22
2914  */
2915 function wfResetSessionID() {
2916         wfDeprecated( __FUNCTION__, '1.27' );
2917         $session = SessionManager::getGlobalSession();
2918         $delay = $session->delaySave();
2919
2920         $session->resetId();
2921
2922         // Make sure a session is started, since that's what the old
2923         // wfResetSessionID() did.
2924         if ( session_id() !== $session->getId() ) {
2925                 wfSetupSession( $session->getId() );
2926         }
2927
2928         ScopedCallback::consume( $delay );
2929 }
2930
2931 /**
2932  * Initialise php session
2933  *
2934  * @deprecated since 1.27, use MediaWiki\Session\SessionManager instead.
2935  *  Generally, "using" SessionManager will be calling ->getSessionById() or
2936  *  ::getGlobalSession() (depending on whether you were passing $sessionId
2937  *  here), then calling $session->persist().
2938  * @param bool|string $sessionId
2939  */
2940 function wfSetupSession( $sessionId = false ) {
2941         wfDeprecated( __FUNCTION__, '1.27' );
2942
2943         if ( $sessionId ) {
2944                 session_id( $sessionId );
2945         }
2946
2947         $session = SessionManager::getGlobalSession();
2948         $session->persist();
2949
2950         if ( session_id() !== $session->getId() ) {
2951                 session_id( $session->getId() );
2952         }
2953         MediaWiki\quietCall( 'session_start' );
2954 }
2955
2956 /**
2957  * Get an object from the precompiled serialized directory
2958  *
2959  * @param string $name
2960  * @return mixed The variable on success, false on failure
2961  */
2962 function wfGetPrecompiledData( $name ) {
2963         global $IP;
2964
2965         $file = "$IP/serialized/$name";
2966         if ( file_exists( $file ) ) {
2967                 $blob = file_get_contents( $file );
2968                 if ( $blob ) {
2969                         return unserialize( $blob );
2970                 }
2971         }
2972         return false;
2973 }
2974
2975 /**
2976  * Make a cache key for the local wiki.
2977  *
2978  * @param string $args,...
2979  * @return string
2980  */
2981 function wfMemcKey( /*...*/ ) {
2982         return call_user_func_array(
2983                 [ ObjectCache::getLocalClusterInstance(), 'makeKey' ],
2984                 func_get_args()
2985         );
2986 }
2987
2988 /**
2989  * Make a cache key for a foreign DB.
2990  *
2991  * Must match what wfMemcKey() would produce in context of the foreign wiki.
2992  *
2993  * @param string $db
2994  * @param string $prefix
2995  * @param string $args,...
2996  * @return string
2997  */
2998 function wfForeignMemcKey( $db, $prefix /*...*/ ) {
2999         $args = array_slice( func_get_args(), 2 );
3000         $keyspace = $prefix ? "$db-$prefix" : $db;
3001         return call_user_func_array(
3002                 [ ObjectCache::getLocalClusterInstance(), 'makeKeyInternal' ],
3003                 [ $keyspace, $args ]
3004         );
3005 }
3006
3007 /**
3008  * Make a cache key with database-agnostic prefix.
3009  *
3010  * Doesn't have a wiki-specific namespace. Uses a generic 'global' prefix
3011  * instead. Must have a prefix as otherwise keys that use a database name
3012  * in the first segment will clash with wfMemcKey/wfForeignMemcKey.
3013  *
3014  * @since 1.26
3015  * @param string $args,...
3016  * @return string
3017  */
3018 function wfGlobalCacheKey( /*...*/ ) {
3019         return call_user_func_array(
3020                 [ ObjectCache::getLocalClusterInstance(), 'makeGlobalKey' ],
3021                 func_get_args()
3022         );
3023 }
3024
3025 /**
3026  * Get an ASCII string identifying this wiki
3027  * This is used as a prefix in memcached keys
3028  *
3029  * @return string
3030  */
3031 function wfWikiID() {
3032         global $wgDBprefix, $wgDBname;
3033         if ( $wgDBprefix ) {
3034                 return "$wgDBname-$wgDBprefix";
3035         } else {
3036                 return $wgDBname;
3037         }
3038 }
3039
3040 /**
3041  * Split a wiki ID into DB name and table prefix
3042  *
3043  * @param string $wiki
3044  *
3045  * @return array
3046  */
3047 function wfSplitWikiID( $wiki ) {
3048         $bits = explode( '-', $wiki, 2 );
3049         if ( count( $bits ) < 2 ) {
3050                 $bits[] = '';
3051         }
3052         return $bits;
3053 }
3054
3055 /**
3056  * Get a Database object.
3057  *
3058  * @param int $db Index of the connection to get. May be DB_MASTER for the
3059  *            master (for write queries), DB_REPLICA for potentially lagged read
3060  *            queries, or an integer >= 0 for a particular server.
3061  *
3062  * @param string|string[] $groups Query groups. An array of group names that this query
3063  *                belongs to. May contain a single string if the query is only
3064  *                in one group.
3065  *
3066  * @param string|bool $wiki The wiki ID, or false for the current wiki
3067  *
3068  * Note: multiple calls to wfGetDB(DB_REPLICA) during the course of one request
3069  * will always return the same object, unless the underlying connection or load
3070  * balancer is manually destroyed.
3071  *
3072  * Note 2: use $this->getDB() in maintenance scripts that may be invoked by
3073  * updater to ensure that a proper database is being updated.
3074  *
3075  * @todo Replace calls to wfGetDB with calls to LoadBalancer::getConnection()
3076  *       on an injected instance of LoadBalancer.
3077  *
3078  * @return \Wikimedia\Rdbms\Database
3079  */
3080 function wfGetDB( $db, $groups = [], $wiki = false ) {
3081         return wfGetLB( $wiki )->getConnection( $db, $groups, $wiki );
3082 }
3083
3084 /**
3085  * Get a load balancer object.
3086  *
3087  * @deprecated since 1.27, use MediaWikiServices::getDBLoadBalancer()
3088  *              or MediaWikiServices::getDBLoadBalancerFactory() instead.
3089  *
3090  * @param string|bool $wiki Wiki ID, or false for the current wiki
3091  * @return \Wikimedia\Rdbms\LoadBalancer
3092  */
3093 function wfGetLB( $wiki = false ) {
3094         if ( $wiki === false ) {
3095                 return \MediaWiki\MediaWikiServices::getInstance()->getDBLoadBalancer();
3096         } else {
3097                 $factory = \MediaWiki\MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
3098                 return $factory->getMainLB( $wiki );
3099         }
3100 }
3101
3102 /**
3103  * Get the load balancer factory object
3104  *
3105  * @deprecated since 1.27, use MediaWikiServices::getDBLoadBalancerFactory() instead.
3106  *
3107  * @return \Wikimedia\Rdbms\LBFactory
3108  */
3109 function wfGetLBFactory() {
3110         return \MediaWiki\MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
3111 }
3112
3113 /**
3114  * Find a file.
3115  * Shortcut for RepoGroup::singleton()->findFile()
3116  *
3117  * @param string $title String or Title object
3118  * @param array $options Associative array of options (see RepoGroup::findFile)
3119  * @return File|bool File, or false if the file does not exist
3120  */
3121 function wfFindFile( $title, $options = [] ) {
3122         return RepoGroup::singleton()->findFile( $title, $options );
3123 }
3124
3125 /**
3126  * Get an object referring to a locally registered file.
3127  * Returns a valid placeholder object if the file does not exist.
3128  *
3129  * @param Title|string $title
3130  * @return LocalFile|null A File, or null if passed an invalid Title
3131  */
3132 function wfLocalFile( $title ) {
3133         return RepoGroup::singleton()->getLocalRepo()->newFile( $title );
3134 }
3135
3136 /**
3137  * Should low-performance queries be disabled?
3138  *
3139  * @return bool
3140  * @codeCoverageIgnore
3141  */
3142 function wfQueriesMustScale() {
3143         global $wgMiserMode;
3144         return $wgMiserMode
3145                 || ( SiteStats::pages() > 100000
3146                 && SiteStats::edits() > 1000000
3147                 && SiteStats::users() > 10000 );
3148 }
3149
3150 /**
3151  * Get the path to a specified script file, respecting file
3152  * extensions; this is a wrapper around $wgScriptPath etc.
3153  * except for 'index' and 'load' which use $wgScript/$wgLoadScript
3154  *
3155  * @param string $script Script filename, sans extension
3156  * @return string
3157  */
3158 function wfScript( $script = 'index' ) {
3159         global $wgScriptPath, $wgScript, $wgLoadScript;
3160         if ( $script === 'index' ) {
3161                 return $wgScript;
3162         } elseif ( $script === 'load' ) {
3163                 return $wgLoadScript;
3164         } else {
3165                 return "{$wgScriptPath}/{$script}.php";
3166         }
3167 }
3168
3169 /**
3170  * Get the script URL.
3171  *
3172  * @return string Script URL
3173  */
3174 function wfGetScriptUrl() {
3175         if ( isset( $_SERVER['SCRIPT_NAME'] ) ) {
3176                 /* as it was called, minus the query string.
3177                  *
3178                  * Some sites use Apache rewrite rules to handle subdomains,
3179                  * and have PHP set up in a weird way that causes PHP_SELF
3180                  * to contain the rewritten URL instead of the one that the
3181                  * outside world sees.
3182                  *
3183                  * If in this mode, use SCRIPT_URL instead, which mod_rewrite
3184                  * provides containing the "before" URL.
3185                  */
3186                 return $_SERVER['SCRIPT_NAME'];
3187         } else {
3188                 return $_SERVER['URL'];
3189         }
3190 }
3191
3192 /**
3193  * Convenience function converts boolean values into "true"
3194  * or "false" (string) values
3195  *
3196  * @param bool $value
3197  * @return string
3198  */
3199 function wfBoolToStr( $value ) {
3200         return $value ? 'true' : 'false';
3201 }
3202
3203 /**
3204  * Get a platform-independent path to the null file, e.g. /dev/null
3205  *
3206  * @return string
3207  */
3208 function wfGetNull() {
3209         return wfIsWindows() ? 'NUL' : '/dev/null';
3210 }
3211
3212 /**
3213  * Waits for the replica DBs to catch up to the master position
3214  *
3215  * Use this when updating very large numbers of rows, as in maintenance scripts,
3216  * to avoid causing too much lag. Of course, this is a no-op if there are no replica DBs.
3217  *
3218  * By default this waits on the main DB cluster of the current wiki.
3219  * If $cluster is set to "*" it will wait on all DB clusters, including
3220  * external ones. If the lag being waiting on is caused by the code that
3221  * does this check, it makes since to use $ifWritesSince, particularly if
3222  * cluster is "*", to avoid excess overhead.
3223  *
3224  * Never call this function after a big DB write that is still in a transaction.
3225  * This only makes sense after the possible lag inducing changes were committed.
3226  *
3227  * @param float|null $ifWritesSince Only wait if writes were done since this UNIX timestamp
3228  * @param string|bool $wiki Wiki identifier accepted by wfGetLB
3229  * @param string|bool $cluster Cluster name accepted by LBFactory. Default: false.
3230  * @param int|null $timeout Max wait time. Default: 1 day (cli), ~10 seconds (web)
3231  * @return bool Success (able to connect and no timeouts reached)
3232  * @deprecated since 1.27 Use LBFactory::waitForReplication
3233  */
3234 function wfWaitForSlaves(
3235         $ifWritesSince = null, $wiki = false, $cluster = false, $timeout = null
3236 ) {
3237         if ( $timeout === null ) {
3238                 $timeout = ( PHP_SAPI === 'cli' ) ? 86400 : 10;
3239         }
3240
3241         if ( $cluster === '*' ) {
3242                 $cluster = false;
3243                 $wiki = false;
3244         } elseif ( $wiki === false ) {
3245                 $wiki = wfWikiID();
3246         }
3247
3248         try {
3249                 wfGetLBFactory()->waitForReplication( [
3250                         'wiki' => $wiki,
3251                         'cluster' => $cluster,
3252                         'timeout' => $timeout,
3253                         // B/C: first argument used to be "max seconds of lag"; ignore such values
3254                         'ifWritesSince' => ( $ifWritesSince > 1e9 ) ? $ifWritesSince : null
3255                 ] );
3256         } catch ( DBReplicationWaitError $e ) {
3257                 return false;
3258         }
3259
3260         return true;
3261 }
3262
3263 /**
3264  * Count down from $seconds to zero on the terminal, with a one-second pause
3265  * between showing each number. For use in command-line scripts.
3266  *
3267  * @codeCoverageIgnore
3268  * @param int $seconds
3269  */
3270 function wfCountDown( $seconds ) {
3271         for ( $i = $seconds; $i >= 0; $i-- ) {
3272                 if ( $i != $seconds ) {
3273                         echo str_repeat( "\x08", strlen( $i + 1 ) );
3274                 }
3275                 echo $i;
3276                 flush();
3277                 if ( $i ) {
3278                         sleep( 1 );
3279                 }
3280         }
3281         echo "\n";
3282 }
3283
3284 /**
3285  * Replace all invalid characters with '-'.
3286  * Additional characters can be defined in $wgIllegalFileChars (see T22489).
3287  * By default, $wgIllegalFileChars includes ':', '/', '\'.
3288  *
3289  * @param string $name Filename to process
3290  * @return string
3291  */
3292 function wfStripIllegalFilenameChars( $name ) {
3293         global $wgIllegalFileChars;
3294         $illegalFileChars = $wgIllegalFileChars ? "|[" . $wgIllegalFileChars . "]" : '';
3295         $name = preg_replace(
3296                 "/[^" . Title::legalChars() . "]" . $illegalFileChars . "/",
3297                 '-',
3298                 $name
3299         );
3300         // $wgIllegalFileChars may not include '/' and '\', so we still need to do this
3301         $name = wfBaseName( $name );
3302         return $name;
3303 }
3304
3305 /**
3306  * Set PHP's memory limit to the larger of php.ini or $wgMemoryLimit
3307  *
3308  * @return int Resulting value of the memory limit.
3309  */
3310 function wfMemoryLimit() {
3311         global $wgMemoryLimit;
3312         $memlimit = wfShorthandToInteger( ini_get( 'memory_limit' ) );
3313         if ( $memlimit != -1 ) {
3314                 $conflimit = wfShorthandToInteger( $wgMemoryLimit );
3315                 if ( $conflimit == -1 ) {
3316                         wfDebug( "Removing PHP's memory limit\n" );
3317                         MediaWiki\suppressWarnings();
3318                         ini_set( 'memory_limit', $conflimit );
3319                         MediaWiki\restoreWarnings();
3320                         return $conflimit;
3321                 } elseif ( $conflimit > $memlimit ) {
3322                         wfDebug( "Raising PHP's memory limit to $conflimit bytes\n" );
3323                         MediaWiki\suppressWarnings();
3324                         ini_set( 'memory_limit', $conflimit );
3325                         MediaWiki\restoreWarnings();
3326                         return $conflimit;
3327                 }
3328         }
3329         return $memlimit;
3330 }
3331
3332 /**
3333  * Set PHP's time limit to the larger of php.ini or $wgTransactionalTimeLimit
3334  *
3335  * @return int Prior time limit
3336  * @since 1.26
3337  */
3338 function wfTransactionalTimeLimit() {
3339         global $wgTransactionalTimeLimit;
3340
3341         $timeLimit = ini_get( 'max_execution_time' );
3342         // Note that CLI scripts use 0
3343         if ( $timeLimit > 0 && $wgTransactionalTimeLimit > $timeLimit ) {
3344                 set_time_limit( $wgTransactionalTimeLimit );
3345         }
3346
3347         ignore_user_abort( true ); // ignore client disconnects
3348
3349         return $timeLimit;
3350 }
3351
3352 /**
3353  * Converts shorthand byte notation to integer form
3354  *
3355  * @param string $string
3356  * @param int $default Returned if $string is empty
3357  * @return int
3358  */
3359 function wfShorthandToInteger( $string = '', $default = -1 ) {
3360         $string = trim( $string );
3361         if ( $string === '' ) {
3362                 return $default;
3363         }
3364         $last = $string[strlen( $string ) - 1];
3365         $val = intval( $string );
3366         switch ( $last ) {
3367                 case 'g':
3368                 case 'G':
3369                         $val *= 1024;
3370                         // break intentionally missing
3371                 case 'm':
3372                 case 'M':
3373                         $val *= 1024;
3374                         // break intentionally missing
3375                 case 'k':
3376                 case 'K':
3377                         $val *= 1024;
3378         }
3379
3380         return $val;
3381 }
3382
3383 /**
3384  * Get the normalised IETF language tag
3385  * See unit test for examples.
3386  *
3387  * @param string $code The language code.
3388  * @return string The language code which complying with BCP 47 standards.
3389  */
3390 function wfBCP47( $code ) {
3391         $codeSegment = explode( '-', $code );
3392         $codeBCP = [];
3393         foreach ( $codeSegment as $segNo => $seg ) {
3394                 // when previous segment is x, it is a private segment and should be lc
3395                 if ( $segNo > 0 && strtolower( $codeSegment[( $segNo - 1 )] ) == 'x' ) {
3396                         $codeBCP[$segNo] = strtolower( $seg );
3397                 // ISO 3166 country code
3398                 } elseif ( ( strlen( $seg ) == 2 ) && ( $segNo > 0 ) ) {
3399                         $codeBCP[$segNo] = strtoupper( $seg );
3400                 // ISO 15924 script code
3401                 } elseif ( ( strlen( $seg ) == 4 ) && ( $segNo > 0 ) ) {
3402                         $codeBCP[$segNo] = ucfirst( strtolower( $seg ) );
3403                 // Use lowercase for other cases
3404                 } else {
3405                         $codeBCP[$segNo] = strtolower( $seg );
3406                 }
3407         }
3408         $langCode = implode( '-', $codeBCP );
3409         return $langCode;
3410 }
3411
3412 /**
3413  * Get a specific cache object.
3414  *
3415  * @param int|string $cacheType A CACHE_* constants, or other key in $wgObjectCaches
3416  * @return BagOStuff
3417  */
3418 function wfGetCache( $cacheType ) {
3419         return ObjectCache::getInstance( $cacheType );
3420 }
3421
3422 /**
3423  * Get the main cache object
3424  *
3425  * @return BagOStuff
3426  */
3427 function wfGetMainCache() {
3428         global $wgMainCacheType;
3429         return ObjectCache::getInstance( $wgMainCacheType );
3430 }
3431
3432 /**
3433  * Get the cache object used by the message cache
3434  *
3435  * @return BagOStuff
3436  */
3437 function wfGetMessageCacheStorage() {
3438         global $wgMessageCacheType;
3439         return ObjectCache::getInstance( $wgMessageCacheType );
3440 }
3441
3442 /**
3443  * Get the cache object used by the parser cache
3444  *
3445  * @return BagOStuff
3446  */
3447 function wfGetParserCacheStorage() {
3448         global $wgParserCacheType;
3449         return ObjectCache::getInstance( $wgParserCacheType );
3450 }
3451
3452 /**
3453  * Call hook functions defined in $wgHooks
3454  *
3455  * @param string $event Event name
3456  * @param array $args Parameters passed to hook functions
3457  * @param string|null $deprecatedVersion Optionally mark hook as deprecated with version number
3458  *
3459  * @return bool True if no handler aborted the hook
3460  * @deprecated since 1.25 - use Hooks::run
3461  */
3462 function wfRunHooks( $event, array $args = [], $deprecatedVersion = null ) {
3463         return Hooks::run( $event, $args, $deprecatedVersion );
3464 }
3465
3466 /**
3467  * Wrapper around php's unpack.
3468  *
3469  * @param string $format The format string (See php's docs)
3470  * @param string $data A binary string of binary data
3471  * @param int|bool $length The minimum length of $data or false. This is to
3472  *      prevent reading beyond the end of $data. false to disable the check.
3473  *
3474  * Also be careful when using this function to read unsigned 32 bit integer
3475  * because php might make it negative.
3476  *
3477  * @throws MWException If $data not long enough, or if unpack fails
3478  * @return array Associative array of the extracted data
3479  */
3480 function wfUnpack( $format, $data, $length = false ) {
3481         if ( $length !== false ) {
3482                 $realLen = strlen( $data );
3483                 if ( $realLen < $length ) {
3484                         throw new MWException( "Tried to use wfUnpack on a "
3485                                 . "string of length $realLen, but needed one "
3486                                 . "of at least length $length."
3487                         );
3488                 }
3489         }
3490
3491         MediaWiki\suppressWarnings();
3492         $result = unpack( $format, $data );
3493         MediaWiki\restoreWarnings();
3494
3495         if ( $result === false ) {
3496                 // If it cannot extract the packed data.
3497                 throw new MWException( "unpack could not unpack binary data" );
3498         }
3499         return $result;
3500 }
3501
3502 /**
3503  * Determine if an image exists on the 'bad image list'.
3504  *
3505  * The format of MediaWiki:Bad_image_list is as follows:
3506  *    * Only list items (lines starting with "*") are considered
3507  *    * The first link on a line must be a link to a bad image
3508  *    * Any subsequent links on the same line are considered to be exceptions,
3509  *      i.e. articles where the image may occur inline.
3510  *
3511  * @param string $name The image name to check
3512  * @param Title|bool $contextTitle The page on which the image occurs, if known
3513  * @param string $blacklist Wikitext of a file blacklist
3514  * @return bool
3515  */
3516 function wfIsBadImage( $name, $contextTitle = false, $blacklist = null ) {
3517         # Handle redirects; callers almost always hit wfFindFile() anyway,
3518         # so just use that method because it has a fast process cache.
3519         $file = wfFindFile( $name ); // get the final name
3520         $name = $file ? $file->getTitle()->getDBkey() : $name;
3521
3522         # Run the extension hook
3523         $bad = false;
3524         if ( !Hooks::run( 'BadImage', [ $name, &$bad ] ) ) {
3525                 return (bool)$bad;
3526         }
3527
3528         $cache = ObjectCache::getLocalServerInstance( 'hash' );
3529         $key = wfMemcKey( 'bad-image-list', ( $blacklist === null ) ? 'default' : md5( $blacklist ) );
3530         $badImages = $cache->get( $key );
3531
3532         if ( $badImages === false ) { // cache miss
3533                 if ( $blacklist === null ) {
3534                         $blacklist = wfMessage( 'bad_image_list' )->inContentLanguage()->plain(); // site list
3535                 }
3536                 # Build the list now
3537                 $badImages = [];
3538                 $lines = explode( "\n", $blacklist );
3539                 foreach ( $lines as $line ) {
3540                         # List items only
3541                         if ( substr( $line, 0, 1 ) !== '*' ) {
3542                                 continue;
3543                         }
3544
3545                         # Find all links
3546                         $m = [];
3547                         if ( !preg_match_all( '/\[\[:?(.*?)\]\]/', $line, $m ) ) {
3548                                 continue;
3549                         }
3550
3551                         $exceptions = [];
3552                         $imageDBkey = false;
3553                         foreach ( $m[1] as $i => $titleText ) {
3554                                 $title = Title::newFromText( $titleText );
3555                                 if ( !is_null( $title ) ) {
3556                                         if ( $i == 0 ) {
3557                                                 $imageDBkey = $title->getDBkey();
3558                                         } else {
3559                                                 $exceptions[$title->getPrefixedDBkey()] = true;
3560                                         }
3561                                 }
3562                         }
3563
3564                         if ( $imageDBkey !== false ) {
3565                                 $badImages[$imageDBkey] = $exceptions;
3566                         }
3567                 }
3568                 $cache->set( $key, $badImages, 60 );
3569         }
3570
3571         $contextKey = $contextTitle ? $contextTitle->getPrefixedDBkey() : false;
3572         $bad = isset( $badImages[$name] ) && !isset( $badImages[$name][$contextKey] );
3573
3574         return $bad;
3575 }
3576
3577 /**
3578  * Determine whether the client at a given source IP is likely to be able to
3579  * access the wiki via HTTPS.
3580  *
3581  * @param string $ip The IPv4/6 address in the normal human-readable form
3582  * @return bool
3583  */
3584 function wfCanIPUseHTTPS( $ip ) {
3585         $canDo = true;
3586         Hooks::run( 'CanIPUseHTTPS', [ $ip, &$canDo ] );
3587         return !!$canDo;
3588 }
3589
3590 /**
3591  * Determine input string is represents as infinity
3592  *
3593  * @param string $str The string to determine
3594  * @return bool
3595  * @since 1.25
3596  */
3597 function wfIsInfinity( $str ) {
3598         $infinityValues = [ 'infinite', 'indefinite', 'infinity', 'never' ];
3599         return in_array( $str, $infinityValues );
3600 }
3601
3602 /**
3603  * Returns true if these thumbnail parameters match one that MediaWiki
3604  * requests from file description pages and/or parser output.
3605  *
3606  * $params is considered non-standard if they involve a non-standard
3607  * width or any non-default parameters aside from width and page number.
3608  * The number of possible files with standard parameters is far less than
3609  * that of all combinations; rate-limiting for them can thus be more generious.
3610  *
3611  * @param File $file
3612  * @param array $params
3613  * @return bool
3614  * @since 1.24 Moved from thumb.php to GlobalFunctions in 1.25
3615  */
3616 function wfThumbIsStandard( File $file, array $params ) {
3617         global $wgThumbLimits, $wgImageLimits, $wgResponsiveImages;
3618
3619         $multipliers = [ 1 ];
3620         if ( $wgResponsiveImages ) {
3621                 // These available sizes are hardcoded currently elsewhere in MediaWiki.
3622                 // @see Linker::processResponsiveImages
3623                 $multipliers[] = 1.5;
3624                 $multipliers[] = 2;
3625         }
3626
3627         $handler = $file->getHandler();
3628         if ( !$handler || !isset( $params['width'] ) ) {
3629                 return false;
3630         }
3631
3632         $basicParams = [];
3633         if ( isset( $params['page'] ) ) {
3634                 $basicParams['page'] = $params['page'];
3635         }
3636
3637         $thumbLimits = [];
3638         $imageLimits = [];
3639         // Expand limits to account for multipliers
3640         foreach ( $multipliers as $multiplier ) {
3641                 $thumbLimits = array_merge( $thumbLimits, array_map(
3642                         function ( $width ) use ( $multiplier ) {
3643                                 return round( $width * $multiplier );
3644                         }, $wgThumbLimits )
3645                 );
3646                 $imageLimits = array_merge( $imageLimits, array_map(
3647                         function ( $pair ) use ( $multiplier ) {
3648                                 return [
3649                                         round( $pair[0] * $multiplier ),
3650                                         round( $pair[1] * $multiplier ),
3651                                 ];
3652                         }, $wgImageLimits )
3653                 );
3654         }
3655
3656         // Check if the width matches one of $wgThumbLimits
3657         if ( in_array( $params['width'], $thumbLimits ) ) {
3658                 $normalParams = $basicParams + [ 'width' => $params['width'] ];
3659                 // Append any default values to the map (e.g. "lossy", "lossless", ...)
3660                 $handler->normaliseParams( $file, $normalParams );
3661         } else {
3662                 // If not, then check if the width matchs one of $wgImageLimits
3663                 $match = false;
3664                 foreach ( $imageLimits as $pair ) {
3665                         $normalParams = $basicParams + [ 'width' => $pair[0], 'height' => $pair[1] ];
3666                         // Decide whether the thumbnail should be scaled on width or height.
3667                         // Also append any default values to the map (e.g. "lossy", "lossless", ...)
3668                         $handler->normaliseParams( $file, $normalParams );
3669                         // Check if this standard thumbnail size maps to the given width
3670                         if ( $normalParams['width'] == $params['width'] ) {
3671                                 $match = true;
3672                                 break;
3673                         }
3674                 }
3675                 if ( !$match ) {
3676                         return false; // not standard for description pages
3677                 }
3678         }
3679
3680         // Check that the given values for non-page, non-width, params are just defaults
3681         foreach ( $params as $key => $value ) {
3682                 if ( !isset( $normalParams[$key] ) || $normalParams[$key] != $value ) {
3683                         return false;
3684                 }
3685         }
3686
3687         return true;
3688 }
3689
3690 /**
3691  * Merges two (possibly) 2 dimensional arrays into the target array ($baseArray).
3692  *
3693  * Values that exist in both values will be combined with += (all values of the array
3694  * of $newValues will be added to the values of the array of $baseArray, while values,
3695  * that exists in both, the value of $baseArray will be used).
3696  *
3697  * @param array $baseArray The array where you want to add the values of $newValues to
3698  * @param array $newValues An array with new values
3699  * @return array The combined array
3700  * @since 1.26
3701  */
3702 function wfArrayPlus2d( array $baseArray, array $newValues ) {
3703         // First merge items that are in both arrays
3704         foreach ( $baseArray as $name => &$groupVal ) {
3705                 if ( isset( $newValues[$name] ) ) {
3706                         $groupVal += $newValues[$name];
3707                 }
3708         }
3709         // Now add items that didn't exist yet
3710         $baseArray += $newValues;
3711
3712         return $baseArray;
3713 }