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
  31 // Hide compatibility functions from Doxygen
  32 /// @cond
  33 /**
  34  * Compatibility functions
  35  *
  36  * We support PHP 5.5.9 and up.
  37  * Re-implementations of newer functions or functions in non-standard
  38  * PHP extensions may be included here.
  39  */
  40
  41 // hash_equals function only exists in PHP >= 5.6.0
  42 // http://php.net/hash_equals
  43 if ( !function_exists( 'hash_equals' ) ) {
  44         /**
  45          * Check whether a user-provided string is equal to a fixed-length secret string
  46          * without revealing bytes of the secret string through timing differences.
  47          *
  48          * The usual way to compare strings (PHP's === operator or the underlying memcmp()
  49          * function in C) is to compare corresponding bytes and stop at the first difference,
  50          * which would take longer for a partial match than for a complete mismatch. This
  51          * is not secure when one of the strings (e.g. an HMAC or token) must remain secret
  52          * and the other may come from an attacker. Statistical analysis of timing measurements
  53          * over many requests may allow the attacker to guess the string's bytes one at a time
  54          * (and check his guesses) even if the timing differences are extremely small.
  55          *
  56          * When making such a security-sensitive comparison, it is essential that the sequence
  57          * in which instructions are executed and memory locations are accessed not depend on
  58          * the secret string's value. HOWEVER, for simplicity, we do not attempt to minimize
  59          * the inevitable leakage of the string's length. That is generally known anyway as
  60          * a chararacteristic of the hash function used to compute the secret value.
  61          *
  62          * Longer explanation: http://www.emerose.com/timing-attacks-explained
  63          *
  64          * @codeCoverageIgnore
  65          * @param string $known_string Fixed-length secret string to compare against
  66          * @param string $user_string User-provided string
  67          * @return bool True if the strings are the same, false otherwise
  68          */
  69         function hash_equals( $known_string, $user_string ) {
  70                 // Strict type checking as in PHP's native implementation
  71                 if ( !is_string( $known_string ) ) {
  72                         trigger_error( 'hash_equals(): Expected known_string to be a string, ' .
  73                                 gettype( $known_string ) . ' given', E_USER_WARNING );
  74
  75                         return false;
  76                 }
  77
  78                 if ( !is_string( $user_string ) ) {
  79                         trigger_error( 'hash_equals(): Expected user_string to be a string, ' .
  80                                 gettype( $user_string ) . ' given', E_USER_WARNING );
  81
  82                         return false;
  83                 }
  84
  85                 $known_string_len = strlen( $known_string );
  86                 if ( $known_string_len !== strlen( $user_string ) ) {
  87                         return false;
  88                 }
  89
  90                 $result = 0;
  91                 for ( $i = 0; $i < $known_string_len; $i++ ) {
  92                         $result |= ord( $known_string[$i] ) ^ ord( $user_string[$i] );
  93                 }
  94
  95                 return ( $result === 0 );
  96         }
  97 }
  98 /// @endcond
  99
 100 /**
 101  * Load an extension
 102  *
 103  * This queues an extension to be loaded through
 104  * the ExtensionRegistry system.
 105  *
 106  * @param string $ext Name of the extension to load
 107  * @param string|null $path Absolute path of where to find the extension.json file
 108  * @since 1.25
 109  */
 110 function wfLoadExtension( $ext, $path = null ) {
 111         if ( !$path ) {
 112                 global $wgExtensionDirectory;
 113                 $path = "$wgExtensionDirectory/$ext/extension.json";
 114         }
 115         ExtensionRegistry::getInstance()->queue( $path );
 116 }
 117
 118 /**
 119  * Load multiple extensions at once
 120  *
 121  * Same as wfLoadExtension, but more efficient if you
 122  * are loading multiple extensions.
 123  *
 124  * If you want to specify custom paths, you should interact with
 125  * ExtensionRegistry directly.
 126  *
 127  * @see wfLoadExtension
 128  * @param string[] $exts Array of extension names to load
 129  * @since 1.25
 130  */
 131 function wfLoadExtensions( array $exts ) {
 132         global $wgExtensionDirectory;
 133         $registry = ExtensionRegistry::getInstance();
 134         foreach ( $exts as $ext ) {
 135                 $registry->queue( "$wgExtensionDirectory/$ext/extension.json" );
 136         }
 137 }
 138
 139 /**
 140  * Load a skin
 141  *
 142  * @see wfLoadExtension
 143  * @param string $skin Name of the extension to load
 144  * @param string|null $path Absolute path of where to find the skin.json file
 145  * @since 1.25
 146  */
 147 function wfLoadSkin( $skin, $path = null ) {
 148         if ( !$path ) {
 149                 global $wgStyleDirectory;
 150                 $path = "$wgStyleDirectory/$skin/skin.json";
 151         }
 152         ExtensionRegistry::getInstance()->queue( $path );
 153 }
 154
 155 /**
 156  * Load multiple skins at once
 157  *
 158  * @see wfLoadExtensions
 159  * @param string[] $skins Array of extension names to load
 160  * @since 1.25
 161  */
 162 function wfLoadSkins( array $skins ) {
 163         global $wgStyleDirectory;
 164         $registry = ExtensionRegistry::getInstance();
 165         foreach ( $skins as $skin ) {
 166                 $registry->queue( "$wgStyleDirectory/$skin/skin.json" );
 167         }
 168 }
 169
 170 /**
 171  * Like array_diff( $a, $b ) except that it works with two-dimensional arrays.
 172  * @param array $a
 173  * @param array $b
 174  * @return array
 175  */
 176 function wfArrayDiff2( $a, $b ) {
 177         return array_udiff( $a, $b, 'wfArrayDiff2_cmp' );
 178 }
 179
 180 /**
 181  * @param array|string $a
 182  * @param array|string $b
 183  * @return int
 184  */
 185 function wfArrayDiff2_cmp( $a, $b ) {
 186         if ( is_string( $a ) && is_string( $b ) ) {
 187                 return strcmp( $a, $b );
 188         } elseif ( count( $a ) !== count( $b ) ) {
 189                 return count( $a ) < count( $b ) ? -1 : 1;
 190         } else {
 191                 reset( $a );
 192                 reset( $b );
 193                 while ( ( list( , $valueA ) = each( $a ) ) && ( list( , $valueB ) = each( $b ) ) ) {
 194                         $cmp = strcmp( $valueA, $valueB );
 195                         if ( $cmp !== 0 ) {
 196                                 return $cmp;
 197                         }
 198                 }
 199                 return 0;
 200         }
 201 }
 202
 203 /**
 204  * Appends to second array if $value differs from that in $default
 205  *
 206  * @param string|int $key
 207  * @param mixed $value
 208  * @param mixed $default
 209  * @param array $changed Array to alter
 210  * @throws MWException
 211  */
 212 function wfAppendToArrayIfNotDefault( $key, $value, $default, &$changed ) {
 213         if ( is_null( $changed ) ) {
 214                 throw new MWException( 'GlobalFunctions::wfAppendToArrayIfNotDefault got null' );
 215         }
 216         if ( $default[$key] !== $value ) {
 217                 $changed[$key] = $value;
 218         }
 219 }
 220
 221 /**
 222  * Merge arrays in the style of getUserPermissionsErrors, with duplicate removal
 223  * e.g.
 224  *      wfMergeErrorArrays(
 225  *              [ [ 'x' ] ],
 226  *              [ [ 'x', '2' ] ],
 227  *              [ [ 'x' ] ],
 228  *              [ [ 'y' ] ]
 229  *      );
 230  * returns:
 231  *              [
 232  *              [ 'x', '2' ],
 233  *              [ 'x' ],
 234  *              [ 'y' ]
 235  *      ]
 236  *
 237  * @param array $array1,...
 238  * @return array
 239  */
 240 function wfMergeErrorArrays( /*...*/ ) {
 241         $args = func_get_args();
 242         $out = [];
 243         foreach ( $args as $errors ) {
 244                 foreach ( $errors as $params ) {
 245                         $originalParams = $params;
 246                         if ( $params[0] instanceof MessageSpecifier ) {
 247                                 $msg = $params[0];
 248                                 $params = array_merge( [ $msg->getKey() ], $msg->getParams() );
 249                         }
 250                         # @todo FIXME: Sometimes get nested arrays for $params,
 251                         # which leads to E_NOTICEs
 252                         $spec = implode( "\t", $params );
 253                         $out[$spec] = $originalParams;
 254                 }
 255         }
 256         return array_values( $out );
 257 }
 258
 259 /**
 260  * Insert array into another array after the specified *KEY*
 261  *
 262  * @param array $array The array.
 263  * @param array $insert The array to insert.
 264  * @param mixed $after The key to insert after
 265  * @return array
 266  */
 267 function wfArrayInsertAfter( array $array, array $insert, $after ) {
 268         // Find the offset of the element to insert after.
 269         $keys = array_keys( $array );
 270         $offsetByKey = array_flip( $keys );
 271
 272         $offset = $offsetByKey[$after];
 273
 274         // Insert at the specified offset
 275         $before = array_slice( $array, 0, $offset + 1, true );
 276         $after = array_slice( $array, $offset + 1, count( $array ) - $offset, true );
 277
 278         $output = $before + $insert + $after;
 279
 280         return $output;
 281 }
 282
 283 /**
 284  * Recursively converts the parameter (an object) to an array with the same data
 285  *
 286  * @param object|array $objOrArray
 287  * @param bool $recursive
 288  * @return array
 289  */
 290 function wfObjectToArray( $objOrArray, $recursive = true ) {
 291         $array = [];
 292         if ( is_object( $objOrArray ) ) {
 293                 $objOrArray = get_object_vars( $objOrArray );
 294         }
 295         foreach ( $objOrArray as $key => $value ) {
 296                 if ( $recursive && ( is_object( $value ) || is_array( $value ) ) ) {
 297                         $value = wfObjectToArray( $value );
 298                 }
 299
 300                 $array[$key] = $value;
 301         }
 302
 303         return $array;
 304 }
 305
 306 /**
 307  * Get a random decimal value between 0 and 1, in a way
 308  * not likely to give duplicate values for any realistic
 309  * number of articles.
 310  *
 311  * @note This is designed for use in relation to Special:RandomPage
 312  *       and the page_random database field.
 313  *
 314  * @return string
 315  */
 316 function wfRandom() {
 317         // The maximum random value is "only" 2^31-1, so get two random
 318         // values to reduce the chance of dupes
 319         $max = mt_getrandmax() + 1;
 320         $rand = number_format( ( mt_rand() * $max + mt_rand() ) / $max / $max, 12, '.', '' );
 321         return $rand;
 322 }
 323
 324 /**
 325  * Get a random string containing a number of pseudo-random hex characters.
 326  *
 327  * @note This is not secure, if you are trying to generate some sort
 328  *       of token please use MWCryptRand instead.
 329  *
 330  * @param int $length The length of the string to generate
 331  * @return string
 332  * @since 1.20
 333  */
 334 function wfRandomString( $length = 32 ) {
 335         $str = '';
 336         for ( $n = 0; $n < $length; $n += 7 ) {
 337                 $str .= sprintf( '%07x', mt_rand() & 0xfffffff );
 338         }
 339         return substr( $str, 0, $length );
 340 }
 341
 342 /**
 343  * We want some things to be included as literal characters in our title URLs
 344  * for prettiness, which urlencode encodes by default.  According to RFC 1738,
 345  * all of the following should be safe:
 346  *
 347  * ;:@&=$-_.+!*'(),
 348  *
 349  * RFC 1738 says ~ is unsafe, however RFC 3986 considers it an unreserved
 350  * character which should not be encoded. More importantly, google chrome
 351  * always converts %7E back to ~, and converting it in this function can
 352  * cause a redirect loop (T105265).
 353  *
 354  * But + is not safe because it's used to indicate a space; &= are only safe in
 355  * paths and not in queries (and we don't distinguish here); ' seems kind of
 356  * scary; and urlencode() doesn't touch -_. to begin with.  Plus, although /
 357  * is reserved, we don't care.  So the list we unescape is:
 358  *
 359  * ;:@$!*(),/~
 360  *
 361  * However, IIS7 redirects fail when the url contains a colon (Bug 22709),
 362  * so no fancy : for IIS7.
 363  *
 364  * %2F in the page titles seems to fatally break for some reason.
 365  *
 366  * @param string $s
 367  * @return string
 368  */
 369 function wfUrlencode( $s ) {
 370         static $needle;
 371
 372         if ( is_null( $s ) ) {
 373                 $needle = null;
 374                 return '';
 375         }
 376
 377         if ( is_null( $needle ) ) {
 378                 $needle = [ '%3B', '%40', '%24', '%21', '%2A', '%28', '%29', '%2C', '%2F', '%7E' ];
 379                 if ( !isset( $_SERVER['SERVER_SOFTWARE'] ) ||
 380                         ( strpos( $_SERVER['SERVER_SOFTWARE'], 'Microsoft-IIS/7' ) === false )
 381                 ) {
 382                         $needle[] = '%3A';
 383                 }
 384         }
 385
 386         $s = urlencode( $s );
 387         $s = str_ireplace(
 388                 $needle,
 389                 [ ';', '@', '$', '!', '*', '(', ')', ',', '/', '~', ':' ],
 390                 $s
 391         );
 392
 393         return $s;
 394 }
 395
 396 /**
 397  * This function takes one or two arrays as input, and returns a CGI-style string, e.g.
 398  * "days=7&limit=100". Options in the first array override options in the second.
 399  * Options set to null or false will not be output.
 400  *
 401  * @param array $array1 ( String|Array )
 402  * @param array|null $array2 ( String|Array )
 403  * @param string $prefix
 404  * @return string
 405  */
 406 function wfArrayToCgi( $array1, $array2 = null, $prefix = '' ) {
 407         if ( !is_null( $array2 ) ) {
 408                 $array1 = $array1 + $array2;
 409         }
 410
 411         $cgi = '';
 412         foreach ( $array1 as $key => $value ) {
 413                 if ( !is_null( $value ) && $value !== false ) {
 414                         if ( $cgi != '' ) {
 415                                 $cgi .= '&';
 416                         }
 417                         if ( $prefix !== '' ) {
 418                                 $key = $prefix . "[$key]";
 419                         }
 420                         if ( is_array( $value ) ) {
 421                                 $firstTime = true;
 422                                 foreach ( $value as $k => $v ) {
 423                                         $cgi .= $firstTime ? '' : '&';
 424                                         if ( is_array( $v ) ) {
 425                                                 $cgi .= wfArrayToCgi( $v, null, $key . "[$k]" );
 426                                         } else {
 427                                                 $cgi .= urlencode( $key . "[$k]" ) . '=' . urlencode( $v );
 428                                         }
 429                                         $firstTime = false;
 430                                 }
 431                         } else {
 432                                 if ( is_object( $value ) ) {
 433                                         $value = $value->__toString();
 434                                 }
 435                                 $cgi .= urlencode( $key ) . '=' . urlencode( $value );
 436                         }
 437                 }
 438         }
 439         return $cgi;
 440 }
 441
 442 /**
 443  * This is the logical opposite of wfArrayToCgi(): it accepts a query string as
 444  * its argument and returns the same string in array form.  This allows compatibility
 445  * with legacy functions that accept raw query strings instead of nice
 446  * arrays.  Of course, keys and values are urldecode()d.
 447  *
 448  * @param string $query Query string
 449  * @return string[] Array version of input
 450  */
 451 function wfCgiToArray( $query ) {
 452         if ( isset( $query[0] ) && $query[0] == '?' ) {
 453                 $query = substr( $query, 1 );
 454         }
 455         $bits = explode( '&', $query );
 456         $ret = [];
 457         foreach ( $bits as $bit ) {
 458                 if ( $bit === '' ) {
 459                         continue;
 460                 }
 461                 if ( strpos( $bit, '=' ) === false ) {
 462                         // Pieces like &qwerty become 'qwerty' => '' (at least this is what php does)
 463                         $key = $bit;
 464                         $value = '';
 465                 } else {
 466                         list( $key, $value ) = explode( '=', $bit );
 467                 }
 468                 $key = urldecode( $key );
 469                 $value = urldecode( $value );
 470                 if ( strpos( $key, '[' ) !== false ) {
 471                         $keys = array_reverse( explode( '[', $key ) );
 472                         $key = array_pop( $keys );
 473                         $temp = $value;
 474                         foreach ( $keys as $k ) {
 475                                 $k = substr( $k, 0, -1 );
 476                                 $temp = [ $k => $temp ];
 477                         }
 478                         if ( isset( $ret[$key] ) ) {
 479                                 $ret[$key] = array_merge( $ret[$key], $temp );
 480                         } else {
 481                                 $ret[$key] = $temp;
 482                         }
 483                 } else {
 484                         $ret[$key] = $value;
 485                 }
 486         }
 487         return $ret;
 488 }
 489
 490 /**
 491  * Append a query string to an existing URL, which may or may not already
 492  * have query string parameters already. If so, they will be combined.
 493  *
 494  * @param string $url
 495  * @param string|string[] $query String or associative array
 496  * @return string
 497  */
 498 function wfAppendQuery( $url, $query ) {
 499         if ( is_array( $query ) ) {
 500                 $query = wfArrayToCgi( $query );
 501         }
 502         if ( $query != '' ) {
 503                 // Remove the fragment, if there is one
 504                 $fragment = false;
 505                 $hashPos = strpos( $url, '#' );
 506                 if ( $hashPos !== false ) {
 507                         $fragment = substr( $url, $hashPos );
 508                         $url = substr( $url, 0, $hashPos );
 509                 }
 510
 511                 // Add parameter
 512                 if ( false === strpos( $url, '?' ) ) {
 513                         $url .= '?';
 514                 } else {
 515                         $url .= '&';
 516                 }
 517                 $url .= $query;
 518
 519                 // Put the fragment back
 520                 if ( $fragment !== false ) {
 521                         $url .= $fragment;
 522                 }
 523         }
 524         return $url;
 525 }
 526
 527 /**
 528  * Expand a potentially local URL to a fully-qualified URL. Assumes $wgServer
 529  * is correct.
 530  *
 531  * The meaning of the PROTO_* constants is as follows:
 532  * PROTO_HTTP: Output a URL starting with http://
 533  * PROTO_HTTPS: Output a URL starting with https://
 534  * PROTO_RELATIVE: Output a URL starting with // (protocol-relative URL)
 535  * PROTO_CURRENT: Output a URL starting with either http:// or https:// , depending
 536  *    on which protocol was used for the current incoming request
 537  * PROTO_CANONICAL: For URLs without a domain, like /w/index.php , use $wgCanonicalServer.
 538  *    For protocol-relative URLs, use the protocol of $wgCanonicalServer
 539  * PROTO_INTERNAL: Like PROTO_CANONICAL, but uses $wgInternalServer instead of $wgCanonicalServer
 540  *
 541  * @todo this won't work with current-path-relative URLs
 542  * like "subdir/foo.html", etc.
 543  *
 544  * @param string $url Either fully-qualified or a local path + query
 545  * @param string $defaultProto One of the PROTO_* constants. Determines the
 546  *    protocol to use if $url or $wgServer is protocol-relative
 547  * @return string Fully-qualified URL, current-path-relative URL or false if
 548  *    no valid URL can be constructed
 549  */
 550 function wfExpandUrl( $url, $defaultProto = PROTO_CURRENT ) {
 551         global $wgServer, $wgCanonicalServer, $wgInternalServer, $wgRequest,
 552                 $wgHttpsPort;
 553         if ( $defaultProto === PROTO_CANONICAL ) {
 554                 $serverUrl = $wgCanonicalServer;
 555         } elseif ( $defaultProto === PROTO_INTERNAL && $wgInternalServer !== false ) {
 556                 // Make $wgInternalServer fall back to $wgServer if not set
 557                 $serverUrl = $wgInternalServer;
 558         } else {
 559                 $serverUrl = $wgServer;
 560                 if ( $defaultProto === PROTO_CURRENT ) {
 561                         $defaultProto = $wgRequest->getProtocol() . '://';
 562                 }
 563         }
 564
 565         // Analyze $serverUrl to obtain its protocol
 566         $bits = wfParseUrl( $serverUrl );
 567         $serverHasProto = $bits && $bits['scheme'] != '';
 568
 569         if ( $defaultProto === PROTO_CANONICAL || $defaultProto === PROTO_INTERNAL ) {
 570                 if ( $serverHasProto ) {
 571                         $defaultProto = $bits['scheme'] . '://';
 572                 } else {
 573                         // $wgCanonicalServer or $wgInternalServer doesn't have a protocol.
 574                         // This really isn't supposed to happen. Fall back to HTTP in this
 575                         // ridiculous case.
 576                         $defaultProto = PROTO_HTTP;
 577                 }
 578         }
 579
 580         $defaultProtoWithoutSlashes = substr( $defaultProto, 0, -2 );
 581
 582         if ( substr( $url, 0, 2 ) == '//' ) {
 583                 $url = $defaultProtoWithoutSlashes . $url;
 584         } elseif ( substr( $url, 0, 1 ) == '/' ) {
 585                 // If $serverUrl is protocol-relative, prepend $defaultProtoWithoutSlashes,
 586                 // otherwise leave it alone.
 587                 $url = ( $serverHasProto ? '' : $defaultProtoWithoutSlashes ) . $serverUrl . $url;
 588         }
 589
 590         $bits = wfParseUrl( $url );
 591
 592         // ensure proper port for HTTPS arrives in URL
 593         // https://phabricator.wikimedia.org/T67184
 594         if ( $defaultProto === PROTO_HTTPS && $wgHttpsPort != 443 ) {
 595                 $bits['port'] = $wgHttpsPort;
 596         }
 597
 598         if ( $bits && isset( $bits['path'] ) ) {
 599                 $bits['path'] = wfRemoveDotSegments( $bits['path'] );
 600                 return wfAssembleUrl( $bits );
 601         } elseif ( $bits ) {
 602                 # No path to expand
 603                 return $url;
 604         } elseif ( substr( $url, 0, 1 ) != '/' ) {
 605                 # URL is a relative path
 606                 return wfRemoveDotSegments( $url );
 607         }
 608
 609         # Expanded URL is not valid.
 610         return false;
 611 }
 612
 613 /**
 614  * This function will reassemble a URL parsed with wfParseURL.  This is useful
 615  * if you need to edit part of a URL and put it back together.
 616  *
 617  * This is the basic structure used (brackets contain keys for $urlParts):
 618  * [scheme][delimiter][user]:[pass]@[host]:[port][path]?[query]#[fragment]
 619  *
 620  * @todo Need to integrate this into wfExpandUrl (bug 32168)
 621  *
 622  * @since 1.19
 623  * @param array $urlParts URL parts, as output from wfParseUrl
 624  * @return string URL assembled from its component parts
 625  */
 626 function wfAssembleUrl( $urlParts ) {
 627         $result = '';
 628
 629         if ( isset( $urlParts['delimiter'] ) ) {
 630                 if ( isset( $urlParts['scheme'] ) ) {
 631                         $result .= $urlParts['scheme'];
 632                 }
 633
 634                 $result .= $urlParts['delimiter'];
 635         }
 636
 637         if ( isset( $urlParts['host'] ) ) {
 638                 if ( isset( $urlParts['user'] ) ) {
 639                         $result .= $urlParts['user'];
 640                         if ( isset( $urlParts['pass'] ) ) {
 641                                 $result .= ':' . $urlParts['pass'];
 642                         }
 643                         $result .= '@';
 644                 }
 645
 646                 $result .= $urlParts['host'];
 647
 648                 if ( isset( $urlParts['port'] ) ) {
 649                         $result .= ':' . $urlParts['port'];
 650                 }
 651         }
 652
 653         if ( isset( $urlParts['path'] ) ) {
 654                 $result .= $urlParts['path'];
 655         }
 656
 657         if ( isset( $urlParts['query'] ) ) {
 658                 $result .= '?' . $urlParts['query'];
 659         }
 660
 661         if ( isset( $urlParts['fragment'] ) ) {
 662                 $result .= '#' . $urlParts['fragment'];
 663         }
 664
 665         return $result;
 666 }
 667
 668 /**
 669  * Remove all dot-segments in the provided URL path.  For example,
 670  * '/a/./b/../c/' becomes '/a/c/'.  For details on the algorithm, please see
 671  * RFC3986 section 5.2.4.
 672  *
 673  * @todo Need to integrate this into wfExpandUrl (bug 32168)
 674  *
 675  * @param string $urlPath URL path, potentially containing dot-segments
 676  * @return string URL path with all dot-segments removed
 677  */
 678 function wfRemoveDotSegments( $urlPath ) {
 679         $output = '';
 680         $inputOffset = 0;
 681         $inputLength = strlen( $urlPath );
 682
 683         while ( $inputOffset < $inputLength ) {
 684                 $prefixLengthOne = substr( $urlPath, $inputOffset, 1 );
 685                 $prefixLengthTwo = substr( $urlPath, $inputOffset, 2 );
 686                 $prefixLengthThree = substr( $urlPath, $inputOffset, 3 );
 687                 $prefixLengthFour = substr( $urlPath, $inputOffset, 4 );
 688                 $trimOutput = false;
 689
 690                 if ( $prefixLengthTwo == './' ) {
 691                         # Step A, remove leading "./"
 692                         $inputOffset += 2;
 693                 } elseif ( $prefixLengthThree == '../' ) {
 694                         # Step A, remove leading "../"
 695                         $inputOffset += 3;
 696                 } elseif ( ( $prefixLengthTwo == '/.' ) && ( $inputOffset + 2 == $inputLength ) ) {
 697                         # Step B, replace leading "/.$" with "/"
 698                         $inputOffset += 1;
 699                         $urlPath[$inputOffset] = '/';
 700                 } elseif ( $prefixLengthThree == '/./' ) {
 701                         # Step B, replace leading "/./" with "/"
 702                         $inputOffset += 2;
 703                 } elseif ( $prefixLengthThree == '/..' && ( $inputOffset + 3 == $inputLength ) ) {
 704                         # Step C, replace leading "/..$" with "/" and
 705                         # remove last path component in output
 706                         $inputOffset += 2;
 707                         $urlPath[$inputOffset] = '/';
 708                         $trimOutput = true;
 709                 } elseif ( $prefixLengthFour == '/../' ) {
 710                         # Step C, replace leading "/../" with "/" and
 711                         # remove last path component in output
 712                         $inputOffset += 3;
 713                         $trimOutput = true;
 714                 } elseif ( ( $prefixLengthOne == '.' ) && ( $inputOffset + 1 == $inputLength ) ) {
 715                         # Step D, remove "^.$"
 716                         $inputOffset += 1;
 717                 } elseif ( ( $prefixLengthTwo == '..' ) && ( $inputOffset + 2 == $inputLength ) ) {
 718                         # Step D, remove "^..$"
 719                         $inputOffset += 2;
 720                 } else {
 721                         # Step E, move leading path segment to output
 722                         if ( $prefixLengthOne == '/' ) {
 723                                 $slashPos = strpos( $urlPath, '/', $inputOffset + 1 );
 724                         } else {
 725                                 $slashPos = strpos( $urlPath, '/', $inputOffset );
 726                         }
 727                         if ( $slashPos === false ) {
 728                                 $output .= substr( $urlPath, $inputOffset );
 729                                 $inputOffset = $inputLength;
 730                         } else {
 731                                 $output .= substr( $urlPath, $inputOffset, $slashPos - $inputOffset );
 732                                 $inputOffset += $slashPos - $inputOffset;
 733                         }
 734                 }
 735
 736                 if ( $trimOutput ) {
 737                         $slashPos = strrpos( $output, '/' );
 738                         if ( $slashPos === false ) {
 739                                 $output = '';
 740                         } else {
 741                                 $output = substr( $output, 0, $slashPos );
 742                         }
 743                 }
 744         }
 745
 746         return $output;
 747 }
 748
 749 /**
 750  * Returns a regular expression of url protocols
 751  *
 752  * @param bool $includeProtocolRelative If false, remove '//' from the returned protocol list.
 753  *        DO NOT USE this directly, use wfUrlProtocolsWithoutProtRel() instead
 754  * @return string
 755  */
 756 function wfUrlProtocols( $includeProtocolRelative = true ) {
 757         global $wgUrlProtocols;
 758
 759         // Cache return values separately based on $includeProtocolRelative
 760         static $withProtRel = null, $withoutProtRel = null;
 761         $cachedValue = $includeProtocolRelative ? $withProtRel : $withoutProtRel;
 762         if ( !is_null( $cachedValue ) ) {
 763                 return $cachedValue;
 764         }
 765
 766         // Support old-style $wgUrlProtocols strings, for backwards compatibility
 767         // with LocalSettings files from 1.5
 768         if ( is_array( $wgUrlProtocols ) ) {
 769                 $protocols = [];
 770                 foreach ( $wgUrlProtocols as $protocol ) {
 771                         // Filter out '//' if !$includeProtocolRelative
 772                         if ( $includeProtocolRelative || $protocol !== '//' ) {
 773                                 $protocols[] = preg_quote( $protocol, '/' );
 774                         }
 775                 }
 776
 777                 $retval = implode( '|', $protocols );
 778         } else {
 779                 // Ignore $includeProtocolRelative in this case
 780                 // This case exists for pre-1.6 compatibility, and we can safely assume
 781                 // that '//' won't appear in a pre-1.6 config because protocol-relative
 782                 // URLs weren't supported until 1.18
 783                 $retval = $wgUrlProtocols;
 784         }
 785
 786         // Cache return value
 787         if ( $includeProtocolRelative ) {
 788                 $withProtRel = $retval;
 789         } else {
 790                 $withoutProtRel = $retval;
 791         }
 792         return $retval;
 793 }
 794
 795 /**
 796  * Like wfUrlProtocols(), but excludes '//' from the protocol list. Use this if
 797  * you need a regex that matches all URL protocols but does not match protocol-
 798  * relative URLs
 799  * @return string
 800  */
 801 function wfUrlProtocolsWithoutProtRel() {
 802         return wfUrlProtocols( false );
 803 }
 804
 805 /**
 806  * parse_url() work-alike, but non-broken.  Differences:
 807  *
 808  * 1) Does not raise warnings on bad URLs (just returns false).
 809  * 2) Handles protocols that don't use :// (e.g., mailto: and news:, as well as
 810  *    protocol-relative URLs) correctly.
 811  * 3) Adds a "delimiter" element to the array, either '://', ':' or '//' (see (2)).
 812  *
 813  * @param string $url A URL to parse
 814  * @return string[]|bool Bits of the URL in an associative array, per PHP docs, false on failure
 815  */
 816 function wfParseUrl( $url ) {
 817         global $wgUrlProtocols; // Allow all protocols defined in DefaultSettings/LocalSettings.php
 818
 819         // Protocol-relative URLs are handled really badly by parse_url(). It's so
 820         // bad that the easiest way to handle them is to just prepend 'http:' and
 821         // strip the protocol out later.
 822         $wasRelative = substr( $url, 0, 2 ) == '//';
 823         if ( $wasRelative ) {
 824                 $url = "http:$url";
 825         }
 826         MediaWiki\suppressWarnings();
 827         $bits = parse_url( $url );
 828         MediaWiki\restoreWarnings();
 829         // parse_url() returns an array without scheme for some invalid URLs, e.g.
 830         // parse_url("%0Ahttp://example.com") == [ 'host' => '%0Ahttp', 'path' => 'example.com' ]
 831         if ( !$bits || !isset( $bits['scheme'] ) ) {
 832                 return false;
 833         }
 834
 835         // parse_url() incorrectly handles schemes case-sensitively. Convert it to lowercase.
 836         $bits['scheme'] = strtolower( $bits['scheme'] );
 837
 838         // most of the protocols are followed by ://, but mailto: and sometimes news: not, check for it
 839         if ( in_array( $bits['scheme'] . '://', $wgUrlProtocols ) ) {
 840                 $bits['delimiter'] = '://';
 841         } elseif ( in_array( $bits['scheme'] . ':', $wgUrlProtocols ) ) {
 842                 $bits['delimiter'] = ':';
 843                 // parse_url detects for news: and mailto: the host part of an url as path
 844                 // We have to correct this wrong detection
 845                 if ( isset( $bits['path'] ) ) {
 846                         $bits['host'] = $bits['path'];
 847                         $bits['path'] = '';
 848                 }
 849         } else {
 850                 return false;
 851         }
 852
 853         /* Provide an empty host for eg. file:/// urls (see bug 28627) */
 854         if ( !isset( $bits['host'] ) ) {
 855                 $bits['host'] = '';
 856
 857                 // bug 45069
 858                 if ( isset( $bits['path'] ) ) {
 859                         /* parse_url loses the third / for file:///c:/ urls (but not on variants) */
 860                         if ( substr( $bits['path'], 0, 1 ) !== '/' ) {
 861                                 $bits['path'] = '/' . $bits['path'];
 862                         }
 863                 } else {
 864                         $bits['path'] = '';
 865                 }
 866         }
 867
 868         // If the URL was protocol-relative, fix scheme and delimiter
 869         if ( $wasRelative ) {
 870                 $bits['scheme'] = '';
 871                 $bits['delimiter'] = '//';
 872         }
 873         return $bits;
 874 }
 875
 876 /**
 877  * Take a URL, make sure it's expanded to fully qualified, and replace any
 878  * encoded non-ASCII Unicode characters with their UTF-8 original forms
 879  * for more compact display and legibility for local audiences.
 880  *
 881  * @todo handle punycode domains too
 882  *
 883  * @param string $url
 884  * @return string
 885  */
 886 function wfExpandIRI( $url ) {
 887         return preg_replace_callback(
 888                 '/((?:%[89A-F][0-9A-F])+)/i',
 889                 'wfExpandIRI_callback',
 890                 wfExpandUrl( $url )
 891         );
 892 }
 893
 894 /**
 895  * Private callback for wfExpandIRI
 896  * @param array $matches
 897  * @return string
 898  */
 899 function wfExpandIRI_callback( $matches ) {
 900         return urldecode( $matches[1] );
 901 }
 902
 903 /**
 904  * Make URL indexes, appropriate for the el_index field of externallinks.
 905  *
 906  * @param string $url
 907  * @return array
 908  */
 909 function wfMakeUrlIndexes( $url ) {
 910         $bits = wfParseUrl( $url );
 911
 912         // Reverse the labels in the hostname, convert to lower case
 913         // For emails reverse domainpart only
 914         if ( $bits['scheme'] == 'mailto' ) {
 915                 $mailparts = explode( '@', $bits['host'], 2 );
 916                 if ( count( $mailparts ) === 2 ) {
 917                         $domainpart = strtolower( implode( '.', array_reverse( explode( '.', $mailparts[1] ) ) ) );
 918                 } else {
 919                         // No domain specified, don't mangle it
 920                         $domainpart = '';
 921                 }
 922                 $reversedHost = $domainpart . '@' . $mailparts[0];
 923         } else {
 924                 $reversedHost = strtolower( implode( '.', array_reverse( explode( '.', $bits['host'] ) ) ) );
 925         }
 926         // Add an extra dot to the end
 927         // Why? Is it in wrong place in mailto links?
 928         if ( substr( $reversedHost, -1, 1 ) !== '.' ) {
 929                 $reversedHost .= '.';
 930         }
 931         // Reconstruct the pseudo-URL
 932         $prot = $bits['scheme'];
 933         $index = $prot . $bits['delimiter'] . $reversedHost;
 934         // Leave out user and password. Add the port, path, query and fragment
 935         if ( isset( $bits['port'] ) ) {
 936                 $index .= ':' . $bits['port'];
 937         }
 938         if ( isset( $bits['path'] ) ) {
 939                 $index .= $bits['path'];
 940         } else {
 941                 $index .= '/';
 942         }
 943         if ( isset( $bits['query'] ) ) {
 944                 $index .= '?' . $bits['query'];
 945         }
 946         if ( isset( $bits['fragment'] ) ) {
 947                 $index .= '#' . $bits['fragment'];
 948         }
 949
 950         if ( $prot == '' ) {
 951                 return [ "http:$index", "https:$index" ];
 952         } else {
 953                 return [ $index ];
 954         }
 955 }
 956
 957 /**
 958  * Check whether a given URL has a domain that occurs in a given set of domains
 959  * @param string $url URL
 960  * @param array $domains Array of domains (strings)
 961  * @return bool True if the host part of $url ends in one of the strings in $domains
 962  */
 963 function wfMatchesDomainList( $url, $domains ) {
 964         $bits = wfParseUrl( $url );
 965         if ( is_array( $bits ) && isset( $bits['host'] ) ) {
 966                 $host = '.' . $bits['host'];
 967                 foreach ( (array)$domains as $domain ) {
 968                         $domain = '.' . $domain;
 969                         if ( substr( $host, -strlen( $domain ) ) === $domain ) {
 970                                 return true;
 971                         }
 972                 }
 973         }
 974         return false;
 975 }
 976
 977 /**
 978  * Sends a line to the debug log if enabled or, optionally, to a comment in output.
 979  * In normal operation this is a NOP.
 980  *
 981  * Controlling globals:
 982  * $wgDebugLogFile - points to the log file
 983  * $wgDebugRawPage - if false, 'action=raw' hits will not result in debug output.
 984  * $wgDebugComments - if on, some debug items may appear in comments in the HTML output.
 985  *
 986  * @since 1.25 support for additional context data
 987  *
 988  * @param string $text
 989  * @param string|bool $dest Destination of the message:
 990  *     - 'all': both to the log and HTML (debug toolbar or HTML comments)
 991  *     - 'private': excluded from HTML output
 992  *   For backward compatibility, it can also take a boolean:
 993  *     - true: same as 'all'
 994  *     - false: same as 'private'
 995  * @param array $context Additional logging context data
 996  */
 997 function wfDebug( $text, $dest = 'all', array $context = [] ) {
 998         global $wgDebugRawPage, $wgDebugLogPrefix;
 999         global $wgDebugTimestamps, $wgRequestTime;
1000
1001         if ( !$wgDebugRawPage && wfIsDebugRawPage() ) {
1002                 return;
1003         }
1004
1005         $text = trim( $text );
1006
1007         if ( $wgDebugTimestamps ) {
1008                 $context['seconds_elapsed'] = sprintf(
1009                         '%6.4f',
1010                         microtime( true ) - $wgRequestTime
1011                 );
1012                 $context['memory_used'] = sprintf(
1013                         '%5.1fM',
1014                         ( memory_get_usage( true ) / ( 1024 * 1024 ) )
1015                 );
1016         }
1017
1018         if ( $wgDebugLogPrefix !== '' ) {
1019                 $context['prefix'] = $wgDebugLogPrefix;
1020         }
1021         $context['private'] = ( $dest === false || $dest === 'private' );
1022
1023         $logger = LoggerFactory::getInstance( 'wfDebug' );
1024         $logger->debug( $text, $context );
1025 }
1026
1027 /**
1028  * Returns true if debug logging should be suppressed if $wgDebugRawPage = false
1029  * @return bool
1030  */
1031 function wfIsDebugRawPage() {
1032         static $cache;
1033         if ( $cache !== null ) {
1034                 return $cache;
1035         }
1036         # Check for raw action using $_GET not $wgRequest, since the latter might not be initialised yet
1037         if ( ( isset( $_GET['action'] ) && $_GET['action'] == 'raw' )
1038                 || (
1039                         isset( $_SERVER['SCRIPT_NAME'] )
1040                         && substr( $_SERVER['SCRIPT_NAME'], -8 ) == 'load.php'
1041                 )
1042         ) {
1043                 $cache = true;
1044         } else {
1045                 $cache = false;
1046         }
1047         return $cache;
1048 }
1049
1050 /**
1051  * Send a line giving PHP memory usage.
1052  *
1053  * @param bool $exact Print exact byte values instead of kibibytes (default: false)
1054  */
1055 function wfDebugMem( $exact = false ) {
1056         $mem = memory_get_usage();
1057         if ( !$exact ) {
1058                 $mem = floor( $mem / 1024 ) . ' KiB';
1059         } else {
1060                 $mem .= ' B';
1061         }
1062         wfDebug( "Memory usage: $mem\n" );
1063 }
1064
1065 /**
1066  * Send a line to a supplementary debug log file, if configured, or main debug
1067  * log if not.
1068  *
1069  * To configure a supplementary log file, set $wgDebugLogGroups[$logGroup] to
1070  * a string filename or an associative array mapping 'destination' to the
1071  * desired filename. The associative array may also contain a 'sample' key
1072  * with an integer value, specifying a sampling factor. Sampled log events
1073  * will be emitted with a 1 in N random chance.
1074  *
1075  * @since 1.23 support for sampling log messages via $wgDebugLogGroups.
1076  * @since 1.25 support for additional context data
1077  * @since 1.25 sample behavior dependent on configured $wgMWLoggerDefaultSpi
1078  *
1079  * @param string $logGroup
1080  * @param string $text
1081  * @param string|bool $dest Destination of the message:
1082  *     - 'all': both to the log and HTML (debug toolbar or HTML comments)
1083  *     - 'private': only to the specific log if set in $wgDebugLogGroups and
1084  *       discarded otherwise
1085  *   For backward compatibility, it can also take a boolean:
1086  *     - true: same as 'all'
1087  *     - false: same as 'private'
1088  * @param array $context Additional logging context data
1089  */
1090 function wfDebugLog(
1091         $logGroup, $text, $dest = 'all', array $context = []
1092 ) {
1093         $text = trim( $text );
1094
1095         $logger = LoggerFactory::getInstance( $logGroup );
1096         $context['private'] = ( $dest === false || $dest === 'private' );
1097         $logger->info( $text, $context );
1098 }
1099
1100 /**
1101  * Log for database errors
1102  *
1103  * @since 1.25 support for additional context data
1104  *
1105  * @param string $text Database error message.
1106  * @param array $context Additional logging context data
1107  */
1108 function wfLogDBError( $text, array $context = [] ) {
1109         $logger = LoggerFactory::getInstance( 'wfLogDBError' );
1110         $logger->error( trim( $text ), $context );
1111 }
1112
1113 /**
1114  * Throws a warning that $function is deprecated
1115  *
1116  * @param string $function
1117  * @param string|bool $version Version of MediaWiki that the function
1118  *    was deprecated in (Added in 1.19).
1119  * @param string|bool $component Added in 1.19.
1120  * @param int $callerOffset How far up the call stack is the original
1121  *    caller. 2 = function that called the function that called
1122  *    wfDeprecated (Added in 1.20)
1123  *
1124  * @return null
1125  */
1126 function wfDeprecated( $function, $version = false, $component = false, $callerOffset = 2 ) {
1127         MWDebug::deprecated( $function, $version, $component, $callerOffset + 1 );
1128 }
1129
1130 /**
1131  * Send a warning either to the debug log or in a PHP error depending on
1132  * $wgDevelopmentWarnings. To log warnings in production, use wfLogWarning() instead.
1133  *
1134  * @param string $msg Message to send
1135  * @param int $callerOffset Number of items to go back in the backtrace to
1136  *        find the correct caller (1 = function calling wfWarn, ...)
1137  * @param int $level PHP error level; defaults to E_USER_NOTICE;
1138  *        only used when $wgDevelopmentWarnings is true
1139  */
1140 function wfWarn( $msg, $callerOffset = 1, $level = E_USER_NOTICE ) {
1141         MWDebug::warning( $msg, $callerOffset + 1, $level, 'auto' );
1142 }
1143
1144 /**
1145  * Send a warning as a PHP error and the debug log. This is intended for logging
1146  * warnings in production. For logging development warnings, use WfWarn instead.
1147  *
1148  * @param string $msg Message to send
1149  * @param int $callerOffset Number of items to go back in the backtrace to
1150  *        find the correct caller (1 = function calling wfLogWarning, ...)
1151  * @param int $level PHP error level; defaults to E_USER_WARNING
1152  */
1153 function wfLogWarning( $msg, $callerOffset = 1, $level = E_USER_WARNING ) {
1154         MWDebug::warning( $msg, $callerOffset + 1, $level, 'production' );
1155 }
1156
1157 /**
1158  * Log to a file without getting "file size exceeded" signals.
1159  *
1160  * Can also log to TCP or UDP with the syntax udp://host:port/prefix. This will
1161  * send lines to the specified port, prefixed by the specified prefix and a space.
1162  * @since 1.25 support for additional context data
1163  *
1164  * @param string $text
1165  * @param string $file Filename
1166  * @param array $context Additional logging context data
1167  * @throws MWException
1168  * @deprecated since 1.25 Use \MediaWiki\Logger\LegacyLogger::emit or UDPTransport
1169  */
1170 function wfErrorLog( $text, $file, array $context = [] ) {
1171         wfDeprecated( __METHOD__, '1.25' );
1172         $logger = LoggerFactory::getInstance( 'wfErrorLog' );
1173         $context['destination'] = $file;
1174         $logger->info( trim( $text ), $context );
1175 }
1176
1177 /**
1178  * @todo document
1179  */
1180 function wfLogProfilingData() {
1181         global $wgDebugLogGroups, $wgDebugRawPage;
1182
1183         $context = RequestContext::getMain();
1184         $request = $context->getRequest();
1185
1186         $profiler = Profiler::instance();
1187         $profiler->setContext( $context );
1188         $profiler->logData();
1189
1190         $config = $context->getConfig();
1191         if ( $config->get( 'StatsdServer' ) ) {
1192                 try {
1193                         $statsdServer = explode( ':', $config->get( 'StatsdServer' ) );
1194                         $statsdHost = $statsdServer[0];
1195                         $statsdPort = isset( $statsdServer[1] ) ? $statsdServer[1] : 8125;
1196                         $statsdSender = new SocketSender( $statsdHost, $statsdPort );
1197                         $statsdClient = new SamplingStatsdClient( $statsdSender, true, false );
1198                         $statsdClient->send( $context->getStats()->getBuffer() );
1199                 } catch ( Exception $ex ) {
1200                         MWExceptionHandler::logException( $ex );
1201                 }
1202         }
1203
1204         # Profiling must actually be enabled...
1205         if ( $profiler instanceof ProfilerStub ) {
1206                 return;
1207         }
1208
1209         if ( isset( $wgDebugLogGroups['profileoutput'] )
1210                 && $wgDebugLogGroups['profileoutput'] === false
1211         ) {
1212                 // Explicitly disabled
1213                 return;
1214         }
1215         if ( !$wgDebugRawPage && wfIsDebugRawPage() ) {
1216                 return;
1217         }
1218
1219         $ctx = [ 'elapsed' => $request->getElapsedTime() ];
1220         if ( !empty( $_SERVER['HTTP_X_FORWARDED_FOR'] ) ) {
1221                 $ctx['forwarded_for'] = $_SERVER['HTTP_X_FORWARDED_FOR'];
1222         }
1223         if ( !empty( $_SERVER['HTTP_CLIENT_IP'] ) ) {
1224                 $ctx['client_ip'] = $_SERVER['HTTP_CLIENT_IP'];
1225         }
1226         if ( !empty( $_SERVER['HTTP_FROM'] ) ) {
1227                 $ctx['from'] = $_SERVER['HTTP_FROM'];
1228         }
1229         if ( isset( $ctx['forwarded_for'] ) ||
1230                 isset( $ctx['client_ip'] ) ||
1231                 isset( $ctx['from'] ) ) {
1232                 $ctx['proxy'] = $_SERVER['REMOTE_ADDR'];
1233         }
1234
1235         // Don't load $wgUser at this late stage just for statistics purposes
1236         // @todo FIXME: We can detect some anons even if it is not loaded.
1237         // See User::getId()
1238         $user = $context->getUser();
1239         $ctx['anon'] = $user->isItemLoaded( 'id' ) && $user->isAnon();
1240
1241         // Command line script uses a FauxRequest object which does not have
1242         // any knowledge about an URL and throw an exception instead.
1243         try {
1244                 $ctx['url'] = urldecode( $request->getRequestURL() );
1245         } catch ( Exception $ignored ) {
1246                 // no-op
1247         }
1248
1249         $ctx['output'] = $profiler->getOutput();
1250
1251         $log = LoggerFactory::getInstance( 'profileoutput' );
1252         $log->info( "Elapsed: {elapsed}; URL: <{url}>\n{output}", $ctx );
1253 }
1254
1255 /**
1256  * Increment a statistics counter
1257  *
1258  * @param string $key
1259  * @param int $count
1260  * @return void
1261  */
1262 function wfIncrStats( $key, $count = 1 ) {
1263         $stats = RequestContext::getMain()->getStats();
1264         $stats->updateCount( $key, $count );
1265 }
1266
1267 /**
1268  * Check whether the wiki is in read-only mode.
1269  *
1270  * @return bool
1271  */
1272 function wfReadOnly() {
1273         return wfReadOnlyReason() !== false;
1274 }
1275
1276 /**
1277  * Check if the site is in read-only mode and return the message if so
1278  *
1279  * This checks wfConfiguredReadOnlyReason() and the main load balancer
1280  * for slave lag. This may result in DB_SLAVE connection being made.
1281  *
1282  * @return string|bool String when in read-only mode; false otherwise
1283  */
1284 function wfReadOnlyReason() {
1285         $readOnly = wfConfiguredReadOnlyReason();
1286         if ( $readOnly !== false ) {
1287                 return $readOnly;
1288         }
1289
1290         static $lbReadOnly = null;
1291         if ( $lbReadOnly === null ) {
1292                 // Callers use this method to be aware that data presented to a user
1293                 // may be very stale and thus allowing submissions can be problematic.
1294                 $lbReadOnly = wfGetLB()->getReadOnlyReason();
1295         }
1296
1297         return $lbReadOnly;
1298 }
1299
1300 /**
1301  * Get the value of $wgReadOnly or the contents of $wgReadOnlyFile.
1302  *
1303  * @return string|bool String when in read-only mode; false otherwise
1304  * @since 1.27
1305  */
1306 function wfConfiguredReadOnlyReason() {
1307         global $wgReadOnly, $wgReadOnlyFile;
1308
1309         if ( $wgReadOnly === null ) {
1310                 // Set $wgReadOnly for faster access next time
1311                 if ( is_file( $wgReadOnlyFile ) && filesize( $wgReadOnlyFile ) > 0 ) {
1312                         $wgReadOnly = file_get_contents( $wgReadOnlyFile );
1313                 } else {
1314                         $wgReadOnly = false;
1315                 }
1316         }
1317
1318         return $wgReadOnly;
1319 }
1320
1321 /**
1322  * Return a Language object from $langcode
1323  *
1324  * @param Language|string|bool $langcode Either:
1325  *                  - a Language object
1326  *                  - code of the language to get the message for, if it is
1327  *                    a valid code create a language for that language, if
1328  *                    it is a string but not a valid code then make a basic
1329  *                    language object
1330  *                  - a boolean: if it's false then use the global object for
1331  *                    the current user's language (as a fallback for the old parameter
1332  *                    functionality), or if it is true then use global object
1333  *                    for the wiki's content language.
1334  * @return Language
1335  */
1336 function wfGetLangObj( $langcode = false ) {
1337         # Identify which language to get or create a language object for.
1338         # Using is_object here due to Stub objects.
1339         if ( is_object( $langcode ) ) {
1340                 # Great, we already have the object (hopefully)!
1341                 return $langcode;
1342         }
1343
1344         global $wgContLang, $wgLanguageCode;
1345         if ( $langcode === true || $langcode === $wgLanguageCode ) {
1346                 # $langcode is the language code of the wikis content language object.
1347                 # or it is a boolean and value is true
1348                 return $wgContLang;
1349         }
1350
1351         global $wgLang;
1352         if ( $langcode === false || $langcode === $wgLang->getCode() ) {
1353                 # $langcode is the language code of user language object.
1354                 # or it was a boolean and value is false
1355                 return $wgLang;
1356         }
1357
1358         $validCodes = array_keys( Language::fetchLanguageNames() );
1359         if ( in_array( $langcode, $validCodes ) ) {
1360                 # $langcode corresponds to a valid language.
1361                 return Language::factory( $langcode );
1362         }
1363
1364         # $langcode is a string, but not a valid language code; use content language.
1365         wfDebug( "Invalid language code passed to wfGetLangObj, falling back to content language.\n" );
1366         return $wgContLang;
1367 }
1368
1369 /**
1370  * This is the function for getting translated interface messages.
1371  *
1372  * @see Message class for documentation how to use them.
1373  * @see https://www.mediawiki.org/wiki/Manual:Messages_API
1374  *
1375  * This function replaces all old wfMsg* functions.
1376  *
1377  * @param string|string[]|MessageSpecifier $key Message key, or array of keys, or a MessageSpecifier
1378  * @param mixed $params,... Normal message parameters
1379  * @return Message
1380  *
1381  * @since 1.17
1382  *
1383  * @see Message::__construct
1384  */
1385 function wfMessage( $key /*...*/ ) {
1386         $params = func_get_args();
1387         array_shift( $params );
1388         if ( isset( $params[0] ) && is_array( $params[0] ) ) {
1389                 $params = $params[0];
1390         }
1391         return new Message( $key, $params );
1392 }
1393
1394 /**
1395  * This function accepts multiple message keys and returns a message instance
1396  * for the first message which is non-empty. If all messages are empty then an
1397  * instance of the first message key is returned.
1398  *
1399  * @param string|string[] $keys,... Message keys
1400  * @return Message
1401  *
1402  * @since 1.18
1403  *
1404  * @see Message::newFallbackSequence
1405  */
1406 function wfMessageFallback( /*...*/ ) {
1407         $args = func_get_args();
1408         return call_user_func_array( 'Message::newFallbackSequence', $args );
1409 }
1410
1411 /**
1412  * Replace message parameter keys on the given formatted output.
1413  *
1414  * @param string $message
1415  * @param array $args
1416  * @return string
1417  * @private
1418  */
1419 function wfMsgReplaceArgs( $message, $args ) {
1420         # Fix windows line-endings
1421         # Some messages are split with explode("\n", $msg)
1422         $message = str_replace( "\r", '', $message );
1423
1424         // Replace arguments
1425         if ( is_array( $args ) && $args ) {
1426                 if ( is_array( $args[0] ) ) {
1427                         $args = array_values( $args[0] );
1428                 }
1429                 $replacementKeys = [];
1430                 foreach ( $args as $n => $param ) {
1431                         $replacementKeys['$' . ( $n + 1 )] = $param;
1432                 }
1433                 $message = strtr( $message, $replacementKeys );
1434         }
1435
1436         return $message;
1437 }
1438
1439 /**
1440  * Fetch server name for use in error reporting etc.
1441  * Use real server name if available, so we know which machine
1442  * in a server farm generated the current page.
1443  *
1444  * @return string
1445  */
1446 function wfHostname() {
1447         static $host;
1448         if ( is_null( $host ) ) {
1449
1450                 # Hostname overriding
1451                 global $wgOverrideHostname;
1452                 if ( $wgOverrideHostname !== false ) {
1453                         # Set static and skip any detection
1454                         $host = $wgOverrideHostname;
1455                         return $host;
1456                 }
1457
1458                 if ( function_exists( 'posix_uname' ) ) {
1459                         // This function not present on Windows
1460                         $uname = posix_uname();
1461                 } else {
1462                         $uname = false;
1463                 }
1464                 if ( is_array( $uname ) && isset( $uname['nodename'] ) ) {
1465                         $host = $uname['nodename'];
1466                 } elseif ( getenv( 'COMPUTERNAME' ) ) {
1467                         # Windows computer name
1468                         $host = getenv( 'COMPUTERNAME' );
1469                 } else {
1470                         # This may be a virtual server.
1471                         $host = $_SERVER['SERVER_NAME'];
1472                 }
1473         }
1474         return $host;
1475 }
1476
1477 /**
1478  * Returns a script tag that stores the amount of time it took MediaWiki to
1479  * handle the request in milliseconds as 'wgBackendResponseTime'.
1480  *
1481  * If $wgShowHostnames is true, the script will also set 'wgHostname' to the
1482  * hostname of the server handling the request.
1483  *
1484  * @return string
1485  */
1486 function wfReportTime() {
1487         global $wgRequestTime, $wgShowHostnames;
1488
1489         $responseTime = round( ( microtime( true ) - $wgRequestTime ) * 1000 );
1490         $reportVars = [ 'wgBackendResponseTime' => $responseTime ];
1491         if ( $wgShowHostnames ) {
1492                 $reportVars['wgHostname'] = wfHostname();
1493         }
1494         return Skin::makeVariablesScript( $reportVars );
1495 }
1496
1497 /**
1498  * Safety wrapper for debug_backtrace().
1499  *
1500  * Will return an empty array if debug_backtrace is disabled, otherwise
1501  * the output from debug_backtrace() (trimmed).
1502  *
1503  * @param int $limit This parameter can be used to limit the number of stack frames returned
1504  *
1505  * @return array Array of backtrace information
1506  */
1507 function wfDebugBacktrace( $limit = 0 ) {
1508         static $disabled = null;
1509
1510         if ( is_null( $disabled ) ) {
1511                 $disabled = !function_exists( 'debug_backtrace' );
1512                 if ( $disabled ) {
1513                         wfDebug( "debug_backtrace() is disabled\n" );
1514                 }
1515         }
1516         if ( $disabled ) {
1517                 return [];
1518         }
1519
1520         if ( $limit ) {
1521                 return array_slice( debug_backtrace( DEBUG_BACKTRACE_PROVIDE_OBJECT, $limit + 1 ), 1 );
1522         } else {
1523                 return array_slice( debug_backtrace(), 1 );
1524         }
1525 }
1526
1527 /**
1528  * Get a debug backtrace as a string
1529  *
1530  * @param bool|null $raw If true, the return value is plain text. If false, HTML.
1531  *   Defaults to $wgCommandLineMode if unset.
1532  * @return string
1533  * @since 1.25 Supports $raw parameter.
1534  */
1535 function wfBacktrace( $raw = null ) {
1536         global $wgCommandLineMode;
1537
1538         if ( $raw === null ) {
1539                 $raw = $wgCommandLineMode;
1540         }
1541
1542         if ( $raw ) {
1543                 $frameFormat = "%s line %s calls %s()\n";
1544                 $traceFormat = "%s";
1545         } else {
1546                 $frameFormat = "<li>%s line %s calls %s()</li>\n";
1547                 $traceFormat = "<ul>\n%s</ul>\n";
1548         }
1549
1550         $frames = array_map( function ( $frame ) use ( $frameFormat ) {
1551                 $file = !empty( $frame['file'] ) ? basename( $frame['file'] ) : '-';
1552                 $line = isset( $frame['line'] ) ? $frame['line'] : '-';
1553                 $call = $frame['function'];
1554                 if ( !empty( $frame['class'] ) ) {
1555                         $call = $frame['class'] . $frame['type'] . $call;
1556                 }
1557                 return sprintf( $frameFormat, $file, $line, $call );
1558         }, wfDebugBacktrace() );
1559
1560         return sprintf( $traceFormat, implode( '', $frames ) );
1561 }
1562
1563 /**
1564  * Get the name of the function which called this function
1565  * wfGetCaller( 1 ) is the function with the wfGetCaller() call (ie. __FUNCTION__)
1566  * wfGetCaller( 2 ) [default] is the caller of the function running wfGetCaller()
1567  * wfGetCaller( 3 ) is the parent of that.
1568  *
1569  * @param int $level
1570  * @return string
1571  */
1572 function wfGetCaller( $level = 2 ) {
1573         $backtrace = wfDebugBacktrace( $level + 1 );
1574         if ( isset( $backtrace[$level] ) ) {
1575                 return wfFormatStackFrame( $backtrace[$level] );
1576         } else {
1577                 return 'unknown';
1578         }
1579 }
1580
1581 /**
1582  * Return a string consisting of callers in the stack. Useful sometimes
1583  * for profiling specific points.
1584  *
1585  * @param int $limit The maximum depth of the stack frame to return, or false for the entire stack.
1586  * @return string
1587  */
1588 function wfGetAllCallers( $limit = 3 ) {
1589         $trace = array_reverse( wfDebugBacktrace() );
1590         if ( !$limit || $limit > count( $trace ) - 1 ) {
1591                 $limit = count( $trace ) - 1;
1592         }
1593         $trace = array_slice( $trace, -$limit - 1, $limit );
1594         return implode( '/', array_map( 'wfFormatStackFrame', $trace ) );
1595 }
1596
1597 /**
1598  * Return a string representation of frame
1599  *
1600  * @param array $frame
1601  * @return string
1602  */
1603 function wfFormatStackFrame( $frame ) {
1604         if ( !isset( $frame['function'] ) ) {
1605                 return 'NO_FUNCTION_GIVEN';
1606         }
1607         return isset( $frame['class'] ) && isset( $frame['type'] ) ?
1608                 $frame['class'] . $frame['type'] . $frame['function'] :
1609                 $frame['function'];
1610 }
1611
1612 /* Some generic result counters, pulled out of SearchEngine */
1613
1614 /**
1615  * @todo document
1616  *
1617  * @param int $offset
1618  * @param int $limit
1619  * @return string
1620  */
1621 function wfShowingResults( $offset, $limit ) {
1622         return wfMessage( 'showingresults' )->numParams( $limit, $offset + 1 )->parse();
1623 }
1624
1625 /**
1626  * @todo document
1627  * @todo FIXME: We may want to blacklist some broken browsers
1628  *
1629  * @param bool $force
1630  * @return bool Whereas client accept gzip compression
1631  */
1632 function wfClientAcceptsGzip( $force = false ) {
1633         static $result = null;
1634         if ( $result === null || $force ) {
1635                 $result = false;
1636                 if ( isset( $_SERVER['HTTP_ACCEPT_ENCODING'] ) ) {
1637                         # @todo FIXME: We may want to blacklist some broken browsers
1638                         $m = [];
1639                         if ( preg_match(
1640                                         '/\bgzip(?:;(q)=([0-9]+(?:\.[0-9]+)))?\b/',
1641                                         $_SERVER['HTTP_ACCEPT_ENCODING'],
1642                                         $m
1643                                 )
1644                         ) {
1645                                 if ( isset( $m[2] ) && ( $m[1] == 'q' ) && ( $m[2] == 0 ) ) {
1646                                         $result = false;
1647                                         return $result;
1648                                 }
1649                                 wfDebug( "wfClientAcceptsGzip: client accepts gzip.\n" );
1650                                 $result = true;
1651                         }
1652                 }
1653         }
1654         return $result;
1655 }
1656
1657 /**
1658  * Escapes the given text so that it may be output using addWikiText()
1659  * without any linking, formatting, etc. making its way through. This
1660  * is achieved by substituting certain characters with HTML entities.
1661  * As required by the callers, "<nowiki>" is not used.
1662  *
1663  * @param string $text Text to be escaped
1664  * @return string
1665  */
1666 function wfEscapeWikiText( $text ) {
1667         static $repl = null, $repl2 = null;
1668         if ( $repl === null ) {
1669                 $repl = [
1670                         '"' => '&#34;', '&' => '&#38;', "'" => '&#39;', '<' => '&#60;',
1671                         '=' => '&#61;', '>' => '&#62;', '[' => '&#91;', ']' => '&#93;',
1672                         '{' => '&#123;', '|' => '&#124;', '}' => '&#125;', ';' => '&#59;',
1673                         "\n#" => "\n&#35;", "\r#" => "\r&#35;",
1674                         "\n*" => "\n&#42;", "\r*" => "\r&#42;",
1675                         "\n:" => "\n&#58;", "\r:" => "\r&#58;",
1676                         "\n " => "\n&#32;", "\r " => "\r&#32;",
1677                         "\n\n" => "\n&#10;", "\r\n" => "&#13;\n",
1678                         "\n\r" => "\n&#13;", "\r\r" => "\r&#13;",
1679                         "\n\t" => "\n&#9;", "\r\t" => "\r&#9;", // "\n\t\n" is treated like "\n\n"
1680                         "\n----" => "\n&#45;---", "\r----" => "\r&#45;---",
1681                         '__' => '_&#95;', '://' => '&#58;//',
1682                 ];
1683
1684                 // We have to catch everything "\s" matches in PCRE
1685                 foreach ( [ 'ISBN', 'RFC', 'PMID' ] as $magic ) {
1686                         $repl["$magic "] = "$magic&#32;";
1687                         $repl["$magic\t"] = "$magic&#9;";
1688                         $repl["$magic\r"] = "$magic&#13;";
1689                         $repl["$magic\n"] = "$magic&#10;";
1690                         $repl["$magic\f"] = "$magic&#12;";
1691                 }
1692
1693                 // And handle protocols that don't use "://"
1694                 global $wgUrlProtocols;
1695                 $repl2 = [];
1696                 foreach ( $wgUrlProtocols as $prot ) {
1697                         if ( substr( $prot, -1 ) === ':' ) {
1698                                 $repl2[] = preg_quote( substr( $prot, 0, -1 ), '/' );
1699                         }
1700                 }
1701                 $repl2 = $repl2 ? '/\b(' . implode( '|', $repl2 ) . '):/i' : '/^(?!)/';
1702         }
1703         $text = substr( strtr( "\n$text", $repl ), 1 );
1704         $text = preg_replace( $repl2, '$1&#58;', $text );
1705         return $text;
1706 }
1707
1708 /**
1709  * Sets dest to source and returns the original value of dest
1710  * If source is NULL, it just returns the value, it doesn't set the variable
1711  * If force is true, it will set the value even if source is NULL
1712  *
1713  * @param mixed $dest
1714  * @param mixed $source
1715  * @param bool $force
1716  * @return mixed
1717  */
1718 function wfSetVar( &$dest, $source, $force = false ) {
1719         $temp = $dest;
1720         if ( !is_null( $source ) || $force ) {
1721                 $dest = $source;
1722         }
1723         return $temp;
1724 }
1725
1726 /**
1727  * As for wfSetVar except setting a bit
1728  *
1729  * @param int $dest
1730  * @param int $bit
1731  * @param bool $state
1732  *
1733  * @return bool
1734  */
1735 function wfSetBit( &$dest, $bit, $state = true ) {
1736         $temp = (bool)( $dest & $bit );
1737         if ( !is_null( $state ) ) {
1738                 if ( $state ) {
1739                         $dest |= $bit;
1740                 } else {
1741                         $dest &= ~$bit;
1742                 }
1743         }
1744         return $temp;
1745 }
1746
1747 /**
1748  * A wrapper around the PHP function var_export().
1749  * Either print it or add it to the regular output ($wgOut).
1750  *
1751  * @param mixed $var A PHP variable to dump.
1752  */
1753 function wfVarDump( $var ) {
1754         global $wgOut;
1755         $s = str_replace( "\n", "<br />\n", var_export( $var, true ) . "\n" );
1756         if ( headers_sent() || !isset( $wgOut ) || !is_object( $wgOut ) ) {
1757                 print $s;
1758         } else {
1759                 $wgOut->addHTML( $s );
1760         }
1761 }
1762
1763 /**
1764  * Provide a simple HTTP error.
1765  *
1766  * @param int|string $code
1767  * @param string $label
1768  * @param string $desc
1769  */
1770 function wfHttpError( $code, $label, $desc ) {
1771         global $wgOut;
1772         HttpStatus::header( $code );
1773         if ( $wgOut ) {
1774                 $wgOut->disable();
1775                 $wgOut->sendCacheControl();
1776         }
1777
1778         header( 'Content-type: text/html; charset=utf-8' );
1779         print '<!DOCTYPE html>' .
1780                 '<html><head><title>' .
1781                 htmlspecialchars( $label ) .
1782                 '</title></head><body><h1>' .
1783                 htmlspecialchars( $label ) .
1784                 '</h1><p>' .
1785                 nl2br( htmlspecialchars( $desc ) ) .
1786                 "</p></body></html>\n";
1787 }
1788
1789 /**
1790  * Clear away any user-level output buffers, discarding contents.
1791  *
1792  * Suitable for 'starting afresh', for instance when streaming
1793  * relatively large amounts of data without buffering, or wanting to
1794  * output image files without ob_gzhandler's compression.
1795  *
1796  * The optional $resetGzipEncoding parameter controls suppression of
1797  * the Content-Encoding header sent by ob_gzhandler; by default it
1798  * is left. See comments for wfClearOutputBuffers() for why it would
1799  * be used.
1800  *
1801  * Note that some PHP configuration options may add output buffer
1802  * layers which cannot be removed; these are left in place.
1803  *
1804  * @param bool $resetGzipEncoding
1805  */
1806 function wfResetOutputBuffers( $resetGzipEncoding = true ) {
1807         if ( $resetGzipEncoding ) {
1808                 // Suppress Content-Encoding and Content-Length
1809                 // headers from 1.10+s wfOutputHandler
1810                 global $wgDisableOutputCompression;
1811                 $wgDisableOutputCompression = true;
1812         }
1813         while ( $status = ob_get_status() ) {
1814                 if ( isset( $status['flags'] ) ) {
1815                         $flags = PHP_OUTPUT_HANDLER_CLEANABLE | PHP_OUTPUT_HANDLER_REMOVABLE;
1816                         $deleteable = ( $status['flags'] & $flags ) === $flags;
1817                 } elseif ( isset( $status['del'] ) ) {
1818                         $deleteable = $status['del'];
1819                 } else {
1820                         // Guess that any PHP-internal setting can't be removed.
1821                         $deleteable = $status['type'] !== 0; /* PHP_OUTPUT_HANDLER_INTERNAL */
1822                 }
1823                 if ( !$deleteable ) {
1824                         // Give up, and hope the result doesn't break
1825                         // output behavior.
1826                         break;
1827                 }
1828                 if ( $status['name'] === 'MediaWikiTestCase::wfResetOutputBuffersBarrier' ) {
1829                         // Unit testing barrier to prevent this function from breaking PHPUnit.
1830                         break;
1831                 }
1832                 if ( !ob_end_clean() ) {
1833                         // Could not remove output buffer handler; abort now
1834                         // to avoid getting in some kind of infinite loop.
1835                         break;
1836                 }
1837                 if ( $resetGzipEncoding ) {
1838                         if ( $status['name'] == 'ob_gzhandler' ) {
1839                                 // Reset the 'Content-Encoding' field set by this handler
1840                                 // so we can start fresh.
1841                                 header_remove( 'Content-Encoding' );
1842                                 break;
1843                         }
1844                 }
1845         }
1846 }
1847
1848 /**
1849  * More legible than passing a 'false' parameter to wfResetOutputBuffers():
1850  *
1851  * Clear away output buffers, but keep the Content-Encoding header
1852  * produced by ob_gzhandler, if any.
1853  *
1854  * This should be used for HTTP 304 responses, where you need to
1855  * preserve the Content-Encoding header of the real result, but
1856  * also need to suppress the output of ob_gzhandler to keep to spec
1857  * and avoid breaking Firefox in rare cases where the headers and
1858  * body are broken over two packets.
1859  */
1860 function wfClearOutputBuffers() {
1861         wfResetOutputBuffers( false );
1862 }
1863
1864 /**
1865  * Converts an Accept-* header into an array mapping string values to quality
1866  * factors
1867  *
1868  * @param string $accept
1869  * @param string $def Default
1870  * @return float[] Associative array of string => float pairs
1871  */
1872 function wfAcceptToPrefs( $accept, $def = '*/*' ) {
1873         # No arg means accept anything (per HTTP spec)
1874         if ( !$accept ) {
1875                 return [ $def => 1.0 ];
1876         }
1877
1878         $prefs = [];
1879
1880         $parts = explode( ',', $accept );
1881
1882         foreach ( $parts as $part ) {
1883                 # @todo FIXME: Doesn't deal with params like 'text/html; level=1'
1884                 $values = explode( ';', trim( $part ) );
1885                 $match = [];
1886                 if ( count( $values ) == 1 ) {
1887                         $prefs[$values[0]] = 1.0;
1888                 } elseif ( preg_match( '/q\s*=\s*(\d*\.\d+)/', $values[1], $match ) ) {
1889                         $prefs[$values[0]] = floatval( $match[1] );
1890                 }
1891         }
1892
1893         return $prefs;
1894 }
1895
1896 /**
1897  * Checks if a given MIME type matches any of the keys in the given
1898  * array. Basic wildcards are accepted in the array keys.
1899  *
1900  * Returns the matching MIME type (or wildcard) if a match, otherwise
1901  * NULL if no match.
1902  *
1903  * @param string $type
1904  * @param array $avail
1905  * @return string
1906  * @private
1907  */
1908 function mimeTypeMatch( $type, $avail ) {
1909         if ( array_key_exists( $type, $avail ) ) {
1910                 return $type;
1911         } else {
1912                 $mainType = explode( '/', $type )[0];
1913                 if ( array_key_exists( "$mainType/*", $avail ) ) {
1914                         return "$mainType/*";
1915                 } elseif ( array_key_exists( '*/*', $avail ) ) {
1916                         return '*/*';
1917                 } else {
1918                         return null;
1919                 }
1920         }
1921 }
1922
1923 /**
1924  * Returns the 'best' match between a client's requested internet media types
1925  * and the server's list of available types. Each list should be an associative
1926  * array of type to preference (preference is a float between 0.0 and 1.0).
1927  * Wildcards in the types are acceptable.
1928  *
1929  * @param array $cprefs Client's acceptable type list
1930  * @param array $sprefs Server's offered types
1931  * @return string
1932  *
1933  * @todo FIXME: Doesn't handle params like 'text/plain; charset=UTF-8'
1934  * XXX: generalize to negotiate other stuff
1935  */
1936 function wfNegotiateType( $cprefs, $sprefs ) {
1937         $combine = [];
1938
1939         foreach ( array_keys( $sprefs ) as $type ) {
1940                 $subType = explode( '/', $type )[1];
1941                 if ( $subType != '*' ) {
1942                         $ckey = mimeTypeMatch( $type, $cprefs );
1943                         if ( $ckey ) {
1944                                 $combine[$type] = $sprefs[$type] * $cprefs[$ckey];
1945                         }
1946                 }
1947         }
1948
1949         foreach ( array_keys( $cprefs ) as $type ) {
1950                 $subType = explode( '/', $type )[1];
1951                 if ( $subType != '*' && !array_key_exists( $type, $sprefs ) ) {
1952                         $skey = mimeTypeMatch( $type, $sprefs );
1953                         if ( $skey ) {
1954                                 $combine[$type] = $sprefs[$skey] * $cprefs[$type];
1955                         }
1956                 }
1957         }
1958
1959         $bestq = 0;
1960         $besttype = null;
1961
1962         foreach ( array_keys( $combine ) as $type ) {
1963                 if ( $combine[$type] > $bestq ) {
1964                         $besttype = $type;
1965                         $bestq = $combine[$type];
1966                 }
1967         }
1968
1969         return $besttype;
1970 }
1971
1972 /**
1973  * Reference-counted warning suppression
1974  *
1975  * @deprecated since 1.26, use MediaWiki\suppressWarnings() directly
1976  * @param bool $end
1977  */
1978 function wfSuppressWarnings( $end = false ) {
1979         MediaWiki\suppressWarnings( $end );
1980 }
1981
1982 /**
1983  * @deprecated since 1.26, use MediaWiki\restoreWarnings() directly
1984  * Restore error level to previous value
1985  */
1986 function wfRestoreWarnings() {
1987         MediaWiki\suppressWarnings( true );
1988 }
1989
1990 # Autodetect, convert and provide timestamps of various types
1991
1992 /**
1993  * Unix time - the number of seconds since 1970-01-01 00:00:00 UTC
1994  */
1995 define( 'TS_UNIX', 0 );
1996
1997 /**
1998  * MediaWiki concatenated string timestamp (YYYYMMDDHHMMSS)
1999  */
2000 define( 'TS_MW', 1 );
2001
2002 /**
2003  * MySQL DATETIME (YYYY-MM-DD HH:MM:SS)
2004  */
2005 define( 'TS_DB', 2 );
2006
2007 /**
2008  * RFC 2822 format, for E-mail and HTTP headers
2009  */
2010 define( 'TS_RFC2822', 3 );
2011
2012 /**
2013  * ISO 8601 format with no timezone: 1986-02-09T20:00:00Z
2014  *
2015  * This is used by Special:Export
2016  */
2017 define( 'TS_ISO_8601', 4 );
2018
2019 /**
2020  * An Exif timestamp (YYYY:MM:DD HH:MM:SS)
2021  *
2022  * @see http://exif.org/Exif2-2.PDF The Exif 2.2 spec, see page 28 for the
2023  *       DateTime tag and page 36 for the DateTimeOriginal and
2024  *       DateTimeDigitized tags.
2025  */
2026 define( 'TS_EXIF', 5 );
2027
2028 /**
2029  * Oracle format time.
2030  */
2031 define( 'TS_ORACLE', 6 );
2032
2033 /**
2034  * Postgres format time.
2035  */
2036 define( 'TS_POSTGRES', 7 );
2037
2038 /**
2039  * ISO 8601 basic format with no timezone: 19860209T200000Z.  This is used by ResourceLoader
2040  */
2041 define( 'TS_ISO_8601_BASIC', 9 );
2042
2043 /**
2044  * Get a timestamp string in one of various formats
2045  *
2046  * @param mixed $outputtype A timestamp in one of the supported formats, the
2047  *   function will autodetect which format is supplied and act accordingly.
2048  * @param mixed $ts Optional timestamp to convert, default 0 for the current time
2049  * @return string|bool String / false The same date in the format specified in $outputtype or false
2050  */
2051 function wfTimestamp( $outputtype = TS_UNIX, $ts = 0 ) {
2052         try {
2053                 $timestamp = new MWTimestamp( $ts );
2054                 return $timestamp->getTimestamp( $outputtype );
2055         } catch ( TimestampException $e ) {
2056                 wfDebug( "wfTimestamp() fed bogus time value: TYPE=$outputtype; VALUE=$ts\n" );
2057                 return false;
2058         }
2059 }
2060
2061 /**
2062  * Return a formatted timestamp, or null if input is null.
2063  * For dealing with nullable timestamp columns in the database.
2064  *
2065  * @param int $outputtype
2066  * @param string $ts
2067  * @return string
2068  */
2069 function wfTimestampOrNull( $outputtype = TS_UNIX, $ts = null ) {
2070         if ( is_null( $ts ) ) {
2071                 return null;
2072         } else {
2073                 return wfTimestamp( $outputtype, $ts );
2074         }
2075 }
2076
2077 /**
2078  * Convenience function; returns MediaWiki timestamp for the present time.
2079  *
2080  * @return string
2081  */
2082 function wfTimestampNow() {
2083         # return NOW
2084         return wfTimestamp( TS_MW, time() );
2085 }
2086
2087 /**
2088  * Check if the operating system is Windows
2089  *
2090  * @return bool True if it's Windows, false otherwise.
2091  */
2092 function wfIsWindows() {
2093         static $isWindows = null;
2094         if ( $isWindows === null ) {
2095                 $isWindows = strtoupper( substr( PHP_OS, 0, 3 ) ) === 'WIN';
2096         }
2097         return $isWindows;
2098 }
2099
2100 /**
2101  * Check if we are running under HHVM
2102  *
2103  * @return bool
2104  */
2105 function wfIsHHVM() {
2106         return defined( 'HHVM_VERSION' );
2107 }
2108
2109 /**
2110  * Tries to get the system directory for temporary files. First
2111  * $wgTmpDirectory is checked, and then the TMPDIR, TMP, and TEMP
2112  * environment variables are then checked in sequence, then
2113  * sys_get_temp_dir(), then upload_tmp_dir from php.ini.
2114  *
2115  * NOTE: When possible, use instead the tmpfile() function to create
2116  * temporary files to avoid race conditions on file creation, etc.
2117  *
2118  * @return string
2119  */
2120 function wfTempDir() {
2121         global $wgTmpDirectory;
2122
2123         if ( $wgTmpDirectory !== false ) {
2124                 return $wgTmpDirectory;
2125         }
2126
2127         $tmpDir = array_map( "getenv", [ 'TMPDIR', 'TMP', 'TEMP' ] );
2128         $tmpDir[] = sys_get_temp_dir();
2129         $tmpDir[] = ini_get( 'upload_tmp_dir' );
2130
2131         foreach ( $tmpDir as $tmp ) {
2132                 if ( $tmp && file_exists( $tmp ) && is_dir( $tmp ) && is_writable( $tmp ) ) {
2133                         return $tmp;
2134                 }
2135         }
2136
2137         /**
2138          * PHP on Windows will detect C:\Windows\Temp as not writable even though PHP can write to it
2139          * so create a directory within that called 'mwtmp' with a suffix of the user running the
2140          * current process.
2141          * The user is included as if various scripts are run by different users they will likely
2142          * not be able to access each others temporary files.
2143          */
2144         if ( wfIsWindows() ) {
2145                 $tmp = sys_get_temp_dir() . DIRECTORY_SEPARATOR . 'mwtmp' . '-' . get_current_user();
2146                 if ( !file_exists( $tmp ) ) {
2147                         mkdir( $tmp );
2148                 }
2149                 if ( file_exists( $tmp ) && is_dir( $tmp ) && is_writable( $tmp ) ) {
2150                         return $tmp;
2151                 }
2152         }
2153
2154         throw new MWException( 'No writable temporary directory could be found. ' .
2155                 'Please set $wgTmpDirectory to a writable directory.' );
2156 }
2157
2158 /**
2159  * Make directory, and make all parent directories if they don't exist
2160  *
2161  * @param string $dir Full path to directory to create
2162  * @param int $mode Chmod value to use, default is $wgDirectoryMode
2163  * @param string $caller Optional caller param for debugging.
2164  * @throws MWException
2165  * @return bool
2166  */
2167 function wfMkdirParents( $dir, $mode = null, $caller = null ) {
2168         global $wgDirectoryMode;
2169
2170         if ( FileBackend::isStoragePath( $dir ) ) { // sanity
2171                 throw new MWException( __FUNCTION__ . " given storage path '$dir'." );
2172         }
2173
2174         if ( !is_null( $caller ) ) {
2175                 wfDebug( "$caller: called wfMkdirParents($dir)\n" );
2176         }
2177
2178         if ( strval( $dir ) === '' || is_dir( $dir ) ) {
2179                 return true;
2180         }
2181
2182         $dir = str_replace( [ '\\', '/' ], DIRECTORY_SEPARATOR, $dir );
2183
2184         if ( is_null( $mode ) ) {
2185                 $mode = $wgDirectoryMode;
2186         }
2187
2188         // Turn off the normal warning, we're doing our own below
2189         MediaWiki\suppressWarnings();
2190         $ok = mkdir( $dir, $mode, true ); // PHP5 <3
2191         MediaWiki\restoreWarnings();
2192
2193         if ( !$ok ) {
2194                 // directory may have been created on another request since we last checked
2195                 if ( is_dir( $dir ) ) {
2196                         return true;
2197                 }
2198
2199                 // PHP doesn't report the path in its warning message, so add our own to aid in diagnosis.
2200                 wfLogWarning( sprintf( "failed to mkdir \"%s\" mode 0%o", $dir, $mode ) );
2201         }
2202         return $ok;
2203 }
2204
2205 /**
2206  * Remove a directory and all its content.
2207  * Does not hide error.
2208  * @param string $dir
2209  */
2210 function wfRecursiveRemoveDir( $dir ) {
2211         wfDebug( __FUNCTION__ . "( $dir )\n" );
2212         // taken from http://de3.php.net/manual/en/function.rmdir.php#98622
2213         if ( is_dir( $dir ) ) {
2214                 $objects = scandir( $dir );
2215                 foreach ( $objects as $object ) {
2216                         if ( $object != "." && $object != ".." ) {
2217                                 if ( filetype( $dir . '/' . $object ) == "dir" ) {
2218                                         wfRecursiveRemoveDir( $dir . '/' . $object );
2219                                 } else {
2220                                         unlink( $dir . '/' . $object );
2221                                 }
2222                         }
2223                 }
2224                 reset( $objects );
2225                 rmdir( $dir );
2226         }
2227 }
2228
2229 /**
2230  * @param int $nr The number to format
2231  * @param int $acc The number of digits after the decimal point, default 2
2232  * @param bool $round Whether or not to round the value, default true
2233  * @return string
2234  */
2235 function wfPercent( $nr, $acc = 2, $round = true ) {
2236         $ret = sprintf( "%.${acc}f", $nr );
2237         return $round ? round( $ret, $acc ) . '%' : "$ret%";
2238 }
2239
2240 /**
2241  * Safety wrapper around ini_get() for boolean settings.
2242  * The values returned from ini_get() are pre-normalized for settings
2243  * set via php.ini or php_flag/php_admin_flag... but *not*
2244  * for those set via php_value/php_admin_value.
2245  *
2246  * It's fairly common for people to use php_value instead of php_flag,
2247  * which can leave you with an 'off' setting giving a false positive
2248  * for code that just takes the ini_get() return value as a boolean.
2249  *
2250  * To make things extra interesting, setting via php_value accepts
2251  * "true" and "yes" as true, but php.ini and php_flag consider them false. :)
2252  * Unrecognized values go false... again opposite PHP's own coercion
2253  * from string to bool.
2254  *
2255  * Luckily, 'properly' set settings will always come back as '0' or '1',
2256  * so we only have to worry about them and the 'improper' settings.
2257  *
2258  * I frickin' hate PHP... :P
2259  *
2260  * @param string $setting
2261  * @return bool
2262  */
2263 function wfIniGetBool( $setting ) {
2264         $val = strtolower( ini_get( $setting ) );
2265         // 'on' and 'true' can't have whitespace around them, but '1' can.
2266         return $val == 'on'
2267                 || $val == 'true'
2268                 || $val == 'yes'
2269                 || preg_match( "/^\s*[+-]?0*[1-9]/", $val ); // approx C atoi() function
2270 }
2271
2272 /**
2273  * Windows-compatible version of escapeshellarg()
2274  * Windows doesn't recognise single-quotes in the shell, but the escapeshellarg()
2275  * function puts single quotes in regardless of OS.
2276  *
2277  * Also fixes the locale problems on Linux in PHP 5.2.6+ (bug backported to
2278  * earlier distro releases of PHP)
2279  *
2280  * @param string ... strings to escape and glue together, or a single array of strings parameter
2281  * @return string
2282  */
2283 function wfEscapeShellArg( /*...*/ ) {
2284         wfInitShellLocale();
2285
2286         $args = func_get_args();
2287         if ( count( $args ) === 1 && is_array( reset( $args ) ) ) {
2288                 // If only one argument has been passed, and that argument is an array,
2289                 // treat it as a list of arguments
2290                 $args = reset( $args );
2291         }
2292
2293         $first = true;
2294         $retVal = '';
2295         foreach ( $args as $arg ) {
2296                 if ( !$first ) {
2297                         $retVal .= ' ';
2298                 } else {
2299                         $first = false;
2300                 }
2301
2302                 if ( wfIsWindows() ) {
2303                         // Escaping for an MSVC-style command line parser and CMD.EXE
2304                         // @codingStandardsIgnoreStart For long URLs
2305                         // Refs:
2306                         //  * http://web.archive.org/web/20020708081031/http://mailman.lyra.org/pipermail/scite-interest/2002-March/000436.html
2307                         //  * http://technet.microsoft.com/en-us/library/cc723564.aspx
2308                         //  * Bug #13518
2309                         //  * CR r63214
2310                         // Double the backslashes before any double quotes. Escape the double quotes.
2311                         // @codingStandardsIgnoreEnd
2312                         $tokens = preg_split( '/(\\\\*")/', $arg, -1, PREG_SPLIT_DELIM_CAPTURE );
2313                         $arg = '';
2314                         $iteration = 0;
2315                         foreach ( $tokens as $token ) {
2316                                 if ( $iteration % 2 == 1 ) {
2317                                         // Delimiter, a double quote preceded by zero or more slashes
2318                                         $arg .= str_replace( '\\', '\\\\', substr( $token, 0, -1 ) ) . '\\"';
2319                                 } elseif ( $iteration % 4 == 2 ) {
2320                                         // ^ in $token will be outside quotes, need to be escaped
2321                                         $arg .= str_replace( '^', '^^', $token );
2322                                 } else { // $iteration % 4 == 0
2323                                         // ^ in $token will appear inside double quotes, so leave as is
2324                                         $arg .= $token;
2325                                 }
2326                                 $iteration++;
2327                         }
2328                         // Double the backslashes before the end of the string, because
2329                         // we will soon add a quote
2330                         $m = [];
2331                         if ( preg_match( '/^(.*?)(\\\\+)$/', $arg, $m ) ) {
2332                                 $arg = $m[1] . str_replace( '\\', '\\\\', $m[2] );
2333                         }
2334
2335                         // Add surrounding quotes
2336                         $retVal .= '"' . $arg . '"';
2337                 } else {
2338                         $retVal .= escapeshellarg( $arg );
2339                 }
2340         }
2341         return $retVal;
2342 }
2343
2344 /**
2345  * Check if wfShellExec() is effectively disabled via php.ini config
2346  *
2347  * @return bool|string False or 'disabled'
2348  * @since 1.22
2349  */
2350 function wfShellExecDisabled() {
2351         static $disabled = null;
2352         if ( is_null( $disabled ) ) {
2353                 if ( !function_exists( 'proc_open' ) ) {
2354                         wfDebug( "proc_open() is disabled\n" );
2355                         $disabled = 'disabled';
2356                 } else {
2357                         $disabled = false;
2358                 }
2359         }
2360         return $disabled;
2361 }
2362
2363 /**
2364  * Execute a shell command, with time and memory limits mirrored from the PHP
2365  * configuration if supported.
2366  *
2367  * @param string|string[] $cmd If string, a properly shell-escaped command line,
2368  *   or an array of unescaped arguments, in which case each value will be escaped
2369  *   Example:   [ 'convert', '-font', 'font name' ] would produce "'convert' '-font' 'font name'"
2370  * @param null|mixed &$retval Optional, will receive the program's exit code.
2371  *   (non-zero is usually failure). If there is an error from
2372  *   read, select, or proc_open(), this will be set to -1.
2373  * @param array $environ Optional environment variables which should be
2374  *   added to the executed command environment.
2375  * @param array $limits Optional array with limits(filesize, memory, time, walltime)
2376  *   this overwrites the global wgMaxShell* limits.
2377  * @param array $options Array of options:
2378  *   - duplicateStderr: Set this to true to duplicate stderr to stdout,
2379  *     including errors from limit.sh
2380  *   - profileMethod: By default this function will profile based on the calling
2381  *     method. Set this to a string for an alternative method to profile from
2382  *
2383  * @return string Collected stdout as a string
2384  */
2385 function wfShellExec( $cmd, &$retval = null, $environ = [],
2386         $limits = [], $options = []
2387 ) {
2388         global $IP, $wgMaxShellMemory, $wgMaxShellFileSize, $wgMaxShellTime,
2389                 $wgMaxShellWallClockTime, $wgShellCgroup;
2390
2391         $disabled = wfShellExecDisabled();
2392         if ( $disabled ) {
2393                 $retval = 1;
2394                 return 'Unable to run external programs, proc_open() is disabled.';
2395         }
2396
2397         $includeStderr = isset( $options['duplicateStderr'] ) && $options['duplicateStderr'];
2398         $profileMethod = isset( $options['profileMethod'] ) ? $options['profileMethod'] : wfGetCaller();
2399
2400         wfInitShellLocale();
2401
2402         $envcmd = '';
2403         foreach ( $environ as $k => $v ) {
2404                 if ( wfIsWindows() ) {
2405                         /* Surrounding a set in quotes (method used by wfEscapeShellArg) makes the quotes themselves
2406                          * appear in the environment variable, so we must use carat escaping as documented in
2407                          * http://technet.microsoft.com/en-us/library/cc723564.aspx
2408                          * Note however that the quote isn't listed there, but is needed, and the parentheses
2409                          * are listed there but doesn't appear to need it.
2410                          */
2411                         $envcmd .= "set $k=" . preg_replace( '/([&|()<>^"])/', '^\\1', $v ) . '&& ';
2412                 } else {
2413                         /* Assume this is a POSIX shell, thus required to accept variable assignments before the command
2414                          * http://www.opengroup.org/onlinepubs/009695399/utilities/xcu_chap02.html#tag_02_09_01
2415                          */
2416                         $envcmd .= "$k=" . escapeshellarg( $v ) . ' ';
2417                 }
2418         }
2419         if ( is_array( $cmd ) ) {
2420                 $cmd = wfEscapeShellArg( $cmd );
2421         }
2422
2423         $cmd = $envcmd . $cmd;
2424
2425         $useLogPipe = false;
2426         if ( is_executable( '/bin/bash' ) ) {
2427                 $time = intval( isset( $limits['time'] ) ? $limits['time'] : $wgMaxShellTime );
2428                 if ( isset( $limits['walltime'] ) ) {
2429                         $wallTime = intval( $limits['walltime'] );
2430                 } elseif ( isset( $limits['time'] ) ) {
2431                         $wallTime = $time;
2432                 } else {
2433                         $wallTime = intval( $wgMaxShellWallClockTime );
2434                 }
2435                 $mem = intval( isset( $limits['memory'] ) ? $limits['memory'] : $wgMaxShellMemory );
2436                 $filesize = intval( isset( $limits['filesize'] ) ? $limits['filesize'] : $wgMaxShellFileSize );
2437
2438                 if ( $time > 0 || $mem > 0 || $filesize > 0 || $wallTime > 0 ) {
2439                         $cmd = '/bin/bash ' . escapeshellarg( "$IP/includes/limit.sh" ) . ' ' .
2440                                 escapeshellarg( $cmd ) . ' ' .
2441                                 escapeshellarg(
2442                                         "MW_INCLUDE_STDERR=" . ( $includeStderr ? '1' : '' ) . ';' .
2443                                         "MW_CPU_LIMIT=$time; " .
2444                                         'MW_CGROUP=' . escapeshellarg( $wgShellCgroup ) . '; ' .
2445                                         "MW_MEM_LIMIT=$mem; " .
2446                                         "MW_FILE_SIZE_LIMIT=$filesize; " .
2447                                         "MW_WALL_CLOCK_LIMIT=$wallTime; " .
2448                                         "MW_USE_LOG_PIPE=yes"
2449                                 );
2450                         $useLogPipe = true;
2451                 } elseif ( $includeStderr ) {
2452                         $cmd .= ' 2>&1';
2453                 }
2454         } elseif ( $includeStderr ) {
2455                 $cmd .= ' 2>&1';
2456         }
2457         wfDebug( "wfShellExec: $cmd\n" );
2458
2459         // Don't try to execute commands that exceed Linux's MAX_ARG_STRLEN.
2460         // Other platforms may be more accomodating, but we don't want to be
2461         // accomodating, because very long commands probably include user
2462         // input. See T129506.
2463         if ( strlen( $cmd ) > SHELL_MAX_ARG_STRLEN ) {
2464                 throw new Exception( __METHOD__ .
2465                         '(): total length of $cmd must not exceed SHELL_MAX_ARG_STRLEN' );
2466         }
2467
2468         $desc = [
2469                 0 => [ 'file', 'php://stdin', 'r' ],
2470                 1 => [ 'pipe', 'w' ],
2471                 2 => [ 'file', 'php://stderr', 'w' ] ];
2472         if ( $useLogPipe ) {
2473                 $desc[3] = [ 'pipe', 'w' ];
2474         }
2475         $pipes = null;
2476         $scoped = Profiler::instance()->scopedProfileIn( __FUNCTION__ . '-' . $profileMethod );
2477         $proc = proc_open( $cmd, $desc, $pipes );
2478         if ( !$proc ) {
2479                 wfDebugLog( 'exec', "proc_open() failed: $cmd" );
2480                 $retval = -1;
2481                 return '';
2482         }
2483         $outBuffer = $logBuffer = '';
2484         $emptyArray = [];
2485         $status = false;
2486         $logMsg = false;
2487
2488         /* According to the documentation, it is possible for stream_select()
2489          * to fail due to EINTR. I haven't managed to induce this in testing
2490          * despite sending various signals. If it did happen, the error
2491          * message would take the form:
2492          *
2493          * stream_select(): unable to select [4]: Interrupted system call (max_fd=5)
2494          *
2495          * where [4] is the value of the macro EINTR and "Interrupted system
2496          * call" is string which according to the Linux manual is "possibly"
2497          * localised according to LC_MESSAGES.
2498          */
2499         $eintr = defined( 'SOCKET_EINTR' ) ? SOCKET_EINTR : 4;
2500         $eintrMessage = "stream_select(): unable to select [$eintr]";
2501
2502         $running = true;
2503         $timeout = null;
2504         $numReadyPipes = 0;
2505
2506         while ( $running === true || $numReadyPipes !== 0 ) {
2507                 if ( $running ) {
2508                         $status = proc_get_status( $proc );
2509                         // If the process has terminated, switch to nonblocking selects
2510                         // for getting any data still waiting to be read.
2511                         if ( !$status['running'] ) {
2512                                 $running = false;
2513                                 $timeout = 0;
2514                         }
2515                 }
2516
2517                 $readyPipes = $pipes;
2518
2519                 // Clear last error
2520                 // @codingStandardsIgnoreStart Generic.PHP.NoSilencedErrors.Discouraged
2521                 @trigger_error( '' );
2522                 $numReadyPipes = @stream_select( $readyPipes, $emptyArray, $emptyArray, $timeout );
2523                 if ( $numReadyPipes === false ) {
2524                         // @codingStandardsIgnoreEnd
2525                         $error = error_get_last();
2526                         if ( strncmp( $error['message'], $eintrMessage, strlen( $eintrMessage ) ) == 0 ) {
2527                                 continue;
2528                         } else {
2529                                 trigger_error( $error['message'], E_USER_WARNING );
2530                                 $logMsg = $error['message'];
2531                                 break;
2532                         }
2533                 }
2534                 foreach ( $readyPipes as $fd => $pipe ) {
2535                         $block = fread( $pipe, 65536 );
2536                         if ( $block === '' ) {
2537                                 // End of file
2538                                 fclose( $pipes[$fd] );
2539                                 unset( $pipes[$fd] );
2540                                 if ( !$pipes ) {
2541                                         break 2;
2542                                 }
2543                         } elseif ( $block === false ) {
2544                                 // Read error
2545                                 $logMsg = "Error reading from pipe";
2546                                 break 2;
2547                         } elseif ( $fd == 1 ) {
2548                                 // From stdout
2549                                 $outBuffer .= $block;
2550                         } elseif ( $fd == 3 ) {
2551                                 // From log FD
2552                                 $logBuffer .= $block;
2553                                 if ( strpos( $block, "\n" ) !== false ) {
2554                                         $lines = explode( "\n", $logBuffer );
2555                                         $logBuffer = array_pop( $lines );
2556                                         foreach ( $lines as $line ) {
2557                                                 wfDebugLog( 'exec', $line );
2558                                         }
2559                                 }
2560                         }
2561                 }
2562         }
2563
2564         foreach ( $pipes as $pipe ) {
2565                 fclose( $pipe );
2566         }
2567
2568         // Use the status previously collected if possible, since proc_get_status()
2569         // just calls waitpid() which will not return anything useful the second time.
2570         if ( $running ) {
2571                 $status = proc_get_status( $proc );
2572         }
2573
2574         if ( $logMsg !== false ) {
2575                 // Read/select error
2576                 $retval = -1;
2577                 proc_close( $proc );
2578         } elseif ( $status['signaled'] ) {
2579                 $logMsg = "Exited with signal {$status['termsig']}";
2580                 $retval = 128 + $status['termsig'];
2581                 proc_close( $proc );
2582         } else {
2583                 if ( $status['running'] ) {
2584                         $retval = proc_close( $proc );
2585                 } else {
2586                         $retval = $status['exitcode'];
2587                         proc_close( $proc );
2588                 }
2589                 if ( $retval == 127 ) {
2590                         $logMsg = "Possibly missing executable file";
2591                 } elseif ( $retval >= 129 && $retval <= 192 ) {
2592                         $logMsg = "Probably exited with signal " . ( $retval - 128 );
2593                 }
2594         }
2595
2596         if ( $logMsg !== false ) {
2597                 wfDebugLog( 'exec', "$logMsg: $cmd" );
2598         }
2599
2600         return $outBuffer;
2601 }
2602
2603 /**
2604  * Execute a shell command, returning both stdout and stderr. Convenience
2605  * function, as all the arguments to wfShellExec can become unwieldy.
2606  *
2607  * @note This also includes errors from limit.sh, e.g. if $wgMaxShellFileSize is exceeded.
2608  * @param string|string[] $cmd If string, a properly shell-escaped command line,
2609  *   or an array of unescaped arguments, in which case each value will be escaped
2610  *   Example:   [ 'convert', '-font', 'font name' ] would produce "'convert' '-font' 'font name'"
2611  * @param null|mixed &$retval Optional, will receive the program's exit code.
2612  *   (non-zero is usually failure)
2613  * @param array $environ Optional environment variables which should be
2614  *   added to the executed command environment.
2615  * @param array $limits Optional array with limits(filesize, memory, time, walltime)
2616  *   this overwrites the global wgMaxShell* limits.
2617  * @return string Collected stdout and stderr as a string
2618  */
2619 function wfShellExecWithStderr( $cmd, &$retval = null, $environ = [], $limits = [] ) {
2620         return wfShellExec( $cmd, $retval, $environ, $limits,
2621                 [ 'duplicateStderr' => true, 'profileMethod' => wfGetCaller() ] );
2622 }
2623
2624 /**
2625  * Workaround for http://bugs.php.net/bug.php?id=45132
2626  * escapeshellarg() destroys non-ASCII characters if LANG is not a UTF-8 locale
2627  */
2628 function wfInitShellLocale() {
2629         static $done = false;
2630         if ( $done ) {
2631                 return;
2632         }
2633         $done = true;
2634         global $wgShellLocale;
2635         putenv( "LC_CTYPE=$wgShellLocale" );
2636         setlocale( LC_CTYPE, $wgShellLocale );
2637 }
2638
2639 /**
2640  * Generate a shell-escaped command line string to run a MediaWiki cli script.
2641  * Note that $parameters should be a flat array and an option with an argument
2642  * should consist of two consecutive items in the array (do not use "--option value").
2643  *
2644  * @param string $script MediaWiki cli script path
2645  * @param array $parameters Arguments and options to the script
2646  * @param array $options Associative array of options:
2647  *              'php': The path to the php executable
2648  *              'wrapper': Path to a PHP wrapper to handle the maintenance script
2649  * @return string
2650  */
2651 function wfShellWikiCmd( $script, array $parameters = [], array $options = [] ) {
2652         global $wgPhpCli;
2653         // Give site config file a chance to run the script in a wrapper.
2654         // The caller may likely want to call wfBasename() on $script.
2655         Hooks::run( 'wfShellWikiCmd', [ &$script, &$parameters, &$options ] );
2656         $cmd = isset( $options['php'] ) ? [ $options['php'] ] : [ $wgPhpCli ];
2657         if ( isset( $options['wrapper'] ) ) {
2658                 $cmd[] = $options['wrapper'];
2659         }
2660         $cmd[] = $script;
2661         // Escape each parameter for shell
2662         return wfEscapeShellArg( array_merge( $cmd, $parameters ) );
2663 }
2664
2665 /**
2666  * wfMerge attempts to merge differences between three texts.
2667  * Returns true for a clean merge and false for failure or a conflict.
2668  *
2669  * @param string $old
2670  * @param string $mine
2671  * @param string $yours
2672  * @param string $result
2673  * @return bool
2674  */
2675 function wfMerge( $old, $mine, $yours, &$result ) {
2676         global $wgDiff3;
2677
2678         # This check may also protect against code injection in
2679         # case of broken installations.
2680         MediaWiki\suppressWarnings();
2681         $haveDiff3 = $wgDiff3 && file_exists( $wgDiff3 );
2682         MediaWiki\restoreWarnings();
2683
2684         if ( !$haveDiff3 ) {
2685                 wfDebug( "diff3 not found\n" );
2686                 return false;
2687         }
2688
2689         # Make temporary files
2690         $td = wfTempDir();
2691         $oldtextFile = fopen( $oldtextName = tempnam( $td, 'merge-old-' ), 'w' );
2692         $mytextFile = fopen( $mytextName = tempnam( $td, 'merge-mine-' ), 'w' );
2693         $yourtextFile = fopen( $yourtextName = tempnam( $td, 'merge-your-' ), 'w' );
2694
2695         # NOTE: diff3 issues a warning to stderr if any of the files does not end with
2696         #       a newline character. To avoid this, we normalize the trailing whitespace before
2697         #       creating the diff.
2698
2699         fwrite( $oldtextFile, rtrim( $old ) . "\n" );
2700         fclose( $oldtextFile );
2701         fwrite( $mytextFile, rtrim( $mine ) . "\n" );
2702         fclose( $mytextFile );
2703         fwrite( $yourtextFile, rtrim( $yours ) . "\n" );
2704         fclose( $yourtextFile );
2705
2706         # Check for a conflict
2707         $cmd = wfEscapeShellArg( $wgDiff3, '-a', '--overlap-only', $mytextName,
2708                 $oldtextName, $yourtextName );
2709         $handle = popen( $cmd, 'r' );
2710
2711         if ( fgets( $handle, 1024 ) ) {
2712                 $conflict = true;
2713         } else {
2714                 $conflict = false;
2715         }
2716         pclose( $handle );
2717
2718         # Merge differences
2719         $cmd = wfEscapeShellArg( $wgDiff3, '-a', '-e', '--merge', $mytextName,
2720                 $oldtextName, $yourtextName );
2721         $handle = popen( $cmd, 'r' );
2722         $result = '';
2723         do {
2724                 $data = fread( $handle, 8192 );
2725                 if ( strlen( $data ) == 0 ) {
2726                         break;
2727                 }
2728                 $result .= $data;
2729         } while ( true );
2730         pclose( $handle );
2731         unlink( $mytextName );
2732         unlink( $oldtextName );
2733         unlink( $yourtextName );
2734
2735         if ( $result === '' && $old !== '' && !$conflict ) {
2736                 wfDebug( "Unexpected null result from diff3. Command: $cmd\n" );
2737                 $conflict = true;
2738         }
2739         return !$conflict;
2740 }
2741
2742 /**
2743  * Returns unified plain-text diff of two texts.
2744  * "Useful" for machine processing of diffs.
2745  *
2746  * @deprecated since 1.25, use DiffEngine/UnifiedDiffFormatter directly
2747  *
2748  * @param string $before The text before the changes.
2749  * @param string $after The text after the changes.
2750  * @param string $params Command-line options for the diff command.
2751  * @return string Unified diff of $before and $after
2752  */
2753 function wfDiff( $before, $after, $params = '-u' ) {
2754         if ( $before == $after ) {
2755                 return '';
2756         }
2757
2758         global $wgDiff;
2759         MediaWiki\suppressWarnings();
2760         $haveDiff = $wgDiff && file_exists( $wgDiff );
2761         MediaWiki\restoreWarnings();
2762
2763         # This check may also protect against code injection in
2764         # case of broken installations.
2765         if ( !$haveDiff ) {
2766                 wfDebug( "diff executable not found\n" );
2767                 $diffs = new Diff( explode( "\n", $before ), explode( "\n", $after ) );
2768                 $format = new UnifiedDiffFormatter();
2769                 return $format->format( $diffs );
2770         }
2771
2772         # Make temporary files
2773         $td = wfTempDir();
2774         $oldtextFile = fopen( $oldtextName = tempnam( $td, 'merge-old-' ), 'w' );
2775         $newtextFile = fopen( $newtextName = tempnam( $td, 'merge-your-' ), 'w' );
2776
2777         fwrite( $oldtextFile, $before );
2778         fclose( $oldtextFile );
2779         fwrite( $newtextFile, $after );
2780         fclose( $newtextFile );
2781
2782         // Get the diff of the two files
2783         $cmd = "$wgDiff " . $params . ' ' . wfEscapeShellArg( $oldtextName, $newtextName );
2784
2785         $h = popen( $cmd, 'r' );
2786         if ( !$h ) {
2787                 unlink( $oldtextName );
2788                 unlink( $newtextName );
2789                 throw new Exception( __METHOD__ . '(): popen() failed' );
2790         }
2791
2792         $diff = '';
2793
2794         do {
2795                 $data = fread( $h, 8192 );
2796                 if ( strlen( $data ) == 0 ) {
2797                         break;
2798                 }
2799                 $diff .= $data;
2800         } while ( true );
2801
2802         // Clean up
2803         pclose( $h );
2804         unlink( $oldtextName );
2805         unlink( $newtextName );
2806
2807         // Kill the --- and +++ lines. They're not useful.
2808         $diff_lines = explode( "\n", $diff );
2809         if ( isset( $diff_lines[0] ) && strpos( $diff_lines[0], '---' ) === 0 ) {
2810                 unset( $diff_lines[0] );
2811         }
2812         if ( isset( $diff_lines[1] ) && strpos( $diff_lines[1], '+++' ) === 0 ) {
2813                 unset( $diff_lines[1] );
2814         }
2815
2816         $diff = implode( "\n", $diff_lines );
2817
2818         return $diff;
2819 }
2820
2821 /**
2822  * This function works like "use VERSION" in Perl, the program will die with a
2823  * backtrace if the current version of PHP is less than the version provided
2824  *
2825  * This is useful for extensions which due to their nature are not kept in sync
2826  * with releases, and might depend on other versions of PHP than the main code
2827  *
2828  * Note: PHP might die due to parsing errors in some cases before it ever
2829  *       manages to call this function, such is life
2830  *
2831  * @see perldoc -f use
2832  *
2833  * @param string|int|float $req_ver The version to check, can be a string, an integer, or a float
2834  * @throws MWException
2835  */
2836 function wfUsePHP( $req_ver ) {
2837         $php_ver = PHP_VERSION;
2838
2839         if ( version_compare( $php_ver, (string)$req_ver, '<' ) ) {
2840                 throw new MWException( "PHP $req_ver required--this is only $php_ver" );
2841         }
2842 }
2843
2844 /**
2845  * This function works like "use VERSION" in Perl except it checks the version
2846  * of MediaWiki, the program will die with a backtrace if the current version
2847  * of MediaWiki is less than the version provided.
2848  *
2849  * This is useful for extensions which due to their nature are not kept in sync
2850  * with releases
2851  *
2852  * Note: Due to the behavior of PHP's version_compare() which is used in this
2853  * function, if you want to allow the 'wmf' development versions add a 'c' (or
2854  * any single letter other than 'a', 'b' or 'p') as a post-fix to your
2855  * targeted version number. For example if you wanted to allow any variation
2856  * of 1.22 use `wfUseMW( '1.22c' )`. Using an 'a' or 'b' instead of 'c' will
2857  * not result in the same comparison due to the internal logic of
2858  * version_compare().
2859  *
2860  * @see perldoc -f use
2861  *
2862  * @deprecated since 1.26, use the "requires' property of extension.json
2863  * @param string|int|float $req_ver The version to check, can be a string, an integer, or a float
2864  * @throws MWException
2865  */
2866 function wfUseMW( $req_ver ) {
2867         global $wgVersion;
2868
2869         if ( version_compare( $wgVersion, (string)$req_ver, '<' ) ) {
2870                 throw new MWException( "MediaWiki $req_ver required--this is only $wgVersion" );
2871         }
2872 }
2873
2874 /**
2875  * Return the final portion of a pathname.
2876  * Reimplemented because PHP5's "basename()" is buggy with multibyte text.
2877  * http://bugs.php.net/bug.php?id=33898
2878  *
2879  * PHP's basename() only considers '\' a pathchar on Windows and Netware.
2880  * We'll consider it so always, as we don't want '\s' in our Unix paths either.
2881  *
2882  * @param string $path
2883  * @param string $suffix String to remove if present
2884  * @return string
2885  */
2886 function wfBaseName( $path, $suffix = '' ) {
2887         if ( $suffix == '' ) {
2888                 $encSuffix = '';
2889         } else {
2890                 $encSuffix = '(?:' . preg_quote( $suffix, '#' ) . ')?';
2891         }
2892
2893         $matches = [];
2894         if ( preg_match( "#([^/\\\\]*?){$encSuffix}[/\\\\]*$#", $path, $matches ) ) {
2895                 return $matches[1];
2896         } else {
2897                 return '';
2898         }
2899 }
2900
2901 /**
2902  * Generate a relative path name to the given file.
2903  * May explode on non-matching case-insensitive paths,
2904  * funky symlinks, etc.
2905  *
2906  * @param string $path Absolute destination path including target filename
2907  * @param string $from Absolute source path, directory only
2908  * @return string
2909  */
2910 function wfRelativePath( $path, $from ) {
2911         // Normalize mixed input on Windows...
2912         $path = str_replace( '/', DIRECTORY_SEPARATOR, $path );
2913         $from = str_replace( '/', DIRECTORY_SEPARATOR, $from );
2914
2915         // Trim trailing slashes -- fix for drive root
2916         $path = rtrim( $path, DIRECTORY_SEPARATOR );
2917         $from = rtrim( $from, DIRECTORY_SEPARATOR );
2918
2919         $pieces = explode( DIRECTORY_SEPARATOR, dirname( $path ) );
2920         $against = explode( DIRECTORY_SEPARATOR, $from );
2921
2922         if ( $pieces[0] !== $against[0] ) {
2923                 // Non-matching Windows drive letters?
2924                 // Return a full path.
2925                 return $path;
2926         }
2927
2928         // Trim off common prefix
2929         while ( count( $pieces ) && count( $against )
2930                 && $pieces[0] == $against[0] ) {
2931                 array_shift( $pieces );
2932                 array_shift( $against );
2933         }
2934
2935         // relative dots to bump us to the parent
2936         while ( count( $against ) ) {
2937                 array_unshift( $pieces, '..' );
2938                 array_shift( $against );
2939         }
2940
2941         array_push( $pieces, wfBaseName( $path ) );
2942
2943         return implode( DIRECTORY_SEPARATOR, $pieces );
2944 }
2945
2946 /**
2947  * Convert an arbitrarily-long digit string from one numeric base
2948  * to another, optionally zero-padding to a minimum column width.
2949  *
2950  * Supports base 2 through 36; digit values 10-36 are represented
2951  * as lowercase letters a-z. Input is case-insensitive.
2952  *
2953  * @deprecated since 1.27 Use Wikimedia\base_convert() directly
2954  *
2955  * @param string $input Input number
2956  * @param int $sourceBase Base of the input number
2957  * @param int $destBase Desired base of the output
2958  * @param int $pad Minimum number of digits in the output (pad with zeroes)
2959  * @param bool $lowercase Whether to output in lowercase or uppercase
2960  * @param string $engine Either "gmp", "bcmath", or "php"
2961  * @return string|bool The output number as a string, or false on error
2962  */
2963 function wfBaseConvert( $input, $sourceBase, $destBase, $pad = 1,
2964         $lowercase = true, $engine = 'auto'
2965 ) {
2966         return Wikimedia\base_convert( $input, $sourceBase, $destBase, $pad, $lowercase, $engine );
2967 }
2968
2969 /**
2970  * @deprecated since 1.27, PHP's session generation isn't used with
2971  *  MediaWiki\Session\SessionManager
2972  */
2973 function wfFixSessionID() {
2974         wfDeprecated( __FUNCTION__, '1.27' );
2975 }
2976
2977 /**
2978  * Reset the session id
2979  *
2980  * @deprecated since 1.27, use MediaWiki\Session\SessionManager instead
2981  * @since 1.22
2982  */
2983 function wfResetSessionID() {
2984         wfDeprecated( __FUNCTION__, '1.27' );
2985         $session = SessionManager::getGlobalSession();
2986         $delay = $session->delaySave();
2987
2988         $session->resetId();
2989
2990         // Make sure a session is started, since that's what the old
2991         // wfResetSessionID() did.
2992         if ( session_id() !== $session->getId() ) {
2993                 wfSetupSession( $session->getId() );
2994         }
2995
2996         ScopedCallback::consume( $delay );
2997 }
2998
2999 /**
3000  * Initialise php session
3001  *
3002  * @deprecated since 1.27, use MediaWiki\Session\SessionManager instead.
3003  *  Generally, "using" SessionManager will be calling ->getSessionById() or
3004  *  ::getGlobalSession() (depending on whether you were passing $sessionId
3005  *  here), then calling $session->persist().
3006  * @param bool|string $sessionId
3007  */
3008 function wfSetupSession( $sessionId = false ) {
3009         wfDeprecated( __FUNCTION__, '1.27' );
3010
3011         if ( $sessionId ) {
3012                 session_id( $sessionId );
3013         }
3014
3015         $session = SessionManager::getGlobalSession();
3016         $session->persist();
3017
3018         if ( session_id() !== $session->getId() ) {
3019                 session_id( $session->getId() );
3020         }
3021         MediaWiki\quietCall( 'session_start' );
3022 }
3023
3024 /**
3025  * Get an object from the precompiled serialized directory
3026  *
3027  * @param string $name
3028  * @return mixed The variable on success, false on failure
3029  */
3030 function wfGetPrecompiledData( $name ) {
3031         global $IP;
3032
3033         $file = "$IP/serialized/$name";
3034         if ( file_exists( $file ) ) {
3035                 $blob = file_get_contents( $file );
3036                 if ( $blob ) {
3037                         return unserialize( $blob );
3038                 }
3039         }
3040         return false;
3041 }
3042
3043 /**
3044  * Make a cache key for the local wiki.
3045  *
3046  * @param string $args,...
3047  * @return string
3048  */
3049 function wfMemcKey( /*...*/ ) {
3050         return call_user_func_array(
3051                 [ ObjectCache::getLocalClusterInstance(), 'makeKey' ],
3052                 func_get_args()
3053         );
3054 }
3055
3056 /**
3057  * Make a cache key for a foreign DB.
3058  *
3059  * Must match what wfMemcKey() would produce in context of the foreign wiki.
3060  *
3061  * @param string $db
3062  * @param string $prefix
3063  * @param string $args,...
3064  * @return string
3065  */
3066 function wfForeignMemcKey( $db, $prefix /*...*/ ) {
3067         $args = array_slice( func_get_args(), 2 );
3068         $keyspace = $prefix ? "$db-$prefix" : $db;
3069         return call_user_func_array(
3070                 [ ObjectCache::getLocalClusterInstance(), 'makeKeyInternal' ],
3071                 [ $keyspace, $args ]
3072         );
3073 }
3074
3075 /**
3076  * Make a cache key with database-agnostic prefix.
3077  *
3078  * Doesn't have a wiki-specific namespace. Uses a generic 'global' prefix
3079  * instead. Must have a prefix as otherwise keys that use a database name
3080  * in the first segment will clash with wfMemcKey/wfForeignMemcKey.
3081  *
3082  * @since 1.26
3083  * @param string $args,...
3084  * @return string
3085  */
3086 function wfGlobalCacheKey( /*...*/ ) {
3087         return call_user_func_array(
3088                 [ ObjectCache::getLocalClusterInstance(), 'makeGlobalKey' ],
3089                 func_get_args()
3090         );
3091 }
3092
3093 /**
3094  * Get an ASCII string identifying this wiki
3095  * This is used as a prefix in memcached keys
3096  *
3097  * @return string
3098  */
3099 function wfWikiID() {
3100         global $wgDBprefix, $wgDBname;
3101         if ( $wgDBprefix ) {
3102                 return "$wgDBname-$wgDBprefix";
3103         } else {
3104                 return $wgDBname;
3105         }
3106 }
3107
3108 /**
3109  * Split a wiki ID into DB name and table prefix
3110  *
3111  * @param string $wiki
3112  *
3113  * @return array
3114  */
3115 function wfSplitWikiID( $wiki ) {
3116         $bits = explode( '-', $wiki, 2 );
3117         if ( count( $bits ) < 2 ) {
3118                 $bits[] = '';
3119         }
3120         return $bits;
3121 }
3122
3123 /**
3124  * Get a Database object.
3125  *
3126  * @param int $db Index of the connection to get. May be DB_MASTER for the
3127  *            master (for write queries), DB_SLAVE for potentially lagged read
3128  *            queries, or an integer >= 0 for a particular server.
3129  *
3130  * @param string|string[] $groups Query groups. An array of group names that this query
3131  *                belongs to. May contain a single string if the query is only
3132  *                in one group.
3133  *
3134  * @param string|bool $wiki The wiki ID, or false for the current wiki
3135  *
3136  * Note: multiple calls to wfGetDB(DB_SLAVE) during the course of one request
3137  * will always return the same object, unless the underlying connection or load
3138  * balancer is manually destroyed.
3139  *
3140  * Note 2: use $this->getDB() in maintenance scripts that may be invoked by
3141  * updater to ensure that a proper database is being updated.
3142  *
3143  * @todo Replace calls to wfGetDB with calls to LoadBalancer::getConnection()
3144  *       on an injected instance of LoadBalancer.
3145  *
3146  * @return DatabaseBase
3147  */
3148 function wfGetDB( $db, $groups = [], $wiki = false ) {
3149         return wfGetLB( $wiki )->getConnection( $db, $groups, $wiki );
3150 }
3151
3152 /**
3153  * Get a load balancer object.
3154  *
3155  * @deprecated since 1.27, use MediaWikiServices::getDBLoadBalancer()
3156  *              or MediaWikiServices::getDBLoadBalancerFactory() instead.
3157  *
3158  * @param string|bool $wiki Wiki ID, or false for the current wiki
3159  * @return LoadBalancer
3160  */
3161 function wfGetLB( $wiki = false ) {
3162         if ( $wiki === false ) {
3163                 return \MediaWiki\MediaWikiServices::getInstance()->getDBLoadBalancer();
3164         } else {
3165                 $factory = \MediaWiki\MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
3166                 return $factory->getMainLB( $wiki );
3167         }
3168 }
3169
3170 /**
3171  * Get the load balancer factory object
3172  *
3173  * @deprecated since 1.27, use MediaWikiServices::getDBLoadBalancerFactory() instead.
3174  *
3175  * @return LBFactory
3176  */
3177 function wfGetLBFactory() {
3178         return \MediaWiki\MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
3179 }
3180
3181 /**
3182  * Find a file.
3183  * Shortcut for RepoGroup::singleton()->findFile()
3184  *
3185  * @param string $title String or Title object
3186  * @param array $options Associative array of options (see RepoGroup::findFile)
3187  * @return File|bool File, or false if the file does not exist
3188  */
3189 function wfFindFile( $title, $options = [] ) {
3190         return RepoGroup::singleton()->findFile( $title, $options );
3191 }
3192
3193 /**
3194  * Get an object referring to a locally registered file.
3195  * Returns a valid placeholder object if the file does not exist.
3196  *
3197  * @param Title|string $title
3198  * @return LocalFile|null A File, or null if passed an invalid Title
3199  */
3200 function wfLocalFile( $title ) {
3201         return RepoGroup::singleton()->getLocalRepo()->newFile( $title );
3202 }
3203
3204 /**
3205  * Should low-performance queries be disabled?
3206  *
3207  * @return bool
3208  * @codeCoverageIgnore
3209  */
3210 function wfQueriesMustScale() {
3211         global $wgMiserMode;
3212         return $wgMiserMode
3213                 || ( SiteStats::pages() > 100000
3214                 && SiteStats::edits() > 1000000
3215                 && SiteStats::users() > 10000 );
3216 }
3217
3218 /**
3219  * Get the path to a specified script file, respecting file
3220  * extensions; this is a wrapper around $wgScriptPath etc.
3221  * except for 'index' and 'load' which use $wgScript/$wgLoadScript
3222  *
3223  * @param string $script Script filename, sans extension
3224  * @return string
3225  */
3226 function wfScript( $script = 'index' ) {
3227         global $wgScriptPath, $wgScript, $wgLoadScript;
3228         if ( $script === 'index' ) {
3229                 return $wgScript;
3230         } elseif ( $script === 'load' ) {
3231                 return $wgLoadScript;
3232         } else {
3233                 return "{$wgScriptPath}/{$script}.php";
3234         }
3235 }
3236
3237 /**
3238  * Get the script URL.
3239  *
3240  * @return string Script URL
3241  */
3242 function wfGetScriptUrl() {
3243         if ( isset( $_SERVER['SCRIPT_NAME'] ) ) {
3244                 /* as it was called, minus the query string.
3245                  *
3246                  * Some sites use Apache rewrite rules to handle subdomains,
3247                  * and have PHP set up in a weird way that causes PHP_SELF
3248                  * to contain the rewritten URL instead of the one that the
3249                  * outside world sees.
3250                  *
3251                  * If in this mode, use SCRIPT_URL instead, which mod_rewrite
3252                  * provides containing the "before" URL.
3253                  */
3254                 return $_SERVER['SCRIPT_NAME'];
3255         } else {
3256                 return $_SERVER['URL'];
3257         }
3258 }
3259
3260 /**
3261  * Convenience function converts boolean values into "true"
3262  * or "false" (string) values
3263  *
3264  * @param bool $value
3265  * @return string
3266  */
3267 function wfBoolToStr( $value ) {
3268         return $value ? 'true' : 'false';
3269 }
3270
3271 /**
3272  * Get a platform-independent path to the null file, e.g. /dev/null
3273  *
3274  * @return string
3275  */
3276 function wfGetNull() {
3277         return wfIsWindows() ? 'NUL' : '/dev/null';
3278 }
3279
3280 /**
3281  * Waits for the slaves to catch up to the master position
3282  *
3283  * Use this when updating very large numbers of rows, as in maintenance scripts,
3284  * to avoid causing too much lag. Of course, this is a no-op if there are no slaves.
3285  *
3286  * By default this waits on the main DB cluster of the current wiki.
3287  * If $cluster is set to "*" it will wait on all DB clusters, including
3288  * external ones. If the lag being waiting on is caused by the code that
3289  * does this check, it makes since to use $ifWritesSince, particularly if
3290  * cluster is "*", to avoid excess overhead.
3291  *
3292  * Never call this function after a big DB write that is still in a transaction.
3293  * This only makes sense after the possible lag inducing changes were committed.
3294  *
3295  * @param float|null $ifWritesSince Only wait if writes were done since this UNIX timestamp
3296  * @param string|bool $wiki Wiki identifier accepted by wfGetLB
3297  * @param string|bool $cluster Cluster name accepted by LBFactory. Default: false.
3298  * @param int|null $timeout Max wait time. Default: 1 day (cli), ~10 seconds (web)
3299  * @return bool Success (able to connect and no timeouts reached)
3300  * @deprecated since 1.27 Use LBFactory::waitForReplication
3301  */
3302 function wfWaitForSlaves(
3303         $ifWritesSince = null, $wiki = false, $cluster = false, $timeout = null
3304 ) {
3305         if ( $timeout === null ) {
3306                 $timeout = ( PHP_SAPI === 'cli' ) ? 86400 : 10;
3307         }
3308
3309         if ( $cluster === '*' ) {
3310                 $cluster = false;
3311                 $wiki = false;
3312         } elseif ( $wiki === false ) {
3313                 $wiki = wfWikiID();
3314         }
3315
3316         try {
3317                 wfGetLBFactory()->waitForReplication( [
3318                         'wiki' => $wiki,
3319                         'cluster' => $cluster,
3320                         'timeout' => $timeout,
3321                         // B/C: first argument used to be "max seconds of lag"; ignore such values
3322                         'ifWritesSince' => ( $ifWritesSince > 1e9 ) ? $ifWritesSince : null
3323                 ] );
3324         } catch ( DBReplicationWaitError $e ) {
3325                 return false;
3326         }
3327
3328         return true;
3329 }
3330
3331 /**
3332  * Count down from $seconds to zero on the terminal, with a one-second pause
3333  * between showing each number. For use in command-line scripts.
3334  *
3335  * @codeCoverageIgnore
3336  * @param int $seconds
3337  */
3338 function wfCountDown( $seconds ) {
3339         for ( $i = $seconds; $i >= 0; $i-- ) {
3340                 if ( $i != $seconds ) {
3341                         echo str_repeat( "\x08", strlen( $i + 1 ) );
3342                 }
3343                 echo $i;
3344                 flush();
3345                 if ( $i ) {
3346                         sleep( 1 );
3347                 }
3348         }
3349         echo "\n";
3350 }
3351
3352 /**
3353  * Replace all invalid characters with '-'.
3354  * Additional characters can be defined in $wgIllegalFileChars (see T22489).
3355  * By default, $wgIllegalFileChars includes ':', '/', '\'.
3356  *
3357  * @param string $name Filename to process
3358  * @return string
3359  */
3360 function wfStripIllegalFilenameChars( $name ) {
3361         global $wgIllegalFileChars;
3362         $illegalFileChars = $wgIllegalFileChars ? "|[" . $wgIllegalFileChars . "]" : '';
3363         $name = preg_replace(
3364                 "/[^" . Title::legalChars() . "]" . $illegalFileChars . "/",
3365                 '-',
3366                 $name
3367         );
3368         // $wgIllegalFileChars may not include '/' and '\', so we still need to do this
3369         $name = wfBaseName( $name );
3370         return $name;
3371 }
3372
3373 /**
3374  * Set PHP's memory limit to the larger of php.ini or $wgMemoryLimit
3375  *
3376  * @return int Resulting value of the memory limit.
3377  */
3378 function wfMemoryLimit() {
3379         global $wgMemoryLimit;
3380         $memlimit = wfShorthandToInteger( ini_get( 'memory_limit' ) );
3381         if ( $memlimit != -1 ) {
3382                 $conflimit = wfShorthandToInteger( $wgMemoryLimit );
3383                 if ( $conflimit == -1 ) {
3384                         wfDebug( "Removing PHP's memory limit\n" );
3385                         MediaWiki\suppressWarnings();
3386                         ini_set( 'memory_limit', $conflimit );
3387                         MediaWiki\restoreWarnings();
3388                         return $conflimit;
3389                 } elseif ( $conflimit > $memlimit ) {
3390                         wfDebug( "Raising PHP's memory limit to $conflimit bytes\n" );
3391                         MediaWiki\suppressWarnings();
3392                         ini_set( 'memory_limit', $conflimit );
3393                         MediaWiki\restoreWarnings();
3394                         return $conflimit;
3395                 }
3396         }
3397         return $memlimit;
3398 }
3399
3400 /**
3401  * Set PHP's time limit to the larger of php.ini or $wgTransactionalTimeLimit
3402  *
3403  * @return int Prior time limit
3404  * @since 1.26
3405  */
3406 function wfTransactionalTimeLimit() {
3407         global $wgTransactionalTimeLimit;
3408
3409         $timeLimit = ini_get( 'max_execution_time' );
3410         // Note that CLI scripts use 0
3411         if ( $timeLimit > 0 && $wgTransactionalTimeLimit > $timeLimit ) {
3412                 set_time_limit( $wgTransactionalTimeLimit );
3413         }
3414
3415         ignore_user_abort( true ); // ignore client disconnects
3416
3417         return $timeLimit;
3418 }
3419
3420 /**
3421  * Converts shorthand byte notation to integer form
3422  *
3423  * @param string $string
3424  * @param int $default Returned if $string is empty
3425  * @return int
3426  */
3427 function wfShorthandToInteger( $string = '', $default = -1 ) {
3428         $string = trim( $string );
3429         if ( $string === '' ) {
3430                 return $default;
3431         }
3432         $last = $string[strlen( $string ) - 1];
3433         $val = intval( $string );
3434         switch ( $last ) {
3435                 case 'g':
3436                 case 'G':
3437                         $val *= 1024;
3438                         // break intentionally missing
3439                 case 'm':
3440                 case 'M':
3441                         $val *= 1024;
3442                         // break intentionally missing
3443                 case 'k':
3444                 case 'K':
3445                         $val *= 1024;
3446         }
3447
3448         return $val;
3449 }
3450
3451 /**
3452  * Get the normalised IETF language tag
3453  * See unit test for examples.
3454  *
3455  * @param string $code The language code.
3456  * @return string The language code which complying with BCP 47 standards.
3457  */
3458 function wfBCP47( $code ) {
3459         $codeSegment = explode( '-', $code );
3460         $codeBCP = [];
3461         foreach ( $codeSegment as $segNo => $seg ) {
3462                 // when previous segment is x, it is a private segment and should be lc
3463                 if ( $segNo > 0 && strtolower( $codeSegment[( $segNo - 1 )] ) == 'x' ) {
3464                         $codeBCP[$segNo] = strtolower( $seg );
3465                 // ISO 3166 country code
3466                 } elseif ( ( strlen( $seg ) == 2 ) && ( $segNo > 0 ) ) {
3467                         $codeBCP[$segNo] = strtoupper( $seg );
3468                 // ISO 15924 script code
3469                 } elseif ( ( strlen( $seg ) == 4 ) && ( $segNo > 0 ) ) {
3470                         $codeBCP[$segNo] = ucfirst( strtolower( $seg ) );
3471                 // Use lowercase for other cases
3472                 } else {
3473                         $codeBCP[$segNo] = strtolower( $seg );
3474                 }
3475         }
3476         $langCode = implode( '-', $codeBCP );
3477         return $langCode;
3478 }
3479
3480 /**
3481  * Get a specific cache object.
3482  *
3483  * @param int|string $cacheType A CACHE_* constants, or other key in $wgObjectCaches
3484  * @return BagOStuff
3485  */
3486 function wfGetCache( $cacheType ) {
3487         return ObjectCache::getInstance( $cacheType );
3488 }
3489
3490 /**
3491  * Get the main cache object
3492  *
3493  * @return BagOStuff
3494  */
3495 function wfGetMainCache() {
3496         global $wgMainCacheType;
3497         return ObjectCache::getInstance( $wgMainCacheType );
3498 }
3499
3500 /**
3501  * Get the cache object used by the message cache
3502  *
3503  * @return BagOStuff
3504  */
3505 function wfGetMessageCacheStorage() {
3506         global $wgMessageCacheType;
3507         return ObjectCache::getInstance( $wgMessageCacheType );
3508 }
3509
3510 /**
3511  * Get the cache object used by the parser cache
3512  *
3513  * @return BagOStuff
3514  */
3515 function wfGetParserCacheStorage() {
3516         global $wgParserCacheType;
3517         return ObjectCache::getInstance( $wgParserCacheType );
3518 }
3519
3520 /**
3521  * Call hook functions defined in $wgHooks
3522  *
3523  * @param string $event Event name
3524  * @param array $args Parameters passed to hook functions
3525  * @param string|null $deprecatedVersion Optionally mark hook as deprecated with version number
3526  *
3527  * @return bool True if no handler aborted the hook
3528  * @deprecated since 1.25 - use Hooks::run
3529  */
3530 function wfRunHooks( $event, array $args = [], $deprecatedVersion = null ) {
3531         return Hooks::run( $event, $args, $deprecatedVersion );
3532 }
3533
3534 /**
3535  * Wrapper around php's unpack.
3536  *
3537  * @param string $format The format string (See php's docs)
3538  * @param string $data A binary string of binary data
3539  * @param int|bool $length The minimum length of $data or false. This is to
3540  *      prevent reading beyond the end of $data. false to disable the check.
3541  *
3542  * Also be careful when using this function to read unsigned 32 bit integer
3543  * because php might make it negative.
3544  *
3545  * @throws MWException If $data not long enough, or if unpack fails
3546  * @return array Associative array of the extracted data
3547  */
3548 function wfUnpack( $format, $data, $length = false ) {
3549         if ( $length !== false ) {
3550                 $realLen = strlen( $data );
3551                 if ( $realLen < $length ) {
3552                         throw new MWException( "Tried to use wfUnpack on a "
3553                                 . "string of length $realLen, but needed one "
3554                                 . "of at least length $length."
3555                         );
3556                 }
3557         }
3558
3559         MediaWiki\suppressWarnings();
3560         $result = unpack( $format, $data );
3561         MediaWiki\restoreWarnings();
3562
3563         if ( $result === false ) {
3564                 // If it cannot extract the packed data.
3565                 throw new MWException( "unpack could not unpack binary data" );
3566         }
3567         return $result;
3568 }
3569
3570 /**
3571  * Determine if an image exists on the 'bad image list'.
3572  *
3573  * The format of MediaWiki:Bad_image_list is as follows:
3574  *    * Only list items (lines starting with "*") are considered
3575  *    * The first link on a line must be a link to a bad image
3576  *    * Any subsequent links on the same line are considered to be exceptions,
3577  *      i.e. articles where the image may occur inline.
3578  *
3579  * @param string $name The image name to check
3580  * @param Title|bool $contextTitle The page on which the image occurs, if known
3581  * @param string $blacklist Wikitext of a file blacklist
3582  * @return bool
3583  */
3584 function wfIsBadImage( $name, $contextTitle = false, $blacklist = null ) {
3585         # Handle redirects; callers almost always hit wfFindFile() anyway,
3586         # so just use that method because it has a fast process cache.
3587         $file = wfFindFile( $name ); // get the final name
3588         $name = $file ? $file->getTitle()->getDBkey() : $name;
3589
3590         # Run the extension hook
3591         $bad = false;
3592         if ( !Hooks::run( 'BadImage', [ $name, &$bad ] ) ) {
3593                 return $bad;
3594         }
3595
3596         $cache = ObjectCache::getLocalServerInstance( 'hash' );
3597         $key = wfMemcKey( 'bad-image-list', ( $blacklist === null ) ? 'default' : md5( $blacklist ) );
3598         $badImages = $cache->get( $key );
3599
3600         if ( $badImages === false ) { // cache miss
3601                 if ( $blacklist === null ) {
3602                         $blacklist = wfMessage( 'bad_image_list' )->inContentLanguage()->plain(); // site list
3603                 }
3604                 # Build the list now
3605                 $badImages = [];
3606                 $lines = explode( "\n", $blacklist );
3607                 foreach ( $lines as $line ) {
3608                         # List items only
3609                         if ( substr( $line, 0, 1 ) !== '*' ) {
3610                                 continue;
3611                         }
3612
3613                         # Find all links
3614                         $m = [];
3615                         if ( !preg_match_all( '/\[\[:?(.*?)\]\]/', $line, $m ) ) {
3616                                 continue;
3617                         }
3618
3619                         $exceptions = [];
3620                         $imageDBkey = false;
3621                         foreach ( $m[1] as $i => $titleText ) {
3622                                 $title = Title::newFromText( $titleText );
3623                                 if ( !is_null( $title ) ) {
3624                                         if ( $i == 0 ) {
3625                                                 $imageDBkey = $title->getDBkey();
3626                                         } else {
3627                                                 $exceptions[$title->getPrefixedDBkey()] = true;
3628                                         }
3629                                 }
3630                         }
3631
3632                         if ( $imageDBkey !== false ) {
3633                                 $badImages[$imageDBkey] = $exceptions;
3634                         }
3635                 }
3636                 $cache->set( $key, $badImages, 60 );
3637         }
3638
3639         $contextKey = $contextTitle ? $contextTitle->getPrefixedDBkey() : false;
3640         $bad = isset( $badImages[$name] ) && !isset( $badImages[$name][$contextKey] );
3641
3642         return $bad;
3643 }
3644
3645 /**
3646  * Determine whether the client at a given source IP is likely to be able to
3647  * access the wiki via HTTPS.
3648  *
3649  * @param string $ip The IPv4/6 address in the normal human-readable form
3650  * @return bool
3651  */
3652 function wfCanIPUseHTTPS( $ip ) {
3653         $canDo = true;
3654         Hooks::run( 'CanIPUseHTTPS', [ $ip, &$canDo ] );
3655         return !!$canDo;
3656 }
3657
3658 /**
3659  * Determine input string is represents as infinity
3660  *
3661  * @param string $str The string to determine
3662  * @return bool
3663  * @since 1.25
3664  */
3665 function wfIsInfinity( $str ) {
3666         $infinityValues = [ 'infinite', 'indefinite', 'infinity', 'never' ];
3667         return in_array( $str, $infinityValues );
3668 }
3669
3670 /**
3671  * Returns true if these thumbnail parameters match one that MediaWiki
3672  * requests from file description pages and/or parser output.
3673  *
3674  * $params is considered non-standard if they involve a non-standard
3675  * width or any non-default parameters aside from width and page number.
3676  * The number of possible files with standard parameters is far less than
3677  * that of all combinations; rate-limiting for them can thus be more generious.
3678  *
3679  * @param File $file
3680  * @param array $params
3681  * @return bool
3682  * @since 1.24 Moved from thumb.php to GlobalFunctions in 1.25
3683  */
3684 function wfThumbIsStandard( File $file, array $params ) {
3685         global $wgThumbLimits, $wgImageLimits, $wgResponsiveImages;
3686
3687         $multipliers = [ 1 ];
3688         if ( $wgResponsiveImages ) {
3689                 // These available sizes are hardcoded currently elsewhere in MediaWiki.
3690                 // @see Linker::processResponsiveImages
3691                 $multipliers[] = 1.5;
3692                 $multipliers[] = 2;
3693         }
3694
3695         $handler = $file->getHandler();
3696         if ( !$handler || !isset( $params['width'] ) ) {
3697                 return false;
3698         }
3699
3700         $basicParams = [];
3701         if ( isset( $params['page'] ) ) {
3702                 $basicParams['page'] = $params['page'];
3703         }
3704
3705         $thumbLimits = [];
3706         $imageLimits = [];
3707         // Expand limits to account for multipliers
3708         foreach ( $multipliers as $multiplier ) {
3709                 $thumbLimits = array_merge( $thumbLimits, array_map(
3710                         function ( $width ) use ( $multiplier ) {
3711                                 return round( $width * $multiplier );
3712                         }, $wgThumbLimits )
3713                 );
3714                 $imageLimits = array_merge( $imageLimits, array_map(
3715                         function ( $pair ) use ( $multiplier ) {
3716                                 return [
3717                                         round( $pair[0] * $multiplier ),
3718                                         round( $pair[1] * $multiplier ),
3719                                 ];
3720                         }, $wgImageLimits )
3721                 );
3722         }
3723
3724         // Check if the width matches one of $wgThumbLimits
3725         if ( in_array( $params['width'], $thumbLimits ) ) {
3726                 $normalParams = $basicParams + [ 'width' => $params['width'] ];
3727                 // Append any default values to the map (e.g. "lossy", "lossless", ...)
3728                 $handler->normaliseParams( $file, $normalParams );
3729         } else {
3730                 // If not, then check if the width matchs one of $wgImageLimits
3731                 $match = false;
3732                 foreach ( $imageLimits as $pair ) {
3733                         $normalParams = $basicParams + [ 'width' => $pair[0], 'height' => $pair[1] ];
3734                         // Decide whether the thumbnail should be scaled on width or height.
3735                         // Also append any default values to the map (e.g. "lossy", "lossless", ...)
3736                         $handler->normaliseParams( $file, $normalParams );
3737                         // Check if this standard thumbnail size maps to the given width
3738                         if ( $normalParams['width'] == $params['width'] ) {
3739                                 $match = true;
3740                                 break;
3741                         }
3742                 }
3743                 if ( !$match ) {
3744                         return false; // not standard for description pages
3745                 }
3746         }
3747
3748         // Check that the given values for non-page, non-width, params are just defaults
3749         foreach ( $params as $key => $value ) {
3750                 if ( !isset( $normalParams[$key] ) || $normalParams[$key] != $value ) {
3751                         return false;
3752                 }
3753         }
3754
3755         return true;
3756 }
3757
3758 /**
3759  * Merges two (possibly) 2 dimensional arrays into the target array ($baseArray).
3760  *
3761  * Values that exist in both values will be combined with += (all values of the array
3762  * of $newValues will be added to the values of the array of $baseArray, while values,
3763  * that exists in both, the value of $baseArray will be used).
3764  *
3765  * @param array $baseArray The array where you want to add the values of $newValues to
3766  * @param array $newValues An array with new values
3767  * @return array The combined array
3768  * @since 1.26
3769  */
3770 function wfArrayPlus2d( array $baseArray, array $newValues ) {
3771         // First merge items that are in both arrays
3772         foreach ( $baseArray as $name => &$groupVal ) {
3773                 if ( isset( $newValues[$name] ) ) {
3774                         $groupVal += $newValues[$name];
3775                 }
3776         }
3777         // Now add items that didn't exist yet
3778         $baseArray += $newValues;
3779
3780         return $baseArray;
3781 }