includes/GlobalFunctions.php

   1 <?php
   2 /**
   3  * Global functions used everywhere.
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  */
  22
  23 if ( !defined( 'MEDIAWIKI' ) ) {
  24         die( "This file is part of MediaWiki, it is not a valid entry point" );
  25 }
  26
  27 use Liuggio\StatsdClient\Sender\SocketSender;
  28 use MediaWiki\Logger\LoggerFactory;
  29 use MediaWiki\Session\SessionManager;
  30 use MediaWiki\MediaWikiServices;
  31 use Wikimedia\ScopedCallback;
  32 use Wikimedia\Rdbms\DBReplicationWaitError;
  33
  34 // Hide compatibility functions from Doxygen
  35 /// @cond
  36 /**
  37  * Compatibility functions
  38  *
  39  * We support PHP 5.5.9 and up.
  40  * Re-implementations of newer functions or functions in non-standard
  41  * PHP extensions may be included here.
  42  */
  43
  44 // hash_equals function only exists in PHP >= 5.6.0
  45 // https://secure.php.net/hash_equals
  46 if ( !function_exists( 'hash_equals' ) ) {
  47         /**
  48          * Check whether a user-provided string is equal to a fixed-length secret string
  49          * without revealing bytes of the secret string through timing differences.
  50          *
  51          * The usual way to compare strings (PHP's === operator or the underlying memcmp()
  52          * function in C) is to compare corresponding bytes and stop at the first difference,
  53          * which would take longer for a partial match than for a complete mismatch. This
  54          * is not secure when one of the strings (e.g. an HMAC or token) must remain secret
  55          * and the other may come from an attacker. Statistical analysis of timing measurements
  56          * over many requests may allow the attacker to guess the string's bytes one at a time
  57          * (and check his guesses) even if the timing differences are extremely small.
  58          *
  59          * When making such a security-sensitive comparison, it is essential that the sequence
  60          * in which instructions are executed and memory locations are accessed not depend on
  61          * the secret string's value. HOWEVER, for simplicity, we do not attempt to minimize
  62          * the inevitable leakage of the string's length. That is generally known anyway as
  63          * a chararacteristic of the hash function used to compute the secret value.
  64          *
  65          * Longer explanation: http://www.emerose.com/timing-attacks-explained
  66          *
  67          * @codeCoverageIgnore
  68          * @param string $known_string Fixed-length secret string to compare against
  69          * @param string $user_string User-provided string
  70          * @return bool True if the strings are the same, false otherwise
  71          */
  72         function hash_equals( $known_string, $user_string ) {
  73                 // Strict type checking as in PHP's native implementation
  74                 if ( !is_string( $known_string ) ) {
  75                         trigger_error( 'hash_equals(): Expected known_string to be a string, ' .
  76                                 gettype( $known_string ) . ' given', E_USER_WARNING );
  77
  78                         return false;
  79                 }
  80
  81                 if ( !is_string( $user_string ) ) {
  82                         trigger_error( 'hash_equals(): Expected user_string to be a string, ' .
  83                                 gettype( $user_string ) . ' given', E_USER_WARNING );
  84
  85                         return false;
  86                 }
  87
  88                 $known_string_len = strlen( $known_string );
  89                 if ( $known_string_len !== strlen( $user_string ) ) {
  90                         return false;
  91                 }
  92
  93                 $result = 0;
  94                 for ( $i = 0; $i < $known_string_len; $i++ ) {
  95                         $result |= ord( $known_string[$i] ) ^ ord( $user_string[$i] );
  96                 }
  97
  98                 return ( $result === 0 );
  99         }
 100 }
 101 /// @endcond
 102
 103 /**
 104  * Load an extension
 105  *
 106  * This queues an extension to be loaded through
 107  * the ExtensionRegistry system.
 108  *
 109  * @param string $ext Name of the extension to load
 110  * @param string|null $path Absolute path of where to find the extension.json file
 111  * @since 1.25
 112  */
 113 function wfLoadExtension( $ext, $path = null ) {
 114         if ( !$path ) {
 115                 global $wgExtensionDirectory;
 116                 $path = "$wgExtensionDirectory/$ext/extension.json";
 117         }
 118         ExtensionRegistry::getInstance()->queue( $path );
 119 }
 120
 121 /**
 122  * Load multiple extensions at once
 123  *
 124  * Same as wfLoadExtension, but more efficient if you
 125  * are loading multiple extensions.
 126  *
 127  * If you want to specify custom paths, you should interact with
 128  * ExtensionRegistry directly.
 129  *
 130  * @see wfLoadExtension
 131  * @param string[] $exts Array of extension names to load
 132  * @since 1.25
 133  */
 134 function wfLoadExtensions( array $exts ) {
 135         global $wgExtensionDirectory;
 136         $registry = ExtensionRegistry::getInstance();
 137         foreach ( $exts as $ext ) {
 138                 $registry->queue( "$wgExtensionDirectory/$ext/extension.json" );
 139         }
 140 }
 141
 142 /**
 143  * Load a skin
 144  *
 145  * @see wfLoadExtension
 146  * @param string $skin Name of the extension to load
 147  * @param string|null $path Absolute path of where to find the skin.json file
 148  * @since 1.25
 149  */
 150 function wfLoadSkin( $skin, $path = null ) {
 151         if ( !$path ) {
 152                 global $wgStyleDirectory;
 153                 $path = "$wgStyleDirectory/$skin/skin.json";
 154         }
 155         ExtensionRegistry::getInstance()->queue( $path );
 156 }
 157
 158 /**
 159  * Load multiple skins at once
 160  *
 161  * @see wfLoadExtensions
 162  * @param string[] $skins Array of extension names to load
 163  * @since 1.25
 164  */
 165 function wfLoadSkins( array $skins ) {
 166         global $wgStyleDirectory;
 167         $registry = ExtensionRegistry::getInstance();
 168         foreach ( $skins as $skin ) {
 169                 $registry->queue( "$wgStyleDirectory/$skin/skin.json" );
 170         }
 171 }
 172
 173 /**
 174  * Like array_diff( $a, $b ) except that it works with two-dimensional arrays.
 175  * @param array $a
 176  * @param array $b
 177  * @return array
 178  */
 179 function wfArrayDiff2( $a, $b ) {
 180         return array_udiff( $a, $b, 'wfArrayDiff2_cmp' );
 181 }
 182
 183 /**
 184  * @param array|string $a
 185  * @param array|string $b
 186  * @return int
 187  */
 188 function wfArrayDiff2_cmp( $a, $b ) {
 189         if ( is_string( $a ) && is_string( $b ) ) {
 190                 return strcmp( $a, $b );
 191         } elseif ( count( $a ) !== count( $b ) ) {
 192                 return count( $a ) < count( $b ) ? -1 : 1;
 193         } else {
 194                 reset( $a );
 195                 reset( $b );
 196                 while ( ( list( , $valueA ) = each( $a ) ) && ( list( , $valueB ) = each( $b ) ) ) {
 197                         $cmp = strcmp( $valueA, $valueB );
 198                         if ( $cmp !== 0 ) {
 199                                 return $cmp;
 200                         }
 201                 }
 202                 return 0;
 203         }
 204 }
 205
 206 /**
 207  * Appends to second array if $value differs from that in $default
 208  *
 209  * @param string|int $key
 210  * @param mixed $value
 211  * @param mixed $default
 212  * @param array $changed Array to alter
 213  * @throws MWException
 214  */
 215 function wfAppendToArrayIfNotDefault( $key, $value, $default, &$changed ) {
 216         if ( is_null( $changed ) ) {
 217                 throw new MWException( 'GlobalFunctions::wfAppendToArrayIfNotDefault got null' );
 218         }
 219         if ( $default[$key] !== $value ) {
 220                 $changed[$key] = $value;
 221         }
 222 }
 223
 224 /**
 225  * Merge arrays in the style of getUserPermissionsErrors, with duplicate removal
 226  * e.g.
 227  *     wfMergeErrorArrays(
 228  *       [ [ 'x' ] ],
 229  *       [ [ 'x', '2' ] ],
 230  *       [ [ 'x' ] ],
 231  *       [ [ 'y' ] ]
 232  *     );
 233  * returns:
 234  *     [
 235  *       [ 'x', '2' ],
 236  *       [ 'x' ],
 237  *       [ 'y' ]
 238  *     ]
 239  *
 240  * @param array $array1,...
 241  * @return array
 242  */
 243 function wfMergeErrorArrays( /*...*/ ) {
 244         $args = func_get_args();
 245         $out = [];
 246         foreach ( $args as $errors ) {
 247                 foreach ( $errors as $params ) {
 248                         $originalParams = $params;
 249                         if ( $params[0] instanceof MessageSpecifier ) {
 250                                 $msg = $params[0];
 251                                 $params = array_merge( [ $msg->getKey() ], $msg->getParams() );
 252                         }
 253                         # @todo FIXME: Sometimes get nested arrays for $params,
 254                         # which leads to E_NOTICEs
 255                         $spec = implode( "\t", $params );
 256                         $out[$spec] = $originalParams;
 257                 }
 258         }
 259         return array_values( $out );
 260 }
 261
 262 /**
 263  * Insert array into another array after the specified *KEY*
 264  *
 265  * @param array $array The array.
 266  * @param array $insert The array to insert.
 267  * @param mixed $after The key to insert after
 268  * @return array
 269  */
 270 function wfArrayInsertAfter( array $array, array $insert, $after ) {
 271         // Find the offset of the element to insert after.
 272         $keys = array_keys( $array );
 273         $offsetByKey = array_flip( $keys );
 274
 275         $offset = $offsetByKey[$after];
 276
 277         // Insert at the specified offset
 278         $before = array_slice( $array, 0, $offset + 1, true );
 279         $after = array_slice( $array, $offset + 1, count( $array ) - $offset, true );
 280
 281         $output = $before + $insert + $after;
 282
 283         return $output;
 284 }
 285
 286 /**
 287  * Recursively converts the parameter (an object) to an array with the same data
 288  *
 289  * @param object|array $objOrArray
 290  * @param bool $recursive
 291  * @return array
 292  */
 293 function wfObjectToArray( $objOrArray, $recursive = true ) {
 294         $array = [];
 295         if ( is_object( $objOrArray ) ) {
 296                 $objOrArray = get_object_vars( $objOrArray );
 297         }
 298         foreach ( $objOrArray as $key => $value ) {
 299                 if ( $recursive && ( is_object( $value ) || is_array( $value ) ) ) {
 300                         $value = wfObjectToArray( $value );
 301                 }
 302
 303                 $array[$key] = $value;
 304         }
 305
 306         return $array;
 307 }
 308
 309 /**
 310  * Get a random decimal value between 0 and 1, in a way
 311  * not likely to give duplicate values for any realistic
 312  * number of articles.
 313  *
 314  * @note This is designed for use in relation to Special:RandomPage
 315  *       and the page_random database field.
 316  *
 317  * @return string
 318  */
 319 function wfRandom() {
 320         // The maximum random value is "only" 2^31-1, so get two random
 321         // values to reduce the chance of dupes
 322         $max = mt_getrandmax() + 1;
 323         $rand = number_format( ( mt_rand() * $max + mt_rand() ) / $max / $max, 12, '.', '' );
 324         return $rand;
 325 }
 326
 327 /**
 328  * Get a random string containing a number of pseudo-random hex characters.
 329  *
 330  * @note This is not secure, if you are trying to generate some sort
 331  *       of token please use MWCryptRand instead.
 332  *
 333  * @param int $length The length of the string to generate
 334  * @return string
 335  * @since 1.20
 336  */
 337 function wfRandomString( $length = 32 ) {
 338         $str = '';
 339         for ( $n = 0; $n < $length; $n += 7 ) {
 340                 $str .= sprintf( '%07x', mt_rand() & 0xfffffff );
 341         }
 342         return substr( $str, 0, $length );
 343 }
 344
 345 /**
 346  * We want some things to be included as literal characters in our title URLs
 347  * for prettiness, which urlencode encodes by default.  According to RFC 1738,
 348  * all of the following should be safe:
 349  *
 350  * ;:@&=$-_.+!*'(),
 351  *
 352  * RFC 1738 says ~ is unsafe, however RFC 3986 considers it an unreserved
 353  * character which should not be encoded. More importantly, google chrome
 354  * always converts %7E back to ~, and converting it in this function can
 355  * cause a redirect loop (T105265).
 356  *
 357  * But + is not safe because it's used to indicate a space; &= are only safe in
 358  * paths and not in queries (and we don't distinguish here); ' seems kind of
 359  * scary; and urlencode() doesn't touch -_. to begin with.  Plus, although /
 360  * is reserved, we don't care.  So the list we unescape is:
 361  *
 362  * ;:@$!*(),/~
 363  *
 364  * However, IIS7 redirects fail when the url contains a colon (see T24709),
 365  * so no fancy : for IIS7.
 366  *
 367  * %2F in the page titles seems to fatally break for some reason.
 368  *
 369  * @param string $s
 370  * @return string
 371  */
 372 function wfUrlencode( $s ) {
 373         static $needle;
 374
 375         if ( is_null( $s ) ) {
 376                 $needle = null;
 377                 return '';
 378         }
 379
 380         if ( is_null( $needle ) ) {
 381                 $needle = [ '%3B', '%40', '%24', '%21', '%2A', '%28', '%29', '%2C', '%2F', '%7E' ];
 382                 if ( !isset( $_SERVER['SERVER_SOFTWARE'] ) ||
 383                         ( strpos( $_SERVER['SERVER_SOFTWARE'], 'Microsoft-IIS/7' ) === false )
 384                 ) {
 385                         $needle[] = '%3A';
 386                 }
 387         }
 388
 389         $s = urlencode( $s );
 390         $s = str_ireplace(
 391                 $needle,
 392                 [ ';', '@', '$', '!', '*', '(', ')', ',', '/', '~', ':' ],
 393                 $s
 394         );
 395
 396         return $s;
 397 }
 398
 399 /**
 400  * This function takes one or two arrays as input, and returns a CGI-style string, e.g.
 401  * "days=7&limit=100". Options in the first array override options in the second.
 402  * Options set to null or false will not be output.
 403  *
 404  * @param array $array1 ( String|Array )
 405  * @param array|null $array2 ( String|Array )
 406  * @param string $prefix
 407  * @return string
 408  */
 409 function wfArrayToCgi( $array1, $array2 = null, $prefix = '' ) {
 410         if ( !is_null( $array2 ) ) {
 411                 $array1 = $array1 + $array2;
 412         }
 413
 414         $cgi = '';
 415         foreach ( $array1 as $key => $value ) {
 416                 if ( !is_null( $value ) && $value !== false ) {
 417                         if ( $cgi != '' ) {
 418                                 $cgi .= '&';
 419                         }
 420                         if ( $prefix !== '' ) {
 421                                 $key = $prefix . "[$key]";
 422                         }
 423                         if ( is_array( $value ) ) {
 424                                 $firstTime = true;
 425                                 foreach ( $value as $k => $v ) {
 426                                         $cgi .= $firstTime ? '' : '&';
 427                                         if ( is_array( $v ) ) {
 428                                                 $cgi .= wfArrayToCgi( $v, null, $key . "[$k]" );
 429                                         } else {
 430                                                 $cgi .= urlencode( $key . "[$k]" ) . '=' . urlencode( $v );
 431                                         }
 432                                         $firstTime = false;
 433                                 }
 434                         } else {
 435                                 if ( is_object( $value ) ) {
 436                                         $value = $value->__toString();
 437                                 }
 438                                 $cgi .= urlencode( $key ) . '=' . urlencode( $value );
 439                         }
 440                 }
 441         }
 442         return $cgi;
 443 }
 444
 445 /**
 446  * This is the logical opposite of wfArrayToCgi(): it accepts a query string as
 447  * its argument and returns the same string in array form.  This allows compatibility
 448  * with legacy functions that accept raw query strings instead of nice
 449  * arrays.  Of course, keys and values are urldecode()d.
 450  *
 451  * @param string $query Query string
 452  * @return string[] Array version of input
 453  */
 454 function wfCgiToArray( $query ) {
 455         if ( isset( $query[0] ) && $query[0] == '?' ) {
 456                 $query = substr( $query, 1 );
 457         }
 458         $bits = explode( '&', $query );
 459         $ret = [];
 460         foreach ( $bits as $bit ) {
 461                 if ( $bit === '' ) {
 462                         continue;
 463                 }
 464                 if ( strpos( $bit, '=' ) === false ) {
 465                         // Pieces like &qwerty become 'qwerty' => '' (at least this is what php does)
 466                         $key = $bit;
 467                         $value = '';
 468                 } else {
 469                         list( $key, $value ) = explode( '=', $bit );
 470                 }
 471                 $key = urldecode( $key );
 472                 $value = urldecode( $value );
 473                 if ( strpos( $key, '[' ) !== false ) {
 474                         $keys = array_reverse( explode( '[', $key ) );
 475                         $key = array_pop( $keys );
 476                         $temp = $value;
 477                         foreach ( $keys as $k ) {
 478                                 $k = substr( $k, 0, -1 );
 479                                 $temp = [ $k => $temp ];
 480                         }
 481                         if ( isset( $ret[$key] ) ) {
 482                                 $ret[$key] = array_merge( $ret[$key], $temp );
 483                         } else {
 484                                 $ret[$key] = $temp;
 485                         }
 486                 } else {
 487                         $ret[$key] = $value;
 488                 }
 489         }
 490         return $ret;
 491 }
 492
 493 /**
 494  * Append a query string to an existing URL, which may or may not already
 495  * have query string parameters already. If so, they will be combined.
 496  *
 497  * @param string $url
 498  * @param string|string[] $query String or associative array
 499  * @return string
 500  */
 501 function wfAppendQuery( $url, $query ) {
 502         if ( is_array( $query ) ) {
 503                 $query = wfArrayToCgi( $query );
 504         }
 505         if ( $query != '' ) {
 506                 // Remove the fragment, if there is one
 507                 $fragment = false;
 508                 $hashPos = strpos( $url, '#' );
 509                 if ( $hashPos !== false ) {
 510                         $fragment = substr( $url, $hashPos );
 511                         $url = substr( $url, 0, $hashPos );
 512                 }
 513
 514                 // Add parameter
 515                 if ( false === strpos( $url, '?' ) ) {
 516                         $url .= '?';
 517                 } else {
 518                         $url .= '&';
 519                 }
 520                 $url .= $query;
 521
 522                 // Put the fragment back
 523                 if ( $fragment !== false ) {
 524                         $url .= $fragment;
 525                 }
 526         }
 527         return $url;
 528 }
 529
 530 /**
 531  * Expand a potentially local URL to a fully-qualified URL. Assumes $wgServer
 532  * is correct.
 533  *
 534  * The meaning of the PROTO_* constants is as follows:
 535  * PROTO_HTTP: Output a URL starting with http://
 536  * PROTO_HTTPS: Output a URL starting with https://
 537  * PROTO_RELATIVE: Output a URL starting with // (protocol-relative URL)
 538  * PROTO_CURRENT: Output a URL starting with either http:// or https:// , depending
 539  *    on which protocol was used for the current incoming request
 540  * PROTO_CANONICAL: For URLs without a domain, like /w/index.php , use $wgCanonicalServer.
 541  *    For protocol-relative URLs, use the protocol of $wgCanonicalServer
 542  * PROTO_INTERNAL: Like PROTO_CANONICAL, but uses $wgInternalServer instead of $wgCanonicalServer
 543  *
 544  * @todo this won't work with current-path-relative URLs
 545  * like "subdir/foo.html", etc.
 546  *
 547  * @param string $url Either fully-qualified or a local path + query
 548  * @param string $defaultProto One of the PROTO_* constants. Determines the
 549  *    protocol to use if $url or $wgServer is protocol-relative
 550  * @return string|false Fully-qualified URL, current-path-relative URL or false if
 551  *    no valid URL can be constructed
 552  */
 553 function wfExpandUrl( $url, $defaultProto = PROTO_CURRENT ) {
 554         global $wgServer, $wgCanonicalServer, $wgInternalServer, $wgRequest,
 555                 $wgHttpsPort;
 556         if ( $defaultProto === PROTO_CANONICAL ) {
 557                 $serverUrl = $wgCanonicalServer;
 558         } elseif ( $defaultProto === PROTO_INTERNAL && $wgInternalServer !== false ) {
 559                 // Make $wgInternalServer fall back to $wgServer if not set
 560                 $serverUrl = $wgInternalServer;
 561         } else {
 562                 $serverUrl = $wgServer;
 563                 if ( $defaultProto === PROTO_CURRENT ) {
 564                         $defaultProto = $wgRequest->getProtocol() . '://';
 565                 }
 566         }
 567
 568         // Analyze $serverUrl to obtain its protocol
 569         $bits = wfParseUrl( $serverUrl );
 570         $serverHasProto = $bits && $bits['scheme'] != '';
 571
 572         if ( $defaultProto === PROTO_CANONICAL || $defaultProto === PROTO_INTERNAL ) {
 573                 if ( $serverHasProto ) {
 574                         $defaultProto = $bits['scheme'] . '://';
 575                 } else {
 576                         // $wgCanonicalServer or $wgInternalServer doesn't have a protocol.
 577                         // This really isn't supposed to happen. Fall back to HTTP in this
 578                         // ridiculous case.
 579                         $defaultProto = PROTO_HTTP;
 580                 }
 581         }
 582
 583         $defaultProtoWithoutSlashes = substr( $defaultProto, 0, -2 );
 584
 585         if ( substr( $url, 0, 2 ) == '//' ) {
 586                 $url = $defaultProtoWithoutSlashes . $url;
 587         } elseif ( substr( $url, 0, 1 ) == '/' ) {
 588                 // If $serverUrl is protocol-relative, prepend $defaultProtoWithoutSlashes,
 589                 // otherwise leave it alone.
 590                 $url = ( $serverHasProto ? '' : $defaultProtoWithoutSlashes ) . $serverUrl . $url;
 591         }
 592
 593         $bits = wfParseUrl( $url );
 594
 595         // ensure proper port for HTTPS arrives in URL
 596         // https://phabricator.wikimedia.org/T67184
 597         if ( $defaultProto === PROTO_HTTPS && $wgHttpsPort != 443 ) {
 598                 $bits['port'] = $wgHttpsPort;
 599         }
 600
 601         if ( $bits && isset( $bits['path'] ) ) {
 602                 $bits['path'] = wfRemoveDotSegments( $bits['path'] );
 603                 return wfAssembleUrl( $bits );
 604         } elseif ( $bits ) {
 605                 # No path to expand
 606                 return $url;
 607         } elseif ( substr( $url, 0, 1 ) != '/' ) {
 608                 # URL is a relative path
 609                 return wfRemoveDotSegments( $url );
 610         }
 611
 612         # Expanded URL is not valid.
 613         return false;
 614 }
 615
 616 /**
 617  * This function will reassemble a URL parsed with wfParseURL.  This is useful
 618  * if you need to edit part of a URL and put it back together.
 619  *
 620  * This is the basic structure used (brackets contain keys for $urlParts):
 621  * [scheme][delimiter][user]:[pass]@[host]:[port][path]?[query]#[fragment]
 622  *
 623  * @todo Need to integrate this into wfExpandUrl (see T34168)
 624  *
 625  * @since 1.19
 626  * @param array $urlParts URL parts, as output from wfParseUrl
 627  * @return string URL assembled from its component parts
 628  */
 629 function wfAssembleUrl( $urlParts ) {
 630         $result = '';
 631
 632         if ( isset( $urlParts['delimiter'] ) ) {
 633                 if ( isset( $urlParts['scheme'] ) ) {
 634                         $result .= $urlParts['scheme'];
 635                 }
 636
 637                 $result .= $urlParts['delimiter'];
 638         }
 639
 640         if ( isset( $urlParts['host'] ) ) {
 641                 if ( isset( $urlParts['user'] ) ) {
 642                         $result .= $urlParts['user'];
 643                         if ( isset( $urlParts['pass'] ) ) {
 644                                 $result .= ':' . $urlParts['pass'];
 645                         }
 646                         $result .= '@';
 647                 }
 648
 649                 $result .= $urlParts['host'];
 650
 651                 if ( isset( $urlParts['port'] ) ) {
 652                         $result .= ':' . $urlParts['port'];
 653                 }
 654         }
 655
 656         if ( isset( $urlParts['path'] ) ) {
 657                 $result .= $urlParts['path'];
 658         }
 659
 660         if ( isset( $urlParts['query'] ) ) {
 661                 $result .= '?' . $urlParts['query'];
 662         }
 663
 664         if ( isset( $urlParts['fragment'] ) ) {
 665                 $result .= '#' . $urlParts['fragment'];
 666         }
 667
 668         return $result;
 669 }
 670
 671 /**
 672  * Remove all dot-segments in the provided URL path.  For example,
 673  * '/a/./b/../c/' becomes '/a/c/'.  For details on the algorithm, please see
 674  * RFC3986 section 5.2.4.
 675  *
 676  * @todo Need to integrate this into wfExpandUrl (see T34168)
 677  *
 678  * @param string $urlPath URL path, potentially containing dot-segments
 679  * @return string URL path with all dot-segments removed
 680  */
 681 function wfRemoveDotSegments( $urlPath ) {
 682         $output = '';
 683         $inputOffset = 0;
 684         $inputLength = strlen( $urlPath );
 685
 686         while ( $inputOffset < $inputLength ) {
 687                 $prefixLengthOne = substr( $urlPath, $inputOffset, 1 );
 688                 $prefixLengthTwo = substr( $urlPath, $inputOffset, 2 );
 689                 $prefixLengthThree = substr( $urlPath, $inputOffset, 3 );
 690                 $prefixLengthFour = substr( $urlPath, $inputOffset, 4 );
 691                 $trimOutput = false;
 692
 693                 if ( $prefixLengthTwo == './' ) {
 694                         # Step A, remove leading "./"
 695                         $inputOffset += 2;
 696                 } elseif ( $prefixLengthThree == '../' ) {
 697                         # Step A, remove leading "../"
 698                         $inputOffset += 3;
 699                 } elseif ( ( $prefixLengthTwo == '/.' ) && ( $inputOffset + 2 == $inputLength ) ) {
 700                         # Step B, replace leading "/.$" with "/"
 701                         $inputOffset += 1;
 702                         $urlPath[$inputOffset] = '/';
 703                 } elseif ( $prefixLengthThree == '/./' ) {
 704                         # Step B, replace leading "/./" with "/"
 705                         $inputOffset += 2;
 706                 } elseif ( $prefixLengthThree == '/..' && ( $inputOffset + 3 == $inputLength ) ) {
 707                         # Step C, replace leading "/..$" with "/" and
 708                         # remove last path component in output
 709                         $inputOffset += 2;
 710                         $urlPath[$inputOffset] = '/';
 711                         $trimOutput = true;
 712                 } elseif ( $prefixLengthFour == '/../' ) {
 713                         # Step C, replace leading "/../" with "/" and
 714                         # remove last path component in output
 715                         $inputOffset += 3;
 716                         $trimOutput = true;
 717                 } elseif ( ( $prefixLengthOne == '.' ) && ( $inputOffset + 1 == $inputLength ) ) {
 718                         # Step D, remove "^.$"
 719                         $inputOffset += 1;
 720                 } elseif ( ( $prefixLengthTwo == '..' ) && ( $inputOffset + 2 == $inputLength ) ) {
 721                         # Step D, remove "^..$"
 722                         $inputOffset += 2;
 723                 } else {
 724                         # Step E, move leading path segment to output
 725                         if ( $prefixLengthOne == '/' ) {
 726                                 $slashPos = strpos( $urlPath, '/', $inputOffset + 1 );
 727                         } else {
 728                                 $slashPos = strpos( $urlPath, '/', $inputOffset );
 729                         }
 730                         if ( $slashPos === false ) {
 731                                 $output .= substr( $urlPath, $inputOffset );
 732                                 $inputOffset = $inputLength;
 733                         } else {
 734                                 $output .= substr( $urlPath, $inputOffset, $slashPos - $inputOffset );
 735                                 $inputOffset += $slashPos - $inputOffset;
 736                         }
 737                 }
 738
 739                 if ( $trimOutput ) {
 740                         $slashPos = strrpos( $output, '/' );
 741                         if ( $slashPos === false ) {
 742                                 $output = '';
 743                         } else {
 744                                 $output = substr( $output, 0, $slashPos );
 745                         }
 746                 }
 747         }
 748
 749         return $output;
 750 }
 751
 752 /**
 753  * Returns a regular expression of url protocols
 754  *
 755  * @param bool $includeProtocolRelative If false, remove '//' from the returned protocol list.
 756  *        DO NOT USE this directly, use wfUrlProtocolsWithoutProtRel() instead
 757  * @return string
 758  */
 759 function wfUrlProtocols( $includeProtocolRelative = true ) {
 760         global $wgUrlProtocols;
 761
 762         // Cache return values separately based on $includeProtocolRelative
 763         static $withProtRel = null, $withoutProtRel = null;
 764         $cachedValue = $includeProtocolRelative ? $withProtRel : $withoutProtRel;
 765         if ( !is_null( $cachedValue ) ) {
 766                 return $cachedValue;
 767         }
 768
 769         // Support old-style $wgUrlProtocols strings, for backwards compatibility
 770         // with LocalSettings files from 1.5
 771         if ( is_array( $wgUrlProtocols ) ) {
 772                 $protocols = [];
 773                 foreach ( $wgUrlProtocols as $protocol ) {
 774                         // Filter out '//' if !$includeProtocolRelative
 775                         if ( $includeProtocolRelative || $protocol !== '//' ) {
 776                                 $protocols[] = preg_quote( $protocol, '/' );
 777                         }
 778                 }
 779
 780                 $retval = implode( '|', $protocols );
 781         } else {
 782                 // Ignore $includeProtocolRelative in this case
 783                 // This case exists for pre-1.6 compatibility, and we can safely assume
 784                 // that '//' won't appear in a pre-1.6 config because protocol-relative
 785                 // URLs weren't supported until 1.18
 786                 $retval = $wgUrlProtocols;
 787         }
 788
 789         // Cache return value
 790         if ( $includeProtocolRelative ) {
 791                 $withProtRel = $retval;
 792         } else {
 793                 $withoutProtRel = $retval;
 794         }
 795         return $retval;
 796 }
 797
 798 /**
 799  * Like wfUrlProtocols(), but excludes '//' from the protocol list. Use this if
 800  * you need a regex that matches all URL protocols but does not match protocol-
 801  * relative URLs
 802  * @return string
 803  */
 804 function wfUrlProtocolsWithoutProtRel() {
 805         return wfUrlProtocols( false );
 806 }
 807
 808 /**
 809  * parse_url() work-alike, but non-broken.  Differences:
 810  *
 811  * 1) Does not raise warnings on bad URLs (just returns false).
 812  * 2) Handles protocols that don't use :// (e.g., mailto: and news:, as well as
 813  *    protocol-relative URLs) correctly.
 814  * 3) Adds a "delimiter" element to the array, either '://', ':' or '//' (see (2)).
 815  *
 816  * @param string $url A URL to parse
 817  * @return string[]|bool Bits of the URL in an associative array, per PHP docs, false on failure
 818  */
 819 function wfParseUrl( $url ) {
 820         global $wgUrlProtocols; // Allow all protocols defined in DefaultSettings/LocalSettings.php
 821
 822         // Protocol-relative URLs are handled really badly by parse_url(). It's so
 823         // bad that the easiest way to handle them is to just prepend 'http:' and
 824         // strip the protocol out later.
 825         $wasRelative = substr( $url, 0, 2 ) == '//';
 826         if ( $wasRelative ) {
 827                 $url = "http:$url";
 828         }
 829         MediaWiki\suppressWarnings();
 830         $bits = parse_url( $url );
 831         MediaWiki\restoreWarnings();
 832         // parse_url() returns an array without scheme for some invalid URLs, e.g.
 833         // parse_url("%0Ahttp://example.com") == [ 'host' => '%0Ahttp', 'path' => 'example.com' ]
 834         if ( !$bits || !isset( $bits['scheme'] ) ) {
 835                 return false;
 836         }
 837
 838         // parse_url() incorrectly handles schemes case-sensitively. Convert it to lowercase.
 839         $bits['scheme'] = strtolower( $bits['scheme'] );
 840
 841         // most of the protocols are followed by ://, but mailto: and sometimes news: not, check for it
 842         if ( in_array( $bits['scheme'] . '://', $wgUrlProtocols ) ) {
 843                 $bits['delimiter'] = '://';
 844         } elseif ( in_array( $bits['scheme'] . ':', $wgUrlProtocols ) ) {
 845                 $bits['delimiter'] = ':';
 846                 // parse_url detects for news: and mailto: the host part of an url as path
 847                 // We have to correct this wrong detection
 848                 if ( isset( $bits['path'] ) ) {
 849                         $bits['host'] = $bits['path'];
 850                         $bits['path'] = '';
 851                 }
 852         } else {
 853                 return false;
 854         }
 855
 856         /* Provide an empty host for eg. file:/// urls (see T30627) */
 857         if ( !isset( $bits['host'] ) ) {
 858                 $bits['host'] = '';
 859
 860                 // See T47069
 861                 if ( isset( $bits['path'] ) ) {
 862                         /* parse_url loses the third / for file:///c:/ urls (but not on variants) */
 863                         if ( substr( $bits['path'], 0, 1 ) !== '/' ) {
 864                                 $bits['path'] = '/' . $bits['path'];
 865                         }
 866                 } else {
 867                         $bits['path'] = '';
 868                 }
 869         }
 870
 871         // If the URL was protocol-relative, fix scheme and delimiter
 872         if ( $wasRelative ) {
 873                 $bits['scheme'] = '';
 874                 $bits['delimiter'] = '//';
 875         }
 876         return $bits;
 877 }
 878
 879 /**
 880  * Take a URL, make sure it's expanded to fully qualified, and replace any
 881  * encoded non-ASCII Unicode characters with their UTF-8 original forms
 882  * for more compact display and legibility for local audiences.
 883  *
 884  * @todo handle punycode domains too
 885  *
 886  * @param string $url
 887  * @return string
 888  */
 889 function wfExpandIRI( $url ) {
 890         return preg_replace_callback(
 891                 '/((?:%[89A-F][0-9A-F])+)/i',
 892                 'wfExpandIRI_callback',
 893                 wfExpandUrl( $url )
 894         );
 895 }
 896
 897 /**
 898  * Private callback for wfExpandIRI
 899  * @param array $matches
 900  * @return string
 901  */
 902 function wfExpandIRI_callback( $matches ) {
 903         return urldecode( $matches[1] );
 904 }
 905
 906 /**
 907  * Make URL indexes, appropriate for the el_index field of externallinks.
 908  *
 909  * @param string $url
 910  * @return array
 911  */
 912 function wfMakeUrlIndexes( $url ) {
 913         $bits = wfParseUrl( $url );
 914
 915         // Reverse the labels in the hostname, convert to lower case
 916         // For emails reverse domainpart only
 917         if ( $bits['scheme'] == 'mailto' ) {
 918                 $mailparts = explode( '@', $bits['host'], 2 );
 919                 if ( count( $mailparts ) === 2 ) {
 920                         $domainpart = strtolower( implode( '.', array_reverse( explode( '.', $mailparts[1] ) ) ) );
 921                 } else {
 922                         // No domain specified, don't mangle it
 923                         $domainpart = '';
 924                 }
 925                 $reversedHost = $domainpart . '@' . $mailparts[0];
 926         } else {
 927                 $reversedHost = strtolower( implode( '.', array_reverse( explode( '.', $bits['host'] ) ) ) );
 928         }
 929         // Add an extra dot to the end
 930         // Why? Is it in wrong place in mailto links?
 931         if ( substr( $reversedHost, -1, 1 ) !== '.' ) {
 932                 $reversedHost .= '.';
 933         }
 934         // Reconstruct the pseudo-URL
 935         $prot = $bits['scheme'];
 936         $index = $prot . $bits['delimiter'] . $reversedHost;
 937         // Leave out user and password. Add the port, path, query and fragment
 938         if ( isset( $bits['port'] ) ) {
 939                 $index .= ':' . $bits['port'];
 940         }
 941         if ( isset( $bits['path'] ) ) {
 942                 $index .= $bits['path'];
 943         } else {
 944                 $index .= '/';
 945         }
 946         if ( isset( $bits['query'] ) ) {
 947                 $index .= '?' . $bits['query'];
 948         }
 949         if ( isset( $bits['fragment'] ) ) {
 950                 $index .= '#' . $bits['fragment'];
 951         }
 952
 953         if ( $prot == '' ) {
 954                 return [ "http:$index", "https:$index" ];
 955         } else {
 956                 return [ $index ];
 957         }
 958 }
 959
 960 /**
 961  * Check whether a given URL has a domain that occurs in a given set of domains
 962  * @param string $url URL
 963  * @param array $domains Array of domains (strings)
 964  * @return bool True if the host part of $url ends in one of the strings in $domains
 965  */
 966 function wfMatchesDomainList( $url, $domains ) {
 967         $bits = wfParseUrl( $url );
 968         if ( is_array( $bits ) && isset( $bits['host'] ) ) {
 969                 $host = '.' . $bits['host'];
 970                 foreach ( (array)$domains as $domain ) {
 971                         $domain = '.' . $domain;
 972                         if ( substr( $host, -strlen( $domain ) ) === $domain ) {
 973                                 return true;
 974                         }
 975                 }
 976         }
 977         return false;
 978 }
 979
 980 /**
 981  * Sends a line to the debug log if enabled or, optionally, to a comment in output.
 982  * In normal operation this is a NOP.
 983  *
 984  * Controlling globals:
 985  * $wgDebugLogFile - points to the log file
 986  * $wgDebugRawPage - if false, 'action=raw' hits will not result in debug output.
 987  * $wgDebugComments - if on, some debug items may appear in comments in the HTML output.
 988  *
 989  * @since 1.25 support for additional context data
 990  *
 991  * @param string $text
 992  * @param string|bool $dest Destination of the message:
 993  *     - 'all': both to the log and HTML (debug toolbar or HTML comments)
 994  *     - 'private': excluded from HTML output
 995  *   For backward compatibility, it can also take a boolean:
 996  *     - true: same as 'all'
 997  *     - false: same as 'private'
 998  * @param array $context Additional logging context data
 999  */
1000 function wfDebug( $text, $dest = 'all', array $context = [] ) {
1001         global $wgDebugRawPage, $wgDebugLogPrefix;
1002         global $wgDebugTimestamps, $wgRequestTime;
1003
1004         if ( !$wgDebugRawPage && wfIsDebugRawPage() ) {
1005                 return;
1006         }
1007
1008         $text = trim( $text );
1009
1010         if ( $wgDebugTimestamps ) {
1011                 $context['seconds_elapsed'] = sprintf(
1012                         '%6.4f',
1013                         microtime( true ) - $wgRequestTime
1014                 );
1015                 $context['memory_used'] = sprintf(
1016                         '%5.1fM',
1017                         ( memory_get_usage( true ) / ( 1024 * 1024 ) )
1018                 );
1019         }
1020
1021         if ( $wgDebugLogPrefix !== '' ) {
1022                 $context['prefix'] = $wgDebugLogPrefix;
1023         }
1024         $context['private'] = ( $dest === false || $dest === 'private' );
1025
1026         $logger = LoggerFactory::getInstance( 'wfDebug' );
1027         $logger->debug( $text, $context );
1028 }
1029
1030 /**
1031  * Returns true if debug logging should be suppressed if $wgDebugRawPage = false
1032  * @return bool
1033  */
1034 function wfIsDebugRawPage() {
1035         static $cache;
1036         if ( $cache !== null ) {
1037                 return $cache;
1038         }
1039         # Check for raw action using $_GET not $wgRequest, since the latter might not be initialised yet
1040         if ( ( isset( $_GET['action'] ) && $_GET['action'] == 'raw' )
1041                 || (
1042                         isset( $_SERVER['SCRIPT_NAME'] )
1043                         && substr( $_SERVER['SCRIPT_NAME'], -8 ) == 'load.php'
1044                 )
1045         ) {
1046                 $cache = true;
1047         } else {
1048                 $cache = false;
1049         }
1050         return $cache;
1051 }
1052
1053 /**
1054  * Send a line giving PHP memory usage.
1055  *
1056  * @param bool $exact Print exact byte values instead of kibibytes (default: false)
1057  */
1058 function wfDebugMem( $exact = false ) {
1059         $mem = memory_get_usage();
1060         if ( !$exact ) {
1061                 $mem = floor( $mem / 1024 ) . ' KiB';
1062         } else {
1063                 $mem .= ' B';
1064         }
1065         wfDebug( "Memory usage: $mem\n" );
1066 }
1067
1068 /**
1069  * Send a line to a supplementary debug log file, if configured, or main debug
1070  * log if not.
1071  *
1072  * To configure a supplementary log file, set $wgDebugLogGroups[$logGroup] to
1073  * a string filename or an associative array mapping 'destination' to the
1074  * desired filename. The associative array may also contain a 'sample' key
1075  * with an integer value, specifying a sampling factor. Sampled log events
1076  * will be emitted with a 1 in N random chance.
1077  *
1078  * @since 1.23 support for sampling log messages via $wgDebugLogGroups.
1079  * @since 1.25 support for additional context data
1080  * @since 1.25 sample behavior dependent on configured $wgMWLoggerDefaultSpi
1081  *
1082  * @param string $logGroup
1083  * @param string $text
1084  * @param string|bool $dest Destination of the message:
1085  *     - 'all': both to the log and HTML (debug toolbar or HTML comments)
1086  *     - 'private': only to the specific log if set in $wgDebugLogGroups and
1087  *       discarded otherwise
1088  *   For backward compatibility, it can also take a boolean:
1089  *     - true: same as 'all'
1090  *     - false: same as 'private'
1091  * @param array $context Additional logging context data
1092  */
1093 function wfDebugLog(
1094         $logGroup, $text, $dest = 'all', array $context = []
1095 ) {
1096         $text = trim( $text );
1097
1098         $logger = LoggerFactory::getInstance( $logGroup );
1099         $context['private'] = ( $dest === false || $dest === 'private' );
1100         $logger->info( $text, $context );
1101 }
1102
1103 /**
1104  * Log for database errors
1105  *
1106  * @since 1.25 support for additional context data
1107  *
1108  * @param string $text Database error message.
1109  * @param array $context Additional logging context data
1110  */
1111 function wfLogDBError( $text, array $context = [] ) {
1112         $logger = LoggerFactory::getInstance( 'wfLogDBError' );
1113         $logger->error( trim( $text ), $context );
1114 }
1115
1116 /**
1117  * Throws a warning that $function is deprecated
1118  *
1119  * @param string $function
1120  * @param string|bool $version Version of MediaWiki that the function
1121  *    was deprecated in (Added in 1.19).
1122  * @param string|bool $component Added in 1.19.
1123  * @param int $callerOffset How far up the call stack is the original
1124  *    caller. 2 = function that called the function that called
1125  *    wfDeprecated (Added in 1.20)
1126  *
1127  * @return null
1128  */
1129 function wfDeprecated( $function, $version = false, $component = false, $callerOffset = 2 ) {
1130         MWDebug::deprecated( $function, $version, $component, $callerOffset + 1 );
1131 }
1132
1133 /**
1134  * Send a warning either to the debug log or in a PHP error depending on
1135  * $wgDevelopmentWarnings. To log warnings in production, use wfLogWarning() instead.
1136  *
1137  * @param string $msg Message to send
1138  * @param int $callerOffset Number of items to go back in the backtrace to
1139  *        find the correct caller (1 = function calling wfWarn, ...)
1140  * @param int $level PHP error level; defaults to E_USER_NOTICE;
1141  *        only used when $wgDevelopmentWarnings is true
1142  */
1143 function wfWarn( $msg, $callerOffset = 1, $level = E_USER_NOTICE ) {
1144         MWDebug::warning( $msg, $callerOffset + 1, $level, 'auto' );
1145 }
1146
1147 /**
1148  * Send a warning as a PHP error and the debug log. This is intended for logging
1149  * warnings in production. For logging development warnings, use WfWarn instead.
1150  *
1151  * @param string $msg Message to send
1152  * @param int $callerOffset Number of items to go back in the backtrace to
1153  *        find the correct caller (1 = function calling wfLogWarning, ...)
1154  * @param int $level PHP error level; defaults to E_USER_WARNING
1155  */
1156 function wfLogWarning( $msg, $callerOffset = 1, $level = E_USER_WARNING ) {
1157         MWDebug::warning( $msg, $callerOffset + 1, $level, 'production' );
1158 }
1159
1160 /**
1161  * Log to a file without getting "file size exceeded" signals.
1162  *
1163  * Can also log to TCP or UDP with the syntax udp://host:port/prefix. This will
1164  * send lines to the specified port, prefixed by the specified prefix and a space.
1165  * @since 1.25 support for additional context data
1166  *
1167  * @param string $text
1168  * @param string $file Filename
1169  * @param array $context Additional logging context data
1170  * @throws MWException
1171  * @deprecated since 1.25 Use \MediaWiki\Logger\LegacyLogger::emit or UDPTransport
1172  */
1173 function wfErrorLog( $text, $file, array $context = [] ) {
1174         wfDeprecated( __METHOD__, '1.25' );
1175         $logger = LoggerFactory::getInstance( 'wfErrorLog' );
1176         $context['destination'] = $file;
1177         $logger->info( trim( $text ), $context );
1178 }
1179
1180 /**
1181  * @todo document
1182  */
1183 function wfLogProfilingData() {
1184         global $wgDebugLogGroups, $wgDebugRawPage;
1185
1186         $context = RequestContext::getMain();
1187         $request = $context->getRequest();
1188
1189         $profiler = Profiler::instance();
1190         $profiler->setContext( $context );
1191         $profiler->logData();
1192
1193         $config = $context->getConfig();
1194         if ( $config->get( 'StatsdServer' ) ) {
1195                 try {
1196                         $statsdServer = explode( ':', $config->get( 'StatsdServer' ) );
1197                         $statsdHost = $statsdServer[0];
1198                         $statsdPort = isset( $statsdServer[1] ) ? $statsdServer[1] : 8125;
1199                         $statsdSender = new SocketSender( $statsdHost, $statsdPort );
1200                         $statsdClient = new SamplingStatsdClient( $statsdSender, true, false );
1201                         $statsdClient->setSamplingRates( $config->get( 'StatsdSamplingRates' ) );
1202                         $statsdClient->send(
1203                                 MediaWikiServices::getInstance()->getStatsdDataFactory()->getBuffer()
1204                         );
1205                 } catch ( Exception $ex ) {
1206                         MWExceptionHandler::logException( $ex );
1207                 }
1208         }
1209
1210         # Profiling must actually be enabled...
1211         if ( $profiler instanceof ProfilerStub ) {
1212                 return;
1213         }
1214
1215         if ( isset( $wgDebugLogGroups['profileoutput'] )
1216                 && $wgDebugLogGroups['profileoutput'] === false
1217         ) {
1218                 // Explicitly disabled
1219                 return;
1220         }
1221         if ( !$wgDebugRawPage && wfIsDebugRawPage() ) {
1222                 return;
1223         }
1224
1225         $ctx = [ 'elapsed' => $request->getElapsedTime() ];
1226         if ( !empty( $_SERVER['HTTP_X_FORWARDED_FOR'] ) ) {
1227                 $ctx['forwarded_for'] = $_SERVER['HTTP_X_FORWARDED_FOR'];
1228         }
1229         if ( !empty( $_SERVER['HTTP_CLIENT_IP'] ) ) {
1230                 $ctx['client_ip'] = $_SERVER['HTTP_CLIENT_IP'];
1231         }
1232         if ( !empty( $_SERVER['HTTP_FROM'] ) ) {
1233                 $ctx['from'] = $_SERVER['HTTP_FROM'];
1234         }
1235         if ( isset( $ctx['forwarded_for'] ) ||
1236                 isset( $ctx['client_ip'] ) ||
1237                 isset( $ctx['from'] ) ) {
1238                 $ctx['proxy'] = $_SERVER['REMOTE_ADDR'];
1239         }
1240
1241         // Don't load $wgUser at this late stage just for statistics purposes
1242         // @todo FIXME: We can detect some anons even if it is not loaded.
1243         // See User::getId()
1244         $user = $context->getUser();
1245         $ctx['anon'] = $user->isItemLoaded( 'id' ) && $user->isAnon();
1246
1247         // Command line script uses a FauxRequest object which does not have
1248         // any knowledge about an URL and throw an exception instead.
1249         try {
1250                 $ctx['url'] = urldecode( $request->getRequestURL() );
1251         } catch ( Exception $ignored ) {
1252                 // no-op
1253         }
1254
1255         $ctx['output'] = $profiler->getOutput();
1256
1257         $log = LoggerFactory::getInstance( 'profileoutput' );
1258         $log->info( "Elapsed: {elapsed}; URL: <{url}>\n{output}", $ctx );
1259 }
1260
1261 /**
1262  * Increment a statistics counter
1263  *
1264  * @param string $key
1265  * @param int $count
1266  * @return void
1267  */
1268 function wfIncrStats( $key, $count = 1 ) {
1269         $stats = MediaWikiServices::getInstance()->getStatsdDataFactory();
1270         $stats->updateCount( $key, $count );
1271 }
1272
1273 /**
1274  * Check whether the wiki is in read-only mode.
1275  *
1276  * @return bool
1277  */
1278 function wfReadOnly() {
1279         return MediaWikiServices::getInstance()->getReadOnlyMode()
1280                 ->isReadOnly();
1281 }
1282
1283 /**
1284  * Check if the site is in read-only mode and return the message if so
1285  *
1286  * This checks wfConfiguredReadOnlyReason() and the main load balancer
1287  * for replica DB lag. This may result in DB connection being made.
1288  *
1289  * @return string|bool String when in read-only mode; false otherwise
1290  */
1291 function wfReadOnlyReason() {
1292         return MediaWikiServices::getInstance()->getReadOnlyMode()
1293                 ->getReason();
1294 }
1295
1296 /**
1297  * Get the value of $wgReadOnly or the contents of $wgReadOnlyFile.
1298  *
1299  * @return string|bool String when in read-only mode; false otherwise
1300  * @since 1.27
1301  */
1302 function wfConfiguredReadOnlyReason() {
1303         return MediaWikiServices::getInstance()->getConfiguredReadOnlyMode()
1304                 ->getReason();
1305 }
1306
1307 /**
1308  * Return a Language object from $langcode
1309  *
1310  * @param Language|string|bool $langcode Either:
1311  *                  - a Language object
1312  *                  - code of the language to get the message for, if it is
1313  *                    a valid code create a language for that language, if
1314  *                    it is a string but not a valid code then make a basic
1315  *                    language object
1316  *                  - a boolean: if it's false then use the global object for
1317  *                    the current user's language (as a fallback for the old parameter
1318  *                    functionality), or if it is true then use global object
1319  *                    for the wiki's content language.
1320  * @return Language
1321  */
1322 function wfGetLangObj( $langcode = false ) {
1323         # Identify which language to get or create a language object for.
1324         # Using is_object here due to Stub objects.
1325         if ( is_object( $langcode ) ) {
1326                 # Great, we already have the object (hopefully)!
1327                 return $langcode;
1328         }
1329
1330         global $wgContLang, $wgLanguageCode;
1331         if ( $langcode === true || $langcode === $wgLanguageCode ) {
1332                 # $langcode is the language code of the wikis content language object.
1333                 # or it is a boolean and value is true
1334                 return $wgContLang;
1335         }
1336
1337         global $wgLang;
1338         if ( $langcode === false || $langcode === $wgLang->getCode() ) {
1339                 # $langcode is the language code of user language object.
1340                 # or it was a boolean and value is false
1341                 return $wgLang;
1342         }
1343
1344         $validCodes = array_keys( Language::fetchLanguageNames() );
1345         if ( in_array( $langcode, $validCodes ) ) {
1346                 # $langcode corresponds to a valid language.
1347                 return Language::factory( $langcode );
1348         }
1349
1350         # $langcode is a string, but not a valid language code; use content language.
1351         wfDebug( "Invalid language code passed to wfGetLangObj, falling back to content language.\n" );
1352         return $wgContLang;
1353 }
1354
1355 /**
1356  * This is the function for getting translated interface messages.
1357  *
1358  * @see Message class for documentation how to use them.
1359  * @see https://www.mediawiki.org/wiki/Manual:Messages_API
1360  *
1361  * This function replaces all old wfMsg* functions.
1362  *
1363  * @param string|string[]|MessageSpecifier $key Message key, or array of keys, or a MessageSpecifier
1364  * @param mixed $params,... Normal message parameters
1365  * @return Message
1366  *
1367  * @since 1.17
1368  *
1369  * @see Message::__construct
1370  */
1371 function wfMessage( $key /*...*/ ) {
1372         $message = new Message( $key );
1373
1374         // We call Message::params() to reduce code duplication
1375         $params = func_get_args();
1376         array_shift( $params );
1377         if ( $params ) {
1378                 call_user_func_array( [ $message, 'params' ], $params );
1379         }
1380
1381         return $message;
1382 }
1383
1384 /**
1385  * This function accepts multiple message keys and returns a message instance
1386  * for the first message which is non-empty. If all messages are empty then an
1387  * instance of the first message key is returned.
1388  *
1389  * @param string|string[] $keys,... Message keys
1390  * @return Message
1391  *
1392  * @since 1.18
1393  *
1394  * @see Message::newFallbackSequence
1395  */
1396 function wfMessageFallback( /*...*/ ) {
1397         $args = func_get_args();
1398         return call_user_func_array( 'Message::newFallbackSequence', $args );
1399 }
1400
1401 /**
1402  * Replace message parameter keys on the given formatted output.
1403  *
1404  * @param string $message
1405  * @param array $args
1406  * @return string
1407  * @private
1408  */
1409 function wfMsgReplaceArgs( $message, $args ) {
1410         # Fix windows line-endings
1411         # Some messages are split with explode("\n", $msg)
1412         $message = str_replace( "\r", '', $message );
1413
1414         // Replace arguments
1415         if ( is_array( $args ) && $args ) {
1416                 if ( is_array( $args[0] ) ) {
1417                         $args = array_values( $args[0] );
1418                 }
1419                 $replacementKeys = [];
1420                 foreach ( $args as $n => $param ) {
1421                         $replacementKeys['$' . ( $n + 1 )] = $param;
1422                 }
1423                 $message = strtr( $message, $replacementKeys );
1424         }
1425
1426         return $message;
1427 }
1428
1429 /**
1430  * Fetch server name for use in error reporting etc.
1431  * Use real server name if available, so we know which machine
1432  * in a server farm generated the current page.
1433  *
1434  * @return string
1435  */
1436 function wfHostname() {
1437         static $host;
1438         if ( is_null( $host ) ) {
1439                 # Hostname overriding
1440                 global $wgOverrideHostname;
1441                 if ( $wgOverrideHostname !== false ) {
1442                         # Set static and skip any detection
1443                         $host = $wgOverrideHostname;
1444                         return $host;
1445                 }
1446
1447                 if ( function_exists( 'posix_uname' ) ) {
1448                         // This function not present on Windows
1449                         $uname = posix_uname();
1450                 } else {
1451                         $uname = false;
1452                 }
1453                 if ( is_array( $uname ) && isset( $uname['nodename'] ) ) {
1454                         $host = $uname['nodename'];
1455                 } elseif ( getenv( 'COMPUTERNAME' ) ) {
1456                         # Windows computer name
1457                         $host = getenv( 'COMPUTERNAME' );
1458                 } else {
1459                         # This may be a virtual server.
1460                         $host = $_SERVER['SERVER_NAME'];
1461                 }
1462         }
1463         return $host;
1464 }
1465
1466 /**
1467  * Returns a script tag that stores the amount of time it took MediaWiki to
1468  * handle the request in milliseconds as 'wgBackendResponseTime'.
1469  *
1470  * If $wgShowHostnames is true, the script will also set 'wgHostname' to the
1471  * hostname of the server handling the request.
1472  *
1473  * @return string
1474  */
1475 function wfReportTime() {
1476         global $wgRequestTime, $wgShowHostnames;
1477
1478         $responseTime = round( ( microtime( true ) - $wgRequestTime ) * 1000 );
1479         $reportVars = [ 'wgBackendResponseTime' => $responseTime ];
1480         if ( $wgShowHostnames ) {
1481                 $reportVars['wgHostname'] = wfHostname();
1482         }
1483         return Skin::makeVariablesScript( $reportVars );
1484 }
1485
1486 /**
1487  * Safety wrapper for debug_backtrace().
1488  *
1489  * Will return an empty array if debug_backtrace is disabled, otherwise
1490  * the output from debug_backtrace() (trimmed).
1491  *
1492  * @param int $limit This parameter can be used to limit the number of stack frames returned
1493  *
1494  * @return array Array of backtrace information
1495  */
1496 function wfDebugBacktrace( $limit = 0 ) {
1497         static $disabled = null;
1498
1499         if ( is_null( $disabled ) ) {
1500                 $disabled = !function_exists( 'debug_backtrace' );
1501                 if ( $disabled ) {
1502                         wfDebug( "debug_backtrace() is disabled\n" );
1503                 }
1504         }
1505         if ( $disabled ) {
1506                 return [];
1507         }
1508
1509         if ( $limit ) {
1510                 return array_slice( debug_backtrace( DEBUG_BACKTRACE_PROVIDE_OBJECT, $limit + 1 ), 1 );
1511         } else {
1512                 return array_slice( debug_backtrace(), 1 );
1513         }
1514 }
1515
1516 /**
1517  * Get a debug backtrace as a string
1518  *
1519  * @param bool|null $raw If true, the return value is plain text. If false, HTML.
1520  *   Defaults to $wgCommandLineMode if unset.
1521  * @return string
1522  * @since 1.25 Supports $raw parameter.
1523  */
1524 function wfBacktrace( $raw = null ) {
1525         global $wgCommandLineMode;
1526
1527         if ( $raw === null ) {
1528                 $raw = $wgCommandLineMode;
1529         }
1530
1531         if ( $raw ) {
1532                 $frameFormat = "%s line %s calls %s()\n";
1533                 $traceFormat = "%s";
1534         } else {
1535                 $frameFormat = "<li>%s line %s calls %s()</li>\n";
1536                 $traceFormat = "<ul>\n%s</ul>\n";
1537         }
1538
1539         $frames = array_map( function ( $frame ) use ( $frameFormat ) {
1540                 $file = !empty( $frame['file'] ) ? basename( $frame['file'] ) : '-';
1541                 $line = isset( $frame['line'] ) ? $frame['line'] : '-';
1542                 $call = $frame['function'];
1543                 if ( !empty( $frame['class'] ) ) {
1544                         $call = $frame['class'] . $frame['type'] . $call;
1545                 }
1546                 return sprintf( $frameFormat, $file, $line, $call );
1547         }, wfDebugBacktrace() );
1548
1549         return sprintf( $traceFormat, implode( '', $frames ) );
1550 }
1551
1552 /**
1553  * Get the name of the function which called this function
1554  * wfGetCaller( 1 ) is the function with the wfGetCaller() call (ie. __FUNCTION__)
1555  * wfGetCaller( 2 ) [default] is the caller of the function running wfGetCaller()
1556  * wfGetCaller( 3 ) is the parent of that.
1557  *
1558  * @param int $level
1559  * @return string
1560  */
1561 function wfGetCaller( $level = 2 ) {
1562         $backtrace = wfDebugBacktrace( $level + 1 );
1563         if ( isset( $backtrace[$level] ) ) {
1564                 return wfFormatStackFrame( $backtrace[$level] );
1565         } else {
1566                 return 'unknown';
1567         }
1568 }
1569
1570 /**
1571  * Return a string consisting of callers in the stack. Useful sometimes
1572  * for profiling specific points.
1573  *
1574  * @param int $limit The maximum depth of the stack frame to return, or false for the entire stack.
1575  * @return string
1576  */
1577 function wfGetAllCallers( $limit = 3 ) {
1578         $trace = array_reverse( wfDebugBacktrace() );
1579         if ( !$limit || $limit > count( $trace ) - 1 ) {
1580                 $limit = count( $trace ) - 1;
1581         }
1582         $trace = array_slice( $trace, -$limit - 1, $limit );
1583         return implode( '/', array_map( 'wfFormatStackFrame', $trace ) );
1584 }
1585
1586 /**
1587  * Return a string representation of frame
1588  *
1589  * @param array $frame
1590  * @return string
1591  */
1592 function wfFormatStackFrame( $frame ) {
1593         if ( !isset( $frame['function'] ) ) {
1594                 return 'NO_FUNCTION_GIVEN';
1595         }
1596         return isset( $frame['class'] ) && isset( $frame['type'] ) ?
1597                 $frame['class'] . $frame['type'] . $frame['function'] :
1598                 $frame['function'];
1599 }
1600
1601 /* Some generic result counters, pulled out of SearchEngine */
1602
1603 /**
1604  * @todo document
1605  *
1606  * @param int $offset
1607  * @param int $limit
1608  * @return string
1609  */
1610 function wfShowingResults( $offset, $limit ) {
1611         return wfMessage( 'showingresults' )->numParams( $limit, $offset + 1 )->parse();
1612 }
1613
1614 /**
1615  * Whether the client accept gzip encoding
1616  *
1617  * Uses the Accept-Encoding header to check if the client supports gzip encoding.
1618  * Use this when considering to send a gzip-encoded response to the client.
1619  *
1620  * @param bool $force Forces another check even if we already have a cached result.
1621  * @return bool
1622  */
1623 function wfClientAcceptsGzip( $force = false ) {
1624         static $result = null;
1625         if ( $result === null || $force ) {
1626                 $result = false;
1627                 if ( isset( $_SERVER['HTTP_ACCEPT_ENCODING'] ) ) {
1628                         # @todo FIXME: We may want to blacklist some broken browsers
1629                         $m = [];
1630                         if ( preg_match(
1631                                         '/\bgzip(?:;(q)=([0-9]+(?:\.[0-9]+)))?\b/',
1632                                         $_SERVER['HTTP_ACCEPT_ENCODING'],
1633                                         $m
1634                                 )
1635                         ) {
1636                                 if ( isset( $m[2] ) && ( $m[1] == 'q' ) && ( $m[2] == 0 ) ) {
1637                                         $result = false;
1638                                         return $result;
1639                                 }
1640                                 wfDebug( "wfClientAcceptsGzip: client accepts gzip.\n" );
1641                                 $result = true;
1642                         }
1643                 }
1644         }
1645         return $result;
1646 }
1647
1648 /**
1649  * Escapes the given text so that it may be output using addWikiText()
1650  * without any linking, formatting, etc. making its way through. This
1651  * is achieved by substituting certain characters with HTML entities.
1652  * As required by the callers, "<nowiki>" is not used.
1653  *
1654  * @param string $text Text to be escaped
1655  * @return string
1656  */
1657 function wfEscapeWikiText( $text ) {
1658         global $wgEnableMagicLinks;
1659         static $repl = null, $repl2 = null;
1660         if ( $repl === null || defined( 'MW_PARSER_TEST' ) || defined( 'MW_PHPUNIT_TEST' ) ) {
1661                 // Tests depend upon being able to change $wgEnableMagicLinks, so don't cache
1662                 // in those situations
1663                 $repl = [
1664                         '"' => '&#34;', '&' => '&#38;', "'" => '&#39;', '<' => '&#60;',
1665                         '=' => '&#61;', '>' => '&#62;', '[' => '&#91;', ']' => '&#93;',
1666                         '{' => '&#123;', '|' => '&#124;', '}' => '&#125;', ';' => '&#59;',
1667                         "\n#" => "\n&#35;", "\r#" => "\r&#35;",
1668                         "\n*" => "\n&#42;", "\r*" => "\r&#42;",
1669                         "\n:" => "\n&#58;", "\r:" => "\r&#58;",
1670                         "\n " => "\n&#32;", "\r " => "\r&#32;",
1671                         "\n\n" => "\n&#10;", "\r\n" => "&#13;\n",
1672                         "\n\r" => "\n&#13;", "\r\r" => "\r&#13;",
1673                         "\n\t" => "\n&#9;", "\r\t" => "\r&#9;", // "\n\t\n" is treated like "\n\n"
1674                         "\n----" => "\n&#45;---", "\r----" => "\r&#45;---",
1675                         '__' => '_&#95;', '://' => '&#58;//',
1676                 ];
1677
1678                 $magicLinks = array_keys( array_filter( $wgEnableMagicLinks ) );
1679                 // We have to catch everything "\s" matches in PCRE
1680                 foreach ( $magicLinks as $magic ) {
1681                         $repl["$magic "] = "$magic&#32;";
1682                         $repl["$magic\t"] = "$magic&#9;";
1683                         $repl["$magic\r"] = "$magic&#13;";
1684                         $repl["$magic\n"] = "$magic&#10;";
1685                         $repl["$magic\f"] = "$magic&#12;";
1686                 }
1687
1688                 // And handle protocols that don't use "://"
1689                 global $wgUrlProtocols;
1690                 $repl2 = [];
1691                 foreach ( $wgUrlProtocols as $prot ) {
1692                         if ( substr( $prot, -1 ) === ':' ) {
1693                                 $repl2[] = preg_quote( substr( $prot, 0, -1 ), '/' );
1694                         }
1695                 }
1696                 $repl2 = $repl2 ? '/\b(' . implode( '|', $repl2 ) . '):/i' : '/^(?!)/';
1697         }
1698         $text = substr( strtr( "\n$text", $repl ), 1 );
1699         $text = preg_replace( $repl2, '$1&#58;', $text );
1700         return $text;
1701 }
1702
1703 /**
1704  * Sets dest to source and returns the original value of dest
1705  * If source is NULL, it just returns the value, it doesn't set the variable
1706  * If force is true, it will set the value even if source is NULL
1707  *
1708  * @param mixed $dest
1709  * @param mixed $source
1710  * @param bool $force
1711  * @return mixed
1712  */
1713 function wfSetVar( &$dest, $source, $force = false ) {
1714         $temp = $dest;
1715         if ( !is_null( $source ) || $force ) {
1716                 $dest = $source;
1717         }
1718         return $temp;
1719 }
1720
1721 /**
1722  * As for wfSetVar except setting a bit
1723  *
1724  * @param int $dest
1725  * @param int $bit
1726  * @param bool $state
1727  *
1728  * @return bool
1729  */
1730 function wfSetBit( &$dest, $bit, $state = true ) {
1731         $temp = (bool)( $dest & $bit );
1732         if ( !is_null( $state ) ) {
1733                 if ( $state ) {
1734                         $dest |= $bit;
1735                 } else {
1736                         $dest &= ~$bit;
1737                 }
1738         }
1739         return $temp;
1740 }
1741
1742 /**
1743  * A wrapper around the PHP function var_export().
1744  * Either print it or add it to the regular output ($wgOut).
1745  *
1746  * @param mixed $var A PHP variable to dump.
1747  */
1748 function wfVarDump( $var ) {
1749         global $wgOut;
1750         $s = str_replace( "\n", "<br />\n", var_export( $var, true ) . "\n" );
1751         if ( headers_sent() || !isset( $wgOut ) || !is_object( $wgOut ) ) {
1752                 print $s;
1753         } else {
1754                 $wgOut->addHTML( $s );
1755         }
1756 }
1757
1758 /**
1759  * Provide a simple HTTP error.
1760  *
1761  * @param int|string $code
1762  * @param string $label
1763  * @param string $desc
1764  */
1765 function wfHttpError( $code, $label, $desc ) {
1766         global $wgOut;
1767         HttpStatus::header( $code );
1768         if ( $wgOut ) {
1769                 $wgOut->disable();
1770                 $wgOut->sendCacheControl();
1771         }
1772
1773         MediaWiki\HeaderCallback::warnIfHeadersSent();
1774         header( 'Content-type: text/html; charset=utf-8' );
1775         print '<!DOCTYPE html>' .
1776                 '<html><head><title>' .
1777                 htmlspecialchars( $label ) .
1778                 '</title></head><body><h1>' .
1779                 htmlspecialchars( $label ) .
1780                 '</h1><p>' .
1781                 nl2br( htmlspecialchars( $desc ) ) .
1782                 "</p></body></html>\n";
1783 }
1784
1785 /**
1786  * Clear away any user-level output buffers, discarding contents.
1787  *
1788  * Suitable for 'starting afresh', for instance when streaming
1789  * relatively large amounts of data without buffering, or wanting to
1790  * output image files without ob_gzhandler's compression.
1791  *
1792  * The optional $resetGzipEncoding parameter controls suppression of
1793  * the Content-Encoding header sent by ob_gzhandler; by default it
1794  * is left. See comments for wfClearOutputBuffers() for why it would
1795  * be used.
1796  *
1797  * Note that some PHP configuration options may add output buffer
1798  * layers which cannot be removed; these are left in place.
1799  *
1800  * @param bool $resetGzipEncoding
1801  */
1802 function wfResetOutputBuffers( $resetGzipEncoding = true ) {
1803         if ( $resetGzipEncoding ) {
1804                 // Suppress Content-Encoding and Content-Length
1805                 // headers from 1.10+s wfOutputHandler
1806                 global $wgDisableOutputCompression;
1807                 $wgDisableOutputCompression = true;
1808         }
1809         while ( $status = ob_get_status() ) {
1810                 if ( isset( $status['flags'] ) ) {
1811                         $flags = PHP_OUTPUT_HANDLER_CLEANABLE | PHP_OUTPUT_HANDLER_REMOVABLE;
1812                         $deleteable = ( $status['flags'] & $flags ) === $flags;
1813                 } elseif ( isset( $status['del'] ) ) {
1814                         $deleteable = $status['del'];
1815                 } else {
1816                         // Guess that any PHP-internal setting can't be removed.
1817                         $deleteable = $status['type'] !== 0; /* PHP_OUTPUT_HANDLER_INTERNAL */
1818                 }
1819                 if ( !$deleteable ) {
1820                         // Give up, and hope the result doesn't break
1821                         // output behavior.
1822                         break;
1823                 }
1824                 if ( $status['name'] === 'MediaWikiTestCase::wfResetOutputBuffersBarrier' ) {
1825                         // Unit testing barrier to prevent this function from breaking PHPUnit.
1826                         break;
1827                 }
1828                 if ( !ob_end_clean() ) {
1829                         // Could not remove output buffer handler; abort now
1830                         // to avoid getting in some kind of infinite loop.
1831                         break;
1832                 }
1833                 if ( $resetGzipEncoding ) {
1834                         if ( $status['name'] == 'ob_gzhandler' ) {
1835                                 // Reset the 'Content-Encoding' field set by this handler
1836                                 // so we can start fresh.
1837                                 header_remove( 'Content-Encoding' );
1838                                 break;
1839                         }
1840                 }
1841         }
1842 }
1843
1844 /**
1845  * More legible than passing a 'false' parameter to wfResetOutputBuffers():
1846  *
1847  * Clear away output buffers, but keep the Content-Encoding header
1848  * produced by ob_gzhandler, if any.
1849  *
1850  * This should be used for HTTP 304 responses, where you need to
1851  * preserve the Content-Encoding header of the real result, but
1852  * also need to suppress the output of ob_gzhandler to keep to spec
1853  * and avoid breaking Firefox in rare cases where the headers and
1854  * body are broken over two packets.
1855  */
1856 function wfClearOutputBuffers() {
1857         wfResetOutputBuffers( false );
1858 }
1859
1860 /**
1861  * Converts an Accept-* header into an array mapping string values to quality
1862  * factors
1863  *
1864  * @param string $accept
1865  * @param string $def Default
1866  * @return float[] Associative array of string => float pairs
1867  */
1868 function wfAcceptToPrefs( $accept, $def = '*/*' ) {
1869         # No arg means accept anything (per HTTP spec)
1870         if ( !$accept ) {
1871                 return [ $def => 1.0 ];
1872         }
1873
1874         $prefs = [];
1875
1876         $parts = explode( ',', $accept );
1877
1878         foreach ( $parts as $part ) {
1879                 # @todo FIXME: Doesn't deal with params like 'text/html; level=1'
1880                 $values = explode( ';', trim( $part ) );
1881                 $match = [];
1882                 if ( count( $values ) == 1 ) {
1883                         $prefs[$values[0]] = 1.0;
1884                 } elseif ( preg_match( '/q\s*=\s*(\d*\.\d+)/', $values[1], $match ) ) {
1885                         $prefs[$values[0]] = floatval( $match[1] );
1886                 }
1887         }
1888
1889         return $prefs;
1890 }
1891
1892 /**
1893  * Checks if a given MIME type matches any of the keys in the given
1894  * array. Basic wildcards are accepted in the array keys.
1895  *
1896  * Returns the matching MIME type (or wildcard) if a match, otherwise
1897  * NULL if no match.
1898  *
1899  * @param string $type
1900  * @param array $avail
1901  * @return string
1902  * @private
1903  */
1904 function mimeTypeMatch( $type, $avail ) {
1905         if ( array_key_exists( $type, $avail ) ) {
1906                 return $type;
1907         } else {
1908                 $mainType = explode( '/', $type )[0];
1909                 if ( array_key_exists( "$mainType/*", $avail ) ) {
1910                         return "$mainType/*";
1911                 } elseif ( array_key_exists( '*/*', $avail ) ) {
1912                         return '*/*';
1913                 } else {
1914                         return null;
1915                 }
1916         }
1917 }
1918
1919 /**
1920  * Returns the 'best' match between a client's requested internet media types
1921  * and the server's list of available types. Each list should be an associative
1922  * array of type to preference (preference is a float between 0.0 and 1.0).
1923  * Wildcards in the types are acceptable.
1924  *
1925  * @param array $cprefs Client's acceptable type list
1926  * @param array $sprefs Server's offered types
1927  * @return string
1928  *
1929  * @todo FIXME: Doesn't handle params like 'text/plain; charset=UTF-8'
1930  * XXX: generalize to negotiate other stuff
1931  */
1932 function wfNegotiateType( $cprefs, $sprefs ) {
1933         $combine = [];
1934
1935         foreach ( array_keys( $sprefs ) as $type ) {
1936                 $subType = explode( '/', $type )[1];
1937                 if ( $subType != '*' ) {
1938                         $ckey = mimeTypeMatch( $type, $cprefs );
1939                         if ( $ckey ) {
1940                                 $combine[$type] = $sprefs[$type] * $cprefs[$ckey];
1941                         }
1942                 }
1943         }
1944
1945         foreach ( array_keys( $cprefs ) as $type ) {
1946                 $subType = explode( '/', $type )[1];
1947                 if ( $subType != '*' && !array_key_exists( $type, $sprefs ) ) {
1948                         $skey = mimeTypeMatch( $type, $sprefs );
1949                         if ( $skey ) {
1950                                 $combine[$type] = $sprefs[$skey] * $cprefs[$type];
1951                         }
1952                 }
1953         }
1954
1955         $bestq = 0;
1956         $besttype = null;
1957
1958         foreach ( array_keys( $combine ) as $type ) {
1959                 if ( $combine[$type] > $bestq ) {
1960                         $besttype = $type;
1961                         $bestq = $combine[$type];
1962                 }
1963         }
1964
1965         return $besttype;
1966 }
1967
1968 /**
1969  * Reference-counted warning suppression
1970  *
1971  * @deprecated since 1.26, use MediaWiki\suppressWarnings() directly
1972  * @param bool $end
1973  */
1974 function wfSuppressWarnings( $end = false ) {
1975         MediaWiki\suppressWarnings( $end );
1976 }
1977
1978 /**
1979  * @deprecated since 1.26, use MediaWiki\restoreWarnings() directly
1980  * Restore error level to previous value
1981  */
1982 function wfRestoreWarnings() {
1983         MediaWiki\suppressWarnings( true );
1984 }
1985
1986 /**
1987  * Get a timestamp string in one of various formats
1988  *
1989  * @param mixed $outputtype A timestamp in one of the supported formats, the
1990  *   function will autodetect which format is supplied and act accordingly.
1991  * @param mixed $ts Optional timestamp to convert, default 0 for the current time
1992  * @return string|bool String / false The same date in the format specified in $outputtype or false
1993  */
1994 function wfTimestamp( $outputtype = TS_UNIX, $ts = 0 ) {
1995         $ret = MWTimestamp::convert( $outputtype, $ts );
1996         if ( $ret === false ) {
1997                 wfDebug( "wfTimestamp() fed bogus time value: TYPE=$outputtype; VALUE=$ts\n" );
1998         }
1999         return $ret;
2000 }
2001
2002 /**
2003  * Return a formatted timestamp, or null if input is null.
2004  * For dealing with nullable timestamp columns in the database.
2005  *
2006  * @param int $outputtype
2007  * @param string $ts
2008  * @return string
2009  */
2010 function wfTimestampOrNull( $outputtype = TS_UNIX, $ts = null ) {
2011         if ( is_null( $ts ) ) {
2012                 return null;
2013         } else {
2014                 return wfTimestamp( $outputtype, $ts );
2015         }
2016 }
2017
2018 /**
2019  * Convenience function; returns MediaWiki timestamp for the present time.
2020  *
2021  * @return string
2022  */
2023 function wfTimestampNow() {
2024         # return NOW
2025         return MWTimestamp::now( TS_MW );
2026 }
2027
2028 /**
2029  * Check if the operating system is Windows
2030  *
2031  * @return bool True if it's Windows, false otherwise.
2032  */
2033 function wfIsWindows() {
2034         static $isWindows = null;
2035         if ( $isWindows === null ) {
2036                 $isWindows = strtoupper( substr( PHP_OS, 0, 3 ) ) === 'WIN';
2037         }
2038         return $isWindows;
2039 }
2040
2041 /**
2042  * Check if we are running under HHVM
2043  *
2044  * @return bool
2045  */
2046 function wfIsHHVM() {
2047         return defined( 'HHVM_VERSION' );
2048 }
2049
2050 /**
2051  * Tries to get the system directory for temporary files. First
2052  * $wgTmpDirectory is checked, and then the TMPDIR, TMP, and TEMP
2053  * environment variables are then checked in sequence, then
2054  * sys_get_temp_dir(), then upload_tmp_dir from php.ini.
2055  *
2056  * NOTE: When possible, use instead the tmpfile() function to create
2057  * temporary files to avoid race conditions on file creation, etc.
2058  *
2059  * @return string
2060  */
2061 function wfTempDir() {
2062         global $wgTmpDirectory;
2063
2064         if ( $wgTmpDirectory !== false ) {
2065                 return $wgTmpDirectory;
2066         }
2067
2068         return TempFSFile::getUsableTempDirectory();
2069 }
2070
2071 /**
2072  * Make directory, and make all parent directories if they don't exist
2073  *
2074  * @param string $dir Full path to directory to create
2075  * @param int $mode Chmod value to use, default is $wgDirectoryMode
2076  * @param string $caller Optional caller param for debugging.
2077  * @throws MWException
2078  * @return bool
2079  */
2080 function wfMkdirParents( $dir, $mode = null, $caller = null ) {
2081         global $wgDirectoryMode;
2082
2083         if ( FileBackend::isStoragePath( $dir ) ) { // sanity
2084                 throw new MWException( __FUNCTION__ . " given storage path '$dir'." );
2085         }
2086
2087         if ( !is_null( $caller ) ) {
2088                 wfDebug( "$caller: called wfMkdirParents($dir)\n" );
2089         }
2090
2091         if ( strval( $dir ) === '' || is_dir( $dir ) ) {
2092                 return true;
2093         }
2094
2095         $dir = str_replace( [ '\\', '/' ], DIRECTORY_SEPARATOR, $dir );
2096
2097         if ( is_null( $mode ) ) {
2098                 $mode = $wgDirectoryMode;
2099         }
2100
2101         // Turn off the normal warning, we're doing our own below
2102         MediaWiki\suppressWarnings();
2103         $ok = mkdir( $dir, $mode, true ); // PHP5 <3
2104         MediaWiki\restoreWarnings();
2105
2106         if ( !$ok ) {
2107                 // directory may have been created on another request since we last checked
2108                 if ( is_dir( $dir ) ) {
2109                         return true;
2110                 }
2111
2112                 // PHP doesn't report the path in its warning message, so add our own to aid in diagnosis.
2113                 wfLogWarning( sprintf( "failed to mkdir \"%s\" mode 0%o", $dir, $mode ) );
2114         }
2115         return $ok;
2116 }
2117
2118 /**
2119  * Remove a directory and all its content.
2120  * Does not hide error.
2121  * @param string $dir
2122  */
2123 function wfRecursiveRemoveDir( $dir ) {
2124         wfDebug( __FUNCTION__ . "( $dir )\n" );
2125         // taken from https://secure.php.net/manual/en/function.rmdir.php#98622
2126         if ( is_dir( $dir ) ) {
2127                 $objects = scandir( $dir );
2128                 foreach ( $objects as $object ) {
2129                         if ( $object != "." && $object != ".." ) {
2130                                 if ( filetype( $dir . '/' . $object ) == "dir" ) {
2131                                         wfRecursiveRemoveDir( $dir . '/' . $object );
2132                                 } else {
2133                                         unlink( $dir . '/' . $object );
2134                                 }
2135                         }
2136                 }
2137                 reset( $objects );
2138                 rmdir( $dir );
2139         }
2140 }
2141
2142 /**
2143  * @param int $nr The number to format
2144  * @param int $acc The number of digits after the decimal point, default 2
2145  * @param bool $round Whether or not to round the value, default true
2146  * @return string
2147  */
2148 function wfPercent( $nr, $acc = 2, $round = true ) {
2149         $ret = sprintf( "%.${acc}f", $nr );
2150         return $round ? round( $ret, $acc ) . '%' : "$ret%";
2151 }
2152
2153 /**
2154  * Safety wrapper around ini_get() for boolean settings.
2155  * The values returned from ini_get() are pre-normalized for settings
2156  * set via php.ini or php_flag/php_admin_flag... but *not*
2157  * for those set via php_value/php_admin_value.
2158  *
2159  * It's fairly common for people to use php_value instead of php_flag,
2160  * which can leave you with an 'off' setting giving a false positive
2161  * for code that just takes the ini_get() return value as a boolean.
2162  *
2163  * To make things extra interesting, setting via php_value accepts
2164  * "true" and "yes" as true, but php.ini and php_flag consider them false. :)
2165  * Unrecognized values go false... again opposite PHP's own coercion
2166  * from string to bool.
2167  *
2168  * Luckily, 'properly' set settings will always come back as '0' or '1',
2169  * so we only have to worry about them and the 'improper' settings.
2170  *
2171  * I frickin' hate PHP... :P
2172  *
2173  * @param string $setting
2174  * @return bool
2175  */
2176 function wfIniGetBool( $setting ) {
2177         $val = strtolower( ini_get( $setting ) );
2178         // 'on' and 'true' can't have whitespace around them, but '1' can.
2179         return $val == 'on'
2180                 || $val == 'true'
2181                 || $val == 'yes'
2182                 || preg_match( "/^\s*[+-]?0*[1-9]/", $val ); // approx C atoi() function
2183 }
2184
2185 /**
2186  * Version of escapeshellarg() that works better on Windows.
2187  *
2188  * Originally, this fixed the incorrect use of single quotes on Windows
2189  * (https://bugs.php.net/bug.php?id=26285) and the locale problems on Linux in
2190  * PHP 5.2.6+ (bug backported to earlier distro releases of PHP).
2191  *
2192  * @param string ... strings to escape and glue together, or a single array of strings parameter
2193  * @return string
2194  */
2195 function wfEscapeShellArg( /*...*/ ) {
2196         wfInitShellLocale();
2197
2198         $args = func_get_args();
2199         if ( count( $args ) === 1 && is_array( reset( $args ) ) ) {
2200                 // If only one argument has been passed, and that argument is an array,
2201                 // treat it as a list of arguments
2202                 $args = reset( $args );
2203         }
2204
2205         $first = true;
2206         $retVal = '';
2207         foreach ( $args as $arg ) {
2208                 if ( !$first ) {
2209                         $retVal .= ' ';
2210                 } else {
2211                         $first = false;
2212                 }
2213
2214                 if ( wfIsWindows() ) {
2215                         // Escaping for an MSVC-style command line parser and CMD.EXE
2216                         // @codingStandardsIgnoreStart For long URLs
2217                         // Refs:
2218                         //  * https://web.archive.org/web/20020708081031/http://mailman.lyra.org/pipermail/scite-interest/2002-March/000436.html
2219                         //  * https://technet.microsoft.com/en-us/library/cc723564.aspx
2220                         //  * T15518
2221                         //  * CR r63214
2222                         // Double the backslashes before any double quotes. Escape the double quotes.
2223                         // @codingStandardsIgnoreEnd
2224                         $tokens = preg_split( '/(\\\\*")/', $arg, -1, PREG_SPLIT_DELIM_CAPTURE );
2225                         $arg = '';
2226                         $iteration = 0;
2227                         foreach ( $tokens as $token ) {
2228                                 if ( $iteration % 2 == 1 ) {
2229                                         // Delimiter, a double quote preceded by zero or more slashes
2230                                         $arg .= str_replace( '\\', '\\\\', substr( $token, 0, -1 ) ) . '\\"';
2231                                 } elseif ( $iteration % 4 == 2 ) {
2232                                         // ^ in $token will be outside quotes, need to be escaped
2233                                         $arg .= str_replace( '^', '^^', $token );
2234                                 } else { // $iteration % 4 == 0
2235                                         // ^ in $token will appear inside double quotes, so leave as is
2236                                         $arg .= $token;
2237                                 }
2238                                 $iteration++;
2239                         }
2240                         // Double the backslashes before the end of the string, because
2241                         // we will soon add a quote
2242                         $m = [];
2243                         if ( preg_match( '/^(.*?)(\\\\+)$/', $arg, $m ) ) {
2244                                 $arg = $m[1] . str_replace( '\\', '\\\\', $m[2] );
2245                         }
2246
2247                         // Add surrounding quotes
2248                         $retVal .= '"' . $arg . '"';
2249                 } else {
2250                         $retVal .= escapeshellarg( $arg );
2251                 }
2252         }
2253         return $retVal;
2254 }
2255
2256 /**
2257  * Check if wfShellExec() is effectively disabled via php.ini config
2258  *
2259  * @return bool|string False or 'disabled'
2260  * @since 1.22
2261  */
2262 function wfShellExecDisabled() {
2263         static $disabled = null;
2264         if ( is_null( $disabled ) ) {
2265                 if ( !function_exists( 'proc_open' ) ) {
2266                         wfDebug( "proc_open() is disabled\n" );
2267                         $disabled = 'disabled';
2268                 } else {
2269                         $disabled = false;
2270                 }
2271         }
2272         return $disabled;
2273 }
2274
2275 /**
2276  * Execute a shell command, with time and memory limits mirrored from the PHP
2277  * configuration if supported.
2278  *
2279  * @param string|string[] $cmd If string, a properly shell-escaped command line,
2280  *   or an array of unescaped arguments, in which case each value will be escaped
2281  *   Example:   [ 'convert', '-font', 'font name' ] would produce "'convert' '-font' 'font name'"
2282  * @param null|mixed &$retval Optional, will receive the program's exit code.
2283  *   (non-zero is usually failure). If there is an error from
2284  *   read, select, or proc_open(), this will be set to -1.
2285  * @param array $environ Optional environment variables which should be
2286  *   added to the executed command environment.
2287  * @param array $limits Optional array with limits(filesize, memory, time, walltime)
2288  *   this overwrites the global wgMaxShell* limits.
2289  * @param array $options Array of options:
2290  *   - duplicateStderr: Set this to true to duplicate stderr to stdout,
2291  *     including errors from limit.sh
2292  *   - profileMethod: By default this function will profile based on the calling
2293  *     method. Set this to a string for an alternative method to profile from
2294  *
2295  * @return string Collected stdout as a string
2296  */
2297 function wfShellExec( $cmd, &$retval = null, $environ = [],
2298         $limits = [], $options = []
2299 ) {
2300         global $IP, $wgMaxShellMemory, $wgMaxShellFileSize, $wgMaxShellTime,
2301                 $wgMaxShellWallClockTime, $wgShellCgroup;
2302
2303         $disabled = wfShellExecDisabled();
2304         if ( $disabled ) {
2305                 $retval = 1;
2306                 return 'Unable to run external programs, proc_open() is disabled.';
2307         }
2308
2309         $includeStderr = isset( $options['duplicateStderr'] ) && $options['duplicateStderr'];
2310         $profileMethod = isset( $options['profileMethod'] ) ? $options['profileMethod'] : wfGetCaller();
2311
2312         wfInitShellLocale();
2313
2314         $envcmd = '';
2315         foreach ( $environ as $k => $v ) {
2316                 if ( wfIsWindows() ) {
2317                         /* Surrounding a set in quotes (method used by wfEscapeShellArg) makes the quotes themselves
2318                          * appear in the environment variable, so we must use carat escaping as documented in
2319                          * https://technet.microsoft.com/en-us/library/cc723564.aspx
2320                          * Note however that the quote isn't listed there, but is needed, and the parentheses
2321                          * are listed there but doesn't appear to need it.
2322                          */
2323                         $envcmd .= "set $k=" . preg_replace( '/([&|()<>^"])/', '^\\1', $v ) . '&& ';
2324                 } else {
2325                         /* Assume this is a POSIX shell, thus required to accept variable assignments before the command
2326                          * http://www.opengroup.org/onlinepubs/009695399/utilities/xcu_chap02.html#tag_02_09_01
2327                          */
2328                         $envcmd .= "$k=" . escapeshellarg( $v ) . ' ';
2329                 }
2330         }
2331         if ( is_array( $cmd ) ) {
2332                 $cmd = wfEscapeShellArg( $cmd );
2333         }
2334
2335         $cmd = $envcmd . $cmd;
2336
2337         $useLogPipe = false;
2338         if ( is_executable( '/bin/bash' ) ) {
2339                 $time = intval( isset( $limits['time'] ) ? $limits['time'] : $wgMaxShellTime );
2340                 if ( isset( $limits['walltime'] ) ) {
2341                         $wallTime = intval( $limits['walltime'] );
2342                 } elseif ( isset( $limits['time'] ) ) {
2343                         $wallTime = $time;
2344                 } else {
2345                         $wallTime = intval( $wgMaxShellWallClockTime );
2346                 }
2347                 $mem = intval( isset( $limits['memory'] ) ? $limits['memory'] : $wgMaxShellMemory );
2348                 $filesize = intval( isset( $limits['filesize'] ) ? $limits['filesize'] : $wgMaxShellFileSize );
2349
2350                 if ( $time > 0 || $mem > 0 || $filesize > 0 || $wallTime > 0 ) {
2351                         $cmd = '/bin/bash ' . escapeshellarg( "$IP/includes/limit.sh" ) . ' ' .
2352                                 escapeshellarg( $cmd ) . ' ' .
2353                                 escapeshellarg(
2354                                         "MW_INCLUDE_STDERR=" . ( $includeStderr ? '1' : '' ) . ';' .
2355                                         "MW_CPU_LIMIT=$time; " .
2356                                         'MW_CGROUP=' . escapeshellarg( $wgShellCgroup ) . '; ' .
2357                                         "MW_MEM_LIMIT=$mem; " .
2358                                         "MW_FILE_SIZE_LIMIT=$filesize; " .
2359                                         "MW_WALL_CLOCK_LIMIT=$wallTime; " .
2360                                         "MW_USE_LOG_PIPE=yes"
2361                                 );
2362                         $useLogPipe = true;
2363                 } elseif ( $includeStderr ) {
2364                         $cmd .= ' 2>&1';
2365                 }
2366         } elseif ( $includeStderr ) {
2367                 $cmd .= ' 2>&1';
2368         }
2369         wfDebug( "wfShellExec: $cmd\n" );
2370
2371         // Don't try to execute commands that exceed Linux's MAX_ARG_STRLEN.
2372         // Other platforms may be more accomodating, but we don't want to be
2373         // accomodating, because very long commands probably include user
2374         // input. See T129506.
2375         if ( strlen( $cmd ) > SHELL_MAX_ARG_STRLEN ) {
2376                 throw new Exception( __METHOD__ .
2377                         '(): total length of $cmd must not exceed SHELL_MAX_ARG_STRLEN' );
2378         }
2379
2380         $desc = [
2381                 0 => [ 'file', 'php://stdin', 'r' ],
2382                 1 => [ 'pipe', 'w' ],
2383                 2 => [ 'file', 'php://stderr', 'w' ] ];
2384         if ( $useLogPipe ) {
2385                 $desc[3] = [ 'pipe', 'w' ];
2386         }
2387         $pipes = null;
2388         $scoped = Profiler::instance()->scopedProfileIn( __FUNCTION__ . '-' . $profileMethod );
2389         $proc = proc_open( $cmd, $desc, $pipes );
2390         if ( !$proc ) {
2391                 wfDebugLog( 'exec', "proc_open() failed: $cmd" );
2392                 $retval = -1;
2393                 return '';
2394         }
2395         $outBuffer = $logBuffer = '';
2396         $emptyArray = [];
2397         $status = false;
2398         $logMsg = false;
2399
2400         /* According to the documentation, it is possible for stream_select()
2401          * to fail due to EINTR. I haven't managed to induce this in testing
2402          * despite sending various signals. If it did happen, the error
2403          * message would take the form:
2404          *
2405          * stream_select(): unable to select [4]: Interrupted system call (max_fd=5)
2406          *
2407          * where [4] is the value of the macro EINTR and "Interrupted system
2408          * call" is string which according to the Linux manual is "possibly"
2409          * localised according to LC_MESSAGES.
2410          */
2411         $eintr = defined( 'SOCKET_EINTR' ) ? SOCKET_EINTR : 4;
2412         $eintrMessage = "stream_select(): unable to select [$eintr]";
2413
2414         $running = true;
2415         $timeout = null;
2416         $numReadyPipes = 0;
2417
2418         while ( $running === true || $numReadyPipes !== 0 ) {
2419                 if ( $running ) {
2420                         $status = proc_get_status( $proc );
2421                         // If the process has terminated, switch to nonblocking selects
2422                         // for getting any data still waiting to be read.
2423                         if ( !$status['running'] ) {
2424                                 $running = false;
2425                                 $timeout = 0;
2426                         }
2427                 }
2428
2429                 $readyPipes = $pipes;
2430
2431                 // Clear last error
2432                 // @codingStandardsIgnoreStart Generic.PHP.NoSilencedErrors.Discouraged
2433                 @trigger_error( '' );
2434                 $numReadyPipes = @stream_select( $readyPipes, $emptyArray, $emptyArray, $timeout );
2435                 if ( $numReadyPipes === false ) {
2436                         // @codingStandardsIgnoreEnd
2437                         $error = error_get_last();
2438                         if ( strncmp( $error['message'], $eintrMessage, strlen( $eintrMessage ) ) == 0 ) {
2439                                 continue;
2440                         } else {
2441                                 trigger_error( $error['message'], E_USER_WARNING );
2442                                 $logMsg = $error['message'];
2443                                 break;
2444                         }
2445                 }
2446                 foreach ( $readyPipes as $fd => $pipe ) {
2447                         $block = fread( $pipe, 65536 );
2448                         if ( $block === '' ) {
2449                                 // End of file
2450                                 fclose( $pipes[$fd] );
2451                                 unset( $pipes[$fd] );
2452                                 if ( !$pipes ) {
2453                                         break 2;
2454                                 }
2455                         } elseif ( $block === false ) {
2456                                 // Read error
2457                                 $logMsg = "Error reading from pipe";
2458                                 break 2;
2459                         } elseif ( $fd == 1 ) {
2460                                 // From stdout
2461                                 $outBuffer .= $block;
2462                         } elseif ( $fd == 3 ) {
2463                                 // From log FD
2464                                 $logBuffer .= $block;
2465                                 if ( strpos( $block, "\n" ) !== false ) {
2466                                         $lines = explode( "\n", $logBuffer );
2467                                         $logBuffer = array_pop( $lines );
2468                                         foreach ( $lines as $line ) {
2469                                                 wfDebugLog( 'exec', $line );
2470                                         }
2471                                 }
2472                         }
2473                 }
2474         }
2475
2476         foreach ( $pipes as $pipe ) {
2477                 fclose( $pipe );
2478         }
2479
2480         // Use the status previously collected if possible, since proc_get_status()
2481         // just calls waitpid() which will not return anything useful the second time.
2482         if ( $running ) {
2483                 $status = proc_get_status( $proc );
2484         }
2485
2486         if ( $logMsg !== false ) {
2487                 // Read/select error
2488                 $retval = -1;
2489                 proc_close( $proc );
2490         } elseif ( $status['signaled'] ) {
2491                 $logMsg = "Exited with signal {$status['termsig']}";
2492                 $retval = 128 + $status['termsig'];
2493                 proc_close( $proc );
2494         } else {
2495                 if ( $status['running'] ) {
2496                         $retval = proc_close( $proc );
2497                 } else {
2498                         $retval = $status['exitcode'];
2499                         proc_close( $proc );
2500                 }
2501                 if ( $retval == 127 ) {
2502                         $logMsg = "Possibly missing executable file";
2503                 } elseif ( $retval >= 129 && $retval <= 192 ) {
2504                         $logMsg = "Probably exited with signal " . ( $retval - 128 );
2505                 }
2506         }
2507
2508         if ( $logMsg !== false ) {
2509                 wfDebugLog( 'exec', "$logMsg: $cmd" );
2510         }
2511
2512         return $outBuffer;
2513 }
2514
2515 /**
2516  * Execute a shell command, returning both stdout and stderr. Convenience
2517  * function, as all the arguments to wfShellExec can become unwieldy.
2518  *
2519  * @note This also includes errors from limit.sh, e.g. if $wgMaxShellFileSize is exceeded.
2520  * @param string|string[] $cmd If string, a properly shell-escaped command line,
2521  *   or an array of unescaped arguments, in which case each value will be escaped
2522  *   Example:   [ 'convert', '-font', 'font name' ] would produce "'convert' '-font' 'font name'"
2523  * @param null|mixed &$retval Optional, will receive the program's exit code.
2524  *   (non-zero is usually failure)
2525  * @param array $environ Optional environment variables which should be
2526  *   added to the executed command environment.
2527  * @param array $limits Optional array with limits(filesize, memory, time, walltime)
2528  *   this overwrites the global wgMaxShell* limits.
2529  * @return string Collected stdout and stderr as a string
2530  */
2531 function wfShellExecWithStderr( $cmd, &$retval = null, $environ = [], $limits = [] ) {
2532         return wfShellExec( $cmd, $retval, $environ, $limits,
2533                 [ 'duplicateStderr' => true, 'profileMethod' => wfGetCaller() ] );
2534 }
2535
2536 /**
2537  * Set the locale for locale-sensitive operations
2538  *
2539  * Sets LC_ALL to a known value to work around issues like the following:
2540  * - https://bugs.php.net/bug.php?id=45132 escapeshellarg() destroys non-ASCII
2541  *   characters if LANG is not a UTF-8 locale
2542  * - T107128 Scribunto string comparison works case insensitive while the
2543  *   standard Lua case sensitive
2544  *
2545  * @see $wgShellLocale
2546  */
2547 function wfInitShellLocale() {
2548         static $done = false;
2549         if ( $done ) {
2550                 return;
2551         }
2552         $done = true;
2553         global $wgShellLocale;
2554         putenv( "LC_ALL=$wgShellLocale" );
2555         setlocale( LC_ALL, $wgShellLocale );
2556 }
2557
2558 /**
2559  * Generate a shell-escaped command line string to run a MediaWiki cli script.
2560  * Note that $parameters should be a flat array and an option with an argument
2561  * should consist of two consecutive items in the array (do not use "--option value").
2562  *
2563  * @param string $script MediaWiki cli script path
2564  * @param array $parameters Arguments and options to the script
2565  * @param array $options Associative array of options:
2566  *     'php': The path to the php executable
2567  *     'wrapper': Path to a PHP wrapper to handle the maintenance script
2568  * @return string
2569  */
2570 function wfShellWikiCmd( $script, array $parameters = [], array $options = [] ) {
2571         global $wgPhpCli;
2572         // Give site config file a chance to run the script in a wrapper.
2573         // The caller may likely want to call wfBasename() on $script.
2574         Hooks::run( 'wfShellWikiCmd', [ &$script, &$parameters, &$options ] );
2575         $cmd = isset( $options['php'] ) ? [ $options['php'] ] : [ $wgPhpCli ];
2576         if ( isset( $options['wrapper'] ) ) {
2577                 $cmd[] = $options['wrapper'];
2578         }
2579         $cmd[] = $script;
2580         // Escape each parameter for shell
2581         return wfEscapeShellArg( array_merge( $cmd, $parameters ) );
2582 }
2583
2584 /**
2585  * wfMerge attempts to merge differences between three texts.
2586  * Returns true for a clean merge and false for failure or a conflict.
2587  *
2588  * @param string $old
2589  * @param string $mine
2590  * @param string $yours
2591  * @param string $result
2592  * @return bool
2593  */
2594 function wfMerge( $old, $mine, $yours, &$result ) {
2595         global $wgDiff3;
2596
2597         # This check may also protect against code injection in
2598         # case of broken installations.
2599         MediaWiki\suppressWarnings();
2600         $haveDiff3 = $wgDiff3 && file_exists( $wgDiff3 );
2601         MediaWiki\restoreWarnings();
2602
2603         if ( !$haveDiff3 ) {
2604                 wfDebug( "diff3 not found\n" );
2605                 return false;
2606         }
2607
2608         # Make temporary files
2609         $td = wfTempDir();
2610         $oldtextFile = fopen( $oldtextName = tempnam( $td, 'merge-old-' ), 'w' );
2611         $mytextFile = fopen( $mytextName = tempnam( $td, 'merge-mine-' ), 'w' );
2612         $yourtextFile = fopen( $yourtextName = tempnam( $td, 'merge-your-' ), 'w' );
2613
2614         # NOTE: diff3 issues a warning to stderr if any of the files does not end with
2615         #       a newline character. To avoid this, we normalize the trailing whitespace before
2616         #       creating the diff.
2617
2618         fwrite( $oldtextFile, rtrim( $old ) . "\n" );
2619         fclose( $oldtextFile );
2620         fwrite( $mytextFile, rtrim( $mine ) . "\n" );
2621         fclose( $mytextFile );
2622         fwrite( $yourtextFile, rtrim( $yours ) . "\n" );
2623         fclose( $yourtextFile );
2624
2625         # Check for a conflict
2626         $cmd = wfEscapeShellArg( $wgDiff3, '-a', '--overlap-only', $mytextName,
2627                 $oldtextName, $yourtextName );
2628         $handle = popen( $cmd, 'r' );
2629
2630         if ( fgets( $handle, 1024 ) ) {
2631                 $conflict = true;
2632         } else {
2633                 $conflict = false;
2634         }
2635         pclose( $handle );
2636
2637         # Merge differences
2638         $cmd = wfEscapeShellArg( $wgDiff3, '-a', '-e', '--merge', $mytextName,
2639                 $oldtextName, $yourtextName );
2640         $handle = popen( $cmd, 'r' );
2641         $result = '';
2642         do {
2643                 $data = fread( $handle, 8192 );
2644                 if ( strlen( $data ) == 0 ) {
2645                         break;
2646                 }
2647                 $result .= $data;
2648         } while ( true );
2649         pclose( $handle );
2650         unlink( $mytextName );
2651         unlink( $oldtextName );
2652         unlink( $yourtextName );
2653
2654         if ( $result === '' && $old !== '' && !$conflict ) {
2655                 wfDebug( "Unexpected null result from diff3. Command: $cmd\n" );
2656                 $conflict = true;
2657         }
2658         return !$conflict;
2659 }
2660
2661 /**
2662  * Returns unified plain-text diff of two texts.
2663  * "Useful" for machine processing of diffs.
2664  *
2665  * @deprecated since 1.25, use DiffEngine/UnifiedDiffFormatter directly
2666  *
2667  * @param string $before The text before the changes.
2668  * @param string $after The text after the changes.
2669  * @param string $params Command-line options for the diff command.
2670  * @return string Unified diff of $before and $after
2671  */
2672 function wfDiff( $before, $after, $params = '-u' ) {
2673         if ( $before == $after ) {
2674                 return '';
2675         }
2676
2677         global $wgDiff;
2678         MediaWiki\suppressWarnings();
2679         $haveDiff = $wgDiff && file_exists( $wgDiff );
2680         MediaWiki\restoreWarnings();
2681
2682         # This check may also protect against code injection in
2683         # case of broken installations.
2684         if ( !$haveDiff ) {
2685                 wfDebug( "diff executable not found\n" );
2686                 $diffs = new Diff( explode( "\n", $before ), explode( "\n", $after ) );
2687                 $format = new UnifiedDiffFormatter();
2688                 return $format->format( $diffs );
2689         }
2690
2691         # Make temporary files
2692         $td = wfTempDir();
2693         $oldtextFile = fopen( $oldtextName = tempnam( $td, 'merge-old-' ), 'w' );
2694         $newtextFile = fopen( $newtextName = tempnam( $td, 'merge-your-' ), 'w' );
2695
2696         fwrite( $oldtextFile, $before );
2697         fclose( $oldtextFile );
2698         fwrite( $newtextFile, $after );
2699         fclose( $newtextFile );
2700
2701         // Get the diff of the two files
2702         $cmd = "$wgDiff " . $params . ' ' . wfEscapeShellArg( $oldtextName, $newtextName );
2703
2704         $h = popen( $cmd, 'r' );
2705         if ( !$h ) {
2706                 unlink( $oldtextName );
2707                 unlink( $newtextName );
2708                 throw new Exception( __METHOD__ . '(): popen() failed' );
2709         }
2710
2711         $diff = '';
2712
2713         do {
2714                 $data = fread( $h, 8192 );
2715                 if ( strlen( $data ) == 0 ) {
2716                         break;
2717                 }
2718                 $diff .= $data;
2719         } while ( true );
2720
2721         // Clean up
2722         pclose( $h );
2723         unlink( $oldtextName );
2724         unlink( $newtextName );
2725
2726         // Kill the --- and +++ lines. They're not useful.
2727         $diff_lines = explode( "\n", $diff );
2728         if ( isset( $diff_lines[0] ) && strpos( $diff_lines[0], '---' ) === 0 ) {
2729                 unset( $diff_lines[0] );
2730         }
2731         if ( isset( $diff_lines[1] ) && strpos( $diff_lines[1], '+++' ) === 0 ) {
2732                 unset( $diff_lines[1] );
2733         }
2734
2735         $diff = implode( "\n", $diff_lines );
2736
2737         return $diff;
2738 }
2739
2740 /**
2741  * This function works like "use VERSION" in Perl, the program will die with a
2742  * backtrace if the current version of PHP is less than the version provided
2743  *
2744  * This is useful for extensions which due to their nature are not kept in sync
2745  * with releases, and might depend on other versions of PHP than the main code
2746  *
2747  * Note: PHP might die due to parsing errors in some cases before it ever
2748  *       manages to call this function, such is life
2749  *
2750  * @see perldoc -f use
2751  *
2752  * @param string|int|float $req_ver The version to check, can be a string, an integer, or a float
2753  * @throws MWException
2754  */
2755 function wfUsePHP( $req_ver ) {
2756         $php_ver = PHP_VERSION;
2757
2758         if ( version_compare( $php_ver, (string)$req_ver, '<' ) ) {
2759                 throw new MWException( "PHP $req_ver required--this is only $php_ver" );
2760         }
2761 }
2762
2763 /**
2764  * This function works like "use VERSION" in Perl except it checks the version
2765  * of MediaWiki, the program will die with a backtrace if the current version
2766  * of MediaWiki is less than the version provided.
2767  *
2768  * This is useful for extensions which due to their nature are not kept in sync
2769  * with releases
2770  *
2771  * Note: Due to the behavior of PHP's version_compare() which is used in this
2772  * function, if you want to allow the 'wmf' development versions add a 'c' (or
2773  * any single letter other than 'a', 'b' or 'p') as a post-fix to your
2774  * targeted version number. For example if you wanted to allow any variation
2775  * of 1.22 use `wfUseMW( '1.22c' )`. Using an 'a' or 'b' instead of 'c' will
2776  * not result in the same comparison due to the internal logic of
2777  * version_compare().
2778  *
2779  * @see perldoc -f use
2780  *
2781  * @deprecated since 1.26, use the "requires' property of extension.json
2782  * @param string|int|float $req_ver The version to check, can be a string, an integer, or a float
2783  * @throws MWException
2784  */
2785 function wfUseMW( $req_ver ) {
2786         global $wgVersion;
2787
2788         if ( version_compare( $wgVersion, (string)$req_ver, '<' ) ) {
2789                 throw new MWException( "MediaWiki $req_ver required--this is only $wgVersion" );
2790         }
2791 }
2792
2793 /**
2794  * Return the final portion of a pathname.
2795  * Reimplemented because PHP5's "basename()" is buggy with multibyte text.
2796  * https://bugs.php.net/bug.php?id=33898
2797  *
2798  * PHP's basename() only considers '\' a pathchar on Windows and Netware.
2799  * We'll consider it so always, as we don't want '\s' in our Unix paths either.
2800  *
2801  * @param string $path
2802  * @param string $suffix String to remove if present
2803  * @return string
2804  */
2805 function wfBaseName( $path, $suffix = '' ) {
2806         if ( $suffix == '' ) {
2807                 $encSuffix = '';
2808         } else {
2809                 $encSuffix = '(?:' . preg_quote( $suffix, '#' ) . ')?';
2810         }
2811
2812         $matches = [];
2813         if ( preg_match( "#([^/\\\\]*?){$encSuffix}[/\\\\]*$#", $path, $matches ) ) {
2814                 return $matches[1];
2815         } else {
2816                 return '';
2817         }
2818 }
2819
2820 /**
2821  * Generate a relative path name to the given file.
2822  * May explode on non-matching case-insensitive paths,
2823  * funky symlinks, etc.
2824  *
2825  * @param string $path Absolute destination path including target filename
2826  * @param string $from Absolute source path, directory only
2827  * @return string
2828  */
2829 function wfRelativePath( $path, $from ) {
2830         // Normalize mixed input on Windows...
2831         $path = str_replace( '/', DIRECTORY_SEPARATOR, $path );
2832         $from = str_replace( '/', DIRECTORY_SEPARATOR, $from );
2833
2834         // Trim trailing slashes -- fix for drive root
2835         $path = rtrim( $path, DIRECTORY_SEPARATOR );
2836         $from = rtrim( $from, DIRECTORY_SEPARATOR );
2837
2838         $pieces = explode( DIRECTORY_SEPARATOR, dirname( $path ) );
2839         $against = explode( DIRECTORY_SEPARATOR, $from );
2840
2841         if ( $pieces[0] !== $against[0] ) {
2842                 // Non-matching Windows drive letters?
2843                 // Return a full path.
2844                 return $path;
2845         }
2846
2847         // Trim off common prefix
2848         while ( count( $pieces ) && count( $against )
2849                 && $pieces[0] == $against[0] ) {
2850                 array_shift( $pieces );
2851                 array_shift( $against );
2852         }
2853
2854         // relative dots to bump us to the parent
2855         while ( count( $against ) ) {
2856                 array_unshift( $pieces, '..' );
2857                 array_shift( $against );
2858         }
2859
2860         array_push( $pieces, wfBaseName( $path ) );
2861
2862         return implode( DIRECTORY_SEPARATOR, $pieces );
2863 }
2864
2865 /**
2866  * Convert an arbitrarily-long digit string from one numeric base
2867  * to another, optionally zero-padding to a minimum column width.
2868  *
2869  * Supports base 2 through 36; digit values 10-36 are represented
2870  * as lowercase letters a-z. Input is case-insensitive.
2871  *
2872  * @deprecated since 1.27 Use Wikimedia\base_convert() directly
2873  *
2874  * @param string $input Input number
2875  * @param int $sourceBase Base of the input number
2876  * @param int $destBase Desired base of the output
2877  * @param int $pad Minimum number of digits in the output (pad with zeroes)
2878  * @param bool $lowercase Whether to output in lowercase or uppercase
2879  * @param string $engine Either "gmp", "bcmath", or "php"
2880  * @return string|bool The output number as a string, or false on error
2881  */
2882 function wfBaseConvert( $input, $sourceBase, $destBase, $pad = 1,
2883         $lowercase = true, $engine = 'auto'
2884 ) {
2885         return Wikimedia\base_convert( $input, $sourceBase, $destBase, $pad, $lowercase, $engine );
2886 }
2887
2888 /**
2889  * @deprecated since 1.27, PHP's session generation isn't used with
2890  *  MediaWiki\Session\SessionManager
2891  */
2892 function wfFixSessionID() {
2893         wfDeprecated( __FUNCTION__, '1.27' );
2894 }
2895
2896 /**
2897  * Reset the session id
2898  *
2899  * @deprecated since 1.27, use MediaWiki\Session\SessionManager instead
2900  * @since 1.22
2901  */
2902 function wfResetSessionID() {
2903         wfDeprecated( __FUNCTION__, '1.27' );
2904         $session = SessionManager::getGlobalSession();
2905         $delay = $session->delaySave();
2906
2907         $session->resetId();
2908
2909         // Make sure a session is started, since that's what the old
2910         // wfResetSessionID() did.
2911         if ( session_id() !== $session->getId() ) {
2912                 wfSetupSession( $session->getId() );
2913         }
2914
2915         ScopedCallback::consume( $delay );
2916 }
2917
2918 /**
2919  * Initialise php session
2920  *
2921  * @deprecated since 1.27, use MediaWiki\Session\SessionManager instead.
2922  *  Generally, "using" SessionManager will be calling ->getSessionById() or
2923  *  ::getGlobalSession() (depending on whether you were passing $sessionId
2924  *  here), then calling $session->persist().
2925  * @param bool|string $sessionId
2926  */
2927 function wfSetupSession( $sessionId = false ) {
2928         wfDeprecated( __FUNCTION__, '1.27' );
2929
2930         if ( $sessionId ) {
2931                 session_id( $sessionId );
2932         }
2933
2934         $session = SessionManager::getGlobalSession();
2935         $session->persist();
2936
2937         if ( session_id() !== $session->getId() ) {
2938                 session_id( $session->getId() );
2939         }
2940         MediaWiki\quietCall( 'session_start' );
2941 }
2942
2943 /**
2944  * Get an object from the precompiled serialized directory
2945  *
2946  * @param string $name
2947  * @return mixed The variable on success, false on failure
2948  */
2949 function wfGetPrecompiledData( $name ) {
2950         global $IP;
2951
2952         $file = "$IP/serialized/$name";
2953         if ( file_exists( $file ) ) {
2954                 $blob = file_get_contents( $file );
2955                 if ( $blob ) {
2956                         return unserialize( $blob );
2957                 }
2958         }
2959         return false;
2960 }
2961
2962 /**
2963  * Make a cache key for the local wiki.
2964  *
2965  * @deprecated since 1.30 Call makeKey on a BagOStuff instance
2966  * @param string $args,...
2967  * @return string
2968  */
2969 function wfMemcKey( /*...*/ ) {
2970         return call_user_func_array(
2971                 [ ObjectCache::getLocalClusterInstance(), 'makeKey' ],
2972                 func_get_args()
2973         );
2974 }
2975
2976 /**
2977  * Make a cache key for a foreign DB.
2978  *
2979  * Must match what wfMemcKey() would produce in context of the foreign wiki.
2980  *
2981  * @param string $db
2982  * @param string $prefix
2983  * @param string $args,...
2984  * @return string
2985  */
2986 function wfForeignMemcKey( $db, $prefix /*...*/ ) {
2987         $args = array_slice( func_get_args(), 2 );
2988         $keyspace = $prefix ? "$db-$prefix" : $db;
2989         return call_user_func_array(
2990                 [ ObjectCache::getLocalClusterInstance(), 'makeKeyInternal' ],
2991                 [ $keyspace, $args ]
2992         );
2993 }
2994
2995 /**
2996  * Make a cache key with database-agnostic prefix.
2997  *
2998  * Doesn't have a wiki-specific namespace. Uses a generic 'global' prefix
2999  * instead. Must have a prefix as otherwise keys that use a database name
3000  * in the first segment will clash with wfMemcKey/wfForeignMemcKey.
3001  *
3002  * @deprecated since 1.30 Call makeGlobalKey on a BagOStuff instance
3003  * @since 1.26
3004  * @param string $args,...
3005  * @return string
3006  */
3007 function wfGlobalCacheKey( /*...*/ ) {
3008         return call_user_func_array(
3009                 [ ObjectCache::getLocalClusterInstance(), 'makeGlobalKey' ],
3010                 func_get_args()
3011         );
3012 }
3013
3014 /**
3015  * Get an ASCII string identifying this wiki
3016  * This is used as a prefix in memcached keys
3017  *
3018  * @return string
3019  */
3020 function wfWikiID() {
3021         global $wgDBprefix, $wgDBname;
3022         if ( $wgDBprefix ) {
3023                 return "$wgDBname-$wgDBprefix";
3024         } else {
3025                 return $wgDBname;
3026         }
3027 }
3028
3029 /**
3030  * Split a wiki ID into DB name and table prefix
3031  *
3032  * @param string $wiki
3033  *
3034  * @return array
3035  */
3036 function wfSplitWikiID( $wiki ) {
3037         $bits = explode( '-', $wiki, 2 );
3038         if ( count( $bits ) < 2 ) {
3039                 $bits[] = '';
3040         }
3041         return $bits;
3042 }
3043
3044 /**
3045  * Get a Database object.
3046  *
3047  * @param int $db Index of the connection to get. May be DB_MASTER for the
3048  *            master (for write queries), DB_REPLICA for potentially lagged read
3049  *            queries, or an integer >= 0 for a particular server.
3050  *
3051  * @param string|string[] $groups Query groups. An array of group names that this query
3052  *                belongs to. May contain a single string if the query is only
3053  *                in one group.
3054  *
3055  * @param string|bool $wiki The wiki ID, or false for the current wiki
3056  *
3057  * Note: multiple calls to wfGetDB(DB_REPLICA) during the course of one request
3058  * will always return the same object, unless the underlying connection or load
3059  * balancer is manually destroyed.
3060  *
3061  * Note 2: use $this->getDB() in maintenance scripts that may be invoked by
3062  * updater to ensure that a proper database is being updated.
3063  *
3064  * @todo Replace calls to wfGetDB with calls to LoadBalancer::getConnection()
3065  *       on an injected instance of LoadBalancer.
3066  *
3067  * @return \Wikimedia\Rdbms\Database
3068  */
3069 function wfGetDB( $db, $groups = [], $wiki = false ) {
3070         return wfGetLB( $wiki )->getConnection( $db, $groups, $wiki );
3071 }
3072
3073 /**
3074  * Get a load balancer object.
3075  *
3076  * @deprecated since 1.27, use MediaWikiServices::getDBLoadBalancer()
3077  *              or MediaWikiServices::getDBLoadBalancerFactory() instead.
3078  *
3079  * @param string|bool $wiki Wiki ID, or false for the current wiki
3080  * @return \Wikimedia\Rdbms\LoadBalancer
3081  */
3082 function wfGetLB( $wiki = false ) {
3083         if ( $wiki === false ) {
3084                 return MediaWikiServices::getInstance()->getDBLoadBalancer();
3085         } else {
3086                 $factory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
3087                 return $factory->getMainLB( $wiki );
3088         }
3089 }
3090
3091 /**
3092  * Get the load balancer factory object
3093  *
3094  * @deprecated since 1.27, use MediaWikiServices::getDBLoadBalancerFactory() instead.
3095  *
3096  * @return \Wikimedia\Rdbms\LBFactory
3097  */
3098 function wfGetLBFactory() {
3099         return MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
3100 }
3101
3102 /**
3103  * Find a file.
3104  * Shortcut for RepoGroup::singleton()->findFile()
3105  *
3106  * @param string $title String or Title object
3107  * @param array $options Associative array of options (see RepoGroup::findFile)
3108  * @return File|bool File, or false if the file does not exist
3109  */
3110 function wfFindFile( $title, $options = [] ) {
3111         return RepoGroup::singleton()->findFile( $title, $options );
3112 }
3113
3114 /**
3115  * Get an object referring to a locally registered file.
3116  * Returns a valid placeholder object if the file does not exist.
3117  *
3118  * @param Title|string $title
3119  * @return LocalFile|null A File, or null if passed an invalid Title
3120  */
3121 function wfLocalFile( $title ) {
3122         return RepoGroup::singleton()->getLocalRepo()->newFile( $title );
3123 }
3124
3125 /**
3126  * Should low-performance queries be disabled?
3127  *
3128  * @return bool
3129  * @codeCoverageIgnore
3130  */
3131 function wfQueriesMustScale() {
3132         global $wgMiserMode;
3133         return $wgMiserMode
3134                 || ( SiteStats::pages() > 100000
3135                 && SiteStats::edits() > 1000000
3136                 && SiteStats::users() > 10000 );
3137 }
3138
3139 /**
3140  * Get the path to a specified script file, respecting file
3141  * extensions; this is a wrapper around $wgScriptPath etc.
3142  * except for 'index' and 'load' which use $wgScript/$wgLoadScript
3143  *
3144  * @param string $script Script filename, sans extension
3145  * @return string
3146  */
3147 function wfScript( $script = 'index' ) {
3148         global $wgScriptPath, $wgScript, $wgLoadScript;
3149         if ( $script === 'index' ) {
3150                 return $wgScript;
3151         } elseif ( $script === 'load' ) {
3152                 return $wgLoadScript;
3153         } else {
3154                 return "{$wgScriptPath}/{$script}.php";
3155         }
3156 }
3157
3158 /**
3159  * Get the script URL.
3160  *
3161  * @return string Script URL
3162  */
3163 function wfGetScriptUrl() {
3164         if ( isset( $_SERVER['SCRIPT_NAME'] ) ) {
3165                 /* as it was called, minus the query string.
3166                  *
3167                  * Some sites use Apache rewrite rules to handle subdomains,
3168                  * and have PHP set up in a weird way that causes PHP_SELF
3169                  * to contain the rewritten URL instead of the one that the
3170                  * outside world sees.
3171                  *
3172                  * If in this mode, use SCRIPT_URL instead, which mod_rewrite
3173                  * provides containing the "before" URL.
3174                  */
3175                 return $_SERVER['SCRIPT_NAME'];
3176         } else {
3177                 return $_SERVER['URL'];
3178         }
3179 }
3180
3181 /**
3182  * Convenience function converts boolean values into "true"
3183  * or "false" (string) values
3184  *
3185  * @param bool $value
3186  * @return string
3187  */
3188 function wfBoolToStr( $value ) {
3189         return $value ? 'true' : 'false';
3190 }
3191
3192 /**
3193  * Get a platform-independent path to the null file, e.g. /dev/null
3194  *
3195  * @return string
3196  */
3197 function wfGetNull() {
3198         return wfIsWindows() ? 'NUL' : '/dev/null';
3199 }
3200
3201 /**
3202  * Waits for the replica DBs to catch up to the master position
3203  *
3204  * Use this when updating very large numbers of rows, as in maintenance scripts,
3205  * to avoid causing too much lag. Of course, this is a no-op if there are no replica DBs.
3206  *
3207  * By default this waits on the main DB cluster of the current wiki.
3208  * If $cluster is set to "*" it will wait on all DB clusters, including
3209  * external ones. If the lag being waiting on is caused by the code that
3210  * does this check, it makes since to use $ifWritesSince, particularly if
3211  * cluster is "*", to avoid excess overhead.
3212  *
3213  * Never call this function after a big DB write that is still in a transaction.
3214  * This only makes sense after the possible lag inducing changes were committed.
3215  *
3216  * @param float|null $ifWritesSince Only wait if writes were done since this UNIX timestamp
3217  * @param string|bool $wiki Wiki identifier accepted by wfGetLB
3218  * @param string|bool $cluster Cluster name accepted by LBFactory. Default: false.
3219  * @param int|null $timeout Max wait time. Default: 1 day (cli), ~10 seconds (web)
3220  * @return bool Success (able to connect and no timeouts reached)
3221  * @deprecated since 1.27 Use LBFactory::waitForReplication
3222  */
3223 function wfWaitForSlaves(
3224         $ifWritesSince = null, $wiki = false, $cluster = false, $timeout = null
3225 ) {
3226         if ( $timeout === null ) {
3227                 $timeout = ( PHP_SAPI === 'cli' ) ? 86400 : 10;
3228         }
3229
3230         if ( $cluster === '*' ) {
3231                 $cluster = false;
3232                 $wiki = false;
3233         } elseif ( $wiki === false ) {
3234                 $wiki = wfWikiID();
3235         }
3236
3237         try {
3238                 wfGetLBFactory()->waitForReplication( [
3239                         'wiki' => $wiki,
3240                         'cluster' => $cluster,
3241                         'timeout' => $timeout,
3242                         // B/C: first argument used to be "max seconds of lag"; ignore such values
3243                         'ifWritesSince' => ( $ifWritesSince > 1e9 ) ? $ifWritesSince : null
3244                 ] );
3245         } catch ( DBReplicationWaitError $e ) {
3246                 return false;
3247         }
3248
3249         return true;
3250 }
3251
3252 /**
3253  * Count down from $seconds to zero on the terminal, with a one-second pause
3254  * between showing each number. For use in command-line scripts.
3255  *
3256  * @codeCoverageIgnore
3257  * @param int $seconds
3258  */
3259 function wfCountDown( $seconds ) {
3260         for ( $i = $seconds; $i >= 0; $i-- ) {
3261                 if ( $i != $seconds ) {
3262                         echo str_repeat( "\x08", strlen( $i + 1 ) );
3263                 }
3264                 echo $i;
3265                 flush();
3266                 if ( $i ) {
3267                         sleep( 1 );
3268                 }
3269         }
3270         echo "\n";
3271 }
3272
3273 /**
3274  * Replace all invalid characters with '-'.
3275  * Additional characters can be defined in $wgIllegalFileChars (see T22489).
3276  * By default, $wgIllegalFileChars includes ':', '/', '\'.
3277  *
3278  * @param string $name Filename to process
3279  * @return string
3280  */
3281 function wfStripIllegalFilenameChars( $name ) {
3282         global $wgIllegalFileChars;
3283         $illegalFileChars = $wgIllegalFileChars ? "|[" . $wgIllegalFileChars . "]" : '';
3284         $name = preg_replace(
3285                 "/[^" . Title::legalChars() . "]" . $illegalFileChars . "/",
3286                 '-',
3287                 $name
3288         );
3289         // $wgIllegalFileChars may not include '/' and '\', so we still need to do this
3290         $name = wfBaseName( $name );
3291         return $name;
3292 }
3293
3294 /**
3295  * Set PHP's memory limit to the larger of php.ini or $wgMemoryLimit
3296  *
3297  * @return int Resulting value of the memory limit.
3298  */
3299 function wfMemoryLimit() {
3300         global $wgMemoryLimit;
3301         $memlimit = wfShorthandToInteger( ini_get( 'memory_limit' ) );
3302         if ( $memlimit != -1 ) {
3303                 $conflimit = wfShorthandToInteger( $wgMemoryLimit );
3304                 if ( $conflimit == -1 ) {
3305                         wfDebug( "Removing PHP's memory limit\n" );
3306                         MediaWiki\suppressWarnings();
3307                         ini_set( 'memory_limit', $conflimit );
3308                         MediaWiki\restoreWarnings();
3309                         return $conflimit;
3310                 } elseif ( $conflimit > $memlimit ) {
3311                         wfDebug( "Raising PHP's memory limit to $conflimit bytes\n" );
3312                         MediaWiki\suppressWarnings();
3313                         ini_set( 'memory_limit', $conflimit );
3314                         MediaWiki\restoreWarnings();
3315                         return $conflimit;
3316                 }
3317         }
3318         return $memlimit;
3319 }
3320
3321 /**
3322  * Set PHP's time limit to the larger of php.ini or $wgTransactionalTimeLimit
3323  *
3324  * @return int Prior time limit
3325  * @since 1.26
3326  */
3327 function wfTransactionalTimeLimit() {
3328         global $wgTransactionalTimeLimit;
3329
3330         $timeLimit = ini_get( 'max_execution_time' );
3331         // Note that CLI scripts use 0
3332         if ( $timeLimit > 0 && $wgTransactionalTimeLimit > $timeLimit ) {
3333                 set_time_limit( $wgTransactionalTimeLimit );
3334         }
3335
3336         ignore_user_abort( true ); // ignore client disconnects
3337
3338         return $timeLimit;
3339 }
3340
3341 /**
3342  * Converts shorthand byte notation to integer form
3343  *
3344  * @param string $string
3345  * @param int $default Returned if $string is empty
3346  * @return int
3347  */
3348 function wfShorthandToInteger( $string = '', $default = -1 ) {
3349         $string = trim( $string );
3350         if ( $string === '' ) {
3351                 return $default;
3352         }
3353         $last = $string[strlen( $string ) - 1];
3354         $val = intval( $string );
3355         switch ( $last ) {
3356                 case 'g':
3357                 case 'G':
3358                         $val *= 1024;
3359                         // break intentionally missing
3360                 case 'm':
3361                 case 'M':
3362                         $val *= 1024;
3363                         // break intentionally missing
3364                 case 'k':
3365                 case 'K':
3366                         $val *= 1024;
3367         }
3368
3369         return $val;
3370 }
3371
3372 /**
3373  * Get the normalised IETF language tag
3374  * See unit test for examples.
3375  *
3376  * @param string $code The language code.
3377  * @return string The language code which complying with BCP 47 standards.
3378  */
3379 function wfBCP47( $code ) {
3380         $codeSegment = explode( '-', $code );
3381         $codeBCP = [];
3382         foreach ( $codeSegment as $segNo => $seg ) {
3383                 // when previous segment is x, it is a private segment and should be lc
3384                 if ( $segNo > 0 && strtolower( $codeSegment[( $segNo - 1 )] ) == 'x' ) {
3385                         $codeBCP[$segNo] = strtolower( $seg );
3386                 // ISO 3166 country code
3387                 } elseif ( ( strlen( $seg ) == 2 ) && ( $segNo > 0 ) ) {
3388                         $codeBCP[$segNo] = strtoupper( $seg );
3389                 // ISO 15924 script code
3390                 } elseif ( ( strlen( $seg ) == 4 ) && ( $segNo > 0 ) ) {
3391                         $codeBCP[$segNo] = ucfirst( strtolower( $seg ) );
3392                 // Use lowercase for other cases
3393                 } else {
3394                         $codeBCP[$segNo] = strtolower( $seg );
3395                 }
3396         }
3397         $langCode = implode( '-', $codeBCP );
3398         return $langCode;
3399 }
3400
3401 /**
3402  * Get a specific cache object.
3403  *
3404  * @param int|string $cacheType A CACHE_* constants, or other key in $wgObjectCaches
3405  * @return BagOStuff
3406  */
3407 function wfGetCache( $cacheType ) {
3408         return ObjectCache::getInstance( $cacheType );
3409 }
3410
3411 /**
3412  * Get the main cache object
3413  *
3414  * @return BagOStuff
3415  */
3416 function wfGetMainCache() {
3417         global $wgMainCacheType;
3418         return ObjectCache::getInstance( $wgMainCacheType );
3419 }
3420
3421 /**
3422  * Get the cache object used by the message cache
3423  *
3424  * @return BagOStuff
3425  */
3426 function wfGetMessageCacheStorage() {
3427         global $wgMessageCacheType;
3428         return ObjectCache::getInstance( $wgMessageCacheType );
3429 }
3430
3431 /**
3432  * Get the cache object used by the parser cache
3433  *
3434  * @return BagOStuff
3435  */
3436 function wfGetParserCacheStorage() {
3437         global $wgParserCacheType;
3438         return ObjectCache::getInstance( $wgParserCacheType );
3439 }
3440
3441 /**
3442  * Call hook functions defined in $wgHooks
3443  *
3444  * @param string $event Event name
3445  * @param array $args Parameters passed to hook functions
3446  * @param string|null $deprecatedVersion Optionally mark hook as deprecated with version number
3447  *
3448  * @return bool True if no handler aborted the hook
3449  * @deprecated since 1.25 - use Hooks::run
3450  */
3451 function wfRunHooks( $event, array $args = [], $deprecatedVersion = null ) {
3452         return Hooks::run( $event, $args, $deprecatedVersion );
3453 }
3454
3455 /**
3456  * Wrapper around php's unpack.
3457  *
3458  * @param string $format The format string (See php's docs)
3459  * @param string $data A binary string of binary data
3460  * @param int|bool $length The minimum length of $data or false. This is to
3461  *      prevent reading beyond the end of $data. false to disable the check.
3462  *
3463  * Also be careful when using this function to read unsigned 32 bit integer
3464  * because php might make it negative.
3465  *
3466  * @throws MWException If $data not long enough, or if unpack fails
3467  * @return array Associative array of the extracted data
3468  */
3469 function wfUnpack( $format, $data, $length = false ) {
3470         if ( $length !== false ) {
3471                 $realLen = strlen( $data );
3472                 if ( $realLen < $length ) {
3473                         throw new MWException( "Tried to use wfUnpack on a "
3474                                 . "string of length $realLen, but needed one "
3475                                 . "of at least length $length."
3476                         );
3477                 }
3478         }
3479
3480         MediaWiki\suppressWarnings();
3481         $result = unpack( $format, $data );
3482         MediaWiki\restoreWarnings();
3483
3484         if ( $result === false ) {
3485                 // If it cannot extract the packed data.
3486                 throw new MWException( "unpack could not unpack binary data" );
3487         }
3488         return $result;
3489 }
3490
3491 /**
3492  * Determine if an image exists on the 'bad image list'.
3493  *
3494  * The format of MediaWiki:Bad_image_list is as follows:
3495  *    * Only list items (lines starting with "*") are considered
3496  *    * The first link on a line must be a link to a bad image
3497  *    * Any subsequent links on the same line are considered to be exceptions,
3498  *      i.e. articles where the image may occur inline.
3499  *
3500  * @param string $name The image name to check
3501  * @param Title|bool $contextTitle The page on which the image occurs, if known
3502  * @param string $blacklist Wikitext of a file blacklist
3503  * @return bool
3504  */
3505 function wfIsBadImage( $name, $contextTitle = false, $blacklist = null ) {
3506         # Handle redirects; callers almost always hit wfFindFile() anyway,
3507         # so just use that method because it has a fast process cache.
3508         $file = wfFindFile( $name ); // get the final name
3509         $name = $file ? $file->getTitle()->getDBkey() : $name;
3510
3511         # Run the extension hook
3512         $bad = false;
3513         if ( !Hooks::run( 'BadImage', [ $name, &$bad ] ) ) {
3514                 return (bool)$bad;
3515         }
3516
3517         $cache = ObjectCache::getLocalServerInstance( 'hash' );
3518         $key = $cache->makeKey(
3519                 'bad-image-list', ( $blacklist === null ) ? 'default' : md5( $blacklist )
3520         );
3521         $badImages = $cache->get( $key );
3522
3523         if ( $badImages === false ) { // cache miss
3524                 if ( $blacklist === null ) {
3525                         $blacklist = wfMessage( 'bad_image_list' )->inContentLanguage()->plain(); // site list
3526                 }
3527                 # Build the list now
3528                 $badImages = [];
3529                 $lines = explode( "\n", $blacklist );
3530                 foreach ( $lines as $line ) {
3531                         # List items only
3532                         if ( substr( $line, 0, 1 ) !== '*' ) {
3533                                 continue;
3534                         }
3535
3536                         # Find all links
3537                         $m = [];
3538                         if ( !preg_match_all( '/\[\[:?(.*?)\]\]/', $line, $m ) ) {
3539                                 continue;
3540                         }
3541
3542                         $exceptions = [];
3543                         $imageDBkey = false;
3544                         foreach ( $m[1] as $i => $titleText ) {
3545                                 $title = Title::newFromText( $titleText );
3546                                 if ( !is_null( $title ) ) {
3547                                         if ( $i == 0 ) {
3548                                                 $imageDBkey = $title->getDBkey();
3549                                         } else {
3550                                                 $exceptions[$title->getPrefixedDBkey()] = true;
3551                                         }
3552                                 }
3553                         }
3554
3555                         if ( $imageDBkey !== false ) {
3556                                 $badImages[$imageDBkey] = $exceptions;
3557                         }
3558                 }
3559                 $cache->set( $key, $badImages, 60 );
3560         }
3561
3562         $contextKey = $contextTitle ? $contextTitle->getPrefixedDBkey() : false;
3563         $bad = isset( $badImages[$name] ) && !isset( $badImages[$name][$contextKey] );
3564
3565         return $bad;
3566 }
3567
3568 /**
3569  * Determine whether the client at a given source IP is likely to be able to
3570  * access the wiki via HTTPS.
3571  *
3572  * @param string $ip The IPv4/6 address in the normal human-readable form
3573  * @return bool
3574  */
3575 function wfCanIPUseHTTPS( $ip ) {
3576         $canDo = true;
3577         Hooks::run( 'CanIPUseHTTPS', [ $ip, &$canDo ] );
3578         return !!$canDo;
3579 }
3580
3581 /**
3582  * Determine input string is represents as infinity
3583  *
3584  * @param string $str The string to determine
3585  * @return bool
3586  * @since 1.25
3587  */
3588 function wfIsInfinity( $str ) {
3589         $infinityValues = [ 'infinite', 'indefinite', 'infinity', 'never' ];
3590         return in_array( $str, $infinityValues );
3591 }
3592
3593 /**
3594  * Returns true if these thumbnail parameters match one that MediaWiki
3595  * requests from file description pages and/or parser output.
3596  *
3597  * $params is considered non-standard if they involve a non-standard
3598  * width or any non-default parameters aside from width and page number.
3599  * The number of possible files with standard parameters is far less than
3600  * that of all combinations; rate-limiting for them can thus be more generious.
3601  *
3602  * @param File $file
3603  * @param array $params
3604  * @return bool
3605  * @since 1.24 Moved from thumb.php to GlobalFunctions in 1.25
3606  */
3607 function wfThumbIsStandard( File $file, array $params ) {
3608         global $wgThumbLimits, $wgImageLimits, $wgResponsiveImages;
3609
3610         $multipliers = [ 1 ];
3611         if ( $wgResponsiveImages ) {
3612                 // These available sizes are hardcoded currently elsewhere in MediaWiki.
3613                 // @see Linker::processResponsiveImages
3614                 $multipliers[] = 1.5;
3615                 $multipliers[] = 2;
3616         }
3617
3618         $handler = $file->getHandler();
3619         if ( !$handler || !isset( $params['width'] ) ) {
3620                 return false;
3621         }
3622
3623         $basicParams = [];
3624         if ( isset( $params['page'] ) ) {
3625                 $basicParams['page'] = $params['page'];
3626         }
3627
3628         $thumbLimits = [];
3629         $imageLimits = [];
3630         // Expand limits to account for multipliers
3631         foreach ( $multipliers as $multiplier ) {
3632                 $thumbLimits = array_merge( $thumbLimits, array_map(
3633                         function ( $width ) use ( $multiplier ) {
3634                                 return round( $width * $multiplier );
3635                         }, $wgThumbLimits )
3636                 );
3637                 $imageLimits = array_merge( $imageLimits, array_map(
3638                         function ( $pair ) use ( $multiplier ) {
3639                                 return [
3640                                         round( $pair[0] * $multiplier ),
3641                                         round( $pair[1] * $multiplier ),
3642                                 ];
3643                         }, $wgImageLimits )
3644                 );
3645         }
3646
3647         // Check if the width matches one of $wgThumbLimits
3648         if ( in_array( $params['width'], $thumbLimits ) ) {
3649                 $normalParams = $basicParams + [ 'width' => $params['width'] ];
3650                 // Append any default values to the map (e.g. "lossy", "lossless", ...)
3651                 $handler->normaliseParams( $file, $normalParams );
3652         } else {
3653                 // If not, then check if the width matchs one of $wgImageLimits
3654                 $match = false;
3655                 foreach ( $imageLimits as $pair ) {
3656                         $normalParams = $basicParams + [ 'width' => $pair[0], 'height' => $pair[1] ];
3657                         // Decide whether the thumbnail should be scaled on width or height.
3658                         // Also append any default values to the map (e.g. "lossy", "lossless", ...)
3659                         $handler->normaliseParams( $file, $normalParams );
3660                         // Check if this standard thumbnail size maps to the given width
3661                         if ( $normalParams['width'] == $params['width'] ) {
3662                                 $match = true;
3663                                 break;
3664                         }
3665                 }
3666                 if ( !$match ) {
3667                         return false; // not standard for description pages
3668                 }
3669         }
3670
3671         // Check that the given values for non-page, non-width, params are just defaults
3672         foreach ( $params as $key => $value ) {
3673                 if ( !isset( $normalParams[$key] ) || $normalParams[$key] != $value ) {
3674                         return false;
3675                 }
3676         }
3677
3678         return true;
3679 }
3680
3681 /**
3682  * Merges two (possibly) 2 dimensional arrays into the target array ($baseArray).
3683  *
3684  * Values that exist in both values will be combined with += (all values of the array
3685  * of $newValues will be added to the values of the array of $baseArray, while values,
3686  * that exists in both, the value of $baseArray will be used).
3687  *
3688  * @param array $baseArray The array where you want to add the values of $newValues to
3689  * @param array $newValues An array with new values
3690  * @return array The combined array
3691  * @since 1.26
3692  */
3693 function wfArrayPlus2d( array $baseArray, array $newValues ) {
3694         // First merge items that are in both arrays
3695         foreach ( $baseArray as $name => &$groupVal ) {
3696                 if ( isset( $newValues[$name] ) ) {
3697                         $groupVal += $newValues[$name];
3698                 }
3699         }
3700         // Now add items that didn't exist yet
3701         $baseArray += $newValues;
3702
3703         return $baseArray;
3704 }