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 // Hide compatibility functions from Doxygen
  28 /// @cond
  29
  30 /**
  31  * Compatibility functions
  32  *
  33  * We support PHP 5.3.3 and up.
  34  * Re-implementations of newer functions or functions in non-standard
  35  * PHP extensions may be included here.
  36  */
  37
  38 if ( !function_exists( 'mb_substr' ) ) {
  39         /**
  40          * @codeCoverageIgnore
  41          * @see Fallback::mb_substr
  42          * @return string
  43          */
  44         function mb_substr( $str, $start, $count = 'end' ) {
  45                 return Fallback::mb_substr( $str, $start, $count );
  46         }
  47
  48         /**
  49          * @codeCoverageIgnore
  50          * @see Fallback::mb_substr_split_unicode
  51          * @return int
  52          */
  53         function mb_substr_split_unicode( $str, $splitPos ) {
  54                 return Fallback::mb_substr_split_unicode( $str, $splitPos );
  55         }
  56 }
  57
  58 if ( !function_exists( 'mb_strlen' ) ) {
  59         /**
  60          * @codeCoverageIgnore
  61          * @see Fallback::mb_strlen
  62          * @return int
  63          */
  64         function mb_strlen( $str, $enc = '' ) {
  65                 return Fallback::mb_strlen( $str, $enc );
  66         }
  67 }
  68
  69 if ( !function_exists( 'mb_strpos' ) ) {
  70         /**
  71          * @codeCoverageIgnore
  72          * @see Fallback::mb_strpos
  73          * @return int
  74          */
  75         function mb_strpos( $haystack, $needle, $offset = 0, $encoding = '' ) {
  76                 return Fallback::mb_strpos( $haystack, $needle, $offset, $encoding );
  77         }
  78 }
  79
  80 if ( !function_exists( 'mb_strrpos' ) ) {
  81         /**
  82          * @codeCoverageIgnore
  83          * @see Fallback::mb_strrpos
  84          * @return int
  85          */
  86         function mb_strrpos( $haystack, $needle, $offset = 0, $encoding = '' ) {
  87                 return Fallback::mb_strrpos( $haystack, $needle, $offset, $encoding );
  88         }
  89 }
  90
  91 // gzdecode function only exists in PHP >= 5.4.0
  92 // http://php.net/gzdecode
  93 if ( !function_exists( 'gzdecode' ) ) {
  94         /**
  95          * @codeCoverageIgnore
  96          * @param string $data
  97          * @return string
  98          */
  99         function gzdecode( $data ) {
 100                 return gzinflate( substr( $data, 10, -8 ) );
 101         }
 102 }
 103
 104 // hash_equals function only exists in PHP >= 5.6.0
 105 // http://php.net/hash_equals
 106 if ( !function_exists( 'hash_equals' ) ) {
 107         /**
 108          * Check whether a user-provided string is equal to a fixed-length secret string
 109          * without revealing bytes of the secret string through timing differences.
 110          *
 111          * The usual way to compare strings (PHP's === operator or the underlying memcmp()
 112          * function in C) is to compare corresponding bytes and stop at the first difference,
 113          * which would take longer for a partial match than for a complete mismatch. This
 114          * is not secure when one of the strings (e.g. an HMAC or token) must remain secret
 115          * and the other may come from an attacker. Statistical analysis of timing measurements
 116          * over many requests may allow the attacker to guess the string's bytes one at a time
 117          * (and check his guesses) even if the timing differences are extremely small.
 118          *
 119          * When making such a security-sensitive comparison, it is essential that the sequence
 120          * in which instructions are executed and memory locations are accessed not depend on
 121          * the secret string's value. HOWEVER, for simplicity, we do not attempt to minimize
 122          * the inevitable leakage of the string's length. That is generally known anyway as
 123          * a chararacteristic of the hash function used to compute the secret value.
 124          *
 125          * Longer explanation: http://www.emerose.com/timing-attacks-explained
 126          *
 127          * @codeCoverageIgnore
 128          * @param string $known_string Fixed-length secret string to compare against
 129          * @param string $user_string User-provided string
 130          * @return bool True if the strings are the same, false otherwise
 131          */
 132         function hash_equals( $known_string, $user_string ) {
 133                 // Strict type checking as in PHP's native implementation
 134                 if ( !is_string( $known_string ) ) {
 135                         trigger_error( 'hash_equals(): Expected known_string to be a string, ' .
 136                                 gettype( $known_string ) . ' given', E_USER_WARNING );
 137
 138                         return false;
 139                 }
 140
 141                 if ( !is_string( $user_string ) ) {
 142                         trigger_error( 'hash_equals(): Expected user_string to be a string, ' .
 143                                 gettype( $user_string ) . ' given', E_USER_WARNING );
 144
 145                         return false;
 146                 }
 147
 148                 $known_string_len = strlen( $known_string );
 149                 if ( $known_string_len !== strlen( $user_string ) ) {
 150                         return false;
 151                 }
 152
 153                 $result = 0;
 154                 for ( $i = 0; $i < $known_string_len; $i++ ) {
 155                         $result |= ord( $known_string[$i] ) ^ ord( $user_string[$i] );
 156                 }
 157
 158                 return ( $result === 0 );
 159         }
 160 }
 161 /// @endcond
 162
 163 /**
 164  * Like array_diff( $a, $b ) except that it works with two-dimensional arrays.
 165  * @param array $a
 166  * @param array $b
 167  * @return array
 168  */
 169 function wfArrayDiff2( $a, $b ) {
 170         return array_udiff( $a, $b, 'wfArrayDiff2_cmp' );
 171 }
 172
 173 /**
 174  * @param array|string $a
 175  * @param array|string $b
 176  * @return int
 177  */
 178 function wfArrayDiff2_cmp( $a, $b ) {
 179         if ( is_string( $a ) && is_string( $b ) ) {
 180                 return strcmp( $a, $b );
 181         } elseif ( count( $a ) !== count( $b ) ) {
 182                 return count( $a ) < count( $b ) ? -1 : 1;
 183         } else {
 184                 reset( $a );
 185                 reset( $b );
 186                 while ( ( list( , $valueA ) = each( $a ) ) && ( list( , $valueB ) = each( $b ) ) ) {
 187                         $cmp = strcmp( $valueA, $valueB );
 188                         if ( $cmp !== 0 ) {
 189                                 return $cmp;
 190                         }
 191                 }
 192                 return 0;
 193         }
 194 }
 195
 196 /**
 197  * Appends to second array if $value differs from that in $default
 198  *
 199  * @param string|int $key
 200  * @param mixed $value
 201  * @param mixed $default
 202  * @param array $changed Array to alter
 203  * @throws MWException
 204  */
 205 function wfAppendToArrayIfNotDefault( $key, $value, $default, &$changed ) {
 206         if ( is_null( $changed ) ) {
 207                 throw new MWException( 'GlobalFunctions::wfAppendToArrayIfNotDefault got null' );
 208         }
 209         if ( $default[$key] !== $value ) {
 210                 $changed[$key] = $value;
 211         }
 212 }
 213
 214 /**
 215  * Merge arrays in the style of getUserPermissionsErrors, with duplicate removal
 216  * e.g.
 217  *      wfMergeErrorArrays(
 218  *              array( array( 'x' ) ),
 219  *              array( array( 'x', '2' ) ),
 220  *              array( array( 'x' ) ),
 221  *              array( array( 'y' ) )
 222  *      );
 223  * returns:
 224  *              array(
 225  *              array( 'x', '2' ),
 226  *              array( 'x' ),
 227  *              array( 'y' )
 228  *      )
 229  *
 230  * @param array $array1,...
 231  * @return array
 232  */
 233 function wfMergeErrorArrays( /*...*/ ) {
 234         $args = func_get_args();
 235         $out = array();
 236         foreach ( $args as $errors ) {
 237                 foreach ( $errors as $params ) {
 238                         # @todo FIXME: Sometimes get nested arrays for $params,
 239                         # which leads to E_NOTICEs
 240                         $spec = implode( "\t", $params );
 241                         $out[$spec] = $params;
 242                 }
 243         }
 244         return array_values( $out );
 245 }
 246
 247 /**
 248  * Insert array into another array after the specified *KEY*
 249  *
 250  * @param array $array The array.
 251  * @param array $insert The array to insert.
 252  * @param mixed $after The key to insert after
 253  * @return array
 254  */
 255 function wfArrayInsertAfter( array $array, array $insert, $after ) {
 256         // Find the offset of the element to insert after.
 257         $keys = array_keys( $array );
 258         $offsetByKey = array_flip( $keys );
 259
 260         $offset = $offsetByKey[$after];
 261
 262         // Insert at the specified offset
 263         $before = array_slice( $array, 0, $offset + 1, true );
 264         $after = array_slice( $array, $offset + 1, count( $array ) - $offset, true );
 265
 266         $output = $before + $insert + $after;
 267
 268         return $output;
 269 }
 270
 271 /**
 272  * Recursively converts the parameter (an object) to an array with the same data
 273  *
 274  * @param object|array $objOrArray
 275  * @param bool $recursive
 276  * @return array
 277  */
 278 function wfObjectToArray( $objOrArray, $recursive = true ) {
 279         $array = array();
 280         if ( is_object( $objOrArray ) ) {
 281                 $objOrArray = get_object_vars( $objOrArray );
 282         }
 283         foreach ( $objOrArray as $key => $value ) {
 284                 if ( $recursive && ( is_object( $value ) || is_array( $value ) ) ) {
 285                         $value = wfObjectToArray( $value );
 286                 }
 287
 288                 $array[$key] = $value;
 289         }
 290
 291         return $array;
 292 }
 293
 294 /**
 295  * Get a random decimal value between 0 and 1, in a way
 296  * not likely to give duplicate values for any realistic
 297  * number of articles.
 298  *
 299  * @return string
 300  */
 301 function wfRandom() {
 302         # The maximum random value is "only" 2^31-1, so get two random
 303         # values to reduce the chance of dupes
 304         $max = mt_getrandmax() + 1;
 305         $rand = number_format( ( mt_rand() * $max + mt_rand() ) / $max / $max, 12, '.', '' );
 306
 307         return $rand;
 308 }
 309
 310 /**
 311  * Get a random string containing a number of pseudo-random hex
 312  * characters.
 313  * @note This is not secure, if you are trying to generate some sort
 314  *       of token please use MWCryptRand instead.
 315  *
 316  * @param int $length The length of the string to generate
 317  * @return string
 318  * @since 1.20
 319  */
 320 function wfRandomString( $length = 32 ) {
 321         $str = '';
 322         for ( $n = 0; $n < $length; $n += 7 ) {
 323                 $str .= sprintf( '%07x', mt_rand() & 0xfffffff );
 324         }
 325         return substr( $str, 0, $length );
 326 }
 327
 328 /**
 329  * We want some things to be included as literal characters in our title URLs
 330  * for prettiness, which urlencode encodes by default.  According to RFC 1738,
 331  * all of the following should be safe:
 332  *
 333  * ;:@&=$-_.+!*'(),
 334  *
 335  * But + is not safe because it's used to indicate a space; &= are only safe in
 336  * paths and not in queries (and we don't distinguish here); ' seems kind of
 337  * scary; and urlencode() doesn't touch -_. to begin with.  Plus, although /
 338  * is reserved, we don't care.  So the list we unescape is:
 339  *
 340  * ;:@$!*(),/
 341  *
 342  * However, IIS7 redirects fail when the url contains a colon (Bug 22709),
 343  * so no fancy : for IIS7.
 344  *
 345  * %2F in the page titles seems to fatally break for some reason.
 346  *
 347  * @param string $s
 348  * @return string
 349  */
 350 function wfUrlencode( $s ) {
 351         static $needle;
 352
 353         if ( is_null( $s ) ) {
 354                 $needle = null;
 355                 return '';
 356         }
 357
 358         if ( is_null( $needle ) ) {
 359                 $needle = array( '%3B', '%40', '%24', '%21', '%2A', '%28', '%29', '%2C', '%2F' );
 360                 if ( !isset( $_SERVER['SERVER_SOFTWARE'] ) ||
 361                         ( strpos( $_SERVER['SERVER_SOFTWARE'], 'Microsoft-IIS/7' ) === false )
 362                 ) {
 363                         $needle[] = '%3A';
 364                 }
 365         }
 366
 367         $s = urlencode( $s );
 368         $s = str_ireplace(
 369                 $needle,
 370                 array( ';', '@', '$', '!', '*', '(', ')', ',', '/', ':' ),
 371                 $s
 372         );
 373
 374         return $s;
 375 }
 376
 377 /**
 378  * This function takes two arrays as input, and returns a CGI-style string, e.g.
 379  * "days=7&limit=100". Options in the first array override options in the second.
 380  * Options set to null or false will not be output.
 381  *
 382  * @param array $array1 ( String|Array )
 383  * @param array $array2 ( String|Array )
 384  * @param string $prefix
 385  * @return string
 386  */
 387 function wfArrayToCgi( $array1, $array2 = null, $prefix = '' ) {
 388         if ( !is_null( $array2 ) ) {
 389                 $array1 = $array1 + $array2;
 390         }
 391
 392         $cgi = '';
 393         foreach ( $array1 as $key => $value ) {
 394                 if ( !is_null( $value ) && $value !== false ) {
 395                         if ( $cgi != '' ) {
 396                                 $cgi .= '&';
 397                         }
 398                         if ( $prefix !== '' ) {
 399                                 $key = $prefix . "[$key]";
 400                         }
 401                         if ( is_array( $value ) ) {
 402                                 $firstTime = true;
 403                                 foreach ( $value as $k => $v ) {
 404                                         $cgi .= $firstTime ? '' : '&';
 405                                         if ( is_array( $v ) ) {
 406                                                 $cgi .= wfArrayToCgi( $v, null, $key . "[$k]" );
 407                                         } else {
 408                                                 $cgi .= urlencode( $key . "[$k]" ) . '=' . urlencode( $v );
 409                                         }
 410                                         $firstTime = false;
 411                                 }
 412                         } else {
 413                                 if ( is_object( $value ) ) {
 414                                         $value = $value->__toString();
 415                                 }
 416                                 $cgi .= urlencode( $key ) . '=' . urlencode( $value );
 417                         }
 418                 }
 419         }
 420         return $cgi;
 421 }
 422
 423 /**
 424  * This is the logical opposite of wfArrayToCgi(): it accepts a query string as
 425  * its argument and returns the same string in array form.  This allows compatibility
 426  * with legacy functions that accept raw query strings instead of nice
 427  * arrays.  Of course, keys and values are urldecode()d.
 428  *
 429  * @param string $query Query string
 430  * @return string[] Array version of input
 431  */
 432 function wfCgiToArray( $query ) {
 433         if ( isset( $query[0] ) && $query[0] == '?' ) {
 434                 $query = substr( $query, 1 );
 435         }
 436         $bits = explode( '&', $query );
 437         $ret = array();
 438         foreach ( $bits as $bit ) {
 439                 if ( $bit === '' ) {
 440                         continue;
 441                 }
 442                 if ( strpos( $bit, '=' ) === false ) {
 443                         // Pieces like &qwerty become 'qwerty' => '' (at least this is what php does)
 444                         $key = $bit;
 445                         $value = '';
 446                 } else {
 447                         list( $key, $value ) = explode( '=', $bit );
 448                 }
 449                 $key = urldecode( $key );
 450                 $value = urldecode( $value );
 451                 if ( strpos( $key, '[' ) !== false ) {
 452                         $keys = array_reverse( explode( '[', $key ) );
 453                         $key = array_pop( $keys );
 454                         $temp = $value;
 455                         foreach ( $keys as $k ) {
 456                                 $k = substr( $k, 0, -1 );
 457                                 $temp = array( $k => $temp );
 458                         }
 459                         if ( isset( $ret[$key] ) ) {
 460                                 $ret[$key] = array_merge( $ret[$key], $temp );
 461                         } else {
 462                                 $ret[$key] = $temp;
 463                         }
 464                 } else {
 465                         $ret[$key] = $value;
 466                 }
 467         }
 468         return $ret;
 469 }
 470
 471 /**
 472  * Append a query string to an existing URL, which may or may not already
 473  * have query string parameters already. If so, they will be combined.
 474  *
 475  * @param string $url
 476  * @param string|string[] $query String or associative array
 477  * @return string
 478  */
 479 function wfAppendQuery( $url, $query ) {
 480         if ( is_array( $query ) ) {
 481                 $query = wfArrayToCgi( $query );
 482         }
 483         if ( $query != '' ) {
 484                 if ( false === strpos( $url, '?' ) ) {
 485                         $url .= '?';
 486                 } else {
 487                         $url .= '&';
 488                 }
 489                 $url .= $query;
 490         }
 491         return $url;
 492 }
 493
 494 /**
 495  * Expand a potentially local URL to a fully-qualified URL. Assumes $wgServer
 496  * is correct.
 497  *
 498  * The meaning of the PROTO_* constants is as follows:
 499  * PROTO_HTTP: Output a URL starting with http://
 500  * PROTO_HTTPS: Output a URL starting with https://
 501  * PROTO_RELATIVE: Output a URL starting with // (protocol-relative URL)
 502  * PROTO_CURRENT: Output a URL starting with either http:// or https:// , depending
 503  *    on which protocol was used for the current incoming request
 504  * PROTO_CANONICAL: For URLs without a domain, like /w/index.php , use $wgCanonicalServer.
 505  *    For protocol-relative URLs, use the protocol of $wgCanonicalServer
 506  * PROTO_INTERNAL: Like PROTO_CANONICAL, but uses $wgInternalServer instead of $wgCanonicalServer
 507  *
 508  * @todo this won't work with current-path-relative URLs
 509  * like "subdir/foo.html", etc.
 510  *
 511  * @param string $url Either fully-qualified or a local path + query
 512  * @param string $defaultProto One of the PROTO_* constants. Determines the
 513  *    protocol to use if $url or $wgServer is protocol-relative
 514  * @return string Fully-qualified URL, current-path-relative URL or false if
 515  *    no valid URL can be constructed
 516  */
 517 function wfExpandUrl( $url, $defaultProto = PROTO_CURRENT ) {
 518         global $wgServer, $wgCanonicalServer, $wgInternalServer, $wgRequest,
 519                 $wgHttpsPort;
 520         if ( $defaultProto === PROTO_CANONICAL ) {
 521                 $serverUrl = $wgCanonicalServer;
 522         } elseif ( $defaultProto === PROTO_INTERNAL && $wgInternalServer !== false ) {
 523                 // Make $wgInternalServer fall back to $wgServer if not set
 524                 $serverUrl = $wgInternalServer;
 525         } else {
 526                 $serverUrl = $wgServer;
 527                 if ( $defaultProto === PROTO_CURRENT ) {
 528                         $defaultProto = $wgRequest->getProtocol() . '://';
 529                 }
 530         }
 531
 532         // Analyze $serverUrl to obtain its protocol
 533         $bits = wfParseUrl( $serverUrl );
 534         $serverHasProto = $bits && $bits['scheme'] != '';
 535
 536         if ( $defaultProto === PROTO_CANONICAL || $defaultProto === PROTO_INTERNAL ) {
 537                 if ( $serverHasProto ) {
 538                         $defaultProto = $bits['scheme'] . '://';
 539                 } else {
 540                         // $wgCanonicalServer or $wgInternalServer doesn't have a protocol.
 541                         // This really isn't supposed to happen. Fall back to HTTP in this
 542                         // ridiculous case.
 543                         $defaultProto = PROTO_HTTP;
 544                 }
 545         }
 546
 547         $defaultProtoWithoutSlashes = substr( $defaultProto, 0, -2 );
 548
 549         if ( substr( $url, 0, 2 ) == '//' ) {
 550                 $url = $defaultProtoWithoutSlashes . $url;
 551         } elseif ( substr( $url, 0, 1 ) == '/' ) {
 552                 // If $serverUrl is protocol-relative, prepend $defaultProtoWithoutSlashes,
 553                 // otherwise leave it alone.
 554                 $url = ( $serverHasProto ? '' : $defaultProtoWithoutSlashes ) . $serverUrl . $url;
 555         }
 556
 557         $bits = wfParseUrl( $url );
 558
 559         // ensure proper port for HTTPS arrives in URL
 560         // https://bugzilla.wikimedia.org/show_bug.cgi?id=65184
 561         if ( $defaultProto === PROTO_HTTPS && $wgHttpsPort != 443 ) {
 562                 $bits['port'] = $wgHttpsPort;
 563         }
 564
 565         if ( $bits && isset( $bits['path'] ) ) {
 566                 $bits['path'] = wfRemoveDotSegments( $bits['path'] );
 567                 return wfAssembleUrl( $bits );
 568         } elseif ( $bits ) {
 569                 # No path to expand
 570                 return $url;
 571         } elseif ( substr( $url, 0, 1 ) != '/' ) {
 572                 # URL is a relative path
 573                 return wfRemoveDotSegments( $url );
 574         }
 575
 576         # Expanded URL is not valid.
 577         return false;
 578 }
 579
 580 /**
 581  * This function will reassemble a URL parsed with wfParseURL.  This is useful
 582  * if you need to edit part of a URL and put it back together.
 583  *
 584  * This is the basic structure used (brackets contain keys for $urlParts):
 585  * [scheme][delimiter][user]:[pass]@[host]:[port][path]?[query]#[fragment]
 586  *
 587  * @todo Need to integrate this into wfExpandUrl (bug 32168)
 588  *
 589  * @since 1.19
 590  * @param array $urlParts URL parts, as output from wfParseUrl
 591  * @return string URL assembled from its component parts
 592  */
 593 function wfAssembleUrl( $urlParts ) {
 594         $result = '';
 595
 596         if ( isset( $urlParts['delimiter'] ) ) {
 597                 if ( isset( $urlParts['scheme'] ) ) {
 598                         $result .= $urlParts['scheme'];
 599                 }
 600
 601                 $result .= $urlParts['delimiter'];
 602         }
 603
 604         if ( isset( $urlParts['host'] ) ) {
 605                 if ( isset( $urlParts['user'] ) ) {
 606                         $result .= $urlParts['user'];
 607                         if ( isset( $urlParts['pass'] ) ) {
 608                                 $result .= ':' . $urlParts['pass'];
 609                         }
 610                         $result .= '@';
 611                 }
 612
 613                 $result .= $urlParts['host'];
 614
 615                 if ( isset( $urlParts['port'] ) ) {
 616                         $result .= ':' . $urlParts['port'];
 617                 }
 618         }
 619
 620         if ( isset( $urlParts['path'] ) ) {
 621                 $result .= $urlParts['path'];
 622         }
 623
 624         if ( isset( $urlParts['query'] ) ) {
 625                 $result .= '?' . $urlParts['query'];
 626         }
 627
 628         if ( isset( $urlParts['fragment'] ) ) {
 629                 $result .= '#' . $urlParts['fragment'];
 630         }
 631
 632         return $result;
 633 }
 634
 635 /**
 636  * Remove all dot-segments in the provided URL path.  For example,
 637  * '/a/./b/../c/' becomes '/a/c/'.  For details on the algorithm, please see
 638  * RFC3986 section 5.2.4.
 639  *
 640  * @todo Need to integrate this into wfExpandUrl (bug 32168)
 641  *
 642  * @param string $urlPath URL path, potentially containing dot-segments
 643  * @return string URL path with all dot-segments removed
 644  */
 645 function wfRemoveDotSegments( $urlPath ) {
 646         $output = '';
 647         $inputOffset = 0;
 648         $inputLength = strlen( $urlPath );
 649
 650         while ( $inputOffset < $inputLength ) {
 651                 $prefixLengthOne = substr( $urlPath, $inputOffset, 1 );
 652                 $prefixLengthTwo = substr( $urlPath, $inputOffset, 2 );
 653                 $prefixLengthThree = substr( $urlPath, $inputOffset, 3 );
 654                 $prefixLengthFour = substr( $urlPath, $inputOffset, 4 );
 655                 $trimOutput = false;
 656
 657                 if ( $prefixLengthTwo == './' ) {
 658                         # Step A, remove leading "./"
 659                         $inputOffset += 2;
 660                 } elseif ( $prefixLengthThree == '../' ) {
 661                         # Step A, remove leading "../"
 662                         $inputOffset += 3;
 663                 } elseif ( ( $prefixLengthTwo == '/.' ) && ( $inputOffset + 2 == $inputLength ) ) {
 664                         # Step B, replace leading "/.$" with "/"
 665                         $inputOffset += 1;
 666                         $urlPath[$inputOffset] = '/';
 667                 } elseif ( $prefixLengthThree == '/./' ) {
 668                         # Step B, replace leading "/./" with "/"
 669                         $inputOffset += 2;
 670                 } elseif ( $prefixLengthThree == '/..' && ( $inputOffset + 3 == $inputLength ) ) {
 671                         # Step C, replace leading "/..$" with "/" and
 672                         # remove last path component in output
 673                         $inputOffset += 2;
 674                         $urlPath[$inputOffset] = '/';
 675                         $trimOutput = true;
 676                 } elseif ( $prefixLengthFour == '/../' ) {
 677                         # Step C, replace leading "/../" with "/" and
 678                         # remove last path component in output
 679                         $inputOffset += 3;
 680                         $trimOutput = true;
 681                 } elseif ( ( $prefixLengthOne == '.' ) && ( $inputOffset + 1 == $inputLength ) ) {
 682                         # Step D, remove "^.$"
 683                         $inputOffset += 1;
 684                 } elseif ( ( $prefixLengthTwo == '..' ) && ( $inputOffset + 2 == $inputLength ) ) {
 685                         # Step D, remove "^..$"
 686                         $inputOffset += 2;
 687                 } else {
 688                         # Step E, move leading path segment to output
 689                         if ( $prefixLengthOne == '/' ) {
 690                                 $slashPos = strpos( $urlPath, '/', $inputOffset + 1 );
 691                         } else {
 692                                 $slashPos = strpos( $urlPath, '/', $inputOffset );
 693                         }
 694                         if ( $slashPos === false ) {
 695                                 $output .= substr( $urlPath, $inputOffset );
 696                                 $inputOffset = $inputLength;
 697                         } else {
 698                                 $output .= substr( $urlPath, $inputOffset, $slashPos - $inputOffset );
 699                                 $inputOffset += $slashPos - $inputOffset;
 700                         }
 701                 }
 702
 703                 if ( $trimOutput ) {
 704                         $slashPos = strrpos( $output, '/' );
 705                         if ( $slashPos === false ) {
 706                                 $output = '';
 707                         } else {
 708                                 $output = substr( $output, 0, $slashPos );
 709                         }
 710                 }
 711         }
 712
 713         return $output;
 714 }
 715
 716 /**
 717  * Returns a regular expression of url protocols
 718  *
 719  * @param bool $includeProtocolRelative If false, remove '//' from the returned protocol list.
 720  *        DO NOT USE this directly, use wfUrlProtocolsWithoutProtRel() instead
 721  * @return string
 722  */
 723 function wfUrlProtocols( $includeProtocolRelative = true ) {
 724         global $wgUrlProtocols;
 725
 726         // Cache return values separately based on $includeProtocolRelative
 727         static $withProtRel = null, $withoutProtRel = null;
 728         $cachedValue = $includeProtocolRelative ? $withProtRel : $withoutProtRel;
 729         if ( !is_null( $cachedValue ) ) {
 730                 return $cachedValue;
 731         }
 732
 733         // Support old-style $wgUrlProtocols strings, for backwards compatibility
 734         // with LocalSettings files from 1.5
 735         if ( is_array( $wgUrlProtocols ) ) {
 736                 $protocols = array();
 737                 foreach ( $wgUrlProtocols as $protocol ) {
 738                         // Filter out '//' if !$includeProtocolRelative
 739                         if ( $includeProtocolRelative || $protocol !== '//' ) {
 740                                 $protocols[] = preg_quote( $protocol, '/' );
 741                         }
 742                 }
 743
 744                 $retval = implode( '|', $protocols );
 745         } else {
 746                 // Ignore $includeProtocolRelative in this case
 747                 // This case exists for pre-1.6 compatibility, and we can safely assume
 748                 // that '//' won't appear in a pre-1.6 config because protocol-relative
 749                 // URLs weren't supported until 1.18
 750                 $retval = $wgUrlProtocols;
 751         }
 752
 753         // Cache return value
 754         if ( $includeProtocolRelative ) {
 755                 $withProtRel = $retval;
 756         } else {
 757                 $withoutProtRel = $retval;
 758         }
 759         return $retval;
 760 }
 761
 762 /**
 763  * Like wfUrlProtocols(), but excludes '//' from the protocol list. Use this if
 764  * you need a regex that matches all URL protocols but does not match protocol-
 765  * relative URLs
 766  * @return string
 767  */
 768 function wfUrlProtocolsWithoutProtRel() {
 769         return wfUrlProtocols( false );
 770 }
 771
 772 /**
 773  * parse_url() work-alike, but non-broken.  Differences:
 774  *
 775  * 1) Does not raise warnings on bad URLs (just returns false).
 776  * 2) Handles protocols that don't use :// (e.g., mailto: and news:, as well as
 777  *    protocol-relative URLs) correctly.
 778  * 3) Adds a "delimiter" element to the array, either '://', ':' or '//' (see (2)).
 779  *
 780  * @param string $url A URL to parse
 781  * @return string[] Bits of the URL in an associative array, per PHP docs
 782  */
 783 function wfParseUrl( $url ) {
 784         global $wgUrlProtocols; // Allow all protocols defined in DefaultSettings/LocalSettings.php
 785
 786         // Protocol-relative URLs are handled really badly by parse_url(). It's so
 787         // bad that the easiest way to handle them is to just prepend 'http:' and
 788         // strip the protocol out later.
 789         $wasRelative = substr( $url, 0, 2 ) == '//';
 790         if ( $wasRelative ) {
 791                 $url = "http:$url";
 792         }
 793         wfSuppressWarnings();
 794         $bits = parse_url( $url );
 795         wfRestoreWarnings();
 796         // parse_url() returns an array without scheme for some invalid URLs, e.g.
 797         // parse_url("%0Ahttp://example.com") == array( 'host' => '%0Ahttp', 'path' => 'example.com' )
 798         if ( !$bits || !isset( $bits['scheme'] ) ) {
 799                 return false;
 800         }
 801
 802         // parse_url() incorrectly handles schemes case-sensitively. Convert it to lowercase.
 803         $bits['scheme'] = strtolower( $bits['scheme'] );
 804
 805         // most of the protocols are followed by ://, but mailto: and sometimes news: not, check for it
 806         if ( in_array( $bits['scheme'] . '://', $wgUrlProtocols ) ) {
 807                 $bits['delimiter'] = '://';
 808         } elseif ( in_array( $bits['scheme'] . ':', $wgUrlProtocols ) ) {
 809                 $bits['delimiter'] = ':';
 810                 // parse_url detects for news: and mailto: the host part of an url as path
 811                 // We have to correct this wrong detection
 812                 if ( isset( $bits['path'] ) ) {
 813                         $bits['host'] = $bits['path'];
 814                         $bits['path'] = '';
 815                 }
 816         } else {
 817                 return false;
 818         }
 819
 820         /* Provide an empty host for eg. file:/// urls (see bug 28627) */
 821         if ( !isset( $bits['host'] ) ) {
 822                 $bits['host'] = '';
 823
 824                 // bug 45069
 825                 if ( isset( $bits['path'] ) ) {
 826                         /* parse_url loses the third / for file:///c:/ urls (but not on variants) */
 827                         if ( substr( $bits['path'], 0, 1 ) !== '/' ) {
 828                                 $bits['path'] = '/' . $bits['path'];
 829                         }
 830                 } else {
 831                         $bits['path'] = '';
 832                 }
 833         }
 834
 835         // If the URL was protocol-relative, fix scheme and delimiter
 836         if ( $wasRelative ) {
 837                 $bits['scheme'] = '';
 838                 $bits['delimiter'] = '//';
 839         }
 840         return $bits;
 841 }
 842
 843 /**
 844  * Take a URL, make sure it's expanded to fully qualified, and replace any
 845  * encoded non-ASCII Unicode characters with their UTF-8 original forms
 846  * for more compact display and legibility for local audiences.
 847  *
 848  * @todo handle punycode domains too
 849  *
 850  * @param string $url
 851  * @return string
 852  */
 853 function wfExpandIRI( $url ) {
 854         return preg_replace_callback(
 855                 '/((?:%[89A-F][0-9A-F])+)/i',
 856                 'wfExpandIRI_callback',
 857                 wfExpandUrl( $url )
 858         );
 859 }
 860
 861 /**
 862  * Private callback for wfExpandIRI
 863  * @param array $matches
 864  * @return string
 865  */
 866 function wfExpandIRI_callback( $matches ) {
 867         return urldecode( $matches[1] );
 868 }
 869
 870 /**
 871  * Make URL indexes, appropriate for the el_index field of externallinks.
 872  *
 873  * @param string $url
 874  * @return array
 875  */
 876 function wfMakeUrlIndexes( $url ) {
 877         $bits = wfParseUrl( $url );
 878
 879         // Reverse the labels in the hostname, convert to lower case
 880         // For emails reverse domainpart only
 881         if ( $bits['scheme'] == 'mailto' ) {
 882                 $mailparts = explode( '@', $bits['host'], 2 );
 883                 if ( count( $mailparts ) === 2 ) {
 884                         $domainpart = strtolower( implode( '.', array_reverse( explode( '.', $mailparts[1] ) ) ) );
 885                 } else {
 886                         // No domain specified, don't mangle it
 887                         $domainpart = '';
 888                 }
 889                 $reversedHost = $domainpart . '@' . $mailparts[0];
 890         } else {
 891                 $reversedHost = strtolower( implode( '.', array_reverse( explode( '.', $bits['host'] ) ) ) );
 892         }
 893         // Add an extra dot to the end
 894         // Why? Is it in wrong place in mailto links?
 895         if ( substr( $reversedHost, -1, 1 ) !== '.' ) {
 896                 $reversedHost .= '.';
 897         }
 898         // Reconstruct the pseudo-URL
 899         $prot = $bits['scheme'];
 900         $index = $prot . $bits['delimiter'] . $reversedHost;
 901         // Leave out user and password. Add the port, path, query and fragment
 902         if ( isset( $bits['port'] ) ) {
 903                 $index .= ':' . $bits['port'];
 904         }
 905         if ( isset( $bits['path'] ) ) {
 906                 $index .= $bits['path'];
 907         } else {
 908                 $index .= '/';
 909         }
 910         if ( isset( $bits['query'] ) ) {
 911                 $index .= '?' . $bits['query'];
 912         }
 913         if ( isset( $bits['fragment'] ) ) {
 914                 $index .= '#' . $bits['fragment'];
 915         }
 916
 917         if ( $prot == '' ) {
 918                 return array( "http:$index", "https:$index" );
 919         } else {
 920                 return array( $index );
 921         }
 922 }
 923
 924 /**
 925  * Check whether a given URL has a domain that occurs in a given set of domains
 926  * @param string $url URL
 927  * @param array $domains Array of domains (strings)
 928  * @return bool True if the host part of $url ends in one of the strings in $domains
 929  */
 930 function wfMatchesDomainList( $url, $domains ) {
 931         $bits = wfParseUrl( $url );
 932         if ( is_array( $bits ) && isset( $bits['host'] ) ) {
 933                 $host = '.' . $bits['host'];
 934                 foreach ( (array)$domains as $domain ) {
 935                         $domain = '.' . $domain;
 936                         if ( substr( $host, -strlen( $domain ) ) === $domain ) {
 937                                 return true;
 938                         }
 939                 }
 940         }
 941         return false;
 942 }
 943
 944 /**
 945  * Sends a line to the debug log if enabled or, optionally, to a comment in output.
 946  * In normal operation this is a NOP.
 947  *
 948  * Controlling globals:
 949  * $wgDebugLogFile - points to the log file
 950  * $wgDebugRawPage - if false, 'action=raw' hits will not result in debug output.
 951  * $wgDebugComments - if on, some debug items may appear in comments in the HTML output.
 952  *
 953  * @param string $text
 954  * @param string|bool $dest Destination of the message:
 955  *     - 'all': both to the log and HTML (debug toolbar or HTML comments)
 956  *     - 'log': only to the log and not in HTML
 957  *   For backward compatibility, it can also take a boolean:
 958  *     - true: same as 'all'
 959  *     - false: same as 'log'
 960  */
 961 function wfDebug( $text, $dest = 'all' ) {
 962         global $wgDebugRawPage, $wgDebugLogPrefix;
 963
 964         if ( !$wgDebugRawPage && wfIsDebugRawPage() ) {
 965                 return;
 966         }
 967
 968         // Turn $dest into a string if it's a boolean (for b/c)
 969         if ( $dest === true ) {
 970                 $dest = 'all';
 971         } elseif ( $dest === false ) {
 972                 $dest = 'log';
 973         }
 974
 975         $timer = wfDebugTimer();
 976         if ( $timer !== '' ) {
 977                 // Prepend elapsed request time and real memory usage to each line
 978                 $text = preg_replace( '/[^\n]/', $timer . '\0', $text, 1 );
 979         }
 980
 981         if ( $dest === 'all' ) {
 982                 MWDebug::debugMsg( $text );
 983         }
 984
 985         $ctx = array();
 986         if ( $wgDebugLogPrefix !== '' ) {
 987                 $ctx['prefix'] = $wgDebugLogPrefix;
 988         }
 989
 990         $logger = MWLogger::getInstance( 'wfDebug' );
 991         $logger->debug( rtrim( $text, "\n" ), $ctx );
 992 }
 993
 994 /**
 995  * Returns true if debug logging should be suppressed if $wgDebugRawPage = false
 996  * @return bool
 997  */
 998 function wfIsDebugRawPage() {
 999         static $cache;
1000         if ( $cache !== null ) {
1001                 return $cache;
1002         }
1003         # Check for raw action using $_GET not $wgRequest, since the latter might not be initialised yet
1004         if ( ( isset( $_GET['action'] ) && $_GET['action'] == 'raw' )
1005                 || (
1006                         isset( $_SERVER['SCRIPT_NAME'] )
1007                         && substr( $_SERVER['SCRIPT_NAME'], -8 ) == 'load.php'
1008                 )
1009         ) {
1010                 $cache = true;
1011         } else {
1012                 $cache = false;
1013         }
1014         return $cache;
1015 }
1016
1017 /**
1018  * Get microsecond timestamps for debug logs
1019  *
1020  * @return string
1021  */
1022 function wfDebugTimer() {
1023         global $wgDebugTimestamps, $wgRequestTime;
1024
1025         if ( !$wgDebugTimestamps ) {
1026                 return '';
1027         }
1028
1029         $prefix = sprintf( "%6.4f", microtime( true ) - $wgRequestTime );
1030         $mem = sprintf( "%5.1fM", ( memory_get_usage( true ) / ( 1024 * 1024 ) ) );
1031         return "$prefix $mem  ";
1032 }
1033
1034 /**
1035  * Send a line giving PHP memory usage.
1036  *
1037  * @param bool $exact Print exact byte values instead of kibibytes (default: false)
1038  */
1039 function wfDebugMem( $exact = false ) {
1040         $mem = memory_get_usage();
1041         if ( !$exact ) {
1042                 $mem = floor( $mem / 1024 ) . ' KiB';
1043         } else {
1044                 $mem .= ' B';
1045         }
1046         wfDebug( "Memory usage: $mem\n" );
1047 }
1048
1049 /**
1050  * Send a line to a supplementary debug log file, if configured, or main debug log if not.
1051  * To configure a supplementary log file, set $wgDebugLogGroups[$logGroup] to a string
1052  * filename or an associative array mapping 'destination' to the desired filename. The
1053  * associative array may also contain a 'sample' key with an integer value, specifying
1054  * a sampling factor.
1055  *
1056  * @since 1.23 support for sampling log messages via $wgDebugLogGroups.
1057  *
1058  * @param string $logGroup
1059  * @param string $text
1060  * @param string|bool $dest Destination of the message:
1061  *     - 'all': both to the log and HTML (debug toolbar or HTML comments)
1062  *     - 'log': only to the log and not in HTML
1063  *     - 'private': only to the specifc log if set in $wgDebugLogGroups and
1064  *       discarded otherwise
1065  *   For backward compatibility, it can also take a boolean:
1066  *     - true: same as 'all'
1067  *     - false: same as 'private'
1068  */
1069 function wfDebugLog( $logGroup, $text, $dest = 'all' ) {
1070         global $wgDebugLogGroups;
1071
1072         // Turn $dest into a string if it's a boolean (for b/c)
1073         if ( $dest === true ) {
1074                 $dest = 'all';
1075         } elseif ( $dest === false ) {
1076                 $dest = 'private';
1077         }
1078
1079         $text = trim( $text );
1080
1081         if ( $dest === 'all' ) {
1082                 MWDebug::debugMsg( "[{$logGroup}] {$text}\n" );
1083         }
1084
1085         $logger = MWLogger::getInstance( $logGroup );
1086         $logger->debug( $text, array(
1087                 'private' => ( $dest === 'private' ),
1088         ) );
1089 }
1090
1091 /**
1092  * Log for database errors
1093  *
1094  * @param string $text Database error message.
1095  */
1096 function wfLogDBError( $text ) {
1097         $logger = MWLogger::getInstance( 'wfLogDBError' );
1098         $logger->error( trim( $text ) );
1099 }
1100
1101 /**
1102  * Throws a warning that $function is deprecated
1103  *
1104  * @param string $function
1105  * @param string|bool $version Version of MediaWiki that the function
1106  *    was deprecated in (Added in 1.19).
1107  * @param string|bool $component Added in 1.19.
1108  * @param int $callerOffset How far up the call stack is the original
1109  *    caller. 2 = function that called the function that called
1110  *    wfDeprecated (Added in 1.20)
1111  *
1112  * @return null
1113  */
1114 function wfDeprecated( $function, $version = false, $component = false, $callerOffset = 2 ) {
1115         MWDebug::deprecated( $function, $version, $component, $callerOffset + 1 );
1116 }
1117
1118 /**
1119  * Send a warning either to the debug log or in a PHP error depending on
1120  * $wgDevelopmentWarnings. To log warnings in production, use wfLogWarning() instead.
1121  *
1122  * @param string $msg Message to send
1123  * @param int $callerOffset Number of items to go back in the backtrace to
1124  *        find the correct caller (1 = function calling wfWarn, ...)
1125  * @param int $level PHP error level; defaults to E_USER_NOTICE;
1126  *        only used when $wgDevelopmentWarnings is true
1127  */
1128 function wfWarn( $msg, $callerOffset = 1, $level = E_USER_NOTICE ) {
1129         MWDebug::warning( $msg, $callerOffset + 1, $level, 'auto' );
1130 }
1131
1132 /**
1133  * Send a warning as a PHP error and the debug log. This is intended for logging
1134  * warnings in production. For logging development warnings, use WfWarn instead.
1135  *
1136  * @param string $msg Message to send
1137  * @param int $callerOffset Number of items to go back in the backtrace to
1138  *        find the correct caller (1 = function calling wfLogWarning, ...)
1139  * @param int $level PHP error level; defaults to E_USER_WARNING
1140  */
1141 function wfLogWarning( $msg, $callerOffset = 1, $level = E_USER_WARNING ) {
1142         MWDebug::warning( $msg, $callerOffset + 1, $level, 'production' );
1143 }
1144
1145 /**
1146  * Log to a file without getting "file size exceeded" signals.
1147  *
1148  * Can also log to TCP or UDP with the syntax udp://host:port/prefix. This will
1149  * send lines to the specified port, prefixed by the specified prefix and a space.
1150  *
1151  * @param string $text
1152  * @param string $file Filename
1153  * @throws MWException
1154  */
1155 function wfErrorLog( $text, $file ) {
1156         $logger = MWLogger::getInstance( 'wfErrorLog' );
1157         $logger->info( trim( $text ), array(
1158                 'destination' => $file,
1159         ) );
1160 }
1161
1162 /**
1163  * @todo document
1164  */
1165 function wfLogProfilingData() {
1166         global $wgRequestTime, $wgDebugLogFile, $wgDebugLogGroups, $wgDebugRawPage;
1167         global $wgProfileLimit, $wgUser, $wgRequest;
1168
1169         StatCounter::singleton()->flush();
1170
1171         $profiler = Profiler::instance();
1172
1173         # Profiling must actually be enabled...
1174         if ( $profiler->isStub() ) {
1175                 return;
1176         }
1177
1178         // Get total page request time and only show pages that longer than
1179         // $wgProfileLimit time (default is 0)
1180         $elapsed = microtime( true ) - $wgRequestTime;
1181         if ( $elapsed <= $wgProfileLimit ) {
1182                 return;
1183         }
1184
1185         $profiler->logData();
1186
1187         // Check whether this should be logged in the debug file.
1188         if ( isset( $wgDebugLogGroups['profileoutput'] )
1189                 && $wgDebugLogGroups['profileoutput'] === false
1190         ) {
1191                 // Explicitely disabled
1192                 return;
1193         }
1194         if ( !isset( $wgDebugLogGroups['profileoutput'] ) && $wgDebugLogFile == '' ) {
1195                 // Logging not enabled; no point going further
1196                 return;
1197         }
1198         if ( !$wgDebugRawPage && wfIsDebugRawPage() ) {
1199                 return;
1200         }
1201
1202         $forward = '';
1203         if ( !empty( $_SERVER['HTTP_X_FORWARDED_FOR'] ) ) {
1204                 $forward = ' forwarded for ' . $_SERVER['HTTP_X_FORWARDED_FOR'];
1205         }
1206         if ( !empty( $_SERVER['HTTP_CLIENT_IP'] ) ) {
1207                 $forward .= ' client IP ' . $_SERVER['HTTP_CLIENT_IP'];
1208         }
1209         if ( !empty( $_SERVER['HTTP_FROM'] ) ) {
1210                 $forward .= ' from ' . $_SERVER['HTTP_FROM'];
1211         }
1212         if ( $forward ) {
1213                 $forward = "\t(proxied via {$_SERVER['REMOTE_ADDR']}{$forward})";
1214         }
1215         // Don't load $wgUser at this late stage just for statistics purposes
1216         // @todo FIXME: We can detect some anons even if it is not loaded. See User::getId()
1217         if ( $wgUser->isItemLoaded( 'id' ) && $wgUser->isAnon() ) {
1218                 $forward .= ' anon';
1219         }
1220
1221         // Command line script uses a FauxRequest object which does not have
1222         // any knowledge about an URL and throw an exception instead.
1223         try {
1224                 $requestUrl = $wgRequest->getRequestURL();
1225         } catch ( MWException $e ) {
1226                 $requestUrl = 'n/a';
1227         }
1228
1229         $log = sprintf( "%s\t%04.3f\t%s\n",
1230                 gmdate( 'YmdHis' ), $elapsed,
1231                 urldecode( $requestUrl . $forward ) );
1232
1233         wfDebugLog( 'profileoutput', $log . $profiler->getOutput() );
1234 }
1235
1236 /**
1237  * Increment a statistics counter
1238  *
1239  * @param string $key
1240  * @param int $count
1241  * @return void
1242  */
1243 function wfIncrStats( $key, $count = 1 ) {
1244         StatCounter::singleton()->incr( $key, $count );
1245 }
1246
1247 /**
1248  * Check whether the wiki is in read-only mode.
1249  *
1250  * @return bool
1251  */
1252 function wfReadOnly() {
1253         return wfReadOnlyReason() !== false;
1254 }
1255
1256 /**
1257  * Get the value of $wgReadOnly or the contents of $wgReadOnlyFile.
1258  *
1259  * @return string|bool String when in read-only mode; false otherwise
1260  */
1261 function wfReadOnlyReason() {
1262         global $wgReadOnly, $wgReadOnlyFile;
1263
1264         if ( $wgReadOnly === null ) {
1265                 // Set $wgReadOnly for faster access next time
1266                 if ( is_file( $wgReadOnlyFile ) && filesize( $wgReadOnlyFile ) > 0 ) {
1267                         $wgReadOnly = file_get_contents( $wgReadOnlyFile );
1268                 } else {
1269                         $wgReadOnly = false;
1270                 }
1271         }
1272
1273         return $wgReadOnly;
1274 }
1275
1276 /**
1277  * Return a Language object from $langcode
1278  *
1279  * @param Language|string|bool $langcode Either:
1280  *                  - a Language object
1281  *                  - code of the language to get the message for, if it is
1282  *                    a valid code create a language for that language, if
1283  *                    it is a string but not a valid code then make a basic
1284  *                    language object
1285  *                  - a boolean: if it's false then use the global object for
1286  *                    the current user's language (as a fallback for the old parameter
1287  *                    functionality), or if it is true then use global object
1288  *                    for the wiki's content language.
1289  * @return Language
1290  */
1291 function wfGetLangObj( $langcode = false ) {
1292         # Identify which language to get or create a language object for.
1293         # Using is_object here due to Stub objects.
1294         if ( is_object( $langcode ) ) {
1295                 # Great, we already have the object (hopefully)!
1296                 return $langcode;
1297         }
1298
1299         global $wgContLang, $wgLanguageCode;
1300         if ( $langcode === true || $langcode === $wgLanguageCode ) {
1301                 # $langcode is the language code of the wikis content language object.
1302                 # or it is a boolean and value is true
1303                 return $wgContLang;
1304         }
1305
1306         global $wgLang;
1307         if ( $langcode === false || $langcode === $wgLang->getCode() ) {
1308                 # $langcode is the language code of user language object.
1309                 # or it was a boolean and value is false
1310                 return $wgLang;
1311         }
1312
1313         $validCodes = array_keys( Language::fetchLanguageNames() );
1314         if ( in_array( $langcode, $validCodes ) ) {
1315                 # $langcode corresponds to a valid language.
1316                 return Language::factory( $langcode );
1317         }
1318
1319         # $langcode is a string, but not a valid language code; use content language.
1320         wfDebug( "Invalid language code passed to wfGetLangObj, falling back to content language.\n" );
1321         return $wgContLang;
1322 }
1323
1324 /**
1325  * This is the function for getting translated interface messages.
1326  *
1327  * @see Message class for documentation how to use them.
1328  * @see https://www.mediawiki.org/wiki/Manual:Messages_API
1329  *
1330  * This function replaces all old wfMsg* functions.
1331  *
1332  * @param string|string[] $key Message key, or array of keys
1333  * @param mixed $params,... Normal message parameters
1334  * @return Message
1335  *
1336  * @since 1.17
1337  *
1338  * @see Message::__construct
1339  */
1340 function wfMessage( $key /*...*/ ) {
1341         $params = func_get_args();
1342         array_shift( $params );
1343         if ( isset( $params[0] ) && is_array( $params[0] ) ) {
1344                 $params = $params[0];
1345         }
1346         return new Message( $key, $params );
1347 }
1348
1349 /**
1350  * This function accepts multiple message keys and returns a message instance
1351  * for the first message which is non-empty. If all messages are empty then an
1352  * instance of the first message key is returned.
1353  *
1354  * @param string|string[] $keys,... Message keys
1355  * @return Message
1356  *
1357  * @since 1.18
1358  *
1359  * @see Message::newFallbackSequence
1360  */
1361 function wfMessageFallback( /*...*/ ) {
1362         $args = func_get_args();
1363         return call_user_func_array( 'Message::newFallbackSequence', $args );
1364 }
1365
1366 /**
1367  * Get a message from anywhere, for the current user language.
1368  *
1369  * Use wfMsgForContent() instead if the message should NOT
1370  * change depending on the user preferences.
1371  *
1372  * @deprecated since 1.18
1373  *
1374  * @param string $key Lookup key for the message, usually
1375  *    defined in languages/Language.php
1376  *
1377  * Parameters to the message, which can be used to insert variable text into
1378  * it, can be passed to this function in the following formats:
1379  * - One per argument, starting at the second parameter
1380  * - As an array in the second parameter
1381  * These are not shown in the function definition.
1382  *
1383  * @return string
1384  */
1385 function wfMsg( $key ) {
1386         wfDeprecated( __METHOD__, '1.21' );
1387
1388         $args = func_get_args();
1389         array_shift( $args );
1390         return wfMsgReal( $key, $args );
1391 }
1392
1393 /**
1394  * Same as above except doesn't transform the message
1395  *
1396  * @deprecated since 1.18
1397  *
1398  * @param string $key
1399  * @return string
1400  */
1401 function wfMsgNoTrans( $key ) {
1402         wfDeprecated( __METHOD__, '1.21' );
1403
1404         $args = func_get_args();
1405         array_shift( $args );
1406         return wfMsgReal( $key, $args, true, false, false );
1407 }
1408
1409 /**
1410  * Get a message from anywhere, for the current global language
1411  * set with $wgLanguageCode.
1412  *
1413  * Use this if the message should NOT change dependent on the
1414  * language set in the user's preferences. This is the case for
1415  * most text written into logs, as well as link targets (such as
1416  * the name of the copyright policy page). Link titles, on the
1417  * other hand, should be shown in the UI language.
1418  *
1419  * Note that MediaWiki allows users to change the user interface
1420  * language in their preferences, but a single installation
1421  * typically only contains content in one language.
1422  *
1423  * Be wary of this distinction: If you use wfMsg() where you should
1424  * use wfMsgForContent(), a user of the software may have to
1425  * customize potentially hundreds of messages in
1426  * order to, e.g., fix a link in every possible language.
1427  *
1428  * @deprecated since 1.18
1429  *
1430  * @param string $key Lookup key for the message, usually
1431  *     defined in languages/Language.php
1432  * @return string
1433  */
1434 function wfMsgForContent( $key ) {
1435         wfDeprecated( __METHOD__, '1.21' );
1436
1437         global $wgForceUIMsgAsContentMsg;
1438         $args = func_get_args();
1439         array_shift( $args );
1440         $forcontent = true;
1441         if ( is_array( $wgForceUIMsgAsContentMsg )
1442                 && in_array( $key, $wgForceUIMsgAsContentMsg )
1443         ) {
1444                 $forcontent = false;
1445         }
1446         return wfMsgReal( $key, $args, true, $forcontent );
1447 }
1448
1449 /**
1450  * Same as above except doesn't transform the message
1451  *
1452  * @deprecated since 1.18
1453  *
1454  * @param string $key
1455  * @return string
1456  */
1457 function wfMsgForContentNoTrans( $key ) {
1458         wfDeprecated( __METHOD__, '1.21' );
1459
1460         global $wgForceUIMsgAsContentMsg;
1461         $args = func_get_args();
1462         array_shift( $args );
1463         $forcontent = true;
1464         if ( is_array( $wgForceUIMsgAsContentMsg )
1465                 && in_array( $key, $wgForceUIMsgAsContentMsg )
1466         ) {
1467                 $forcontent = false;
1468         }
1469         return wfMsgReal( $key, $args, true, $forcontent, false );
1470 }
1471
1472 /**
1473  * Really get a message
1474  *
1475  * @deprecated since 1.18
1476  *
1477  * @param string $key Key to get.
1478  * @param array $args
1479  * @param bool $useDB
1480  * @param string|bool $forContent Language code, or false for user lang, true for content lang.
1481  * @param bool $transform Whether or not to transform the message.
1482  * @return string The requested message.
1483  */
1484 function wfMsgReal( $key, $args, $useDB = true, $forContent = false, $transform = true ) {
1485         wfDeprecated( __METHOD__, '1.21' );
1486
1487         wfProfileIn( __METHOD__ );
1488         $message = wfMsgGetKey( $key, $useDB, $forContent, $transform );
1489         $message = wfMsgReplaceArgs( $message, $args );
1490         wfProfileOut( __METHOD__ );
1491         return $message;
1492 }
1493
1494 /**
1495  * Fetch a message string value, but don't replace any keys yet.
1496  *
1497  * @deprecated since 1.18
1498  *
1499  * @param string $key
1500  * @param bool $useDB
1501  * @param string|bool $langCode Code of the language to get the message for, or
1502  *   behaves as a content language switch if it is a boolean.
1503  * @param bool $transform Whether to parse magic words, etc.
1504  * @return string
1505  */
1506 function wfMsgGetKey( $key, $useDB = true, $langCode = false, $transform = true ) {
1507         wfDeprecated( __METHOD__, '1.21' );
1508
1509         wfRunHooks( 'NormalizeMessageKey', array( &$key, &$useDB, &$langCode, &$transform ) );
1510
1511         $cache = MessageCache::singleton();
1512         $message = $cache->get( $key, $useDB, $langCode );
1513         if ( $message === false ) {
1514                 $message = '&lt;' . htmlspecialchars( $key ) . '&gt;';
1515         } elseif ( $transform ) {
1516                 $message = $cache->transform( $message );
1517         }
1518         return $message;
1519 }
1520
1521 /**
1522  * Replace message parameter keys on the given formatted output.
1523  *
1524  * @param string $message
1525  * @param array $args
1526  * @return string
1527  * @private
1528  */
1529 function wfMsgReplaceArgs( $message, $args ) {
1530         # Fix windows line-endings
1531         # Some messages are split with explode("\n", $msg)
1532         $message = str_replace( "\r", '', $message );
1533
1534         // Replace arguments
1535         if ( count( $args ) ) {
1536                 if ( is_array( $args[0] ) ) {
1537                         $args = array_values( $args[0] );
1538                 }
1539                 $replacementKeys = array();
1540                 foreach ( $args as $n => $param ) {
1541                         $replacementKeys['$' . ( $n + 1 )] = $param;
1542                 }
1543                 $message = strtr( $message, $replacementKeys );
1544         }
1545
1546         return $message;
1547 }
1548
1549 /**
1550  * Return an HTML-escaped version of a message.
1551  * Parameter replacements, if any, are done *after* the HTML-escaping,
1552  * so parameters may contain HTML (eg links or form controls). Be sure
1553  * to pre-escape them if you really do want plaintext, or just wrap
1554  * the whole thing in htmlspecialchars().
1555  *
1556  * @deprecated since 1.18
1557  *
1558  * @param string $key
1559  * @param string $args,... Parameters
1560  * @return string
1561  */
1562 function wfMsgHtml( $key ) {
1563         wfDeprecated( __METHOD__, '1.21' );
1564
1565         $args = func_get_args();
1566         array_shift( $args );
1567         return wfMsgReplaceArgs( htmlspecialchars( wfMsgGetKey( $key ) ), $args );
1568 }
1569
1570 /**
1571  * Return an HTML version of message
1572  * Parameter replacements, if any, are done *after* parsing the wiki-text message,
1573  * so parameters may contain HTML (eg links or form controls). Be sure
1574  * to pre-escape them if you really do want plaintext, or just wrap
1575  * the whole thing in htmlspecialchars().
1576  *
1577  * @deprecated since 1.18
1578  *
1579  * @param string $key
1580  * @param string $args,... Parameters
1581  * @return string
1582  */
1583 function wfMsgWikiHtml( $key ) {
1584         wfDeprecated( __METHOD__, '1.21' );
1585
1586         $args = func_get_args();
1587         array_shift( $args );
1588         return wfMsgReplaceArgs(
1589                 MessageCache::singleton()->parse( wfMsgGetKey( $key ), null,
1590                 /* can't be set to false */ true, /* interface */ true )->getText(),
1591                 $args );
1592 }
1593
1594 /**
1595  * Returns message in the requested format
1596  *
1597  * @deprecated since 1.18
1598  *
1599  * @param string $key Key of the message
1600  * @param array $options Processing rules.
1601  *   Can take the following options:
1602  *     parse: parses wikitext to HTML
1603  *     parseinline: parses wikitext to HTML and removes the surrounding
1604  *       p's added by parser or tidy
1605  *     escape: filters message through htmlspecialchars
1606  *     escapenoentities: same, but allows entity references like &#160; through
1607  *     replaceafter: parameters are substituted after parsing or escaping
1608  *     parsemag: transform the message using magic phrases
1609  *     content: fetch message for content language instead of interface
1610  *   Also can accept a single associative argument, of the form 'language' => 'xx':
1611  *     language: Language object or language code to fetch message for
1612  *       (overridden by content).
1613  * Behavior for conflicting options (e.g., parse+parseinline) is undefined.
1614  *
1615  * @return string
1616  */
1617 function wfMsgExt( $key, $options ) {
1618         wfDeprecated( __METHOD__, '1.21' );
1619
1620         $args = func_get_args();
1621         array_shift( $args );
1622         array_shift( $args );
1623         $options = (array)$options;
1624
1625         foreach ( $options as $arrayKey => $option ) {
1626                 if ( !preg_match( '/^[0-9]+|language$/', $arrayKey ) ) {
1627                         # An unknown index, neither numeric nor "language"
1628                         wfWarn( "wfMsgExt called with incorrect parameter key $arrayKey", 1, E_USER_WARNING );
1629                 } elseif ( preg_match( '/^[0-9]+$/', $arrayKey ) && !in_array( $option,
1630                 array( 'parse', 'parseinline', 'escape', 'escapenoentities',
1631                 'replaceafter', 'parsemag', 'content' ) ) ) {
1632                         # A numeric index with unknown value
1633                         wfWarn( "wfMsgExt called with incorrect parameter $option", 1, E_USER_WARNING );
1634                 }
1635         }
1636
1637         if ( in_array( 'content', $options, true ) ) {
1638                 $forContent = true;
1639                 $langCode = true;
1640                 $langCodeObj = null;
1641         } elseif ( array_key_exists( 'language', $options ) ) {
1642                 $forContent = false;
1643                 $langCode = wfGetLangObj( $options['language'] );
1644                 $langCodeObj = $langCode;
1645         } else {
1646                 $forContent = false;
1647                 $langCode = false;
1648                 $langCodeObj = null;
1649         }
1650
1651         $string = wfMsgGetKey( $key, /*DB*/true, $langCode, /*Transform*/false );
1652
1653         if ( !in_array( 'replaceafter', $options, true ) ) {
1654                 $string = wfMsgReplaceArgs( $string, $args );
1655         }
1656
1657         $messageCache = MessageCache::singleton();
1658         $parseInline = in_array( 'parseinline', $options, true );
1659         if ( in_array( 'parse', $options, true ) || $parseInline ) {
1660                 $string = $messageCache->parse( $string, null, true, !$forContent, $langCodeObj );
1661                 if ( $string instanceof ParserOutput ) {
1662                         $string = $string->getText();
1663                 }
1664
1665                 if ( $parseInline ) {
1666                         $string = Parser::stripOuterParagraph( $string );
1667                 }
1668         } elseif ( in_array( 'parsemag', $options, true ) ) {
1669                 $string = $messageCache->transform( $string,
1670                                 !$forContent, $langCodeObj );
1671         }
1672
1673         if ( in_array( 'escape', $options, true ) ) {
1674                 $string = htmlspecialchars ( $string );
1675         } elseif ( in_array( 'escapenoentities', $options, true ) ) {
1676                 $string = Sanitizer::escapeHtmlAllowEntities( $string );
1677         }
1678
1679         if ( in_array( 'replaceafter', $options, true ) ) {
1680                 $string = wfMsgReplaceArgs( $string, $args );
1681         }
1682
1683         return $string;
1684 }
1685
1686 /**
1687  * Since wfMsg() and co suck, they don't return false if the message key they
1688  * looked up didn't exist but instead the key wrapped in <>'s, this function checks for the
1689  * nonexistence of messages by checking the MessageCache::get() result directly.
1690  *
1691  * @deprecated since 1.18. Use Message::isDisabled().
1692  *
1693  * @param string $key The message key looked up
1694  * @return bool True if the message *doesn't* exist.
1695  */
1696 function wfEmptyMsg( $key ) {
1697         wfDeprecated( __METHOD__, '1.21' );
1698
1699         return MessageCache::singleton()->get( $key, /*useDB*/true, /*content*/false ) === false;
1700 }
1701
1702 /**
1703  * Fetch server name for use in error reporting etc.
1704  * Use real server name if available, so we know which machine
1705  * in a server farm generated the current page.
1706  *
1707  * @return string
1708  */
1709 function wfHostname() {
1710         static $host;
1711         if ( is_null( $host ) ) {
1712
1713                 # Hostname overriding
1714                 global $wgOverrideHostname;
1715                 if ( $wgOverrideHostname !== false ) {
1716                         # Set static and skip any detection
1717                         $host = $wgOverrideHostname;
1718                         return $host;
1719                 }
1720
1721                 if ( function_exists( 'posix_uname' ) ) {
1722                         // This function not present on Windows
1723                         $uname = posix_uname();
1724                 } else {
1725                         $uname = false;
1726                 }
1727                 if ( is_array( $uname ) && isset( $uname['nodename'] ) ) {
1728                         $host = $uname['nodename'];
1729                 } elseif ( getenv( 'COMPUTERNAME' ) ) {
1730                         # Windows computer name
1731                         $host = getenv( 'COMPUTERNAME' );
1732                 } else {
1733                         # This may be a virtual server.
1734                         $host = $_SERVER['SERVER_NAME'];
1735                 }
1736         }
1737         return $host;
1738 }
1739
1740 /**
1741  * Returns a script tag that stores the amount of time it took MediaWiki to
1742  * handle the request in milliseconds as 'wgBackendResponseTime'.
1743  *
1744  * If $wgShowHostnames is true, the script will also set 'wgHostname' to the
1745  * hostname of the server handling the request.
1746  *
1747  * @return string
1748  */
1749 function wfReportTime() {
1750         global $wgRequestTime, $wgShowHostnames;
1751
1752         $responseTime = round( ( microtime( true ) - $wgRequestTime ) * 1000 );
1753         $reportVars = array( 'wgBackendResponseTime' => $responseTime );
1754         if ( $wgShowHostnames ) {
1755                 $reportVars['wgHostname'] = wfHostname();
1756         }
1757         return Skin::makeVariablesScript( $reportVars );
1758 }
1759
1760 /**
1761  * Safety wrapper for debug_backtrace().
1762  *
1763  * Will return an empty array if debug_backtrace is disabled, otherwise
1764  * the output from debug_backtrace() (trimmed).
1765  *
1766  * @param int $limit This parameter can be used to limit the number of stack frames returned
1767  *
1768  * @return array Array of backtrace information
1769  */
1770 function wfDebugBacktrace( $limit = 0 ) {
1771         static $disabled = null;
1772
1773         if ( is_null( $disabled ) ) {
1774                 $disabled = !function_exists( 'debug_backtrace' );
1775                 if ( $disabled ) {
1776                         wfDebug( "debug_backtrace() is disabled\n" );
1777                 }
1778         }
1779         if ( $disabled ) {
1780                 return array();
1781         }
1782
1783         if ( $limit && version_compare( PHP_VERSION, '5.4.0', '>=' ) ) {
1784                 return array_slice( debug_backtrace( DEBUG_BACKTRACE_PROVIDE_OBJECT, $limit + 1 ), 1 );
1785         } else {
1786                 return array_slice( debug_backtrace(), 1 );
1787         }
1788 }
1789
1790 /**
1791  * Get a debug backtrace as a string
1792  *
1793  * @return string
1794  */
1795 function wfBacktrace() {
1796         global $wgCommandLineMode;
1797
1798         if ( $wgCommandLineMode ) {
1799                 $msg = '';
1800         } else {
1801                 $msg = "<ul>\n";
1802         }
1803         $backtrace = wfDebugBacktrace();
1804         foreach ( $backtrace as $call ) {
1805                 if ( isset( $call['file'] ) ) {
1806                         $f = explode( DIRECTORY_SEPARATOR, $call['file'] );
1807                         $file = $f[count( $f ) - 1];
1808                 } else {
1809                         $file = '-';
1810                 }
1811                 if ( isset( $call['line'] ) ) {
1812                         $line = $call['line'];
1813                 } else {
1814                         $line = '-';
1815                 }
1816                 if ( $wgCommandLineMode ) {
1817                         $msg .= "$file line $line calls ";
1818                 } else {
1819                         $msg .= '<li>' . $file . ' line ' . $line . ' calls ';
1820                 }
1821                 if ( !empty( $call['class'] ) ) {
1822                         $msg .= $call['class'] . $call['type'];
1823                 }
1824                 $msg .= $call['function'] . '()';
1825
1826                 if ( $wgCommandLineMode ) {
1827                         $msg .= "\n";
1828                 } else {
1829                         $msg .= "</li>\n";
1830                 }
1831         }
1832         if ( $wgCommandLineMode ) {
1833                 $msg .= "\n";
1834         } else {
1835                 $msg .= "</ul>\n";
1836         }
1837
1838         return $msg;
1839 }
1840
1841 /**
1842  * Get the name of the function which called this function
1843  * wfGetCaller( 1 ) is the function with the wfGetCaller() call (ie. __FUNCTION__)
1844  * wfGetCaller( 2 ) [default] is the caller of the function running wfGetCaller()
1845  * wfGetCaller( 3 ) is the parent of that.
1846  *
1847  * @param int $level
1848  * @return string
1849  */
1850 function wfGetCaller( $level = 2 ) {
1851         $backtrace = wfDebugBacktrace( $level + 1 );
1852         if ( isset( $backtrace[$level] ) ) {
1853                 return wfFormatStackFrame( $backtrace[$level] );
1854         } else {
1855                 return 'unknown';
1856         }
1857 }
1858
1859 /**
1860  * Return a string consisting of callers in the stack. Useful sometimes
1861  * for profiling specific points.
1862  *
1863  * @param int $limit The maximum depth of the stack frame to return, or false for the entire stack.
1864  * @return string
1865  */
1866 function wfGetAllCallers( $limit = 3 ) {
1867         $trace = array_reverse( wfDebugBacktrace() );
1868         if ( !$limit || $limit > count( $trace ) - 1 ) {
1869                 $limit = count( $trace ) - 1;
1870         }
1871         $trace = array_slice( $trace, -$limit - 1, $limit );
1872         return implode( '/', array_map( 'wfFormatStackFrame', $trace ) );
1873 }
1874
1875 /**
1876  * Return a string representation of frame
1877  *
1878  * @param array $frame
1879  * @return string
1880  */
1881 function wfFormatStackFrame( $frame ) {
1882         return isset( $frame['class'] ) ?
1883                 $frame['class'] . '::' . $frame['function'] :
1884                 $frame['function'];
1885 }
1886
1887 /* Some generic result counters, pulled out of SearchEngine */
1888
1889 /**
1890  * @todo document
1891  *
1892  * @param int $offset
1893  * @param int $limit
1894  * @return string
1895  */
1896 function wfShowingResults( $offset, $limit ) {
1897         return wfMessage( 'showingresults' )->numParams( $limit, $offset + 1 )->parse();
1898 }
1899
1900 /**
1901  * @todo document
1902  * @todo FIXME: We may want to blacklist some broken browsers
1903  *
1904  * @param bool $force
1905  * @return bool Whereas client accept gzip compression
1906  */
1907 function wfClientAcceptsGzip( $force = false ) {
1908         static $result = null;
1909         if ( $result === null || $force ) {
1910                 $result = false;
1911                 if ( isset( $_SERVER['HTTP_ACCEPT_ENCODING'] ) ) {
1912                         # @todo FIXME: We may want to blacklist some broken browsers
1913                         $m = array();
1914                         if ( preg_match(
1915                                         '/\bgzip(?:;(q)=([0-9]+(?:\.[0-9]+)))?\b/',
1916                                         $_SERVER['HTTP_ACCEPT_ENCODING'],
1917                                         $m
1918                                 )
1919                         ) {
1920                                 if ( isset( $m[2] ) && ( $m[1] == 'q' ) && ( $m[2] == 0 ) ) {
1921                                         $result = false;
1922                                         return $result;
1923                                 }
1924                                 wfDebug( "wfClientAcceptsGzip: client accepts gzip.\n" );
1925                                 $result = true;
1926                         }
1927                 }
1928         }
1929         return $result;
1930 }
1931
1932 /**
1933  * Obtain the offset and limit values from the request string;
1934  * used in special pages
1935  *
1936  * @param int $deflimit Default limit if none supplied
1937  * @param string $optionname Name of a user preference to check against
1938  * @return array
1939  * @deprecated since 1.24, just call WebRequest::getLimitOffset() directly
1940  */
1941 function wfCheckLimits( $deflimit = 50, $optionname = 'rclimit' ) {
1942         global $wgRequest;
1943         wfDeprecated( __METHOD__, '1.24' );
1944         return $wgRequest->getLimitOffset( $deflimit, $optionname );
1945 }
1946
1947 /**
1948  * Escapes the given text so that it may be output using addWikiText()
1949  * without any linking, formatting, etc. making its way through. This
1950  * is achieved by substituting certain characters with HTML entities.
1951  * As required by the callers, "<nowiki>" is not used.
1952  *
1953  * @param string $text Text to be escaped
1954  * @return string
1955  */
1956 function wfEscapeWikiText( $text ) {
1957         static $repl = null, $repl2 = null;
1958         if ( $repl === null ) {
1959                 $repl = array(
1960                         '"' => '&#34;', '&' => '&#38;', "'" => '&#39;', '<' => '&#60;',
1961                         '=' => '&#61;', '>' => '&#62;', '[' => '&#91;', ']' => '&#93;',
1962                         '{' => '&#123;', '|' => '&#124;', '}' => '&#125;', ';' => '&#59;',
1963                         "\n#" => "\n&#35;", "\r#" => "\r&#35;",
1964                         "\n*" => "\n&#42;", "\r*" => "\r&#42;",
1965                         "\n:" => "\n&#58;", "\r:" => "\r&#58;",
1966                         "\n " => "\n&#32;", "\r " => "\r&#32;",
1967                         "\n\n" => "\n&#10;", "\r\n" => "&#13;\n",
1968                         "\n\r" => "\n&#13;", "\r\r" => "\r&#13;",
1969                         "\n\t" => "\n&#9;", "\r\t" => "\r&#9;", // "\n\t\n" is treated like "\n\n"
1970                         "\n----" => "\n&#45;---", "\r----" => "\r&#45;---",
1971                         '__' => '_&#95;', '://' => '&#58;//',
1972                 );
1973
1974                 // We have to catch everything "\s" matches in PCRE
1975                 foreach ( array( 'ISBN', 'RFC', 'PMID' ) as $magic ) {
1976                         $repl["$magic "] = "$magic&#32;";
1977                         $repl["$magic\t"] = "$magic&#9;";
1978                         $repl["$magic\r"] = "$magic&#13;";
1979                         $repl["$magic\n"] = "$magic&#10;";
1980                         $repl["$magic\f"] = "$magic&#12;";
1981                 }
1982
1983                 // And handle protocols that don't use "://"
1984                 global $wgUrlProtocols;
1985                 $repl2 = array();
1986                 foreach ( $wgUrlProtocols as $prot ) {
1987                         if ( substr( $prot, -1 ) === ':' ) {
1988                                 $repl2[] = preg_quote( substr( $prot, 0, -1 ), '/' );
1989                         }
1990                 }
1991                 $repl2 = $repl2 ? '/\b(' . join( '|', $repl2 ) . '):/i' : '/^(?!)/';
1992         }
1993         $text = substr( strtr( "\n$text", $repl ), 1 );
1994         $text = preg_replace( $repl2, '$1&#58;', $text );
1995         return $text;
1996 }
1997
1998 /**
1999  * Sets dest to source and returns the original value of dest
2000  * If source is NULL, it just returns the value, it doesn't set the variable
2001  * If force is true, it will set the value even if source is NULL
2002  *
2003  * @param mixed $dest
2004  * @param mixed $source
2005  * @param bool $force
2006  * @return mixed
2007  */
2008 function wfSetVar( &$dest, $source, $force = false ) {
2009         $temp = $dest;
2010         if ( !is_null( $source ) || $force ) {
2011                 $dest = $source;
2012         }
2013         return $temp;
2014 }
2015
2016 /**
2017  * As for wfSetVar except setting a bit
2018  *
2019  * @param int $dest
2020  * @param int $bit
2021  * @param bool $state
2022  *
2023  * @return bool
2024  */
2025 function wfSetBit( &$dest, $bit, $state = true ) {
2026         $temp = (bool)( $dest & $bit );
2027         if ( !is_null( $state ) ) {
2028                 if ( $state ) {
2029                         $dest |= $bit;
2030                 } else {
2031                         $dest &= ~$bit;
2032                 }
2033         }
2034         return $temp;
2035 }
2036
2037 /**
2038  * A wrapper around the PHP function var_export().
2039  * Either print it or add it to the regular output ($wgOut).
2040  *
2041  * @param mixed $var A PHP variable to dump.
2042  */
2043 function wfVarDump( $var ) {
2044         global $wgOut;
2045         $s = str_replace( "\n", "<br />\n", var_export( $var, true ) . "\n" );
2046         if ( headers_sent() || !isset( $wgOut ) || !is_object( $wgOut ) ) {
2047                 print $s;
2048         } else {
2049                 $wgOut->addHTML( $s );
2050         }
2051 }
2052
2053 /**
2054  * Provide a simple HTTP error.
2055  *
2056  * @param int|string $code
2057  * @param string $label
2058  * @param string $desc
2059  */
2060 function wfHttpError( $code, $label, $desc ) {
2061         global $wgOut;
2062         $wgOut->disable();
2063         header( "HTTP/1.0 $code $label" );
2064         header( "Status: $code $label" );
2065         $wgOut->sendCacheControl();
2066
2067         header( 'Content-type: text/html; charset=utf-8' );
2068         print "<!doctype html>" .
2069                 '<html><head><title>' .
2070                 htmlspecialchars( $label ) .
2071                 '</title></head><body><h1>' .
2072                 htmlspecialchars( $label ) .
2073                 '</h1><p>' .
2074                 nl2br( htmlspecialchars( $desc ) ) .
2075                 "</p></body></html>\n";
2076 }
2077
2078 /**
2079  * Clear away any user-level output buffers, discarding contents.
2080  *
2081  * Suitable for 'starting afresh', for instance when streaming
2082  * relatively large amounts of data without buffering, or wanting to
2083  * output image files without ob_gzhandler's compression.
2084  *
2085  * The optional $resetGzipEncoding parameter controls suppression of
2086  * the Content-Encoding header sent by ob_gzhandler; by default it
2087  * is left. See comments for wfClearOutputBuffers() for why it would
2088  * be used.
2089  *
2090  * Note that some PHP configuration options may add output buffer
2091  * layers which cannot be removed; these are left in place.
2092  *
2093  * @param bool $resetGzipEncoding
2094  */
2095 function wfResetOutputBuffers( $resetGzipEncoding = true ) {
2096         if ( $resetGzipEncoding ) {
2097                 // Suppress Content-Encoding and Content-Length
2098                 // headers from 1.10+s wfOutputHandler
2099                 global $wgDisableOutputCompression;
2100                 $wgDisableOutputCompression = true;
2101         }
2102         while ( $status = ob_get_status() ) {
2103                 if ( $status['type'] == 0 /* PHP_OUTPUT_HANDLER_INTERNAL */ ) {
2104                         // Probably from zlib.output_compression or other
2105                         // PHP-internal setting which can't be removed.
2106                         //
2107                         // Give up, and hope the result doesn't break
2108                         // output behavior.
2109                         break;
2110                 }
2111                 if ( !ob_end_clean() ) {
2112                         // Could not remove output buffer handler; abort now
2113                         // to avoid getting in some kind of infinite loop.
2114                         break;
2115                 }
2116                 if ( $resetGzipEncoding ) {
2117                         if ( $status['name'] == 'ob_gzhandler' ) {
2118                                 // Reset the 'Content-Encoding' field set by this handler
2119                                 // so we can start fresh.
2120                                 header_remove( 'Content-Encoding' );
2121                                 break;
2122                         }
2123                 }
2124         }
2125 }
2126
2127 /**
2128  * More legible than passing a 'false' parameter to wfResetOutputBuffers():
2129  *
2130  * Clear away output buffers, but keep the Content-Encoding header
2131  * produced by ob_gzhandler, if any.
2132  *
2133  * This should be used for HTTP 304 responses, where you need to
2134  * preserve the Content-Encoding header of the real result, but
2135  * also need to suppress the output of ob_gzhandler to keep to spec
2136  * and avoid breaking Firefox in rare cases where the headers and
2137  * body are broken over two packets.
2138  */
2139 function wfClearOutputBuffers() {
2140         wfResetOutputBuffers( false );
2141 }
2142
2143 /**
2144  * Converts an Accept-* header into an array mapping string values to quality
2145  * factors
2146  *
2147  * @param string $accept
2148  * @param string $def Default
2149  * @return float[] Associative array of string => float pairs
2150  */
2151 function wfAcceptToPrefs( $accept, $def = '*/*' ) {
2152         # No arg means accept anything (per HTTP spec)
2153         if ( !$accept ) {
2154                 return array( $def => 1.0 );
2155         }
2156
2157         $prefs = array();
2158
2159         $parts = explode( ',', $accept );
2160
2161         foreach ( $parts as $part ) {
2162                 # @todo FIXME: Doesn't deal with params like 'text/html; level=1'
2163                 $values = explode( ';', trim( $part ) );
2164                 $match = array();
2165                 if ( count( $values ) == 1 ) {
2166                         $prefs[$values[0]] = 1.0;
2167                 } elseif ( preg_match( '/q\s*=\s*(\d*\.\d+)/', $values[1], $match ) ) {
2168                         $prefs[$values[0]] = floatval( $match[1] );
2169                 }
2170         }
2171
2172         return $prefs;
2173 }
2174
2175 /**
2176  * Checks if a given MIME type matches any of the keys in the given
2177  * array. Basic wildcards are accepted in the array keys.
2178  *
2179  * Returns the matching MIME type (or wildcard) if a match, otherwise
2180  * NULL if no match.
2181  *
2182  * @param string $type
2183  * @param array $avail
2184  * @return string
2185  * @private
2186  */
2187 function mimeTypeMatch( $type, $avail ) {
2188         if ( array_key_exists( $type, $avail ) ) {
2189                 return $type;
2190         } else {
2191                 $parts = explode( '/', $type );
2192                 if ( array_key_exists( $parts[0] . '/*', $avail ) ) {
2193                         return $parts[0] . '/*';
2194                 } elseif ( array_key_exists( '*/*', $avail ) ) {
2195                         return '*/*';
2196                 } else {
2197                         return null;
2198                 }
2199         }
2200 }
2201
2202 /**
2203  * Returns the 'best' match between a client's requested internet media types
2204  * and the server's list of available types. Each list should be an associative
2205  * array of type to preference (preference is a float between 0.0 and 1.0).
2206  * Wildcards in the types are acceptable.
2207  *
2208  * @param array $cprefs Client's acceptable type list
2209  * @param array $sprefs Server's offered types
2210  * @return string
2211  *
2212  * @todo FIXME: Doesn't handle params like 'text/plain; charset=UTF-8'
2213  * XXX: generalize to negotiate other stuff
2214  */
2215 function wfNegotiateType( $cprefs, $sprefs ) {
2216         $combine = array();
2217
2218         foreach ( array_keys( $sprefs ) as $type ) {
2219                 $parts = explode( '/', $type );
2220                 if ( $parts[1] != '*' ) {
2221                         $ckey = mimeTypeMatch( $type, $cprefs );
2222                         if ( $ckey ) {
2223                                 $combine[$type] = $sprefs[$type] * $cprefs[$ckey];
2224                         }
2225                 }
2226         }
2227
2228         foreach ( array_keys( $cprefs ) as $type ) {
2229                 $parts = explode( '/', $type );
2230                 if ( $parts[1] != '*' && !array_key_exists( $type, $sprefs ) ) {
2231                         $skey = mimeTypeMatch( $type, $sprefs );
2232                         if ( $skey ) {
2233                                 $combine[$type] = $sprefs[$skey] * $cprefs[$type];
2234                         }
2235                 }
2236         }
2237
2238         $bestq = 0;
2239         $besttype = null;
2240
2241         foreach ( array_keys( $combine ) as $type ) {
2242                 if ( $combine[$type] > $bestq ) {
2243                         $besttype = $type;
2244                         $bestq = $combine[$type];
2245                 }
2246         }
2247
2248         return $besttype;
2249 }
2250
2251 /**
2252  * Reference-counted warning suppression
2253  *
2254  * @param bool $end
2255  */
2256 function wfSuppressWarnings( $end = false ) {
2257         static $suppressCount = 0;
2258         static $originalLevel = false;
2259
2260         if ( $end ) {
2261                 if ( $suppressCount ) {
2262                         --$suppressCount;
2263                         if ( !$suppressCount ) {
2264                                 error_reporting( $originalLevel );
2265                         }
2266                 }
2267         } else {
2268                 if ( !$suppressCount ) {
2269                         $originalLevel = error_reporting( E_ALL & ~(
2270                                 E_WARNING |
2271                                 E_NOTICE |
2272                                 E_USER_WARNING |
2273                                 E_USER_NOTICE |
2274                                 E_DEPRECATED |
2275                                 E_USER_DEPRECATED |
2276                                 E_STRICT
2277                         ) );
2278                 }
2279                 ++$suppressCount;
2280         }
2281 }
2282
2283 /**
2284  * Restore error level to previous value
2285  */
2286 function wfRestoreWarnings() {
2287         wfSuppressWarnings( true );
2288 }
2289
2290 # Autodetect, convert and provide timestamps of various types
2291
2292 /**
2293  * Unix time - the number of seconds since 1970-01-01 00:00:00 UTC
2294  */
2295 define( 'TS_UNIX', 0 );
2296
2297 /**
2298  * MediaWiki concatenated string timestamp (YYYYMMDDHHMMSS)
2299  */
2300 define( 'TS_MW', 1 );
2301
2302 /**
2303  * MySQL DATETIME (YYYY-MM-DD HH:MM:SS)
2304  */
2305 define( 'TS_DB', 2 );
2306
2307 /**
2308  * RFC 2822 format, for E-mail and HTTP headers
2309  */
2310 define( 'TS_RFC2822', 3 );
2311
2312 /**
2313  * ISO 8601 format with no timezone: 1986-02-09T20:00:00Z
2314  *
2315  * This is used by Special:Export
2316  */
2317 define( 'TS_ISO_8601', 4 );
2318
2319 /**
2320  * An Exif timestamp (YYYY:MM:DD HH:MM:SS)
2321  *
2322  * @see http://exif.org/Exif2-2.PDF The Exif 2.2 spec, see page 28 for the
2323  *       DateTime tag and page 36 for the DateTimeOriginal and
2324  *       DateTimeDigitized tags.
2325  */
2326 define( 'TS_EXIF', 5 );
2327
2328 /**
2329  * Oracle format time.
2330  */
2331 define( 'TS_ORACLE', 6 );
2332
2333 /**
2334  * Postgres format time.
2335  */
2336 define( 'TS_POSTGRES', 7 );
2337
2338 /**
2339  * ISO 8601 basic format with no timezone: 19860209T200000Z.  This is used by ResourceLoader
2340  */
2341 define( 'TS_ISO_8601_BASIC', 9 );
2342
2343 /**
2344  * Get a timestamp string in one of various formats
2345  *
2346  * @param mixed $outputtype A timestamp in one of the supported formats, the
2347  *   function will autodetect which format is supplied and act accordingly.
2348  * @param mixed $ts Optional timestamp to convert, default 0 for the current time
2349  * @return string|bool String / false The same date in the format specified in $outputtype or false
2350  */
2351 function wfTimestamp( $outputtype = TS_UNIX, $ts = 0 ) {
2352         try {
2353                 $timestamp = new MWTimestamp( $ts );
2354                 return $timestamp->getTimestamp( $outputtype );
2355         } catch ( TimestampException $e ) {
2356                 wfDebug( "wfTimestamp() fed bogus time value: TYPE=$outputtype; VALUE=$ts\n" );
2357                 return false;
2358         }
2359 }
2360
2361 /**
2362  * Return a formatted timestamp, or null if input is null.
2363  * For dealing with nullable timestamp columns in the database.
2364  *
2365  * @param int $outputtype
2366  * @param string $ts
2367  * @return string
2368  */
2369 function wfTimestampOrNull( $outputtype = TS_UNIX, $ts = null ) {
2370         if ( is_null( $ts ) ) {
2371                 return null;
2372         } else {
2373                 return wfTimestamp( $outputtype, $ts );
2374         }
2375 }
2376
2377 /**
2378  * Convenience function; returns MediaWiki timestamp for the present time.
2379  *
2380  * @return string
2381  */
2382 function wfTimestampNow() {
2383         # return NOW
2384         return wfTimestamp( TS_MW, time() );
2385 }
2386
2387 /**
2388  * Check if the operating system is Windows
2389  *
2390  * @return bool True if it's Windows, false otherwise.
2391  */
2392 function wfIsWindows() {
2393         static $isWindows = null;
2394         if ( $isWindows === null ) {
2395                 $isWindows = substr( php_uname(), 0, 7 ) == 'Windows';
2396         }
2397         return $isWindows;
2398 }
2399
2400 /**
2401  * Check if we are running under HHVM
2402  *
2403  * @return bool
2404  */
2405 function wfIsHHVM() {
2406         return defined( 'HHVM_VERSION' );
2407 }
2408
2409 /**
2410  * Swap two variables
2411  *
2412  * @deprecated since 1.24
2413  * @param mixed $x
2414  * @param mixed $y
2415  */
2416 function swap( &$x, &$y ) {
2417         wfDeprecated( __FUNCTION__, '1.24' );
2418         $z = $x;
2419         $x = $y;
2420         $y = $z;
2421 }
2422
2423 /**
2424  * Tries to get the system directory for temporary files. First
2425  * $wgTmpDirectory is checked, and then the TMPDIR, TMP, and TEMP
2426  * environment variables are then checked in sequence, and if none are
2427  * set try sys_get_temp_dir().
2428  *
2429  * NOTE: When possible, use instead the tmpfile() function to create
2430  * temporary files to avoid race conditions on file creation, etc.
2431  *
2432  * @return string
2433  */
2434 function wfTempDir() {
2435         global $wgTmpDirectory;
2436
2437         if ( $wgTmpDirectory !== false ) {
2438                 return $wgTmpDirectory;
2439         }
2440
2441         $tmpDir = array_map( "getenv", array( 'TMPDIR', 'TMP', 'TEMP' ) );
2442
2443         foreach ( $tmpDir as $tmp ) {
2444                 if ( $tmp && file_exists( $tmp ) && is_dir( $tmp ) && is_writable( $tmp ) ) {
2445                         return $tmp;
2446                 }
2447         }
2448         return sys_get_temp_dir();
2449 }
2450
2451 /**
2452  * Make directory, and make all parent directories if they don't exist
2453  *
2454  * @param string $dir Full path to directory to create
2455  * @param int $mode Chmod value to use, default is $wgDirectoryMode
2456  * @param string $caller Optional caller param for debugging.
2457  * @throws MWException
2458  * @return bool
2459  */
2460 function wfMkdirParents( $dir, $mode = null, $caller = null ) {
2461         global $wgDirectoryMode;
2462
2463         if ( FileBackend::isStoragePath( $dir ) ) { // sanity
2464                 throw new MWException( __FUNCTION__ . " given storage path '$dir'." );
2465         }
2466
2467         if ( !is_null( $caller ) ) {
2468                 wfDebug( "$caller: called wfMkdirParents($dir)\n" );
2469         }
2470
2471         if ( strval( $dir ) === '' || ( file_exists( $dir ) && is_dir( $dir ) ) ) {
2472                 return true;
2473         }
2474
2475         $dir = str_replace( array( '\\', '/' ), DIRECTORY_SEPARATOR, $dir );
2476
2477         if ( is_null( $mode ) ) {
2478                 $mode = $wgDirectoryMode;
2479         }
2480
2481         // Turn off the normal warning, we're doing our own below
2482         wfSuppressWarnings();
2483         $ok = mkdir( $dir, $mode, true ); // PHP5 <3
2484         wfRestoreWarnings();
2485
2486         if ( !$ok ) {
2487                 //directory may have been created on another request since we last checked
2488                 if ( is_dir( $dir ) ) {
2489                         return true;
2490                 }
2491
2492                 // PHP doesn't report the path in its warning message, so add our own to aid in diagnosis.
2493                 wfLogWarning( sprintf( "failed to mkdir \"%s\" mode 0%o", $dir, $mode ) );
2494         }
2495         return $ok;
2496 }
2497
2498 /**
2499  * Remove a directory and all its content.
2500  * Does not hide error.
2501  * @param string $dir
2502  */
2503 function wfRecursiveRemoveDir( $dir ) {
2504         wfDebug( __FUNCTION__ . "( $dir )\n" );
2505         // taken from http://de3.php.net/manual/en/function.rmdir.php#98622
2506         if ( is_dir( $dir ) ) {
2507                 $objects = scandir( $dir );
2508                 foreach ( $objects as $object ) {
2509                         if ( $object != "." && $object != ".." ) {
2510                                 if ( filetype( $dir . '/' . $object ) == "dir" ) {
2511                                         wfRecursiveRemoveDir( $dir . '/' . $object );
2512                                 } else {
2513                                         unlink( $dir . '/' . $object );
2514                                 }
2515                         }
2516                 }
2517                 reset( $objects );
2518                 rmdir( $dir );
2519         }
2520 }
2521
2522 /**
2523  * @param int $nr The number to format
2524  * @param int $acc The number of digits after the decimal point, default 2
2525  * @param bool $round Whether or not to round the value, default true
2526  * @return string
2527  */
2528 function wfPercent( $nr, $acc = 2, $round = true ) {
2529         $ret = sprintf( "%.${acc}f", $nr );
2530         return $round ? round( $ret, $acc ) . '%' : "$ret%";
2531 }
2532
2533 /**
2534  * Safety wrapper around ini_get() for boolean settings.
2535  * The values returned from ini_get() are pre-normalized for settings
2536  * set via php.ini or php_flag/php_admin_flag... but *not*
2537  * for those set via php_value/php_admin_value.
2538  *
2539  * It's fairly common for people to use php_value instead of php_flag,
2540  * which can leave you with an 'off' setting giving a false positive
2541  * for code that just takes the ini_get() return value as a boolean.
2542  *
2543  * To make things extra interesting, setting via php_value accepts
2544  * "true" and "yes" as true, but php.ini and php_flag consider them false. :)
2545  * Unrecognized values go false... again opposite PHP's own coercion
2546  * from string to bool.
2547  *
2548  * Luckily, 'properly' set settings will always come back as '0' or '1',
2549  * so we only have to worry about them and the 'improper' settings.
2550  *
2551  * I frickin' hate PHP... :P
2552  *
2553  * @param string $setting
2554  * @return bool
2555  */
2556 function wfIniGetBool( $setting ) {
2557         $val = strtolower( ini_get( $setting ) );
2558         // 'on' and 'true' can't have whitespace around them, but '1' can.
2559         return $val == 'on'
2560                 || $val == 'true'
2561                 || $val == 'yes'
2562                 || preg_match( "/^\s*[+-]?0*[1-9]/", $val ); // approx C atoi() function
2563 }
2564
2565 /**
2566  * Windows-compatible version of escapeshellarg()
2567  * Windows doesn't recognise single-quotes in the shell, but the escapeshellarg()
2568  * function puts single quotes in regardless of OS.
2569  *
2570  * Also fixes the locale problems on Linux in PHP 5.2.6+ (bug backported to
2571  * earlier distro releases of PHP)
2572  *
2573  * @param string $args,...
2574  * @return string
2575  */
2576 function wfEscapeShellArg( /*...*/ ) {
2577         wfInitShellLocale();
2578
2579         $args = func_get_args();
2580         $first = true;
2581         $retVal = '';
2582         foreach ( $args as $arg ) {
2583                 if ( !$first ) {
2584                         $retVal .= ' ';
2585                 } else {
2586                         $first = false;
2587                 }
2588
2589                 if ( wfIsWindows() ) {
2590                         // Escaping for an MSVC-style command line parser and CMD.EXE
2591                         // @codingStandardsIgnoreStart For long URLs
2592                         // Refs:
2593                         //  * http://web.archive.org/web/20020708081031/http://mailman.lyra.org/pipermail/scite-interest/2002-March/000436.html
2594                         //  * http://technet.microsoft.com/en-us/library/cc723564.aspx
2595                         //  * Bug #13518
2596                         //  * CR r63214
2597                         // Double the backslashes before any double quotes. Escape the double quotes.
2598                         // @codingStandardsIgnoreEnd
2599                         $tokens = preg_split( '/(\\\\*")/', $arg, -1, PREG_SPLIT_DELIM_CAPTURE );
2600                         $arg = '';
2601                         $iteration = 0;
2602                         foreach ( $tokens as $token ) {
2603                                 if ( $iteration % 2 == 1 ) {
2604                                         // Delimiter, a double quote preceded by zero or more slashes
2605                                         $arg .= str_replace( '\\', '\\\\', substr( $token, 0, -1 ) ) . '\\"';
2606                                 } elseif ( $iteration % 4 == 2 ) {
2607                                         // ^ in $token will be outside quotes, need to be escaped
2608                                         $arg .= str_replace( '^', '^^', $token );
2609                                 } else { // $iteration % 4 == 0
2610                                         // ^ in $token will appear inside double quotes, so leave as is
2611                                         $arg .= $token;
2612                                 }
2613                                 $iteration++;
2614                         }
2615                         // Double the backslashes before the end of the string, because
2616                         // we will soon add a quote
2617                         $m = array();
2618                         if ( preg_match( '/^(.*?)(\\\\+)$/', $arg, $m ) ) {
2619                                 $arg = $m[1] . str_replace( '\\', '\\\\', $m[2] );
2620                         }
2621
2622                         // Add surrounding quotes
2623                         $retVal .= '"' . $arg . '"';
2624                 } else {
2625                         $retVal .= escapeshellarg( $arg );
2626                 }
2627         }
2628         return $retVal;
2629 }
2630
2631 /**
2632  * Check if wfShellExec() is effectively disabled via php.ini config
2633  *
2634  * @return bool|string False or one of (safemode,disabled)
2635  * @since 1.22
2636  */
2637 function wfShellExecDisabled() {
2638         static $disabled = null;
2639         if ( is_null( $disabled ) ) {
2640                 if ( wfIniGetBool( 'safe_mode' ) ) {
2641                         wfDebug( "wfShellExec can't run in safe_mode, PHP's exec functions are too broken.\n" );
2642                         $disabled = 'safemode';
2643                 } elseif ( !function_exists( 'proc_open' ) ) {
2644                         wfDebug( "proc_open() is disabled\n" );
2645                         $disabled = 'disabled';
2646                 } else {
2647                         $disabled = false;
2648                 }
2649         }
2650         return $disabled;
2651 }
2652
2653 /**
2654  * Execute a shell command, with time and memory limits mirrored from the PHP
2655  * configuration if supported.
2656  *
2657  * @param string|string[] $cmd If string, a properly shell-escaped command line,
2658  *   or an array of unescaped arguments, in which case each value will be escaped
2659  *   Example:   [ 'convert', '-font', 'font name' ] would produce "'convert' '-font' 'font name'"
2660  * @param null|mixed &$retval Optional, will receive the program's exit code.
2661  *   (non-zero is usually failure). If there is an error from
2662  *   read, select, or proc_open(), this will be set to -1.
2663  * @param array $environ Optional environment variables which should be
2664  *   added to the executed command environment.
2665  * @param array $limits Optional array with limits(filesize, memory, time, walltime)
2666  *   this overwrites the global wgMaxShell* limits.
2667  * @param array $options Array of options:
2668  *   - duplicateStderr: Set this to true to duplicate stderr to stdout,
2669  *     including errors from limit.sh
2670  *
2671  * @return string Collected stdout as a string
2672  */
2673 function wfShellExec( $cmd, &$retval = null, $environ = array(),
2674         $limits = array(), $options = array()
2675 ) {
2676         global $IP, $wgMaxShellMemory, $wgMaxShellFileSize, $wgMaxShellTime,
2677                 $wgMaxShellWallClockTime, $wgShellCgroup;
2678
2679         $disabled = wfShellExecDisabled();
2680         if ( $disabled ) {
2681                 $retval = 1;
2682                 return $disabled == 'safemode' ?
2683                         'Unable to run external programs in safe mode.' :
2684                         'Unable to run external programs, proc_open() is disabled.';
2685         }
2686
2687         $includeStderr = isset( $options['duplicateStderr'] ) && $options['duplicateStderr'];
2688
2689         wfInitShellLocale();
2690
2691         $envcmd = '';
2692         foreach ( $environ as $k => $v ) {
2693                 if ( wfIsWindows() ) {
2694                         /* Surrounding a set in quotes (method used by wfEscapeShellArg) makes the quotes themselves
2695                          * appear in the environment variable, so we must use carat escaping as documented in
2696                          * http://technet.microsoft.com/en-us/library/cc723564.aspx
2697                          * Note however that the quote isn't listed there, but is needed, and the parentheses
2698                          * are listed there but doesn't appear to need it.
2699                          */
2700                         $envcmd .= "set $k=" . preg_replace( '/([&|()<>^"])/', '^\\1', $v ) . '&& ';
2701                 } else {
2702                         /* Assume this is a POSIX shell, thus required to accept variable assignments before the command
2703                          * http://www.opengroup.org/onlinepubs/009695399/utilities/xcu_chap02.html#tag_02_09_01
2704                          */
2705                         $envcmd .= "$k=" . escapeshellarg( $v ) . ' ';
2706                 }
2707         }
2708         if ( is_array( $cmd ) ) {
2709                 // Command line may be given as an array, escape each value and glue them together with a space
2710                 $cmdVals = array();
2711                 foreach ( $cmd as $val ) {
2712                         $cmdVals[] = wfEscapeShellArg( $val );
2713                 }
2714                 $cmd = implode( ' ', $cmdVals );
2715         }
2716
2717         $cmd = $envcmd . $cmd;
2718
2719         $useLogPipe = false;
2720         if ( is_executable( '/bin/bash' ) ) {
2721                 $time = intval ( isset( $limits['time'] ) ? $limits['time'] : $wgMaxShellTime );
2722                 if ( isset( $limits['walltime'] ) ) {
2723                         $wallTime = intval( $limits['walltime'] );
2724                 } elseif ( isset( $limits['time'] ) ) {
2725                         $wallTime = $time;
2726                 } else {
2727                         $wallTime = intval( $wgMaxShellWallClockTime );
2728                 }
2729                 $mem = intval ( isset( $limits['memory'] ) ? $limits['memory'] : $wgMaxShellMemory );
2730                 $filesize = intval ( isset( $limits['filesize'] ) ? $limits['filesize'] : $wgMaxShellFileSize );
2731
2732                 if ( $time > 0 || $mem > 0 || $filesize > 0 || $wallTime > 0 ) {
2733                         $cmd = '/bin/bash ' . escapeshellarg( "$IP/includes/limit.sh" ) . ' ' .
2734                                 escapeshellarg( $cmd ) . ' ' .
2735                                 escapeshellarg(
2736                                         "MW_INCLUDE_STDERR=" . ( $includeStderr ? '1' : '' ) . ';' .
2737                                         "MW_CPU_LIMIT=$time; " .
2738                                         'MW_CGROUP=' . escapeshellarg( $wgShellCgroup ) . '; ' .
2739                                         "MW_MEM_LIMIT=$mem; " .
2740                                         "MW_FILE_SIZE_LIMIT=$filesize; " .
2741                                         "MW_WALL_CLOCK_LIMIT=$wallTime; " .
2742                                         "MW_USE_LOG_PIPE=yes"
2743                                 );
2744                         $useLogPipe = true;
2745                 } elseif ( $includeStderr ) {
2746                         $cmd .= ' 2>&1';
2747                 }
2748         } elseif ( $includeStderr ) {
2749                 $cmd .= ' 2>&1';
2750         }
2751         wfDebug( "wfShellExec: $cmd\n" );
2752
2753         $desc = array(
2754                 0 => array( 'file', 'php://stdin', 'r' ),
2755                 1 => array( 'pipe', 'w' ),
2756                 2 => array( 'file', 'php://stderr', 'w' ) );
2757         if ( $useLogPipe ) {
2758                 $desc[3] = array( 'pipe', 'w' );
2759         }
2760         $pipes = null;
2761         $proc = proc_open( $cmd, $desc, $pipes );
2762         if ( !$proc ) {
2763                 wfDebugLog( 'exec', "proc_open() failed: $cmd" );
2764                 $retval = -1;
2765                 return '';
2766         }
2767         $outBuffer = $logBuffer = '';
2768         $emptyArray = array();
2769         $status = false;
2770         $logMsg = false;
2771
2772         // According to the documentation, it is possible for stream_select()
2773         // to fail due to EINTR. I haven't managed to induce this in testing
2774         // despite sending various signals. If it did happen, the error
2775         // message would take the form:
2776         //
2777         // stream_select(): unable to select [4]: Interrupted system call (max_fd=5)
2778         //
2779         // where [4] is the value of the macro EINTR and "Interrupted system
2780         // call" is string which according to the Linux manual is "possibly"
2781         // localised according to LC_MESSAGES.
2782         $eintr = defined( 'SOCKET_EINTR' ) ? SOCKET_EINTR : 4;
2783         $eintrMessage = "stream_select(): unable to select [$eintr]";
2784
2785         // Build a table mapping resource IDs to pipe FDs to work around a
2786         // PHP 5.3 issue in which stream_select() does not preserve array keys
2787         // <https://bugs.php.net/bug.php?id=53427>.
2788         $fds = array();
2789         foreach ( $pipes as $fd => $pipe ) {
2790                 $fds[(int)$pipe] = $fd;
2791         }
2792
2793         $running = true;
2794         $timeout = null;
2795         $numReadyPipes = 0;
2796
2797         while ( $running === true || $numReadyPipes !== 0 ) {
2798                 if ( $running ) {
2799                         $status = proc_get_status( $proc );
2800                         // If the process has terminated, switch to nonblocking selects
2801                         // for getting any data still waiting to be read.
2802                         if ( !$status['running'] ) {
2803                                 $running = false;
2804                                 $timeout = 0;
2805                         }
2806                 }
2807
2808                 $readyPipes = $pipes;
2809
2810                 // Clear last error
2811                 // @codingStandardsIgnoreStart Generic.PHP.NoSilencedErrors.Discouraged
2812                 @trigger_error( '' );
2813                 $numReadyPipes = @stream_select( $readyPipes, $emptyArray, $emptyArray, $timeout );
2814                 if ( $numReadyPipes === false ) {
2815                         // @codingStandardsIgnoreEnd
2816                         $error = error_get_last();
2817                         if ( strncmp( $error['message'], $eintrMessage, strlen( $eintrMessage ) ) == 0 ) {
2818                                 continue;
2819                         } else {
2820                                 trigger_error( $error['message'], E_USER_WARNING );
2821                                 $logMsg = $error['message'];
2822                                 break;
2823                         }
2824                 }
2825                 foreach ( $readyPipes as $pipe ) {
2826                         $block = fread( $pipe, 65536 );
2827                         $fd = $fds[(int)$pipe];
2828                         if ( $block === '' ) {
2829                                 // End of file
2830                                 fclose( $pipes[$fd] );
2831                                 unset( $pipes[$fd] );
2832                                 if ( !$pipes ) {
2833                                         break 2;
2834                                 }
2835                         } elseif ( $block === false ) {
2836                                 // Read error
2837                                 $logMsg = "Error reading from pipe";
2838                                 break 2;
2839                         } elseif ( $fd == 1 ) {
2840                                 // From stdout
2841                                 $outBuffer .= $block;
2842                         } elseif ( $fd == 3 ) {
2843                                 // From log FD
2844                                 $logBuffer .= $block;
2845                                 if ( strpos( $block, "\n" ) !== false ) {
2846                                         $lines = explode( "\n", $logBuffer );
2847                                         $logBuffer = array_pop( $lines );
2848                                         foreach ( $lines as $line ) {
2849                                                 wfDebugLog( 'exec', $line );
2850                                         }
2851                                 }
2852                         }
2853                 }
2854         }
2855
2856         foreach ( $pipes as $pipe ) {
2857                 fclose( $pipe );
2858         }
2859
2860         // Use the status previously collected if possible, since proc_get_status()
2861         // just calls waitpid() which will not return anything useful the second time.
2862         if ( $running ) {
2863                 $status = proc_get_status( $proc );
2864         }
2865
2866         if ( $logMsg !== false ) {
2867                 // Read/select error
2868                 $retval = -1;
2869                 proc_close( $proc );
2870         } elseif ( $status['signaled'] ) {
2871                 $logMsg = "Exited with signal {$status['termsig']}";
2872                 $retval = 128 + $status['termsig'];
2873                 proc_close( $proc );
2874         } else {
2875                 if ( $status['running'] ) {
2876                         $retval = proc_close( $proc );
2877                 } else {
2878                         $retval = $status['exitcode'];
2879                         proc_close( $proc );
2880                 }
2881                 if ( $retval == 127 ) {
2882                         $logMsg = "Possibly missing executable file";
2883                 } elseif ( $retval >= 129 && $retval <= 192 ) {
2884                         $logMsg = "Probably exited with signal " . ( $retval - 128 );
2885                 }
2886         }
2887
2888         if ( $logMsg !== false ) {
2889                 wfDebugLog( 'exec', "$logMsg: $cmd" );
2890         }
2891
2892         return $outBuffer;
2893 }
2894
2895 /**
2896  * Execute a shell command, returning both stdout and stderr. Convenience
2897  * function, as all the arguments to wfShellExec can become unwieldy.
2898  *
2899  * @note This also includes errors from limit.sh, e.g. if $wgMaxShellFileSize is exceeded.
2900  * @param string|string[] $cmd If string, a properly shell-escaped command line,
2901  *   or an array of unescaped arguments, in which case each value will be escaped
2902  *   Example:   [ 'convert', '-font', 'font name' ] would produce "'convert' '-font' 'font name'"
2903  * @param null|mixed &$retval Optional, will receive the program's exit code.
2904  *   (non-zero is usually failure)
2905  * @param array $environ Optional environment variables which should be
2906  *   added to the executed command environment.
2907  * @param array $limits Optional array with limits(filesize, memory, time, walltime)
2908  *   this overwrites the global wgMaxShell* limits.
2909  * @return string Collected stdout and stderr as a string
2910  */
2911 function wfShellExecWithStderr( $cmd, &$retval = null, $environ = array(), $limits = array() ) {
2912         return wfShellExec( $cmd, $retval, $environ, $limits, array( 'duplicateStderr' => true ) );
2913 }
2914
2915 /**
2916  * Workaround for http://bugs.php.net/bug.php?id=45132
2917  * escapeshellarg() destroys non-ASCII characters if LANG is not a UTF-8 locale
2918  */
2919 function wfInitShellLocale() {
2920         static $done = false;
2921         if ( $done ) {
2922                 return;
2923         }
2924         $done = true;
2925         global $wgShellLocale;
2926         if ( !wfIniGetBool( 'safe_mode' ) ) {
2927                 putenv( "LC_CTYPE=$wgShellLocale" );
2928                 setlocale( LC_CTYPE, $wgShellLocale );
2929         }
2930 }
2931
2932 /**
2933  * Alias to wfShellWikiCmd()
2934  *
2935  * @see wfShellWikiCmd()
2936  */
2937 function wfShellMaintenanceCmd( $script, array $parameters = array(), array $options = array() ) {
2938         return wfShellWikiCmd( $script, $parameters, $options );
2939 }
2940
2941 /**
2942  * Generate a shell-escaped command line string to run a MediaWiki cli script.
2943  * Note that $parameters should be a flat array and an option with an argument
2944  * should consist of two consecutive items in the array (do not use "--option value").
2945  *
2946  * @param string $script MediaWiki cli script path
2947  * @param array $parameters Arguments and options to the script
2948  * @param array $options Associative array of options:
2949  *              'php': The path to the php executable
2950  *              'wrapper': Path to a PHP wrapper to handle the maintenance script
2951  * @return string
2952  */
2953 function wfShellWikiCmd( $script, array $parameters = array(), array $options = array() ) {
2954         global $wgPhpCli;
2955         // Give site config file a chance to run the script in a wrapper.
2956         // The caller may likely want to call wfBasename() on $script.
2957         wfRunHooks( 'wfShellWikiCmd', array( &$script, &$parameters, &$options ) );
2958         $cmd = isset( $options['php'] ) ? array( $options['php'] ) : array( $wgPhpCli );
2959         if ( isset( $options['wrapper'] ) ) {
2960                 $cmd[] = $options['wrapper'];
2961         }
2962         $cmd[] = $script;
2963         // Escape each parameter for shell
2964         return implode( " ", array_map( 'wfEscapeShellArg', array_merge( $cmd, $parameters ) ) );
2965 }
2966
2967 /**
2968  * wfMerge attempts to merge differences between three texts.
2969  * Returns true for a clean merge and false for failure or a conflict.
2970  *
2971  * @param string $old
2972  * @param string $mine
2973  * @param string $yours
2974  * @param string $result
2975  * @return bool
2976  */
2977 function wfMerge( $old, $mine, $yours, &$result ) {
2978         global $wgDiff3;
2979
2980         # This check may also protect against code injection in
2981         # case of broken installations.
2982         wfSuppressWarnings();
2983         $haveDiff3 = $wgDiff3 && file_exists( $wgDiff3 );
2984         wfRestoreWarnings();
2985
2986         if ( !$haveDiff3 ) {
2987                 wfDebug( "diff3 not found\n" );
2988                 return false;
2989         }
2990
2991         # Make temporary files
2992         $td = wfTempDir();
2993         $oldtextFile = fopen( $oldtextName = tempnam( $td, 'merge-old-' ), 'w' );
2994         $mytextFile = fopen( $mytextName = tempnam( $td, 'merge-mine-' ), 'w' );
2995         $yourtextFile = fopen( $yourtextName = tempnam( $td, 'merge-your-' ), 'w' );
2996
2997         # NOTE: diff3 issues a warning to stderr if any of the files does not end with
2998         #       a newline character. To avoid this, we normalize the trailing whitespace before
2999         #       creating the diff.
3000
3001         fwrite( $oldtextFile, rtrim( $old ) . "\n" );
3002         fclose( $oldtextFile );
3003         fwrite( $mytextFile, rtrim( $mine ) . "\n" );
3004         fclose( $mytextFile );
3005         fwrite( $yourtextFile, rtrim( $yours ) . "\n" );
3006         fclose( $yourtextFile );
3007
3008         # Check for a conflict
3009         $cmd = wfEscapeShellArg( $wgDiff3 ) . ' -a --overlap-only ' .
3010                 wfEscapeShellArg( $mytextName ) . ' ' .
3011                 wfEscapeShellArg( $oldtextName ) . ' ' .
3012                 wfEscapeShellArg( $yourtextName );
3013         $handle = popen( $cmd, 'r' );
3014
3015         if ( fgets( $handle, 1024 ) ) {
3016                 $conflict = true;
3017         } else {
3018                 $conflict = false;
3019         }
3020         pclose( $handle );
3021
3022         # Merge differences
3023         $cmd = wfEscapeShellArg( $wgDiff3 ) . ' -a -e --merge ' .
3024                 wfEscapeShellArg( $mytextName, $oldtextName, $yourtextName );
3025         $handle = popen( $cmd, 'r' );
3026         $result = '';
3027         do {
3028                 $data = fread( $handle, 8192 );
3029                 if ( strlen( $data ) == 0 ) {
3030                         break;
3031                 }
3032                 $result .= $data;
3033         } while ( true );
3034         pclose( $handle );
3035         unlink( $mytextName );
3036         unlink( $oldtextName );
3037         unlink( $yourtextName );
3038
3039         if ( $result === '' && $old !== '' && !$conflict ) {
3040                 wfDebug( "Unexpected null result from diff3. Command: $cmd\n" );
3041                 $conflict = true;
3042         }
3043         return !$conflict;
3044 }
3045
3046 /**
3047  * Returns unified plain-text diff of two texts.
3048  * Useful for machine processing of diffs.
3049  *
3050  * @param string $before The text before the changes.
3051  * @param string $after The text after the changes.
3052  * @param string $params Command-line options for the diff command.
3053  * @return string Unified diff of $before and $after
3054  */
3055 function wfDiff( $before, $after, $params = '-u' ) {
3056         if ( $before == $after ) {
3057                 return '';
3058         }
3059
3060         global $wgDiff;
3061         wfSuppressWarnings();
3062         $haveDiff = $wgDiff && file_exists( $wgDiff );
3063         wfRestoreWarnings();
3064
3065         # This check may also protect against code injection in
3066         # case of broken installations.
3067         if ( !$haveDiff ) {
3068                 wfDebug( "diff executable not found\n" );
3069                 $diffs = new Diff( explode( "\n", $before ), explode( "\n", $after ) );
3070                 $format = new UnifiedDiffFormatter();
3071                 return $format->format( $diffs );
3072         }
3073
3074         # Make temporary files
3075         $td = wfTempDir();
3076         $oldtextFile = fopen( $oldtextName = tempnam( $td, 'merge-old-' ), 'w' );
3077         $newtextFile = fopen( $newtextName = tempnam( $td, 'merge-your-' ), 'w' );
3078
3079         fwrite( $oldtextFile, $before );
3080         fclose( $oldtextFile );
3081         fwrite( $newtextFile, $after );
3082         fclose( $newtextFile );
3083
3084         // Get the diff of the two files
3085         $cmd = "$wgDiff " . $params . ' ' . wfEscapeShellArg( $oldtextName, $newtextName );
3086
3087         $h = popen( $cmd, 'r' );
3088
3089         $diff = '';
3090
3091         do {
3092                 $data = fread( $h, 8192 );
3093                 if ( strlen( $data ) == 0 ) {
3094                         break;
3095                 }
3096                 $diff .= $data;
3097         } while ( true );
3098
3099         // Clean up
3100         pclose( $h );
3101         unlink( $oldtextName );
3102         unlink( $newtextName );
3103
3104         // Kill the --- and +++ lines. They're not useful.
3105         $diff_lines = explode( "\n", $diff );
3106         if ( isset( $diff_lines[0] ) && strpos( $diff_lines[0], '---' ) === 0 ) {
3107                 unset( $diff_lines[0] );
3108         }
3109         if ( isset( $diff_lines[1] ) && strpos( $diff_lines[1], '+++' ) === 0 ) {
3110                 unset( $diff_lines[1] );
3111         }
3112
3113         $diff = implode( "\n", $diff_lines );
3114
3115         return $diff;
3116 }
3117
3118 /**
3119  * This function works like "use VERSION" in Perl, the program will die with a
3120  * backtrace if the current version of PHP is less than the version provided
3121  *
3122  * This is useful for extensions which due to their nature are not kept in sync
3123  * with releases, and might depend on other versions of PHP than the main code
3124  *
3125  * Note: PHP might die due to parsing errors in some cases before it ever
3126  *       manages to call this function, such is life
3127  *
3128  * @see perldoc -f use
3129  *
3130  * @param string|int|float $req_ver The version to check, can be a string, an integer, or a float
3131  * @throws MWException
3132  */
3133 function wfUsePHP( $req_ver ) {
3134         $php_ver = PHP_VERSION;
3135
3136         if ( version_compare( $php_ver, (string)$req_ver, '<' ) ) {
3137                 throw new MWException( "PHP $req_ver required--this is only $php_ver" );
3138         }
3139 }
3140
3141 /**
3142  * This function works like "use VERSION" in Perl except it checks the version
3143  * of MediaWiki, the program will die with a backtrace if the current version
3144  * of MediaWiki is less than the version provided.
3145  *
3146  * This is useful for extensions which due to their nature are not kept in sync
3147  * with releases
3148  *
3149  * Note: Due to the behavior of PHP's version_compare() which is used in this
3150  * function, if you want to allow the 'wmf' development versions add a 'c' (or
3151  * any single letter other than 'a', 'b' or 'p') as a post-fix to your
3152  * targeted version number. For example if you wanted to allow any variation
3153  * of 1.22 use `wfUseMW( '1.22c' )`. Using an 'a' or 'b' instead of 'c' will
3154  * not result in the same comparison due to the internal logic of
3155  * version_compare().
3156  *
3157  * @see perldoc -f use
3158  *
3159  * @param string|int|float $req_ver The version to check, can be a string, an integer, or a float
3160  * @throws MWException
3161  */
3162 function wfUseMW( $req_ver ) {
3163         global $wgVersion;
3164
3165         if ( version_compare( $wgVersion, (string)$req_ver, '<' ) ) {
3166                 throw new MWException( "MediaWiki $req_ver required--this is only $wgVersion" );
3167         }
3168 }
3169
3170 /**
3171  * Return the final portion of a pathname.
3172  * Reimplemented because PHP5's "basename()" is buggy with multibyte text.
3173  * http://bugs.php.net/bug.php?id=33898
3174  *
3175  * PHP's basename() only considers '\' a pathchar on Windows and Netware.
3176  * We'll consider it so always, as we don't want '\s' in our Unix paths either.
3177  *
3178  * @param string $path
3179  * @param string $suffix String to remove if present
3180  * @return string
3181  */
3182 function wfBaseName( $path, $suffix = '' ) {
3183         if ( $suffix == '' ) {
3184                 $encSuffix = '';
3185         } else {
3186                 $encSuffix = '(?:' . preg_quote( $suffix, '#' ) . ')?';
3187         }
3188
3189         $matches = array();
3190         if ( preg_match( "#([^/\\\\]*?){$encSuffix}[/\\\\]*$#", $path, $matches ) ) {
3191                 return $matches[1];
3192         } else {
3193                 return '';
3194         }
3195 }
3196
3197 /**
3198  * Generate a relative path name to the given file.
3199  * May explode on non-matching case-insensitive paths,
3200  * funky symlinks, etc.
3201  *
3202  * @param string $path Absolute destination path including target filename
3203  * @param string $from Absolute source path, directory only
3204  * @return string
3205  */
3206 function wfRelativePath( $path, $from ) {
3207         // Normalize mixed input on Windows...
3208         $path = str_replace( '/', DIRECTORY_SEPARATOR, $path );
3209         $from = str_replace( '/', DIRECTORY_SEPARATOR, $from );
3210
3211         // Trim trailing slashes -- fix for drive root
3212         $path = rtrim( $path, DIRECTORY_SEPARATOR );
3213         $from = rtrim( $from, DIRECTORY_SEPARATOR );
3214
3215         $pieces = explode( DIRECTORY_SEPARATOR, dirname( $path ) );
3216         $against = explode( DIRECTORY_SEPARATOR, $from );
3217
3218         if ( $pieces[0] !== $against[0] ) {
3219                 // Non-matching Windows drive letters?
3220                 // Return a full path.
3221                 return $path;
3222         }
3223
3224         // Trim off common prefix
3225         while ( count( $pieces ) && count( $against )
3226                 && $pieces[0] == $against[0] ) {
3227                 array_shift( $pieces );
3228                 array_shift( $against );
3229         }
3230
3231         // relative dots to bump us to the parent
3232         while ( count( $against ) ) {
3233                 array_unshift( $pieces, '..' );
3234                 array_shift( $against );
3235         }
3236
3237         array_push( $pieces, wfBaseName( $path ) );
3238
3239         return implode( DIRECTORY_SEPARATOR, $pieces );
3240 }
3241
3242 /**
3243  * Convert an arbitrarily-long digit string from one numeric base
3244  * to another, optionally zero-padding to a minimum column width.
3245  *
3246  * Supports base 2 through 36; digit values 10-36 are represented
3247  * as lowercase letters a-z. Input is case-insensitive.
3248  *
3249  * @param string $input Input number
3250  * @param int $sourceBase Base of the input number
3251  * @param int $destBase Desired base of the output
3252  * @param int $pad Minimum number of digits in the output (pad with zeroes)
3253  * @param bool $lowercase Whether to output in lowercase or uppercase
3254  * @param string $engine Either "gmp", "bcmath", or "php"
3255  * @return string|bool The output number as a string, or false on error
3256  */
3257 function wfBaseConvert( $input, $sourceBase, $destBase, $pad = 1,
3258         $lowercase = true, $engine = 'auto'
3259 ) {
3260         $input = (string)$input;
3261         if (
3262                 $sourceBase < 2 ||
3263                 $sourceBase > 36 ||
3264                 $destBase < 2 ||
3265                 $destBase > 36 ||
3266                 $sourceBase != (int)$sourceBase ||
3267                 $destBase != (int)$destBase ||
3268                 $pad != (int)$pad ||
3269                 !preg_match(
3270                         "/^[" . substr( '0123456789abcdefghijklmnopqrstuvwxyz', 0, $sourceBase ) . "]+$/i",
3271                         $input
3272                 )
3273         ) {
3274                 return false;
3275         }
3276
3277         static $baseChars = array(
3278                 10 => 'a', 11 => 'b', 12 => 'c', 13 => 'd', 14 => 'e', 15 => 'f',
3279                 16 => 'g', 17 => 'h', 18 => 'i', 19 => 'j', 20 => 'k', 21 => 'l',
3280                 22 => 'm', 23 => 'n', 24 => 'o', 25 => 'p', 26 => 'q', 27 => 'r',
3281                 28 => 's', 29 => 't', 30 => 'u', 31 => 'v', 32 => 'w', 33 => 'x',
3282                 34 => 'y', 35 => 'z',
3283
3284                 '0' => 0, '1' => 1, '2' => 2, '3' => 3, '4' => 4, '5' => 5,
3285                 '6' => 6, '7' => 7, '8' => 8, '9' => 9, 'a' => 10, 'b' => 11,
3286                 'c' => 12, 'd' => 13, 'e' => 14, 'f' => 15, 'g' => 16, 'h' => 17,
3287                 'i' => 18, 'j' => 19, 'k' => 20, 'l' => 21, 'm' => 22, 'n' => 23,
3288                 'o' => 24, 'p' => 25, 'q' => 26, 'r' => 27, 's' => 28, 't' => 29,
3289                 'u' => 30, 'v' => 31, 'w' => 32, 'x' => 33, 'y' => 34, 'z' => 35
3290         );
3291
3292         if ( extension_loaded( 'gmp' ) && ( $engine == 'auto' || $engine == 'gmp' ) ) {
3293                 // Removing leading zeros works around broken base detection code in
3294                 // some PHP versions (see <https://bugs.php.net/bug.php?id=50175> and
3295                 // <https://bugs.php.net/bug.php?id=55398>).
3296                 $result = gmp_strval( gmp_init( ltrim( $input, '0' ), $sourceBase ), $destBase );
3297         } elseif ( extension_loaded( 'bcmath' ) && ( $engine == 'auto' || $engine == 'bcmath' ) ) {
3298                 $decimal = '0';
3299                 foreach ( str_split( strtolower( $input ) ) as $char ) {
3300                         $decimal = bcmul( $decimal, $sourceBase );
3301                         $decimal = bcadd( $decimal, $baseChars[$char] );
3302                 }
3303
3304                 // @codingStandardsIgnoreStart Generic.CodeAnalysis.ForLoopWithTestFunctionCall.NotAllowed
3305                 for ( $result = ''; bccomp( $decimal, 0 ); $decimal = bcdiv( $decimal, $destBase, 0 ) ) {
3306                         $result .= $baseChars[bcmod( $decimal, $destBase )];
3307                 }
3308                 // @codingStandardsIgnoreEnd
3309
3310                 $result = strrev( $result );
3311         } else {
3312                 $inDigits = array();
3313                 foreach ( str_split( strtolower( $input ) ) as $char ) {
3314                         $inDigits[] = $baseChars[$char];
3315                 }
3316
3317                 // Iterate over the input, modulo-ing out an output digit
3318                 // at a time until input is gone.
3319                 $result = '';
3320                 while ( $inDigits ) {
3321                         $work = 0;
3322                         $workDigits = array();
3323
3324                         // Long division...
3325                         foreach ( $inDigits as $digit ) {
3326                                 $work *= $sourceBase;
3327                                 $work += $digit;
3328
3329                                 if ( $workDigits || $work >= $destBase ) {
3330                                         $workDigits[] = (int)( $work / $destBase );
3331                                 }
3332                                 $work %= $destBase;
3333                         }
3334
3335                         // All that division leaves us with a remainder,
3336                         // which is conveniently our next output digit.
3337                         $result .= $baseChars[$work];
3338
3339                         // And we continue!
3340                         $inDigits = $workDigits;
3341                 }
3342
3343                 $result = strrev( $result );
3344         }
3345
3346         if ( !$lowercase ) {
3347                 $result = strtoupper( $result );
3348         }
3349
3350         return str_pad( $result, $pad, '0', STR_PAD_LEFT );
3351 }
3352
3353 /**
3354  * Check if there is sufficient entropy in php's built-in session generation
3355  *
3356  * @return bool True = there is sufficient entropy
3357  */
3358 function wfCheckEntropy() {
3359         return (
3360                         ( wfIsWindows() && version_compare( PHP_VERSION, '5.3.3', '>=' ) )
3361                         || ini_get( 'session.entropy_file' )
3362                 )
3363                 && intval( ini_get( 'session.entropy_length' ) ) >= 32;
3364 }
3365
3366 /**
3367  * Override session_id before session startup if php's built-in
3368  * session generation code is not secure.
3369  */
3370 function wfFixSessionID() {
3371         // If the cookie or session id is already set we already have a session and should abort
3372         if ( isset( $_COOKIE[session_name()] ) || session_id() ) {
3373                 return;
3374         }
3375
3376         // PHP's built-in session entropy is enabled if:
3377         // - entropy_file is set or you're on Windows with php 5.3.3+
3378         // - AND entropy_length is > 0
3379         // We treat it as disabled if it doesn't have an entropy length of at least 32
3380         $entropyEnabled = wfCheckEntropy();
3381
3382         // If built-in entropy is not enabled or not sufficient override PHP's
3383         // built in session id generation code
3384         if ( !$entropyEnabled ) {
3385                 wfDebug( __METHOD__ . ": PHP's built in entropy is disabled or not sufficient, " .
3386                         "overriding session id generation using our cryptrand source.\n" );
3387                 session_id( MWCryptRand::generateHex( 32 ) );
3388         }
3389 }
3390
3391 /**
3392  * Reset the session_id
3393  *
3394  * @since 1.22
3395  */
3396 function wfResetSessionID() {
3397         global $wgCookieSecure;
3398         $oldSessionId = session_id();
3399         $cookieParams = session_get_cookie_params();
3400         if ( wfCheckEntropy() && $wgCookieSecure == $cookieParams['secure'] ) {
3401                 session_regenerate_id( false );
3402         } else {
3403                 $tmp = $_SESSION;
3404                 session_destroy();
3405                 wfSetupSession( MWCryptRand::generateHex( 32 ) );
3406                 $_SESSION = $tmp;
3407         }
3408         $newSessionId = session_id();
3409         wfRunHooks( 'ResetSessionID', array( $oldSessionId, $newSessionId ) );
3410 }
3411
3412 /**
3413  * Initialise php session
3414  *
3415  * @param bool $sessionId
3416  */
3417 function wfSetupSession( $sessionId = false ) {
3418         global $wgSessionsInMemcached, $wgSessionsInObjectCache, $wgCookiePath, $wgCookieDomain,
3419                         $wgCookieSecure, $wgCookieHttpOnly, $wgSessionHandler;
3420         if ( $wgSessionsInObjectCache || $wgSessionsInMemcached ) {
3421                 ObjectCacheSessionHandler::install();
3422         } elseif ( $wgSessionHandler && $wgSessionHandler != ini_get( 'session.save_handler' ) ) {
3423                 # Only set this if $wgSessionHandler isn't null and session.save_handler
3424                 # hasn't already been set to the desired value (that causes errors)
3425                 ini_set( 'session.save_handler', $wgSessionHandler );
3426         }
3427         session_set_cookie_params(
3428                 0, $wgCookiePath, $wgCookieDomain, $wgCookieSecure, $wgCookieHttpOnly );
3429         session_cache_limiter( 'private, must-revalidate' );
3430         if ( $sessionId ) {
3431                 session_id( $sessionId );
3432         } else {
3433                 wfFixSessionID();
3434         }
3435         wfSuppressWarnings();
3436         session_start();
3437         wfRestoreWarnings();
3438 }
3439
3440 /**
3441  * Get an object from the precompiled serialized directory
3442  *
3443  * @param string $name
3444  * @return mixed The variable on success, false on failure
3445  */
3446 function wfGetPrecompiledData( $name ) {
3447         global $IP;
3448
3449         $file = "$IP/serialized/$name";
3450         if ( file_exists( $file ) ) {
3451                 $blob = file_get_contents( $file );
3452                 if ( $blob ) {
3453                         return unserialize( $blob );
3454                 }
3455         }
3456         return false;
3457 }
3458
3459 /**
3460  * Get a cache key
3461  *
3462  * @param string $args,...
3463  * @return string
3464  */
3465 function wfMemcKey( /*...*/ ) {
3466         global $wgCachePrefix;
3467         $prefix = $wgCachePrefix === false ? wfWikiID() : $wgCachePrefix;
3468         $args = func_get_args();
3469         $key = $prefix . ':' . implode( ':', $args );
3470         $key = str_replace( ' ', '_', $key );
3471         return $key;
3472 }
3473
3474 /**
3475  * Get a cache key for a foreign DB
3476  *
3477  * @param string $db
3478  * @param string $prefix
3479  * @param string $args,...
3480  * @return string
3481  */
3482 function wfForeignMemcKey( $db, $prefix /*...*/ ) {
3483         $args = array_slice( func_get_args(), 2 );
3484         if ( $prefix ) {
3485                 $key = "$db-$prefix:" . implode( ':', $args );
3486         } else {
3487                 $key = $db . ':' . implode( ':', $args );
3488         }
3489         return str_replace( ' ', '_', $key );
3490 }
3491
3492 /**
3493  * Get an ASCII string identifying this wiki
3494  * This is used as a prefix in memcached keys
3495  *
3496  * @return string
3497  */
3498 function wfWikiID() {
3499         global $wgDBprefix, $wgDBname;
3500         if ( $wgDBprefix ) {
3501                 return "$wgDBname-$wgDBprefix";
3502         } else {
3503                 return $wgDBname;
3504         }
3505 }
3506
3507 /**
3508  * Split a wiki ID into DB name and table prefix
3509  *
3510  * @param string $wiki
3511  *
3512  * @return array
3513  */
3514 function wfSplitWikiID( $wiki ) {
3515         $bits = explode( '-', $wiki, 2 );
3516         if ( count( $bits ) < 2 ) {
3517                 $bits[] = '';
3518         }
3519         return $bits;
3520 }
3521
3522 /**
3523  * Get a Database object.
3524  *
3525  * @param int $db Index of the connection to get. May be DB_MASTER for the
3526  *            master (for write queries), DB_SLAVE for potentially lagged read
3527  *            queries, or an integer >= 0 for a particular server.
3528  *
3529  * @param string|string[] $groups Query groups. An array of group names that this query
3530  *                belongs to. May contain a single string if the query is only
3531  *                in one group.
3532  *
3533  * @param string|bool $wiki The wiki ID, or false for the current wiki
3534  *
3535  * Note: multiple calls to wfGetDB(DB_SLAVE) during the course of one request
3536  * will always return the same object, unless the underlying connection or load
3537  * balancer is manually destroyed.
3538  *
3539  * Note 2: use $this->getDB() in maintenance scripts that may be invoked by
3540  * updater to ensure that a proper database is being updated.
3541  *
3542  * @return DatabaseBase
3543  */
3544 function &wfGetDB( $db, $groups = array(), $wiki = false ) {
3545         return wfGetLB( $wiki )->getConnection( $db, $groups, $wiki );
3546 }
3547
3548 /**
3549  * Get a load balancer object.
3550  *
3551  * @param string|bool $wiki Wiki ID, or false for the current wiki
3552  * @return LoadBalancer
3553  */
3554 function wfGetLB( $wiki = false ) {
3555         return wfGetLBFactory()->getMainLB( $wiki );
3556 }
3557
3558 /**
3559  * Get the load balancer factory object
3560  *
3561  * @return LBFactory
3562  */
3563 function &wfGetLBFactory() {
3564         return LBFactory::singleton();
3565 }
3566
3567 /**
3568  * Find a file.
3569  * Shortcut for RepoGroup::singleton()->findFile()
3570  *
3571  * @param string $title String or Title object
3572  * @param array $options Associative array of options:
3573  *     time:           requested time for an archived image, or false for the
3574  *                     current version. An image object will be returned which was
3575  *                     created at the specified time.
3576  *
3577  *     ignoreRedirect: If true, do not follow file redirects
3578  *
3579  *     private:        If true, return restricted (deleted) files if the current
3580  *                     user is allowed to view them. Otherwise, such files will not
3581  *                     be found.
3582  *
3583  *     bypassCache:    If true, do not use the process-local cache of File objects
3584  *
3585  * @return File|bool File, or false if the file does not exist
3586  */
3587 function wfFindFile( $title, $options = array() ) {
3588         return RepoGroup::singleton()->findFile( $title, $options );
3589 }
3590
3591 /**
3592  * Get an object referring to a locally registered file.
3593  * Returns a valid placeholder object if the file does not exist.
3594  *
3595  * @param Title|string $title
3596  * @return LocalFile|null A File, or null if passed an invalid Title
3597  */
3598 function wfLocalFile( $title ) {
3599         return RepoGroup::singleton()->getLocalRepo()->newFile( $title );
3600 }
3601
3602 /**
3603  * Should low-performance queries be disabled?
3604  *
3605  * @return bool
3606  * @codeCoverageIgnore
3607  */
3608 function wfQueriesMustScale() {
3609         global $wgMiserMode;
3610         return $wgMiserMode
3611                 || ( SiteStats::pages() > 100000
3612                 && SiteStats::edits() > 1000000
3613                 && SiteStats::users() > 10000 );
3614 }
3615
3616 /**
3617  * Get the path to a specified script file, respecting file
3618  * extensions; this is a wrapper around $wgScriptExtension etc.
3619  * except for 'index' and 'load' which use $wgScript/$wgLoadScript
3620  *
3621  * @param string $script Script filename, sans extension
3622  * @return string
3623  */
3624 function wfScript( $script = 'index' ) {
3625         global $wgScriptPath, $wgScriptExtension, $wgScript, $wgLoadScript;
3626         if ( $script === 'index' ) {
3627                 return $wgScript;
3628         } elseif ( $script === 'load' ) {
3629                 return $wgLoadScript;
3630         } else {
3631                 return "{$wgScriptPath}/{$script}{$wgScriptExtension}";
3632         }
3633 }
3634
3635 /**
3636  * Get the script URL.
3637  *
3638  * @return string Script URL
3639  */
3640 function wfGetScriptUrl() {
3641         if ( isset( $_SERVER['SCRIPT_NAME'] ) ) {
3642                 #
3643                 # as it was called, minus the query string.
3644                 #
3645                 # Some sites use Apache rewrite rules to handle subdomains,
3646                 # and have PHP set up in a weird way that causes PHP_SELF
3647                 # to contain the rewritten URL instead of the one that the
3648                 # outside world sees.
3649                 #
3650                 # If in this mode, use SCRIPT_URL instead, which mod_rewrite
3651                 # provides containing the "before" URL.
3652                 return $_SERVER['SCRIPT_NAME'];
3653         } else {
3654                 return $_SERVER['URL'];
3655         }
3656 }
3657
3658 /**
3659  * Convenience function converts boolean values into "true"
3660  * or "false" (string) values
3661  *
3662  * @param bool $value
3663  * @return string
3664  */
3665 function wfBoolToStr( $value ) {
3666         return $value ? 'true' : 'false';
3667 }
3668
3669 /**
3670  * Get a platform-independent path to the null file, e.g. /dev/null
3671  *
3672  * @return string
3673  */
3674 function wfGetNull() {
3675         return wfIsWindows() ? 'NUL' : '/dev/null';
3676 }
3677
3678 /**
3679  * Waits for the slaves to catch up to the master position
3680  *
3681  * Use this when updating very large numbers of rows, as in maintenance scripts,
3682  * to avoid causing too much lag. Of course, this is a no-op if there are no slaves.
3683  *
3684  * By default this waits on the main DB cluster of the current wiki.
3685  * If $cluster is set to "*" it will wait on all DB clusters, including
3686  * external ones. If the lag being waiting on is caused by the code that
3687  * does this check, it makes since to use $ifWritesSince, particularly if
3688  * cluster is "*", to avoid excess overhead.
3689  *
3690  * Never call this function after a big DB write that is still in a transaction.
3691  * This only makes sense after the possible lag inducing changes were committed.
3692  *
3693  * @param float|null $ifWritesSince Only wait if writes were done since this UNIX timestamp
3694  * @param string|bool $wiki Wiki identifier accepted by wfGetLB
3695  * @param string|bool $cluster Cluster name accepted by LBFactory. Default: false.
3696  * @param int|null $timeout Max wait time. Default: 1 day (cli), ~10 seconds (web)
3697  * @return bool Success (able to connect and no timeouts reached)
3698  */
3699 function wfWaitForSlaves(
3700         $ifWritesSince = null, $wiki = false, $cluster = false, $timeout = null
3701 ) {
3702         // B/C: first argument used to be "max seconds of lag"; ignore such values
3703         $ifWritesSince = ( $ifWritesSince > 1e9 ) ? $ifWritesSince : null;
3704
3705         if ( $timeout === null ) {
3706                 $timeout = ( PHP_SAPI === 'cli' ) ? 86400 : 10;
3707         }
3708
3709         // Figure out which clusters need to be checked
3710         $lbs = array();
3711         if ( $cluster === '*' ) {
3712                 wfGetLBFactory()->forEachLB( function ( LoadBalancer $lb ) use ( &$lbs ) {
3713                         $lbs[] = $lb;
3714                 } );
3715         } elseif ( $cluster !== false ) {
3716                 $lbs[] = wfGetLBFactory()->getExternalLB( $cluster );
3717         } else {
3718                 $lbs[] = wfGetLB( $wiki );
3719         }
3720
3721         // Get all the master positions of applicable DBs right now.
3722         // This can be faster since waiting on one cluster reduces the
3723         // time needed to wait on the next clusters.
3724         $masterPositions = array_fill( 0, count( $lbs ), false );
3725         foreach ( $lbs as $i => $lb ) {
3726                 // bug 27975 - Don't try to wait for slaves if there are none
3727                 // Prevents permission error when getting master position
3728                 if ( $lb->getServerCount() > 1 ) {
3729                         if ( $ifWritesSince && !$lb->hasMasterConnection() ) {
3730                                 continue; // assume no writes done
3731                         }
3732                         // Use the empty string to not trigger selectDB() since the connection
3733                         // may have been to a server that does not have a DB for the current wiki.
3734                         $dbw = $lb->getConnection( DB_MASTER, array(), '' );
3735                         if ( $ifWritesSince && $dbw->lastDoneWrites() < $ifWritesSince ) {
3736                                 continue; // no writes since the last wait
3737                         }
3738                         $masterPositions[$i] = $dbw->getMasterPos();
3739                 }
3740         }
3741
3742         $ok = true;
3743         foreach ( $lbs as $i => $lb ) {
3744                 if ( $masterPositions[$i] ) {
3745                         // The DBMS may not support getMasterPos() or the whole
3746                         // load balancer might be fake (e.g. $wgAllDBsAreLocalhost).
3747                         $ok = $lb->waitForAll( $masterPositions[$i], $timeout ) && $ok;
3748                 }
3749         }
3750
3751         return $ok;
3752 }
3753
3754 /**
3755  * Count down from $seconds to zero on the terminal, with a one-second pause
3756  * between showing each number. For use in command-line scripts.
3757  *
3758  * @codeCoverageIgnore
3759  * @param int $seconds
3760  */
3761 function wfCountDown( $seconds ) {
3762         for ( $i = $seconds; $i >= 0; $i-- ) {
3763                 if ( $i != $seconds ) {
3764                         echo str_repeat( "\x08", strlen( $i + 1 ) );
3765                 }
3766                 echo $i;
3767                 flush();
3768                 if ( $i ) {
3769                         sleep( 1 );
3770                 }
3771         }
3772         echo "\n";
3773 }
3774
3775 /**
3776  * Replace all invalid characters with -
3777  * Additional characters can be defined in $wgIllegalFileChars (see bug 20489)
3778  * By default, $wgIllegalFileChars = ':'
3779  *
3780  * @param string $name Filename to process
3781  * @return string
3782  */
3783 function wfStripIllegalFilenameChars( $name ) {
3784         global $wgIllegalFileChars;
3785         $illegalFileChars = $wgIllegalFileChars ? "|[" . $wgIllegalFileChars . "]" : '';
3786         $name = wfBaseName( $name );
3787         $name = preg_replace(
3788                 "/[^" . Title::legalChars() . "]" . $illegalFileChars . "/",
3789                 '-',
3790                 $name
3791         );
3792         return $name;
3793 }
3794
3795 /**
3796  * Set PHP's memory limit to the larger of php.ini or $wgMemoryLimit;
3797  *
3798  * @return int Value the memory limit was set to.
3799  */
3800 function wfMemoryLimit() {
3801         global $wgMemoryLimit;
3802         $memlimit = wfShorthandToInteger( ini_get( 'memory_limit' ) );
3803         if ( $memlimit != -1 ) {
3804                 $conflimit = wfShorthandToInteger( $wgMemoryLimit );
3805                 if ( $conflimit == -1 ) {
3806                         wfDebug( "Removing PHP's memory limit\n" );
3807                         wfSuppressWarnings();
3808                         ini_set( 'memory_limit', $conflimit );
3809                         wfRestoreWarnings();
3810                         return $conflimit;
3811                 } elseif ( $conflimit > $memlimit ) {
3812                         wfDebug( "Raising PHP's memory limit to $conflimit bytes\n" );
3813                         wfSuppressWarnings();
3814                         ini_set( 'memory_limit', $conflimit );
3815                         wfRestoreWarnings();
3816                         return $conflimit;
3817                 }
3818         }
3819         return $memlimit;
3820 }
3821
3822 /**
3823  * Converts shorthand byte notation to integer form
3824  *
3825  * @param string $string
3826  * @return int
3827  */
3828 function wfShorthandToInteger( $string = '' ) {
3829         $string = trim( $string );
3830         if ( $string === '' ) {
3831                 return -1;
3832         }
3833         $last = $string[strlen( $string ) - 1];
3834         $val = intval( $string );
3835         switch ( $last ) {
3836                 case 'g':
3837                 case 'G':
3838                         $val *= 1024;
3839                         // break intentionally missing
3840                 case 'm':
3841                 case 'M':
3842                         $val *= 1024;
3843                         // break intentionally missing
3844                 case 'k':
3845                 case 'K':
3846                         $val *= 1024;
3847         }
3848
3849         return $val;
3850 }
3851
3852 /**
3853  * Get the normalised IETF language tag
3854  * See unit test for examples.
3855  *
3856  * @param string $code The language code.
3857  * @return string The language code which complying with BCP 47 standards.
3858  */
3859 function wfBCP47( $code ) {
3860         $codeSegment = explode( '-', $code );
3861         $codeBCP = array();
3862         foreach ( $codeSegment as $segNo => $seg ) {
3863                 // when previous segment is x, it is a private segment and should be lc
3864                 if ( $segNo > 0 && strtolower( $codeSegment[( $segNo - 1 )] ) == 'x' ) {
3865                         $codeBCP[$segNo] = strtolower( $seg );
3866                 // ISO 3166 country code
3867                 } elseif ( ( strlen( $seg ) == 2 ) && ( $segNo > 0 ) ) {
3868                         $codeBCP[$segNo] = strtoupper( $seg );
3869                 // ISO 15924 script code
3870                 } elseif ( ( strlen( $seg ) == 4 ) && ( $segNo > 0 ) ) {
3871                         $codeBCP[$segNo] = ucfirst( strtolower( $seg ) );
3872                 // Use lowercase for other cases
3873                 } else {
3874                         $codeBCP[$segNo] = strtolower( $seg );
3875                 }
3876         }
3877         $langCode = implode( '-', $codeBCP );
3878         return $langCode;
3879 }
3880
3881 /**
3882  * Get a cache object.
3883  *
3884  * @param int $inputType Cache type, one the the CACHE_* constants.
3885  * @return BagOStuff
3886  */
3887 function wfGetCache( $inputType ) {
3888         return ObjectCache::getInstance( $inputType );
3889 }
3890
3891 /**
3892  * Get the main cache object
3893  *
3894  * @return BagOStuff
3895  */
3896 function wfGetMainCache() {
3897         global $wgMainCacheType;
3898         return ObjectCache::getInstance( $wgMainCacheType );
3899 }
3900
3901 /**
3902  * Get the cache object used by the message cache
3903  *
3904  * @return BagOStuff
3905  */
3906 function wfGetMessageCacheStorage() {
3907         global $wgMessageCacheType;
3908         return ObjectCache::getInstance( $wgMessageCacheType );
3909 }
3910
3911 /**
3912  * Get the cache object used by the parser cache
3913  *
3914  * @return BagOStuff
3915  */
3916 function wfGetParserCacheStorage() {
3917         global $wgParserCacheType;
3918         return ObjectCache::getInstance( $wgParserCacheType );
3919 }
3920
3921 /**
3922  * Get the cache object used by the language converter
3923  *
3924  * @return BagOStuff
3925  */
3926 function wfGetLangConverterCacheStorage() {
3927         global $wgLanguageConverterCacheType;
3928         return ObjectCache::getInstance( $wgLanguageConverterCacheType );
3929 }
3930
3931 /**
3932  * Call hook functions defined in $wgHooks
3933  *
3934  * @param string $event Event name
3935  * @param array $args Parameters passed to hook functions
3936  * @param string|null $deprecatedVersion Optionally mark hook as deprecated with version number
3937  *
3938  * @return bool True if no handler aborted the hook
3939  */
3940 function wfRunHooks( $event, array $args = array(), $deprecatedVersion = null ) {
3941         return Hooks::run( $event, $args, $deprecatedVersion );
3942 }
3943
3944 /**
3945  * Wrapper around php's unpack.
3946  *
3947  * @param string $format The format string (See php's docs)
3948  * @param string $data A binary string of binary data
3949  * @param int|bool $length The minimum length of $data or false. This is to
3950  *      prevent reading beyond the end of $data. false to disable the check.
3951  *
3952  * Also be careful when using this function to read unsigned 32 bit integer
3953  * because php might make it negative.
3954  *
3955  * @throws MWException If $data not long enough, or if unpack fails
3956  * @return array Associative array of the extracted data
3957  */
3958 function wfUnpack( $format, $data, $length = false ) {
3959         if ( $length !== false ) {
3960                 $realLen = strlen( $data );
3961                 if ( $realLen < $length ) {
3962                         throw new MWException( "Tried to use wfUnpack on a "
3963                                 . "string of length $realLen, but needed one "
3964                                 . "of at least length $length."
3965                         );
3966                 }
3967         }
3968
3969         wfSuppressWarnings();
3970         $result = unpack( $format, $data );
3971         wfRestoreWarnings();
3972
3973         if ( $result === false ) {
3974                 // If it cannot extract the packed data.
3975                 throw new MWException( "unpack could not unpack binary data" );
3976         }
3977         return $result;
3978 }
3979
3980 /**
3981  * Determine if an image exists on the 'bad image list'.
3982  *
3983  * The format of MediaWiki:Bad_image_list is as follows:
3984  *    * Only list items (lines starting with "*") are considered
3985  *    * The first link on a line must be a link to a bad image
3986  *    * Any subsequent links on the same line are considered to be exceptions,
3987  *      i.e. articles where the image may occur inline.
3988  *
3989  * @param string $name The image name to check
3990  * @param Title|bool $contextTitle The page on which the image occurs, if known
3991  * @param string $blacklist Wikitext of a file blacklist
3992  * @return bool
3993  */
3994 function wfIsBadImage( $name, $contextTitle = false, $blacklist = null ) {
3995         static $badImageCache = null; // based on bad_image_list msg
3996         wfProfileIn( __METHOD__ );
3997
3998         # Handle redirects
3999         $redirectTitle = RepoGroup::singleton()->checkRedirect( Title::makeTitle( NS_FILE, $name ) );
4000         if ( $redirectTitle ) {
4001                 $name = $redirectTitle->getDBkey();
4002         }
4003
4004         # Run the extension hook
4005         $bad = false;
4006         if ( !wfRunHooks( 'BadImage', array( $name, &$bad ) ) ) {
4007                 wfProfileOut( __METHOD__ );
4008                 return $bad;
4009         }
4010
4011         $cacheable = ( $blacklist === null );
4012         if ( $cacheable && $badImageCache !== null ) {
4013                 $badImages = $badImageCache;
4014         } else { // cache miss
4015                 if ( $blacklist === null ) {
4016                         $blacklist = wfMessage( 'bad_image_list' )->inContentLanguage()->plain(); // site list
4017                 }
4018                 # Build the list now
4019                 $badImages = array();
4020                 $lines = explode( "\n", $blacklist );
4021                 foreach ( $lines as $line ) {
4022                         # List items only
4023                         if ( substr( $line, 0, 1 ) !== '*' ) {
4024                                 continue;
4025                         }
4026
4027                         # Find all links
4028                         $m = array();
4029                         if ( !preg_match_all( '/\[\[:?(.*?)\]\]/', $line, $m ) ) {
4030                                 continue;
4031                         }
4032
4033                         $exceptions = array();
4034                         $imageDBkey = false;
4035                         foreach ( $m[1] as $i => $titleText ) {
4036                                 $title = Title::newFromText( $titleText );
4037                                 if ( !is_null( $title ) ) {
4038                                         if ( $i == 0 ) {
4039                                                 $imageDBkey = $title->getDBkey();
4040                                         } else {
4041                                                 $exceptions[$title->getPrefixedDBkey()] = true;
4042                                         }
4043                                 }
4044                         }
4045
4046                         if ( $imageDBkey !== false ) {
4047                                 $badImages[$imageDBkey] = $exceptions;
4048                         }
4049                 }
4050                 if ( $cacheable ) {
4051                         $badImageCache = $badImages;
4052                 }
4053         }
4054
4055         $contextKey = $contextTitle ? $contextTitle->getPrefixedDBkey() : false;
4056         $bad = isset( $badImages[$name] ) && !isset( $badImages[$name][$contextKey] );
4057         wfProfileOut( __METHOD__ );
4058         return $bad;
4059 }
4060
4061 /**
4062  * Determine whether the client at a given source IP is likely to be able to
4063  * access the wiki via HTTPS.
4064  *
4065  * @param string $ip The IPv4/6 address in the normal human-readable form
4066  * @return bool
4067  */
4068 function wfCanIPUseHTTPS( $ip ) {
4069         $canDo = true;
4070         wfRunHooks( 'CanIPUseHTTPS', array( $ip, &$canDo ) );
4071         return !!$canDo;
4072 }
4073
4074 /**
4075  * Work out the IP address based on various globals
4076  * For trusted proxies, use the XFF client IP (first of the chain)
4077  *
4078  * @deprecated since 1.19; call $wgRequest->getIP() directly.
4079  * @return string
4080  */
4081 function wfGetIP() {
4082         wfDeprecated( __METHOD__, '1.19' );
4083         global $wgRequest;
4084         return $wgRequest->getIP();
4085 }
4086
4087 /**
4088  * Checks if an IP is a trusted proxy provider.
4089  * Useful to tell if X-Forwarded-For data is possibly bogus.
4090  * Squid cache servers for the site are whitelisted.
4091  * @deprecated Since 1.24, use IP::isTrustedProxy()
4092  *
4093  * @param string $ip
4094  * @return bool
4095  */
4096 function wfIsTrustedProxy( $ip ) {
4097         return IP::isTrustedProxy( $ip );
4098 }
4099
4100 /**
4101  * Checks if an IP matches a proxy we've configured.
4102  * @deprecated Since 1.24, use IP::isConfiguredProxy()
4103  *
4104  * @param string $ip
4105  * @return bool
4106  * @since 1.23 Supports CIDR ranges in $wgSquidServersNoPurge
4107  */
4108 function wfIsConfiguredProxy( $ip ) {
4109         return IP::isConfiguredProxy( $ip );
4110 }