includes/MagicWord.php

   1 <?php
   2 /**
   3  * File for magic words
   4  * See docs/magicword.txt
   5  *
   6  * @file
   7  * @ingroup Parser
   8  */
   9
  10 /**
  11  * This class encapsulates "magic words" such as #redirect, __NOTOC__, etc.
  12  * Usage:
  13  *     if (MagicWord::get( 'redirect' )->match( $text ) )
  14  *
  15  * Possible future improvements:
  16  *   * Simultaneous searching for a number of magic words
  17  *   * MagicWord::$mObjects in shared memory
  18  *
  19  * Please avoid reading the data out of one of these objects and then writing
  20  * special case code. If possible, add another match()-like function here.
  21  *
  22  * To add magic words in an extension, use the LanguageGetMagic hook. For
  23  * magic words which are also Parser variables, add a MagicWordwgVariableIDs
  24  * hook. Use string keys.
  25  *
  26  * @ingroup Parser
  27  */
  28 class MagicWord {
  29         /**#@+
  30          * @private
  31          */
  32         var $mId, $mSynonyms, $mCaseSensitive, $mRegex;
  33         var $mRegexStart, $mBaseRegex, $mVariableRegex;
  34         var $mModified, $mFound;
  35
  36         static public $mVariableIDsInitialised = false;
  37         static public $mVariableIDs = array(
  38                 'currentmonth',
  39                 'currentmonthname',
  40                 'currentmonthnamegen',
  41                 'currentmonthabbrev',
  42                 'currentday',
  43                 'currentday2',
  44                 'currentdayname',
  45                 'currentyear',
  46                 'currenttime',
  47                 'currenthour',
  48                 'localmonth',
  49                 'localmonthname',
  50                 'localmonthnamegen',
  51                 'localmonthabbrev',
  52                 'localday',
  53                 'localday2',
  54                 'localdayname',
  55                 'localyear',
  56                 'localtime',
  57                 'localhour',
  58                 'numberofarticles',
  59                 'numberoffiles',
  60                 'numberofedits',
  61                 'sitename',
  62                 'server',
  63                 'servername',
  64                 'scriptpath',
  65                 'pagename',
  66                 'pagenamee',
  67                 'fullpagename',
  68                 'fullpagenamee',
  69                 'namespace',
  70                 'namespacee',
  71                 'currentweek',
  72                 'currentdow',
  73                 'localweek',
  74                 'localdow',
  75                 'revisionid',
  76                 'revisionday',
  77                 'revisionday2',
  78                 'revisionmonth',
  79                 'revisionyear',
  80                 'revisiontimestamp',
  81                 'subpagename',
  82                 'subpagenamee',
  83                 'displaytitle',
  84                 'talkspace',
  85                 'talkspacee',
  86                 'subjectspace',
  87                 'subjectspacee',
  88                 'talkpagename',
  89                 'talkpagenamee',
  90                 'subjectpagename',
  91                 'subjectpagenamee',
  92                 'numberofusers',
  93                 'newsectionlink',
  94                 'numberofpages',
  95                 'currentversion',
  96                 'basepagename',
  97                 'basepagenamee',
  98                 'urlencode',
  99                 'currenttimestamp',
 100                 'localtimestamp',
 101                 'directionmark',
 102                 'language',
 103                 'contentlanguage',
 104                 'pagesinnamespace',
 105                 'numberofadmins',
 106                 'defaultsort',
 107                 'pagesincategory',
 108                 'index',
 109                 'noindex',
 110                 'numberingroup',
 111         );
 112
 113         /* Array of caching hints for ParserCache */
 114         static public $mCacheTTLs = array (
 115                 'currentmonth' => 86400,
 116                 'currentmonthname' => 86400,
 117                 'currentmonthnamegen' => 86400,
 118                 'currentmonthabbrev' => 86400,
 119                 'currentday' => 3600,
 120                 'currentday2' => 3600,
 121                 'currentdayname' => 3600,
 122                 'currentyear' => 86400,
 123                 'currenttime' => 3600,
 124                 'currenthour' => 3600,
 125                 'localmonth' => 86400,
 126                 'localmonthname' => 86400,
 127                 'localmonthnamegen' => 86400,
 128                 'localmonthabbrev' => 86400,
 129                 'localday' => 3600,
 130                 'localday2' => 3600,
 131                 'localdayname' => 3600,
 132                 'localyear' => 86400,
 133                 'localtime' => 3600,
 134                 'localhour' => 3600,
 135                 'numberofarticles' => 3600,
 136                 'numberoffiles' => 3600,
 137                 'numberofedits' => 3600,
 138                 'currentweek' => 3600,
 139                 'currentdow' => 3600,
 140                 'localweek' => 3600,
 141                 'localdow' => 3600,
 142                 'numberofusers' => 3600,
 143                 'numberofpages' => 3600,
 144                 'currentversion' => 86400,
 145                 'currenttimestamp' => 3600,
 146                 'localtimestamp' => 3600,
 147                 'pagesinnamespace' => 3600,
 148                 'numberofadmins' => 3600,
 149                 'numberingroup' => 3600,
 150                 );
 151
 152         static public $mDoubleUnderscoreIDs = array(
 153                 'notoc',
 154                 'nogallery',
 155                 'forcetoc',
 156                 'toc',
 157                 'noeditsection',
 158                 'newsectionlink',
 159                 'hiddencat',
 160                 'index',
 161                 'noindex',
 162                 'staticredirect',
 163                 'noheader',
 164         );
 165
 166
 167         static public $mObjects = array();
 168         static public $mDoubleUnderscoreArray = null;
 169
 170         /**#@-*/
 171
 172         function __construct($id = 0, $syn = '', $cs = false) {
 173                 $this->mId = $id;
 174                 $this->mSynonyms = (array)$syn;
 175                 $this->mCaseSensitive = $cs;
 176                 $this->mRegex = '';
 177                 $this->mRegexStart = '';
 178                 $this->mVariableRegex = '';
 179                 $this->mVariableStartToEndRegex = '';
 180                 $this->mModified = false;
 181         }
 182
 183         /**
 184          * Factory: creates an object representing an ID
 185          * @static
 186          */
 187         static function &get( $id ) {
 188                 wfProfileIn( __METHOD__ );
 189                 if (!array_key_exists( $id, self::$mObjects ) ) {
 190                         $mw = new MagicWord();
 191                         $mw->load( $id );
 192                         self::$mObjects[$id] = $mw;
 193                 }
 194                 wfProfileOut( __METHOD__ );
 195                 return self::$mObjects[$id];
 196         }
 197
 198         /**
 199          * Get an array of parser variable IDs
 200          */
 201         static function getVariableIDs() {
 202                 if ( !self::$mVariableIDsInitialised ) {
 203                         # Deprecated constant definition hook, available for extensions that need it
 204                         $magicWords = array();
 205                         wfRunHooks( 'MagicWordMagicWords', array( &$magicWords ) );
 206                         foreach ( $magicWords as $word ) {
 207                                 define( $word, $word );
 208                         }
 209
 210                         # Get variable IDs
 211                         wfRunHooks( 'MagicWordwgVariableIDs', array( &self::$mVariableIDs ) );
 212                         self::$mVariableIDsInitialised = true;
 213                 }
 214                 return self::$mVariableIDs;
 215         }
 216
 217         /* Allow external reads of TTL array */
 218         static function getCacheTTL($id) {
 219                 if (array_key_exists($id,self::$mCacheTTLs)) {
 220                         return self::$mCacheTTLs[$id];
 221                 } else {
 222                         return -1;
 223                 }
 224         }
 225
 226         /** Get a MagicWordArray of double-underscore entities */
 227         static function getDoubleUnderscoreArray() {
 228                 if ( is_null( self::$mDoubleUnderscoreArray ) ) {
 229                         self::$mDoubleUnderscoreArray = new MagicWordArray( self::$mDoubleUnderscoreIDs );
 230                 }
 231                 return self::$mDoubleUnderscoreArray;
 232         }
 233
 234         # Initialises this object with an ID
 235         function load( $id ) {
 236                 global $wgContLang;
 237                 $this->mId = $id;
 238                 $wgContLang->getMagic( $this );
 239                 if ( !$this->mSynonyms ) {
 240                         $this->mSynonyms = array( 'dkjsagfjsgashfajsh' );
 241                         #throw new MWException( "Error: invalid magic word '$id'" );
 242                         wfDebugLog( 'exception', "Error: invalid magic word '$id'\n" );
 243                 }
 244         }
 245
 246         /**
 247          * Preliminary initialisation
 248          * @private
 249          */
 250         function initRegex() {
 251                 #$variableClass = Title::legalChars();
 252                 # This was used for matching "$1" variables, but different uses of the feature will have
 253                 # different restrictions, which should be checked *after* the MagicWord has been matched,
 254                 # not here. - IMSoP
 255
 256                 $escSyn = array();
 257                 foreach ( $this->mSynonyms as $synonym )
 258                         // In case a magic word contains /, like that's going to happen;)
 259                         $escSyn[] = preg_quote( $synonym, '/' );
 260                 $this->mBaseRegex = implode( '|', $escSyn );
 261
 262                 $case = $this->mCaseSensitive ? '' : 'iu';
 263                 $this->mRegex = "/{$this->mBaseRegex}/{$case}";
 264                 $this->mRegexStart = "/^(?:{$this->mBaseRegex})/{$case}";
 265                 $this->mVariableRegex = str_replace( "\\$1", "(.*?)", $this->mRegex );
 266                 $this->mVariableStartToEndRegex = str_replace( "\\$1", "(.*?)",
 267                         "/^(?:{$this->mBaseRegex})$/{$case}" );
 268         }
 269
 270         /**
 271          * Gets a regex representing matching the word
 272          */
 273         function getRegex() {
 274                 if ($this->mRegex == '' ) {
 275                         $this->initRegex();
 276                 }
 277                 return $this->mRegex;
 278         }
 279
 280         /**
 281          * Gets the regexp case modifier to use, i.e. i or nothing, to be used if
 282          * one is using MagicWord::getBaseRegex(), otherwise it'll be included in
 283          * the complete expression
 284          */
 285         function getRegexCase() {
 286                 if ( $this->mRegex === '' )
 287                         $this->initRegex();
 288
 289                 return $this->mCaseSensitive ? '' : 'iu';
 290         }
 291
 292         /**
 293          * Gets a regex matching the word, if it is at the string start
 294          */
 295         function getRegexStart() {
 296                 if ($this->mRegex == '' ) {
 297                         $this->initRegex();
 298                 }
 299                 return $this->mRegexStart;
 300         }
 301
 302         /**
 303          * regex without the slashes and what not
 304          */
 305         function getBaseRegex() {
 306                 if ($this->mRegex == '') {
 307                         $this->initRegex();
 308                 }
 309                 return $this->mBaseRegex;
 310         }
 311
 312         /**
 313          * Returns true if the text contains the word
 314          * @return bool
 315          */
 316         function match( $text ) {
 317                 return preg_match( $this->getRegex(), $text );
 318         }
 319
 320         /**
 321          * Returns true if the text starts with the word
 322          * @return bool
 323          */
 324         function matchStart( $text ) {
 325                 return preg_match( $this->getRegexStart(), $text );
 326         }
 327
 328         /**
 329          * Returns NULL if there's no match, the value of $1 otherwise
 330          * The return code is the matched string, if there's no variable
 331          * part in the regex and the matched variable part ($1) if there
 332          * is one.
 333          */
 334         function matchVariableStartToEnd( $text ) {
 335                 $matches = array();
 336                 $matchcount = preg_match( $this->getVariableStartToEndRegex(), $text, $matches );
 337                 if ( $matchcount == 0 ) {
 338                         return NULL;
 339                 } else {
 340                         # multiple matched parts (variable match); some will be empty because of
 341                         # synonyms. The variable will be the second non-empty one so remove any
 342                         # blank elements and re-sort the indices.
 343                         # See also bug 6526
 344
 345                         $matches = array_values(array_filter($matches));
 346
 347                         if ( count($matches) == 1 ) { return $matches[0]; }
 348                         else { return $matches[1]; }
 349                 }
 350         }
 351
 352
 353         /**
 354          * Returns true if the text matches the word, and alters the
 355          * input string, removing all instances of the word
 356          */
 357         function matchAndRemove( &$text ) {
 358                 $this->mFound = false;
 359                 $text = preg_replace_callback( $this->getRegex(), array( &$this, 'pregRemoveAndRecord' ), $text );
 360                 return $this->mFound;
 361         }
 362
 363         function matchStartAndRemove( &$text ) {
 364                 $this->mFound = false;
 365                 $text = preg_replace_callback( $this->getRegexStart(), array( &$this, 'pregRemoveAndRecord' ), $text );
 366                 return $this->mFound;
 367         }
 368
 369         /**
 370          * Used in matchAndRemove()
 371          * @private
 372          **/
 373         function pregRemoveAndRecord( ) {
 374                 $this->mFound = true;
 375                 return '';
 376         }
 377
 378         /**
 379          * Replaces the word with something else
 380          */
 381         function replace( $replacement, $subject, $limit=-1 ) {
 382                 $res = preg_replace( $this->getRegex(), StringUtils::escapeRegexReplacement( $replacement ), $subject, $limit );
 383                 $this->mModified = !($res === $subject);
 384                 return $res;
 385         }
 386
 387         /**
 388          * Variable handling: {{SUBST:xxx}} style words
 389          * Calls back a function to determine what to replace xxx with
 390          * Input word must contain $1
 391          */
 392         function substituteCallback( $text, $callback ) {
 393                 $res = preg_replace_callback( $this->getVariableRegex(), $callback, $text );
 394                 $this->mModified = !($res === $text);
 395                 return $res;
 396         }
 397
 398         /**
 399          * Matches the word, where $1 is a wildcard
 400          */
 401         function getVariableRegex()     {
 402                 if ( $this->mVariableRegex == '' ) {
 403                         $this->initRegex();
 404                 }
 405                 return $this->mVariableRegex;
 406         }
 407
 408         /**
 409          * Matches the entire string, where $1 is a wildcard
 410          */
 411         function getVariableStartToEndRegex() {
 412                 if ( $this->mVariableStartToEndRegex == '' ) {
 413                         $this->initRegex();
 414                 }
 415                 return $this->mVariableStartToEndRegex;
 416         }
 417
 418         /**
 419          * Accesses the synonym list directly
 420          */
 421         function getSynonym( $i ) {
 422                 return $this->mSynonyms[$i];
 423         }
 424
 425         function getSynonyms() {
 426                 return $this->mSynonyms;
 427         }
 428
 429         /**
 430          * Returns true if the last call to replace() or substituteCallback()
 431          * returned a modified text, otherwise false.
 432          */
 433         function getWasModified(){
 434                 return $this->mModified;
 435         }
 436
 437         /**
 438          * $magicarr is an associative array of (magic word ID => replacement)
 439          * This method uses the php feature to do several replacements at the same time,
 440          * thereby gaining some efficiency. The result is placed in the out variable
 441          * $result. The return value is true if something was replaced.
 442          * @static
 443          **/
 444         function replaceMultiple( $magicarr, $subject, &$result ){
 445                 $search = array();
 446                 $replace = array();
 447                 foreach( $magicarr as $id => $replacement ){
 448                         $mw = MagicWord::get( $id );
 449                         $search[] = $mw->getRegex();
 450                         $replace[] = $replacement;
 451                 }
 452
 453                 $result = preg_replace( $search, $replace, $subject );
 454                 return !($result === $subject);
 455         }
 456
 457         /**
 458          * Adds all the synonyms of this MagicWord to an array, to allow quick
 459          * lookup in a list of magic words
 460          */
 461         function addToArray( &$array, $value ) {
 462                 global $wgContLang;
 463                 foreach ( $this->mSynonyms as $syn ) {
 464                         $array[$wgContLang->lc($syn)] = $value;
 465                 }
 466         }
 467
 468         function isCaseSensitive() {
 469                 return $this->mCaseSensitive;
 470         }
 471
 472         function getId() {
 473                 return $this->mId;
 474         }
 475 }
 476
 477 /**
 478  * Class for handling an array of magic words
 479  * @ingroup Parser
 480  */
 481 class MagicWordArray {
 482         var $names = array();
 483         var $hash;
 484         var $baseRegex, $regex;
 485         var $matches;
 486
 487         function __construct( $names = array() ) {
 488                 $this->names = $names;
 489         }
 490
 491         /**
 492          * Add a magic word by name
 493          */
 494         public function add( $name ) {
 495                 global $wgContLang;
 496                 $this->names[] = $name;
 497                 $this->hash = $this->baseRegex = $this->regex = null;
 498         }
 499
 500         /**
 501          * Add a number of magic words by name
 502          */
 503         public function addArray( $names ) {
 504                 $this->names = array_merge( $this->names, array_values( $names ) );
 505                 $this->hash = $this->baseRegex = $this->regex = null;
 506         }
 507
 508         /**
 509          * Get a 2-d hashtable for this array
 510          */
 511         function getHash() {
 512                 if ( is_null( $this->hash ) ) {
 513                         global $wgContLang;
 514                         $this->hash = array( 0 => array(), 1 => array() );
 515                         foreach ( $this->names as $name ) {
 516                                 $magic = MagicWord::get( $name );
 517                                 $case = intval( $magic->isCaseSensitive() );
 518                                 foreach ( $magic->getSynonyms() as $syn ) {
 519                                         if ( !$case ) {
 520                                                 $syn = $wgContLang->lc( $syn );
 521                                         }
 522                                         $this->hash[$case][$syn] = $name;
 523                                 }
 524                         }
 525                 }
 526                 return $this->hash;
 527         }
 528
 529         /**
 530          * Get the base regex
 531          */
 532         function getBaseRegex() {
 533                 if ( is_null( $this->baseRegex ) ) {
 534                         $this->baseRegex = array( 0 => '', 1 => '' );
 535                         foreach ( $this->names as $name ) {
 536                                 $magic = MagicWord::get( $name );
 537                                 $case = intval( $magic->isCaseSensitive() );
 538                                 foreach ( $magic->getSynonyms() as $i => $syn ) {
 539                                         $group = "(?P<{$i}_{$name}>" . preg_quote( $syn, '/' ) . ')';
 540                                         if ( $this->baseRegex[$case] === '' ) {
 541                                                 $this->baseRegex[$case] = $group;
 542                                         } else {
 543                                                 $this->baseRegex[$case] .= '|' . $group;
 544                                         }
 545                                 }
 546                         }
 547                 }
 548                 return $this->baseRegex;
 549         }
 550
 551         /**
 552          * Get an unanchored regex
 553          */
 554         function getRegex() {
 555                 if ( is_null( $this->regex ) ) {
 556                         $base = $this->getBaseRegex();
 557                         $this->regex = array( '', '' );
 558                         if ( $this->baseRegex[0] !== '' ) {
 559                                 $this->regex[0] = "/{$base[0]}/iuS";
 560                         }
 561                         if ( $this->baseRegex[1] !== '' ) {
 562                                 $this->regex[1] = "/{$base[1]}/S";
 563                         }
 564                 }
 565                 return $this->regex;
 566         }
 567
 568         /**
 569          * Get a regex for matching variables
 570          */
 571         function getVariableRegex() {
 572                 return str_replace( "\\$1", "(.*?)", $this->getRegex() );
 573         }
 574
 575         /**
 576          * Get an anchored regex for matching variables
 577          */
 578         function getVariableStartToEndRegex() {
 579                 $base = $this->getBaseRegex();
 580                 $newRegex = array( '', '' );
 581                 if ( $base[0] !== '' ) {
 582                         $newRegex[0] = str_replace( "\\$1", "(.*?)", "/^(?:{$base[0]})$/iuS" );
 583                 }
 584                 if ( $base[1] !== '' ) {
 585                         $newRegex[1] = str_replace( "\\$1", "(.*?)", "/^(?:{$base[1]})$/S" );
 586                 }
 587                 return $newRegex;
 588         }
 589
 590         /**
 591          * Parse a match array from preg_match
 592          * Returns array(magic word ID, parameter value)
 593          * If there is no parameter value, that element will be false.
 594          */
 595         function parseMatch( $m ) {
 596                 reset( $m );
 597                 while ( list( $key, $value ) = each( $m ) ) {
 598                         if ( $key === 0 || $value === '' ) {
 599                                 continue;
 600                         }
 601                         $parts = explode( '_', $key, 2 );
 602                         if ( count( $parts ) != 2 ) {
 603                                 // This shouldn't happen
 604                                 // continue;
 605                                 throw new MWException( __METHOD__ . ': bad parameter name' );
 606                         }
 607                         list( /* $synIndex */, $magicName ) = $parts;
 608                         $paramValue = next( $m );
 609                         return array( $magicName, $paramValue );
 610                 }
 611                 // This shouldn't happen either
 612                 throw new MWException( __METHOD__.': parameter not found' );
 613                 return array( false, false );
 614         }
 615
 616         /**
 617          * Match some text, with parameter capture
 618          * Returns an array with the magic word name in the first element and the
 619          * parameter in the second element.
 620          * Both elements are false if there was no match.
 621          */
 622         public function matchVariableStartToEnd( $text ) {
 623                 global $wgContLang;
 624                 $regexes = $this->getVariableStartToEndRegex();
 625                 foreach ( $regexes as $regex ) {
 626                         if ( $regex !== '' ) {
 627                                 $m = false;
 628                                 if ( preg_match( $regex, $text, $m ) ) {
 629                                         return $this->parseMatch( $m );
 630                                 }
 631                         }
 632                 }
 633                 return array( false, false );
 634         }
 635
 636         /**
 637          * Match some text, without parameter capture
 638          * Returns the magic word name, or false if there was no capture
 639          */
 640         public function matchStartToEnd( $text ) {
 641                 $hash = $this->getHash();
 642                 if ( isset( $hash[1][$text] ) ) {
 643                         return $hash[1][$text];
 644                 }
 645                 global $wgContLang;
 646                 $lc = $wgContLang->lc( $text );
 647                 if ( isset( $hash[0][$lc] ) ) {
 648                         return $hash[0][$lc];
 649                 }
 650                 return false;
 651         }
 652
 653         /**
 654          * Returns an associative array, ID => param value, for all items that match
 655          * Removes the matched items from the input string (passed by reference)
 656          */
 657         public function matchAndRemove( &$text ) {
 658                 $found = array();
 659                 $regexes = $this->getRegex();
 660                 foreach ( $regexes as $regex ) {
 661                         if ( $regex === '' ) {
 662                                 continue;
 663                         }
 664                         preg_match_all( $regex, $text, $matches, PREG_SET_ORDER );
 665                         foreach ( $matches as $m ) {
 666                                 list( $name, $param ) = $this->parseMatch( $m );
 667                                 $found[$name] = $param;
 668                         }
 669                         $text = preg_replace( $regex, '', $text );
 670                 }
 671                 return $found;
 672         }
 673 }