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