includes/MagicWord.php

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