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                 'numberofactiveusers',
  94                 'newsectionlink',
  95                 'nonewsectionlink',
  96                 'numberofpages',
  97                 'currentversion',
  98                 'basepagename',
  99                 'basepagenamee',
 100                 'urlencode',
 101                 'currenttimestamp',
 102                 'localtimestamp',
 103                 'directionmark',
 104                 'language',
 105                 'contentlanguage',
 106                 'pagesinnamespace',
 107                 'numberofadmins',
 108                 'numberofviews',
 109                 'defaultsort',
 110                 'pagesincategory',
 111                 'index',
 112                 'noindex',
 113                 'numberingroup',
 114         );
 115
 116         /* Array of caching hints for ParserCache */
 117         static public $mCacheTTLs = array (
 118                 'currentmonth' => 86400,
 119                 'currentmonthname' => 86400,
 120                 'currentmonthnamegen' => 86400,
 121                 'currentmonthabbrev' => 86400,
 122                 'currentday' => 3600,
 123                 'currentday2' => 3600,
 124                 'currentdayname' => 3600,
 125                 'currentyear' => 86400,
 126                 'currenttime' => 3600,
 127                 'currenthour' => 3600,
 128                 'localmonth' => 86400,
 129                 'localmonthname' => 86400,
 130                 'localmonthnamegen' => 86400,
 131                 'localmonthabbrev' => 86400,
 132                 'localday' => 3600,
 133                 'localday2' => 3600,
 134                 'localdayname' => 3600,
 135                 'localyear' => 86400,
 136                 'localtime' => 3600,
 137                 'localhour' => 3600,
 138                 'numberofarticles' => 3600,
 139                 'numberoffiles' => 3600,
 140                 'numberofedits' => 3600,
 141                 'currentweek' => 3600,
 142                 'currentdow' => 3600,
 143                 'localweek' => 3600,
 144                 'localdow' => 3600,
 145                 'numberofusers' => 3600,
 146                 'numberofactiveusers' => 3600,
 147                 'numberofpages' => 3600,
 148                 'currentversion' => 86400,
 149                 'currenttimestamp' => 3600,
 150                 'localtimestamp' => 3600,
 151                 'pagesinnamespace' => 3600,
 152                 'numberofadmins' => 3600,
 153                 'numberofviews' => 3600,
 154                 'numberingroup' => 3600,
 155                 );
 156
 157         static public $mDoubleUnderscoreIDs = array(
 158                 'notoc',
 159                 'nogallery',
 160                 'forcetoc',
 161                 'toc',
 162                 'noeditsection',
 163                 'newsectionlink',
 164                 'nonewsectionlink',
 165                 'hiddencat',
 166                 'index',
 167                 'noindex',
 168                 'staticredirect',
 169         );
 170
 171
 172         static public $mObjects = array();
 173         static public $mDoubleUnderscoreArray = null;
 174
 175         /**#@-*/
 176
 177         function __construct($id = 0, $syn = '', $cs = false) {
 178                 $this->mId = $id;
 179                 $this->mSynonyms = (array)$syn;
 180                 $this->mCaseSensitive = $cs;
 181                 $this->mRegex = '';
 182                 $this->mRegexStart = '';
 183                 $this->mVariableRegex = '';
 184                 $this->mVariableStartToEndRegex = '';
 185                 $this->mModified = false;
 186         }
 187
 188         /**
 189          * Factory: creates an object representing an ID
 190          * @static
 191          */
 192         static function &get( $id ) {
 193                 wfProfileIn( __METHOD__ );
 194                 if (!array_key_exists( $id, self::$mObjects ) ) {
 195                         $mw = new MagicWord();
 196                         $mw->load( $id );
 197                         self::$mObjects[$id] = $mw;
 198                 }
 199                 wfProfileOut( __METHOD__ );
 200                 return self::$mObjects[$id];
 201         }
 202
 203         /**
 204          * Get an array of parser variable IDs
 205          */
 206         static function getVariableIDs() {
 207                 if ( !self::$mVariableIDsInitialised ) {
 208                         # Deprecated constant definition hook, available for extensions that need it
 209                         $magicWords = array();
 210                         wfRunHooks( 'MagicWordMagicWords', array( &$magicWords ) );
 211                         foreach ( $magicWords as $word ) {
 212                                 define( $word, $word );
 213                         }
 214
 215                         # Get variable IDs
 216                         wfRunHooks( 'MagicWordwgVariableIDs', array( &self::$mVariableIDs ) );
 217                         self::$mVariableIDsInitialised = true;
 218                 }
 219                 return self::$mVariableIDs;
 220         }
 221
 222         /* Allow external reads of TTL array */
 223         static function getCacheTTL($id) {
 224                 if (array_key_exists($id,self::$mCacheTTLs)) {
 225                         return self::$mCacheTTLs[$id];
 226                 } else {
 227                         return -1;
 228                 }
 229         }
 230
 231         /** Get a MagicWordArray of double-underscore entities */
 232         static function getDoubleUnderscoreArray() {
 233                 if ( is_null( self::$mDoubleUnderscoreArray ) ) {
 234                         self::$mDoubleUnderscoreArray = new MagicWordArray( self::$mDoubleUnderscoreIDs );
 235                 }
 236                 return self::$mDoubleUnderscoreArray;
 237         }
 238
 239         # Initialises this object with an ID
 240         function load( $id ) {
 241                 global $wgContLang;
 242                 $this->mId = $id;
 243                 $wgContLang->getMagic( $this );
 244                 if ( !$this->mSynonyms ) {
 245                         $this->mSynonyms = array( 'dkjsagfjsgashfajsh' );
 246                         #throw new MWException( "Error: invalid magic word '$id'" );
 247                         wfDebugLog( 'exception', "Error: invalid magic word '$id'\n" );
 248                 }
 249         }
 250
 251         /**
 252          * Preliminary initialisation
 253          * @private
 254          */
 255         function initRegex() {
 256                 #$variableClass = Title::legalChars();
 257                 # This was used for matching "$1" variables, but different uses of the feature will have
 258                 # different restrictions, which should be checked *after* the MagicWord has been matched,
 259                 # not here. - IMSoP
 260
 261                 $escSyn = array();
 262                 foreach ( $this->mSynonyms as $synonym )
 263                         // In case a magic word contains /, like that's going to happen;)
 264                         $escSyn[] = preg_quote( $synonym, '/' );
 265                 $this->mBaseRegex = implode( '|', $escSyn );
 266
 267                 $case = $this->mCaseSensitive ? '' : 'iu';
 268                 $this->mRegex = "/{$this->mBaseRegex}/{$case}";
 269                 $this->mRegexStart = "/^(?:{$this->mBaseRegex})/{$case}";
 270                 $this->mVariableRegex = str_replace( "\\$1", "(.*?)", $this->mRegex );
 271                 $this->mVariableStartToEndRegex = str_replace( "\\$1", "(.*?)",
 272                         "/^(?:{$this->mBaseRegex})$/{$case}" );
 273         }
 274
 275         /**
 276          * Gets a regex representing matching the word
 277          */
 278         function getRegex() {
 279                 if ($this->mRegex == '' ) {
 280                         $this->initRegex();
 281                 }
 282                 return $this->mRegex;
 283         }
 284
 285         /**
 286          * Gets the regexp case modifier to use, i.e. i or nothing, to be used if
 287          * one is using MagicWord::getBaseRegex(), otherwise it'll be included in
 288          * the complete expression
 289          */
 290         function getRegexCase() {
 291                 if ( $this->mRegex === '' )
 292                         $this->initRegex();
 293
 294                 return $this->mCaseSensitive ? '' : 'iu';
 295         }
 296
 297         /**
 298          * Gets a regex matching the word, if it is at the string start
 299          */
 300         function getRegexStart() {
 301                 if ($this->mRegex == '' ) {
 302                         $this->initRegex();
 303                 }
 304                 return $this->mRegexStart;
 305         }
 306
 307         /**
 308          * regex without the slashes and what not
 309          */
 310         function getBaseRegex() {
 311                 if ($this->mRegex == '') {
 312                         $this->initRegex();
 313                 }
 314                 return $this->mBaseRegex;
 315         }
 316
 317         /**
 318          * Returns true if the text contains the word
 319          * @return bool
 320          */
 321         function match( $text ) {
 322                 return preg_match( $this->getRegex(), $text );
 323         }
 324
 325         /**
 326          * Returns true if the text starts with the word
 327          * @return bool
 328          */
 329         function matchStart( $text ) {
 330                 return preg_match( $this->getRegexStart(), $text );
 331         }
 332
 333         /**
 334          * Returns NULL if there's no match, the value of $1 otherwise
 335          * The return code is the matched string, if there's no variable
 336          * part in the regex and the matched variable part ($1) if there
 337          * is one.
 338          */
 339         function matchVariableStartToEnd( $text ) {
 340                 $matches = array();
 341                 $matchcount = preg_match( $this->getVariableStartToEndRegex(), $text, $matches );
 342                 if ( $matchcount == 0 ) {
 343                         return NULL;
 344                 } else {
 345                         # multiple matched parts (variable match); some will be empty because of
 346                         # synonyms. The variable will be the second non-empty one so remove any
 347                         # blank elements and re-sort the indices.
 348                         # See also bug 6526
 349
 350                         $matches = array_values(array_filter($matches));
 351
 352                         if ( count($matches) == 1 ) { return $matches[0]; }
 353                         else { return $matches[1]; }
 354                 }
 355         }
 356
 357
 358         /**
 359          * Returns true if the text matches the word, and alters the
 360          * input string, removing all instances of the word
 361          */
 362         function matchAndRemove( &$text ) {
 363                 $this->mFound = false;
 364                 $text = preg_replace_callback( $this->getRegex(), array( &$this, 'pregRemoveAndRecord' ), $text );
 365                 return $this->mFound;
 366         }
 367
 368         function matchStartAndRemove( &$text ) {
 369                 $this->mFound = false;
 370                 $text = preg_replace_callback( $this->getRegexStart(), array( &$this, 'pregRemoveAndRecord' ), $text );
 371                 return $this->mFound;
 372         }
 373
 374         /**
 375          * Used in matchAndRemove()
 376          * @private
 377          **/
 378         function pregRemoveAndRecord( ) {
 379                 $this->mFound = true;
 380                 return '';
 381         }
 382
 383         /**
 384          * Replaces the word with something else
 385          */
 386         function replace( $replacement, $subject, $limit=-1 ) {
 387                 $res = preg_replace( $this->getRegex(), StringUtils::escapeRegexReplacement( $replacement ), $subject, $limit );
 388                 $this->mModified = !($res === $subject);
 389                 return $res;
 390         }
 391
 392         /**
 393          * Variable handling: {{SUBST:xxx}} style words
 394          * Calls back a function to determine what to replace xxx with
 395          * Input word must contain $1
 396          */
 397         function substituteCallback( $text, $callback ) {
 398                 $res = preg_replace_callback( $this->getVariableRegex(), $callback, $text );
 399                 $this->mModified = !($res === $text);
 400                 return $res;
 401         }
 402
 403         /**
 404          * Matches the word, where $1 is a wildcard
 405          */
 406         function getVariableRegex()     {
 407                 if ( $this->mVariableRegex == '' ) {
 408                         $this->initRegex();
 409                 }
 410                 return $this->mVariableRegex;
 411         }
 412
 413         /**
 414          * Matches the entire string, where $1 is a wildcard
 415          */
 416         function getVariableStartToEndRegex() {
 417                 if ( $this->mVariableStartToEndRegex == '' ) {
 418                         $this->initRegex();
 419                 }
 420                 return $this->mVariableStartToEndRegex;
 421         }
 422
 423         /**
 424          * Accesses the synonym list directly
 425          */
 426         function getSynonym( $i ) {
 427                 return $this->mSynonyms[$i];
 428         }
 429
 430         function getSynonyms() {
 431                 return $this->mSynonyms;
 432         }
 433
 434         /**
 435          * Returns true if the last call to replace() or substituteCallback()
 436          * returned a modified text, otherwise false.
 437          */
 438         function getWasModified(){
 439                 return $this->mModified;
 440         }
 441
 442         /**
 443          * $magicarr is an associative array of (magic word ID => replacement)
 444          * This method uses the php feature to do several replacements at the same time,
 445          * thereby gaining some efficiency. The result is placed in the out variable
 446          * $result. The return value is true if something was replaced.
 447          * @static
 448          **/
 449         function replaceMultiple( $magicarr, $subject, &$result ){
 450                 $search = array();
 451                 $replace = array();
 452                 foreach( $magicarr as $id => $replacement ){
 453                         $mw = MagicWord::get( $id );
 454                         $search[] = $mw->getRegex();
 455                         $replace[] = $replacement;
 456                 }
 457
 458                 $result = preg_replace( $search, $replace, $subject );
 459                 return !($result === $subject);
 460         }
 461
 462         /**
 463          * Adds all the synonyms of this MagicWord to an array, to allow quick
 464          * lookup in a list of magic words
 465          */
 466         function addToArray( &$array, $value ) {
 467                 global $wgContLang;
 468                 foreach ( $this->mSynonyms as $syn ) {
 469                         $array[$wgContLang->lc($syn)] = $value;
 470                 }
 471         }
 472
 473         function isCaseSensitive() {
 474                 return $this->mCaseSensitive;
 475         }
 476
 477         function getId() {
 478                 return $this->mId;
 479         }
 480 }
 481
 482 /**
 483  * Class for handling an array of magic words
 484  * @ingroup Parser
 485  */
 486 class MagicWordArray {
 487         var $names = array();
 488         var $hash;
 489         var $baseRegex, $regex;
 490         var $matches;
 491
 492         function __construct( $names = array() ) {
 493                 $this->names = $names;
 494         }
 495
 496         /**
 497          * Add a magic word by name
 498          */
 499         public function add( $name ) {
 500                 global $wgContLang;
 501                 $this->names[] = $name;
 502                 $this->hash = $this->baseRegex = $this->regex = null;
 503         }
 504
 505         /**
 506          * Add a number of magic words by name
 507          */
 508         public function addArray( $names ) {
 509                 $this->names = array_merge( $this->names, array_values( $names ) );
 510                 $this->hash = $this->baseRegex = $this->regex = null;
 511         }
 512
 513         /**
 514          * Get a 2-d hashtable for this array
 515          */
 516         function getHash() {
 517                 if ( is_null( $this->hash ) ) {
 518                         global $wgContLang;
 519                         $this->hash = array( 0 => array(), 1 => array() );
 520                         foreach ( $this->names as $name ) {
 521                                 $magic = MagicWord::get( $name );
 522                                 $case = intval( $magic->isCaseSensitive() );
 523                                 foreach ( $magic->getSynonyms() as $syn ) {
 524                                         if ( !$case ) {
 525                                                 $syn = $wgContLang->lc( $syn );
 526                                         }
 527                                         $this->hash[$case][$syn] = $name;
 528                                 }
 529                         }
 530                 }
 531                 return $this->hash;
 532         }
 533
 534         /**
 535          * Get the base regex
 536          */
 537         function getBaseRegex() {
 538                 if ( is_null( $this->baseRegex ) ) {
 539                         $this->baseRegex = array( 0 => '', 1 => '' );
 540                         foreach ( $this->names as $name ) {
 541                                 $magic = MagicWord::get( $name );
 542                                 $case = intval( $magic->isCaseSensitive() );
 543                                 foreach ( $magic->getSynonyms() as $i => $syn ) {
 544                                         $group = "(?P<{$i}_{$name}>" . preg_quote( $syn, '/' ) . ')';
 545                                         if ( $this->baseRegex[$case] === '' ) {
 546                                                 $this->baseRegex[$case] = $group;
 547                                         } else {
 548                                                 $this->baseRegex[$case] .= '|' . $group;
 549                                         }
 550                                 }
 551                         }
 552                 }
 553                 return $this->baseRegex;
 554         }
 555
 556         /**
 557          * Get an unanchored regex
 558          */
 559         function getRegex() {
 560                 if ( is_null( $this->regex ) ) {
 561                         $base = $this->getBaseRegex();
 562                         $this->regex = array( '', '' );
 563                         if ( $this->baseRegex[0] !== '' ) {
 564                                 $this->regex[0] = "/{$base[0]}/iuS";
 565                         }
 566                         if ( $this->baseRegex[1] !== '' ) {
 567                                 $this->regex[1] = "/{$base[1]}/S";
 568                         }
 569                 }
 570                 return $this->regex;
 571         }
 572
 573         /**
 574          * Get a regex for matching variables
 575          */
 576         function getVariableRegex() {
 577                 return str_replace( "\\$1", "(.*?)", $this->getRegex() );
 578         }
 579
 580         /**
 581          * Get an anchored regex for matching variables
 582          */
 583         function getVariableStartToEndRegex() {
 584                 $base = $this->getBaseRegex();
 585                 $newRegex = array( '', '' );
 586                 if ( $base[0] !== '' ) {
 587                         $newRegex[0] = str_replace( "\\$1", "(.*?)", "/^(?:{$base[0]})$/iuS" );
 588                 }
 589                 if ( $base[1] !== '' ) {
 590                         $newRegex[1] = str_replace( "\\$1", "(.*?)", "/^(?:{$base[1]})$/S" );
 591                 }
 592                 return $newRegex;
 593         }
 594
 595         /**
 596          * Parse a match array from preg_match
 597          * Returns array(magic word ID, parameter value)
 598          * If there is no parameter value, that element will be false.
 599          */
 600         function parseMatch( $m ) {
 601                 reset( $m );
 602                 while ( list( $key, $value ) = each( $m ) ) {
 603                         if ( $key === 0 || $value === '' ) {
 604                                 continue;
 605                         }
 606                         $parts = explode( '_', $key, 2 );
 607                         if ( count( $parts ) != 2 ) {
 608                                 // This shouldn't happen
 609                                 // continue;
 610                                 throw new MWException( __METHOD__ . ': bad parameter name' );
 611                         }
 612                         list( /* $synIndex */, $magicName ) = $parts;
 613                         $paramValue = next( $m );
 614                         return array( $magicName, $paramValue );
 615                 }
 616                 // This shouldn't happen either
 617                 throw new MWException( __METHOD__.': parameter not found' );
 618                 return array( false, false );
 619         }
 620
 621         /**
 622          * Match some text, with parameter capture
 623          * Returns an array with the magic word name in the first element and the
 624          * parameter in the second element.
 625          * Both elements are false if there was no match.
 626          */
 627         public function matchVariableStartToEnd( $text ) {
 628                 global $wgContLang;
 629                 $regexes = $this->getVariableStartToEndRegex();
 630                 foreach ( $regexes as $regex ) {
 631                         if ( $regex !== '' ) {
 632                                 $m = false;
 633                                 if ( preg_match( $regex, $text, $m ) ) {
 634                                         return $this->parseMatch( $m );
 635                                 }
 636                         }
 637                 }
 638                 return array( false, false );
 639         }
 640
 641         /**
 642          * Match some text, without parameter capture
 643          * Returns the magic word name, or false if there was no capture
 644          */
 645         public function matchStartToEnd( $text ) {
 646                 $hash = $this->getHash();
 647                 if ( isset( $hash[1][$text] ) ) {
 648                         return $hash[1][$text];
 649                 }
 650                 global $wgContLang;
 651                 $lc = $wgContLang->lc( $text );
 652                 if ( isset( $hash[0][$lc] ) ) {
 653                         return $hash[0][$lc];
 654                 }
 655                 return false;
 656         }
 657
 658         /**
 659          * Returns an associative array, ID => param value, for all items that match
 660          * Removes the matched items from the input string (passed by reference)
 661          */
 662         public function matchAndRemove( &$text ) {
 663                 $found = array();
 664                 $regexes = $this->getRegex();
 665                 foreach ( $regexes as $regex ) {
 666                         if ( $regex === '' ) {
 667                                 continue;
 668                         }
 669                         preg_match_all( $regex, $text, $matches, PREG_SET_ORDER );
 670                         foreach ( $matches as $m ) {
 671                                 list( $name, $param ) = $this->parseMatch( $m );
 672                                 $found[$name] = $param;
 673                         }
 674                         $text = preg_replace( $regex, '', $text );
 675                 }
 676                 return $found;
 677         }
 678 }