includes/MagicWord.php

   1 <?php
   2 /**
   3  * File for magic words
   4  *
   5  * See docs/magicword.txt
   6  *
   7  * @file
   8  * @ingroup Parser
   9  */
  10
  11 /**
  12  * This class encapsulates "magic words" such as #redirect, __NOTOC__, etc.
  13  * Usage:
  14  *     if (MagicWord::get( 'redirect' )->match( $text ) )
  15  *
  16  * Possible future improvements:
  17  *   * Simultaneous searching for a number of magic words
  18  *   * MagicWord::$mObjects in shared memory
  19  *
  20  * Please avoid reading the data out of one of these objects and then writing
  21  * special case code. If possible, add another match()-like function here.
  22  *
  23  * To add magic words in an extension, use the LanguageGetMagic hook. For
  24  * magic words which are also Parser variables, add a MagicWordwgVariableIDs
  25  * hook. Use string keys.
  26  *
  27  * @ingroup Parser
  28  */
  29 class MagicWord {
  30         /**#@+
  31          * @private
  32          */
  33         var $mId, $mSynonyms, $mCaseSensitive, $mRegex;
  34         var $mRegexStart, $mBaseRegex, $mVariableRegex;
  35         var $mModified, $mFound;
  36
  37         static public $mVariableIDsInitialised = false;
  38         static public $mVariableIDs = array(
  39                 'currentmonth',
  40                 'currentmonth1',
  41                 'currentmonthname',
  42                 'currentmonthnamegen',
  43                 'currentmonthabbrev',
  44                 'currentday',
  45                 'currentday2',
  46                 'currentdayname',
  47                 'currentyear',
  48                 'currenttime',
  49                 'currenthour',
  50                 'localmonth',
  51                 'localmonth1',
  52                 'localmonthname',
  53                 'localmonthnamegen',
  54                 'localmonthabbrev',
  55                 'localday',
  56                 'localday2',
  57                 'localdayname',
  58                 'localyear',
  59                 'localtime',
  60                 'localhour',
  61                 'numberofarticles',
  62                 'numberoffiles',
  63                 'numberofedits',
  64                 'articlepath',
  65                 'sitename',
  66                 'server',
  67                 'servername',
  68                 'scriptpath',
  69                 'stylepath',
  70                 'pagename',
  71                 'pagenamee',
  72                 'fullpagename',
  73                 'fullpagenamee',
  74                 'namespace',
  75                 'namespacee',
  76                 'currentweek',
  77                 'currentdow',
  78                 'localweek',
  79                 'localdow',
  80                 'revisionid',
  81                 'revisionday',
  82                 'revisionday2',
  83                 'revisionmonth',
  84                 'revisionmonth1',
  85                 'revisionyear',
  86                 'revisiontimestamp',
  87                 'revisionuser',
  88                 'subpagename',
  89                 'subpagenamee',
  90                 'talkspace',
  91                 'talkspacee',
  92                 'subjectspace',
  93                 'subjectspacee',
  94                 'talkpagename',
  95                 'talkpagenamee',
  96                 'subjectpagename',
  97                 'subjectpagenamee',
  98                 'numberofusers',
  99                 'numberofactiveusers',
 100                 'numberofpages',
 101                 'currentversion',
 102                 'basepagename',
 103                 'basepagenamee',
 104                 'currenttimestamp',
 105                 'localtimestamp',
 106                 'directionmark',
 107                 'contentlanguage',
 108                 'numberofadmins',
 109                 'numberofviews',
 110         );
 111
 112         /* Array of caching hints for ParserCache */
 113         static public $mCacheTTLs = array (
 114                 'currentmonth' => 86400,
 115                 'currentmonth1' => 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                 'localmonth1' => 86400,
 127                 'localmonthname' => 86400,
 128                 'localmonthnamegen' => 86400,
 129                 'localmonthabbrev' => 86400,
 130                 'localday' => 3600,
 131                 'localday2' => 3600,
 132                 'localdayname' => 3600,
 133                 'localyear' => 86400,
 134                 'localtime' => 3600,
 135                 'localhour' => 3600,
 136                 'numberofarticles' => 3600,
 137                 'numberoffiles' => 3600,
 138                 'numberofedits' => 3600,
 139                 'currentweek' => 3600,
 140                 'currentdow' => 3600,
 141                 'localweek' => 3600,
 142                 'localdow' => 3600,
 143                 'numberofusers' => 3600,
 144                 'numberofactiveusers' => 3600,
 145                 'numberofpages' => 3600,
 146                 'currentversion' => 86400,
 147                 'currenttimestamp' => 3600,
 148                 'localtimestamp' => 3600,
 149                 'pagesinnamespace' => 3600,
 150                 'numberofadmins' => 3600,
 151                 'numberofviews' => 3600,
 152                 'numberingroup' => 3600,
 153                 );
 154
 155         static public $mDoubleUnderscoreIDs = array(
 156                 'notoc',
 157                 'nogallery',
 158                 'forcetoc',
 159                 'toc',
 160                 'noeditsection',
 161                 'newsectionlink',
 162                 'nonewsectionlink',
 163                 'hiddencat',
 164                 'index',
 165                 'noindex',
 166                 'staticredirect',
 167                 'notitleconvert',
 168                 'nocontentconvert',
 169         );
 170
 171         static public $mSubstIDs = array(
 172                 'subst',
 173                 'safesubst',
 174         );
 175
 176         static public $mObjects = array();
 177         static public $mDoubleUnderscoreArray = null;
 178
 179         /**#@-*/
 180
 181         function __construct($id = 0, $syn = '', $cs = false) {
 182                 $this->mId = $id;
 183                 $this->mSynonyms = (array)$syn;
 184                 $this->mCaseSensitive = $cs;
 185                 $this->mRegex = '';
 186                 $this->mRegexStart = '';
 187                 $this->mVariableRegex = '';
 188                 $this->mVariableStartToEndRegex = '';
 189                 $this->mModified = false;
 190         }
 191
 192         /**
 193          * Factory: creates an object representing an ID
 194          * @static
 195          * @return MagicWord
 196          */
 197         static function &get( $id ) {
 198                 wfProfileIn( __METHOD__ );
 199                 if ( !isset( self::$mObjects[$id] ) ) {
 200                         $mw = new MagicWord();
 201                         $mw->load( $id );
 202                         self::$mObjects[$id] = $mw;
 203                 }
 204                 wfProfileOut( __METHOD__ );
 205                 return self::$mObjects[$id];
 206         }
 207
 208         /**
 209          * Get an array of parser variable IDs
 210          */
 211         static function getVariableIDs() {
 212                 if ( !self::$mVariableIDsInitialised ) {
 213                         # Deprecated constant definition hook, available for extensions that need it
 214                         $magicWords = array();
 215                         wfRunHooks( 'MagicWordMagicWords', array( &$magicWords ) );
 216                         foreach ( $magicWords as $word ) {
 217                                 define( $word, $word );
 218                         }
 219
 220                         # Get variable IDs
 221                         wfRunHooks( 'MagicWordwgVariableIDs', array( &self::$mVariableIDs ) );
 222                         self::$mVariableIDsInitialised = true;
 223                 }
 224                 return self::$mVariableIDs;
 225         }
 226
 227         /**
 228          * Get an array of parser substitution modifier IDs
 229          */
 230         static function getSubstIDs() {
 231                 return self::$mSubstIDs;
 232         }
 233
 234         /* Allow external reads of TTL array */
 235         static function getCacheTTL($id) {
 236                 if (array_key_exists($id,self::$mCacheTTLs)) {
 237                         return self::$mCacheTTLs[$id];
 238                 } else {
 239                         return -1;
 240                 }
 241         }
 242
 243         /** Get a MagicWordArray of double-underscore entities */
 244         static function getDoubleUnderscoreArray() {
 245                 if ( is_null( self::$mDoubleUnderscoreArray ) ) {
 246                         self::$mDoubleUnderscoreArray = new MagicWordArray( self::$mDoubleUnderscoreIDs );
 247                 }
 248                 return self::$mDoubleUnderscoreArray;
 249         }
 250
 251         /**
 252          * Clear the self::$mObjects variable
 253          * For use in parser tests
 254          */
 255         public static function clearCache() {
 256                 self::$mObjects = array();
 257         }
 258
 259         # Initialises this object with an ID
 260         function load( $id ) {
 261                 global $wgContLang;
 262                 $this->mId = $id;
 263                 $wgContLang->getMagic( $this );
 264                 if ( !$this->mSynonyms ) {
 265                         $this->mSynonyms = array( 'dkjsagfjsgashfajsh' );
 266                         #throw new MWException( "Error: invalid magic word '$id'" );
 267                         wfDebugLog( 'exception', "Error: invalid magic word '$id'\n" );
 268                 }
 269         }
 270
 271         /**
 272          * Preliminary initialisation
 273          * @private
 274          */
 275         function initRegex() {
 276                 #$variableClass = Title::legalChars();
 277                 # This was used for matching "$1" variables, but different uses of the feature will have
 278                 # different restrictions, which should be checked *after* the MagicWord has been matched,
 279                 # not here. - IMSoP
 280
 281                 $escSyn = array();
 282                 foreach ( $this->mSynonyms as $synonym )
 283                         // In case a magic word contains /, like that's going to happen;)
 284                         $escSyn[] = preg_quote( $synonym, '/' );
 285                 $this->mBaseRegex = implode( '|', $escSyn );
 286
 287                 $case = $this->mCaseSensitive ? '' : 'iu';
 288                 $this->mRegex = "/{$this->mBaseRegex}/{$case}";
 289                 $this->mRegexStart = "/^(?:{$this->mBaseRegex})/{$case}";
 290                 $this->mVariableRegex = str_replace( "\\$1", "(.*?)", $this->mRegex );
 291                 $this->mVariableStartToEndRegex = str_replace( "\\$1", "(.*?)",
 292                         "/^(?:{$this->mBaseRegex})$/{$case}" );
 293         }
 294
 295         /**
 296          * Gets a regex representing matching the word
 297          */
 298         function getRegex() {
 299                 if ($this->mRegex == '' ) {
 300                         $this->initRegex();
 301                 }
 302                 return $this->mRegex;
 303         }
 304
 305         /**
 306          * Gets the regexp case modifier to use, i.e. i or nothing, to be used if
 307          * one is using MagicWord::getBaseRegex(), otherwise it'll be included in
 308          * the complete expression
 309          */
 310         function getRegexCase() {
 311                 if ( $this->mRegex === '' )
 312                         $this->initRegex();
 313
 314                 return $this->mCaseSensitive ? '' : 'iu';
 315         }
 316
 317         /**
 318          * Gets a regex matching the word, if it is at the string start
 319          */
 320         function getRegexStart() {
 321                 if ($this->mRegex == '' ) {
 322                         $this->initRegex();
 323                 }
 324                 return $this->mRegexStart;
 325         }
 326
 327         /**
 328          * regex without the slashes and what not
 329          */
 330         function getBaseRegex() {
 331                 if ($this->mRegex == '') {
 332                         $this->initRegex();
 333                 }
 334                 return $this->mBaseRegex;
 335         }
 336
 337         /**
 338          * Returns true if the text contains the word
 339          * @return bool
 340          */
 341         function match( $text ) {
 342                 return (bool)preg_match( $this->getRegex(), $text );
 343         }
 344
 345         /**
 346          * Returns true if the text starts with the word
 347          * @return bool
 348          */
 349         function matchStart( $text ) {
 350                 return (bool)preg_match( $this->getRegexStart(), $text );
 351         }
 352
 353         /**
 354          * Returns NULL if there's no match, the value of $1 otherwise
 355          * The return code is the matched string, if there's no variable
 356          * part in the regex and the matched variable part ($1) if there
 357          * is one.
 358          */
 359         function matchVariableStartToEnd( $text ) {
 360                 $matches = array();
 361                 $matchcount = preg_match( $this->getVariableStartToEndRegex(), $text, $matches );
 362                 if ( $matchcount == 0 ) {
 363                         return null;
 364                 } else {
 365                         # multiple matched parts (variable match); some will be empty because of
 366                         # synonyms. The variable will be the second non-empty one so remove any
 367                         # blank elements and re-sort the indices.
 368                         # See also bug 6526
 369
 370                         $matches = array_values(array_filter($matches));
 371
 372                         if ( count($matches) == 1 ) { return $matches[0]; }
 373                         else { return $matches[1]; }
 374                 }
 375         }
 376
 377
 378         /**
 379          * Returns true if the text matches the word, and alters the
 380          * input string, removing all instances of the word
 381          */
 382         function matchAndRemove( &$text ) {
 383                 $this->mFound = false;
 384                 $text = preg_replace_callback( $this->getRegex(), array( &$this, 'pregRemoveAndRecord' ), $text );
 385                 return $this->mFound;
 386         }
 387
 388         function matchStartAndRemove( &$text ) {
 389                 $this->mFound = false;
 390                 $text = preg_replace_callback( $this->getRegexStart(), array( &$this, 'pregRemoveAndRecord' ), $text );
 391                 return $this->mFound;
 392         }
 393
 394         /**
 395          * Used in matchAndRemove()
 396          * @private
 397          **/
 398         function pregRemoveAndRecord( ) {
 399                 $this->mFound = true;
 400                 return '';
 401         }
 402
 403         /**
 404          * Replaces the word with something else
 405          */
 406         function replace( $replacement, $subject, $limit=-1 ) {
 407                 $res = preg_replace( $this->getRegex(), StringUtils::escapeRegexReplacement( $replacement ), $subject, $limit );
 408                 $this->mModified = !($res === $subject);
 409                 return $res;
 410         }
 411
 412         /**
 413          * Variable handling: {{SUBST:xxx}} style words
 414          * Calls back a function to determine what to replace xxx with
 415          * Input word must contain $1
 416          */
 417         function substituteCallback( $text, $callback ) {
 418                 $res = preg_replace_callback( $this->getVariableRegex(), $callback, $text );
 419                 $this->mModified = !($res === $text);
 420                 return $res;
 421         }
 422
 423         /**
 424          * Matches the word, where $1 is a wildcard
 425          */
 426         function getVariableRegex()     {
 427                 if ( $this->mVariableRegex == '' ) {
 428                         $this->initRegex();
 429                 }
 430                 return $this->mVariableRegex;
 431         }
 432
 433         /**
 434          * Matches the entire string, where $1 is a wildcard
 435          */
 436         function getVariableStartToEndRegex() {
 437                 if ( $this->mVariableStartToEndRegex == '' ) {
 438                         $this->initRegex();
 439                 }
 440                 return $this->mVariableStartToEndRegex;
 441         }
 442
 443         /**
 444          * Accesses the synonym list directly
 445          */
 446         function getSynonym( $i ) {
 447                 return $this->mSynonyms[$i];
 448         }
 449
 450         function getSynonyms() {
 451                 return $this->mSynonyms;
 452         }
 453
 454         /**
 455          * Returns true if the last call to replace() or substituteCallback()
 456          * returned a modified text, otherwise false.
 457          */
 458         function getWasModified(){
 459                 return $this->mModified;
 460         }
 461
 462         /**
 463          * $magicarr is an associative array of (magic word ID => replacement)
 464          * This method uses the php feature to do several replacements at the same time,
 465          * thereby gaining some efficiency. The result is placed in the out variable
 466          * $result. The return value is true if something was replaced.
 467          * @static
 468          **/
 469         function replaceMultiple( $magicarr, $subject, &$result ){
 470                 $search = array();
 471                 $replace = array();
 472                 foreach( $magicarr as $id => $replacement ){
 473                         $mw = MagicWord::get( $id );
 474                         $search[] = $mw->getRegex();
 475                         $replace[] = $replacement;
 476                 }
 477
 478                 $result = preg_replace( $search, $replace, $subject );
 479                 return !($result === $subject);
 480         }
 481
 482         /**
 483          * Adds all the synonyms of this MagicWord to an array, to allow quick
 484          * lookup in a list of magic words
 485          */
 486         function addToArray( &$array, $value ) {
 487                 global $wgContLang;
 488                 foreach ( $this->mSynonyms as $syn ) {
 489                         $array[$wgContLang->lc($syn)] = $value;
 490                 }
 491         }
 492
 493         function isCaseSensitive() {
 494                 return $this->mCaseSensitive;
 495         }
 496
 497         function getId() {
 498                 return $this->mId;
 499         }
 500 }
 501
 502 /**
 503  * Class for handling an array of magic words
 504  * @ingroup Parser
 505  */
 506 class MagicWordArray {
 507         var $names = array();
 508         var $hash;
 509         var $baseRegex, $regex;
 510         var $matches;
 511
 512         function __construct( $names = array() ) {
 513                 $this->names = $names;
 514         }
 515
 516         /**
 517          * Add a magic word by name
 518          */
 519         public function add( $name ) {
 520                 $this->names[] = $name;
 521                 $this->hash = $this->baseRegex = $this->regex = null;
 522         }
 523
 524         /**
 525          * Add a number of magic words by name
 526          */
 527         public function addArray( $names ) {
 528                 $this->names = array_merge( $this->names, array_values( $names ) );
 529                 $this->hash = $this->baseRegex = $this->regex = null;
 530         }
 531
 532         /**
 533          * Get a 2-d hashtable for this array
 534          */
 535         function getHash() {
 536                 if ( is_null( $this->hash ) ) {
 537                         global $wgContLang;
 538                         $this->hash = array( 0 => array(), 1 => array() );
 539                         foreach ( $this->names as $name ) {
 540                                 $magic = MagicWord::get( $name );
 541                                 $case = intval( $magic->isCaseSensitive() );
 542                                 foreach ( $magic->getSynonyms() as $syn ) {
 543                                         if ( !$case ) {
 544                                                 $syn = $wgContLang->lc( $syn );
 545                                         }
 546                                         $this->hash[$case][$syn] = $name;
 547                                 }
 548                         }
 549                 }
 550                 return $this->hash;
 551         }
 552
 553         /**
 554          * Get the base regex
 555          */
 556         function getBaseRegex() {
 557                 if ( is_null( $this->baseRegex ) ) {
 558                         $this->baseRegex = array( 0 => '', 1 => '' );
 559                         foreach ( $this->names as $name ) {
 560                                 $magic = MagicWord::get( $name );
 561                                 $case = intval( $magic->isCaseSensitive() );
 562                                 foreach ( $magic->getSynonyms() as $i => $syn ) {
 563                                         $group = "(?P<{$i}_{$name}>" . preg_quote( $syn, '/' ) . ')';
 564                                         if ( $this->baseRegex[$case] === '' ) {
 565                                                 $this->baseRegex[$case] = $group;
 566                                         } else {
 567                                                 $this->baseRegex[$case] .= '|' . $group;
 568                                         }
 569                                 }
 570                         }
 571                 }
 572                 return $this->baseRegex;
 573         }
 574
 575         /**
 576          * Get an unanchored regex that does not match parameters
 577          */
 578         function getRegex() {
 579                 if ( is_null( $this->regex ) ) {
 580                         $base = $this->getBaseRegex();
 581                         $this->regex = array( '', '' );
 582                         if ( $this->baseRegex[0] !== '' ) {
 583                                 $this->regex[0] = "/{$base[0]}/iuS";
 584                         }
 585                         if ( $this->baseRegex[1] !== '' ) {
 586                                 $this->regex[1] = "/{$base[1]}/S";
 587                         }
 588                 }
 589                 return $this->regex;
 590         }
 591
 592         /**
 593          * Get a regex for matching variables with parameters
 594          */
 595         function getVariableRegex() {
 596                 return str_replace( "\\$1", "(.*?)", $this->getRegex() );
 597         }
 598
 599         /**
 600          * Get a regex anchored to the start of the string that does not match parameters
 601          */
 602         function getRegexStart() {
 603                 $base = $this->getBaseRegex();
 604                 $newRegex = array( '', '' );
 605                 if ( $base[0] !== '' ) {
 606                         $newRegex[0] = "/^(?:{$base[0]})/iuS";
 607                 }
 608                 if ( $base[1] !== '' ) {
 609                         $newRegex[1] = "/^(?:{$base[1]})/S";
 610                 }
 611                 return $newRegex;
 612         }
 613
 614         /**
 615          * Get an anchored regex for matching variables with parameters
 616          */
 617         function getVariableStartToEndRegex() {
 618                 $base = $this->getBaseRegex();
 619                 $newRegex = array( '', '' );
 620                 if ( $base[0] !== '' ) {
 621                         $newRegex[0] = str_replace( "\\$1", "(.*?)", "/^(?:{$base[0]})$/iuS" );
 622                 }
 623                 if ( $base[1] !== '' ) {
 624                         $newRegex[1] = str_replace( "\\$1", "(.*?)", "/^(?:{$base[1]})$/S" );
 625                 }
 626                 return $newRegex;
 627         }
 628
 629         /**
 630          * Parse a match array from preg_match
 631          * Returns array(magic word ID, parameter value)
 632          * If there is no parameter value, that element will be false.
 633          */
 634         function parseMatch( $m ) {
 635                 reset( $m );
 636                 while ( list( $key, $value ) = each( $m ) ) {
 637                         if ( $key === 0 || $value === '' ) {
 638                                 continue;
 639                         }
 640                         $parts = explode( '_', $key, 2 );
 641                         if ( count( $parts ) != 2 ) {
 642                                 // This shouldn't happen
 643                                 // continue;
 644                                 throw new MWException( __METHOD__ . ': bad parameter name' );
 645                         }
 646                         list( /* $synIndex */, $magicName ) = $parts;
 647                         $paramValue = next( $m );
 648                         return array( $magicName, $paramValue );
 649                 }
 650                 // This shouldn't happen either
 651                 throw new MWException( __METHOD__.': parameter not found' );
 652         }
 653
 654         /**
 655          * Match some text, with parameter capture
 656          * Returns an array with the magic word name in the first element and the
 657          * parameter in the second element.
 658          * Both elements are false if there was no match.
 659          */
 660         public function matchVariableStartToEnd( $text ) {
 661                 $regexes = $this->getVariableStartToEndRegex();
 662                 foreach ( $regexes as $regex ) {
 663                         if ( $regex !== '' ) {
 664                                 $m = false;
 665                                 if ( preg_match( $regex, $text, $m ) ) {
 666                                         return $this->parseMatch( $m );
 667                                 }
 668                         }
 669                 }
 670                 return array( false, false );
 671         }
 672
 673         /**
 674          * Match some text, without parameter capture
 675          * Returns the magic word name, or false if there was no capture
 676          */
 677         public function matchStartToEnd( $text ) {
 678                 $hash = $this->getHash();
 679                 if ( isset( $hash[1][$text] ) ) {
 680                         return $hash[1][$text];
 681                 }
 682                 global $wgContLang;
 683                 $lc = $wgContLang->lc( $text );
 684                 if ( isset( $hash[0][$lc] ) ) {
 685                         return $hash[0][$lc];
 686                 }
 687                 return false;
 688         }
 689
 690         /**
 691          * Returns an associative array, ID => param value, for all items that match
 692          * Removes the matched items from the input string (passed by reference)
 693          */
 694         public function matchAndRemove( &$text ) {
 695                 $found = array();
 696                 $regexes = $this->getRegex();
 697                 foreach ( $regexes as $regex ) {
 698                         if ( $regex === '' ) {
 699                                 continue;
 700                         }
 701                         preg_match_all( $regex, $text, $matches, PREG_SET_ORDER );
 702                         foreach ( $matches as $m ) {
 703                                 list( $name, $param ) = $this->parseMatch( $m );
 704                                 $found[$name] = $param;
 705                         }
 706                         $text = preg_replace( $regex, '', $text );
 707                 }
 708                 return $found;
 709         }
 710
 711         /**
 712          * Return the ID of the magic word at the start of $text, and remove
 713          * the prefix from $text.
 714          * Return false if no match found and $text is not modified.
 715          * Does not match parameters.
 716          */
 717         public function matchStartAndRemove( &$text ) {
 718                 $regexes = $this->getRegexStart();
 719                 foreach ( $regexes as $regex ) {
 720                         if ( $regex === '' ) {
 721                                 continue;
 722                         }
 723                         if ( preg_match( $regex, $text, $m ) ) {
 724                                 list( $id, ) = $this->parseMatch( $m );
 725                                 if ( strlen( $m[0] ) >= strlen( $text ) ) {
 726                                         $text = '';
 727                                 } else {
 728                                         $text = substr( $text, strlen( $m[0] ) );
 729                                 }
 730                                 return $id;
 731                         }
 732                 }
 733                 return false;
 734         }
 735 }