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