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