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