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                 'notitleconvert',
 165                 'nocontentconvert',
 166         );
 167
 168         static public $mSubstIDs = array(
 169                 'subst',
 170                 'safesubst',
 171         );
 172
 173         static public $mObjects = array();
 174         static public $mDoubleUnderscoreArray = null;
 175
 176         /**#@-*/
 177
 178         function __construct($id = 0, $syn = '', $cs = false) {
 179                 $this->mId = $id;
 180                 $this->mSynonyms = (array)$syn;
 181                 $this->mCaseSensitive = $cs;
 182                 $this->mRegex = '';
 183                 $this->mRegexStart = '';
 184                 $this->mVariableRegex = '';
 185                 $this->mVariableStartToEndRegex = '';
 186                 $this->mModified = false;
 187         }
 188
 189         /**
 190          * Factory: creates an object representing an ID
 191          * @static
 192          */
 193         static function &get( $id ) {
 194                 wfProfileIn( __METHOD__ );
 195                 if ( !isset( self::$mObjects[$id] ) ) {
 196                         $mw = new MagicWord();
 197                         $mw->load( $id );
 198                         self::$mObjects[$id] = $mw;
 199                 }
 200                 wfProfileOut( __METHOD__ );
 201                 return self::$mObjects[$id];
 202         }
 203
 204         /**
 205          * Get an array of parser variable IDs
 206          */
 207         static function getVariableIDs() {
 208                 if ( !self::$mVariableIDsInitialised ) {
 209                         # Deprecated constant definition hook, available for extensions that need it
 210                         $magicWords = array();
 211                         wfRunHooks( 'MagicWordMagicWords', array( &$magicWords ) );
 212                         foreach ( $magicWords as $word ) {
 213                                 define( $word, $word );
 214                         }
 215
 216                         # Get variable IDs
 217                         wfRunHooks( 'MagicWordwgVariableIDs', array( &self::$mVariableIDs ) );
 218                         self::$mVariableIDsInitialised = true;
 219                 }
 220                 return self::$mVariableIDs;
 221         }
 222
 223         /**
 224          * Get an array of parser substitution modifier IDs
 225          */
 226         static function getSubstIDs() {
 227                 return self::$mSubstIDs;
 228         }
 229
 230         /* Allow external reads of TTL array */
 231         static function getCacheTTL($id) {
 232                 if (array_key_exists($id,self::$mCacheTTLs)) {
 233                         return self::$mCacheTTLs[$id];
 234                 } else {
 235                         return -1;
 236                 }
 237         }
 238
 239         /** Get a MagicWordArray of double-underscore entities */
 240         static function getDoubleUnderscoreArray() {
 241                 if ( is_null( self::$mDoubleUnderscoreArray ) ) {
 242                         self::$mDoubleUnderscoreArray = new MagicWordArray( self::$mDoubleUnderscoreIDs );
 243                 }
 244                 return self::$mDoubleUnderscoreArray;
 245         }
 246
 247         /**
 248          * Clear the self::$mObjects variable
 249          * For use in parser tests
 250          */
 251         public static function clearCache() {
 252                 self::$mObjects = array();
 253         }
 254
 255         # Initialises this object with an ID
 256         function load( $id ) {
 257                 global $wgContLang;
 258                 $this->mId = $id;
 259                 $wgContLang->getMagic( $this );
 260                 if ( !$this->mSynonyms ) {
 261                         $this->mSynonyms = array( 'dkjsagfjsgashfajsh' );
 262                         #throw new MWException( "Error: invalid magic word '$id'" );
 263                         wfDebugLog( 'exception', "Error: invalid magic word '$id'\n" );
 264                 }
 265         }
 266
 267         /**
 268          * Preliminary initialisation
 269          * @private
 270          */
 271         function initRegex() {
 272                 #$variableClass = Title::legalChars();
 273                 # This was used for matching "$1" variables, but different uses of the feature will have
 274                 # different restrictions, which should be checked *after* the MagicWord has been matched,
 275                 # not here. - IMSoP
 276
 277                 $escSyn = array();
 278                 foreach ( $this->mSynonyms as $synonym )
 279                         // In case a magic word contains /, like that's going to happen;)
 280                         $escSyn[] = preg_quote( $synonym, '/' );
 281                 $this->mBaseRegex = implode( '|', $escSyn );
 282
 283                 $case = $this->mCaseSensitive ? '' : 'iu';
 284                 $this->mRegex = "/{$this->mBaseRegex}/{$case}";
 285                 $this->mRegexStart = "/^(?:{$this->mBaseRegex})/{$case}";
 286                 $this->mVariableRegex = str_replace( "\\$1", "(.*?)", $this->mRegex );
 287                 $this->mVariableStartToEndRegex = str_replace( "\\$1", "(.*?)",
 288                         "/^(?:{$this->mBaseRegex})$/{$case}" );
 289         }
 290
 291         /**
 292          * Gets a regex representing matching the word
 293          */
 294         function getRegex() {
 295                 if ($this->mRegex == '' ) {
 296                         $this->initRegex();
 297                 }
 298                 return $this->mRegex;
 299         }
 300
 301         /**
 302          * Gets the regexp case modifier to use, i.e. i or nothing, to be used if
 303          * one is using MagicWord::getBaseRegex(), otherwise it'll be included in
 304          * the complete expression
 305          */
 306         function getRegexCase() {
 307                 if ( $this->mRegex === '' )
 308                         $this->initRegex();
 309
 310                 return $this->mCaseSensitive ? '' : 'iu';
 311         }
 312
 313         /**
 314          * Gets a regex matching the word, if it is at the string start
 315          */
 316         function getRegexStart() {
 317                 if ($this->mRegex == '' ) {
 318                         $this->initRegex();
 319                 }
 320                 return $this->mRegexStart;
 321         }
 322
 323         /**
 324          * regex without the slashes and what not
 325          */
 326         function getBaseRegex() {
 327                 if ($this->mRegex == '') {
 328                         $this->initRegex();
 329                 }
 330                 return $this->mBaseRegex;
 331         }
 332
 333         /**
 334          * Returns true if the text contains the word
 335          * @return bool
 336          */
 337         function match( $text ) {
 338                 return (bool)preg_match( $this->getRegex(), $text );
 339         }
 340
 341         /**
 342          * Returns true if the text starts with the word
 343          * @return bool
 344          */
 345         function matchStart( $text ) {
 346                 return (bool)preg_match( $this->getRegexStart(), $text );
 347         }
 348
 349         /**
 350          * Returns NULL if there's no match, the value of $1 otherwise
 351          * The return code is the matched string, if there's no variable
 352          * part in the regex and the matched variable part ($1) if there
 353          * is one.
 354          */
 355         function matchVariableStartToEnd( $text ) {
 356                 $matches = array();
 357                 $matchcount = preg_match( $this->getVariableStartToEndRegex(), $text, $matches );
 358                 if ( $matchcount == 0 ) {
 359                         return null;
 360                 } else {
 361                         # multiple matched parts (variable match); some will be empty because of
 362                         # synonyms. The variable will be the second non-empty one so remove any
 363                         # blank elements and re-sort the indices.
 364                         # See also bug 6526
 365
 366                         $matches = array_values(array_filter($matches));
 367
 368                         if ( count($matches) == 1 ) { return $matches[0]; }
 369                         else { return $matches[1]; }
 370                 }
 371         }
 372
 373
 374         /**
 375          * Returns true if the text matches the word, and alters the
 376          * input string, removing all instances of the word
 377          */
 378         function matchAndRemove( &$text ) {
 379                 $this->mFound = false;
 380                 $text = preg_replace_callback( $this->getRegex(), array( &$this, 'pregRemoveAndRecord' ), $text );
 381                 return $this->mFound;
 382         }
 383
 384         function matchStartAndRemove( &$text ) {
 385                 $this->mFound = false;
 386                 $text = preg_replace_callback( $this->getRegexStart(), array( &$this, 'pregRemoveAndRecord' ), $text );
 387                 return $this->mFound;
 388         }
 389
 390         /**
 391          * Used in matchAndRemove()
 392          * @private
 393          **/
 394         function pregRemoveAndRecord( ) {
 395                 $this->mFound = true;
 396                 return '';
 397         }
 398
 399         /**
 400          * Replaces the word with something else
 401          */
 402         function replace( $replacement, $subject, $limit=-1 ) {
 403                 $res = preg_replace( $this->getRegex(), StringUtils::escapeRegexReplacement( $replacement ), $subject, $limit );
 404                 $this->mModified = !($res === $subject);
 405                 return $res;
 406         }
 407
 408         /**
 409          * Variable handling: {{SUBST:xxx}} style words
 410          * Calls back a function to determine what to replace xxx with
 411          * Input word must contain $1
 412          */
 413         function substituteCallback( $text, $callback ) {
 414                 $res = preg_replace_callback( $this->getVariableRegex(), $callback, $text );
 415                 $this->mModified = !($res === $text);
 416                 return $res;
 417         }
 418
 419         /**
 420          * Matches the word, where $1 is a wildcard
 421          */
 422         function getVariableRegex()     {
 423                 if ( $this->mVariableRegex == '' ) {
 424                         $this->initRegex();
 425                 }
 426                 return $this->mVariableRegex;
 427         }
 428
 429         /**
 430          * Matches the entire string, where $1 is a wildcard
 431          */
 432         function getVariableStartToEndRegex() {
 433                 if ( $this->mVariableStartToEndRegex == '' ) {
 434                         $this->initRegex();
 435                 }
 436                 return $this->mVariableStartToEndRegex;
 437         }
 438
 439         /**
 440          * Accesses the synonym list directly
 441          */
 442         function getSynonym( $i ) {
 443                 return $this->mSynonyms[$i];
 444         }
 445
 446         function getSynonyms() {
 447                 return $this->mSynonyms;
 448         }
 449
 450         /**
 451          * Returns true if the last call to replace() or substituteCallback()
 452          * returned a modified text, otherwise false.
 453          */
 454         function getWasModified(){
 455                 return $this->mModified;
 456         }
 457
 458         /**
 459          * $magicarr is an associative array of (magic word ID => replacement)
 460          * This method uses the php feature to do several replacements at the same time,
 461          * thereby gaining some efficiency. The result is placed in the out variable
 462          * $result. The return value is true if something was replaced.
 463          * @static
 464          **/
 465         function replaceMultiple( $magicarr, $subject, &$result ){
 466                 $search = array();
 467                 $replace = array();
 468                 foreach( $magicarr as $id => $replacement ){
 469                         $mw = MagicWord::get( $id );
 470                         $search[] = $mw->getRegex();
 471                         $replace[] = $replacement;
 472                 }
 473
 474                 $result = preg_replace( $search, $replace, $subject );
 475                 return !($result === $subject);
 476         }
 477
 478         /**
 479          * Adds all the synonyms of this MagicWord to an array, to allow quick
 480          * lookup in a list of magic words
 481          */
 482         function addToArray( &$array, $value ) {
 483                 global $wgContLang;
 484                 foreach ( $this->mSynonyms as $syn ) {
 485                         $array[$wgContLang->lc($syn)] = $value;
 486                 }
 487         }
 488
 489         function isCaseSensitive() {
 490                 return $this->mCaseSensitive;
 491         }
 492
 493         function getId() {
 494                 return $this->mId;
 495         }
 496 }
 497
 498 /**
 499  * Class for handling an array of magic words
 500  * @ingroup Parser
 501  */
 502 class MagicWordArray {
 503         var $names = array();
 504         var $hash;
 505         var $baseRegex, $regex;
 506         var $matches;
 507
 508         function __construct( $names = array() ) {
 509                 $this->names = $names;
 510         }
 511
 512         /**
 513          * Add a magic word by name
 514          */
 515         public function add( $name ) {
 516                 global $wgContLang;
 517                 $this->names[] = $name;
 518                 $this->hash = $this->baseRegex = $this->regex = null;
 519         }
 520
 521         /**
 522          * Add a number of magic words by name
 523          */
 524         public function addArray( $names ) {
 525                 $this->names = array_merge( $this->names, array_values( $names ) );
 526                 $this->hash = $this->baseRegex = $this->regex = null;
 527         }
 528
 529         /**
 530          * Get a 2-d hashtable for this array
 531          */
 532         function getHash() {
 533                 if ( is_null( $this->hash ) ) {
 534                         global $wgContLang;
 535                         $this->hash = array( 0 => array(), 1 => array() );
 536                         foreach ( $this->names as $name ) {
 537                                 $magic = MagicWord::get( $name );
 538                                 $case = intval( $magic->isCaseSensitive() );
 539                                 foreach ( $magic->getSynonyms() as $syn ) {
 540                                         if ( !$case ) {
 541                                                 $syn = $wgContLang->lc( $syn );
 542                                         }
 543                                         $this->hash[$case][$syn] = $name;
 544                                 }
 545                         }
 546                 }
 547                 return $this->hash;
 548         }
 549
 550         /**
 551          * Get the base regex
 552          */
 553         function getBaseRegex() {
 554                 if ( is_null( $this->baseRegex ) ) {
 555                         $this->baseRegex = array( 0 => '', 1 => '' );
 556                         foreach ( $this->names as $name ) {
 557                                 $magic = MagicWord::get( $name );
 558                                 $case = intval( $magic->isCaseSensitive() );
 559                                 foreach ( $magic->getSynonyms() as $i => $syn ) {
 560                                         $group = "(?P<{$i}_{$name}>" . preg_quote( $syn, '/' ) . ')';
 561                                         if ( $this->baseRegex[$case] === '' ) {
 562                                                 $this->baseRegex[$case] = $group;
 563                                         } else {
 564                                                 $this->baseRegex[$case] .= '|' . $group;
 565                                         }
 566                                 }
 567                         }
 568                 }
 569                 return $this->baseRegex;
 570         }
 571
 572         /**
 573          * Get an unanchored regex that does not match parameters
 574          */
 575         function getRegex() {
 576                 if ( is_null( $this->regex ) ) {
 577                         $base = $this->getBaseRegex();
 578                         $this->regex = array( '', '' );
 579                         if ( $this->baseRegex[0] !== '' ) {
 580                                 $this->regex[0] = "/{$base[0]}/iuS";
 581                         }
 582                         if ( $this->baseRegex[1] !== '' ) {
 583                                 $this->regex[1] = "/{$base[1]}/S";
 584                         }
 585                 }
 586                 return $this->regex;
 587         }
 588
 589         /**
 590          * Get a regex for matching variables with parameters
 591          */
 592         function getVariableRegex() {
 593                 return str_replace( "\\$1", "(.*?)", $this->getRegex() );
 594         }
 595
 596         /**
 597          * Get a regex anchored to the start of the string that does not match parameters
 598          */
 599         function getRegexStart() {
 600                 $base = $this->getBaseRegex();
 601                 $newRegex = array( '', '' );
 602                 if ( $base[0] !== '' ) {
 603                         $newRegex[0] = "/^(?:{$base[0]})/iuS";
 604                 }
 605                 if ( $base[1] !== '' ) {
 606                         $newRegex[1] = "/^(?:{$base[1]})/S";
 607                 }
 608                 return $newRegex;
 609         }
 610
 611         /**
 612          * Get an anchored regex for matching variables with parameters
 613          */
 614         function getVariableStartToEndRegex() {
 615                 $base = $this->getBaseRegex();
 616                 $newRegex = array( '', '' );
 617                 if ( $base[0] !== '' ) {
 618                         $newRegex[0] = str_replace( "\\$1", "(.*?)", "/^(?:{$base[0]})$/iuS" );
 619                 }
 620                 if ( $base[1] !== '' ) {
 621                         $newRegex[1] = str_replace( "\\$1", "(.*?)", "/^(?:{$base[1]})$/S" );
 622                 }
 623                 return $newRegex;
 624         }
 625
 626         /**
 627          * Parse a match array from preg_match
 628          * Returns array(magic word ID, parameter value)
 629          * If there is no parameter value, that element will be false.
 630          */
 631         function parseMatch( $m ) {
 632                 reset( $m );
 633                 while ( list( $key, $value ) = each( $m ) ) {
 634                         if ( $key === 0 || $value === '' ) {
 635                                 continue;
 636                         }
 637                         $parts = explode( '_', $key, 2 );
 638                         if ( count( $parts ) != 2 ) {
 639                                 // This shouldn't happen
 640                                 // continue;
 641                                 throw new MWException( __METHOD__ . ': bad parameter name' );
 642                         }
 643                         list( /* $synIndex */, $magicName ) = $parts;
 644                         $paramValue = next( $m );
 645                         return array( $magicName, $paramValue );
 646                 }
 647                 // This shouldn't happen either
 648                 throw new MWException( __METHOD__.': parameter not found' );
 649                 return array( false, false );
 650         }
 651
 652         /**
 653          * Match some text, with parameter capture
 654          * Returns an array with the magic word name in the first element and the
 655          * parameter in the second element.
 656          * Both elements are false if there was no match.
 657          */
 658         public function matchVariableStartToEnd( $text ) {
 659                 global $wgContLang;
 660                 $regexes = $this->getVariableStartToEndRegex();
 661                 foreach ( $regexes as $regex ) {
 662                         if ( $regex !== '' ) {
 663                                 $m = false;
 664                                 if ( preg_match( $regex, $text, $m ) ) {
 665                                         return $this->parseMatch( $m );
 666                                 }
 667                         }
 668                 }
 669                 return array( false, false );
 670         }
 671
 672         /**
 673          * Match some text, without parameter capture
 674          * Returns the magic word name, or false if there was no capture
 675          */
 676         public function matchStartToEnd( $text ) {
 677                 $hash = $this->getHash();
 678                 if ( isset( $hash[1][$text] ) ) {
 679                         return $hash[1][$text];
 680                 }
 681                 global $wgContLang;
 682                 $lc = $wgContLang->lc( $text );
 683                 if ( isset( $hash[0][$lc] ) ) {
 684                         return $hash[0][$lc];
 685                 }
 686                 return false;
 687         }
 688
 689         /**
 690          * Returns an associative array, ID => param value, for all items that match
 691          * Removes the matched items from the input string (passed by reference)
 692          */
 693         public function matchAndRemove( &$text ) {
 694                 $found = array();
 695                 $regexes = $this->getRegex();
 696                 foreach ( $regexes as $regex ) {
 697                         if ( $regex === '' ) {
 698                                 continue;
 699                         }
 700                         preg_match_all( $regex, $text, $matches, PREG_SET_ORDER );
 701                         foreach ( $matches as $m ) {
 702                                 list( $name, $param ) = $this->parseMatch( $m );
 703                                 $found[$name] = $param;
 704                         }
 705                         $text = preg_replace( $regex, '', $text );
 706                 }
 707                 return $found;
 708         }
 709
 710         /**
 711          * Return the ID of the magic word at the start of $text, and remove
 712          * the prefix from $text.
 713          * Return false if no match found and $text is not modified.
 714          * Does not match parameters.
 715          */
 716         public function matchStartAndRemove( &$text ) {
 717                 $regexes = $this->getRegexStart();
 718                 foreach ( $regexes as $regex ) {
 719                         if ( $regex === '' ) {
 720                                 continue;
 721                         }
 722                         if ( preg_match( $regex, $text, $m ) ) {
 723                                 list( $id, $param ) = $this->parseMatch( $m );
 724                                 if ( strlen( $m[0] ) >= strlen( $text ) ) {
 725                                         $text = '';
 726                                 } else {
 727                                         $text = substr( $text, strlen( $m[0] ) );
 728                                 }
 729                                 return $id;
 730                         }
 731                 }
 732                 return false;
 733         }
 734 }