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