includes/MagicWord.php

   1 <?php
   2 /**
   3  * File for magic words
   4  *
   5  * See docs/magicword.txt
   6  *
   7  * @file
   8  * @ingroup Parser
   9  */
  10
  11 /**
  12  * This class encapsulates "magic words" such as #redirect, __NOTOC__, etc.
  13  * Usage:
  14  *     if (MagicWord::get( 'redirect' )->match( $text ) )
  15  *
  16  * Possible future improvements:
  17  *   * Simultaneous searching for a number of magic words
  18  *   * MagicWord::$mObjects in shared memory
  19  *
  20  * Please avoid reading the data out of one of these objects and then writing
  21  * special case code. If possible, add another match()-like function here.
  22  *
  23  * To add magic words in an extension, use the LanguageGetMagic hook. For
  24  * magic words which are also Parser variables, add a MagicWordwgVariableIDs
  25  * hook. Use string keys.
  26  *
  27  * @ingroup Parser
  28  */
  29 class MagicWord {
  30         /**#@+
  31          * @private
  32          */
  33         var $mId, $mSynonyms, $mCaseSensitive, $mRegex;
  34         var $mRegexStart, $mBaseRegex, $mVariableRegex;
  35         var $mModified, $mFound;
  36
  37         static public $mVariableIDsInitialised = false;
  38         static public $mVariableIDs = array(
  39                 'currentmonth',
  40                 'currentmonth1',
  41                 'currentmonthname',
  42                 'currentmonthnamegen',
  43                 'currentmonthabbrev',
  44                 'currentday',
  45                 'currentday2',
  46                 'currentdayname',
  47                 'currentyear',
  48                 'currenttime',
  49                 'currenthour',
  50                 'localmonth',
  51                 'localmonth1',
  52                 'localmonthname',
  53                 'localmonthnamegen',
  54                 'localmonthabbrev',
  55                 'localday',
  56                 'localday2',
  57                 'localdayname',
  58                 'localyear',
  59                 'localtime',
  60                 'localhour',
  61                 'numberofarticles',
  62                 'numberoffiles',
  63                 'numberofedits',
  64                 'articlepath',
  65                 'sitename',
  66                 'server',
  67                 'servername',
  68                 'scriptpath',
  69                 'stylepath',
  70                 'pagename',
  71                 'pagenamee',
  72                 'fullpagename',
  73                 'fullpagenamee',
  74                 'namespace',
  75                 'namespacee',
  76                 'currentweek',
  77                 'currentdow',
  78                 'localweek',
  79                 'localdow',
  80                 'revisionid',
  81                 'revisionday',
  82                 'revisionday2',
  83                 'revisionmonth',
  84                 'revisionmonth1',
  85                 'revisionyear',
  86                 'revisiontimestamp',
  87                 'revisionuser',
  88                 'subpagename',
  89                 'subpagenamee',
  90                 'talkspace',
  91                 'talkspacee',
  92                 'subjectspace',
  93                 'subjectspacee',
  94                 'talkpagename',
  95                 'talkpagenamee',
  96                 'subjectpagename',
  97                 'subjectpagenamee',
  98                 'numberofusers',
  99                 'numberofactiveusers',
 100                 'numberofpages',
 101                 'currentversion',
 102                 'basepagename',
 103                 'basepagenamee',
 104                 'currenttimestamp',
 105                 'localtimestamp',
 106                 'directionmark',
 107                 'contentlanguage',
 108                 'numberofadmins',
 109                 'numberofviews',
 110         );
 111
 112         /* Array of caching hints for ParserCache */
 113         static public $mCacheTTLs = array (
 114                 'currentmonth' => 86400,
 115                 'currentmonth1' => 86400,
 116                 'currentmonthname' => 86400,
 117                 'currentmonthnamegen' => 86400,
 118                 'currentmonthabbrev' => 86400,
 119                 'currentday' => 3600,
 120                 'currentday2' => 3600,
 121                 'currentdayname' => 3600,
 122                 'currentyear' => 86400,
 123                 'currenttime' => 3600,
 124                 'currenthour' => 3600,
 125                 'localmonth' => 86400,
 126                 'localmonth1' => 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                 'numberofactiveusers' => 3600,
 145                 'numberofpages' => 3600,
 146                 'currentversion' => 86400,
 147                 'currenttimestamp' => 3600,
 148                 'localtimestamp' => 3600,
 149                 'pagesinnamespace' => 3600,
 150                 'numberofadmins' => 3600,
 151                 'numberofviews' => 3600,
 152                 'numberingroup' => 3600,
 153                 );
 154
 155         static public $mDoubleUnderscoreIDs = array(
 156                 'notoc',
 157                 'nogallery',
 158                 'forcetoc',
 159                 'toc',
 160                 'noeditsection',
 161                 'newsectionlink',
 162                 'nonewsectionlink',
 163                 'hiddencat',
 164                 'index',
 165                 'noindex',
 166                 'staticredirect',
 167                 'notitleconvert',
 168                 'nocontentconvert',
 169         );
 170
 171         static public $mSubstIDs = array(
 172                 'subst',
 173                 'safesubst',
 174         );
 175
 176         static public $mObjects = array();
 177         static public $mDoubleUnderscoreArray = null;
 178
 179         /**#@-*/
 180
 181         function __construct($id = 0, $syn = '', $cs = false) {
 182                 $this->mId = $id;
 183                 $this->mSynonyms = (array)$syn;
 184                 $this->mCaseSensitive = $cs;
 185                 $this->mRegex = '';
 186                 $this->mRegexStart = '';
 187                 $this->mVariableRegex = '';
 188                 $this->mVariableStartToEndRegex = '';
 189                 $this->mModified = false;
 190         }
 191
 192         /**
 193          * Factory: creates an object representing an ID
 194          * @static
 195          */
 196         static function &get( $id ) {
 197                 wfProfileIn( __METHOD__ );
 198                 if ( !isset( self::$mObjects[$id] ) ) {
 199                         $mw = new MagicWord();
 200                         $mw->load( $id );
 201                         self::$mObjects[$id] = $mw;
 202                 }
 203                 wfProfileOut( __METHOD__ );
 204                 return self::$mObjects[$id];
 205         }
 206
 207         /**
 208          * Get an array of parser variable IDs
 209          */
 210         static function getVariableIDs() {
 211                 if ( !self::$mVariableIDsInitialised ) {
 212                         # Deprecated constant definition hook, available for extensions that need it
 213                         $magicWords = array();
 214                         wfRunHooks( 'MagicWordMagicWords', array( &$magicWords ) );
 215                         foreach ( $magicWords as $word ) {
 216                                 define( $word, $word );
 217                         }
 218
 219                         # Get variable IDs
 220                         wfRunHooks( 'MagicWordwgVariableIDs', array( &self::$mVariableIDs ) );
 221                         self::$mVariableIDsInitialised = true;
 222                 }
 223                 return self::$mVariableIDs;
 224         }
 225
 226         /**
 227          * Get an array of parser substitution modifier IDs
 228          */
 229         static function getSubstIDs() {
 230                 return self::$mSubstIDs;
 231         }
 232
 233         /* Allow external reads of TTL array */
 234         static function getCacheTTL($id) {
 235                 if (array_key_exists($id,self::$mCacheTTLs)) {
 236                         return self::$mCacheTTLs[$id];
 237                 } else {
 238                         return -1;
 239                 }
 240         }
 241
 242         /** Get a MagicWordArray of double-underscore entities */
 243         static function getDoubleUnderscoreArray() {
 244                 if ( is_null( self::$mDoubleUnderscoreArray ) ) {
 245                         self::$mDoubleUnderscoreArray = new MagicWordArray( self::$mDoubleUnderscoreIDs );
 246                 }
 247                 return self::$mDoubleUnderscoreArray;
 248         }
 249
 250         /**
 251          * Clear the self::$mObjects variable
 252          * For use in parser tests
 253          */
 254         public static function clearCache() {
 255                 self::$mObjects = array();
 256         }
 257
 258         # Initialises this object with an ID
 259         function load( $id ) {
 260                 global $wgContLang;
 261                 $this->mId = $id;
 262                 $wgContLang->getMagic( $this );
 263                 if ( !$this->mSynonyms ) {
 264                         $this->mSynonyms = array( 'dkjsagfjsgashfajsh' );
 265                         #throw new MWException( "Error: invalid magic word '$id'" );
 266                         wfDebugLog( 'exception', "Error: invalid magic word '$id'\n" );
 267                 }
 268         }
 269
 270         /**
 271          * Preliminary initialisation
 272          * @private
 273          */
 274         function initRegex() {
 275                 #$variableClass = Title::legalChars();
 276                 # This was used for matching "$1" variables, but different uses of the feature will have
 277                 # different restrictions, which should be checked *after* the MagicWord has been matched,
 278                 # not here. - IMSoP
 279
 280                 $escSyn = array();
 281                 foreach ( $this->mSynonyms as $synonym )
 282                         // In case a magic word contains /, like that's going to happen;)
 283                         $escSyn[] = preg_quote( $synonym, '/' );
 284                 $this->mBaseRegex = implode( '|', $escSyn );
 285
 286                 $case = $this->mCaseSensitive ? '' : 'iu';
 287                 $this->mRegex = "/{$this->mBaseRegex}/{$case}";
 288                 $this->mRegexStart = "/^(?:{$this->mBaseRegex})/{$case}";
 289                 $this->mVariableRegex = str_replace( "\\$1", "(.*?)", $this->mRegex );
 290                 $this->mVariableStartToEndRegex = str_replace( "\\$1", "(.*?)",
 291                         "/^(?:{$this->mBaseRegex})$/{$case}" );
 292         }
 293
 294         /**
 295          * Gets a regex representing matching the word
 296          */
 297         function getRegex() {
 298                 if ($this->mRegex == '' ) {
 299                         $this->initRegex();
 300                 }
 301                 return $this->mRegex;
 302         }
 303
 304         /**
 305          * Gets the regexp case modifier to use, i.e. i or nothing, to be used if
 306          * one is using MagicWord::getBaseRegex(), otherwise it'll be included in
 307          * the complete expression
 308          */
 309         function getRegexCase() {
 310                 if ( $this->mRegex === '' )
 311                         $this->initRegex();
 312
 313                 return $this->mCaseSensitive ? '' : 'iu';
 314         }
 315
 316         /**
 317          * Gets a regex matching the word, if it is at the string start
 318          */
 319         function getRegexStart() {
 320                 if ($this->mRegex == '' ) {
 321                         $this->initRegex();
 322                 }
 323                 return $this->mRegexStart;
 324         }
 325
 326         /**
 327          * regex without the slashes and what not
 328          */
 329         function getBaseRegex() {
 330                 if ($this->mRegex == '') {
 331                         $this->initRegex();
 332                 }
 333                 return $this->mBaseRegex;
 334         }
 335
 336         /**
 337          * Returns true if the text contains the word
 338          * @return bool
 339          */
 340         function match( $text ) {
 341                 return (bool)preg_match( $this->getRegex(), $text );
 342         }
 343
 344         /**
 345          * Returns true if the text starts with the word
 346          * @return bool
 347          */
 348         function matchStart( $text ) {
 349                 return (bool)preg_match( $this->getRegexStart(), $text );
 350         }
 351
 352         /**
 353          * Returns NULL if there's no match, the value of $1 otherwise
 354          * The return code is the matched string, if there's no variable
 355          * part in the regex and the matched variable part ($1) if there
 356          * is one.
 357          */
 358         function matchVariableStartToEnd( $text ) {
 359                 $matches = array();
 360                 $matchcount = preg_match( $this->getVariableStartToEndRegex(), $text, $matches );
 361                 if ( $matchcount == 0 ) {
 362                         return null;
 363                 } else {
 364                         # multiple matched parts (variable match); some will be empty because of
 365                         # synonyms. The variable will be the second non-empty one so remove any
 366                         # blank elements and re-sort the indices.
 367                         # See also bug 6526
 368
 369                         $matches = array_values(array_filter($matches));
 370
 371                         if ( count($matches) == 1 ) { return $matches[0]; }
 372                         else { return $matches[1]; }
 373                 }
 374         }
 375
 376
 377         /**
 378          * Returns true if the text matches the word, and alters the
 379          * input string, removing all instances of the word
 380          */
 381         function matchAndRemove( &$text ) {
 382                 $this->mFound = false;
 383                 $text = preg_replace_callback( $this->getRegex(), array( &$this, 'pregRemoveAndRecord' ), $text );
 384                 return $this->mFound;
 385         }
 386
 387         function matchStartAndRemove( &$text ) {
 388                 $this->mFound = false;
 389                 $text = preg_replace_callback( $this->getRegexStart(), array( &$this, 'pregRemoveAndRecord' ), $text );
 390                 return $this->mFound;
 391         }
 392
 393         /**
 394          * Used in matchAndRemove()
 395          * @private
 396          **/
 397         function pregRemoveAndRecord( ) {
 398                 $this->mFound = true;
 399                 return '';
 400         }
 401
 402         /**
 403          * Replaces the word with something else
 404          */
 405         function replace( $replacement, $subject, $limit=-1 ) {
 406                 $res = preg_replace( $this->getRegex(), StringUtils::escapeRegexReplacement( $replacement ), $subject, $limit );
 407                 $this->mModified = !($res === $subject);
 408                 return $res;
 409         }
 410
 411         /**
 412          * Variable handling: {{SUBST:xxx}} style words
 413          * Calls back a function to determine what to replace xxx with
 414          * Input word must contain $1
 415          */
 416         function substituteCallback( $text, $callback ) {
 417                 $res = preg_replace_callback( $this->getVariableRegex(), $callback, $text );
 418                 $this->mModified = !($res === $text);
 419                 return $res;
 420         }
 421
 422         /**
 423          * Matches the word, where $1 is a wildcard
 424          */
 425         function getVariableRegex()     {
 426                 if ( $this->mVariableRegex == '' ) {
 427                         $this->initRegex();
 428                 }
 429                 return $this->mVariableRegex;
 430         }
 431
 432         /**
 433          * Matches the entire string, where $1 is a wildcard
 434          */
 435         function getVariableStartToEndRegex() {
 436                 if ( $this->mVariableStartToEndRegex == '' ) {
 437                         $this->initRegex();
 438                 }
 439                 return $this->mVariableStartToEndRegex;
 440         }
 441
 442         /**
 443          * Accesses the synonym list directly
 444          */
 445         function getSynonym( $i ) {
 446                 return $this->mSynonyms[$i];
 447         }
 448
 449         function getSynonyms() {
 450                 return $this->mSynonyms;
 451         }
 452
 453         /**
 454          * Returns true if the last call to replace() or substituteCallback()
 455          * returned a modified text, otherwise false.
 456          */
 457         function getWasModified(){
 458                 return $this->mModified;
 459         }
 460
 461         /**
 462          * $magicarr is an associative array of (magic word ID => replacement)
 463          * This method uses the php feature to do several replacements at the same time,
 464          * thereby gaining some efficiency. The result is placed in the out variable
 465          * $result. The return value is true if something was replaced.
 466          * @static
 467          **/
 468         function replaceMultiple( $magicarr, $subject, &$result ){
 469                 $search = array();
 470                 $replace = array();
 471                 foreach( $magicarr as $id => $replacement ){
 472                         $mw = MagicWord::get( $id );
 473                         $search[] = $mw->getRegex();
 474                         $replace[] = $replacement;
 475                 }
 476
 477                 $result = preg_replace( $search, $replace, $subject );
 478                 return !($result === $subject);
 479         }
 480
 481         /**
 482          * Adds all the synonyms of this MagicWord to an array, to allow quick
 483          * lookup in a list of magic words
 484          */
 485         function addToArray( &$array, $value ) {
 486                 global $wgContLang;
 487                 foreach ( $this->mSynonyms as $syn ) {
 488                         $array[$wgContLang->lc($syn)] = $value;
 489                 }
 490         }
 491
 492         function isCaseSensitive() {
 493                 return $this->mCaseSensitive;
 494         }
 495
 496         function getId() {
 497                 return $this->mId;
 498         }
 499 }
 500
 501 /**
 502  * Class for handling an array of magic words
 503  * @ingroup Parser
 504  */
 505 class MagicWordArray {
 506         var $names = array();
 507         var $hash;
 508         var $baseRegex, $regex;
 509         var $matches;
 510
 511         function __construct( $names = array() ) {
 512                 $this->names = $names;
 513         }
 514
 515         /**
 516          * Add a magic word by name
 517          */
 518         public function add( $name ) {
 519                 $this->names[] = $name;
 520                 $this->hash = $this->baseRegex = $this->regex = null;
 521         }
 522
 523         /**
 524          * Add a number of magic words by name
 525          */
 526         public function addArray( $names ) {
 527                 $this->names = array_merge( $this->names, array_values( $names ) );
 528                 $this->hash = $this->baseRegex = $this->regex = null;
 529         }
 530
 531         /**
 532          * Get a 2-d hashtable for this array
 533          */
 534         function getHash() {
 535                 if ( is_null( $this->hash ) ) {
 536                         global $wgContLang;
 537                         $this->hash = array( 0 => array(), 1 => array() );
 538                         foreach ( $this->names as $name ) {
 539                                 $magic = MagicWord::get( $name );
 540                                 $case = intval( $magic->isCaseSensitive() );
 541                                 foreach ( $magic->getSynonyms() as $syn ) {
 542                                         if ( !$case ) {
 543                                                 $syn = $wgContLang->lc( $syn );
 544                                         }
 545                                         $this->hash[$case][$syn] = $name;
 546                                 }
 547                         }
 548                 }
 549                 return $this->hash;
 550         }
 551
 552         /**
 553          * Get the base regex
 554          */
 555         function getBaseRegex() {
 556                 if ( is_null( $this->baseRegex ) ) {
 557                         $this->baseRegex = array( 0 => '', 1 => '' );
 558                         foreach ( $this->names as $name ) {
 559                                 $magic = MagicWord::get( $name );
 560                                 $case = intval( $magic->isCaseSensitive() );
 561                                 foreach ( $magic->getSynonyms() as $i => $syn ) {
 562                                         $group = "(?P<{$i}_{$name}>" . preg_quote( $syn, '/' ) . ')';
 563                                         if ( $this->baseRegex[$case] === '' ) {
 564                                                 $this->baseRegex[$case] = $group;
 565                                         } else {
 566                                                 $this->baseRegex[$case] .= '|' . $group;
 567                                         }
 568                                 }
 569                         }
 570                 }
 571                 return $this->baseRegex;
 572         }
 573
 574         /**
 575          * Get an unanchored regex that does not match parameters
 576          */
 577         function getRegex() {
 578                 if ( is_null( $this->regex ) ) {
 579                         $base = $this->getBaseRegex();
 580                         $this->regex = array( '', '' );
 581                         if ( $this->baseRegex[0] !== '' ) {
 582                                 $this->regex[0] = "/{$base[0]}/iuS";
 583                         }
 584                         if ( $this->baseRegex[1] !== '' ) {
 585                                 $this->regex[1] = "/{$base[1]}/S";
 586                         }
 587                 }
 588                 return $this->regex;
 589         }
 590
 591         /**
 592          * Get a regex for matching variables with parameters
 593          */
 594         function getVariableRegex() {
 595                 return str_replace( "\\$1", "(.*?)", $this->getRegex() );
 596         }
 597
 598         /**
 599          * Get a regex anchored to the start of the string that does not match parameters
 600          */
 601         function getRegexStart() {
 602                 $base = $this->getBaseRegex();
 603                 $newRegex = array( '', '' );
 604                 if ( $base[0] !== '' ) {
 605                         $newRegex[0] = "/^(?:{$base[0]})/iuS";
 606                 }
 607                 if ( $base[1] !== '' ) {
 608                         $newRegex[1] = "/^(?:{$base[1]})/S";
 609                 }
 610                 return $newRegex;
 611         }
 612
 613         /**
 614          * Get an anchored regex for matching variables with parameters
 615          */
 616         function getVariableStartToEndRegex() {
 617                 $base = $this->getBaseRegex();
 618                 $newRegex = array( '', '' );
 619                 if ( $base[0] !== '' ) {
 620                         $newRegex[0] = str_replace( "\\$1", "(.*?)", "/^(?:{$base[0]})$/iuS" );
 621                 }
 622                 if ( $base[1] !== '' ) {
 623                         $newRegex[1] = str_replace( "\\$1", "(.*?)", "/^(?:{$base[1]})$/S" );
 624                 }
 625                 return $newRegex;
 626         }
 627
 628         /**
 629          * Parse a match array from preg_match
 630          * Returns array(magic word ID, parameter value)
 631          * If there is no parameter value, that element will be false.
 632          */
 633         function parseMatch( $m ) {
 634                 reset( $m );
 635                 while ( list( $key, $value ) = each( $m ) ) {
 636                         if ( $key === 0 || $value === '' ) {
 637                                 continue;
 638                         }
 639                         $parts = explode( '_', $key, 2 );
 640                         if ( count( $parts ) != 2 ) {
 641                                 // This shouldn't happen
 642                                 // continue;
 643                                 throw new MWException( __METHOD__ . ': bad parameter name' );
 644                         }
 645                         list( /* $synIndex */, $magicName ) = $parts;
 646                         $paramValue = next( $m );
 647                         return array( $magicName, $paramValue );
 648                 }
 649                 // This shouldn't happen either
 650                 throw new MWException( __METHOD__.': parameter not found' );
 651         }
 652
 653         /**
 654          * Match some text, with parameter capture
 655          * Returns an array with the magic word name in the first element and the
 656          * parameter in the second element.
 657          * Both elements are false if there was no match.
 658          */
 659         public function matchVariableStartToEnd( $text ) {
 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, ) = $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 }