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