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                 'revisionmonth1',
  83                 'revisionyear',
  84                 'revisiontimestamp',
  85                 'revisionuser',
  86                 'subpagename',
  87                 'subpagenamee',
  88                 'talkspace',
  89                 'talkspacee',
  90                 'subjectspace',
  91                 'subjectspacee',
  92                 'talkpagename',
  93                 'talkpagenamee',
  94                 'subjectpagename',
  95                 'subjectpagenamee',
  96                 'numberofusers',
  97                 'numberofactiveusers',
  98                 'numberofpages',
  99                 'currentversion',
 100                 'basepagename',
 101                 'basepagenamee',
 102                 'currenttimestamp',
 103                 'localtimestamp',
 104                 'directionmark',
 105                 'contentlanguage',
 106                 'numberofadmins',
 107                 'numberofviews',
 108         );
 109
 110         /* Array of caching hints for ParserCache */
 111         static public $mCacheTTLs = array (
 112                 'currentmonth' => 86400,
 113                 'currentmonth1' => 86400,
 114                 'currentmonthname' => 86400,
 115                 'currentmonthnamegen' => 86400,
 116                 'currentmonthabbrev' => 86400,
 117                 'currentday' => 3600,
 118                 'currentday2' => 3600,
 119                 'currentdayname' => 3600,
 120                 'currentyear' => 86400,
 121                 'currenttime' => 3600,
 122                 'currenthour' => 3600,
 123                 'localmonth' => 86400,
 124                 'localmonth1' => 86400,
 125                 'localmonthname' => 86400,
 126                 'localmonthnamegen' => 86400,
 127                 'localmonthabbrev' => 86400,
 128                 'localday' => 3600,
 129                 'localday2' => 3600,
 130                 'localdayname' => 3600,
 131                 'localyear' => 86400,
 132                 'localtime' => 3600,
 133                 'localhour' => 3600,
 134                 'numberofarticles' => 3600,
 135                 'numberoffiles' => 3600,
 136                 'numberofedits' => 3600,
 137                 'currentweek' => 3600,
 138                 'currentdow' => 3600,
 139                 'localweek' => 3600,
 140                 'localdow' => 3600,
 141                 'numberofusers' => 3600,
 142                 'numberofactiveusers' => 3600,
 143                 'numberofpages' => 3600,
 144                 'currentversion' => 86400,
 145                 'currenttimestamp' => 3600,
 146                 'localtimestamp' => 3600,
 147                 'pagesinnamespace' => 3600,
 148                 'numberofadmins' => 3600,
 149                 'numberofviews' => 3600,
 150                 'numberingroup' => 3600,
 151                 );
 152
 153         static public $mDoubleUnderscoreIDs = array(
 154                 'notoc',
 155                 'nogallery',
 156                 'forcetoc',
 157                 'toc',
 158                 'noeditsection',
 159                 'newsectionlink',
 160                 'nonewsectionlink',
 161                 'hiddencat',
 162                 'index',
 163                 'noindex',
 164                 'staticredirect',
 165                 'notitleconvert',
 166                 'nocontentconvert',
 167         );
 168
 169         static public $mSubstIDs = array(
 170                 'subst',
 171                 'safesubst',
 172         );
 173
 174         static public $mObjects = array();
 175         static public $mDoubleUnderscoreArray = null;
 176
 177         /**#@-*/
 178
 179         function __construct($id = 0, $syn = '', $cs = false) {
 180                 $this->mId = $id;
 181                 $this->mSynonyms = (array)$syn;
 182                 $this->mCaseSensitive = $cs;
 183                 $this->mRegex = '';
 184                 $this->mRegexStart = '';
 185                 $this->mVariableRegex = '';
 186                 $this->mVariableStartToEndRegex = '';
 187                 $this->mModified = false;
 188         }
 189
 190         /**
 191          * Factory: creates an object representing an ID
 192          * @static
 193          */
 194         static function &get( $id ) {
 195                 wfProfileIn( __METHOD__ );
 196                 if ( !isset( self::$mObjects[$id] ) ) {
 197                         $mw = new MagicWord();
 198                         $mw->load( $id );
 199                         self::$mObjects[$id] = $mw;
 200                 }
 201                 wfProfileOut( __METHOD__ );
 202                 return self::$mObjects[$id];
 203         }
 204
 205         /**
 206          * Get an array of parser variable IDs
 207          */
 208         static function getVariableIDs() {
 209                 if ( !self::$mVariableIDsInitialised ) {
 210                         # Deprecated constant definition hook, available for extensions that need it
 211                         $magicWords = array();
 212                         wfRunHooks( 'MagicWordMagicWords', array( &$magicWords ) );
 213                         foreach ( $magicWords as $word ) {
 214                                 define( $word, $word );
 215                         }
 216
 217                         # Get variable IDs
 218                         wfRunHooks( 'MagicWordwgVariableIDs', array( &self::$mVariableIDs ) );
 219                         self::$mVariableIDsInitialised = true;
 220                 }
 221                 return self::$mVariableIDs;
 222         }
 223
 224         /**
 225          * Get an array of parser substitution modifier IDs
 226          */
 227         static function getSubstIDs() {
 228                 return self::$mSubstIDs;
 229         }
 230
 231         /* Allow external reads of TTL array */
 232         static function getCacheTTL($id) {
 233                 if (array_key_exists($id,self::$mCacheTTLs)) {
 234                         return self::$mCacheTTLs[$id];
 235                 } else {
 236                         return -1;
 237                 }
 238         }
 239
 240         /** Get a MagicWordArray of double-underscore entities */
 241         static function getDoubleUnderscoreArray() {
 242                 if ( is_null( self::$mDoubleUnderscoreArray ) ) {
 243                         self::$mDoubleUnderscoreArray = new MagicWordArray( self::$mDoubleUnderscoreIDs );
 244                 }
 245                 return self::$mDoubleUnderscoreArray;
 246         }
 247
 248         /**
 249          * Clear the self::$mObjects variable
 250          * For use in parser tests
 251          */
 252         public static function clearCache() {
 253                 self::$mObjects = array();
 254         }
 255
 256         # Initialises this object with an ID
 257         function load( $id ) {
 258                 global $wgContLang;
 259                 $this->mId = $id;
 260                 $wgContLang->getMagic( $this );
 261                 if ( !$this->mSynonyms ) {
 262                         $this->mSynonyms = array( 'dkjsagfjsgashfajsh' );
 263                         #throw new MWException( "Error: invalid magic word '$id'" );
 264                         wfDebugLog( 'exception', "Error: invalid magic word '$id'\n" );
 265                 }
 266         }
 267
 268         /**
 269          * Preliminary initialisation
 270          * @private
 271          */
 272         function initRegex() {
 273                 #$variableClass = Title::legalChars();
 274                 # This was used for matching "$1" variables, but different uses of the feature will have
 275                 # different restrictions, which should be checked *after* the MagicWord has been matched,
 276                 # not here. - IMSoP
 277
 278                 $escSyn = array();
 279                 foreach ( $this->mSynonyms as $synonym )
 280                         // In case a magic word contains /, like that's going to happen;)
 281                         $escSyn[] = preg_quote( $synonym, '/' );
 282                 $this->mBaseRegex = implode( '|', $escSyn );
 283
 284                 $case = $this->mCaseSensitive ? '' : 'iu';
 285                 $this->mRegex = "/{$this->mBaseRegex}/{$case}";
 286                 $this->mRegexStart = "/^(?:{$this->mBaseRegex})/{$case}";
 287                 $this->mVariableRegex = str_replace( "\\$1", "(.*?)", $this->mRegex );
 288                 $this->mVariableStartToEndRegex = str_replace( "\\$1", "(.*?)",
 289                         "/^(?:{$this->mBaseRegex})$/{$case}" );
 290         }
 291
 292         /**
 293          * Gets a regex representing matching the word
 294          */
 295         function getRegex() {
 296                 if ($this->mRegex == '' ) {
 297                         $this->initRegex();
 298                 }
 299                 return $this->mRegex;
 300         }
 301
 302         /**
 303          * Gets the regexp case modifier to use, i.e. i or nothing, to be used if
 304          * one is using MagicWord::getBaseRegex(), otherwise it'll be included in
 305          * the complete expression
 306          */
 307         function getRegexCase() {
 308                 if ( $this->mRegex === '' )
 309                         $this->initRegex();
 310
 311                 return $this->mCaseSensitive ? '' : 'iu';
 312         }
 313
 314         /**
 315          * Gets a regex matching the word, if it is at the string start
 316          */
 317         function getRegexStart() {
 318                 if ($this->mRegex == '' ) {
 319                         $this->initRegex();
 320                 }
 321                 return $this->mRegexStart;
 322         }
 323
 324         /**
 325          * regex without the slashes and what not
 326          */
 327         function getBaseRegex() {
 328                 if ($this->mRegex == '') {
 329                         $this->initRegex();
 330                 }
 331                 return $this->mBaseRegex;
 332         }
 333
 334         /**
 335          * Returns true if the text contains the word
 336          * @return bool
 337          */
 338         function match( $text ) {
 339                 return (bool)preg_match( $this->getRegex(), $text );
 340         }
 341
 342         /**
 343          * Returns true if the text starts with the word
 344          * @return bool
 345          */
 346         function matchStart( $text ) {
 347                 return (bool)preg_match( $this->getRegexStart(), $text );
 348         }
 349
 350         /**
 351          * Returns NULL if there's no match, the value of $1 otherwise
 352          * The return code is the matched string, if there's no variable
 353          * part in the regex and the matched variable part ($1) if there
 354          * is one.
 355          */
 356         function matchVariableStartToEnd( $text ) {
 357                 $matches = array();
 358                 $matchcount = preg_match( $this->getVariableStartToEndRegex(), $text, $matches );
 359                 if ( $matchcount == 0 ) {
 360                         return null;
 361                 } else {
 362                         # multiple matched parts (variable match); some will be empty because of
 363                         # synonyms. The variable will be the second non-empty one so remove any
 364                         # blank elements and re-sort the indices.
 365                         # See also bug 6526
 366
 367                         $matches = array_values(array_filter($matches));
 368
 369                         if ( count($matches) == 1 ) { return $matches[0]; }
 370                         else { return $matches[1]; }
 371                 }
 372         }
 373
 374
 375         /**
 376          * Returns true if the text matches the word, and alters the
 377          * input string, removing all instances of the word
 378          */
 379         function matchAndRemove( &$text ) {
 380                 $this->mFound = false;
 381                 $text = preg_replace_callback( $this->getRegex(), array( &$this, 'pregRemoveAndRecord' ), $text );
 382                 return $this->mFound;
 383         }
 384
 385         function matchStartAndRemove( &$text ) {
 386                 $this->mFound = false;
 387                 $text = preg_replace_callback( $this->getRegexStart(), array( &$this, 'pregRemoveAndRecord' ), $text );
 388                 return $this->mFound;
 389         }
 390
 391         /**
 392          * Used in matchAndRemove()
 393          * @private
 394          **/
 395         function pregRemoveAndRecord( ) {
 396                 $this->mFound = true;
 397                 return '';
 398         }
 399
 400         /**
 401          * Replaces the word with something else
 402          */
 403         function replace( $replacement, $subject, $limit=-1 ) {
 404                 $res = preg_replace( $this->getRegex(), StringUtils::escapeRegexReplacement( $replacement ), $subject, $limit );
 405                 $this->mModified = !($res === $subject);
 406                 return $res;
 407         }
 408
 409         /**
 410          * Variable handling: {{SUBST:xxx}} style words
 411          * Calls back a function to determine what to replace xxx with
 412          * Input word must contain $1
 413          */
 414         function substituteCallback( $text, $callback ) {
 415                 $res = preg_replace_callback( $this->getVariableRegex(), $callback, $text );
 416                 $this->mModified = !($res === $text);
 417                 return $res;
 418         }
 419
 420         /**
 421          * Matches the word, where $1 is a wildcard
 422          */
 423         function getVariableRegex()     {
 424                 if ( $this->mVariableRegex == '' ) {
 425                         $this->initRegex();
 426                 }
 427                 return $this->mVariableRegex;
 428         }
 429
 430         /**
 431          * Matches the entire string, where $1 is a wildcard
 432          */
 433         function getVariableStartToEndRegex() {
 434                 if ( $this->mVariableStartToEndRegex == '' ) {
 435                         $this->initRegex();
 436                 }
 437                 return $this->mVariableStartToEndRegex;
 438         }
 439
 440         /**
 441          * Accesses the synonym list directly
 442          */
 443         function getSynonym( $i ) {
 444                 return $this->mSynonyms[$i];
 445         }
 446
 447         function getSynonyms() {
 448                 return $this->mSynonyms;
 449         }
 450
 451         /**
 452          * Returns true if the last call to replace() or substituteCallback()
 453          * returned a modified text, otherwise false.
 454          */
 455         function getWasModified(){
 456                 return $this->mModified;
 457         }
 458
 459         /**
 460          * $magicarr is an associative array of (magic word ID => replacement)
 461          * This method uses the php feature to do several replacements at the same time,
 462          * thereby gaining some efficiency. The result is placed in the out variable
 463          * $result. The return value is true if something was replaced.
 464          * @static
 465          **/
 466         function replaceMultiple( $magicarr, $subject, &$result ){
 467                 $search = array();
 468                 $replace = array();
 469                 foreach( $magicarr as $id => $replacement ){
 470                         $mw = MagicWord::get( $id );
 471                         $search[] = $mw->getRegex();
 472                         $replace[] = $replacement;
 473                 }
 474
 475                 $result = preg_replace( $search, $replace, $subject );
 476                 return !($result === $subject);
 477         }
 478
 479         /**
 480          * Adds all the synonyms of this MagicWord to an array, to allow quick
 481          * lookup in a list of magic words
 482          */
 483         function addToArray( &$array, $value ) {
 484                 global $wgContLang;
 485                 foreach ( $this->mSynonyms as $syn ) {
 486                         $array[$wgContLang->lc($syn)] = $value;
 487                 }
 488         }
 489
 490         function isCaseSensitive() {
 491                 return $this->mCaseSensitive;
 492         }
 493
 494         function getId() {
 495                 return $this->mId;
 496         }
 497 }
 498
 499 /**
 500  * Class for handling an array of magic words
 501  * @ingroup Parser
 502  */
 503 class MagicWordArray {
 504         var $names = array();
 505         var $hash;
 506         var $baseRegex, $regex;
 507         var $matches;
 508
 509         function __construct( $names = array() ) {
 510                 $this->names = $names;
 511         }
 512
 513         /**
 514          * Add a magic word by name
 515          */
 516         public function add( $name ) {
 517                 global $wgContLang;
 518                 $this->names[] = $name;
 519                 $this->hash = $this->baseRegex = $this->regex = null;
 520         }
 521
 522         /**
 523          * Add a number of magic words by name
 524          */
 525         public function addArray( $names ) {
 526                 $this->names = array_merge( $this->names, array_values( $names ) );
 527                 $this->hash = $this->baseRegex = $this->regex = null;
 528         }
 529
 530         /**
 531          * Get a 2-d hashtable for this array
 532          */
 533         function getHash() {
 534                 if ( is_null( $this->hash ) ) {
 535                         global $wgContLang;
 536                         $this->hash = array( 0 => array(), 1 => array() );
 537                         foreach ( $this->names as $name ) {
 538                                 $magic = MagicWord::get( $name );
 539                                 $case = intval( $magic->isCaseSensitive() );
 540                                 foreach ( $magic->getSynonyms() as $syn ) {
 541                                         if ( !$case ) {
 542                                                 $syn = $wgContLang->lc( $syn );
 543                                         }
 544                                         $this->hash[$case][$syn] = $name;
 545                                 }
 546                         }
 547                 }
 548                 return $this->hash;
 549         }
 550
 551         /**
 552          * Get the base regex
 553          */
 554         function getBaseRegex() {
 555                 if ( is_null( $this->baseRegex ) ) {
 556                         $this->baseRegex = array( 0 => '', 1 => '' );
 557                         foreach ( $this->names as $name ) {
 558                                 $magic = MagicWord::get( $name );
 559                                 $case = intval( $magic->isCaseSensitive() );
 560                                 foreach ( $magic->getSynonyms() as $i => $syn ) {
 561                                         $group = "(?P<{$i}_{$name}>" . preg_quote( $syn, '/' ) . ')';
 562                                         if ( $this->baseRegex[$case] === '' ) {
 563                                                 $this->baseRegex[$case] = $group;
 564                                         } else {
 565                                                 $this->baseRegex[$case] .= '|' . $group;
 566                                         }
 567                                 }
 568                         }
 569                 }
 570                 return $this->baseRegex;
 571         }
 572
 573         /**
 574          * Get an unanchored regex that does not match parameters
 575          */
 576         function getRegex() {
 577                 if ( is_null( $this->regex ) ) {
 578                         $base = $this->getBaseRegex();
 579                         $this->regex = array( '', '' );
 580                         if ( $this->baseRegex[0] !== '' ) {
 581                                 $this->regex[0] = "/{$base[0]}/iuS";
 582                         }
 583                         if ( $this->baseRegex[1] !== '' ) {
 584                                 $this->regex[1] = "/{$base[1]}/S";
 585                         }
 586                 }
 587                 return $this->regex;
 588         }
 589
 590         /**
 591          * Get a regex for matching variables with parameters
 592          */
 593         function getVariableRegex() {
 594                 return str_replace( "\\$1", "(.*?)", $this->getRegex() );
 595         }
 596
 597         /**
 598          * Get a regex anchored to the start of the string that does not match parameters
 599          */
 600         function getRegexStart() {
 601                 $base = $this->getBaseRegex();
 602                 $newRegex = array( '', '' );
 603                 if ( $base[0] !== '' ) {
 604                         $newRegex[0] = "/^(?:{$base[0]})/iuS";
 605                 }
 606                 if ( $base[1] !== '' ) {
 607                         $newRegex[1] = "/^(?:{$base[1]})/S";
 608                 }
 609                 return $newRegex;
 610         }
 611
 612         /**
 613          * Get an anchored regex for matching variables with parameters
 614          */
 615         function getVariableStartToEndRegex() {
 616                 $base = $this->getBaseRegex();
 617                 $newRegex = array( '', '' );
 618                 if ( $base[0] !== '' ) {
 619                         $newRegex[0] = str_replace( "\\$1", "(.*?)", "/^(?:{$base[0]})$/iuS" );
 620                 }
 621                 if ( $base[1] !== '' ) {
 622                         $newRegex[1] = str_replace( "\\$1", "(.*?)", "/^(?:{$base[1]})$/S" );
 623                 }
 624                 return $newRegex;
 625         }
 626
 627         /**
 628          * Parse a match array from preg_match
 629          * Returns array(magic word ID, parameter value)
 630          * If there is no parameter value, that element will be false.
 631          */
 632         function parseMatch( $m ) {
 633                 reset( $m );
 634                 while ( list( $key, $value ) = each( $m ) ) {
 635                         if ( $key === 0 || $value === '' ) {
 636                                 continue;
 637                         }
 638                         $parts = explode( '_', $key, 2 );
 639                         if ( count( $parts ) != 2 ) {
 640                                 // This shouldn't happen
 641                                 // continue;
 642                                 throw new MWException( __METHOD__ . ': bad parameter name' );
 643                         }
 644                         list( /* $synIndex */, $magicName ) = $parts;
 645                         $paramValue = next( $m );
 646                         return array( $magicName, $paramValue );
 647                 }
 648                 // This shouldn't happen either
 649                 throw new MWException( __METHOD__.': parameter not found' );
 650                 return array( false, false );
 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                 global $wgContLang;
 661                 $regexes = $this->getVariableStartToEndRegex();
 662                 foreach ( $regexes as $regex ) {
 663                         if ( $regex !== '' ) {
 664                                 $m = false;
 665                                 if ( preg_match( $regex, $text, $m ) ) {
 666                                         return $this->parseMatch( $m );
 667                                 }
 668                         }
 669                 }
 670                 return array( false, false );
 671         }
 672
 673         /**
 674          * Match some text, without parameter capture
 675          * Returns the magic word name, or false if there was no capture
 676          */
 677         public function matchStartToEnd( $text ) {
 678                 $hash = $this->getHash();
 679                 if ( isset( $hash[1][$text] ) ) {
 680                         return $hash[1][$text];
 681                 }
 682                 global $wgContLang;
 683                 $lc = $wgContLang->lc( $text );
 684                 if ( isset( $hash[0][$lc] ) ) {
 685                         return $hash[0][$lc];
 686                 }
 687                 return false;
 688         }
 689
 690         /**
 691          * Returns an associative array, ID => param value, for all items that match
 692          * Removes the matched items from the input string (passed by reference)
 693          */
 694         public function matchAndRemove( &$text ) {
 695                 $found = array();
 696                 $regexes = $this->getRegex();
 697                 foreach ( $regexes as $regex ) {
 698                         if ( $regex === '' ) {
 699                                 continue;
 700                         }
 701                         preg_match_all( $regex, $text, $matches, PREG_SET_ORDER );
 702                         foreach ( $matches as $m ) {
 703                                 list( $name, $param ) = $this->parseMatch( $m );
 704                                 $found[$name] = $param;
 705                         }
 706                         $text = preg_replace( $regex, '', $text );
 707                 }
 708                 return $found;
 709         }
 710
 711         /**
 712          * Return the ID of the magic word at the start of $text, and remove
 713          * the prefix from $text.
 714          * Return false if no match found and $text is not modified.
 715          * Does not match parameters.
 716          */
 717         public function matchStartAndRemove( &$text ) {
 718                 $regexes = $this->getRegexStart();
 719                 foreach ( $regexes as $regex ) {
 720                         if ( $regex === '' ) {
 721                                 continue;
 722                         }
 723                         if ( preg_match( $regex, $text, $m ) ) {
 724                                 list( $id, $param ) = $this->parseMatch( $m );
 725                                 if ( strlen( $m[0] ) >= strlen( $text ) ) {
 726                                         $text = '';
 727                                 } else {
 728                                         $text = substr( $text, strlen( $m[0] ) );
 729                                 }
 730                                 return $id;
 731                         }
 732                 }
 733                 return false;
 734         }
 735 }