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                 'articlepath',
  64                 'sitename',
  65                 'server',
  66                 'servername',
  67                 'scriptpath',
  68                 'stylepath',
  69                 'pagename',
  70                 'pagenamee',
  71                 'fullpagename',
  72                 'fullpagenamee',
  73                 'namespace',
  74                 'namespacee',
  75                 'currentweek',
  76                 'currentdow',
  77                 'localweek',
  78                 'localdow',
  79                 'revisionid',
  80                 'revisionday',
  81                 'revisionday2',
  82                 'revisionmonth',
  83                 'revisionmonth1',
  84                 'revisionyear',
  85                 'revisiontimestamp',
  86                 'revisionuser',
  87                 'subpagename',
  88                 'subpagenamee',
  89                 'talkspace',
  90                 'talkspacee',
  91                 'subjectspace',
  92                 'subjectspacee',
  93                 'talkpagename',
  94                 'talkpagenamee',
  95                 'subjectpagename',
  96                 'subjectpagenamee',
  97                 'numberofusers',
  98                 'numberofactiveusers',
  99                 'numberofpages',
 100                 'currentversion',
 101                 'basepagename',
 102                 'basepagenamee',
 103                 'currenttimestamp',
 104                 'localtimestamp',
 105                 'directionmark',
 106                 'contentlanguage',
 107                 'numberofadmins',
 108                 'numberofviews',
 109         );
 110
 111         /* Array of caching hints for ParserCache */
 112         static public $mCacheTTLs = array (
 113                 'currentmonth' => 86400,
 114                 'currentmonth1' => 86400,
 115                 'currentmonthname' => 86400,
 116                 'currentmonthnamegen' => 86400,
 117                 'currentmonthabbrev' => 86400,
 118                 'currentday' => 3600,
 119                 'currentday2' => 3600,
 120                 'currentdayname' => 3600,
 121                 'currentyear' => 86400,
 122                 'currenttime' => 3600,
 123                 'currenthour' => 3600,
 124                 'localmonth' => 86400,
 125                 'localmonth1' => 86400,
 126                 'localmonthname' => 86400,
 127                 'localmonthnamegen' => 86400,
 128                 'localmonthabbrev' => 86400,
 129                 'localday' => 3600,
 130                 'localday2' => 3600,
 131                 'localdayname' => 3600,
 132                 'localyear' => 86400,
 133                 'localtime' => 3600,
 134                 'localhour' => 3600,
 135                 'numberofarticles' => 3600,
 136                 'numberoffiles' => 3600,
 137                 'numberofedits' => 3600,
 138                 'currentweek' => 3600,
 139                 'currentdow' => 3600,
 140                 'localweek' => 3600,
 141                 'localdow' => 3600,
 142                 'numberofusers' => 3600,
 143                 'numberofactiveusers' => 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                 'nonewsectionlink',
 162                 'hiddencat',
 163                 'index',
 164                 'noindex',
 165                 'staticredirect',
 166                 'notitleconvert',
 167                 'nocontentconvert',
 168         );
 169
 170         static public $mSubstIDs = array(
 171                 'subst',
 172                 'safesubst',
 173         );
 174
 175         static public $mObjects = array();
 176         static public $mDoubleUnderscoreArray = null;
 177
 178         /**#@-*/
 179
 180         function __construct($id = 0, $syn = '', $cs = false) {
 181                 $this->mId = $id;
 182                 $this->mSynonyms = (array)$syn;
 183                 $this->mCaseSensitive = $cs;
 184                 $this->mRegex = '';
 185                 $this->mRegexStart = '';
 186                 $this->mVariableRegex = '';
 187                 $this->mVariableStartToEndRegex = '';
 188                 $this->mModified = false;
 189         }
 190
 191         /**
 192          * Factory: creates an object representing an ID
 193          * @static
 194          */
 195         static function &get( $id ) {
 196                 wfProfileIn( __METHOD__ );
 197                 if ( !isset( self::$mObjects[$id] ) ) {
 198                         $mw = new MagicWord();
 199                         $mw->load( $id );
 200                         self::$mObjects[$id] = $mw;
 201                 }
 202                 wfProfileOut( __METHOD__ );
 203                 return self::$mObjects[$id];
 204         }
 205
 206         /**
 207          * Get an array of parser variable IDs
 208          */
 209         static function getVariableIDs() {
 210                 if ( !self::$mVariableIDsInitialised ) {
 211                         # Deprecated constant definition hook, available for extensions that need it
 212                         $magicWords = array();
 213                         wfRunHooks( 'MagicWordMagicWords', array( &$magicWords ) );
 214                         foreach ( $magicWords as $word ) {
 215                                 define( $word, $word );
 216                         }
 217
 218                         # Get variable IDs
 219                         wfRunHooks( 'MagicWordwgVariableIDs', array( &self::$mVariableIDs ) );
 220                         self::$mVariableIDsInitialised = true;
 221                 }
 222                 return self::$mVariableIDs;
 223         }
 224
 225         /**
 226          * Get an array of parser substitution modifier IDs
 227          */
 228         static function getSubstIDs() {
 229                 return self::$mSubstIDs;
 230         }
 231
 232         /* Allow external reads of TTL array */
 233         static function getCacheTTL($id) {
 234                 if (array_key_exists($id,self::$mCacheTTLs)) {
 235                         return self::$mCacheTTLs[$id];
 236                 } else {
 237                         return -1;
 238                 }
 239         }
 240
 241         /** Get a MagicWordArray of double-underscore entities */
 242         static function getDoubleUnderscoreArray() {
 243                 if ( is_null( self::$mDoubleUnderscoreArray ) ) {
 244                         self::$mDoubleUnderscoreArray = new MagicWordArray( self::$mDoubleUnderscoreIDs );
 245                 }
 246                 return self::$mDoubleUnderscoreArray;
 247         }
 248
 249         /**
 250          * Clear the self::$mObjects variable
 251          * For use in parser tests
 252          */
 253         public static function clearCache() {
 254                 self::$mObjects = array();
 255         }
 256
 257         # Initialises this object with an ID
 258         function load( $id ) {
 259                 global $wgContLang;
 260                 $this->mId = $id;
 261                 $wgContLang->getMagic( $this );
 262                 if ( !$this->mSynonyms ) {
 263                         $this->mSynonyms = array( 'dkjsagfjsgashfajsh' );
 264                         #throw new MWException( "Error: invalid magic word '$id'" );
 265                         wfDebugLog( 'exception', "Error: invalid magic word '$id'\n" );
 266                 }
 267         }
 268
 269         /**
 270          * Preliminary initialisation
 271          * @private
 272          */
 273         function initRegex() {
 274                 #$variableClass = Title::legalChars();
 275                 # This was used for matching "$1" variables, but different uses of the feature will have
 276                 # different restrictions, which should be checked *after* the MagicWord has been matched,
 277                 # not here. - IMSoP
 278
 279                 $escSyn = array();
 280                 foreach ( $this->mSynonyms as $synonym )
 281                         // In case a magic word contains /, like that's going to happen;)
 282                         $escSyn[] = preg_quote( $synonym, '/' );
 283                 $this->mBaseRegex = implode( '|', $escSyn );
 284
 285                 $case = $this->mCaseSensitive ? '' : 'iu';
 286                 $this->mRegex = "/{$this->mBaseRegex}/{$case}";
 287                 $this->mRegexStart = "/^(?:{$this->mBaseRegex})/{$case}";
 288                 $this->mVariableRegex = str_replace( "\\$1", "(.*?)", $this->mRegex );
 289                 $this->mVariableStartToEndRegex = str_replace( "\\$1", "(.*?)",
 290                         "/^(?:{$this->mBaseRegex})$/{$case}" );
 291         }
 292
 293         /**
 294          * Gets a regex representing matching the word
 295          */
 296         function getRegex() {
 297                 if ($this->mRegex == '' ) {
 298                         $this->initRegex();
 299                 }
 300                 return $this->mRegex;
 301         }
 302
 303         /**
 304          * Gets the regexp case modifier to use, i.e. i or nothing, to be used if
 305          * one is using MagicWord::getBaseRegex(), otherwise it'll be included in
 306          * the complete expression
 307          */
 308         function getRegexCase() {
 309                 if ( $this->mRegex === '' )
 310                         $this->initRegex();
 311
 312                 return $this->mCaseSensitive ? '' : 'iu';
 313         }
 314
 315         /**
 316          * Gets a regex matching the word, if it is at the string start
 317          */
 318         function getRegexStart() {
 319                 if ($this->mRegex == '' ) {
 320                         $this->initRegex();
 321                 }
 322                 return $this->mRegexStart;
 323         }
 324
 325         /**
 326          * regex without the slashes and what not
 327          */
 328         function getBaseRegex() {
 329                 if ($this->mRegex == '') {
 330                         $this->initRegex();
 331                 }
 332                 return $this->mBaseRegex;
 333         }
 334
 335         /**
 336          * Returns true if the text contains the word
 337          * @return bool
 338          */
 339         function match( $text ) {
 340                 return (bool)preg_match( $this->getRegex(), $text );
 341         }
 342
 343         /**
 344          * Returns true if the text starts with the word
 345          * @return bool
 346          */
 347         function matchStart( $text ) {
 348                 return (bool)preg_match( $this->getRegexStart(), $text );
 349         }
 350
 351         /**
 352          * Returns NULL if there's no match, the value of $1 otherwise
 353          * The return code is the matched string, if there's no variable
 354          * part in the regex and the matched variable part ($1) if there
 355          * is one.
 356          */
 357         function matchVariableStartToEnd( $text ) {
 358                 $matches = array();
 359                 $matchcount = preg_match( $this->getVariableStartToEndRegex(), $text, $matches );
 360                 if ( $matchcount == 0 ) {
 361                         return null;
 362                 } else {
 363                         # multiple matched parts (variable match); some will be empty because of
 364                         # synonyms. The variable will be the second non-empty one so remove any
 365                         # blank elements and re-sort the indices.
 366                         # See also bug 6526
 367
 368                         $matches = array_values(array_filter($matches));
 369
 370                         if ( count($matches) == 1 ) { return $matches[0]; }
 371                         else { return $matches[1]; }
 372                 }
 373         }
 374
 375
 376         /**
 377          * Returns true if the text matches the word, and alters the
 378          * input string, removing all instances of the word
 379          */
 380         function matchAndRemove( &$text ) {
 381                 $this->mFound = false;
 382                 $text = preg_replace_callback( $this->getRegex(), array( &$this, 'pregRemoveAndRecord' ), $text );
 383                 return $this->mFound;
 384         }
 385
 386         function matchStartAndRemove( &$text ) {
 387                 $this->mFound = false;
 388                 $text = preg_replace_callback( $this->getRegexStart(), array( &$this, 'pregRemoveAndRecord' ), $text );
 389                 return $this->mFound;
 390         }
 391
 392         /**
 393          * Used in matchAndRemove()
 394          * @private
 395          **/
 396         function pregRemoveAndRecord( ) {
 397                 $this->mFound = true;
 398                 return '';
 399         }
 400
 401         /**
 402          * Replaces the word with something else
 403          */
 404         function replace( $replacement, $subject, $limit=-1 ) {
 405                 $res = preg_replace( $this->getRegex(), StringUtils::escapeRegexReplacement( $replacement ), $subject, $limit );
 406                 $this->mModified = !($res === $subject);
 407                 return $res;
 408         }
 409
 410         /**
 411          * Variable handling: {{SUBST:xxx}} style words
 412          * Calls back a function to determine what to replace xxx with
 413          * Input word must contain $1
 414          */
 415         function substituteCallback( $text, $callback ) {
 416                 $res = preg_replace_callback( $this->getVariableRegex(), $callback, $text );
 417                 $this->mModified = !($res === $text);
 418                 return $res;
 419         }
 420
 421         /**
 422          * Matches the word, where $1 is a wildcard
 423          */
 424         function getVariableRegex()     {
 425                 if ( $this->mVariableRegex == '' ) {
 426                         $this->initRegex();
 427                 }
 428                 return $this->mVariableRegex;
 429         }
 430
 431         /**
 432          * Matches the entire string, where $1 is a wildcard
 433          */
 434         function getVariableStartToEndRegex() {
 435                 if ( $this->mVariableStartToEndRegex == '' ) {
 436                         $this->initRegex();
 437                 }
 438                 return $this->mVariableStartToEndRegex;
 439         }
 440
 441         /**
 442          * Accesses the synonym list directly
 443          */
 444         function getSynonym( $i ) {
 445                 return $this->mSynonyms[$i];
 446         }
 447
 448         function getSynonyms() {
 449                 return $this->mSynonyms;
 450         }
 451
 452         /**
 453          * Returns true if the last call to replace() or substituteCallback()
 454          * returned a modified text, otherwise false.
 455          */
 456         function getWasModified(){
 457                 return $this->mModified;
 458         }
 459
 460         /**
 461          * $magicarr is an associative array of (magic word ID => replacement)
 462          * This method uses the php feature to do several replacements at the same time,
 463          * thereby gaining some efficiency. The result is placed in the out variable
 464          * $result. The return value is true if something was replaced.
 465          * @static
 466          **/
 467         function replaceMultiple( $magicarr, $subject, &$result ){
 468                 $search = array();
 469                 $replace = array();
 470                 foreach( $magicarr as $id => $replacement ){
 471                         $mw = MagicWord::get( $id );
 472                         $search[] = $mw->getRegex();
 473                         $replace[] = $replacement;
 474                 }
 475
 476                 $result = preg_replace( $search, $replace, $subject );
 477                 return !($result === $subject);
 478         }
 479
 480         /**
 481          * Adds all the synonyms of this MagicWord to an array, to allow quick
 482          * lookup in a list of magic words
 483          */
 484         function addToArray( &$array, $value ) {
 485                 global $wgContLang;
 486                 foreach ( $this->mSynonyms as $syn ) {
 487                         $array[$wgContLang->lc($syn)] = $value;
 488                 }
 489         }
 490
 491         function isCaseSensitive() {
 492                 return $this->mCaseSensitive;
 493         }
 494
 495         function getId() {
 496                 return $this->mId;
 497         }
 498 }
 499
 500 /**
 501  * Class for handling an array of magic words
 502  * @ingroup Parser
 503  */
 504 class MagicWordArray {
 505         var $names = array();
 506         var $hash;
 507         var $baseRegex, $regex;
 508         var $matches;
 509
 510         function __construct( $names = array() ) {
 511                 $this->names = $names;
 512         }
 513
 514         /**
 515          * Add a magic word by name
 516          */
 517         public function add( $name ) {
 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                 $regexes = $this->getVariableStartToEndRegex();
 661                 foreach ( $regexes as $regex ) {
 662                         if ( $regex !== '' ) {
 663                                 $m = false;
 664                                 if ( preg_match( $regex, $text, $m ) ) {
 665                                         return $this->parseMatch( $m );
 666                                 }
 667                         }
 668                 }
 669                 return array( false, false );
 670         }
 671
 672         /**
 673          * Match some text, without parameter capture
 674          * Returns the magic word name, or false if there was no capture
 675          */
 676         public function matchStartToEnd( $text ) {
 677                 $hash = $this->getHash();
 678                 if ( isset( $hash[1][$text] ) ) {
 679                         return $hash[1][$text];
 680                 }
 681                 global $wgContLang;
 682                 $lc = $wgContLang->lc( $text );
 683                 if ( isset( $hash[0][$lc] ) ) {
 684                         return $hash[0][$lc];
 685                 }
 686                 return false;
 687         }
 688
 689         /**
 690          * Returns an associative array, ID => param value, for all items that match
 691          * Removes the matched items from the input string (passed by reference)
 692          */
 693         public function matchAndRemove( &$text ) {
 694                 $found = array();
 695                 $regexes = $this->getRegex();
 696                 foreach ( $regexes as $regex ) {
 697                         if ( $regex === '' ) {
 698                                 continue;
 699                         }
 700                         preg_match_all( $regex, $text, $matches, PREG_SET_ORDER );
 701                         foreach ( $matches as $m ) {
 702                                 list( $name, $param ) = $this->parseMatch( $m );
 703                                 $found[$name] = $param;
 704                         }
 705                         $text = preg_replace( $regex, '', $text );
 706                 }
 707                 return $found;
 708         }
 709
 710         /**
 711          * Return the ID of the magic word at the start of $text, and remove
 712          * the prefix from $text.
 713          * Return false if no match found and $text is not modified.
 714          * Does not match parameters.
 715          */
 716         public function matchStartAndRemove( &$text ) {
 717                 $regexes = $this->getRegexStart();
 718                 foreach ( $regexes as $regex ) {
 719                         if ( $regex === '' ) {
 720                                 continue;
 721                         }
 722                         if ( preg_match( $regex, $text, $m ) ) {
 723                                 list( $id, $param ) = $this->parseMatch( $m );
 724                                 if ( strlen( $m[0] ) >= strlen( $text ) ) {
 725                                         $text = '';
 726                                 } else {
 727                                         $text = substr( $text, strlen( $m[0] ) );
 728                                 }
 729                                 return $id;
 730                         }
 731                 }
 732                 return false;
 733         }
 734 }