includes/MagicWord.php

   1 <?php
   2 /**
   3  * File for magic words
   4  * See docs/magicword.txt
   5  *
   6  * @file
   7  * @ingroup Parser
   8  */
   9
  10 /**
  11  * This class encapsulates "magic words" such as #redirect, __NOTOC__, etc.
  12  * Usage:
  13  *     if (MagicWord::get( 'redirect' )->match( $text ) )
  14  *
  15  * Possible future improvements:
  16  *   * Simultaneous searching for a number of magic words
  17  *   * MagicWord::$mObjects in shared memory
  18  *
  19  * Please avoid reading the data out of one of these objects and then writing
  20  * special case code. If possible, add another match()-like function here.
  21  *
  22  * To add magic words in an extension, use the LanguageGetMagic hook. For
  23  * magic words which are also Parser variables, add a MagicWordwgVariableIDs
  24  * hook. Use string keys.
  25  *
  26  * @ingroup Parser
  27  */
  28 class MagicWord {
  29         /**#@+
  30          * @private
  31          */
  32         var $mId, $mSynonyms, $mCaseSensitive, $mRegex;
  33         var $mRegexStart, $mBaseRegex, $mVariableRegex;
  34         var $mModified, $mFound;
  35
  36         static public $mVariableIDsInitialised = false;
  37         static public $mVariableIDs = array(
  38                 'currentmonth',
  39                 'currentmonthname',
  40                 'currentmonthnamegen',
  41                 'currentmonthabbrev',
  42                 'currentday',
  43                 'currentday2',
  44                 'currentdayname',
  45                 'currentyear',
  46                 'currenttime',
  47                 'currenthour',
  48                 'localmonth',
  49                 'localmonthname',
  50                 'localmonthnamegen',
  51                 'localmonthabbrev',
  52                 'localday',
  53                 'localday2',
  54                 'localdayname',
  55                 'localyear',
  56                 'localtime',
  57                 'localhour',
  58                 'numberofarticles',
  59                 'numberoffiles',
  60                 'numberofedits',
  61                 'sitename',
  62                 'server',
  63                 'servername',
  64                 'scriptpath',
  65                 'pagename',
  66                 'pagenamee',
  67                 'fullpagename',
  68                 'fullpagenamee',
  69                 'namespace',
  70                 'namespacee',
  71                 'currentweek',
  72                 'currentdow',
  73                 'localweek',
  74                 'localdow',
  75                 'revisionid',
  76                 'revisionday',
  77                 'revisionday2',
  78                 'revisionmonth',
  79                 'revisionyear',
  80                 'revisiontimestamp',
  81                 'subpagename',
  82                 'subpagenamee',
  83                 'displaytitle',
  84                 'talkspace',
  85                 'talkspacee',
  86                 'subjectspace',
  87                 'subjectspacee',
  88                 'talkpagename',
  89                 'talkpagenamee',
  90                 'subjectpagename',
  91                 'subjectpagenamee',
  92                 'numberofusers',
  93                 'newsectionlink',
  94                 'numberofpages',
  95                 'currentversion',
  96                 'basepagename',
  97                 'basepagenamee',
  98                 'urlencode',
  99                 'currenttimestamp',
 100                 'localtimestamp',
 101                 'directionmark',
 102                 'language',
 103                 'contentlanguage',
 104                 'pagesinnamespace',
 105                 'numberofadmins',
 106                 'defaultsort',
 107                 'pagesincategory',
 108         );
 109
 110         /* Array of caching hints for ParserCache */
 111         static public $mCacheTTLs = array (
 112                 'currentmonth' => 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                 'localmonthname' => 86400,
 124                 'localmonthnamegen' => 86400,
 125                 'localmonthabbrev' => 86400,
 126                 'localday' => 3600,
 127                 'localday2' => 3600,
 128                 'localdayname' => 3600,
 129                 'localyear' => 86400,
 130                 'localtime' => 3600,
 131                 'localhour' => 3600,
 132                 'numberofarticles' => 3600,
 133                 'numberoffiles' => 3600,
 134                 'numberofedits' => 3600,
 135                 'currentweek' => 3600,
 136                 'currentdow' => 3600,
 137                 'localweek' => 3600,
 138                 'localdow' => 3600,
 139                 'numberofusers' => 3600,
 140                 'numberofpages' => 3600,
 141                 'currentversion' => 86400,
 142                 'currenttimestamp' => 3600,
 143                 'localtimestamp' => 3600,
 144                 'pagesinnamespace' => 3600,
 145                 'numberofadmins' => 3600,
 146                 );
 147
 148         static public $mDoubleUnderscoreIDs = array(
 149                 'notoc',
 150                 'nogallery',
 151                 'forcetoc',
 152                 'toc',
 153                 'noeditsection',
 154                 'newsectionlink',
 155                 'hiddencat',
 156         );
 157
 158
 159         static public $mObjects = array();
 160         static public $mDoubleUnderscoreArray = null;
 161
 162         /**#@-*/
 163
 164         function __construct($id = 0, $syn = '', $cs = false) {
 165                 $this->mId = $id;
 166                 $this->mSynonyms = (array)$syn;
 167                 $this->mCaseSensitive = $cs;
 168                 $this->mRegex = '';
 169                 $this->mRegexStart = '';
 170                 $this->mVariableRegex = '';
 171                 $this->mVariableStartToEndRegex = '';
 172                 $this->mModified = false;
 173         }
 174
 175         /**
 176          * Factory: creates an object representing an ID
 177          * @static
 178          */
 179         static function &get( $id ) {
 180                 wfProfileIn( __METHOD__ );
 181                 if (!array_key_exists( $id, self::$mObjects ) ) {
 182                         $mw = new MagicWord();
 183                         $mw->load( $id );
 184                         self::$mObjects[$id] = $mw;
 185                 }
 186                 wfProfileOut( __METHOD__ );
 187                 return self::$mObjects[$id];
 188         }
 189
 190         /**
 191          * Get an array of parser variable IDs
 192          */
 193         static function getVariableIDs() {
 194                 if ( !self::$mVariableIDsInitialised ) {
 195                         # Deprecated constant definition hook, available for extensions that need it
 196                         $magicWords = array();
 197                         wfRunHooks( 'MagicWordMagicWords', array( &$magicWords ) );
 198                         foreach ( $magicWords as $word ) {
 199                                 define( $word, $word );
 200                         }
 201
 202                         # Get variable IDs
 203                         wfRunHooks( 'MagicWordwgVariableIDs', array( &self::$mVariableIDs ) );
 204                         self::$mVariableIDsInitialised = true;
 205                 }
 206                 return self::$mVariableIDs;
 207         }
 208
 209         /* Allow external reads of TTL array */
 210         static function getCacheTTL($id) {
 211                 if (array_key_exists($id,self::$mCacheTTLs)) {
 212                         return self::$mCacheTTLs[$id];
 213                 } else {
 214                         return -1;
 215                 }
 216         }
 217
 218         /** Get a MagicWordArray of double-underscore entities */
 219         static function getDoubleUnderscoreArray() {
 220                 if ( is_null( self::$mDoubleUnderscoreArray ) ) {
 221                         self::$mDoubleUnderscoreArray = new MagicWordArray( self::$mDoubleUnderscoreIDs );
 222                 }
 223                 return self::$mDoubleUnderscoreArray;
 224         }
 225
 226         # Initialises this object with an ID
 227         function load( $id ) {
 228                 global $wgContLang;
 229                 $this->mId = $id;
 230                 $wgContLang->getMagic( $this );
 231                 if ( !$this->mSynonyms ) {
 232                         $this->mSynonyms = array( 'dkjsagfjsgashfajsh' );
 233                         #throw new MWException( "Error: invalid magic word '$id'" );
 234                         wfDebugLog( 'exception', "Error: invalid magic word '$id'\n" );
 235                 }
 236         }
 237
 238         /**
 239          * Preliminary initialisation
 240          * @private
 241          */
 242         function initRegex() {
 243                 #$variableClass = Title::legalChars();
 244                 # This was used for matching "$1" variables, but different uses of the feature will have
 245                 # different restrictions, which should be checked *after* the MagicWord has been matched,
 246                 # not here. - IMSoP
 247
 248                 $escSyn = array();
 249                 foreach ( $this->mSynonyms as $synonym )
 250                         // In case a magic word contains /, like that's going to happen;)
 251                         $escSyn[] = preg_quote( $synonym, '/' );
 252                 $this->mBaseRegex = implode( '|', $escSyn );
 253
 254                 $case = $this->mCaseSensitive ? '' : 'iu';
 255                 $this->mRegex = "/{$this->mBaseRegex}/{$case}";
 256                 $this->mRegexStart = "/^(?:{$this->mBaseRegex})/{$case}";
 257                 $this->mVariableRegex = str_replace( "\\$1", "(.*?)", $this->mRegex );
 258                 $this->mVariableStartToEndRegex = str_replace( "\\$1", "(.*?)",
 259                         "/^(?:{$this->mBaseRegex})$/{$case}" );
 260         }
 261
 262         /**
 263          * Gets a regex representing matching the word
 264          */
 265         function getRegex() {
 266                 if ($this->mRegex == '' ) {
 267                         $this->initRegex();
 268                 }
 269                 return $this->mRegex;
 270         }
 271
 272         /**
 273          * Gets the regexp case modifier to use, i.e. i or nothing, to be used if
 274          * one is using MagicWord::getBaseRegex(), otherwise it'll be included in
 275          * the complete expression
 276          */
 277         function getRegexCase() {
 278                 if ( $this->mRegex === '' )
 279                         $this->initRegex();
 280
 281                 return $this->mCaseSensitive ? '' : 'iu';
 282         }
 283
 284         /**
 285          * Gets a regex matching the word, if it is at the string start
 286          */
 287         function getRegexStart() {
 288                 if ($this->mRegex == '' ) {
 289                         $this->initRegex();
 290                 }
 291                 return $this->mRegexStart;
 292         }
 293
 294         /**
 295          * regex without the slashes and what not
 296          */
 297         function getBaseRegex() {
 298                 if ($this->mRegex == '') {
 299                         $this->initRegex();
 300                 }
 301                 return $this->mBaseRegex;
 302         }
 303
 304         /**
 305          * Returns true if the text contains the word
 306          * @return bool
 307          */
 308         function match( $text ) {
 309                 return preg_match( $this->getRegex(), $text );
 310         }
 311
 312         /**
 313          * Returns true if the text starts with the word
 314          * @return bool
 315          */
 316         function matchStart( $text ) {
 317                 return preg_match( $this->getRegexStart(), $text );
 318         }
 319
 320         /**
 321          * Returns NULL if there's no match, the value of $1 otherwise
 322          * The return code is the matched string, if there's no variable
 323          * part in the regex and the matched variable part ($1) if there
 324          * is one.
 325          */
 326         function matchVariableStartToEnd( $text ) {
 327                 $matches = array();
 328                 $matchcount = preg_match( $this->getVariableStartToEndRegex(), $text, $matches );
 329                 if ( $matchcount == 0 ) {
 330                         return NULL;
 331                 } else {
 332                         # multiple matched parts (variable match); some will be empty because of
 333                         # synonyms. The variable will be the second non-empty one so remove any
 334                         # blank elements and re-sort the indices.
 335                         # See also bug 6526
 336
 337                         $matches = array_values(array_filter($matches));
 338
 339                         if ( count($matches) == 1 ) { return $matches[0]; }
 340                         else { return $matches[1]; }
 341                 }
 342         }
 343
 344
 345         /**
 346          * Returns true if the text matches the word, and alters the
 347          * input string, removing all instances of the word
 348          */
 349         function matchAndRemove( &$text ) {
 350                 $this->mFound = false;
 351                 $text = preg_replace_callback( $this->getRegex(), array( &$this, 'pregRemoveAndRecord' ), $text );
 352                 return $this->mFound;
 353         }
 354
 355         function matchStartAndRemove( &$text ) {
 356                 $this->mFound = false;
 357                 $text = preg_replace_callback( $this->getRegexStart(), array( &$this, 'pregRemoveAndRecord' ), $text );
 358                 return $this->mFound;
 359         }
 360
 361         /**
 362          * Used in matchAndRemove()
 363          * @private
 364          **/
 365         function pregRemoveAndRecord( ) {
 366                 $this->mFound = true;
 367                 return '';
 368         }
 369
 370         /**
 371          * Replaces the word with something else
 372          */
 373         function replace( $replacement, $subject, $limit=-1 ) {
 374                 $res = preg_replace( $this->getRegex(), StringUtils::escapeRegexReplacement( $replacement ), $subject, $limit );
 375                 $this->mModified = !($res === $subject);
 376                 return $res;
 377         }
 378
 379         /**
 380          * Variable handling: {{SUBST:xxx}} style words
 381          * Calls back a function to determine what to replace xxx with
 382          * Input word must contain $1
 383          */
 384         function substituteCallback( $text, $callback ) {
 385                 $res = preg_replace_callback( $this->getVariableRegex(), $callback, $text );
 386                 $this->mModified = !($res === $text);
 387                 return $res;
 388         }
 389
 390         /**
 391          * Matches the word, where $1 is a wildcard
 392          */
 393         function getVariableRegex()     {
 394                 if ( $this->mVariableRegex == '' ) {
 395                         $this->initRegex();
 396                 }
 397                 return $this->mVariableRegex;
 398         }
 399
 400         /**
 401          * Matches the entire string, where $1 is a wildcard
 402          */
 403         function getVariableStartToEndRegex() {
 404                 if ( $this->mVariableStartToEndRegex == '' ) {
 405                         $this->initRegex();
 406                 }
 407                 return $this->mVariableStartToEndRegex;
 408         }
 409
 410         /**
 411          * Accesses the synonym list directly
 412          */
 413         function getSynonym( $i ) {
 414                 return $this->mSynonyms[$i];
 415         }
 416
 417         function getSynonyms() {
 418                 return $this->mSynonyms;
 419         }
 420
 421         /**
 422          * Returns true if the last call to replace() or substituteCallback()
 423          * returned a modified text, otherwise false.
 424          */
 425         function getWasModified(){
 426                 return $this->mModified;
 427         }
 428
 429         /**
 430          * $magicarr is an associative array of (magic word ID => replacement)
 431          * This method uses the php feature to do several replacements at the same time,
 432          * thereby gaining some efficiency. The result is placed in the out variable
 433          * $result. The return value is true if something was replaced.
 434          * @static
 435          **/
 436         function replaceMultiple( $magicarr, $subject, &$result ){
 437                 $search = array();
 438                 $replace = array();
 439                 foreach( $magicarr as $id => $replacement ){
 440                         $mw = MagicWord::get( $id );
 441                         $search[] = $mw->getRegex();
 442                         $replace[] = $replacement;
 443                 }
 444
 445                 $result = preg_replace( $search, $replace, $subject );
 446                 return !($result === $subject);
 447         }
 448
 449         /**
 450          * Adds all the synonyms of this MagicWord to an array, to allow quick
 451          * lookup in a list of magic words
 452          */
 453         function addToArray( &$array, $value ) {
 454                 global $wgContLang;
 455                 foreach ( $this->mSynonyms as $syn ) {
 456                         $array[$wgContLang->lc($syn)] = $value;
 457                 }
 458         }
 459
 460         function isCaseSensitive() {
 461                 return $this->mCaseSensitive;
 462         }
 463
 464         function getId() {
 465                 return $this->mId;
 466         }
 467 }
 468
 469 /**
 470  * Class for handling an array of magic words
 471  * @ingroup Parser
 472  */
 473 class MagicWordArray {
 474         var $names = array();
 475         var $hash;
 476         var $baseRegex, $regex;
 477         var $matches;
 478
 479         function __construct( $names = array() ) {
 480                 $this->names = $names;
 481         }
 482
 483         /**
 484          * Add a magic word by name
 485          */
 486         public function add( $name ) {
 487                 global $wgContLang;
 488                 $this->names[] = $name;
 489                 $this->hash = $this->baseRegex = $this->regex = null;
 490         }
 491
 492         /**
 493          * Add a number of magic words by name
 494          */
 495         public function addArray( $names ) {
 496                 $this->names = array_merge( $this->names, array_values( $names ) );
 497                 $this->hash = $this->baseRegex = $this->regex = null;
 498         }
 499
 500         /**
 501          * Get a 2-d hashtable for this array
 502          */
 503         function getHash() {
 504                 if ( is_null( $this->hash ) ) {
 505                         global $wgContLang;
 506                         $this->hash = array( 0 => array(), 1 => array() );
 507                         foreach ( $this->names as $name ) {
 508                                 $magic = MagicWord::get( $name );
 509                                 $case = intval( $magic->isCaseSensitive() );
 510                                 foreach ( $magic->getSynonyms() as $syn ) {
 511                                         if ( !$case ) {
 512                                                 $syn = $wgContLang->lc( $syn );
 513                                         }
 514                                         $this->hash[$case][$syn] = $name;
 515                                 }
 516                         }
 517                 }
 518                 return $this->hash;
 519         }
 520
 521         /**
 522          * Get the base regex
 523          */
 524         function getBaseRegex() {
 525                 if ( is_null( $this->baseRegex ) ) {
 526                         $this->baseRegex = array( 0 => '', 1 => '' );
 527                         foreach ( $this->names as $name ) {
 528                                 $magic = MagicWord::get( $name );
 529                                 $case = intval( $magic->isCaseSensitive() );
 530                                 foreach ( $magic->getSynonyms() as $i => $syn ) {
 531                                         $group = "(?P<{$i}_{$name}>" . preg_quote( $syn, '/' ) . ')';
 532                                         if ( $this->baseRegex[$case] === '' ) {
 533                                                 $this->baseRegex[$case] = $group;
 534                                         } else {
 535                                                 $this->baseRegex[$case] .= '|' . $group;
 536                                         }
 537                                 }
 538                         }
 539                 }
 540                 return $this->baseRegex;
 541         }
 542
 543         /**
 544          * Get an unanchored regex
 545          */
 546         function getRegex() {
 547                 if ( is_null( $this->regex ) ) {
 548                         $base = $this->getBaseRegex();
 549                         $this->regex = array( '', '' );
 550                         if ( $this->baseRegex[0] !== '' ) {
 551                                 $this->regex[0] = "/{$base[0]}/iuS";
 552                         }
 553                         if ( $this->baseRegex[1] !== '' ) {
 554                                 $this->regex[1] = "/{$base[1]}/S";
 555                         }
 556                 }
 557                 return $this->regex;
 558         }
 559
 560         /**
 561          * Get a regex for matching variables
 562          */
 563         function getVariableRegex() {
 564                 return str_replace( "\\$1", "(.*?)", $this->getRegex() );
 565         }
 566
 567         /**
 568          * Get an anchored regex for matching variables
 569          */
 570         function getVariableStartToEndRegex() {
 571                 $base = $this->getBaseRegex();
 572                 $newRegex = array( '', '' );
 573                 if ( $base[0] !== '' ) {
 574                         $newRegex[0] = str_replace( "\\$1", "(.*?)", "/^(?:{$base[0]})$/iuS" );
 575                 }
 576                 if ( $base[1] !== '' ) {
 577                         $newRegex[1] = str_replace( "\\$1", "(.*?)", "/^(?:{$base[1]})$/S" );
 578                 }
 579                 return $newRegex;
 580         }
 581
 582         /**
 583          * Parse a match array from preg_match
 584          * Returns array(magic word ID, parameter value)
 585          * If there is no parameter value, that element will be false.
 586          */
 587         function parseMatch( $m ) {
 588                 reset( $m );
 589                 while ( list( $key, $value ) = each( $m ) ) {
 590                         if ( $key === 0 || $value === '' ) {
 591                                 continue;
 592                         }
 593                         $parts = explode( '_', $key, 2 );
 594                         if ( count( $parts ) != 2 ) {
 595                                 // This shouldn't happen
 596                                 // continue;
 597                                 throw new MWException( __METHOD__ . ': bad parameter name' );
 598                         }
 599                         list( /* $synIndex */, $magicName ) = $parts;
 600                         $paramValue = next( $m );
 601                         return array( $magicName, $paramValue );
 602                 }
 603                 // This shouldn't happen either
 604                 throw new MWException( __METHOD__.': parameter not found' );
 605                 return array( false, false );
 606         }
 607
 608         /**
 609          * Match some text, with parameter capture
 610          * Returns an array with the magic word name in the first element and the
 611          * parameter in the second element.
 612          * Both elements are false if there was no match.
 613          */
 614         public function matchVariableStartToEnd( $text ) {
 615                 global $wgContLang;
 616                 $regexes = $this->getVariableStartToEndRegex();
 617                 foreach ( $regexes as $regex ) {
 618                         if ( $regex !== '' ) {
 619                                 $m = false;
 620                                 if ( preg_match( $regex, $text, $m ) ) {
 621                                         return $this->parseMatch( $m );
 622                                 }
 623                         }
 624                 }
 625                 return array( false, false );
 626         }
 627
 628         /**
 629          * Match some text, without parameter capture
 630          * Returns the magic word name, or false if there was no capture
 631          */
 632         public function matchStartToEnd( $text ) {
 633                 $hash = $this->getHash();
 634                 if ( isset( $hash[1][$text] ) ) {
 635                         return $hash[1][$text];
 636                 }
 637                 global $wgContLang;
 638                 $lc = $wgContLang->lc( $text );
 639                 if ( isset( $hash[0][$lc] ) ) {
 640                         return $hash[0][$lc];
 641                 }
 642                 return false;
 643         }
 644
 645         /**
 646          * Returns an associative array, ID => param value, for all items that match
 647          * Removes the matched items from the input string (passed by reference)
 648          */
 649         public function matchAndRemove( &$text ) {
 650                 $found = array();
 651                 $regexes = $this->getRegex();
 652                 foreach ( $regexes as $regex ) {
 653                         if ( $regex === '' ) {
 654                                 continue;
 655                         }
 656                         preg_match_all( $regex, $text, $matches, PREG_SET_ORDER );
 657                         foreach ( $matches as $m ) {
 658                                 list( $name, $param ) = $this->parseMatch( $m );
 659                                 $found[$name] = $param;
 660                         }
 661                         $text = preg_replace( $regex, '', $text );
 662                 }
 663                 return $found;
 664         }
 665 }