includes/MagicWord.php

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