includes/MagicWord.php

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