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                 'numberofviews',
 107                 'defaultsort',
 108                 'pagesincategory',
 109                 'index',
 110                 'noindex',
 111                 'numberingroup',
 112                 'protectionlevel',
 113         );
 114
 115         /* Array of caching hints for ParserCache */
 116         static public $mCacheTTLs = array (
 117                 'currentmonth' => 86400,
 118                 'currentmonthname' => 86400,
 119                 'currentmonthnamegen' => 86400,
 120                 'currentmonthabbrev' => 86400,
 121                 'currentday' => 3600,
 122                 'currentday2' => 3600,
 123                 'currentdayname' => 3600,
 124                 'currentyear' => 86400,
 125                 'currenttime' => 3600,
 126                 'currenthour' => 3600,
 127                 'localmonth' => 86400,
 128                 'localmonthname' => 86400,
 129                 'localmonthnamegen' => 86400,
 130                 'localmonthabbrev' => 86400,
 131                 'localday' => 3600,
 132                 'localday2' => 3600,
 133                 'localdayname' => 3600,
 134                 'localyear' => 86400,
 135                 'localtime' => 3600,
 136                 'localhour' => 3600,
 137                 'numberofarticles' => 3600,
 138                 'numberoffiles' => 3600,
 139                 'numberofedits' => 3600,
 140                 'currentweek' => 3600,
 141                 'currentdow' => 3600,
 142                 'localweek' => 3600,
 143                 'localdow' => 3600,
 144                 'numberofusers' => 3600,
 145                 'numberofpages' => 3600,
 146                 'currentversion' => 86400,
 147                 'currenttimestamp' => 3600,
 148                 'localtimestamp' => 3600,
 149                 'pagesinnamespace' => 3600,
 150                 'numberofadmins' => 3600,
 151                 'numberofviews' => 3600,
 152                 'numberingroup' => 3600,
 153                 );
 154
 155         static public $mDoubleUnderscoreIDs = array(
 156                 'notoc',
 157                 'nogallery',
 158                 'forcetoc',
 159                 'toc',
 160                 'noeditsection',
 161                 'newsectionlink',
 162                 'hiddencat',
 163                 'index',
 164                 'noindex',
 165                 'staticredirect',
 166         );
 167
 168
 169         static public $mObjects = array();
 170         static public $mDoubleUnderscoreArray = null;
 171
 172         /**#@-*/
 173
 174         function __construct($id = 0, $syn = '', $cs = false) {
 175                 $this->mId = $id;
 176                 $this->mSynonyms = (array)$syn;
 177                 $this->mCaseSensitive = $cs;
 178                 $this->mRegex = '';
 179                 $this->mRegexStart = '';
 180                 $this->mVariableRegex = '';
 181                 $this->mVariableStartToEndRegex = '';
 182                 $this->mModified = false;
 183         }
 184
 185         /**
 186          * Factory: creates an object representing an ID
 187          * @static
 188          */
 189         static function &get( $id ) {
 190                 wfProfileIn( __METHOD__ );
 191                 if (!array_key_exists( $id, self::$mObjects ) ) {
 192                         $mw = new MagicWord();
 193                         $mw->load( $id );
 194                         self::$mObjects[$id] = $mw;
 195                 }
 196                 wfProfileOut( __METHOD__ );
 197                 return self::$mObjects[$id];
 198         }
 199
 200         /**
 201          * Get an array of parser variable IDs
 202          */
 203         static function getVariableIDs() {
 204                 if ( !self::$mVariableIDsInitialised ) {
 205                         # Deprecated constant definition hook, available for extensions that need it
 206                         $magicWords = array();
 207                         wfRunHooks( 'MagicWordMagicWords', array( &$magicWords ) );
 208                         foreach ( $magicWords as $word ) {
 209                                 define( $word, $word );
 210                         }
 211
 212                         # Get variable IDs
 213                         wfRunHooks( 'MagicWordwgVariableIDs', array( &self::$mVariableIDs ) );
 214                         self::$mVariableIDsInitialised = true;
 215                 }
 216                 return self::$mVariableIDs;
 217         }
 218
 219         /* Allow external reads of TTL array */
 220         static function getCacheTTL($id) {
 221                 if (array_key_exists($id,self::$mCacheTTLs)) {
 222                         return self::$mCacheTTLs[$id];
 223                 } else {
 224                         return -1;
 225                 }
 226         }
 227
 228         /** Get a MagicWordArray of double-underscore entities */
 229         static function getDoubleUnderscoreArray() {
 230                 if ( is_null( self::$mDoubleUnderscoreArray ) ) {
 231                         self::$mDoubleUnderscoreArray = new MagicWordArray( self::$mDoubleUnderscoreIDs );
 232                 }
 233                 return self::$mDoubleUnderscoreArray;
 234         }
 235
 236         # Initialises this object with an ID
 237         function load( $id ) {
 238                 global $wgContLang;
 239                 $this->mId = $id;
 240                 $wgContLang->getMagic( $this );
 241                 if ( !$this->mSynonyms ) {
 242                         $this->mSynonyms = array( 'dkjsagfjsgashfajsh' );
 243                         #throw new MWException( "Error: invalid magic word '$id'" );
 244                         wfDebugLog( 'exception', "Error: invalid magic word '$id'\n" );
 245                 }
 246         }
 247
 248         /**
 249          * Preliminary initialisation
 250          * @private
 251          */
 252         function initRegex() {
 253                 #$variableClass = Title::legalChars();
 254                 # This was used for matching "$1" variables, but different uses of the feature will have
 255                 # different restrictions, which should be checked *after* the MagicWord has been matched,
 256                 # not here. - IMSoP
 257
 258                 $escSyn = array();
 259                 foreach ( $this->mSynonyms as $synonym )
 260                         // In case a magic word contains /, like that's going to happen;)
 261                         $escSyn[] = preg_quote( $synonym, '/' );
 262                 $this->mBaseRegex = implode( '|', $escSyn );
 263
 264                 $case = $this->mCaseSensitive ? '' : 'iu';
 265                 $this->mRegex = "/{$this->mBaseRegex}/{$case}";
 266                 $this->mRegexStart = "/^(?:{$this->mBaseRegex})/{$case}";
 267                 $this->mVariableRegex = str_replace( "\\$1", "(.*?)", $this->mRegex );
 268                 $this->mVariableStartToEndRegex = str_replace( "\\$1", "(.*?)",
 269                         "/^(?:{$this->mBaseRegex})$/{$case}" );
 270         }
 271
 272         /**
 273          * Gets a regex representing matching the word
 274          */
 275         function getRegex() {
 276                 if ($this->mRegex == '' ) {
 277                         $this->initRegex();
 278                 }
 279                 return $this->mRegex;
 280         }
 281
 282         /**
 283          * Gets the regexp case modifier to use, i.e. i or nothing, to be used if
 284          * one is using MagicWord::getBaseRegex(), otherwise it'll be included in
 285          * the complete expression
 286          */
 287         function getRegexCase() {
 288                 if ( $this->mRegex === '' )
 289                         $this->initRegex();
 290
 291                 return $this->mCaseSensitive ? '' : 'iu';
 292         }
 293
 294         /**
 295          * Gets a regex matching the word, if it is at the string start
 296          */
 297         function getRegexStart() {
 298                 if ($this->mRegex == '' ) {
 299                         $this->initRegex();
 300                 }
 301                 return $this->mRegexStart;
 302         }
 303
 304         /**
 305          * regex without the slashes and what not
 306          */
 307         function getBaseRegex() {
 308                 if ($this->mRegex == '') {
 309                         $this->initRegex();
 310                 }
 311                 return $this->mBaseRegex;
 312         }
 313
 314         /**
 315          * Returns true if the text contains the word
 316          * @return bool
 317          */
 318         function match( $text ) {
 319                 return preg_match( $this->getRegex(), $text );
 320         }
 321
 322         /**
 323          * Returns true if the text starts with the word
 324          * @return bool
 325          */
 326         function matchStart( $text ) {
 327                 return preg_match( $this->getRegexStart(), $text );
 328         }
 329
 330         /**
 331          * Returns NULL if there's no match, the value of $1 otherwise
 332          * The return code is the matched string, if there's no variable
 333          * part in the regex and the matched variable part ($1) if there
 334          * is one.
 335          */
 336         function matchVariableStartToEnd( $text ) {
 337                 $matches = array();
 338                 $matchcount = preg_match( $this->getVariableStartToEndRegex(), $text, $matches );
 339                 if ( $matchcount == 0 ) {
 340                         return NULL;
 341                 } else {
 342                         # multiple matched parts (variable match); some will be empty because of
 343                         # synonyms. The variable will be the second non-empty one so remove any
 344                         # blank elements and re-sort the indices.
 345                         # See also bug 6526
 346
 347                         $matches = array_values(array_filter($matches));
 348
 349                         if ( count($matches) == 1 ) { return $matches[0]; }
 350                         else { return $matches[1]; }
 351                 }
 352         }
 353
 354
 355         /**
 356          * Returns true if the text matches the word, and alters the
 357          * input string, removing all instances of the word
 358          */
 359         function matchAndRemove( &$text ) {
 360                 $this->mFound = false;
 361                 $text = preg_replace_callback( $this->getRegex(), array( &$this, 'pregRemoveAndRecord' ), $text );
 362                 return $this->mFound;
 363         }
 364
 365         function matchStartAndRemove( &$text ) {
 366                 $this->mFound = false;
 367                 $text = preg_replace_callback( $this->getRegexStart(), array( &$this, 'pregRemoveAndRecord' ), $text );
 368                 return $this->mFound;
 369         }
 370
 371         /**
 372          * Used in matchAndRemove()
 373          * @private
 374          **/
 375         function pregRemoveAndRecord( ) {
 376                 $this->mFound = true;
 377                 return '';
 378         }
 379
 380         /**
 381          * Replaces the word with something else
 382          */
 383         function replace( $replacement, $subject, $limit=-1 ) {
 384                 $res = preg_replace( $this->getRegex(), StringUtils::escapeRegexReplacement( $replacement ), $subject, $limit );
 385                 $this->mModified = !($res === $subject);
 386                 return $res;
 387         }
 388
 389         /**
 390          * Variable handling: {{SUBST:xxx}} style words
 391          * Calls back a function to determine what to replace xxx with
 392          * Input word must contain $1
 393          */
 394         function substituteCallback( $text, $callback ) {
 395                 $res = preg_replace_callback( $this->getVariableRegex(), $callback, $text );
 396                 $this->mModified = !($res === $text);
 397                 return $res;
 398         }
 399
 400         /**
 401          * Matches the word, where $1 is a wildcard
 402          */
 403         function getVariableRegex()     {
 404                 if ( $this->mVariableRegex == '' ) {
 405                         $this->initRegex();
 406                 }
 407                 return $this->mVariableRegex;
 408         }
 409
 410         /**
 411          * Matches the entire string, where $1 is a wildcard
 412          */
 413         function getVariableStartToEndRegex() {
 414                 if ( $this->mVariableStartToEndRegex == '' ) {
 415                         $this->initRegex();
 416                 }
 417                 return $this->mVariableStartToEndRegex;
 418         }
 419
 420         /**
 421          * Accesses the synonym list directly
 422          */
 423         function getSynonym( $i ) {
 424                 return $this->mSynonyms[$i];
 425         }
 426
 427         function getSynonyms() {
 428                 return $this->mSynonyms;
 429         }
 430
 431         /**
 432          * Returns true if the last call to replace() or substituteCallback()
 433          * returned a modified text, otherwise false.
 434          */
 435         function getWasModified(){
 436                 return $this->mModified;
 437         }
 438
 439         /**
 440          * $magicarr is an associative array of (magic word ID => replacement)
 441          * This method uses the php feature to do several replacements at the same time,
 442          * thereby gaining some efficiency. The result is placed in the out variable
 443          * $result. The return value is true if something was replaced.
 444          * @static
 445          **/
 446         function replaceMultiple( $magicarr, $subject, &$result ){
 447                 $search = array();
 448                 $replace = array();
 449                 foreach( $magicarr as $id => $replacement ){
 450                         $mw = MagicWord::get( $id );
 451                         $search[] = $mw->getRegex();
 452                         $replace[] = $replacement;
 453                 }
 454
 455                 $result = preg_replace( $search, $replace, $subject );
 456                 return !($result === $subject);
 457         }
 458
 459         /**
 460          * Adds all the synonyms of this MagicWord to an array, to allow quick
 461          * lookup in a list of magic words
 462          */
 463         function addToArray( &$array, $value ) {
 464                 global $wgContLang;
 465                 foreach ( $this->mSynonyms as $syn ) {
 466                         $array[$wgContLang->lc($syn)] = $value;
 467                 }
 468         }
 469
 470         function isCaseSensitive() {
 471                 return $this->mCaseSensitive;
 472         }
 473
 474         function getId() {
 475                 return $this->mId;
 476         }
 477 }
 478
 479 /**
 480  * Class for handling an array of magic words
 481  * @ingroup Parser
 482  */
 483 class MagicWordArray {
 484         var $names = array();
 485         var $hash;
 486         var $baseRegex, $regex;
 487         var $matches;
 488
 489         function __construct( $names = array() ) {
 490                 $this->names = $names;
 491         }
 492
 493         /**
 494          * Add a magic word by name
 495          */
 496         public function add( $name ) {
 497                 global $wgContLang;
 498                 $this->names[] = $name;
 499                 $this->hash = $this->baseRegex = $this->regex = null;
 500         }
 501
 502         /**
 503          * Add a number of magic words by name
 504          */
 505         public function addArray( $names ) {
 506                 $this->names = array_merge( $this->names, array_values( $names ) );
 507                 $this->hash = $this->baseRegex = $this->regex = null;
 508         }
 509
 510         /**
 511          * Get a 2-d hashtable for this array
 512          */
 513         function getHash() {
 514                 if ( is_null( $this->hash ) ) {
 515                         global $wgContLang;
 516                         $this->hash = array( 0 => array(), 1 => array() );
 517                         foreach ( $this->names as $name ) {
 518                                 $magic = MagicWord::get( $name );
 519                                 $case = intval( $magic->isCaseSensitive() );
 520                                 foreach ( $magic->getSynonyms() as $syn ) {
 521                                         if ( !$case ) {
 522                                                 $syn = $wgContLang->lc( $syn );
 523                                         }
 524                                         $this->hash[$case][$syn] = $name;
 525                                 }
 526                         }
 527                 }
 528                 return $this->hash;
 529         }
 530
 531         /**
 532          * Get the base regex
 533          */
 534         function getBaseRegex() {
 535                 if ( is_null( $this->baseRegex ) ) {
 536                         $this->baseRegex = array( 0 => '', 1 => '' );
 537                         foreach ( $this->names as $name ) {
 538                                 $magic = MagicWord::get( $name );
 539                                 $case = intval( $magic->isCaseSensitive() );
 540                                 foreach ( $magic->getSynonyms() as $i => $syn ) {
 541                                         $group = "(?P<{$i}_{$name}>" . preg_quote( $syn, '/' ) . ')';
 542                                         if ( $this->baseRegex[$case] === '' ) {
 543                                                 $this->baseRegex[$case] = $group;
 544                                         } else {
 545                                                 $this->baseRegex[$case] .= '|' . $group;
 546                                         }
 547                                 }
 548                         }
 549                 }
 550                 return $this->baseRegex;
 551         }
 552
 553         /**
 554          * Get an unanchored regex
 555          */
 556         function getRegex() {
 557                 if ( is_null( $this->regex ) ) {
 558                         $base = $this->getBaseRegex();
 559                         $this->regex = array( '', '' );
 560                         if ( $this->baseRegex[0] !== '' ) {
 561                                 $this->regex[0] = "/{$base[0]}/iuS";
 562                         }
 563                         if ( $this->baseRegex[1] !== '' ) {
 564                                 $this->regex[1] = "/{$base[1]}/S";
 565                         }
 566                 }
 567                 return $this->regex;
 568         }
 569
 570         /**
 571          * Get a regex for matching variables
 572          */
 573         function getVariableRegex() {
 574                 return str_replace( "\\$1", "(.*?)", $this->getRegex() );
 575         }
 576
 577         /**
 578          * Get an anchored regex for matching variables
 579          */
 580         function getVariableStartToEndRegex() {
 581                 $base = $this->getBaseRegex();
 582                 $newRegex = array( '', '' );
 583                 if ( $base[0] !== '' ) {
 584                         $newRegex[0] = str_replace( "\\$1", "(.*?)", "/^(?:{$base[0]})$/iuS" );
 585                 }
 586                 if ( $base[1] !== '' ) {
 587                         $newRegex[1] = str_replace( "\\$1", "(.*?)", "/^(?:{$base[1]})$/S" );
 588                 }
 589                 return $newRegex;
 590         }
 591
 592         /**
 593          * Parse a match array from preg_match
 594          * Returns array(magic word ID, parameter value)
 595          * If there is no parameter value, that element will be false.
 596          */
 597         function parseMatch( $m ) {
 598                 reset( $m );
 599                 while ( list( $key, $value ) = each( $m ) ) {
 600                         if ( $key === 0 || $value === '' ) {
 601                                 continue;
 602                         }
 603                         $parts = explode( '_', $key, 2 );
 604                         if ( count( $parts ) != 2 ) {
 605                                 // This shouldn't happen
 606                                 // continue;
 607                                 throw new MWException( __METHOD__ . ': bad parameter name' );
 608                         }
 609                         list( /* $synIndex */, $magicName ) = $parts;
 610                         $paramValue = next( $m );
 611                         return array( $magicName, $paramValue );
 612                 }
 613                 // This shouldn't happen either
 614                 throw new MWException( __METHOD__.': parameter not found' );
 615                 return array( false, false );
 616         }
 617
 618         /**
 619          * Match some text, with parameter capture
 620          * Returns an array with the magic word name in the first element and the
 621          * parameter in the second element.
 622          * Both elements are false if there was no match.
 623          */
 624         public function matchVariableStartToEnd( $text ) {
 625                 global $wgContLang;
 626                 $regexes = $this->getVariableStartToEndRegex();
 627                 foreach ( $regexes as $regex ) {
 628                         if ( $regex !== '' ) {
 629                                 $m = false;
 630                                 if ( preg_match( $regex, $text, $m ) ) {
 631                                         return $this->parseMatch( $m );
 632                                 }
 633                         }
 634                 }
 635                 return array( false, false );
 636         }
 637
 638         /**
 639          * Match some text, without parameter capture
 640          * Returns the magic word name, or false if there was no capture
 641          */
 642         public function matchStartToEnd( $text ) {
 643                 $hash = $this->getHash();
 644                 if ( isset( $hash[1][$text] ) ) {
 645                         return $hash[1][$text];
 646                 }
 647                 global $wgContLang;
 648                 $lc = $wgContLang->lc( $text );
 649                 if ( isset( $hash[0][$lc] ) ) {
 650                         return $hash[0][$lc];
 651                 }
 652                 return false;
 653         }
 654
 655         /**
 656          * Returns an associative array, ID => param value, for all items that match
 657          * Removes the matched items from the input string (passed by reference)
 658          */
 659         public function matchAndRemove( &$text ) {
 660                 $found = array();
 661                 $regexes = $this->getRegex();
 662                 foreach ( $regexes as $regex ) {
 663                         if ( $regex === '' ) {
 664                                 continue;
 665                         }
 666                         preg_match_all( $regex, $text, $matches, PREG_SET_ORDER );
 667                         foreach ( $matches as $m ) {
 668                                 list( $name, $param ) = $this->parseMatch( $m );
 669                                 $found[$name] = $param;
 670                         }
 671                         $text = preg_replace( $regex, '', $text );
 672                 }
 673                 return $found;
 674         }
 675 }