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