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                 'currentmonth1',
  40                 'currentmonthname',
  41                 'currentmonthnamegen',
  42                 'currentmonthabbrev',
  43                 'currentday',
  44                 'currentday2',
  45                 'currentdayname',
  46                 'currentyear',
  47                 'currenttime',
  48                 'currenthour',
  49                 'localmonth',
  50                 'localmonth1',
  51                 'localmonthname',
  52                 'localmonthnamegen',
  53                 'localmonthabbrev',
  54                 'localday',
  55                 'localday2',
  56                 'localdayname',
  57                 'localyear',
  58                 'localtime',
  59                 'localhour',
  60                 'numberofarticles',
  61                 'numberoffiles',
  62                 'numberofedits',
  63                 'sitename',
  64                 'server',
  65                 'servername',
  66                 'scriptpath',
  67                 'pagename',
  68                 'pagenamee',
  69                 'fullpagename',
  70                 'fullpagenamee',
  71                 'namespace',
  72                 'namespacee',
  73                 'currentweek',
  74                 'currentdow',
  75                 'localweek',
  76                 'localdow',
  77                 'revisionid',
  78                 'revisionday',
  79                 'revisionday2',
  80                 'revisionmonth',
  81                 'revisionyear',
  82                 'revisiontimestamp',
  83                 'revisionuser',
  84                 'subpagename',
  85                 'subpagenamee',
  86                 'talkspace',
  87                 'talkspacee',
  88                 'subjectspace',
  89                 'subjectspacee',
  90                 'talkpagename',
  91                 'talkpagenamee',
  92                 'subjectpagename',
  93                 'subjectpagenamee',
  94                 'numberofusers',
  95                 'numberofactiveusers',
  96                 'numberofpages',
  97                 'currentversion',
  98                 'basepagename',
  99                 'basepagenamee',
 100                 'currenttimestamp',
 101                 'localtimestamp',
 102                 'directionmark',
 103                 'contentlanguage',
 104                 'numberofadmins',
 105                 'numberofviews',
 106         );
 107
 108         /* Array of caching hints for ParserCache */
 109         static public $mCacheTTLs = array (
 110                 'currentmonth' => 86400,
 111                 'currentmonth1' => 86400,
 112                 'currentmonthname' => 86400,
 113                 'currentmonthnamegen' => 86400,
 114                 'currentmonthabbrev' => 86400,
 115                 'currentday' => 3600,
 116                 'currentday2' => 3600,
 117                 'currentdayname' => 3600,
 118                 'currentyear' => 86400,
 119                 'currenttime' => 3600,
 120                 'currenthour' => 3600,
 121                 'localmonth' => 86400,
 122                 'localmonth1' => 86400,
 123                 'localmonthname' => 86400,
 124                 'localmonthnamegen' => 86400,
 125                 'localmonthabbrev' => 86400,
 126                 'localday' => 3600,
 127                 'localday2' => 3600,
 128                 'localdayname' => 3600,
 129                 'localyear' => 86400,
 130                 'localtime' => 3600,
 131                 'localhour' => 3600,
 132                 'numberofarticles' => 3600,
 133                 'numberoffiles' => 3600,
 134                 'numberofedits' => 3600,
 135                 'currentweek' => 3600,
 136                 'currentdow' => 3600,
 137                 'localweek' => 3600,
 138                 'localdow' => 3600,
 139                 'numberofusers' => 3600,
 140                 'numberofactiveusers' => 3600,
 141                 'numberofpages' => 3600,
 142                 'currentversion' => 86400,
 143                 'currenttimestamp' => 3600,
 144                 'localtimestamp' => 3600,
 145                 'pagesinnamespace' => 3600,
 146                 'numberofadmins' => 3600,
 147                 'numberofviews' => 3600,
 148                 'numberingroup' => 3600,
 149                 );
 150
 151         static public $mDoubleUnderscoreIDs = array(
 152                 'notoc',
 153                 'nogallery',
 154                 'forcetoc',
 155                 'toc',
 156                 'noeditsection',
 157                 'newsectionlink',
 158                 'nonewsectionlink',
 159                 'hiddencat',
 160                 'index',
 161                 'noindex',
 162                 'staticredirect',
 163         );
 164
 165
 166         static public $mObjects = array();
 167         static public $mDoubleUnderscoreArray = null;
 168
 169         /**#@-*/
 170
 171         function __construct($id = 0, $syn = '', $cs = false) {
 172                 $this->mId = $id;
 173                 $this->mSynonyms = (array)$syn;
 174                 $this->mCaseSensitive = $cs;
 175                 $this->mRegex = '';
 176                 $this->mRegexStart = '';
 177                 $this->mVariableRegex = '';
 178                 $this->mVariableStartToEndRegex = '';
 179                 $this->mModified = false;
 180         }
 181
 182         /**
 183          * Factory: creates an object representing an ID
 184          * @static
 185          */
 186         static function &get( $id ) {
 187                 wfProfileIn( __METHOD__ );
 188                 if ( !isset( self::$mObjects[$id] ) ) {
 189                         $mw = new MagicWord();
 190                         $mw->load( $id );
 191                         self::$mObjects[$id] = $mw;
 192                 }
 193                 wfProfileOut( __METHOD__ );
 194                 return self::$mObjects[$id];
 195         }
 196
 197         /**
 198          * Get an array of parser variable IDs
 199          */
 200         static function getVariableIDs() {
 201                 if ( !self::$mVariableIDsInitialised ) {
 202                         # Deprecated constant definition hook, available for extensions that need it
 203                         $magicWords = array();
 204                         wfRunHooks( 'MagicWordMagicWords', array( &$magicWords ) );
 205                         foreach ( $magicWords as $word ) {
 206                                 define( $word, $word );
 207                         }
 208
 209                         # Get variable IDs
 210                         wfRunHooks( 'MagicWordwgVariableIDs', array( &self::$mVariableIDs ) );
 211                         self::$mVariableIDsInitialised = true;
 212                 }
 213                 return self::$mVariableIDs;
 214         }
 215
 216         /* Allow external reads of TTL array */
 217         static function getCacheTTL($id) {
 218                 if (array_key_exists($id,self::$mCacheTTLs)) {
 219                         return self::$mCacheTTLs[$id];
 220                 } else {
 221                         return -1;
 222                 }
 223         }
 224
 225         /** Get a MagicWordArray of double-underscore entities */
 226         static function getDoubleUnderscoreArray() {
 227                 if ( is_null( self::$mDoubleUnderscoreArray ) ) {
 228                         self::$mDoubleUnderscoreArray = new MagicWordArray( self::$mDoubleUnderscoreIDs );
 229                 }
 230                 return self::$mDoubleUnderscoreArray;
 231         }
 232
 233         /**
 234          * Clear the self::$mObjects variable
 235          * For use in parser tests
 236          */
 237         public static function clearCache() {
 238                 self::$mObjects = array();
 239         }
 240
 241         # Initialises this object with an ID
 242         function load( $id ) {
 243                 global $wgContLang;
 244                 $this->mId = $id;
 245                 $wgContLang->getMagic( $this );
 246                 if ( !$this->mSynonyms ) {
 247                         $this->mSynonyms = array( 'dkjsagfjsgashfajsh' );
 248                         #throw new MWException( "Error: invalid magic word '$id'" );
 249                         wfDebugLog( 'exception', "Error: invalid magic word '$id'\n" );
 250                 }
 251         }
 252
 253         /**
 254          * Preliminary initialisation
 255          * @private
 256          */
 257         function initRegex() {
 258                 #$variableClass = Title::legalChars();
 259                 # This was used for matching "$1" variables, but different uses of the feature will have
 260                 # different restrictions, which should be checked *after* the MagicWord has been matched,
 261                 # not here. - IMSoP
 262
 263                 $escSyn = array();
 264                 foreach ( $this->mSynonyms as $synonym )
 265                         // In case a magic word contains /, like that's going to happen;)
 266                         $escSyn[] = preg_quote( $synonym, '/' );
 267                 $this->mBaseRegex = implode( '|', $escSyn );
 268
 269                 $case = $this->mCaseSensitive ? '' : 'iu';
 270                 $this->mRegex = "/{$this->mBaseRegex}/{$case}";
 271                 $this->mRegexStart = "/^(?:{$this->mBaseRegex})/{$case}";
 272                 $this->mVariableRegex = str_replace( "\\$1", "(.*?)", $this->mRegex );
 273                 $this->mVariableStartToEndRegex = str_replace( "\\$1", "(.*?)",
 274                         "/^(?:{$this->mBaseRegex})$/{$case}" );
 275         }
 276
 277         /**
 278          * Gets a regex representing matching the word
 279          */
 280         function getRegex() {
 281                 if ($this->mRegex == '' ) {
 282                         $this->initRegex();
 283                 }
 284                 return $this->mRegex;
 285         }
 286
 287         /**
 288          * Gets the regexp case modifier to use, i.e. i or nothing, to be used if
 289          * one is using MagicWord::getBaseRegex(), otherwise it'll be included in
 290          * the complete expression
 291          */
 292         function getRegexCase() {
 293                 if ( $this->mRegex === '' )
 294                         $this->initRegex();
 295
 296                 return $this->mCaseSensitive ? '' : 'iu';
 297         }
 298
 299         /**
 300          * Gets a regex matching the word, if it is at the string start
 301          */
 302         function getRegexStart() {
 303                 if ($this->mRegex == '' ) {
 304                         $this->initRegex();
 305                 }
 306                 return $this->mRegexStart;
 307         }
 308
 309         /**
 310          * regex without the slashes and what not
 311          */
 312         function getBaseRegex() {
 313                 if ($this->mRegex == '') {
 314                         $this->initRegex();
 315                 }
 316                 return $this->mBaseRegex;
 317         }
 318
 319         /**
 320          * Returns true if the text contains the word
 321          * @return bool
 322          */
 323         function match( $text ) {
 324                 return preg_match( $this->getRegex(), $text );
 325         }
 326
 327         /**
 328          * Returns true if the text starts with the word
 329          * @return bool
 330          */
 331         function matchStart( $text ) {
 332                 return preg_match( $this->getRegexStart(), $text );
 333         }
 334
 335         /**
 336          * Returns NULL if there's no match, the value of $1 otherwise
 337          * The return code is the matched string, if there's no variable
 338          * part in the regex and the matched variable part ($1) if there
 339          * is one.
 340          */
 341         function matchVariableStartToEnd( $text ) {
 342                 $matches = array();
 343                 $matchcount = preg_match( $this->getVariableStartToEndRegex(), $text, $matches );
 344                 if ( $matchcount == 0 ) {
 345                         return NULL;
 346                 } else {
 347                         # multiple matched parts (variable match); some will be empty because of
 348                         # synonyms. The variable will be the second non-empty one so remove any
 349                         # blank elements and re-sort the indices.
 350                         # See also bug 6526
 351
 352                         $matches = array_values(array_filter($matches));
 353
 354                         if ( count($matches) == 1 ) { return $matches[0]; }
 355                         else { return $matches[1]; }
 356                 }
 357         }
 358
 359
 360         /**
 361          * Returns true if the text matches the word, and alters the
 362          * input string, removing all instances of the word
 363          */
 364         function matchAndRemove( &$text ) {
 365                 $this->mFound = false;
 366                 $text = preg_replace_callback( $this->getRegex(), array( &$this, 'pregRemoveAndRecord' ), $text );
 367                 return $this->mFound;
 368         }
 369
 370         function matchStartAndRemove( &$text ) {
 371                 $this->mFound = false;
 372                 $text = preg_replace_callback( $this->getRegexStart(), array( &$this, 'pregRemoveAndRecord' ), $text );
 373                 return $this->mFound;
 374         }
 375
 376         /**
 377          * Used in matchAndRemove()
 378          * @private
 379          **/
 380         function pregRemoveAndRecord( ) {
 381                 $this->mFound = true;
 382                 return '';
 383         }
 384
 385         /**
 386          * Replaces the word with something else
 387          */
 388         function replace( $replacement, $subject, $limit=-1 ) {
 389                 $res = preg_replace( $this->getRegex(), StringUtils::escapeRegexReplacement( $replacement ), $subject, $limit );
 390                 $this->mModified = !($res === $subject);
 391                 return $res;
 392         }
 393
 394         /**
 395          * Variable handling: {{SUBST:xxx}} style words
 396          * Calls back a function to determine what to replace xxx with
 397          * Input word must contain $1
 398          */
 399         function substituteCallback( $text, $callback ) {
 400                 $res = preg_replace_callback( $this->getVariableRegex(), $callback, $text );
 401                 $this->mModified = !($res === $text);
 402                 return $res;
 403         }
 404
 405         /**
 406          * Matches the word, where $1 is a wildcard
 407          */
 408         function getVariableRegex()     {
 409                 if ( $this->mVariableRegex == '' ) {
 410                         $this->initRegex();
 411                 }
 412                 return $this->mVariableRegex;
 413         }
 414
 415         /**
 416          * Matches the entire string, where $1 is a wildcard
 417          */
 418         function getVariableStartToEndRegex() {
 419                 if ( $this->mVariableStartToEndRegex == '' ) {
 420                         $this->initRegex();
 421                 }
 422                 return $this->mVariableStartToEndRegex;
 423         }
 424
 425         /**
 426          * Accesses the synonym list directly
 427          */
 428         function getSynonym( $i ) {
 429                 return $this->mSynonyms[$i];
 430         }
 431
 432         function getSynonyms() {
 433                 return $this->mSynonyms;
 434         }
 435
 436         /**
 437          * Returns true if the last call to replace() or substituteCallback()
 438          * returned a modified text, otherwise false.
 439          */
 440         function getWasModified(){
 441                 return $this->mModified;
 442         }
 443
 444         /**
 445          * $magicarr is an associative array of (magic word ID => replacement)
 446          * This method uses the php feature to do several replacements at the same time,
 447          * thereby gaining some efficiency. The result is placed in the out variable
 448          * $result. The return value is true if something was replaced.
 449          * @static
 450          **/
 451         function replaceMultiple( $magicarr, $subject, &$result ){
 452                 $search = array();
 453                 $replace = array();
 454                 foreach( $magicarr as $id => $replacement ){
 455                         $mw = MagicWord::get( $id );
 456                         $search[] = $mw->getRegex();
 457                         $replace[] = $replacement;
 458                 }
 459
 460                 $result = preg_replace( $search, $replace, $subject );
 461                 return !($result === $subject);
 462         }
 463
 464         /**
 465          * Adds all the synonyms of this MagicWord to an array, to allow quick
 466          * lookup in a list of magic words
 467          */
 468         function addToArray( &$array, $value ) {
 469                 global $wgContLang;
 470                 foreach ( $this->mSynonyms as $syn ) {
 471                         $array[$wgContLang->lc($syn)] = $value;
 472                 }
 473         }
 474
 475         function isCaseSensitive() {
 476                 return $this->mCaseSensitive;
 477         }
 478
 479         function getId() {
 480                 return $this->mId;
 481         }
 482 }
 483
 484 /**
 485  * Class for handling an array of magic words
 486  * @ingroup Parser
 487  */
 488 class MagicWordArray {
 489         var $names = array();
 490         var $hash;
 491         var $baseRegex, $regex;
 492         var $matches;
 493
 494         function __construct( $names = array() ) {
 495                 $this->names = $names;
 496         }
 497
 498         /**
 499          * Add a magic word by name
 500          */
 501         public function add( $name ) {
 502                 global $wgContLang;
 503                 $this->names[] = $name;
 504                 $this->hash = $this->baseRegex = $this->regex = null;
 505         }
 506
 507         /**
 508          * Add a number of magic words by name
 509          */
 510         public function addArray( $names ) {
 511                 $this->names = array_merge( $this->names, array_values( $names ) );
 512                 $this->hash = $this->baseRegex = $this->regex = null;
 513         }
 514
 515         /**
 516          * Get a 2-d hashtable for this array
 517          */
 518         function getHash() {
 519                 if ( is_null( $this->hash ) ) {
 520                         global $wgContLang;
 521                         $this->hash = array( 0 => array(), 1 => array() );
 522                         foreach ( $this->names as $name ) {
 523                                 $magic = MagicWord::get( $name );
 524                                 $case = intval( $magic->isCaseSensitive() );
 525                                 foreach ( $magic->getSynonyms() as $syn ) {
 526                                         if ( !$case ) {
 527                                                 $syn = $wgContLang->lc( $syn );
 528                                         }
 529                                         $this->hash[$case][$syn] = $name;
 530                                 }
 531                         }
 532                 }
 533                 return $this->hash;
 534         }
 535
 536         /**
 537          * Get the base regex
 538          */
 539         function getBaseRegex() {
 540                 if ( is_null( $this->baseRegex ) ) {
 541                         $this->baseRegex = array( 0 => '', 1 => '' );
 542                         foreach ( $this->names as $name ) {
 543                                 $magic = MagicWord::get( $name );
 544                                 $case = intval( $magic->isCaseSensitive() );
 545                                 foreach ( $magic->getSynonyms() as $i => $syn ) {
 546                                         $group = "(?P<{$i}_{$name}>" . preg_quote( $syn, '/' ) . ')';
 547                                         if ( $this->baseRegex[$case] === '' ) {
 548                                                 $this->baseRegex[$case] = $group;
 549                                         } else {
 550                                                 $this->baseRegex[$case] .= '|' . $group;
 551                                         }
 552                                 }
 553                         }
 554                 }
 555                 return $this->baseRegex;
 556         }
 557
 558         /**
 559          * Get an unanchored regex
 560          */
 561         function getRegex() {
 562                 if ( is_null( $this->regex ) ) {
 563                         $base = $this->getBaseRegex();
 564                         $this->regex = array( '', '' );
 565                         if ( $this->baseRegex[0] !== '' ) {
 566                                 $this->regex[0] = "/{$base[0]}/iuS";
 567                         }
 568                         if ( $this->baseRegex[1] !== '' ) {
 569                                 $this->regex[1] = "/{$base[1]}/S";
 570                         }
 571                 }
 572                 return $this->regex;
 573         }
 574
 575         /**
 576          * Get a regex for matching variables
 577          */
 578         function getVariableRegex() {
 579                 return str_replace( "\\$1", "(.*?)", $this->getRegex() );
 580         }
 581
 582         /**
 583          * Get an anchored regex for matching variables
 584          */
 585         function getVariableStartToEndRegex() {
 586                 $base = $this->getBaseRegex();
 587                 $newRegex = array( '', '' );
 588                 if ( $base[0] !== '' ) {
 589                         $newRegex[0] = str_replace( "\\$1", "(.*?)", "/^(?:{$base[0]})$/iuS" );
 590                 }
 591                 if ( $base[1] !== '' ) {
 592                         $newRegex[1] = str_replace( "\\$1", "(.*?)", "/^(?:{$base[1]})$/S" );
 593                 }
 594                 return $newRegex;
 595         }
 596
 597         /**
 598          * Parse a match array from preg_match
 599          * Returns array(magic word ID, parameter value)
 600          * If there is no parameter value, that element will be false.
 601          */
 602         function parseMatch( $m ) {
 603                 reset( $m );
 604                 while ( list( $key, $value ) = each( $m ) ) {
 605                         if ( $key === 0 || $value === '' ) {
 606                                 continue;
 607                         }
 608                         $parts = explode( '_', $key, 2 );
 609                         if ( count( $parts ) != 2 ) {
 610                                 // This shouldn't happen
 611                                 // continue;
 612                                 throw new MWException( __METHOD__ . ': bad parameter name' );
 613                         }
 614                         list( /* $synIndex */, $magicName ) = $parts;
 615                         $paramValue = next( $m );
 616                         return array( $magicName, $paramValue );
 617                 }
 618                 // This shouldn't happen either
 619                 throw new MWException( __METHOD__.': parameter not found' );
 620                 return array( false, false );
 621         }
 622
 623         /**
 624          * Match some text, with parameter capture
 625          * Returns an array with the magic word name in the first element and the
 626          * parameter in the second element.
 627          * Both elements are false if there was no match.
 628          */
 629         public function matchVariableStartToEnd( $text ) {
 630                 global $wgContLang;
 631                 $regexes = $this->getVariableStartToEndRegex();
 632                 foreach ( $regexes as $regex ) {
 633                         if ( $regex !== '' ) {
 634                                 $m = false;
 635                                 if ( preg_match( $regex, $text, $m ) ) {
 636                                         return $this->parseMatch( $m );
 637                                 }
 638                         }
 639                 }
 640                 return array( false, false );
 641         }
 642
 643         /**
 644          * Match some text, without parameter capture
 645          * Returns the magic word name, or false if there was no capture
 646          */
 647         public function matchStartToEnd( $text ) {
 648                 $hash = $this->getHash();
 649                 if ( isset( $hash[1][$text] ) ) {
 650                         return $hash[1][$text];
 651                 }
 652                 global $wgContLang;
 653                 $lc = $wgContLang->lc( $text );
 654                 if ( isset( $hash[0][$lc] ) ) {
 655                         return $hash[0][$lc];
 656                 }
 657                 return false;
 658         }
 659
 660         /**
 661          * Returns an associative array, ID => param value, for all items that match
 662          * Removes the matched items from the input string (passed by reference)
 663          */
 664         public function matchAndRemove( &$text ) {
 665                 $found = array();
 666                 $regexes = $this->getRegex();
 667                 foreach ( $regexes as $regex ) {
 668                         if ( $regex === '' ) {
 669                                 continue;
 670                         }
 671                         preg_match_all( $regex, $text, $matches, PREG_SET_ORDER );
 672                         foreach ( $matches as $m ) {
 673                                 list( $name, $param ) = $this->parseMatch( $m );
 674                                 $found[$name] = $param;
 675                         }
 676                         $text = preg_replace( $regex, '', $text );
 677                 }
 678                 return $found;
 679         }
 680 }