includes/MagicWord.php

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