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