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         static public $mObjects = array();
 106
 107         /**#@-*/
 108
 109         function __construct($id = 0, $syn = '', $cs = false) {
 110                 $this->mId = $id;
 111                 $this->mSynonyms = (array)$syn;
 112                 $this->mCaseSensitive = $cs;
 113                 $this->mRegex = '';
 114                 $this->mRegexStart = '';
 115                 $this->mVariableRegex = '';
 116                 $this->mVariableStartToEndRegex = '';
 117                 $this->mModified = false;
 118         }
 119
 120         /**
 121          * Factory: creates an object representing an ID
 122          * @static
 123          */
 124         static function &get( $id ) {
 125                 if (!array_key_exists( $id, self::$mObjects ) ) {
 126                         $mw = new MagicWord();
 127                         $mw->load( $id );
 128                         self::$mObjects[$id] = $mw;
 129                 }
 130                 return self::$mObjects[$id];
 131         }
 132
 133         /**
 134          * Get an array of parser variable IDs
 135          */
 136         static function getVariableIDs() {
 137                 if ( !self::$mVariableIDsInitialised ) {
 138                         # Deprecated constant definition hook, available for extensions that need it
 139                         $magicWords = array();
 140                         wfRunHooks( 'MagicWordMagicWords', array( &$magicWords ) );
 141                         foreach ( $magicWords as $word ) {
 142                                 define( $word, $word );
 143                         }
 144
 145                         # Get variable IDs
 146                         wfRunHooks( 'MagicWordwgVariableIDs', array( &self::$mVariableIDs ) );
 147                         self::$mVariableIDsInitialised = true;
 148                 }
 149                 return self::$mVariableIDs;
 150         }
 151
 152         # Initialises this object with an ID
 153         function load( $id ) {
 154                 global $wgContLang;
 155                 $this->mId = $id;
 156                 $wgContLang->getMagic( $this );
 157                 if ( !$this->mSynonyms ) {
 158                         $this->mSynonyms = array( 'dkjsagfjsgashfajsh' );
 159                         #throw new MWException( "Error: invalid magic word '$id'" );
 160                         wfDebugLog( 'exception', "Error: invalid magic word '$id'\n" );
 161                 }
 162         }
 163
 164         /**
 165          * Preliminary initialisation
 166          * @private
 167          */
 168         function initRegex() {
 169                 #$variableClass = Title::legalChars();
 170                 # This was used for matching "$1" variables, but different uses of the feature will have
 171                 # different restrictions, which should be checked *after* the MagicWord has been matched,
 172                 # not here. - IMSoP
 173
 174                 $escSyn = array();
 175                 foreach ( $this->mSynonyms as $synonym )
 176                         // In case a magic word contains /, like that's going to happen;)
 177                         $escSyn[] = preg_quote( $synonym, '/' );
 178                 $this->mBaseRegex = implode( '|', $escSyn );
 179
 180                 $case = $this->mCaseSensitive ? '' : 'iu';
 181                 $this->mRegex = "/{$this->mBaseRegex}/{$case}";
 182                 $this->mRegexStart = "/^(?:{$this->mBaseRegex})/{$case}";
 183                 $this->mVariableRegex = str_replace( "\\$1", "(.*?)", $this->mRegex );
 184                 $this->mVariableStartToEndRegex = str_replace( "\\$1", "(.*?)",
 185                         "/^(?:{$this->mBaseRegex})$/{$case}" );
 186         }
 187
 188         /**
 189          * Gets a regex representing matching the word
 190          */
 191         function getRegex() {
 192                 if ($this->mRegex == '' ) {
 193                         $this->initRegex();
 194                 }
 195                 return $this->mRegex;
 196         }
 197
 198         /**
 199          * Gets the regexp case modifier to use, i.e. i or nothing, to be used if
 200          * one is using MagicWord::getBaseRegex(), otherwise it'll be included in
 201          * the complete expression
 202          */
 203         function getRegexCase() {
 204                 if ( $this->mRegex === '' )
 205                         $this->initRegex();
 206
 207                 return $this->mCaseSensitive ? '' : 'iu';
 208         }
 209
 210         /**
 211          * Gets a regex matching the word, if it is at the string start
 212          */
 213         function getRegexStart() {
 214                 if ($this->mRegex == '' ) {
 215                         $this->initRegex();
 216                 }
 217                 return $this->mRegexStart;
 218         }
 219
 220         /**
 221          * regex without the slashes and what not
 222          */
 223         function getBaseRegex() {
 224                 if ($this->mRegex == '') {
 225                         $this->initRegex();
 226                 }
 227                 return $this->mBaseRegex;
 228         }
 229
 230         /**
 231          * Returns true if the text contains the word
 232          * @return bool
 233          */
 234         function match( $text ) {
 235                 return preg_match( $this->getRegex(), $text );
 236         }
 237
 238         /**
 239          * Returns true if the text starts with the word
 240          * @return bool
 241          */
 242         function matchStart( $text ) {
 243                 return preg_match( $this->getRegexStart(), $text );
 244         }
 245
 246         /**
 247          * Returns NULL if there's no match, the value of $1 otherwise
 248          * The return code is the matched string, if there's no variable
 249          * part in the regex and the matched variable part ($1) if there
 250          * is one.
 251          */
 252         function matchVariableStartToEnd( $text ) {
 253                 $matches = array();
 254                 $matchcount = preg_match( $this->getVariableStartToEndRegex(), $text, $matches );
 255                 if ( $matchcount == 0 ) {
 256                         return NULL;
 257                 } else {
 258                         # multiple matched parts (variable match); some will be empty because of
 259                         # synonyms. The variable will be the second non-empty one so remove any
 260                         # blank elements and re-sort the indices.
 261                         # See also bug 6526
 262
 263                         $matches = array_values(array_filter($matches));
 264
 265                         if ( count($matches) == 1 ) { return $matches[0]; }
 266                         else { return $matches[1]; }
 267                 }
 268         }
 269
 270
 271         /**
 272          * Returns true if the text matches the word, and alters the
 273          * input string, removing all instances of the word
 274          */
 275         function matchAndRemove( &$text ) {
 276                 $this->mFound = false;
 277                 $text = preg_replace_callback( $this->getRegex(), array( &$this, 'pregRemoveAndRecord' ), $text );
 278                 return $this->mFound;
 279         }
 280
 281         function matchStartAndRemove( &$text ) {
 282                 $this->mFound = false;
 283                 $text = preg_replace_callback( $this->getRegexStart(), array( &$this, 'pregRemoveAndRecord' ), $text );
 284                 return $this->mFound;
 285         }
 286
 287         /**
 288          * Used in matchAndRemove()
 289          * @private
 290          **/
 291         function pregRemoveAndRecord( ) {
 292                 $this->mFound = true;
 293                 return '';
 294         }
 295
 296         /**
 297          * Replaces the word with something else
 298          */
 299         function replace( $replacement, $subject, $limit=-1 ) {
 300                 $res = preg_replace( $this->getRegex(), StringUtils::escapeRegexReplacement( $replacement ), $subject, $limit );
 301                 $this->mModified = !($res === $subject);
 302                 return $res;
 303         }
 304
 305         /**
 306          * Variable handling: {{SUBST:xxx}} style words
 307          * Calls back a function to determine what to replace xxx with
 308          * Input word must contain $1
 309          */
 310         function substituteCallback( $text, $callback ) {
 311                 $res = preg_replace_callback( $this->getVariableRegex(), $callback, $text );
 312                 $this->mModified = !($res === $text);
 313                 return $res;
 314         }
 315
 316         /**
 317          * Matches the word, where $1 is a wildcard
 318          */
 319         function getVariableRegex()     {
 320                 if ( $this->mVariableRegex == '' ) {
 321                         $this->initRegex();
 322                 }
 323                 return $this->mVariableRegex;
 324         }
 325
 326         /**
 327          * Matches the entire string, where $1 is a wildcard
 328          */
 329         function getVariableStartToEndRegex() {
 330                 if ( $this->mVariableStartToEndRegex == '' ) {
 331                         $this->initRegex();
 332                 }
 333                 return $this->mVariableStartToEndRegex;
 334         }
 335
 336         /**
 337          * Accesses the synonym list directly
 338          */
 339         function getSynonym( $i ) {
 340                 return $this->mSynonyms[$i];
 341         }
 342
 343         function getSynonyms() {
 344                 return $this->mSynonyms;
 345         }
 346
 347         /**
 348          * Returns true if the last call to replace() or substituteCallback()
 349          * returned a modified text, otherwise false.
 350          */
 351         function getWasModified(){
 352                 return $this->mModified;
 353         }
 354
 355         /**
 356          * $magicarr is an associative array of (magic word ID => replacement)
 357          * This method uses the php feature to do several replacements at the same time,
 358          * thereby gaining some efficiency. The result is placed in the out variable
 359          * $result. The return value is true if something was replaced.
 360          * @static
 361          **/
 362         function replaceMultiple( $magicarr, $subject, &$result ){
 363                 $search = array();
 364                 $replace = array();
 365                 foreach( $magicarr as $id => $replacement ){
 366                         $mw = MagicWord::get( $id );
 367                         $search[] = $mw->getRegex();
 368                         $replace[] = $replacement;
 369                 }
 370
 371                 $result = preg_replace( $search, $replace, $subject );
 372                 return !($result === $subject);
 373         }
 374
 375         /**
 376          * Adds all the synonyms of this MagicWord to an array, to allow quick
 377          * lookup in a list of magic words
 378          */
 379         function addToArray( &$array, $value ) {
 380                 global $wgContLang;
 381                 foreach ( $this->mSynonyms as $syn ) {
 382                         $array[$wgContLang->lc($syn)] = $value;
 383                 }
 384         }
 385
 386         function isCaseSensitive() {
 387                 return $this->mCaseSensitive;
 388         }
 389
 390         function getId() {
 391                 return $this->mId;
 392         }
 393 }
 394
 395 /**
 396  * Class for handling an array of magic words
 397  */
 398 class MagicWordArray {
 399         var $names = array();
 400         var $hash;
 401         var $baseRegex, $regex;
 402
 403         function __construct( $names = array() ) {
 404                 $this->names = $names;
 405         }
 406
 407         /**
 408          * Add a magic word by name
 409          */
 410         public function add( $name ) {
 411                 global $wgContLang;
 412                 $this->names[] = $name;
 413                 $this->hash = $this->baseRegex = $this->regex = null;
 414         }
 415
 416         /**
 417          * Add a number of magic words by name
 418          */
 419         public function addArray( $names ) {
 420                 $this->names = array_merge( $this->names, array_values( $names ) );
 421                 $this->hash = $this->baseRegex = $this->regex = null;
 422         }
 423
 424         /**
 425          * Get a 2-d hashtable for this array
 426          */
 427         function getHash() {
 428                 if ( is_null( $this->hash ) ) {
 429                         global $wgContLang;
 430                         $this->hash = array( 0 => array(), 1 => array() );
 431                         foreach ( $this->names as $name ) {
 432                                 $magic = MagicWord::get( $name );
 433                                 $case = intval( $magic->isCaseSensitive() );
 434                                 foreach ( $magic->getSynonyms() as $syn ) {
 435                                         if ( !$case ) {
 436                                                 $syn = $wgContLang->lc( $syn );
 437                                         }
 438                                         $this->hash[$case][$syn] = $name;
 439                                 }
 440                         }
 441                 }
 442                 return $this->hash;
 443         }
 444
 445         /**
 446          * Get the base regex
 447          */
 448         function getBaseRegex() {
 449                 if ( is_null( $this->baseRegex ) ) {
 450                         $this->baseRegex = array( 0 => '', 1 => '' );
 451                         foreach ( $this->names as $name ) {
 452                                 $magic = MagicWord::get( $name );
 453                                 $case = intval( $magic->isCaseSensitive() );
 454                                 foreach ( $magic->getSynonyms() as $i => $syn ) {
 455                                         $group = "(?P<{$i}_{$name}>" . preg_quote( $syn, '/' ) . ')';
 456                                         if ( $this->baseRegex[$case] === '' ) {
 457                                                 $this->baseRegex[$case] = $group;
 458                                         } else {
 459                                                 $this->baseRegex[$case] .= '|' . $group;
 460                                         }
 461                                 }
 462                         }
 463                 }
 464                 return $this->baseRegex;
 465         }
 466
 467         /**
 468          * Get an unanchored regex
 469          */
 470         function getRegex() {
 471                 if ( is_null( $this->regex ) ) {
 472                         $base = $this->getBaseRegex();
 473                         $this->regex = array( '', '' );
 474                         if ( $this->baseRegex[0] !== '' ) {
 475                                 $this->regex[0] = "/{$base[0]}/iuS";
 476                         }
 477                         if ( $this->baseRegex[1] !== '' ) {
 478                                 $this->regex[1] = "/{$base[1]}/S";
 479                         }
 480                 }
 481                 return $this->regex;
 482         }
 483
 484         /**
 485          * Get a regex for matching variables
 486          */
 487         function getVariableRegex() {
 488                 return str_replace( "\\$1", "(.*?)", $this->getRegex() );
 489         }
 490
 491         /**
 492          * Get an anchored regex for matching variables
 493          */
 494         function getVariableStartToEndRegex() {
 495                 $base = $this->getBaseRegex();
 496                 $newRegex = array( '', '' );
 497                 if ( $base[0] !== '' ) {
 498                         $newRegex[0] = str_replace( "\\$1", "(.*?)", "/^(?:{$base[0]})$/iuS" );
 499                 }
 500                 if ( $base[1] !== '' ) {
 501                         $newRegex[1] = str_replace( "\\$1", "(.*?)", "/^(?:{$base[1]})$/S" );
 502                 }
 503                 return $newRegex;
 504         }
 505
 506         /**
 507          * Parse a match array from preg_match
 508          */
 509         function parseMatch( $m ) {
 510                 reset( $m );
 511                 while ( list( $key, $value ) = each( $m ) ) {
 512                         if ( $key === 0 || $value === '' ) {
 513                                 continue;
 514                         }
 515                         $parts = explode( '_', $key, 2 );
 516                         if ( count( $parts ) != 2 ) {
 517                                 // This shouldn't happen
 518                                 // continue;
 519                                 throw new MWException( __METHOD__ . ': bad parameter name' );
 520                         }
 521                         list( /* $synIndex */, $magicName ) = $parts;
 522                         $paramValue = next( $m );
 523                         return array( $magicName, $paramValue );
 524                 }
 525                 // This shouldn't happen either
 526                 throw new MWException( __METHOD__.': parameter not found' );
 527                 return array( false, false );
 528         }
 529
 530         /**
 531          * Match some text, with parameter capture
 532          * Returns an array with the magic word name in the first element and the
 533          * parameter in the second element.
 534          * Both elements are false if there was no match.
 535          */
 536         public function matchVariableStartToEnd( $text ) {
 537                 global $wgContLang;
 538                 $regexes = $this->getVariableStartToEndRegex();
 539                 foreach ( $regexes as $regex ) {
 540                         if ( $regex !== '' ) {
 541                                 $m = false;
 542                                 if ( preg_match( $regex, $text, $m ) ) {
 543                                         return $this->parseMatch( $m );
 544                                 }
 545                         }
 546                 }
 547                 return array( false, false );
 548         }
 549
 550         /**
 551          * Match some text, without parameter capture
 552          * Returns the magic word name, or false if there was no capture
 553          */
 554         public function matchStartToEnd( $text ) {
 555                 $hash = $this->getHash();
 556                 if ( isset( $hash[1][$text] ) ) {
 557                         return $hash[1][$text];
 558                 }
 559                 global $wgContLang;
 560                 $lc = $wgContLang->lc( $text );
 561                 if ( isset( $hash[0][$lc] ) ) {
 562                         return $hash[0][$lc];
 563                 }
 564                 return false;
 565         }
 566 }