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