includes/parser/LinkHolderArray.php

   1 <?php
   2 /**
   3  * Holder of replacement pairs for wiki links
   4  *
   5  * This program is free software; you can redistribute it and/or modify
   6  * it under the terms of the GNU General Public License as published by
   7  * the Free Software Foundation; either version 2 of the License, or
   8  * (at your option) any later version.
   9  *
  10  * This program is distributed in the hope that it will be useful,
  11  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  12  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  13  * GNU General Public License for more details.
  14  *
  15  * You should have received a copy of the GNU General Public License along
  16  * with this program; if not, write to the Free Software Foundation, Inc.,
  17  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  18  * http://www.gnu.org/copyleft/gpl.html
  19  *
  20  * @file
  21  * @ingroup Parser
  22  */
  23
  24 /**
  25  * @ingroup Parser
  26  */
  27 class LinkHolderArray {
  28         var $internals = array(), $interwikis = array();
  29         var $size = 0;
  30         var $parent;
  31         protected $tempIdOffset;
  32
  33         function __construct( $parent ) {
  34                 $this->parent = $parent;
  35         }
  36
  37         /**
  38          * Reduce memory usage to reduce the impact of circular references
  39          */
  40         function __destruct() {
  41                 foreach ( $this as $name => $value ) {
  42                         unset( $this->$name );
  43                 }
  44         }
  45
  46         /**
  47          * Don't serialize the parent object, it is big, and not needed when it is
  48          * a parameter to mergeForeign(), which is the only application of
  49          * serializing at present.
  50          *
  51          * Compact the titles, only serialize the text form.
  52           * @return array
  53           */
  54         function __sleep() {
  55                 foreach ( $this->internals as &$nsLinks ) {
  56                         foreach ( $nsLinks as &$entry ) {
  57                                 unset( $entry['title'] );
  58                         }
  59                 }
  60                 unset( $nsLinks );
  61                 unset( $entry );
  62
  63                 foreach ( $this->interwikis as &$entry ) {
  64                         unset( $entry['title'] );
  65                 }
  66                 unset( $entry );
  67
  68                 return array( 'internals', 'interwikis', 'size' );
  69         }
  70
  71         /**
  72          * Recreate the Title objects
  73          */
  74         function __wakeup() {
  75                 foreach ( $this->internals as &$nsLinks ) {
  76                         foreach ( $nsLinks as &$entry ) {
  77                                 $entry['title'] = Title::newFromText( $entry['pdbk'] );
  78                         }
  79                 }
  80                 unset( $nsLinks );
  81                 unset( $entry );
  82
  83                 foreach ( $this->interwikis as &$entry ) {
  84                         $entry['title'] = Title::newFromText( $entry['pdbk'] );
  85                 }
  86                 unset( $entry );
  87         }
  88
  89         /**
  90          * Merge another LinkHolderArray into this one
  91          * @param $other LinkHolderArray
  92          */
  93         function merge( $other ) {
  94                 foreach ( $other->internals as $ns => $entries ) {
  95                         $this->size += count( $entries );
  96                         if ( !isset( $this->internals[$ns] ) ) {
  97                                 $this->internals[$ns] = $entries;
  98                         } else {
  99                                 $this->internals[$ns] += $entries;
 100                         }
 101                 }
 102                 $this->interwikis += $other->interwikis;
 103         }
 104
 105         /**
 106          * Merge a LinkHolderArray from another parser instance into this one. The
 107          * keys will not be preserved. Any text which went with the old
 108          * LinkHolderArray and needs to work with the new one should be passed in
 109          * the $texts array. The strings in this array will have their link holders
 110          * converted for use in the destination link holder. The resulting array of
 111          * strings will be returned.
 112          *
 113          * @param $other LinkHolderArray
 114          * @param $texts Array of strings
 115          * @return Array
 116          */
 117         function mergeForeign( $other, $texts ) {
 118                 $this->tempIdOffset = $idOffset = $this->parent->nextLinkID();
 119                 $maxId = 0;
 120
 121                 # Renumber internal links
 122                 foreach ( $other->internals as $ns => $nsLinks ) {
 123                         foreach ( $nsLinks as $key => $entry ) {
 124                                 $newKey = $idOffset + $key;
 125                                 $this->internals[$ns][$newKey] = $entry;
 126                                 $maxId = $newKey > $maxId ? $newKey : $maxId;
 127                         }
 128                 }
 129                 $texts = preg_replace_callback( '/(<!--LINK \d+:)(\d+)(-->)/',
 130                         array( $this, 'mergeForeignCallback' ), $texts );
 131
 132                 # Renumber interwiki links
 133                 foreach ( $other->interwikis as $key => $entry ) {
 134                         $newKey = $idOffset + $key;
 135                         $this->interwikis[$newKey] = $entry;
 136                         $maxId = $newKey > $maxId ? $newKey : $maxId;
 137                 }
 138                 $texts = preg_replace_callback( '/(<!--IWLINK )(\d+)(-->)/',
 139                         array( $this, 'mergeForeignCallback' ), $texts );
 140
 141                 # Set the parent link ID to be beyond the highest used ID
 142                 $this->parent->setLinkID( $maxId + 1 );
 143                 $this->tempIdOffset = null;
 144                 return $texts;
 145         }
 146
 147         protected function mergeForeignCallback( $m ) {
 148                 return $m[1] . ( $m[2] + $this->tempIdOffset ) . $m[3];
 149         }
 150
 151         /**
 152          * Get a subset of the current LinkHolderArray which is sufficient to
 153          * interpret the given text.
 154          * @return LinkHolderArray
 155          */
 156         function getSubArray( $text ) {
 157                 $sub = new LinkHolderArray( $this->parent );
 158
 159                 # Internal links
 160                 $pos = 0;
 161                 while ( $pos < strlen( $text ) ) {
 162                         if ( !preg_match( '/<!--LINK (\d+):(\d+)-->/',
 163                                 $text, $m, PREG_OFFSET_CAPTURE, $pos ) )
 164                         {
 165                                 break;
 166                         }
 167                         $ns = $m[1][0];
 168                         $key = $m[2][0];
 169                         $sub->internals[$ns][$key] = $this->internals[$ns][$key];
 170                         $pos = $m[0][1] + strlen( $m[0][0] );
 171                 }
 172
 173                 # Interwiki links
 174                 $pos = 0;
 175                 while ( $pos < strlen( $text ) ) {
 176                         if ( !preg_match( '/<!--IWLINK (\d+)-->/', $text, $m, PREG_OFFSET_CAPTURE, $pos ) ) {
 177                                 break;
 178                         }
 179                         $key = $m[1][0];
 180                         $sub->interwikis[$key] = $this->interwikis[$key];
 181                         $pos = $m[0][1] + strlen( $m[0][0] );
 182                 }
 183                 return $sub;
 184         }
 185
 186         /**
 187          * Returns true if the memory requirements of this object are getting large
 188          * @return bool
 189          */
 190         function isBig() {
 191                 global $wgLinkHolderBatchSize;
 192                 return $this->size > $wgLinkHolderBatchSize;
 193         }
 194
 195         /**
 196          * Clear all stored link holders.
 197          * Make sure you don't have any text left using these link holders, before you call this
 198          */
 199         function clear() {
 200                 $this->internals = array();
 201                 $this->interwikis = array();
 202                 $this->size = 0;
 203         }
 204
 205         /**
 206          * Make a link placeholder. The text returned can be later resolved to a real link with
 207          * replaceLinkHolders(). This is done for two reasons: firstly to avoid further
 208          * parsing of interwiki links, and secondly to allow all existence checks and
 209          * article length checks (for stub links) to be bundled into a single query.
 210          *
 211          * @param $nt Title
 212          * @return string
 213          */
 214         function makeHolder( $nt, $text = '', $query = array(), $trail = '', $prefix = ''  ) {
 215                 wfProfileIn( __METHOD__ );
 216                 if ( ! is_object($nt) ) {
 217                         # Fail gracefully
 218                         $retVal = "<!-- ERROR -->{$prefix}{$text}{$trail}";
 219                 } else {
 220                         # Separate the link trail from the rest of the link
 221                         list( $inside, $trail ) = Linker::splitTrail( $trail );
 222
 223                         $entry = array(
 224                                 'title' => $nt,
 225                                 'text' => $prefix.$text.$inside,
 226                                 'pdbk' => $nt->getPrefixedDBkey(),
 227                         );
 228                         if ( $query !== array() ) {
 229                                 $entry['query'] = $query;
 230                         }
 231
 232                         if ( $nt->isExternal() ) {
 233                                 // Use a globally unique ID to keep the objects mergable
 234                                 $key = $this->parent->nextLinkID();
 235                                 $this->interwikis[$key] = $entry;
 236                                 $retVal = "<!--IWLINK $key-->{$trail}";
 237                         } else {
 238                                 $key = $this->parent->nextLinkID();
 239                                 $ns = $nt->getNamespace();
 240                                 $this->internals[$ns][$key] = $entry;
 241                                 $retVal = "<!--LINK $ns:$key-->{$trail}";
 242                         }
 243                         $this->size++;
 244                 }
 245                 wfProfileOut( __METHOD__ );
 246                 return $retVal;
 247         }
 248
 249         /**
 250          * @todo FIXME: Update documentation. makeLinkObj() is deprecated.
 251          * Replace <!--LINK--> link placeholders with actual links, in the buffer
 252          * Placeholders created in Skin::makeLinkObj()
 253          * Returns an array of link CSS classes, indexed by PDBK.
 254          */
 255         function replace( &$text ) {
 256                 wfProfileIn( __METHOD__ );
 257
 258                 $colours = $this->replaceInternal( $text );
 259                 $this->replaceInterwiki( $text );
 260
 261                 wfProfileOut( __METHOD__ );
 262                 return $colours;
 263         }
 264
 265         /**
 266          * Replace internal links
 267          */
 268         protected function replaceInternal( &$text ) {
 269                 if ( !$this->internals ) {
 270                         return;
 271                 }
 272
 273                 wfProfileIn( __METHOD__ );
 274                 global $wgContLang;
 275
 276                 $colours = array();
 277                 $linkCache = LinkCache::singleton();
 278                 $output = $this->parent->getOutput();
 279
 280                 wfProfileIn( __METHOD__.'-check' );
 281                 $dbr = wfGetDB( DB_SLAVE );
 282                 $threshold = $this->parent->getOptions()->getStubThreshold();
 283
 284                 # Sort by namespace
 285                 ksort( $this->internals );
 286
 287                 $linkcolour_ids = array();
 288
 289                 # Generate query
 290                 $queries = array();
 291                 foreach ( $this->internals as $ns => $entries ) {
 292                         foreach ( $entries as $entry ) {
 293                                 $title = $entry['title'];
 294                                 $pdbk = $entry['pdbk'];
 295
 296                                 # Skip invalid entries.
 297                                 # Result will be ugly, but prevents crash.
 298                                 if ( is_null( $title ) ) {
 299                                         continue;
 300                                 }
 301
 302                                 # Check if it's a static known link, e.g. interwiki
 303                                 if ( $title->isAlwaysKnown() ) {
 304                                         $colours[$pdbk] = '';
 305                                 } elseif ( $ns == NS_SPECIAL ) {
 306                                         $colours[$pdbk] = 'new';
 307                                 } elseif ( ( $id = $linkCache->getGoodLinkID( $pdbk ) ) != 0 ) {
 308                                         $colours[$pdbk] = Linker::getLinkColour( $title, $threshold );
 309                                         $output->addLink( $title, $id );
 310                                         $linkcolour_ids[$id] = $pdbk;
 311                                 } elseif ( $linkCache->isBadLink( $pdbk ) ) {
 312                                         $colours[$pdbk] = 'new';
 313                                 } else {
 314                                         # Not in the link cache, add it to the query
 315                                         $queries[$ns][] = $title->getDBkey();
 316                                 }
 317                         }
 318                 }
 319                 if ( $queries ) {
 320                         $where = array();
 321                         foreach( $queries as $ns => $pages ){
 322                                 $where[] = $dbr->makeList(
 323                                         array(
 324                                                 'page_namespace' => $ns,
 325                                                 'page_title' => $pages,
 326                                         ),
 327                                         LIST_AND
 328                                 );
 329                         }
 330
 331                         $res = $dbr->select(
 332                                 'page',
 333                                 array( 'page_id', 'page_namespace', 'page_title', 'page_is_redirect', 'page_len', 'page_latest' ),
 334                                 $dbr->makeList( $where, LIST_OR ),
 335                                 __METHOD__
 336                         );
 337
 338                         # Fetch data and form into an associative array
 339                         # non-existent = broken
 340                         foreach ( $res as $s ) {
 341                                 $title = Title::makeTitle( $s->page_namespace, $s->page_title );
 342                                 $pdbk = $title->getPrefixedDBkey();
 343                                 $linkCache->addGoodLinkObjFromRow( $title, $s );
 344                                 $output->addLink( $title, $s->page_id );
 345                                 # @todo FIXME: Convoluted data flow
 346                                 # The redirect status and length is passed to getLinkColour via the LinkCache
 347                                 # Use formal parameters instead
 348                                 $colours[$pdbk] = Linker::getLinkColour( $title, $threshold );
 349                                 //add id to the extension todolist
 350                                 $linkcolour_ids[$s->page_id] = $pdbk;
 351                         }
 352                         unset( $res );
 353                 }
 354                 if ( count($linkcolour_ids) ) {
 355                         //pass an array of page_ids to an extension
 356                         wfRunHooks( 'GetLinkColours', array( $linkcolour_ids, &$colours ) );
 357                 }
 358                 wfProfileOut( __METHOD__.'-check' );
 359
 360                 # Do a second query for different language variants of links and categories
 361                 if($wgContLang->hasVariants()) {
 362                         $this->doVariants( $colours );
 363                 }
 364
 365                 # Construct search and replace arrays
 366                 wfProfileIn( __METHOD__.'-construct' );
 367                 $replacePairs = array();
 368                 foreach ( $this->internals as $ns => $entries ) {
 369                         foreach ( $entries as $index => $entry ) {
 370                                 $pdbk = $entry['pdbk'];
 371                                 $title = $entry['title'];
 372                                 $query = isset( $entry['query'] ) ? $entry['query'] : array();
 373                                 $key = "$ns:$index";
 374                                 $searchkey = "<!--LINK $key-->";
 375                                 $displayText = $entry['text'];
 376                                 if ( $displayText === '' ) {
 377                                         $displayText = null;
 378                                 }
 379                                 if ( !isset( $colours[$pdbk] ) ) {
 380                                         $colours[$pdbk] = 'new';
 381                                 }
 382                                 $attribs = array();
 383                                 if ( $colours[$pdbk] == 'new' ) {
 384                                         $linkCache->addBadLinkObj( $title );
 385                                         $output->addLink( $title, 0 );
 386                                         $type = array( 'broken' );
 387                                 } else {
 388                                         if ( $colours[$pdbk] != '' ) {
 389                                                 $attribs['class'] = $colours[$pdbk];
 390                                         }
 391                                         $type = array( 'known', 'noclasses' );
 392                                 }
 393                                 $replacePairs[$searchkey] = Linker::link( $title, $displayText,
 394                                                 $attribs, $query, $type );
 395                         }
 396                 }
 397                 $replacer = new HashtableReplacer( $replacePairs, 1 );
 398                 wfProfileOut( __METHOD__.'-construct' );
 399
 400                 # Do the thing
 401                 wfProfileIn( __METHOD__.'-replace' );
 402                 $text = preg_replace_callback(
 403                         '/(<!--LINK .*?-->)/',
 404                         $replacer->cb(),
 405                         $text);
 406
 407                 wfProfileOut( __METHOD__.'-replace' );
 408                 wfProfileOut( __METHOD__ );
 409         }
 410
 411         /**
 412          * Replace interwiki links
 413          */
 414         protected function replaceInterwiki( &$text ) {
 415                 if ( empty( $this->interwikis ) ) {
 416                         return;
 417                 }
 418
 419                 wfProfileIn( __METHOD__ );
 420                 # Make interwiki link HTML
 421                 $output = $this->parent->getOutput();
 422                 $replacePairs = array();
 423                 foreach( $this->interwikis as $key => $link ) {
 424                         $replacePairs[$key] = Linker::link( $link['title'], $link['text'] );
 425                         $output->addInterwikiLink( $link['title'] );
 426                 }
 427                 $replacer = new HashtableReplacer( $replacePairs, 1 );
 428
 429                 $text = preg_replace_callback(
 430                         '/<!--IWLINK (.*?)-->/',
 431                         $replacer->cb(),
 432                         $text );
 433                 wfProfileOut( __METHOD__ );
 434         }
 435
 436         /**
 437          * Modify $this->internals and $colours according to language variant linking rules
 438          */
 439         protected function doVariants( &$colours ) {
 440                 global $wgContLang;
 441                 $linkBatch = new LinkBatch();
 442                 $variantMap = array(); // maps $pdbkey_Variant => $keys (of link holders)
 443                 $output = $this->parent->getOutput();
 444                 $linkCache = LinkCache::singleton();
 445                 $threshold = $this->parent->getOptions()->getStubThreshold();
 446                 $titlesToBeConverted = '';
 447                 $titlesAttrs = array();
 448
 449                 // Concatenate titles to a single string, thus we only need auto convert the
 450                 // single string to all variants. This would improve parser's performance
 451                 // significantly.
 452                 foreach ( $this->internals as $ns => $entries ) {
 453                         foreach ( $entries as $index => $entry ) {
 454                                 $pdbk = $entry['pdbk'];
 455                                 // we only deal with new links (in its first query)
 456                                 if ( !isset( $colours[$pdbk] ) ) {
 457                                         $title = $entry['title'];
 458                                         $titleText = $title->getText();
 459                                         $titlesAttrs[] = array(
 460                                                 'ns' => $ns,
 461                                                 'key' => "$ns:$index",
 462                                                 'titleText' => $titleText,
 463                                         );
 464                                         // separate titles with \0 because it would never appears
 465                                         // in a valid title
 466                                         $titlesToBeConverted .= $titleText . "\0";
 467                                 }
 468                         }
 469                 }
 470
 471                 // Now do the conversion and explode string to text of titles
 472                 $titlesAllVariants = $wgContLang->autoConvertToAllVariants( $titlesToBeConverted );
 473                 $allVariantsName = array_keys( $titlesAllVariants );
 474                 foreach ( $titlesAllVariants as &$titlesVariant ) {
 475                         $titlesVariant = explode( "\0", $titlesVariant );
 476                 }
 477                 $l = count( $titlesAttrs );
 478                 // Then add variants of links to link batch
 479                 for ( $i = 0; $i < $l; $i ++ ) {
 480                         foreach ( $allVariantsName as $variantName ) {
 481                                 $textVariant = $titlesAllVariants[$variantName][$i];
 482                                 if ( $textVariant != $titlesAttrs[$i]['titleText'] ) {
 483                                         $variantTitle = Title::makeTitle( $titlesAttrs[$i]['ns'], $textVariant );
 484                                         if( is_null( $variantTitle ) ) {
 485                                                 continue;
 486                                         }
 487                                         $linkBatch->addObj( $variantTitle );
 488                                         $variantMap[$variantTitle->getPrefixedDBkey()][] = $titlesAttrs[$i]['key'];
 489                                 }
 490                         }
 491                 }
 492
 493                 // process categories, check if a category exists in some variant
 494                 $categoryMap = array(); // maps $category_variant => $category (dbkeys)
 495                 $varCategories = array(); // category replacements oldDBkey => newDBkey
 496                 foreach( $output->getCategoryLinks() as $category ){
 497                         $variants = $wgContLang->autoConvertToAllVariants( $category );
 498                         foreach($variants as $variant){
 499                                 if($variant != $category){
 500                                         $variantTitle = Title::newFromDBkey( Title::makeName(NS_CATEGORY,$variant) );
 501                                         if(is_null($variantTitle)) continue;
 502                                         $linkBatch->addObj( $variantTitle );
 503                                         $categoryMap[$variant] = $category;
 504                                 }
 505                         }
 506                 }
 507
 508
 509                 if(!$linkBatch->isEmpty()){
 510                         // construct query
 511                         $dbr = wfGetDB( DB_SLAVE );
 512                         $varRes = $dbr->select( 'page',
 513                                 array( 'page_id', 'page_namespace', 'page_title', 'page_is_redirect', 'page_len', 'page_latest' ),
 514                                 $linkBatch->constructSet( 'page', $dbr ),
 515                                 __METHOD__
 516                         );
 517
 518                         $linkcolour_ids = array();
 519
 520                         // for each found variants, figure out link holders and replace
 521                         foreach ( $varRes as $s ) {
 522
 523                                 $variantTitle = Title::makeTitle( $s->page_namespace, $s->page_title );
 524                                 $varPdbk = $variantTitle->getPrefixedDBkey();
 525                                 $vardbk = $variantTitle->getDBkey();
 526
 527                                 $holderKeys = array();
 528                                 if( isset( $variantMap[$varPdbk] ) ) {
 529                                         $holderKeys = $variantMap[$varPdbk];
 530                                         $linkCache->addGoodLinkObjFromRow( $variantTitle, $s );
 531                                         $output->addLink( $variantTitle, $s->page_id );
 532                                 }
 533
 534                                 // loop over link holders
 535                                 foreach( $holderKeys as $key ) {
 536                                         list( $ns, $index ) = explode( ':', $key, 2 );
 537                                         $entry =& $this->internals[$ns][$index];
 538                                         $pdbk = $entry['pdbk'];
 539
 540                                         if(!isset($colours[$pdbk])){
 541                                                 // found link in some of the variants, replace the link holder data
 542                                                 $entry['title'] = $variantTitle;
 543                                                 $entry['pdbk'] = $varPdbk;
 544
 545                                                 // set pdbk and colour
 546                                                 # @todo FIXME: Convoluted data flow
 547                                                 # The redirect status and length is passed to getLinkColour via the LinkCache
 548                                                 # Use formal parameters instead
 549                                                 $colours[$varPdbk] = Linker::getLinkColour( $variantTitle, $threshold );
 550                                                 $linkcolour_ids[$s->page_id] = $pdbk;
 551                                         }
 552                                 }
 553
 554                                 // check if the object is a variant of a category
 555                                 if(isset($categoryMap[$vardbk])){
 556                                         $oldkey = $categoryMap[$vardbk];
 557                                         if($oldkey != $vardbk)
 558                                                 $varCategories[$oldkey]=$vardbk;
 559                                 }
 560                         }
 561                         wfRunHooks( 'GetLinkColours', array( $linkcolour_ids, &$colours ) );
 562
 563                         // rebuild the categories in original order (if there are replacements)
 564                         if(count($varCategories)>0){
 565                                 $newCats = array();
 566                                 $originalCats = $output->getCategories();
 567                                 foreach($originalCats as $cat => $sortkey){
 568                                         // make the replacement
 569                                         if( array_key_exists($cat,$varCategories) )
 570                                                 $newCats[$varCategories[$cat]] = $sortkey;
 571                                         else $newCats[$cat] = $sortkey;
 572                                 }
 573                                 $output->setCategoryLinks($newCats);
 574                         }
 575                 }
 576         }
 577
 578         /**
 579          * Replace <!--LINK--> link placeholders with plain text of links
 580          * (not HTML-formatted).
 581          *
 582          * @param $text String
 583          * @return String
 584          */
 585         function replaceText( $text ) {
 586                 wfProfileIn( __METHOD__ );
 587
 588                 $text = preg_replace_callback(
 589                         '/<!--(LINK|IWLINK) (.*?)-->/',
 590                         array( &$this, 'replaceTextCallback' ),
 591                         $text );
 592
 593                 wfProfileOut( __METHOD__ );
 594                 return $text;
 595         }
 596
 597         /**
 598          * Callback for replaceText()
 599          *
 600          * @param $matches Array
 601          * @return string
 602          * @private
 603          */
 604         function replaceTextCallback( $matches ) {
 605                 $type = $matches[1];
 606                 $key  = $matches[2];
 607                 if( $type == 'LINK' ) {
 608                         list( $ns, $index ) = explode( ':', $key, 2 );
 609                         if( isset( $this->internals[$ns][$index]['text'] ) ) {
 610                                 return $this->internals[$ns][$index]['text'];
 611                         }
 612                 } elseif( $type == 'IWLINK' ) {
 613                         if( isset( $this->interwikis[$key]['text'] ) ) {
 614                                 return $this->interwikis[$key]['text'];
 615                         }
 616                 }
 617                 return $matches[0];
 618         }
 619 }