Code Coverage |
||||||||||
Lines |
Functions and Methods |
Classes and Traits |
||||||||
| Total | |
75.57% |
167 / 221 |
|
35.71% |
5 / 14 |
CRAP | |
0.00% |
0 / 1 |
| BacklinkCache | |
75.91% |
167 / 220 |
|
35.71% |
5 / 14 |
107.67 | |
0.00% |
0 / 1 |
| __construct | |
100.00% |
8 / 8 |
|
100.00% |
1 / 1 |
1 | |||
| getPage | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| getDB | |
100.00% |
1 / 1 |
|
100.00% |
1 / 1 |
1 | |||
| getLinkPages | |
100.00% |
6 / 6 |
|
100.00% |
1 / 1 |
4 | |||
| queryLinks | |
90.91% |
20 / 22 |
|
0.00% |
0 / 1 |
12.11 | |||
| getPrefix | |
64.29% |
9 / 14 |
|
0.00% |
0 / 1 |
3.41 | |||
| initQueryBuilderForTable | |
65.96% |
31 / 47 |
|
0.00% |
0 / 1 |
19.67 | |||
| hasLinks | |
70.00% |
7 / 10 |
|
0.00% |
0 / 1 |
4.43 | |||
| getNumLinks | |
84.21% |
16 / 19 |
|
0.00% |
0 / 1 |
3.04 | |||
| partition | |
76.92% |
30 / 39 |
|
0.00% |
0 / 1 |
4.20 | |||
| openBatchEnds | |
100.00% |
4 / 4 |
|
100.00% |
1 / 1 |
2 | |||
| partitionResult | |
93.33% |
14 / 15 |
|
0.00% |
0 / 1 |
5.01 | |||
| getCascadeProtectedLinkPages | |
33.33% |
1 / 3 |
|
0.00% |
0 / 1 |
3.19 | |||
| getCascadeProtectedLinksInternal | |
61.29% |
19 / 31 |
|
0.00% |
0 / 1 |
4.93 | |||
| 1 | <?php |
| 2 | /** |
| 3 | * Class for fetching backlink lists, approximate backlink counts and |
| 4 | * partitions. |
| 5 | * |
| 6 | * This program is free software; you can redistribute it and/or modify |
| 7 | * it under the terms of the GNU General Public License as published by |
| 8 | * the Free Software Foundation; either version 2 of the License, or |
| 9 | * (at your option) any later version. |
| 10 | * |
| 11 | * This program is distributed in the hope that it will be useful, |
| 12 | * but WITHOUT ANY WARRANTY; without even the implied warranty of |
| 13 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
| 14 | * GNU General Public License for more details. |
| 15 | * |
| 16 | * You should have received a copy of the GNU General Public License along |
| 17 | * with this program; if not, write to the Free Software Foundation, Inc., |
| 18 | * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA. |
| 19 | * http://www.gnu.org/copyleft/gpl.html |
| 20 | * |
| 21 | * @file |
| 22 | * @author Tim Starling |
| 23 | * @copyright © 2009, Tim Starling, Domas Mituzas |
| 24 | * @copyright © 2010, Max Sem |
| 25 | * @copyright © 2011, Antoine Musso |
| 26 | */ |
| 27 | |
| 28 | namespace MediaWiki\Cache; |
| 29 | |
| 30 | use Iterator; |
| 31 | use LogicException; |
| 32 | use MediaWiki\Config\ServiceOptions; |
| 33 | use MediaWiki\HookContainer\HookContainer; |
| 34 | use MediaWiki\HookContainer\HookRunner; |
| 35 | use MediaWiki\Linker\LinksMigration; |
| 36 | use MediaWiki\MainConfigNames; |
| 37 | use MediaWiki\Page\PageIdentity; |
| 38 | use MediaWiki\Page\PageIdentityValue; |
| 39 | use MediaWiki\Page\PageReference; |
| 40 | use MediaWiki\Title\Title; |
| 41 | use MediaWiki\Title\TitleValue; |
| 42 | use Psr\Log\LoggerInterface; |
| 43 | use RuntimeException; |
| 44 | use stdClass; |
| 45 | use Wikimedia\ObjectCache\WANObjectCache; |
| 46 | use Wikimedia\Rdbms\Database; |
| 47 | use Wikimedia\Rdbms\IConnectionProvider; |
| 48 | use Wikimedia\Rdbms\IReadableDatabase; |
| 49 | use Wikimedia\Rdbms\IResultWrapper; |
| 50 | use Wikimedia\Rdbms\SelectQueryBuilder; |
| 51 | |
| 52 | /** |
| 53 | * Class for fetching backlink lists, approximate backlink counts and |
| 54 | * partitions. This is a shared cache. |
| 55 | * |
| 56 | * Instances of this class should typically be fetched with the method |
| 57 | * ::getBacklinkCache() from the BacklinkCacheFactory service. |
| 58 | * |
| 59 | * Ideally you should only get your backlinks from here when you think |
| 60 | * there is some advantage in caching them. Otherwise, it's just a waste |
| 61 | * of memory. |
| 62 | */ |
| 63 | class BacklinkCache { |
| 64 | /** |
| 65 | * @internal Used by ServiceWiring.php |
| 66 | */ |
| 67 | public const CONSTRUCTOR_OPTIONS = [ |
| 68 | MainConfigNames::UpdateRowsPerJob, |
| 69 | ]; |
| 70 | |
| 71 | /** |
| 72 | * Multi-dimensional array representing batches. Keys are: |
| 73 | * > (string) links table name |
| 74 | * > (int) batch size |
| 75 | * > 'numRows' : Number of rows for this link table |
| 76 | * > 'batches' : [ [ $start, $end ] ] |
| 77 | * |
| 78 | * @see BacklinkCache::partitionResult() |
| 79 | * @var array[] |
| 80 | */ |
| 81 | private $partitionCache = []; |
| 82 | |
| 83 | /** |
| 84 | * Contains the whole links from a database result. |
| 85 | * This is raw data that will be partitioned in $partitionCache |
| 86 | * |
| 87 | * Initialized with BacklinkCache::queryLinks() |
| 88 | * |
| 89 | * @var IResultWrapper[] |
| 90 | */ |
| 91 | private $fullResultCache = []; |
| 92 | |
| 93 | /** |
| 94 | * Cache for hasLinks() |
| 95 | * |
| 96 | * @var bool[] |
| 97 | */ |
| 98 | private $hasLinksCache = []; |
| 99 | |
| 100 | /** @var WANObjectCache */ |
| 101 | private $wanCache; |
| 102 | |
| 103 | /** @var HookRunner */ |
| 104 | private $hookRunner; |
| 105 | |
| 106 | /** |
| 107 | * Local copy of a PageReference object |
| 108 | * @var PageReference |
| 109 | */ |
| 110 | private $page; |
| 111 | |
| 112 | private const CACHE_EXPIRY = 3600; |
| 113 | private IConnectionProvider $dbProvider; |
| 114 | private ServiceOptions $options; |
| 115 | private LinksMigration $linksMigration; |
| 116 | private LoggerInterface $logger; |
| 117 | |
| 118 | /** |
| 119 | * Create a new BacklinkCache |
| 120 | * |
| 121 | * @param ServiceOptions $options |
| 122 | * @param LinksMigration $linksMigration |
| 123 | * @param WANObjectCache $wanCache |
| 124 | * @param HookContainer $hookContainer |
| 125 | * @param IConnectionProvider $dbProvider |
| 126 | * @param LoggerInterface $logger |
| 127 | * @param PageReference $page Page to create a backlink cache for |
| 128 | */ |
| 129 | public function __construct( |
| 130 | ServiceOptions $options, |
| 131 | LinksMigration $linksMigration, |
| 132 | WANObjectCache $wanCache, |
| 133 | HookContainer $hookContainer, |
| 134 | IConnectionProvider $dbProvider, |
| 135 | LoggerInterface $logger, |
| 136 | PageReference $page |
| 137 | ) { |
| 138 | $options->assertRequiredOptions( self::CONSTRUCTOR_OPTIONS ); |
| 139 | $this->options = $options; |
| 140 | $this->linksMigration = $linksMigration; |
| 141 | $this->wanCache = $wanCache; |
| 142 | $this->hookRunner = new HookRunner( $hookContainer ); |
| 143 | $this->dbProvider = $dbProvider; |
| 144 | $this->logger = $logger; |
| 145 | $this->page = $page; |
| 146 | } |
| 147 | |
| 148 | /** |
| 149 | * @since 1.37 |
| 150 | * @return PageReference |
| 151 | */ |
| 152 | public function getPage(): PageReference { |
| 153 | return $this->page; |
| 154 | } |
| 155 | |
| 156 | /** |
| 157 | * Get the replica DB connection to the database |
| 158 | * |
| 159 | * @return IReadableDatabase |
| 160 | */ |
| 161 | private function getDB() { |
| 162 | return $this->dbProvider->getReplicaDatabase(); |
| 163 | } |
| 164 | |
| 165 | /** |
| 166 | * Get the backlinks for a given table. Cached in process memory only. |
| 167 | * @param string $table |
| 168 | * @param int|bool $startId |
| 169 | * @param int|bool $endId |
| 170 | * @param int|float $max Integer, or INF for no max |
| 171 | * @return Iterator<PageIdentity> |
| 172 | * @since 1.37 |
| 173 | */ |
| 174 | public function getLinkPages( |
| 175 | string $table, $startId = false, $endId = false, $max = INF |
| 176 | ): Iterator { |
| 177 | $i = 0; |
| 178 | foreach ( $this->queryLinks( $table, $startId, $endId, $max ) as $row ) { |
| 179 | yield PageIdentityValue::localIdentity( |
| 180 | $row->page_id, $row->page_namespace, $row->page_title ); |
| 181 | |
| 182 | // queryLinks() may return too many rows |
| 183 | if ( is_finite( $max ) && ++$i >= $max ) { |
| 184 | break; |
| 185 | } |
| 186 | } |
| 187 | } |
| 188 | |
| 189 | /** |
| 190 | * Get the backlinks for a given table. Cached in process memory only. |
| 191 | * |
| 192 | * @param string $table |
| 193 | * @param int|bool $startId |
| 194 | * @param int|bool $endId |
| 195 | * @param int|float $max A hint for the maximum number of rows to return. |
| 196 | * May return more rows if there is a previously cached result set. |
| 197 | * @param string $select 'all' or 'ids' |
| 198 | * @return IResultWrapper |
| 199 | */ |
| 200 | private function queryLinks( $table, $startId, $endId, $max, $select = 'all' ) { |
| 201 | if ( !$startId && !$endId && isset( $this->fullResultCache[$table] ) ) { |
| 202 | $this->logger->debug( __METHOD__ . ': got results from cache' ); |
| 203 | return $this->fullResultCache[$table]; |
| 204 | } |
| 205 | |
| 206 | $this->logger->debug( __METHOD__ . ': got results from DB' ); |
| 207 | $queryBuilder = $this->initQueryBuilderForTable( $table, $select ); |
| 208 | $fromField = $this->getPrefix( $table ) . '_from'; |
| 209 | // Use the from field in the condition rather than the joined page_id, |
| 210 | // because databases are stupid and don't necessarily propagate indexes. |
| 211 | if ( $startId ) { |
| 212 | $queryBuilder->where( |
| 213 | $this->getDB()->expr( $fromField, '>=', $startId ) |
| 214 | ); |
| 215 | } |
| 216 | if ( $endId ) { |
| 217 | $queryBuilder->where( |
| 218 | $this->getDB()->expr( $fromField, '<=', $endId ) |
| 219 | ); |
| 220 | } |
| 221 | $queryBuilder->orderBy( $fromField ); |
| 222 | if ( is_finite( $max ) && $max > 0 ) { |
| 223 | $queryBuilder->limit( $max ); |
| 224 | } |
| 225 | |
| 226 | $res = $queryBuilder->caller( __METHOD__ )->fetchResultSet(); |
| 227 | |
| 228 | if ( $select === 'all' && !$startId && !$endId && $res->numRows() < $max ) { |
| 229 | // The full results fit within the limit, so cache them |
| 230 | $this->fullResultCache[$table] = $res; |
| 231 | } else { |
| 232 | $this->logger->debug( __METHOD__ . ": results from DB were uncacheable" ); |
| 233 | } |
| 234 | |
| 235 | return $res; |
| 236 | } |
| 237 | |
| 238 | /** |
| 239 | * Get the field name prefix for a given table |
| 240 | * @param string $table |
| 241 | * @return null|string |
| 242 | */ |
| 243 | private function getPrefix( $table ) { |
| 244 | static $prefixes = [ |
| 245 | 'pagelinks' => 'pl', |
| 246 | 'imagelinks' => 'il', |
| 247 | 'categorylinks' => 'cl', |
| 248 | 'templatelinks' => 'tl', |
| 249 | 'redirect' => 'rd', |
| 250 | ]; |
| 251 | |
| 252 | if ( isset( $prefixes[$table] ) ) { |
| 253 | return $prefixes[$table]; |
| 254 | } else { |
| 255 | $prefix = null; |
| 256 | $this->hookRunner->onBacklinkCacheGetPrefix( $table, $prefix ); |
| 257 | if ( $prefix ) { |
| 258 | return $prefix; |
| 259 | } else { |
| 260 | throw new LogicException( "Invalid table \"$table\" in " . __CLASS__ ); |
| 261 | } |
| 262 | } |
| 263 | } |
| 264 | |
| 265 | /** |
| 266 | * Initialize a new SelectQueryBuilder for selecting backlinks, |
| 267 | * with a join on the page table if needed. |
| 268 | * |
| 269 | * @param string $table |
| 270 | * @param string $select |
| 271 | * @return SelectQueryBuilder |
| 272 | */ |
| 273 | private function initQueryBuilderForTable( string $table, string $select ): SelectQueryBuilder { |
| 274 | $prefix = $this->getPrefix( $table ); |
| 275 | $queryBuilder = $this->getDB()->newSelectQueryBuilder(); |
| 276 | $joinPageTable = $select !== 'ids'; |
| 277 | |
| 278 | if ( $select === 'ids' ) { |
| 279 | $queryBuilder->select( [ 'page_id' => $prefix . '_from' ] ); |
| 280 | } else { |
| 281 | $queryBuilder->select( [ 'page_namespace', 'page_title', 'page_id' ] ); |
| 282 | } |
| 283 | $queryBuilder->from( $table ); |
| 284 | |
| 285 | /* |
| 286 | * If the table is one of the tables known to this method, |
| 287 | * we can use a nice join() method later, always joining on page_id={$prefix}_from. |
| 288 | * If the table is unknown here, and only supported via a hook, |
| 289 | * the hook only produces a single $conds array, |
| 290 | * so we have to use a traditional / ANSI-89 JOIN, |
| 291 | * with the page table just added to the list of tables and the join conds in the WHERE part. |
| 292 | */ |
| 293 | $knownTable = true; |
| 294 | |
| 295 | switch ( $table ) { |
| 296 | case 'pagelinks': |
| 297 | case 'templatelinks': |
| 298 | $queryBuilder->where( |
| 299 | $this->linksMigration->getLinksConditions( $table, TitleValue::newFromPage( $this->page ) ) |
| 300 | ); |
| 301 | break; |
| 302 | case 'redirect': |
| 303 | $queryBuilder->where( [ |
| 304 | "{$prefix}_namespace" => $this->page->getNamespace(), |
| 305 | "{$prefix}_title" => $this->page->getDBkey(), |
| 306 | "{$prefix}_interwiki" => [ '', null ], |
| 307 | ] ); |
| 308 | break; |
| 309 | case 'imagelinks': |
| 310 | case 'categorylinks': |
| 311 | $queryBuilder->where( [ |
| 312 | "{$prefix}_to" => $this->page->getDBkey(), |
| 313 | ] ); |
| 314 | break; |
| 315 | default: |
| 316 | $knownTable = false; |
| 317 | $conds = null; |
| 318 | $this->hookRunner->onBacklinkCacheGetConditions( $table, |
| 319 | Title::newFromPageReference( $this->page ), |
| 320 | $conds |
| 321 | ); |
| 322 | if ( !$conds ) { |
| 323 | throw new LogicException( "Invalid table \"$table\" in " . __CLASS__ ); |
| 324 | } |
| 325 | if ( $joinPageTable ) { |
| 326 | $queryBuilder->table( 'page' ); // join condition in $conds |
| 327 | } else { |
| 328 | // remove any page_id condition from $conds |
| 329 | $conds = array_filter( (array)$conds, static function ( $clause ) { // kind of janky |
| 330 | return !preg_match( '/(\b|=)page_id(\b|=)/', (string)$clause ); |
| 331 | } ); |
| 332 | } |
| 333 | $queryBuilder->where( $conds ); |
| 334 | break; |
| 335 | } |
| 336 | |
| 337 | if ( $knownTable && $joinPageTable ) { |
| 338 | $queryBuilder->join( 'page', null, "page_id={$prefix}_from" ); |
| 339 | } |
| 340 | if ( $joinPageTable ) { |
| 341 | $queryBuilder->straightJoinOption(); |
| 342 | } |
| 343 | |
| 344 | return $queryBuilder; |
| 345 | } |
| 346 | |
| 347 | /** |
| 348 | * Check if there are any backlinks. Only use the process cache, since the |
| 349 | * WAN cache is potentially stale (T368006). |
| 350 | * |
| 351 | * @param string $table |
| 352 | * @return bool |
| 353 | */ |
| 354 | public function hasLinks( $table ) { |
| 355 | if ( isset( $this->hasLinksCache[$table] ) ) { |
| 356 | return $this->hasLinksCache[$table]; |
| 357 | } |
| 358 | if ( isset( $this->partitionCache[$table] ) ) { |
| 359 | $entry = reset( $this->partitionCache[$table] ); |
| 360 | return (bool)$entry['numRows']; |
| 361 | } |
| 362 | if ( isset( $this->fullResultCache[$table] ) ) { |
| 363 | return (bool)$this->fullResultCache[$table]->numRows(); |
| 364 | } |
| 365 | $hasLinks = (bool)$this->queryLinks( $table, false, false, 1 )->numRows(); |
| 366 | $this->hasLinksCache[$table] = $hasLinks; |
| 367 | return $hasLinks; |
| 368 | } |
| 369 | |
| 370 | /** |
| 371 | * Get the approximate number of backlinks |
| 372 | * @param string $table |
| 373 | * @return int |
| 374 | */ |
| 375 | public function getNumLinks( $table ) { |
| 376 | if ( isset( $this->partitionCache[$table] ) ) { |
| 377 | $entry = reset( $this->partitionCache[$table] ); |
| 378 | return $entry['numRows']; |
| 379 | } |
| 380 | |
| 381 | if ( isset( $this->fullResultCache[$table] ) ) { |
| 382 | return $this->fullResultCache[$table]->numRows(); |
| 383 | } |
| 384 | |
| 385 | return $this->wanCache->getWithSetCallback( |
| 386 | $this->wanCache->makeKey( |
| 387 | 'numbacklinks', |
| 388 | CacheKeyHelper::getKeyForPage( $this->page ), |
| 389 | $table |
| 390 | ), |
| 391 | self::CACHE_EXPIRY, |
| 392 | function ( $oldValue, &$ttl, array &$setOpts ) use ( $table ) { |
| 393 | $setOpts += Database::getCacheSetOptions( $this->getDB() ); |
| 394 | |
| 395 | // Use partition() since it will batch the query and skip the JOIN. |
| 396 | // Use $wgUpdateRowsPerJob just to encourage cache reuse for jobs. |
| 397 | $batchSize = $this->options->get( MainConfigNames::UpdateRowsPerJob ); |
| 398 | $this->partition( $table, $batchSize ); |
| 399 | return $this->partitionCache[$table][$batchSize]['numRows']; |
| 400 | } |
| 401 | ); |
| 402 | } |
| 403 | |
| 404 | /** |
| 405 | * Partition the backlinks into batches. |
| 406 | * Returns an array giving the start and end of each range. The first |
| 407 | * batch has a start of false, and the last batch has an end of false. |
| 408 | * |
| 409 | * @param string $table The links table name |
| 410 | * @param int $batchSize |
| 411 | * @return array |
| 412 | */ |
| 413 | public function partition( $table, $batchSize ) { |
| 414 | if ( isset( $this->partitionCache[$table][$batchSize] ) ) { |
| 415 | $this->logger->debug( __METHOD__ . ": got from partition cache" ); |
| 416 | |
| 417 | return $this->partitionCache[$table][$batchSize]['batches']; |
| 418 | } |
| 419 | |
| 420 | $this->partitionCache[$table][$batchSize] = false; |
| 421 | $cacheEntry =& $this->partitionCache[$table][$batchSize]; |
| 422 | |
| 423 | if ( isset( $this->fullResultCache[$table] ) ) { |
| 424 | $res = $this->fullResultCache[$table]; |
| 425 | $numRows = $res->numRows(); |
| 426 | $batches = $this->partitionResult( $res, $numRows, $batchSize ); |
| 427 | $this->openBatchEnds( $batches ); |
| 428 | $cacheEntry = [ 'numRows' => $numRows, 'batches' => $batches ]; |
| 429 | $this->logger->debug( __METHOD__ . ": got from full result cache" ); |
| 430 | |
| 431 | return $cacheEntry['batches']; |
| 432 | } |
| 433 | |
| 434 | $cacheEntry = $this->wanCache->getWithSetCallback( |
| 435 | $this->wanCache->makeKey( |
| 436 | 'backlinks', |
| 437 | CacheKeyHelper::getKeyForPage( $this->page ), |
| 438 | $table, |
| 439 | $batchSize |
| 440 | ), |
| 441 | self::CACHE_EXPIRY, |
| 442 | function ( $oldValue, &$ttl, array &$setOpts ) use ( $table, $batchSize ) { |
| 443 | $setOpts += Database::getCacheSetOptions( $this->getDB() ); |
| 444 | |
| 445 | $value = [ 'numRows' => 0, 'batches' => [] ]; |
| 446 | |
| 447 | // Do the selects in batches to avoid client-side OOMs (T45452). |
| 448 | // Use a LIMIT that plays well with $batchSize to keep equal sized partitions. |
| 449 | $selectSize = max( $batchSize, 200_000 - ( 200_000 % $batchSize ) ); |
| 450 | $start = false; |
| 451 | do { |
| 452 | $res = $this->queryLinks( $table, $start, false, $selectSize, 'ids' ); |
| 453 | $numRows = $res->numRows(); |
| 454 | $batches = $this->partitionResult( $res, $numRows, $batchSize ); |
| 455 | // Merge the link count and range partitions for this chunk |
| 456 | $value['numRows'] += $numRows; |
| 457 | $value['batches'] = array_merge( $value['batches'], $batches ); |
| 458 | if ( count( $batches ) ) { |
| 459 | // pick up after this inclusive range |
| 460 | $start = end( $batches )[1] + 1; |
| 461 | } |
| 462 | } while ( $numRows >= $selectSize ); |
| 463 | // Make sure the first range has start=false and the last one has end=false |
| 464 | $this->openBatchEnds( $value['batches'] ); |
| 465 | |
| 466 | return $value; |
| 467 | } |
| 468 | ); |
| 469 | |
| 470 | return $cacheEntry['batches']; |
| 471 | } |
| 472 | |
| 473 | /** |
| 474 | * Modify an array of batches, setting the start of the first batch to |
| 475 | * false, and the end of the last batch to false, so that the complete |
| 476 | * set of batches covers the entire ID range from 0 to infinity. |
| 477 | * |
| 478 | * @param array &$batches |
| 479 | */ |
| 480 | private function openBatchEnds( array &$batches ) { |
| 481 | if ( !count( $batches ) ) { |
| 482 | $batches = [ [ false, false ] ]; |
| 483 | } else { |
| 484 | $batches[0][0] = false; |
| 485 | $batches[ array_key_last( $batches ) ][1] = false; |
| 486 | } |
| 487 | } |
| 488 | |
| 489 | /** |
| 490 | * Partition a DB result with backlinks in it into batches |
| 491 | * @param IResultWrapper $res Database result |
| 492 | * @param int $numRows The number of rows to use from the result set |
| 493 | * @param int $batchSize |
| 494 | * @return int[][] |
| 495 | */ |
| 496 | private function partitionResult( $res, $numRows, $batchSize ) { |
| 497 | $numBatches = ceil( $numRows / $batchSize ); |
| 498 | $batches = []; |
| 499 | for ( $i = 0; $i < $numBatches; $i++ ) { |
| 500 | $rowNum = $i * $batchSize; |
| 501 | $res->seek( $rowNum ); |
| 502 | $row = $res->fetchObject(); |
| 503 | $start = (int)$row->page_id; |
| 504 | |
| 505 | $rowNum = min( $numRows - 1, ( $i + 1 ) * $batchSize - 1 ); |
| 506 | $res->seek( $rowNum ); |
| 507 | $row = $res->fetchObject(); |
| 508 | $end = (int)$row->page_id; |
| 509 | |
| 510 | // Check order |
| 511 | if ( $start && $end && $start > $end ) { |
| 512 | throw new RuntimeException( __METHOD__ . ': Internal error: query result out of order' ); |
| 513 | } |
| 514 | |
| 515 | $batches[] = [ $start, $end ]; |
| 516 | } |
| 517 | |
| 518 | return $batches; |
| 519 | } |
| 520 | |
| 521 | /** |
| 522 | * Get a PageIdentity iterator for cascade-protected template/file use backlinks |
| 523 | * |
| 524 | * @return Iterator<PageIdentity> |
| 525 | * @since 1.37 |
| 526 | */ |
| 527 | public function getCascadeProtectedLinkPages(): Iterator { |
| 528 | foreach ( $this->getCascadeProtectedLinksInternal() as $row ) { |
| 529 | yield PageIdentityValue::localIdentity( |
| 530 | $row->page_id, $row->page_namespace, $row->page_title ); |
| 531 | } |
| 532 | } |
| 533 | |
| 534 | /** |
| 535 | * Get an array of cascade-protected template/file use backlinks |
| 536 | * |
| 537 | * @return stdClass[] |
| 538 | */ |
| 539 | private function getCascadeProtectedLinksInternal(): array { |
| 540 | $dbr = $this->getDB(); |
| 541 | |
| 542 | // @todo: use UNION without breaking tests that use temp tables |
| 543 | $resSets = []; |
| 544 | $linkConds = $this->linksMigration->getLinksConditions( |
| 545 | 'templatelinks', TitleValue::newFromPage( $this->page ) |
| 546 | ); |
| 547 | $resSets[] = $dbr->newSelectQueryBuilder() |
| 548 | ->select( [ 'page_namespace', 'page_title', 'page_id' ] ) |
| 549 | ->from( 'templatelinks' ) |
| 550 | ->join( 'page_restrictions', null, 'tl_from = pr_page' ) |
| 551 | ->join( 'page', null, 'page_id = tl_from' ) |
| 552 | ->where( $linkConds ) |
| 553 | ->andWhere( [ 'pr_cascade' => 1 ] ) |
| 554 | ->distinct() |
| 555 | ->caller( __METHOD__ )->fetchResultSet(); |
| 556 | if ( $this->page->getNamespace() === NS_FILE ) { |
| 557 | $resSets[] = $dbr->newSelectQueryBuilder() |
| 558 | ->select( [ 'page_namespace', 'page_title', 'page_id' ] ) |
| 559 | ->from( 'imagelinks' ) |
| 560 | ->join( 'page_restrictions', null, 'il_from = pr_page' ) |
| 561 | ->join( 'page', null, 'page_id = il_from' ) |
| 562 | ->where( [ |
| 563 | 'il_to' => $this->page->getDBkey(), |
| 564 | 'pr_cascade' => 1, |
| 565 | ] ) |
| 566 | ->distinct() |
| 567 | ->caller( __METHOD__ )->fetchResultSet(); |
| 568 | } |
| 569 | |
| 570 | // Combine and de-duplicate the results |
| 571 | $mergedRes = []; |
| 572 | foreach ( $resSets as $res ) { |
| 573 | foreach ( $res as $row ) { |
| 574 | // Index by page_id to remove duplicates |
| 575 | $mergedRes[$row->page_id] = $row; |
| 576 | } |
| 577 | } |
| 578 | |
| 579 | // Now that we've de-duplicated, throw away the keys |
| 580 | return array_values( $mergedRes ); |
| 581 | } |
| 582 | } |
| 583 | |
| 584 | /** @deprecated class alias since 1.42 */ |
| 585 | class_alias( BacklinkCache::class, 'BacklinkCache' ); |