includes/api/ApiQueryAllRevisions.php

   1 <?php
   2 /**
   3  * Copyright © 2015 Wikimedia Foundation and contributors
   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  */
  22
  23 use MediaWiki\MediaWikiServices;
  24 use MediaWiki\Revision\RevisionRecord;
  25
  26 /**
  27  * Query module to enumerate all revisions.
  28  *
  29  * @ingroup API
  30  * @since 1.27
  31  */
  32 class ApiQueryAllRevisions extends ApiQueryRevisionsBase {
  33
  34         public function __construct( ApiQuery $query, $moduleName ) {
  35                 parent::__construct( $query, $moduleName, 'arv' );
  36         }
  37
  38         /**
  39          * @param ApiPageSet|null $resultPageSet
  40          * @return void
  41          */
  42         protected function run( ApiPageSet $resultPageSet = null ) {
  43                 global $wgActorTableSchemaMigrationStage;
  44
  45                 $db = $this->getDB();
  46                 $params = $this->extractRequestParams( false );
  47                 $revisionStore = MediaWikiServices::getInstance()->getRevisionStore();
  48
  49                 $result = $this->getResult();
  50
  51                 $this->requireMaxOneParameter( $params, 'user', 'excludeuser' );
  52
  53                 $tsField = 'rev_timestamp';
  54                 $idField = 'rev_id';
  55                 $pageField = 'rev_page';
  56                 if ( $params['user'] !== null &&
  57                         ( $wgActorTableSchemaMigrationStage & SCHEMA_COMPAT_READ_NEW )
  58                 ) {
  59                         // The query is probably best done using the actor_timestamp index on
  60                         // revision_actor_temp. Use the denormalized fields from that table.
  61                         $tsField = 'revactor_timestamp';
  62                         $idField = 'revactor_rev';
  63                         $pageField = 'revactor_page';
  64                 }
  65
  66                 // Namespace check is likely to be desired, but can't be done
  67                 // efficiently in SQL.
  68                 $miser_ns = null;
  69                 $needPageTable = false;
  70                 if ( $params['namespace'] !== null ) {
  71                         $params['namespace'] = array_unique( $params['namespace'] );
  72                         sort( $params['namespace'] );
  73                         if ( $params['namespace'] != MWNamespace::getValidNamespaces() ) {
  74                                 $needPageTable = true;
  75                                 if ( $this->getConfig()->get( 'MiserMode' ) ) {
  76                                         $miser_ns = $params['namespace'];
  77                                 } else {
  78                                         $this->addWhere( [ 'page_namespace' => $params['namespace'] ] );
  79                                 }
  80                         }
  81                 }
  82
  83                 if ( $resultPageSet === null ) {
  84                         $this->parseParameters( $params );
  85                         $revQuery = $revisionStore->getQueryInfo(
  86                                 $this->fetchContent ? [ 'page', 'text' ] : [ 'page' ]
  87                         );
  88                 } else {
  89                         $this->limit = $this->getParameter( 'limit' ) ?: 10;
  90                         $revQuery = [
  91                                 'tables' => [ 'revision' ],
  92                                 'fields' => [ 'rev_timestamp', 'rev_id' ],
  93                                 'joins' => [],
  94                         ];
  95
  96                         if ( $params['generatetitles'] ) {
  97                                 $revQuery['fields'][] = 'rev_page';
  98                         }
  99
 100                         if ( $params['user'] !== null || $params['excludeuser'] !== null ) {
 101                                 $actorQuery = ActorMigration::newMigration()->getJoin( 'rev_user' );
 102                                 $revQuery['tables'] += $actorQuery['tables'];
 103                                 $revQuery['joins'] += $actorQuery['joins'];
 104                         }
 105
 106                         if ( $needPageTable ) {
 107                                 $revQuery['tables'][] = 'page';
 108                                 $revQuery['joins']['page'] = [ 'INNER JOIN', [ "$pageField = page_id" ] ];
 109                                 if ( (bool)$miser_ns ) {
 110                                         $revQuery['fields'][] = 'page_namespace';
 111                                 }
 112                         }
 113                 }
 114
 115                 // If we're going to be using actor_timestamp, we need to swap the order of `revision`
 116                 // and `revision_actor_temp` in the query (for the straight join) and adjust some field aliases.
 117                 if ( $idField !== 'rev_id' && isset( $revQuery['tables']['temp_rev_user'] ) ) {
 118                         $aliasFields = [ 'rev_id' => $idField, 'rev_timestamp' => $tsField, 'rev_page' => $pageField ];
 119                         $revQuery['fields'] = array_merge(
 120                                 $aliasFields,
 121                                 array_diff( $revQuery['fields'], array_keys( $aliasFields ) )
 122                         );
 123                         unset( $revQuery['tables']['temp_rev_user'] );
 124                         $revQuery['tables'] = array_merge(
 125                                 [ 'temp_rev_user' => 'revision_actor_temp' ],
 126                                 $revQuery['tables']
 127                         );
 128                         $revQuery['joins']['revision'] = $revQuery['joins']['temp_rev_user'];
 129                         unset( $revQuery['joins']['temp_rev_user'] );
 130                 }
 131
 132                 $this->addTables( $revQuery['tables'] );
 133                 $this->addFields( $revQuery['fields'] );
 134                 $this->addJoinConds( $revQuery['joins'] );
 135
 136                 // Seems to be needed to avoid a planner bug (T113901)
 137                 $this->addOption( 'STRAIGHT_JOIN' );
 138
 139                 $dir = $params['dir'];
 140                 $this->addTimestampWhereRange( $tsField, $dir, $params['start'], $params['end'] );
 141
 142                 if ( $this->fld_tags ) {
 143                         $this->addTables( 'tag_summary' );
 144                         $this->addJoinConds(
 145                                 [ 'tag_summary' => [ 'LEFT JOIN', [ 'rev_id=ts_rev_id' ] ] ]
 146                         );
 147                         $this->addFields( 'ts_tags' );
 148                 }
 149
 150                 if ( $params['user'] !== null ) {
 151                         $actorQuery = ActorMigration::newMigration()
 152                                 ->getWhere( $db, 'rev_user', User::newFromName( $params['user'], false ) );
 153                         $this->addWhere( $actorQuery['conds'] );
 154                 } elseif ( $params['excludeuser'] !== null ) {
 155                         $actorQuery = ActorMigration::newMigration()
 156                                 ->getWhere( $db, 'rev_user', User::newFromName( $params['excludeuser'], false ) );
 157                         $this->addWhere( 'NOT(' . $actorQuery['conds'] . ')' );
 158                 }
 159
 160                 if ( $params['user'] !== null || $params['excludeuser'] !== null ) {
 161                         // Paranoia: avoid brute force searches (T19342)
 162                         if ( !$this->getUser()->isAllowed( 'deletedhistory' ) ) {
 163                                 $bitmask = RevisionRecord::DELETED_USER;
 164                         } elseif ( !$this->getUser()->isAllowedAny( 'suppressrevision', 'viewsuppressed' ) ) {
 165                                 $bitmask = RevisionRecord::DELETED_USER | RevisionRecord::DELETED_RESTRICTED;
 166                         } else {
 167                                 $bitmask = 0;
 168                         }
 169                         if ( $bitmask ) {
 170                                 $this->addWhere( $db->bitAnd( 'rev_deleted', $bitmask ) . " != $bitmask" );
 171                         }
 172                 }
 173
 174                 if ( $params['continue'] !== null ) {
 175                         $op = ( $dir == 'newer' ? '>' : '<' );
 176                         $cont = explode( '|', $params['continue'] );
 177                         $this->dieContinueUsageIf( count( $cont ) != 2 );
 178                         $ts = $db->addQuotes( $db->timestamp( $cont[0] ) );
 179                         $rev_id = (int)$cont[1];
 180                         $this->dieContinueUsageIf( strval( $rev_id ) !== $cont[1] );
 181                         $this->addWhere( "$tsField $op $ts OR " .
 182                                 "($tsField = $ts AND " .
 183                                 "$idField $op= $rev_id)" );
 184                 }
 185
 186                 $this->addOption( 'LIMIT', $this->limit + 1 );
 187
 188                 $sort = ( $dir == 'newer' ? '' : ' DESC' );
 189                 $orderby = [];
 190                 // Targeting index rev_timestamp, user_timestamp, usertext_timestamp, or actor_timestamp.
 191                 // But 'user' is always constant for the latter three, so it doesn't matter here.
 192                 $orderby[] = "rev_timestamp $sort";
 193                 $orderby[] = "rev_id $sort";
 194                 $this->addOption( 'ORDER BY', $orderby );
 195
 196                 $hookData = [];
 197                 $res = $this->select( __METHOD__, [], $hookData );
 198                 $pageMap = []; // Maps rev_page to array index
 199                 $count = 0;
 200                 $nextIndex = 0;
 201                 $generated = [];
 202                 foreach ( $res as $row ) {
 203                         if ( $count === 0 && $resultPageSet !== null ) {
 204                                 // Set the non-continue since the list of all revisions is
 205                                 // prone to having entries added at the start frequently.
 206                                 $this->getContinuationManager()->addGeneratorNonContinueParam(
 207                                         $this, 'continue', "$row->rev_timestamp|$row->rev_id"
 208                                 );
 209                         }
 210                         if ( ++$count > $this->limit ) {
 211                                 // We've had enough
 212                                 $this->setContinueEnumParameter( 'continue', "$row->rev_timestamp|$row->rev_id" );
 213                                 break;
 214                         }
 215
 216                         // Miser mode namespace check
 217                         if ( $miser_ns !== null && !in_array( $row->page_namespace, $miser_ns ) ) {
 218                                 continue;
 219                         }
 220
 221                         if ( $resultPageSet !== null ) {
 222                                 if ( $params['generatetitles'] ) {
 223                                         $generated[$row->rev_page] = $row->rev_page;
 224                                 } else {
 225                                         $generated[] = $row->rev_id;
 226                                 }
 227                         } else {
 228                                 $revision = $revisionStore->newRevisionFromRow( $row );
 229                                 $rev = $this->extractRevisionInfo( $revision, $row );
 230
 231                                 if ( !isset( $pageMap[$row->rev_page] ) ) {
 232                                         $index = $nextIndex++;
 233                                         $pageMap[$row->rev_page] = $index;
 234                                         $title = Title::newFromLinkTarget( $revision->getPageAsLinkTarget() );
 235                                         $a = [
 236                                                 'pageid' => $title->getArticleID(),
 237                                                 'revisions' => [ $rev ],
 238                                         ];
 239                                         ApiResult::setIndexedTagName( $a['revisions'], 'rev' );
 240                                         ApiQueryBase::addTitleInfo( $a, $title );
 241                                         $fit = $this->processRow( $row, $a['revisions'][0], $hookData ) &&
 242                                                 $result->addValue( [ 'query', $this->getModuleName() ], $index, $a );
 243                                 } else {
 244                                         $index = $pageMap[$row->rev_page];
 245                                         $fit = $this->processRow( $row, $rev, $hookData ) &&
 246                                                 $result->addValue( [ 'query', $this->getModuleName(), $index, 'revisions' ], null, $rev );
 247                                 }
 248                                 if ( !$fit ) {
 249                                         $this->setContinueEnumParameter( 'continue', "$row->rev_timestamp|$row->rev_id" );
 250                                         break;
 251                                 }
 252                         }
 253                 }
 254
 255                 if ( $resultPageSet !== null ) {
 256                         if ( $params['generatetitles'] ) {
 257                                 $resultPageSet->populateFromPageIDs( $generated );
 258                         } else {
 259                                 $resultPageSet->populateFromRevisionIDs( $generated );
 260                         }
 261                 } else {
 262                         $result->addIndexedTagName( [ 'query', $this->getModuleName() ], 'page' );
 263                 }
 264         }
 265
 266         public function getAllowedParams() {
 267                 $ret = parent::getAllowedParams() + [
 268                         'user' => [
 269                                 ApiBase::PARAM_TYPE => 'user',
 270                         ],
 271                         'namespace' => [
 272                                 ApiBase::PARAM_ISMULTI => true,
 273                                 ApiBase::PARAM_TYPE => 'namespace',
 274                                 ApiBase::PARAM_DFLT => null,
 275                         ],
 276                         'start' => [
 277                                 ApiBase::PARAM_TYPE => 'timestamp',
 278                         ],
 279                         'end' => [
 280                                 ApiBase::PARAM_TYPE => 'timestamp',
 281                         ],
 282                         'dir' => [
 283                                 ApiBase::PARAM_TYPE => [
 284                                         'newer',
 285                                         'older'
 286                                 ],
 287                                 ApiBase::PARAM_DFLT => 'older',
 288                                 ApiBase::PARAM_HELP_MSG => 'api-help-param-direction',
 289                         ],
 290                         'excludeuser' => [
 291                                 ApiBase::PARAM_TYPE => 'user',
 292                         ],
 293                         'continue' => [
 294                                 ApiBase::PARAM_HELP_MSG => 'api-help-param-continue',
 295                         ],
 296                         'generatetitles' => [
 297                                 ApiBase::PARAM_DFLT => false,
 298                         ],
 299                 ];
 300
 301                 if ( $this->getConfig()->get( 'MiserMode' ) ) {
 302                         $ret['namespace'][ApiBase::PARAM_HELP_MSG_APPEND] = [
 303                                 'api-help-param-limited-in-miser-mode',
 304                         ];
 305                 }
 306
 307                 return $ret;
 308         }
 309
 310         protected function getExamplesMessages() {
 311                 return [
 312                         'action=query&list=allrevisions&arvuser=Example&arvlimit=50'
 313                                 => 'apihelp-query+allrevisions-example-user',
 314                         'action=query&list=allrevisions&arvdir=newer&arvlimit=50'
 315                                 => 'apihelp-query+allrevisions-example-ns-main',
 316                 ];
 317         }
 318
 319         public function getHelpUrls() {
 320                 return 'https://www.mediawiki.org/wiki/Special:MyLanguage/API:Allrevisions';
 321         }
 322 }