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