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