includes/api/ApiQueryDeletedRevisions.php

   1 <?php
   2 /**
   3  * Copyright © 2014 Wikimedia Foundation and contributors
   4  *
   5  * Heavily based on ApiQueryDeletedrevs,
   6  * Copyright © 2007 Roan Kattouw "<Firstname>.<Lastname>@gmail.com"
   7  *
   8  * This program is free software; you can redistribute it and/or modify
   9  * it under the terms of the GNU General Public License as published by
  10  * the Free Software Foundation; either version 2 of the License, or
  11  * (at your option) any later version.
  12  *
  13  * This program is distributed in the hope that it will be useful,
  14  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  15  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  16  * GNU General Public License for more details.
  17  *
  18  * You should have received a copy of the GNU General Public License along
  19  * with this program; if not, write to the Free Software Foundation, Inc.,
  20  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  21  * http://www.gnu.org/copyleft/gpl.html
  22  *
  23  * @file
  24  */
  25
  26 /**
  27  * Query module to enumerate deleted revisions for pages.
  28  *
  29  * @ingroup API
  30  */
  31 class ApiQueryDeletedRevisions extends ApiQueryRevisionsBase {
  32
  33         public function __construct( ApiQuery $query, $moduleName ) {
  34                 parent::__construct( $query, $moduleName, 'drv' );
  35         }
  36
  37         protected function run( ApiPageSet $resultPageSet = null ) {
  38                 $user = $this->getUser();
  39                 // Before doing anything at all, let's check permissions
  40                 $this->checkUserRightsAny( 'deletedhistory' );
  41
  42                 $pageSet = $this->getPageSet();
  43                 $pageMap = $pageSet->getGoodAndMissingTitlesByNamespace();
  44                 $pageCount = count( $pageSet->getGoodAndMissingTitles() );
  45                 $revCount = $pageSet->getRevisionCount();
  46                 if ( $revCount === 0 && $pageCount === 0 ) {
  47                         // Nothing to do
  48                         return;
  49                 }
  50                 if ( $revCount !== 0 && count( $pageSet->getDeletedRevisionIDs() ) === 0 ) {
  51                         // Nothing to do, revisions were supplied but none are deleted
  52                         return;
  53                 }
  54
  55                 $params = $this->extractRequestParams( false );
  56
  57                 $db = $this->getDB();
  58
  59                 $this->requireMaxOneParameter( $params, 'user', 'excludeuser' );
  60
  61                 if ( $resultPageSet === null ) {
  62                         $this->parseParameters( $params );
  63                         $arQuery = Revision::getArchiveQueryInfo();
  64                         $this->addTables( $arQuery['tables'] );
  65                         $this->addFields( $arQuery['fields'] );
  66                         $this->addJoinConds( $arQuery['joins'] );
  67                         $this->addFields( [ 'ar_title', 'ar_namespace' ] );
  68                 } else {
  69                         $this->limit = $this->getParameter( 'limit' ) ?: 10;
  70                         $this->addTables( 'archive' );
  71                         $this->addFields( [ 'ar_title', 'ar_namespace', 'ar_timestamp', 'ar_rev_id', 'ar_id' ] );
  72                 }
  73
  74                 if ( $this->fld_tags ) {
  75                         $this->addTables( 'tag_summary' );
  76                         $this->addJoinConds(
  77                                 [ 'tag_summary' => [ 'LEFT JOIN', [ 'ar_rev_id=ts_rev_id' ] ] ]
  78                         );
  79                         $this->addFields( 'ts_tags' );
  80                 }
  81
  82                 if ( !is_null( $params['tag'] ) ) {
  83                         $this->addTables( 'change_tag' );
  84                         $this->addJoinConds(
  85                                 [ 'change_tag' => [ 'INNER JOIN', [ 'ar_rev_id=ct_rev_id' ] ] ]
  86                         );
  87                         $this->addWhereFld( 'ct_tag', $params['tag'] );
  88                 }
  89
  90                 if ( $this->fetchContent ) {
  91                         // Modern MediaWiki has the content for deleted revs in the 'text'
  92                         // table using fields old_text and old_flags. But revisions deleted
  93                         // pre-1.5 store the content in the 'archive' table directly using
  94                         // fields ar_text and ar_flags, and no corresponding 'text' row. So
  95                         // we have to LEFT JOIN and fetch all four fields.
  96                         $this->addTables( 'text' );
  97                         $this->addJoinConds(
  98                                 [ 'text' => [ 'LEFT JOIN', [ 'ar_text_id=old_id' ] ] ]
  99                         );
 100                         $this->addFields( [ 'ar_text', 'ar_flags', 'old_text', 'old_flags' ] );
 101
 102                         // This also means stricter restrictions
 103                         $this->checkUserRightsAny( [ 'deletedtext', 'undelete' ] );
 104                 }
 105
 106                 $dir = $params['dir'];
 107
 108                 if ( $revCount !== 0 ) {
 109                         $this->addWhere( [
 110                                 'ar_rev_id' => array_keys( $pageSet->getDeletedRevisionIDs() )
 111                         ] );
 112                 } else {
 113                         // We need a custom WHERE clause that matches all titles.
 114                         $lb = new LinkBatch( $pageSet->getGoodAndMissingTitles() );
 115                         $where = $lb->constructSet( 'ar', $db );
 116                         $this->addWhere( $where );
 117                 }
 118
 119                 if ( !is_null( $params['user'] ) ) {
 120                         // Don't query by user ID here, it might be able to use the ar_usertext_timestamp index.
 121                         $actorQuery = ActorMigration::newMigration()
 122                                 ->getWhere( $db, 'ar_user', User::newFromName( $params['user'], false ), false );
 123                         $this->addTables( $actorQuery['tables'] );
 124                         $this->addJoinConds( $actorQuery['joins'] );
 125                         $this->addWhere( $actorQuery['conds'] );
 126                 } elseif ( !is_null( $params['excludeuser'] ) ) {
 127                         // Here there's no chance of using ar_usertext_timestamp.
 128                         $actorQuery = ActorMigration::newMigration()
 129                                 ->getWhere( $db, 'ar_user', User::newFromName( $params['excludeuser'], false ) );
 130                         $this->addTables( $actorQuery['tables'] );
 131                         $this->addJoinConds( $actorQuery['joins'] );
 132                         $this->addWhere( 'NOT(' . $actorQuery['conds'] . ')' );
 133                 }
 134
 135                 if ( !is_null( $params['user'] ) || !is_null( $params['excludeuser'] ) ) {
 136                         // Paranoia: avoid brute force searches (T19342)
 137                         // (shouldn't be able to get here without 'deletedhistory', but
 138                         // check it again just in case)
 139                         if ( !$user->isAllowed( 'deletedhistory' ) ) {
 140                                 $bitmask = Revision::DELETED_USER;
 141                         } elseif ( !$user->isAllowedAny( 'suppressrevision', 'viewsuppressed' ) ) {
 142                                 $bitmask = Revision::DELETED_USER | Revision::DELETED_RESTRICTED;
 143                         } else {
 144                                 $bitmask = 0;
 145                         }
 146                         if ( $bitmask ) {
 147                                 $this->addWhere( $db->bitAnd( 'ar_deleted', $bitmask ) . " != $bitmask" );
 148                         }
 149                 }
 150
 151                 if ( !is_null( $params['continue'] ) ) {
 152                         $cont = explode( '|', $params['continue'] );
 153                         $op = ( $dir == 'newer' ? '>' : '<' );
 154                         if ( $revCount !== 0 ) {
 155                                 $this->dieContinueUsageIf( count( $cont ) != 2 );
 156                                 $rev = intval( $cont[0] );
 157                                 $this->dieContinueUsageIf( strval( $rev ) !== $cont[0] );
 158                                 $ar_id = (int)$cont[1];
 159                                 $this->dieContinueUsageIf( strval( $ar_id ) !== $cont[1] );
 160                                 $this->addWhere( "ar_rev_id $op $rev OR " .
 161                                         "(ar_rev_id = $rev AND " .
 162                                         "ar_id $op= $ar_id)" );
 163                         } else {
 164                                 $this->dieContinueUsageIf( count( $cont ) != 4 );
 165                                 $ns = intval( $cont[0] );
 166                                 $this->dieContinueUsageIf( strval( $ns ) !== $cont[0] );
 167                                 $title = $db->addQuotes( $cont[1] );
 168                                 $ts = $db->addQuotes( $db->timestamp( $cont[2] ) );
 169                                 $ar_id = (int)$cont[3];
 170                                 $this->dieContinueUsageIf( strval( $ar_id ) !== $cont[3] );
 171                                 $this->addWhere( "ar_namespace $op $ns OR " .
 172                                         "(ar_namespace = $ns AND " .
 173                                         "(ar_title $op $title OR " .
 174                                         "(ar_title = $title AND " .
 175                                         "(ar_timestamp $op $ts OR " .
 176                                         "(ar_timestamp = $ts AND " .
 177                                         "ar_id $op= $ar_id)))))" );
 178                         }
 179                 }
 180
 181                 $this->addOption( 'LIMIT', $this->limit + 1 );
 182
 183                 if ( $revCount !== 0 ) {
 184                         // Sort by ar_rev_id when querying by ar_rev_id
 185                         $this->addWhereRange( 'ar_rev_id', $dir, null, null );
 186                 } else {
 187                         // Sort by ns and title in the same order as timestamp for efficiency
 188                         // But only when not already unique in the query
 189                         if ( count( $pageMap ) > 1 ) {
 190                                 $this->addWhereRange( 'ar_namespace', $dir, null, null );
 191                         }
 192                         $oneTitle = key( reset( $pageMap ) );
 193                         foreach ( $pageMap as $pages ) {
 194                                 if ( count( $pages ) > 1 || key( $pages ) !== $oneTitle ) {
 195                                         $this->addWhereRange( 'ar_title', $dir, null, null );
 196                                         break;
 197                                 }
 198                         }
 199                         $this->addTimestampWhereRange( 'ar_timestamp', $dir, $params['start'], $params['end'] );
 200                 }
 201                 // Include in ORDER BY for uniqueness
 202                 $this->addWhereRange( 'ar_id', $dir, null, null );
 203
 204                 $res = $this->select( __METHOD__ );
 205                 $count = 0;
 206                 $generated = [];
 207                 foreach ( $res as $row ) {
 208                         if ( ++$count > $this->limit ) {
 209                                 // We've had enough
 210                                 $this->setContinueEnumParameter( 'continue',
 211                                         $revCount
 212                                                 ? "$row->ar_rev_id|$row->ar_id"
 213                                                 : "$row->ar_namespace|$row->ar_title|$row->ar_timestamp|$row->ar_id"
 214                                 );
 215                                 break;
 216                         }
 217
 218                         if ( $resultPageSet !== null ) {
 219                                 $generated[] = $row->ar_rev_id;
 220                         } else {
 221                                 if ( !isset( $pageMap[$row->ar_namespace][$row->ar_title] ) ) {
 222                                         // Was it converted?
 223                                         $title = Title::makeTitle( $row->ar_namespace, $row->ar_title );
 224                                         $converted = $pageSet->getConvertedTitles();
 225                                         if ( $title && isset( $converted[$title->getPrefixedText()] ) ) {
 226                                                 $title = Title::newFromText( $converted[$title->getPrefixedText()] );
 227                                                 if ( $title && isset( $pageMap[$title->getNamespace()][$title->getDBkey()] ) ) {
 228                                                         $pageMap[$row->ar_namespace][$row->ar_title] =
 229                                                                 $pageMap[$title->getNamespace()][$title->getDBkey()];
 230                                                 }
 231                                         }
 232                                 }
 233                                 if ( !isset( $pageMap[$row->ar_namespace][$row->ar_title] ) ) {
 234                                         ApiBase::dieDebug(
 235                                                 __METHOD__,
 236                                                 "Found row in archive (ar_id={$row->ar_id}) that didn't get processed by ApiPageSet"
 237                                         );
 238                                 }
 239
 240                                 $fit = $this->addPageSubItem(
 241                                         $pageMap[$row->ar_namespace][$row->ar_title],
 242                                         $this->extractRevisionInfo( Revision::newFromArchiveRow( $row ), $row ),
 243                                         'rev'
 244                                 );
 245                                 if ( !$fit ) {
 246                                         $this->setContinueEnumParameter( 'continue',
 247                                                 $revCount
 248                                                         ? "$row->ar_rev_id|$row->ar_id"
 249                                                         : "$row->ar_namespace|$row->ar_title|$row->ar_timestamp|$row->ar_id"
 250                                         );
 251                                         break;
 252                                 }
 253                         }
 254                 }
 255
 256                 if ( $resultPageSet !== null ) {
 257                         $resultPageSet->populateFromRevisionIDs( $generated );
 258                 }
 259         }
 260
 261         public function getAllowedParams() {
 262                 return parent::getAllowedParams() + [
 263                         'start' => [
 264                                 ApiBase::PARAM_TYPE => 'timestamp',
 265                         ],
 266                         'end' => [
 267                                 ApiBase::PARAM_TYPE => 'timestamp',
 268                         ],
 269                         'dir' => [
 270                                 ApiBase::PARAM_TYPE => [
 271                                         'newer',
 272                                         'older'
 273                                 ],
 274                                 ApiBase::PARAM_DFLT => 'older',
 275                                 ApiBase::PARAM_HELP_MSG => 'api-help-param-direction',
 276                         ],
 277                         'tag' => null,
 278                         'user' => [
 279                                 ApiBase::PARAM_TYPE => 'user'
 280                         ],
 281                         'excludeuser' => [
 282                                 ApiBase::PARAM_TYPE => 'user'
 283                         ],
 284                         'continue' => [
 285                                 ApiBase::PARAM_HELP_MSG => 'api-help-param-continue',
 286                         ],
 287                 ];
 288         }
 289
 290         protected function getExamplesMessages() {
 291                 return [
 292                         'action=query&prop=deletedrevisions&titles=Main%20Page|Talk:Main%20Page&' .
 293                                 'drvprop=user|comment|content'
 294                                 => 'apihelp-query+deletedrevisions-example-titles',
 295                         'action=query&prop=deletedrevisions&revids=123456'
 296                                 => 'apihelp-query+deletedrevisions-example-revids',
 297                 ];
 298         }
 299
 300         public function getHelpUrls() {
 301                 return 'https://www.mediawiki.org/wiki/Special:MyLanguage/API:Deletedrevisions';
 302         }
 303 }