includes/api/ApiQueryAllDeletedRevisions.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 all deleted revisions.
  28  *
  29  * @ingroup API
  30  */
  31 class ApiQueryAllDeletedRevisions extends ApiQueryRevisionsBase {
  32
  33         public function __construct( ApiQuery $query, $moduleName ) {
  34                 parent::__construct( $query, $moduleName, 'adr' );
  35         }
  36
  37         /**
  38          * @param ApiPageSet $resultPageSet
  39          * @return void
  40          */
  41         protected function run( ApiPageSet $resultPageSet = null ) {
  42                 // Before doing anything at all, let's check permissions
  43                 $this->checkUserRightsAny( 'deletedhistory' );
  44
  45                 $user = $this->getUser();
  46                 $db = $this->getDB();
  47                 $params = $this->extractRequestParams( false );
  48
  49                 $result = $this->getResult();
  50
  51                 // If the user wants no namespaces, they get no pages.
  52                 if ( $params['namespace'] === [] ) {
  53                         if ( $resultPageSet === null ) {
  54                                 $result->addValue( 'query', $this->getModuleName(), [] );
  55                         }
  56                         return;
  57                 }
  58
  59                 // This module operates in two modes:
  60                 // 'user': List deleted revs by a certain user
  61                 // 'all': List all deleted revs in NS
  62                 $mode = 'all';
  63                 if ( !is_null( $params['user'] ) ) {
  64                         $mode = 'user';
  65                 }
  66
  67                 if ( $mode == 'user' ) {
  68                         foreach ( [ 'from', 'to', 'prefix', 'excludeuser' ] as $param ) {
  69                                 if ( !is_null( $params[$param] ) ) {
  70                                         $p = $this->getModulePrefix();
  71                                         $this->dieWithError(
  72                                                 [ 'apierror-invalidparammix-cannotusewith', $p.$param, "{$p}user" ],
  73                                                 'invalidparammix'
  74                                         );
  75                                 }
  76                         }
  77                 } else {
  78                         foreach ( [ 'start', 'end' ] as $param ) {
  79                                 if ( !is_null( $params[$param] ) ) {
  80                                         $p = $this->getModulePrefix();
  81                                         $this->dieWithError(
  82                                                 [ 'apierror-invalidparammix-mustusewith', $p.$param, "{$p}user" ],
  83                                                 'invalidparammix'
  84                                         );
  85                                 }
  86                         }
  87                 }
  88
  89                 // If we're generating titles only, we can use DISTINCT for a better
  90                 // query. But we can't do that in 'user' mode (wrong index), and we can
  91                 // only do it when sorting ASC (because MySQL apparently can't use an
  92                 // index backwards for grouping even though it can for ORDER BY, WTF?)
  93                 $dir = $params['dir'];
  94                 $optimizeGenerateTitles = false;
  95                 if ( $mode === 'all' && $params['generatetitles'] && $resultPageSet !== null ) {
  96                         if ( $dir === 'newer' ) {
  97                                 $optimizeGenerateTitles = true;
  98                         } else {
  99                                 $p = $this->getModulePrefix();
 100                                 $this->addWarning( [ 'apiwarn-alldeletedrevisions-performance', $p ], 'performance' );
 101                         }
 102                 }
 103
 104                 if ( $resultPageSet === null ) {
 105                         $this->parseParameters( $params );
 106                         $arQuery = Revision::getArchiveQueryInfo();
 107                         $this->addTables( $arQuery['tables'] );
 108                         $this->addJoinConds( $arQuery['joins'] );
 109                         $this->addFields( $arQuery['fields'] );
 110                         $this->addFields( [ 'ar_title', 'ar_namespace' ] );
 111                 } else {
 112                         $this->limit = $this->getParameter( 'limit' ) ?: 10;
 113                         $this->addTables( 'archive' );
 114                         $this->addFields( [ 'ar_title', 'ar_namespace' ] );
 115                         if ( $optimizeGenerateTitles ) {
 116                                 $this->addOption( 'DISTINCT' );
 117                         } else {
 118                                 $this->addFields( [ 'ar_timestamp', 'ar_rev_id', 'ar_id' ] );
 119                         }
 120                 }
 121
 122                 if ( $this->fld_tags ) {
 123                         $this->addTables( 'tag_summary' );
 124                         $this->addJoinConds(
 125                                 [ 'tag_summary' => [ 'LEFT JOIN', [ 'ar_rev_id=ts_rev_id' ] ] ]
 126                         );
 127                         $this->addFields( 'ts_tags' );
 128                 }
 129
 130                 if ( !is_null( $params['tag'] ) ) {
 131                         $this->addTables( 'change_tag' );
 132                         $this->addJoinConds(
 133                                 [ 'change_tag' => [ 'INNER JOIN', [ 'ar_rev_id=ct_rev_id' ] ] ]
 134                         );
 135                         $this->addWhereFld( 'ct_tag', $params['tag'] );
 136                 }
 137
 138                 if ( $this->fetchContent ) {
 139                         // Modern MediaWiki has the content for deleted revs in the 'text'
 140                         // table using fields old_text and old_flags. But revisions deleted
 141                         // pre-1.5 store the content in the 'archive' table directly using
 142                         // fields ar_text and ar_flags, and no corresponding 'text' row. So
 143                         // we have to LEFT JOIN and fetch all four fields.
 144                         $this->addTables( 'text' );
 145                         $this->addJoinConds(
 146                                 [ 'text' => [ 'LEFT JOIN', [ 'ar_text_id=old_id' ] ] ]
 147                         );
 148                         $this->addFields( [ 'ar_text', 'ar_flags', 'old_text', 'old_flags' ] );
 149
 150                         // This also means stricter restrictions
 151                         $this->checkUserRightsAny( [ 'deletedtext', 'undelete' ] );
 152                 }
 153
 154                 $miser_ns = null;
 155
 156                 if ( $mode == 'all' ) {
 157                         if ( $params['namespace'] !== null ) {
 158                                 $namespaces = $params['namespace'];
 159                         } else {
 160                                 $namespaces = MWNamespace::getValidNamespaces();
 161                         }
 162                         $this->addWhereFld( 'ar_namespace', $namespaces );
 163
 164                         // For from/to/prefix, we have to consider the potential
 165                         // transformations of the title in all specified namespaces.
 166                         // Generally there will be only one transformation, but wikis with
 167                         // some namespaces case-sensitive could have two.
 168                         if ( $params['from'] !== null || $params['to'] !== null ) {
 169                                 $isDirNewer = ( $dir === 'newer' );
 170                                 $after = ( $isDirNewer ? '>=' : '<=' );
 171                                 $before = ( $isDirNewer ? '<=' : '>=' );
 172                                 $where = [];
 173                                 foreach ( $namespaces as $ns ) {
 174                                         $w = [];
 175                                         if ( $params['from'] !== null ) {
 176                                                 $w[] = 'ar_title' . $after .
 177                                                         $db->addQuotes( $this->titlePartToKey( $params['from'], $ns ) );
 178                                         }
 179                                         if ( $params['to'] !== null ) {
 180                                                 $w[] = 'ar_title' . $before .
 181                                                         $db->addQuotes( $this->titlePartToKey( $params['to'], $ns ) );
 182                                         }
 183                                         $w = $db->makeList( $w, LIST_AND );
 184                                         $where[$w][] = $ns;
 185                                 }
 186                                 if ( count( $where ) == 1 ) {
 187                                         $where = key( $where );
 188                                         $this->addWhere( $where );
 189                                 } else {
 190                                         $where2 = [];
 191                                         foreach ( $where as $w => $ns ) {
 192                                                 $where2[] = $db->makeList( [ $w, 'ar_namespace' => $ns ], LIST_AND );
 193                                         }
 194                                         $this->addWhere( $db->makeList( $where2, LIST_OR ) );
 195                                 }
 196                         }
 197
 198                         if ( isset( $params['prefix'] ) ) {
 199                                 $where = [];
 200                                 foreach ( $namespaces as $ns ) {
 201                                         $w = 'ar_title' . $db->buildLike(
 202                                                 $this->titlePartToKey( $params['prefix'], $ns ),
 203                                                 $db->anyString() );
 204                                         $where[$w][] = $ns;
 205                                 }
 206                                 if ( count( $where ) == 1 ) {
 207                                         $where = key( $where );
 208                                         $this->addWhere( $where );
 209                                 } else {
 210                                         $where2 = [];
 211                                         foreach ( $where as $w => $ns ) {
 212                                                 $where2[] = $db->makeList( [ $w, 'ar_namespace' => $ns ], LIST_AND );
 213                                         }
 214                                         $this->addWhere( $db->makeList( $where2, LIST_OR ) );
 215                                 }
 216                         }
 217                 } else {
 218                         if ( $this->getConfig()->get( 'MiserMode' ) ) {
 219                                 $miser_ns = $params['namespace'];
 220                         } else {
 221                                 $this->addWhereFld( 'ar_namespace', $params['namespace'] );
 222                         }
 223                         $this->addTimestampWhereRange( 'ar_timestamp', $dir, $params['start'], $params['end'] );
 224                 }
 225
 226                 if ( !is_null( $params['user'] ) ) {
 227                         // Don't query by user ID here, it might be able to use the ar_usertext_timestamp index.
 228                         $actorQuery = ActorMigration::newMigration()
 229                                 ->getWhere( $db, 'ar_user', User::newFromName( $params['user'], false ), false );
 230                         $this->addTables( $actorQuery['tables'] );
 231                         $this->addJoinConds( $actorQuery['joins'] );
 232                         $this->addWhere( $actorQuery['conds'] );
 233                 } elseif ( !is_null( $params['excludeuser'] ) ) {
 234                         // Here there's no chance of using ar_usertext_timestamp.
 235                         $actorQuery = ActorMigration::newMigration()
 236                                 ->getWhere( $db, 'ar_user', User::newFromName( $params['excludeuser'], false ) );
 237                         $this->addTables( $actorQuery['tables'] );
 238                         $this->addJoinConds( $actorQuery['joins'] );
 239                         $this->addWhere( 'NOT(' . $actorQuery['conds'] . ')' );
 240                 }
 241
 242                 if ( !is_null( $params['user'] ) || !is_null( $params['excludeuser'] ) ) {
 243                         // Paranoia: avoid brute force searches (T19342)
 244                         // (shouldn't be able to get here without 'deletedhistory', but
 245                         // check it again just in case)
 246                         if ( !$user->isAllowed( 'deletedhistory' ) ) {
 247                                 $bitmask = Revision::DELETED_USER;
 248                         } elseif ( !$user->isAllowedAny( 'suppressrevision', 'viewsuppressed' ) ) {
 249                                 $bitmask = Revision::DELETED_USER | Revision::DELETED_RESTRICTED;
 250                         } else {
 251                                 $bitmask = 0;
 252                         }
 253                         if ( $bitmask ) {
 254                                 $this->addWhere( $db->bitAnd( 'ar_deleted', $bitmask ) . " != $bitmask" );
 255                         }
 256                 }
 257
 258                 if ( !is_null( $params['continue'] ) ) {
 259                         $cont = explode( '|', $params['continue'] );
 260                         $op = ( $dir == 'newer' ? '>' : '<' );
 261                         if ( $optimizeGenerateTitles ) {
 262                                 $this->dieContinueUsageIf( count( $cont ) != 2 );
 263                                 $ns = intval( $cont[0] );
 264                                 $this->dieContinueUsageIf( strval( $ns ) !== $cont[0] );
 265                                 $title = $db->addQuotes( $cont[1] );
 266                                 $this->addWhere( "ar_namespace $op $ns OR " .
 267                                         "(ar_namespace = $ns AND ar_title $op= $title)" );
 268                         } elseif ( $mode == 'all' ) {
 269                                 $this->dieContinueUsageIf( count( $cont ) != 4 );
 270                                 $ns = intval( $cont[0] );
 271                                 $this->dieContinueUsageIf( strval( $ns ) !== $cont[0] );
 272                                 $title = $db->addQuotes( $cont[1] );
 273                                 $ts = $db->addQuotes( $db->timestamp( $cont[2] ) );
 274                                 $ar_id = (int)$cont[3];
 275                                 $this->dieContinueUsageIf( strval( $ar_id ) !== $cont[3] );
 276                                 $this->addWhere( "ar_namespace $op $ns OR " .
 277                                         "(ar_namespace = $ns AND " .
 278                                         "(ar_title $op $title OR " .
 279                                         "(ar_title = $title AND " .
 280                                         "(ar_timestamp $op $ts OR " .
 281                                         "(ar_timestamp = $ts AND " .
 282                                         "ar_id $op= $ar_id)))))" );
 283                         } else {
 284                                 $this->dieContinueUsageIf( count( $cont ) != 2 );
 285                                 $ts = $db->addQuotes( $db->timestamp( $cont[0] ) );
 286                                 $ar_id = (int)$cont[1];
 287                                 $this->dieContinueUsageIf( strval( $ar_id ) !== $cont[1] );
 288                                 $this->addWhere( "ar_timestamp $op $ts OR " .
 289                                         "(ar_timestamp = $ts AND " .
 290                                         "ar_id $op= $ar_id)" );
 291                         }
 292                 }
 293
 294                 $this->addOption( 'LIMIT', $this->limit + 1 );
 295
 296                 $sort = ( $dir == 'newer' ? '' : ' DESC' );
 297                 $orderby = [];
 298                 if ( $optimizeGenerateTitles ) {
 299                         // Targeting index name_title_timestamp
 300                         if ( $params['namespace'] === null || count( array_unique( $params['namespace'] ) ) > 1 ) {
 301                                 $orderby[] = "ar_namespace $sort";
 302                         }
 303                         $orderby[] = "ar_title $sort";
 304                 } elseif ( $mode == 'all' ) {
 305                         // Targeting index name_title_timestamp
 306                         if ( $params['namespace'] === null || count( array_unique( $params['namespace'] ) ) > 1 ) {
 307                                 $orderby[] = "ar_namespace $sort";
 308                         }
 309                         $orderby[] = "ar_title $sort";
 310                         $orderby[] = "ar_timestamp $sort";
 311                         $orderby[] = "ar_id $sort";
 312                 } else {
 313                         // Targeting index usertext_timestamp
 314                         // 'user' is always constant.
 315                         $orderby[] = "ar_timestamp $sort";
 316                         $orderby[] = "ar_id $sort";
 317                 }
 318                 $this->addOption( 'ORDER BY', $orderby );
 319
 320                 $res = $this->select( __METHOD__ );
 321                 $pageMap = []; // Maps ns&title to array index
 322                 $count = 0;
 323                 $nextIndex = 0;
 324                 $generated = [];
 325                 foreach ( $res as $row ) {
 326                         if ( ++$count > $this->limit ) {
 327                                 // We've had enough
 328                                 if ( $optimizeGenerateTitles ) {
 329                                         $this->setContinueEnumParameter( 'continue', "$row->ar_namespace|$row->ar_title" );
 330                                 } elseif ( $mode == 'all' ) {
 331                                         $this->setContinueEnumParameter( 'continue',
 332                                                 "$row->ar_namespace|$row->ar_title|$row->ar_timestamp|$row->ar_id"
 333                                         );
 334                                 } else {
 335                                         $this->setContinueEnumParameter( 'continue', "$row->ar_timestamp|$row->ar_id" );
 336                                 }
 337                                 break;
 338                         }
 339
 340                         // Miser mode namespace check
 341                         if ( $miser_ns !== null && !in_array( $row->ar_namespace, $miser_ns ) ) {
 342                                 continue;
 343                         }
 344
 345                         if ( $resultPageSet !== null ) {
 346                                 if ( $params['generatetitles'] ) {
 347                                         $key = "{$row->ar_namespace}:{$row->ar_title}";
 348                                         if ( !isset( $generated[$key] ) ) {
 349                                                 $generated[$key] = Title::makeTitle( $row->ar_namespace, $row->ar_title );
 350                                         }
 351                                 } else {
 352                                         $generated[] = $row->ar_rev_id;
 353                                 }
 354                         } else {
 355                                 $revision = Revision::newFromArchiveRow( $row );
 356                                 $rev = $this->extractRevisionInfo( $revision, $row );
 357
 358                                 if ( !isset( $pageMap[$row->ar_namespace][$row->ar_title] ) ) {
 359                                         $index = $nextIndex++;
 360                                         $pageMap[$row->ar_namespace][$row->ar_title] = $index;
 361                                         $title = $revision->getTitle();
 362                                         $a = [
 363                                                 'pageid' => $title->getArticleID(),
 364                                                 'revisions' => [ $rev ],
 365                                         ];
 366                                         ApiResult::setIndexedTagName( $a['revisions'], 'rev' );
 367                                         ApiQueryBase::addTitleInfo( $a, $title );
 368                                         $fit = $result->addValue( [ 'query', $this->getModuleName() ], $index, $a );
 369                                 } else {
 370                                         $index = $pageMap[$row->ar_namespace][$row->ar_title];
 371                                         $fit = $result->addValue(
 372                                                 [ 'query', $this->getModuleName(), $index, 'revisions' ],
 373                                                 null, $rev );
 374                                 }
 375                                 if ( !$fit ) {
 376                                         if ( $mode == 'all' ) {
 377                                                 $this->setContinueEnumParameter( 'continue',
 378                                                         "$row->ar_namespace|$row->ar_title|$row->ar_timestamp|$row->ar_id"
 379                                                 );
 380                                         } else {
 381                                                 $this->setContinueEnumParameter( 'continue', "$row->ar_timestamp|$row->ar_id" );
 382                                         }
 383                                         break;
 384                                 }
 385                         }
 386                 }
 387
 388                 if ( $resultPageSet !== null ) {
 389                         if ( $params['generatetitles'] ) {
 390                                 $resultPageSet->populateFromTitles( $generated );
 391                         } else {
 392                                 $resultPageSet->populateFromRevisionIDs( $generated );
 393                         }
 394                 } else {
 395                         $result->addIndexedTagName( [ 'query', $this->getModuleName() ], 'page' );
 396                 }
 397         }
 398
 399         public function getAllowedParams() {
 400                 $ret = parent::getAllowedParams() + [
 401                         'user' => [
 402                                 ApiBase::PARAM_TYPE => 'user'
 403                         ],
 404                         'namespace' => [
 405                                 ApiBase::PARAM_ISMULTI => true,
 406                                 ApiBase::PARAM_TYPE => 'namespace',
 407                         ],
 408                         'start' => [
 409                                 ApiBase::PARAM_TYPE => 'timestamp',
 410                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'useronly' ] ],
 411                         ],
 412                         'end' => [
 413                                 ApiBase::PARAM_TYPE => 'timestamp',
 414                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'useronly' ] ],
 415                         ],
 416                         'dir' => [
 417                                 ApiBase::PARAM_TYPE => [
 418                                         'newer',
 419                                         'older'
 420                                 ],
 421                                 ApiBase::PARAM_DFLT => 'older',
 422                                 ApiBase::PARAM_HELP_MSG => 'api-help-param-direction',
 423                         ],
 424                         'from' => [
 425                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'nonuseronly' ] ],
 426                         ],
 427                         'to' => [
 428                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'nonuseronly' ] ],
 429                         ],
 430                         'prefix' => [
 431                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'nonuseronly' ] ],
 432                         ],
 433                         'excludeuser' => [
 434                                 ApiBase::PARAM_TYPE => 'user',
 435                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'nonuseronly' ] ],
 436                         ],
 437                         'tag' => null,
 438                         'continue' => [
 439                                 ApiBase::PARAM_HELP_MSG => 'api-help-param-continue',
 440                         ],
 441                         'generatetitles' => [
 442                                 ApiBase::PARAM_DFLT => false
 443                         ],
 444                 ];
 445
 446                 if ( $this->getConfig()->get( 'MiserMode' ) ) {
 447                         $ret['user'][ApiBase::PARAM_HELP_MSG_APPEND] = [
 448                                 'apihelp-query+alldeletedrevisions-param-miser-user-namespace',
 449                         ];
 450                         $ret['namespace'][ApiBase::PARAM_HELP_MSG_APPEND] = [
 451                                 'apihelp-query+alldeletedrevisions-param-miser-user-namespace',
 452                         ];
 453                 }
 454
 455                 return $ret;
 456         }
 457
 458         protected function getExamplesMessages() {
 459                 return [
 460                         'action=query&list=alldeletedrevisions&adruser=Example&adrlimit=50'
 461                                 => 'apihelp-query+alldeletedrevisions-example-user',
 462                         'action=query&list=alldeletedrevisions&adrdir=newer&adrnamespace=0&adrlimit=50'
 463                                 => 'apihelp-query+alldeletedrevisions-example-ns-main',
 464                 ];
 465         }
 466
 467         public function getHelpUrls() {
 468                 return 'https://www.mediawiki.org/wiki/Special:MyLanguage/API:Alldeletedrevisions';
 469         }
 470 }