includes/api/ApiQueryAllDeletedRevisions.php

   1 <?php
   2 /**
   3  * Created on Oct 3, 2014
   4  *
   5  * Copyright © 2014 Wikimedia Foundation and contributors
   6  *
   7  * Heavily based on ApiQueryDeletedrevs,
   8  * Copyright © 2007 Roan Kattouw "<Firstname>.<Lastname>@gmail.com"
   9  *
  10  * This program is free software; you can redistribute it and/or modify
  11  * it under the terms of the GNU General Public License as published by
  12  * the Free Software Foundation; either version 2 of the License, or
  13  * (at your option) any later version.
  14  *
  15  * This program is distributed in the hope that it will be useful,
  16  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  17  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  18  * GNU General Public License for more details.
  19  *
  20  * You should have received a copy of the GNU General Public License along
  21  * with this program; if not, write to the Free Software Foundation, Inc.,
  22  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  23  * http://www.gnu.org/copyleft/gpl.html
  24  *
  25  * @file
  26  */
  27
  28 /**
  29  * Query module to enumerate all deleted revisions.
  30  *
  31  * @ingroup API
  32  */
  33 class ApiQueryAllDeletedRevisions extends ApiQueryRevisionsBase {
  34
  35         public function __construct( ApiQuery $query, $moduleName ) {
  36                 parent::__construct( $query, $moduleName, 'adr' );
  37         }
  38
  39         /**
  40          * @param ApiPageSet $resultPageSet
  41          * @return void
  42          */
  43         protected function run( ApiPageSet $resultPageSet = null ) {
  44                 // Before doing anything at all, let's check permissions
  45                 $this->checkUserRightsAny( 'deletedhistory' );
  46
  47                 $user = $this->getUser();
  48                 $db = $this->getDB();
  49                 $params = $this->extractRequestParams( false );
  50
  51                 $result = $this->getResult();
  52
  53                 // If the user wants no namespaces, they get no pages.
  54                 if ( $params['namespace'] === [] ) {
  55                         if ( $resultPageSet === null ) {
  56                                 $result->addValue( 'query', $this->getModuleName(), [] );
  57                         }
  58                         return;
  59                 }
  60
  61                 // This module operates in two modes:
  62                 // 'user': List deleted revs by a certain user
  63                 // 'all': List all deleted revs in NS
  64                 $mode = 'all';
  65                 if ( !is_null( $params['user'] ) ) {
  66                         $mode = 'user';
  67                 }
  68
  69                 if ( $mode == 'user' ) {
  70                         foreach ( [ 'from', 'to', 'prefix', 'excludeuser' ] as $param ) {
  71                                 if ( !is_null( $params[$param] ) ) {
  72                                         $p = $this->getModulePrefix();
  73                                         $this->dieWithError(
  74                                                 [ 'apierror-invalidparammix-cannotusewith', $p.$param, "{$p}user" ],
  75                                                 'invalidparammix'
  76                                         );
  77                                 }
  78                         }
  79                 } else {
  80                         foreach ( [ 'start', 'end' ] as $param ) {
  81                                 if ( !is_null( $params[$param] ) ) {
  82                                         $p = $this->getModulePrefix();
  83                                         $this->dieWithError(
  84                                                 [ 'apierror-invalidparammix-mustusewith', $p.$param, "{$p}user" ],
  85                                                 'invalidparammix'
  86                                         );
  87                                 }
  88                         }
  89                 }
  90
  91                 // If we're generating titles only, we can use DISTINCT for a better
  92                 // query. But we can't do that in 'user' mode (wrong index), and we can
  93                 // only do it when sorting ASC (because MySQL apparently can't use an
  94                 // index backwards for grouping even though it can for ORDER BY, WTF?)
  95                 $dir = $params['dir'];
  96                 $optimizeGenerateTitles = false;
  97                 if ( $mode === 'all' && $params['generatetitles'] && $resultPageSet !== null ) {
  98                         if ( $dir === 'newer' ) {
  99                                 $optimizeGenerateTitles = true;
 100                         } else {
 101                                 $p = $this->getModulePrefix();
 102                                 $this->addWarning( [ 'apiwarn-alldeletedrevisions-performance', $p ], 'performance' );
 103                         }
 104                 }
 105
 106                 if ( $resultPageSet === null ) {
 107                         $this->parseParameters( $params );
 108                         $arQuery = Revision::getArchiveQueryInfo();
 109                         $this->addTables( $arQuery['tables'] );
 110                         $this->addJoinConds( $arQuery['joins'] );
 111                         $this->addFields( $arQuery['fields'] );
 112                         $this->addFields( [ 'ar_title', 'ar_namespace' ] );
 113                 } else {
 114                         $this->limit = $this->getParameter( 'limit' ) ?: 10;
 115                         $this->addTables( 'archive' );
 116                         $this->addFields( [ 'ar_title', 'ar_namespace' ] );
 117                         if ( $optimizeGenerateTitles ) {
 118                                 $this->addOption( 'DISTINCT' );
 119                         } else {
 120                                 $this->addFields( [ 'ar_timestamp', 'ar_rev_id', 'ar_id' ] );
 121                         }
 122                 }
 123
 124                 if ( $this->fld_tags ) {
 125                         $this->addTables( 'tag_summary' );
 126                         $this->addJoinConds(
 127                                 [ 'tag_summary' => [ 'LEFT JOIN', [ 'ar_rev_id=ts_rev_id' ] ] ]
 128                         );
 129                         $this->addFields( 'ts_tags' );
 130                 }
 131
 132                 if ( !is_null( $params['tag'] ) ) {
 133                         $this->addTables( 'change_tag' );
 134                         $this->addJoinConds(
 135                                 [ 'change_tag' => [ 'INNER JOIN', [ 'ar_rev_id=ct_rev_id' ] ] ]
 136                         );
 137                         $this->addWhereFld( 'ct_tag', $params['tag'] );
 138                 }
 139
 140                 if ( $this->fetchContent ) {
 141                         // Modern MediaWiki has the content for deleted revs in the 'text'
 142                         // table using fields old_text and old_flags. But revisions deleted
 143                         // pre-1.5 store the content in the 'archive' table directly using
 144                         // fields ar_text and ar_flags, and no corresponding 'text' row. So
 145                         // we have to LEFT JOIN and fetch all four fields.
 146                         $this->addTables( 'text' );
 147                         $this->addJoinConds(
 148                                 [ 'text' => [ 'LEFT JOIN', [ 'ar_text_id=old_id' ] ] ]
 149                         );
 150                         $this->addFields( [ 'ar_text', 'ar_flags', 'old_text', 'old_flags' ] );
 151
 152                         // This also means stricter restrictions
 153                         $this->checkUserRightsAny( [ 'deletedtext', 'undelete' ] );
 154                 }
 155
 156                 $miser_ns = null;
 157
 158                 if ( $mode == 'all' ) {
 159                         if ( $params['namespace'] !== null ) {
 160                                 $namespaces = $params['namespace'];
 161                         } else {
 162                                 $namespaces = MWNamespace::getValidNamespaces();
 163                         }
 164                         $this->addWhereFld( 'ar_namespace', $namespaces );
 165
 166                         // For from/to/prefix, we have to consider the potential
 167                         // transformations of the title in all specified namespaces.
 168                         // Generally there will be only one transformation, but wikis with
 169                         // some namespaces case-sensitive could have two.
 170                         if ( $params['from'] !== null || $params['to'] !== null ) {
 171                                 $isDirNewer = ( $dir === 'newer' );
 172                                 $after = ( $isDirNewer ? '>=' : '<=' );
 173                                 $before = ( $isDirNewer ? '<=' : '>=' );
 174                                 $where = [];
 175                                 foreach ( $namespaces as $ns ) {
 176                                         $w = [];
 177                                         if ( $params['from'] !== null ) {
 178                                                 $w[] = 'ar_title' . $after .
 179                                                         $db->addQuotes( $this->titlePartToKey( $params['from'], $ns ) );
 180                                         }
 181                                         if ( $params['to'] !== null ) {
 182                                                 $w[] = 'ar_title' . $before .
 183                                                         $db->addQuotes( $this->titlePartToKey( $params['to'], $ns ) );
 184                                         }
 185                                         $w = $db->makeList( $w, LIST_AND );
 186                                         $where[$w][] = $ns;
 187                                 }
 188                                 if ( count( $where ) == 1 ) {
 189                                         $where = key( $where );
 190                                         $this->addWhere( $where );
 191                                 } else {
 192                                         $where2 = [];
 193                                         foreach ( $where as $w => $ns ) {
 194                                                 $where2[] = $db->makeList( [ $w, 'ar_namespace' => $ns ], LIST_AND );
 195                                         }
 196                                         $this->addWhere( $db->makeList( $where2, LIST_OR ) );
 197                                 }
 198                         }
 199
 200                         if ( isset( $params['prefix'] ) ) {
 201                                 $where = [];
 202                                 foreach ( $namespaces as $ns ) {
 203                                         $w = 'ar_title' . $db->buildLike(
 204                                                 $this->titlePartToKey( $params['prefix'], $ns ),
 205                                                 $db->anyString() );
 206                                         $where[$w][] = $ns;
 207                                 }
 208                                 if ( count( $where ) == 1 ) {
 209                                         $where = key( $where );
 210                                         $this->addWhere( $where );
 211                                 } else {
 212                                         $where2 = [];
 213                                         foreach ( $where as $w => $ns ) {
 214                                                 $where2[] = $db->makeList( [ $w, 'ar_namespace' => $ns ], LIST_AND );
 215                                         }
 216                                         $this->addWhere( $db->makeList( $where2, LIST_OR ) );
 217                                 }
 218                         }
 219                 } else {
 220                         if ( $this->getConfig()->get( 'MiserMode' ) ) {
 221                                 $miser_ns = $params['namespace'];
 222                         } else {
 223                                 $this->addWhereFld( 'ar_namespace', $params['namespace'] );
 224                         }
 225                         $this->addTimestampWhereRange( 'ar_timestamp', $dir, $params['start'], $params['end'] );
 226                 }
 227
 228                 if ( !is_null( $params['user'] ) ) {
 229                         $this->addWhereFld( 'ar_user_text', $params['user'] );
 230                 } elseif ( !is_null( $params['excludeuser'] ) ) {
 231                         $this->addWhere( 'ar_user_text != ' .
 232                                 $db->addQuotes( $params['excludeuser'] ) );
 233                 }
 234
 235                 if ( !is_null( $params['user'] ) || !is_null( $params['excludeuser'] ) ) {
 236                         // Paranoia: avoid brute force searches (T19342)
 237                         // (shouldn't be able to get here without 'deletedhistory', but
 238                         // check it again just in case)
 239                         if ( !$user->isAllowed( 'deletedhistory' ) ) {
 240                                 $bitmask = Revision::DELETED_USER;
 241                         } elseif ( !$user->isAllowedAny( 'suppressrevision', 'viewsuppressed' ) ) {
 242                                 $bitmask = Revision::DELETED_USER | Revision::DELETED_RESTRICTED;
 243                         } else {
 244                                 $bitmask = 0;
 245                         }
 246                         if ( $bitmask ) {
 247                                 $this->addWhere( $db->bitAnd( 'ar_deleted', $bitmask ) . " != $bitmask" );
 248                         }
 249                 }
 250
 251                 if ( !is_null( $params['continue'] ) ) {
 252                         $cont = explode( '|', $params['continue'] );
 253                         $op = ( $dir == 'newer' ? '>' : '<' );
 254                         if ( $optimizeGenerateTitles ) {
 255                                 $this->dieContinueUsageIf( count( $cont ) != 2 );
 256                                 $ns = intval( $cont[0] );
 257                                 $this->dieContinueUsageIf( strval( $ns ) !== $cont[0] );
 258                                 $title = $db->addQuotes( $cont[1] );
 259                                 $this->addWhere( "ar_namespace $op $ns OR " .
 260                                         "(ar_namespace = $ns AND ar_title $op= $title)" );
 261                         } elseif ( $mode == 'all' ) {
 262                                 $this->dieContinueUsageIf( count( $cont ) != 4 );
 263                                 $ns = intval( $cont[0] );
 264                                 $this->dieContinueUsageIf( strval( $ns ) !== $cont[0] );
 265                                 $title = $db->addQuotes( $cont[1] );
 266                                 $ts = $db->addQuotes( $db->timestamp( $cont[2] ) );
 267                                 $ar_id = (int)$cont[3];
 268                                 $this->dieContinueUsageIf( strval( $ar_id ) !== $cont[3] );
 269                                 $this->addWhere( "ar_namespace $op $ns OR " .
 270                                         "(ar_namespace = $ns AND " .
 271                                         "(ar_title $op $title OR " .
 272                                         "(ar_title = $title AND " .
 273                                         "(ar_timestamp $op $ts OR " .
 274                                         "(ar_timestamp = $ts AND " .
 275                                         "ar_id $op= $ar_id)))))" );
 276                         } else {
 277                                 $this->dieContinueUsageIf( count( $cont ) != 2 );
 278                                 $ts = $db->addQuotes( $db->timestamp( $cont[0] ) );
 279                                 $ar_id = (int)$cont[1];
 280                                 $this->dieContinueUsageIf( strval( $ar_id ) !== $cont[1] );
 281                                 $this->addWhere( "ar_timestamp $op $ts OR " .
 282                                         "(ar_timestamp = $ts AND " .
 283                                         "ar_id $op= $ar_id)" );
 284                         }
 285                 }
 286
 287                 $this->addOption( 'LIMIT', $this->limit + 1 );
 288
 289                 $sort = ( $dir == 'newer' ? '' : ' DESC' );
 290                 $orderby = [];
 291                 if ( $optimizeGenerateTitles ) {
 292                         // Targeting index name_title_timestamp
 293                         if ( $params['namespace'] === null || count( array_unique( $params['namespace'] ) ) > 1 ) {
 294                                 $orderby[] = "ar_namespace $sort";
 295                         }
 296                         $orderby[] = "ar_title $sort";
 297                 } elseif ( $mode == 'all' ) {
 298                         // Targeting index name_title_timestamp
 299                         if ( $params['namespace'] === null || count( array_unique( $params['namespace'] ) ) > 1 ) {
 300                                 $orderby[] = "ar_namespace $sort";
 301                         }
 302                         $orderby[] = "ar_title $sort";
 303                         $orderby[] = "ar_timestamp $sort";
 304                         $orderby[] = "ar_id $sort";
 305                 } else {
 306                         // Targeting index usertext_timestamp
 307                         // 'user' is always constant.
 308                         $orderby[] = "ar_timestamp $sort";
 309                         $orderby[] = "ar_id $sort";
 310                 }
 311                 $this->addOption( 'ORDER BY', $orderby );
 312
 313                 $res = $this->select( __METHOD__ );
 314                 $pageMap = []; // Maps ns&title to array index
 315                 $count = 0;
 316                 $nextIndex = 0;
 317                 $generated = [];
 318                 foreach ( $res as $row ) {
 319                         if ( ++$count > $this->limit ) {
 320                                 // We've had enough
 321                                 if ( $optimizeGenerateTitles ) {
 322                                         $this->setContinueEnumParameter( 'continue', "$row->ar_namespace|$row->ar_title" );
 323                                 } elseif ( $mode == 'all' ) {
 324                                         $this->setContinueEnumParameter( 'continue',
 325                                                 "$row->ar_namespace|$row->ar_title|$row->ar_timestamp|$row->ar_id"
 326                                         );
 327                                 } else {
 328                                         $this->setContinueEnumParameter( 'continue', "$row->ar_timestamp|$row->ar_id" );
 329                                 }
 330                                 break;
 331                         }
 332
 333                         // Miser mode namespace check
 334                         if ( $miser_ns !== null && !in_array( $row->ar_namespace, $miser_ns ) ) {
 335                                 continue;
 336                         }
 337
 338                         if ( $resultPageSet !== null ) {
 339                                 if ( $params['generatetitles'] ) {
 340                                         $key = "{$row->ar_namespace}:{$row->ar_title}";
 341                                         if ( !isset( $generated[$key] ) ) {
 342                                                 $generated[$key] = Title::makeTitle( $row->ar_namespace, $row->ar_title );
 343                                         }
 344                                 } else {
 345                                         $generated[] = $row->ar_rev_id;
 346                                 }
 347                         } else {
 348                                 $revision = Revision::newFromArchiveRow( $row );
 349                                 $rev = $this->extractRevisionInfo( $revision, $row );
 350
 351                                 if ( !isset( $pageMap[$row->ar_namespace][$row->ar_title] ) ) {
 352                                         $index = $nextIndex++;
 353                                         $pageMap[$row->ar_namespace][$row->ar_title] = $index;
 354                                         $title = $revision->getTitle();
 355                                         $a = [
 356                                                 'pageid' => $title->getArticleID(),
 357                                                 'revisions' => [ $rev ],
 358                                         ];
 359                                         ApiResult::setIndexedTagName( $a['revisions'], 'rev' );
 360                                         ApiQueryBase::addTitleInfo( $a, $title );
 361                                         $fit = $result->addValue( [ 'query', $this->getModuleName() ], $index, $a );
 362                                 } else {
 363                                         $index = $pageMap[$row->ar_namespace][$row->ar_title];
 364                                         $fit = $result->addValue(
 365                                                 [ 'query', $this->getModuleName(), $index, 'revisions' ],
 366                                                 null, $rev );
 367                                 }
 368                                 if ( !$fit ) {
 369                                         if ( $mode == 'all' ) {
 370                                                 $this->setContinueEnumParameter( 'continue',
 371                                                         "$row->ar_namespace|$row->ar_title|$row->ar_timestamp|$row->ar_id"
 372                                                 );
 373                                         } else {
 374                                                 $this->setContinueEnumParameter( 'continue', "$row->ar_timestamp|$row->ar_id" );
 375                                         }
 376                                         break;
 377                                 }
 378                         }
 379                 }
 380
 381                 if ( $resultPageSet !== null ) {
 382                         if ( $params['generatetitles'] ) {
 383                                 $resultPageSet->populateFromTitles( $generated );
 384                         } else {
 385                                 $resultPageSet->populateFromRevisionIDs( $generated );
 386                         }
 387                 } else {
 388                         $result->addIndexedTagName( [ 'query', $this->getModuleName() ], 'page' );
 389                 }
 390         }
 391
 392         public function getAllowedParams() {
 393                 $ret = parent::getAllowedParams() + [
 394                         'user' => [
 395                                 ApiBase::PARAM_TYPE => 'user'
 396                         ],
 397                         'namespace' => [
 398                                 ApiBase::PARAM_ISMULTI => true,
 399                                 ApiBase::PARAM_TYPE => 'namespace',
 400                         ],
 401                         'start' => [
 402                                 ApiBase::PARAM_TYPE => 'timestamp',
 403                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'useronly' ] ],
 404                         ],
 405                         'end' => [
 406                                 ApiBase::PARAM_TYPE => 'timestamp',
 407                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'useronly' ] ],
 408                         ],
 409                         'dir' => [
 410                                 ApiBase::PARAM_TYPE => [
 411                                         'newer',
 412                                         'older'
 413                                 ],
 414                                 ApiBase::PARAM_DFLT => 'older',
 415                                 ApiBase::PARAM_HELP_MSG => 'api-help-param-direction',
 416                         ],
 417                         'from' => [
 418                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'nonuseronly' ] ],
 419                         ],
 420                         'to' => [
 421                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'nonuseronly' ] ],
 422                         ],
 423                         'prefix' => [
 424                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'nonuseronly' ] ],
 425                         ],
 426                         'excludeuser' => [
 427                                 ApiBase::PARAM_TYPE => 'user',
 428                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'nonuseronly' ] ],
 429                         ],
 430                         'tag' => null,
 431                         'continue' => [
 432                                 ApiBase::PARAM_HELP_MSG => 'api-help-param-continue',
 433                         ],
 434                         'generatetitles' => [
 435                                 ApiBase::PARAM_DFLT => false
 436                         ],
 437                 ];
 438
 439                 if ( $this->getConfig()->get( 'MiserMode' ) ) {
 440                         $ret['user'][ApiBase::PARAM_HELP_MSG_APPEND] = [
 441                                 'apihelp-query+alldeletedrevisions-param-miser-user-namespace',
 442                         ];
 443                         $ret['namespace'][ApiBase::PARAM_HELP_MSG_APPEND] = [
 444                                 'apihelp-query+alldeletedrevisions-param-miser-user-namespace',
 445                         ];
 446                 }
 447
 448                 return $ret;
 449         }
 450
 451         protected function getExamplesMessages() {
 452                 return [
 453                         'action=query&list=alldeletedrevisions&adruser=Example&adrlimit=50'
 454                                 => 'apihelp-query+alldeletedrevisions-example-user',
 455                         'action=query&list=alldeletedrevisions&adrdir=newer&adrnamespace=0&adrlimit=50'
 456                                 => 'apihelp-query+alldeletedrevisions-example-ns-main',
 457                 ];
 458         }
 459
 460         public function getHelpUrls() {
 461                 return 'https://www.mediawiki.org/wiki/Special:MyLanguage/API:Alldeletedrevisions';
 462         }
 463 }