includes/api/ApiQueryAllDeletedRevisions.php

   1 <?php
   2 /**
   3  * Created on Oct 3, 2014
   4  *
   5  * Copyright © 2014 Brad Jorsch "bjorsch@wikimedia.org"
   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                 $user = $this->getUser();
  45                 // Before doing anything at all, let's check permissions
  46                 if ( !$user->isAllowed( 'deletedhistory' ) ) {
  47                         $this->dieUsage(
  48                                 'You don\'t have permission to view deleted revision information',
  49                                 'permissiondenied'
  50                         );
  51                 }
  52
  53                 $db = $this->getDB();
  54                 $params = $this->extractRequestParams( false );
  55
  56                 $result = $this->getResult();
  57
  58                 // If the user wants no namespaces, they get no pages.
  59                 if ( $params['namespace'] === [] ) {
  60                         if ( $resultPageSet === null ) {
  61                                 $result->addValue( 'query', $this->getModuleName(), [] );
  62                         }
  63                         return;
  64                 }
  65
  66                 // This module operates in two modes:
  67                 // 'user': List deleted revs by a certain user
  68                 // 'all': List all deleted revs in NS
  69                 $mode = 'all';
  70                 if ( !is_null( $params['user'] ) ) {
  71                         $mode = 'user';
  72                 }
  73
  74                 if ( $mode == 'user' ) {
  75                         foreach ( [ 'from', 'to', 'prefix', 'excludeuser' ] as $param ) {
  76                                 if ( !is_null( $params[$param] ) ) {
  77                                         $p = $this->getModulePrefix();
  78                                         $this->dieUsage( "The '{$p}{$param}' parameter cannot be used with '{$p}user'",
  79                                                 'badparams' );
  80                                 }
  81                         }
  82                 } else {
  83                         foreach ( [ 'start', 'end' ] as $param ) {
  84                                 if ( !is_null( $params[$param] ) ) {
  85                                         $p = $this->getModulePrefix();
  86                                         $this->dieUsage( "The '{$p}{$param}' parameter may only be used with '{$p}user'",
  87                                                 'badparams' );
  88                                 }
  89                         }
  90                 }
  91
  92                 // If we're generating titles only, we can use DISTINCT for a better
  93                 // query. But we can't do that in 'user' mode (wrong index), and we can
  94                 // only do it when sorting ASC (because MySQL apparently can't use an
  95                 // index backwards for grouping even though it can for ORDER BY, WTF?)
  96                 $dir = $params['dir'];
  97                 $optimizeGenerateTitles = false;
  98                 if ( $mode === 'all' && $params['generatetitles'] && $resultPageSet !== null ) {
  99                         if ( $dir === 'newer' ) {
 100                                 $optimizeGenerateTitles = true;
 101                         } else {
 102                                 $p = $this->getModulePrefix();
 103                                 $this->setWarning( "For better performance when generating titles, set {$p}dir=newer" );
 104                         }
 105                 }
 106
 107                 $this->addTables( 'archive' );
 108                 if ( $resultPageSet === null ) {
 109                         $this->parseParameters( $params );
 110                         $this->addFields( Revision::selectArchiveFields() );
 111                         $this->addFields( [ 'ar_title', 'ar_namespace' ] );
 112                 } else {
 113                         $this->limit = $this->getParameter( 'limit' ) ?: 10;
 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                         $this->addFields( [ 'ar_text', 'ar_flags', 'old_text', 'old_flags' ] );
 149
 150                         // This also means stricter restrictions
 151                         if ( !$user->isAllowedAny( 'undelete', 'deletedtext' ) ) {
 152                                 $this->dieUsage(
 153                                         'You don\'t have permission to view deleted revision content',
 154                                         'permissiondenied'
 155                                 );
 156                         }
 157                 }
 158
 159                 $miser_ns = null;
 160
 161                 if ( $mode == 'all' ) {
 162                         if ( $params['namespace'] !== null ) {
 163                                 $namespaces = $params['namespace'];
 164                         } else {
 165                                 $namespaces = MWNamespace::getValidNamespaces();
 166                         }
 167                         $this->addWhereFld( 'ar_namespace', $namespaces );
 168
 169                         // For from/to/prefix, we have to consider the potential
 170                         // transformations of the title in all specified namespaces.
 171                         // Generally there will be only one transformation, but wikis with
 172                         // some namespaces case-sensitive could have two.
 173                         if ( $params['from'] !== null || $params['to'] !== null ) {
 174                                 $isDirNewer = ( $dir === 'newer' );
 175                                 $after = ( $isDirNewer ? '>=' : '<=' );
 176                                 $before = ( $isDirNewer ? '<=' : '>=' );
 177                                 $where = [];
 178                                 foreach ( $namespaces as $ns ) {
 179                                         $w = [];
 180                                         if ( $params['from'] !== null ) {
 181                                                 $w[] = 'ar_title' . $after .
 182                                                         $db->addQuotes( $this->titlePartToKey( $params['from'], $ns ) );
 183                                         }
 184                                         if ( $params['to'] !== null ) {
 185                                                 $w[] = 'ar_title' . $before .
 186                                                         $db->addQuotes( $this->titlePartToKey( $params['to'], $ns ) );
 187                                         }
 188                                         $w = $db->makeList( $w, LIST_AND );
 189                                         $where[$w][] = $ns;
 190                                 }
 191                                 if ( count( $where ) == 1 ) {
 192                                         $where = key( $where );
 193                                         $this->addWhere( $where );
 194                                 } else {
 195                                         $where2 = [];
 196                                         foreach ( $where as $w => $ns ) {
 197                                                 $where2[] = $db->makeList( [ $w, 'ar_namespace' => $ns ], LIST_AND );
 198                                         }
 199                                         $this->addWhere( $db->makeList( $where2, LIST_OR ) );
 200                                 }
 201                         }
 202
 203                         if ( isset( $params['prefix'] ) ) {
 204                                 $where = [];
 205                                 foreach ( $namespaces as $ns ) {
 206                                         $w = 'ar_title' . $db->buildLike(
 207                                                 $this->titlePartToKey( $params['prefix'], $ns ),
 208                                                 $db->anyString() );
 209                                         $where[$w][] = $ns;
 210                                 }
 211                                 if ( count( $where ) == 1 ) {
 212                                         $where = key( $where );
 213                                         $this->addWhere( $where );
 214                                 } else {
 215                                         $where2 = [];
 216                                         foreach ( $where as $w => $ns ) {
 217                                                 $where2[] = $db->makeList( [ $w, 'ar_namespace' => $ns ], LIST_AND );
 218                                         }
 219                                         $this->addWhere( $db->makeList( $where2, LIST_OR ) );
 220                                 }
 221                         }
 222                 } else {
 223                         if ( $this->getConfig()->get( 'MiserMode' ) ) {
 224                                 $miser_ns = $params['namespace'];
 225                         } else {
 226                                 $this->addWhereFld( 'ar_namespace', $params['namespace'] );
 227                         }
 228                         $this->addTimestampWhereRange( 'ar_timestamp', $dir, $params['start'], $params['end'] );
 229                 }
 230
 231                 if ( !is_null( $params['user'] ) ) {
 232                         $this->addWhereFld( 'ar_user_text', $params['user'] );
 233                 } elseif ( !is_null( $params['excludeuser'] ) ) {
 234                         $this->addWhere( 'ar_user_text != ' .
 235                                 $db->addQuotes( $params['excludeuser'] ) );
 236                 }
 237
 238                 if ( !is_null( $params['user'] ) || !is_null( $params['excludeuser'] ) ) {
 239                         // Paranoia: avoid brute force searches (bug 17342)
 240                         // (shouldn't be able to get here without 'deletedhistory', but
 241                         // check it again just in case)
 242                         if ( !$user->isAllowed( 'deletedhistory' ) ) {
 243                                 $bitmask = Revision::DELETED_USER;
 244                         } elseif ( !$user->isAllowedAny( 'suppressrevision', 'viewsuppressed' ) ) {
 245                                 $bitmask = Revision::DELETED_USER | Revision::DELETED_RESTRICTED;
 246                         } else {
 247                                 $bitmask = 0;
 248                         }
 249                         if ( $bitmask ) {
 250                                 $this->addWhere( $db->bitAnd( 'ar_deleted', $bitmask ) . " != $bitmask" );
 251                         }
 252                 }
 253
 254                 if ( !is_null( $params['continue'] ) ) {
 255                         $cont = explode( '|', $params['continue'] );
 256                         $op = ( $dir == 'newer' ? '>' : '<' );
 257                         if ( $optimizeGenerateTitles ) {
 258                                 $this->dieContinueUsageIf( count( $cont ) != 2 );
 259                                 $ns = intval( $cont[0] );
 260                                 $this->dieContinueUsageIf( strval( $ns ) !== $cont[0] );
 261                                 $title = $db->addQuotes( $cont[1] );
 262                                 $this->addWhere( "ar_namespace $op $ns OR " .
 263                                         "(ar_namespace = $ns AND ar_title $op= $title)" );
 264                         } elseif ( $mode == 'all' ) {
 265                                 $this->dieContinueUsageIf( count( $cont ) != 4 );
 266                                 $ns = intval( $cont[0] );
 267                                 $this->dieContinueUsageIf( strval( $ns ) !== $cont[0] );
 268                                 $title = $db->addQuotes( $cont[1] );
 269                                 $ts = $db->addQuotes( $db->timestamp( $cont[2] ) );
 270                                 $ar_id = (int)$cont[3];
 271                                 $this->dieContinueUsageIf( strval( $ar_id ) !== $cont[3] );
 272                                 $this->addWhere( "ar_namespace $op $ns OR " .
 273                                         "(ar_namespace = $ns AND " .
 274                                         "(ar_title $op $title OR " .
 275                                         "(ar_title = $title AND " .
 276                                         "(ar_timestamp $op $ts OR " .
 277                                         "(ar_timestamp = $ts AND " .
 278                                         "ar_id $op= $ar_id)))))" );
 279                         } else {
 280                                 $this->dieContinueUsageIf( count( $cont ) != 2 );
 281                                 $ts = $db->addQuotes( $db->timestamp( $cont[0] ) );
 282                                 $ar_id = (int)$cont[1];
 283                                 $this->dieContinueUsageIf( strval( $ar_id ) !== $cont[1] );
 284                                 $this->addWhere( "ar_timestamp $op $ts OR " .
 285                                         "(ar_timestamp = $ts AND " .
 286                                         "ar_id $op= $ar_id)" );
 287                         }
 288                 }
 289
 290                 $this->addOption( 'LIMIT', $this->limit + 1 );
 291
 292                 $sort = ( $dir == 'newer' ? '' : ' DESC' );
 293                 $orderby = [];
 294                 if ( $optimizeGenerateTitles ) {
 295                         // Targeting index name_title_timestamp
 296                         if ( $params['namespace'] === null || count( array_unique( $params['namespace'] ) ) > 1 ) {
 297                                 $orderby[] = "ar_namespace $sort";
 298                         }
 299                         $orderby[] = "ar_title $sort";
 300                 } elseif ( $mode == 'all' ) {
 301                         // Targeting index name_title_timestamp
 302                         if ( $params['namespace'] === null || count( array_unique( $params['namespace'] ) ) > 1 ) {
 303                                 $orderby[] = "ar_namespace $sort";
 304                         }
 305                         $orderby[] = "ar_title $sort";
 306                         $orderby[] = "ar_timestamp $sort";
 307                         $orderby[] = "ar_id $sort";
 308                 } else {
 309                         // Targeting index usertext_timestamp
 310                         // 'user' is always constant.
 311                         $orderby[] = "ar_timestamp $sort";
 312                         $orderby[] = "ar_id $sort";
 313                 }
 314                 $this->addOption( 'ORDER BY', $orderby );
 315
 316                 $res = $this->select( __METHOD__ );
 317                 $pageMap = []; // Maps ns&title to array index
 318                 $count = 0;
 319                 $nextIndex = 0;
 320                 $generated = [];
 321                 foreach ( $res as $row ) {
 322                         if ( ++$count > $this->limit ) {
 323                                 // We've had enough
 324                                 if ( $optimizeGenerateTitles ) {
 325                                         $this->setContinueEnumParameter( 'continue', "$row->ar_namespace|$row->ar_title" );
 326                                 } elseif ( $mode == 'all' ) {
 327                                         $this->setContinueEnumParameter( 'continue',
 328                                                 "$row->ar_namespace|$row->ar_title|$row->ar_timestamp|$row->ar_id"
 329                                         );
 330                                 } else {
 331                                         $this->setContinueEnumParameter( 'continue', "$row->ar_timestamp|$row->ar_id" );
 332                                 }
 333                                 break;
 334                         }
 335
 336                         // Miser mode namespace check
 337                         if ( $miser_ns !== null && !in_array( $row->ar_namespace, $miser_ns ) ) {
 338                                 continue;
 339                         }
 340
 341                         if ( $resultPageSet !== null ) {
 342                                 if ( $params['generatetitles'] ) {
 343                                         $key = "{$row->ar_namespace}:{$row->ar_title}";
 344                                         if ( !isset( $generated[$key] ) ) {
 345                                                 $generated[$key] = Title::makeTitle( $row->ar_namespace, $row->ar_title );
 346                                         }
 347                                 } else {
 348                                         $generated[] = $row->ar_rev_id;
 349                                 }
 350                         } else {
 351                                 $revision = Revision::newFromArchiveRow( $row );
 352                                 $rev = $this->extractRevisionInfo( $revision, $row );
 353
 354                                 if ( !isset( $pageMap[$row->ar_namespace][$row->ar_title] ) ) {
 355                                         $index = $nextIndex++;
 356                                         $pageMap[$row->ar_namespace][$row->ar_title] = $index;
 357                                         $title = $revision->getTitle();
 358                                         $a = [
 359                                                 'pageid' => $title->getArticleID(),
 360                                                 'revisions' => [ $rev ],
 361                                         ];
 362                                         ApiResult::setIndexedTagName( $a['revisions'], 'rev' );
 363                                         ApiQueryBase::addTitleInfo( $a, $title );
 364                                         $fit = $result->addValue( [ 'query', $this->getModuleName() ], $index, $a );
 365                                 } else {
 366                                         $index = $pageMap[$row->ar_namespace][$row->ar_title];
 367                                         $fit = $result->addValue(
 368                                                 [ 'query', $this->getModuleName(), $index, 'revisions' ],
 369                                                 null, $rev );
 370                                 }
 371                                 if ( !$fit ) {
 372                                         if ( $mode == 'all' ) {
 373                                                 $this->setContinueEnumParameter( 'continue',
 374                                                         "$row->ar_namespace|$row->ar_title|$row->ar_timestamp|$row->ar_id"
 375                                                 );
 376                                         } else {
 377                                                 $this->setContinueEnumParameter( 'continue', "$row->ar_timestamp|$row->ar_id" );
 378                                         }
 379                                         break;
 380                                 }
 381                         }
 382                 }
 383
 384                 if ( $resultPageSet !== null ) {
 385                         if ( $params['generatetitles'] ) {
 386                                 $resultPageSet->populateFromTitles( $generated );
 387                         } else {
 388                                 $resultPageSet->populateFromRevisionIDs( $generated );
 389                         }
 390                 } else {
 391                         $result->addIndexedTagName( [ 'query', $this->getModuleName() ], 'page' );
 392                 }
 393         }
 394
 395         public function getAllowedParams() {
 396                 $ret = parent::getAllowedParams() + [
 397                         'user' => [
 398                                 ApiBase::PARAM_TYPE => 'user'
 399                         ],
 400                         'namespace' => [
 401                                 ApiBase::PARAM_ISMULTI => true,
 402                                 ApiBase::PARAM_TYPE => 'namespace',
 403                         ],
 404                         'start' => [
 405                                 ApiBase::PARAM_TYPE => 'timestamp',
 406                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'useronly' ] ],
 407                         ],
 408                         'end' => [
 409                                 ApiBase::PARAM_TYPE => 'timestamp',
 410                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'useronly' ] ],
 411                         ],
 412                         'dir' => [
 413                                 ApiBase::PARAM_TYPE => [
 414                                         'newer',
 415                                         'older'
 416                                 ],
 417                                 ApiBase::PARAM_DFLT => 'older',
 418                                 ApiBase::PARAM_HELP_MSG => 'api-help-param-direction',
 419                         ],
 420                         'from' => [
 421                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'nonuseronly' ] ],
 422                         ],
 423                         'to' => [
 424                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'nonuseronly' ] ],
 425                         ],
 426                         'prefix' => [
 427                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'nonuseronly' ] ],
 428                         ],
 429                         'excludeuser' => [
 430                                 ApiBase::PARAM_TYPE => 'user',
 431                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'nonuseronly' ] ],
 432                         ],
 433                         'tag' => null,
 434                         'continue' => [
 435                                 ApiBase::PARAM_HELP_MSG => 'api-help-param-continue',
 436                         ],
 437                         'generatetitles' => [
 438                                 ApiBase::PARAM_DFLT => false
 439                         ],
 440                 ];
 441
 442                 if ( $this->getConfig()->get( 'MiserMode' ) ) {
 443                         $ret['user'][ApiBase::PARAM_HELP_MSG_APPEND] = [
 444                                 'apihelp-query+alldeletedrevisions-param-miser-user-namespace',
 445                         ];
 446                         $ret['namespace'][ApiBase::PARAM_HELP_MSG_APPEND] = [
 447                                 'apihelp-query+alldeletedrevisions-param-miser-user-namespace',
 448                         ];
 449                 }
 450
 451                 return $ret;
 452         }
 453
 454         protected function getExamplesMessages() {
 455                 return [
 456                         'action=query&list=alldeletedrevisions&adruser=Example&adrlimit=50'
 457                                 => 'apihelp-query+alldeletedrevisions-example-user',
 458                         'action=query&list=alldeletedrevisions&adrdir=newer&adrnamespace=0&adrlimit=50'
 459                                 => 'apihelp-query+alldeletedrevisions-example-ns-main',
 460                 ];
 461         }
 462
 463         public function getHelpUrls() {
 464                 return 'https://www.mediawiki.org/wiki/API:Alldeletedrevisions';
 465         }
 466 }