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                 // 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                 $this->addTables( 'archive' );
 107                 if ( $resultPageSet === null ) {
 108                         $this->parseParameters( $params );
 109                         $this->addFields( Revision::selectArchiveFields() );
 110                         $this->addFields( [ 'ar_title', 'ar_namespace' ] );
 111                 } else {
 112                         $this->limit = $this->getParameter( 'limit' ) ?: 10;
 113                         $this->addFields( [ 'ar_title', 'ar_namespace' ] );
 114                         if ( $optimizeGenerateTitles ) {
 115                                 $this->addOption( 'DISTINCT' );
 116                         } else {
 117                                 $this->addFields( [ 'ar_timestamp', 'ar_rev_id', 'ar_id' ] );
 118                         }
 119                 }
 120
 121                 if ( $this->fld_tags ) {
 122                         $this->addTables( 'tag_summary' );
 123                         $this->addJoinConds(
 124                                 [ 'tag_summary' => [ 'LEFT JOIN', [ 'ar_rev_id=ts_rev_id' ] ] ]
 125                         );
 126                         $this->addFields( 'ts_tags' );
 127                 }
 128
 129                 if ( !is_null( $params['tag'] ) ) {
 130                         $this->addTables( 'change_tag' );
 131                         $this->addJoinConds(
 132                                 [ 'change_tag' => [ 'INNER JOIN', [ 'ar_rev_id=ct_rev_id' ] ] ]
 133                         );
 134                         $this->addWhereFld( 'ct_tag', $params['tag'] );
 135                 }
 136
 137                 if ( $this->fetchContent ) {
 138                         // Modern MediaWiki has the content for deleted revs in the 'text'
 139                         // table using fields old_text and old_flags. But revisions deleted
 140                         // pre-1.5 store the content in the 'archive' table directly using
 141                         // fields ar_text and ar_flags, and no corresponding 'text' row. So
 142                         // we have to LEFT JOIN and fetch all four fields.
 143                         $this->addTables( 'text' );
 144                         $this->addJoinConds(
 145                                 [ 'text' => [ 'LEFT JOIN', [ 'ar_text_id=old_id' ] ] ]
 146                         );
 147                         $this->addFields( [ 'ar_text', 'ar_flags', 'old_text', 'old_flags' ] );
 148
 149                         // This also means stricter restrictions
 150                         $this->checkUserRightsAny( [ 'deletedtext', 'undelete' ] );
 151                 }
 152
 153                 $miser_ns = null;
 154
 155                 if ( $mode == 'all' ) {
 156                         if ( $params['namespace'] !== null ) {
 157                                 $namespaces = $params['namespace'];
 158                         } else {
 159                                 $namespaces = MWNamespace::getValidNamespaces();
 160                         }
 161                         $this->addWhereFld( 'ar_namespace', $namespaces );
 162
 163                         // For from/to/prefix, we have to consider the potential
 164                         // transformations of the title in all specified namespaces.
 165                         // Generally there will be only one transformation, but wikis with
 166                         // some namespaces case-sensitive could have two.
 167                         if ( $params['from'] !== null || $params['to'] !== null ) {
 168                                 $isDirNewer = ( $dir === 'newer' );
 169                                 $after = ( $isDirNewer ? '>=' : '<=' );
 170                                 $before = ( $isDirNewer ? '<=' : '>=' );
 171                                 $where = [];
 172                                 foreach ( $namespaces as $ns ) {
 173                                         $w = [];
 174                                         if ( $params['from'] !== null ) {
 175                                                 $w[] = 'ar_title' . $after .
 176                                                         $db->addQuotes( $this->titlePartToKey( $params['from'], $ns ) );
 177                                         }
 178                                         if ( $params['to'] !== null ) {
 179                                                 $w[] = 'ar_title' . $before .
 180                                                         $db->addQuotes( $this->titlePartToKey( $params['to'], $ns ) );
 181                                         }
 182                                         $w = $db->makeList( $w, LIST_AND );
 183                                         $where[$w][] = $ns;
 184                                 }
 185                                 if ( count( $where ) == 1 ) {
 186                                         $where = key( $where );
 187                                         $this->addWhere( $where );
 188                                 } else {
 189                                         $where2 = [];
 190                                         foreach ( $where as $w => $ns ) {
 191                                                 $where2[] = $db->makeList( [ $w, 'ar_namespace' => $ns ], LIST_AND );
 192                                         }
 193                                         $this->addWhere( $db->makeList( $where2, LIST_OR ) );
 194                                 }
 195                         }
 196
 197                         if ( isset( $params['prefix'] ) ) {
 198                                 $where = [];
 199                                 foreach ( $namespaces as $ns ) {
 200                                         $w = 'ar_title' . $db->buildLike(
 201                                                 $this->titlePartToKey( $params['prefix'], $ns ),
 202                                                 $db->anyString() );
 203                                         $where[$w][] = $ns;
 204                                 }
 205                                 if ( count( $where ) == 1 ) {
 206                                         $where = key( $where );
 207                                         $this->addWhere( $where );
 208                                 } else {
 209                                         $where2 = [];
 210                                         foreach ( $where as $w => $ns ) {
 211                                                 $where2[] = $db->makeList( [ $w, 'ar_namespace' => $ns ], LIST_AND );
 212                                         }
 213                                         $this->addWhere( $db->makeList( $where2, LIST_OR ) );
 214                                 }
 215                         }
 216                 } else {
 217                         if ( $this->getConfig()->get( 'MiserMode' ) ) {
 218                                 $miser_ns = $params['namespace'];
 219                         } else {
 220                                 $this->addWhereFld( 'ar_namespace', $params['namespace'] );
 221                         }
 222                         $this->addTimestampWhereRange( 'ar_timestamp', $dir, $params['start'], $params['end'] );
 223                 }
 224
 225                 if ( !is_null( $params['user'] ) ) {
 226                         $this->addWhereFld( 'ar_user_text', $params['user'] );
 227                 } elseif ( !is_null( $params['excludeuser'] ) ) {
 228                         $this->addWhere( 'ar_user_text != ' .
 229                                 $db->addQuotes( $params['excludeuser'] ) );
 230                 }
 231
 232                 if ( !is_null( $params['user'] ) || !is_null( $params['excludeuser'] ) ) {
 233                         // Paranoia: avoid brute force searches (bug 17342)
 234                         // (shouldn't be able to get here without 'deletedhistory', but
 235                         // check it again just in case)
 236                         if ( !$user->isAllowed( 'deletedhistory' ) ) {
 237                                 $bitmask = Revision::DELETED_USER;
 238                         } elseif ( !$user->isAllowedAny( 'suppressrevision', 'viewsuppressed' ) ) {
 239                                 $bitmask = Revision::DELETED_USER | Revision::DELETED_RESTRICTED;
 240                         } else {
 241                                 $bitmask = 0;
 242                         }
 243                         if ( $bitmask ) {
 244                                 $this->addWhere( $db->bitAnd( 'ar_deleted', $bitmask ) . " != $bitmask" );
 245                         }
 246                 }
 247
 248                 if ( !is_null( $params['continue'] ) ) {
 249                         $cont = explode( '|', $params['continue'] );
 250                         $op = ( $dir == 'newer' ? '>' : '<' );
 251                         if ( $optimizeGenerateTitles ) {
 252                                 $this->dieContinueUsageIf( count( $cont ) != 2 );
 253                                 $ns = intval( $cont[0] );
 254                                 $this->dieContinueUsageIf( strval( $ns ) !== $cont[0] );
 255                                 $title = $db->addQuotes( $cont[1] );
 256                                 $this->addWhere( "ar_namespace $op $ns OR " .
 257                                         "(ar_namespace = $ns AND ar_title $op= $title)" );
 258                         } elseif ( $mode == 'all' ) {
 259                                 $this->dieContinueUsageIf( count( $cont ) != 4 );
 260                                 $ns = intval( $cont[0] );
 261                                 $this->dieContinueUsageIf( strval( $ns ) !== $cont[0] );
 262                                 $title = $db->addQuotes( $cont[1] );
 263                                 $ts = $db->addQuotes( $db->timestamp( $cont[2] ) );
 264                                 $ar_id = (int)$cont[3];
 265                                 $this->dieContinueUsageIf( strval( $ar_id ) !== $cont[3] );
 266                                 $this->addWhere( "ar_namespace $op $ns OR " .
 267                                         "(ar_namespace = $ns AND " .
 268                                         "(ar_title $op $title OR " .
 269                                         "(ar_title = $title AND " .
 270                                         "(ar_timestamp $op $ts OR " .
 271                                         "(ar_timestamp = $ts AND " .
 272                                         "ar_id $op= $ar_id)))))" );
 273                         } else {
 274                                 $this->dieContinueUsageIf( count( $cont ) != 2 );
 275                                 $ts = $db->addQuotes( $db->timestamp( $cont[0] ) );
 276                                 $ar_id = (int)$cont[1];
 277                                 $this->dieContinueUsageIf( strval( $ar_id ) !== $cont[1] );
 278                                 $this->addWhere( "ar_timestamp $op $ts OR " .
 279                                         "(ar_timestamp = $ts AND " .
 280                                         "ar_id $op= $ar_id)" );
 281                         }
 282                 }
 283
 284                 $this->addOption( 'LIMIT', $this->limit + 1 );
 285
 286                 $sort = ( $dir == 'newer' ? '' : ' DESC' );
 287                 $orderby = [];
 288                 if ( $optimizeGenerateTitles ) {
 289                         // Targeting index name_title_timestamp
 290                         if ( $params['namespace'] === null || count( array_unique( $params['namespace'] ) ) > 1 ) {
 291                                 $orderby[] = "ar_namespace $sort";
 292                         }
 293                         $orderby[] = "ar_title $sort";
 294                 } elseif ( $mode == 'all' ) {
 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                         $orderby[] = "ar_timestamp $sort";
 301                         $orderby[] = "ar_id $sort";
 302                 } else {
 303                         // Targeting index usertext_timestamp
 304                         // 'user' is always constant.
 305                         $orderby[] = "ar_timestamp $sort";
 306                         $orderby[] = "ar_id $sort";
 307                 }
 308                 $this->addOption( 'ORDER BY', $orderby );
 309
 310                 $res = $this->select( __METHOD__ );
 311                 $pageMap = []; // Maps ns&title to array index
 312                 $count = 0;
 313                 $nextIndex = 0;
 314                 $generated = [];
 315                 foreach ( $res as $row ) {
 316                         if ( ++$count > $this->limit ) {
 317                                 // We've had enough
 318                                 if ( $optimizeGenerateTitles ) {
 319                                         $this->setContinueEnumParameter( 'continue', "$row->ar_namespace|$row->ar_title" );
 320                                 } elseif ( $mode == 'all' ) {
 321                                         $this->setContinueEnumParameter( 'continue',
 322                                                 "$row->ar_namespace|$row->ar_title|$row->ar_timestamp|$row->ar_id"
 323                                         );
 324                                 } else {
 325                                         $this->setContinueEnumParameter( 'continue', "$row->ar_timestamp|$row->ar_id" );
 326                                 }
 327                                 break;
 328                         }
 329
 330                         // Miser mode namespace check
 331                         if ( $miser_ns !== null && !in_array( $row->ar_namespace, $miser_ns ) ) {
 332                                 continue;
 333                         }
 334
 335                         if ( $resultPageSet !== null ) {
 336                                 if ( $params['generatetitles'] ) {
 337                                         $key = "{$row->ar_namespace}:{$row->ar_title}";
 338                                         if ( !isset( $generated[$key] ) ) {
 339                                                 $generated[$key] = Title::makeTitle( $row->ar_namespace, $row->ar_title );
 340                                         }
 341                                 } else {
 342                                         $generated[] = $row->ar_rev_id;
 343                                 }
 344                         } else {
 345                                 $revision = Revision::newFromArchiveRow( $row );
 346                                 $rev = $this->extractRevisionInfo( $revision, $row );
 347
 348                                 if ( !isset( $pageMap[$row->ar_namespace][$row->ar_title] ) ) {
 349                                         $index = $nextIndex++;
 350                                         $pageMap[$row->ar_namespace][$row->ar_title] = $index;
 351                                         $title = $revision->getTitle();
 352                                         $a = [
 353                                                 'pageid' => $title->getArticleID(),
 354                                                 'revisions' => [ $rev ],
 355                                         ];
 356                                         ApiResult::setIndexedTagName( $a['revisions'], 'rev' );
 357                                         ApiQueryBase::addTitleInfo( $a, $title );
 358                                         $fit = $result->addValue( [ 'query', $this->getModuleName() ], $index, $a );
 359                                 } else {
 360                                         $index = $pageMap[$row->ar_namespace][$row->ar_title];
 361                                         $fit = $result->addValue(
 362                                                 [ 'query', $this->getModuleName(), $index, 'revisions' ],
 363                                                 null, $rev );
 364                                 }
 365                                 if ( !$fit ) {
 366                                         if ( $mode == 'all' ) {
 367                                                 $this->setContinueEnumParameter( 'continue',
 368                                                         "$row->ar_namespace|$row->ar_title|$row->ar_timestamp|$row->ar_id"
 369                                                 );
 370                                         } else {
 371                                                 $this->setContinueEnumParameter( 'continue', "$row->ar_timestamp|$row->ar_id" );
 372                                         }
 373                                         break;
 374                                 }
 375                         }
 376                 }
 377
 378                 if ( $resultPageSet !== null ) {
 379                         if ( $params['generatetitles'] ) {
 380                                 $resultPageSet->populateFromTitles( $generated );
 381                         } else {
 382                                 $resultPageSet->populateFromRevisionIDs( $generated );
 383                         }
 384                 } else {
 385                         $result->addIndexedTagName( [ 'query', $this->getModuleName() ], 'page' );
 386                 }
 387         }
 388
 389         public function getAllowedParams() {
 390                 $ret = parent::getAllowedParams() + [
 391                         'user' => [
 392                                 ApiBase::PARAM_TYPE => 'user'
 393                         ],
 394                         'namespace' => [
 395                                 ApiBase::PARAM_ISMULTI => true,
 396                                 ApiBase::PARAM_TYPE => 'namespace',
 397                         ],
 398                         'start' => [
 399                                 ApiBase::PARAM_TYPE => 'timestamp',
 400                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'useronly' ] ],
 401                         ],
 402                         'end' => [
 403                                 ApiBase::PARAM_TYPE => 'timestamp',
 404                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'useronly' ] ],
 405                         ],
 406                         'dir' => [
 407                                 ApiBase::PARAM_TYPE => [
 408                                         'newer',
 409                                         'older'
 410                                 ],
 411                                 ApiBase::PARAM_DFLT => 'older',
 412                                 ApiBase::PARAM_HELP_MSG => 'api-help-param-direction',
 413                         ],
 414                         'from' => [
 415                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'nonuseronly' ] ],
 416                         ],
 417                         'to' => [
 418                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'nonuseronly' ] ],
 419                         ],
 420                         'prefix' => [
 421                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'nonuseronly' ] ],
 422                         ],
 423                         'excludeuser' => [
 424                                 ApiBase::PARAM_TYPE => 'user',
 425                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'nonuseronly' ] ],
 426                         ],
 427                         'tag' => null,
 428                         'continue' => [
 429                                 ApiBase::PARAM_HELP_MSG => 'api-help-param-continue',
 430                         ],
 431                         'generatetitles' => [
 432                                 ApiBase::PARAM_DFLT => false
 433                         ],
 434                 ];
 435
 436                 if ( $this->getConfig()->get( 'MiserMode' ) ) {
 437                         $ret['user'][ApiBase::PARAM_HELP_MSG_APPEND] = [
 438                                 'apihelp-query+alldeletedrevisions-param-miser-user-namespace',
 439                         ];
 440                         $ret['namespace'][ApiBase::PARAM_HELP_MSG_APPEND] = [
 441                                 'apihelp-query+alldeletedrevisions-param-miser-user-namespace',
 442                         ];
 443                 }
 444
 445                 return $ret;
 446         }
 447
 448         protected function getExamplesMessages() {
 449                 return [
 450                         'action=query&list=alldeletedrevisions&adruser=Example&adrlimit=50'
 451                                 => 'apihelp-query+alldeletedrevisions-example-user',
 452                         'action=query&list=alldeletedrevisions&adrdir=newer&adrnamespace=0&adrlimit=50'
 453                                 => 'apihelp-query+alldeletedrevisions-example-ns-main',
 454                 ];
 455         }
 456
 457         public function getHelpUrls() {
 458                 return 'https://www.mediawiki.org/wiki/API:Alldeletedrevisions';
 459         }
 460 }