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                 // This module operates in two modes:
  59                 // 'user': List deleted revs by a certain user
  60                 // 'all': List all deleted revs in NS
  61                 $mode = 'all';
  62                 if ( !is_null( $params['user'] ) ) {
  63                         $mode = 'user';
  64                 }
  65
  66                 if ( $mode == 'user' ) {
  67                         foreach ( [ 'from', 'to', 'prefix', 'excludeuser' ] as $param ) {
  68                                 if ( !is_null( $params[$param] ) ) {
  69                                         $p = $this->getModulePrefix();
  70                                         $this->dieUsage( "The '{$p}{$param}' parameter cannot be used with '{$p}user'",
  71                                                 'badparams' );
  72                                 }
  73                         }
  74                 } else {
  75                         foreach ( [ 'start', 'end' ] as $param ) {
  76                                 if ( !is_null( $params[$param] ) ) {
  77                                         $p = $this->getModulePrefix();
  78                                         $this->dieUsage( "The '{$p}{$param}' parameter may only be used with '{$p}user'",
  79                                                 'badparams' );
  80                                 }
  81                         }
  82                 }
  83
  84                 // If we're generating titles only, we can use DISTINCT for a better
  85                 // query. But we can't do that in 'user' mode (wrong index), and we can
  86                 // only do it when sorting ASC (because MySQL apparently can't use an
  87                 // index backwards for grouping even though it can for ORDER BY, WTF?)
  88                 $dir = $params['dir'];
  89                 $optimizeGenerateTitles = false;
  90                 if ( $mode === 'all' && $params['generatetitles'] && $resultPageSet !== null ) {
  91                         if ( $dir === 'newer' ) {
  92                                 $optimizeGenerateTitles = true;
  93                         } else {
  94                                 $p = $this->getModulePrefix();
  95                                 $this->setWarning( "For better performance when generating titles, set {$p}dir=newer" );
  96                         }
  97                 }
  98
  99                 $this->addTables( 'archive' );
 100                 if ( $resultPageSet === null ) {
 101                         $this->parseParameters( $params );
 102                         $this->addFields( Revision::selectArchiveFields() );
 103                         $this->addFields( [ 'ar_title', 'ar_namespace' ] );
 104                 } else {
 105                         $this->limit = $this->getParameter( 'limit' ) ?: 10;
 106                         $this->addFields( [ 'ar_title', 'ar_namespace' ] );
 107                         if ( $optimizeGenerateTitles ) {
 108                                 $this->addOption( 'DISTINCT' );
 109                         } else {
 110                                 $this->addFields( [ 'ar_timestamp', 'ar_rev_id', 'ar_id' ] );
 111                         }
 112                 }
 113
 114                 if ( $this->fld_tags ) {
 115                         $this->addTables( 'tag_summary' );
 116                         $this->addJoinConds(
 117                                 [ 'tag_summary' => [ 'LEFT JOIN', [ 'ar_rev_id=ts_rev_id' ] ] ]
 118                         );
 119                         $this->addFields( 'ts_tags' );
 120                 }
 121
 122                 if ( !is_null( $params['tag'] ) ) {
 123                         $this->addTables( 'change_tag' );
 124                         $this->addJoinConds(
 125                                 [ 'change_tag' => [ 'INNER JOIN', [ 'ar_rev_id=ct_rev_id' ] ] ]
 126                         );
 127                         $this->addWhereFld( 'ct_tag', $params['tag'] );
 128                 }
 129
 130                 if ( $this->fetchContent ) {
 131                         // Modern MediaWiki has the content for deleted revs in the 'text'
 132                         // table using fields old_text and old_flags. But revisions deleted
 133                         // pre-1.5 store the content in the 'archive' table directly using
 134                         // fields ar_text and ar_flags, and no corresponding 'text' row. So
 135                         // we have to LEFT JOIN and fetch all four fields.
 136                         $this->addTables( 'text' );
 137                         $this->addJoinConds(
 138                                 [ 'text' => [ 'LEFT JOIN', [ 'ar_text_id=old_id' ] ] ]
 139                         );
 140                         $this->addFields( [ 'ar_text', 'ar_flags', 'old_text', 'old_flags' ] );
 141
 142                         // This also means stricter restrictions
 143                         if ( !$user->isAllowedAny( 'undelete', 'deletedtext' ) ) {
 144                                 $this->dieUsage(
 145                                         'You don\'t have permission to view deleted revision content',
 146                                         'permissiondenied'
 147                                 );
 148                         }
 149                 }
 150
 151                 $miser_ns = null;
 152
 153                 if ( $mode == 'all' ) {
 154                         if ( $params['namespace'] !== null ) {
 155                                 $namespaces = $params['namespace'];
 156                                 $this->addWhereFld( 'ar_namespace', $namespaces );
 157                         } else {
 158                                 $namespaces = MWNamespace::getValidNamespaces();
 159                         }
 160
 161                         // For from/to/prefix, we have to consider the potential
 162                         // transformations of the title in all specified namespaces.
 163                         // Generally there will be only one transformation, but wikis with
 164                         // some namespaces case-sensitive could have two.
 165                         if ( $params['from'] !== null || $params['to'] !== null ) {
 166                                 $isDirNewer = ( $dir === 'newer' );
 167                                 $after = ( $isDirNewer ? '>=' : '<=' );
 168                                 $before = ( $isDirNewer ? '<=' : '>=' );
 169                                 $where = [];
 170                                 foreach ( $namespaces as $ns ) {
 171                                         $w = [];
 172                                         if ( $params['from'] !== null ) {
 173                                                 $w[] = 'ar_title' . $after .
 174                                                         $db->addQuotes( $this->titlePartToKey( $params['from'], $ns ) );
 175                                         }
 176                                         if ( $params['to'] !== null ) {
 177                                                 $w[] = 'ar_title' . $before .
 178                                                         $db->addQuotes( $this->titlePartToKey( $params['to'], $ns ) );
 179                                         }
 180                                         $w = $db->makeList( $w, LIST_AND );
 181                                         $where[$w][] = $ns;
 182                                 }
 183                                 if ( count( $where ) == 1 ) {
 184                                         $where = key( $where );
 185                                         $this->addWhere( $where );
 186                                 } else {
 187                                         $where2 = [];
 188                                         foreach ( $where as $w => $ns ) {
 189                                                 $where2[] = $db->makeList( [ $w, 'ar_namespace' => $ns ], LIST_AND );
 190                                         }
 191                                         $this->addWhere( $db->makeList( $where2, LIST_OR ) );
 192                                 }
 193                         }
 194
 195                         if ( isset( $params['prefix'] ) ) {
 196                                 $where = [];
 197                                 foreach ( $namespaces as $ns ) {
 198                                         $w = 'ar_title' . $db->buildLike(
 199                                                 $this->titlePartToKey( $params['prefix'], $ns ),
 200                                                 $db->anyString() );
 201                                         $where[$w][] = $ns;
 202                                 }
 203                                 if ( count( $where ) == 1 ) {
 204                                         $where = key( $where );
 205                                         $this->addWhere( $where );
 206                                 } else {
 207                                         $where2 = [];
 208                                         foreach ( $where as $w => $ns ) {
 209                                                 $where2[] = $db->makeList( [ $w, 'ar_namespace' => $ns ], LIST_AND );
 210                                         }
 211                                         $this->addWhere( $db->makeList( $where2, LIST_OR ) );
 212                                 }
 213                         }
 214                 } else {
 215                         if ( $this->getConfig()->get( 'MiserMode' ) ) {
 216                                 $miser_ns = $params['namespace'];
 217                         } else {
 218                                 $this->addWhereFld( 'ar_namespace', $params['namespace'] );
 219                         }
 220                         $this->addTimestampWhereRange( 'ar_timestamp', $dir, $params['start'], $params['end'] );
 221                 }
 222
 223                 if ( !is_null( $params['user'] ) ) {
 224                         $this->addWhereFld( 'ar_user_text', $params['user'] );
 225                 } elseif ( !is_null( $params['excludeuser'] ) ) {
 226                         $this->addWhere( 'ar_user_text != ' .
 227                                 $db->addQuotes( $params['excludeuser'] ) );
 228                 }
 229
 230                 if ( !is_null( $params['user'] ) || !is_null( $params['excludeuser'] ) ) {
 231                         // Paranoia: avoid brute force searches (bug 17342)
 232                         // (shouldn't be able to get here without 'deletedhistory', but
 233                         // check it again just in case)
 234                         if ( !$user->isAllowed( 'deletedhistory' ) ) {
 235                                 $bitmask = Revision::DELETED_USER;
 236                         } elseif ( !$user->isAllowedAny( 'suppressrevision', 'viewsuppressed' ) ) {
 237                                 $bitmask = Revision::DELETED_USER | Revision::DELETED_RESTRICTED;
 238                         } else {
 239                                 $bitmask = 0;
 240                         }
 241                         if ( $bitmask ) {
 242                                 $this->addWhere( $db->bitAnd( 'ar_deleted', $bitmask ) . " != $bitmask" );
 243                         }
 244                 }
 245
 246                 if ( !is_null( $params['continue'] ) ) {
 247                         $cont = explode( '|', $params['continue'] );
 248                         $op = ( $dir == 'newer' ? '>' : '<' );
 249                         if ( $optimizeGenerateTitles ) {
 250                                 $this->dieContinueUsageIf( count( $cont ) != 2 );
 251                                 $ns = intval( $cont[0] );
 252                                 $this->dieContinueUsageIf( strval( $ns ) !== $cont[0] );
 253                                 $title = $db->addQuotes( $cont[1] );
 254                                 $this->addWhere( "ar_namespace $op $ns OR " .
 255                                         "(ar_namespace = $ns AND ar_title $op= $title)" );
 256                         } elseif ( $mode == 'all' ) {
 257                                 $this->dieContinueUsageIf( count( $cont ) != 4 );
 258                                 $ns = intval( $cont[0] );
 259                                 $this->dieContinueUsageIf( strval( $ns ) !== $cont[0] );
 260                                 $title = $db->addQuotes( $cont[1] );
 261                                 $ts = $db->addQuotes( $db->timestamp( $cont[2] ) );
 262                                 $ar_id = (int)$cont[3];
 263                                 $this->dieContinueUsageIf( strval( $ar_id ) !== $cont[3] );
 264                                 $this->addWhere( "ar_namespace $op $ns OR " .
 265                                         "(ar_namespace = $ns AND " .
 266                                         "(ar_title $op $title OR " .
 267                                         "(ar_title = $title AND " .
 268                                         "(ar_timestamp $op $ts OR " .
 269                                         "(ar_timestamp = $ts AND " .
 270                                         "ar_id $op= $ar_id)))))" );
 271                         } else {
 272                                 $this->dieContinueUsageIf( count( $cont ) != 2 );
 273                                 $ts = $db->addQuotes( $db->timestamp( $cont[0] ) );
 274                                 $ar_id = (int)$cont[1];
 275                                 $this->dieContinueUsageIf( strval( $ar_id ) !== $cont[1] );
 276                                 $this->addWhere( "ar_timestamp $op $ts OR " .
 277                                         "(ar_timestamp = $ts AND " .
 278                                         "ar_id $op= $ar_id)" );
 279                         }
 280                 }
 281
 282                 $this->addOption( 'LIMIT', $this->limit + 1 );
 283
 284                 $sort = ( $dir == 'newer' ? '' : ' DESC' );
 285                 $orderby = [];
 286                 if ( $optimizeGenerateTitles ) {
 287                         // Targeting index name_title_timestamp
 288                         if ( $params['namespace'] === null || count( array_unique( $params['namespace'] ) ) > 1 ) {
 289                                 $orderby[] = "ar_namespace $sort";
 290                         }
 291                         $orderby[] = "ar_title $sort";
 292                 } elseif ( $mode == 'all' ) {
 293                         // Targeting index name_title_timestamp
 294                         if ( $params['namespace'] === null || count( array_unique( $params['namespace'] ) ) > 1 ) {
 295                                 $orderby[] = "ar_namespace $sort";
 296                         }
 297                         $orderby[] = "ar_title $sort";
 298                         $orderby[] = "ar_timestamp $sort";
 299                         $orderby[] = "ar_id $sort";
 300                 } else {
 301                         // Targeting index usertext_timestamp
 302                         // 'user' is always constant.
 303                         $orderby[] = "ar_timestamp $sort";
 304                         $orderby[] = "ar_id $sort";
 305                 }
 306                 $this->addOption( 'ORDER BY', $orderby );
 307
 308                 $res = $this->select( __METHOD__ );
 309                 $pageMap = []; // Maps ns&title to array index
 310                 $count = 0;
 311                 $nextIndex = 0;
 312                 $generated = [];
 313                 foreach ( $res as $row ) {
 314                         if ( ++$count > $this->limit ) {
 315                                 // We've had enough
 316                                 if ( $optimizeGenerateTitles ) {
 317                                         $this->setContinueEnumParameter( 'continue', "$row->ar_namespace|$row->ar_title" );
 318                                 } elseif ( $mode == 'all' ) {
 319                                         $this->setContinueEnumParameter( 'continue',
 320                                                 "$row->ar_namespace|$row->ar_title|$row->ar_timestamp|$row->ar_id"
 321                                         );
 322                                 } else {
 323                                         $this->setContinueEnumParameter( 'continue', "$row->ar_timestamp|$row->ar_id" );
 324                                 }
 325                                 break;
 326                         }
 327
 328                         // Miser mode namespace check
 329                         if ( $miser_ns !== null && !in_array( $row->ar_namespace, $miser_ns ) ) {
 330                                 continue;
 331                         }
 332
 333                         if ( $resultPageSet !== null ) {
 334                                 if ( $params['generatetitles'] ) {
 335                                         $key = "{$row->ar_namespace}:{$row->ar_title}";
 336                                         if ( !isset( $generated[$key] ) ) {
 337                                                 $generated[$key] = Title::makeTitle( $row->ar_namespace, $row->ar_title );
 338                                         }
 339                                 } else {
 340                                         $generated[] = $row->ar_rev_id;
 341                                 }
 342                         } else {
 343                                 $revision = Revision::newFromArchiveRow( $row );
 344                                 $rev = $this->extractRevisionInfo( $revision, $row );
 345
 346                                 if ( !isset( $pageMap[$row->ar_namespace][$row->ar_title] ) ) {
 347                                         $index = $nextIndex++;
 348                                         $pageMap[$row->ar_namespace][$row->ar_title] = $index;
 349                                         $title = $revision->getTitle();
 350                                         $a = [
 351                                                 'pageid' => $title->getArticleID(),
 352                                                 'revisions' => [ $rev ],
 353                                         ];
 354                                         ApiResult::setIndexedTagName( $a['revisions'], 'rev' );
 355                                         ApiQueryBase::addTitleInfo( $a, $title );
 356                                         $fit = $result->addValue( [ 'query', $this->getModuleName() ], $index, $a );
 357                                 } else {
 358                                         $index = $pageMap[$row->ar_namespace][$row->ar_title];
 359                                         $fit = $result->addValue(
 360                                                 [ 'query', $this->getModuleName(), $index, 'revisions' ],
 361                                                 null, $rev );
 362                                 }
 363                                 if ( !$fit ) {
 364                                         if ( $mode == 'all' ) {
 365                                                 $this->setContinueEnumParameter( 'continue',
 366                                                         "$row->ar_namespace|$row->ar_title|$row->ar_timestamp|$row->ar_id"
 367                                                 );
 368                                         } else {
 369                                                 $this->setContinueEnumParameter( 'continue', "$row->ar_timestamp|$row->ar_id" );
 370                                         }
 371                                         break;
 372                                 }
 373                         }
 374                 }
 375
 376                 if ( $resultPageSet !== null ) {
 377                         if ( $params['generatetitles'] ) {
 378                                 $resultPageSet->populateFromTitles( $generated );
 379                         } else {
 380                                 $resultPageSet->populateFromRevisionIDs( $generated );
 381                         }
 382                 } else {
 383                         $result->addIndexedTagName( [ 'query', $this->getModuleName() ], 'page' );
 384                 }
 385         }
 386
 387         public function getAllowedParams() {
 388                 $ret = parent::getAllowedParams() + [
 389                         'user' => [
 390                                 ApiBase::PARAM_TYPE => 'user'
 391                         ],
 392                         'namespace' => [
 393                                 ApiBase::PARAM_ISMULTI => true,
 394                                 ApiBase::PARAM_TYPE => 'namespace',
 395                         ],
 396                         'start' => [
 397                                 ApiBase::PARAM_TYPE => 'timestamp',
 398                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'useronly' ] ],
 399                         ],
 400                         'end' => [
 401                                 ApiBase::PARAM_TYPE => 'timestamp',
 402                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'useronly' ] ],
 403                         ],
 404                         'dir' => [
 405                                 ApiBase::PARAM_TYPE => [
 406                                         'newer',
 407                                         'older'
 408                                 ],
 409                                 ApiBase::PARAM_DFLT => 'older',
 410                                 ApiBase::PARAM_HELP_MSG => 'api-help-param-direction',
 411                         ],
 412                         'from' => [
 413                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'nonuseronly' ] ],
 414                         ],
 415                         'to' => [
 416                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'nonuseronly' ] ],
 417                         ],
 418                         'prefix' => [
 419                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'nonuseronly' ] ],
 420                         ],
 421                         'excludeuser' => [
 422                                 ApiBase::PARAM_TYPE => 'user',
 423                                 ApiBase::PARAM_HELP_MSG_INFO => [ [ 'nonuseronly' ] ],
 424                         ],
 425                         'tag' => null,
 426                         'continue' => [
 427                                 ApiBase::PARAM_HELP_MSG => 'api-help-param-continue',
 428                         ],
 429                         'generatetitles' => [
 430                                 ApiBase::PARAM_DFLT => false
 431                         ],
 432                 ];
 433
 434                 if ( $this->getConfig()->get( 'MiserMode' ) ) {
 435                         $ret['user'][ApiBase::PARAM_HELP_MSG_APPEND] = [
 436                                 'apihelp-query+alldeletedrevisions-param-miser-user-namespace',
 437                         ];
 438                         $ret['namespace'][ApiBase::PARAM_HELP_MSG_APPEND] = [
 439                                 'apihelp-query+alldeletedrevisions-param-miser-user-namespace',
 440                         ];
 441                 }
 442
 443                 return $ret;
 444         }
 445
 446         protected function getExamplesMessages() {
 447                 return [
 448                         'action=query&list=alldeletedrevisions&adruser=Example&adrlimit=50'
 449                                 => 'apihelp-query+alldeletedrevisions-example-user',
 450                         'action=query&list=alldeletedrevisions&adrdir=newer&adrlimit=50'
 451                                 => 'apihelp-query+alldeletedrevisions-example-ns-main',
 452                 ];
 453         }
 454
 455         public function getHelpUrls() {
 456                 return 'https://www.mediawiki.org/wiki/API:Alldeletedrevisions';
 457         }
 458 }