maintenance/Maintenance.php

   1 <?php
   2 /**
   3  * This program is free software; you can redistribute it and/or modify
   4  * it under the terms of the GNU General Public License as published by
   5  * the Free Software Foundation; either version 2 of the License, or
   6  * (at your option) any later version.
   7  *
   8  * This program is distributed in the hope that it will be useful,
   9  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  10  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  11  * GNU General Public License for more details.
  12  *
  13  * You should have received a copy of the GNU General Public License along
  14  * with this program; if not, write to the Free Software Foundation, Inc.,
  15  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  16  * http://www.gnu.org/copyleft/gpl.html
  17  *
  18  * @file
  19  * @ingroup Maintenance
  20  * @defgroup Maintenance Maintenance
  21  */
  22
  23 define( 'MW_ENTRY_POINT', 'cli' );
  24
  25 // Bail on old versions of PHP, or if composer has not been run yet to install
  26 // dependencies.
  27 require_once __DIR__ . '/../includes/PHPVersionCheck.php';
  28 wfEntryPointCheck( 'text' );
  29
  30 use MediaWiki\Shell\Shell;
  31 use Wikimedia\Rdbms\IResultWrapper;
  32
  33 /**
  34  * @defgroup MaintenanceArchive Maintenance archives
  35  * @ingroup Maintenance
  36  */
  37
  38 // Define this so scripts can easily find doMaintenance.php
  39 define( 'RUN_MAINTENANCE_IF_MAIN', __DIR__ . '/doMaintenance.php' );
  40
  41 /**
  42  * @deprecated since 1.31
  43  */
  44 define( 'DO_MAINTENANCE', RUN_MAINTENANCE_IF_MAIN ); // original name, harmless
  45
  46 $maintClass = false;
  47
  48 // Some extensions rely on MW_INSTALL_PATH to find core files to include. Setting it here helps them
  49 // if they're included by a core script (like DatabaseUpdater) after Maintenance.php has already
  50 // been run.
  51 if ( strval( getenv( 'MW_INSTALL_PATH' ) ) === '' ) {
  52         putenv( 'MW_INSTALL_PATH=' . realpath( __DIR__ . '/..' ) );
  53 }
  54
  55 use Wikimedia\Rdbms\IDatabase;
  56 use MediaWiki\Logger\LoggerFactory;
  57 use MediaWiki\MediaWikiServices;
  58 use Wikimedia\Rdbms\LBFactory;
  59 use Wikimedia\Rdbms\IMaintainableDatabase;
  60
  61 /**
  62  * Abstract maintenance class for quickly writing and churning out
  63  * maintenance scripts with minimal effort. All that _must_ be defined
  64  * is the execute() method. See docs/maintenance.txt for more info
  65  * and a quick demo of how to use it.
  66  *
  67  * Terminology:
  68  *   params: registry of named values that may be passed to the script
  69  *   arg list: registry of positional values that may be passed to the script
  70  *   options: passed param values
  71  *   args: passed positional values
  72  *
  73  * In the command:
  74  *   mwscript somescript.php --foo=bar baz
  75  * foo is a param
  76  * bar is the option value of the option for param foo
  77  * baz is the arg value at index 0 in the arg list
  78  *
  79  * @since 1.16
  80  * @ingroup Maintenance
  81  */
  82 abstract class Maintenance {
  83         /**
  84          * Constants for DB access type
  85          * @see Maintenance::getDbType()
  86          */
  87         const DB_NONE = 0;
  88         const DB_STD = 1;
  89         const DB_ADMIN = 2;
  90
  91         // Const for getStdin()
  92         const STDIN_ALL = 'all';
  93
  94         /**
  95          * Array of desired/allowed params
  96          * @var array[]
  97          * @phan-var array<string,array{desc:string,require:bool,withArg:string,shortName:string,multiOccurrence:bool}>
  98          */
  99         protected $mParams = [];
 100
 101         // Array of mapping short parameters to long ones
 102         protected $mShortParamsMap = [];
 103
 104         // Array of desired/allowed args
 105         protected $mArgList = [];
 106
 107         // This is the list of options that were actually passed
 108         protected $mOptions = [];
 109
 110         // This is the list of arguments that were actually passed
 111         protected $mArgs = [];
 112
 113         // Allow arbitrary options to be passed, or only specified ones?
 114         protected $mAllowUnregisteredOptions = false;
 115
 116         // Name of the script currently running
 117         protected $mSelf;
 118
 119         // Special vars for params that are always used
 120         protected $mQuiet = false;
 121         protected $mDbUser, $mDbPass;
 122
 123         // A description of the script, children should change this via addDescription()
 124         protected $mDescription = '';
 125
 126         // Have we already loaded our user input?
 127         protected $mInputLoaded = false;
 128
 129         /**
 130          * Batch size. If a script supports this, they should set
 131          * a default with setBatchSize()
 132          *
 133          * @var int
 134          */
 135         protected $mBatchSize = null;
 136
 137         /**
 138          * Generic options added by addDefaultParams()
 139          * @var array[]
 140          * @phan-var array<string,array{desc:string,require:bool,withArg:string,shortName:string,multiOccurrence:bool}>
 141          */
 142         private $mGenericParameters = [];
 143         /**
 144          * Generic options which might or not be supported by the script
 145          * @var array[]
 146          * @phan-var array<string,array{desc:string,require:bool,withArg:string,shortName:string,multiOccurrence:bool}>
 147          */
 148         private $mDependantParameters = [];
 149
 150         /**
 151          * Used by getDB() / setDB()
 152          * @var IMaintainableDatabase
 153          */
 154         private $mDb = null;
 155
 156         /** @var float UNIX timestamp */
 157         private $lastReplicationWait = 0.0;
 158
 159         /**
 160          * Used when creating separate schema files.
 161          * @var resource
 162          */
 163         public $fileHandle;
 164
 165         /**
 166          * Accessible via getConfig()
 167          *
 168          * @var Config
 169          */
 170         private $config;
 171
 172         /**
 173          * @see Maintenance::requireExtension
 174          * @var array
 175          */
 176         private $requiredExtensions = [];
 177
 178         /**
 179          * Used to read the options in the order they were passed.
 180          * Useful for option chaining (Ex. dumpBackup.php). It will
 181          * be an empty array if the options are passed in through
 182          * loadParamsAndArgs( $self, $opts, $args ).
 183          *
 184          * This is an array of arrays where
 185          * 0 => the option and 1 => parameter value.
 186          *
 187          * @var array
 188          */
 189         public $orderedOptions = [];
 190
 191         /**
 192          * Default constructor. Children should call this *first* if implementing
 193          * their own constructors
 194          */
 195         public function __construct() {
 196                 global $IP;
 197                 $IP = getenv( 'MW_INSTALL_PATH' );
 198
 199                 $this->addDefaultParams();
 200                 register_shutdown_function( [ $this, 'outputChanneled' ], false );
 201         }
 202
 203         /**
 204          * Should we execute the maintenance script, or just allow it to be included
 205          * as a standalone class? It checks that the call stack only includes this
 206          * function and "requires" (meaning was called from the file scope)
 207          *
 208          * @return bool
 209          */
 210         public static function shouldExecute() {
 211                 global $wgCommandLineMode;
 212
 213                 if ( !function_exists( 'debug_backtrace' ) ) {
 214                         // If someone has a better idea...
 215                         return $wgCommandLineMode;
 216                 }
 217
 218                 $bt = debug_backtrace();
 219                 $count = count( $bt );
 220                 if ( $count < 2 ) {
 221                         return false; // sanity
 222                 }
 223                 if ( $bt[0]['class'] !== self::class || $bt[0]['function'] !== 'shouldExecute' ) {
 224                         return false; // last call should be to this function
 225                 }
 226                 $includeFuncs = [ 'require_once', 'require', 'include', 'include_once' ];
 227                 for ( $i = 1; $i < $count; $i++ ) {
 228                         if ( !in_array( $bt[$i]['function'], $includeFuncs ) ) {
 229                                 return false; // previous calls should all be "requires"
 230                         }
 231                 }
 232
 233                 return true;
 234         }
 235
 236         /**
 237          * Do the actual work. All child classes will need to implement this
 238          *
 239          * @return bool|null|void True for success, false for failure. Not returning
 240          *   a value, or returning null, is also interpreted as success. Returning
 241          *   false for failure will cause doMaintenance.php to exit the process
 242          *   with a non-zero exit status.
 243          */
 244         abstract public function execute();
 245
 246         /**
 247          * Checks to see if a particular option in supported.  Normally this means it
 248          * has been registered by the script via addOption.
 249          * @param string $name The name of the option
 250          * @return bool true if the option exists, false otherwise
 251          */
 252         protected function supportsOption( $name ) {
 253                 return isset( $this->mParams[$name] );
 254         }
 255
 256         /**
 257          * Add a parameter to the script. Will be displayed on --help
 258          * with the associated description
 259          *
 260          * @param string $name The name of the param (help, version, etc)
 261          * @param string $description The description of the param to show on --help
 262          * @param bool $required Is the param required?
 263          * @param bool $withArg Is an argument required with this option?
 264          * @param string|bool $shortName Character to use as short name
 265          * @param bool $multiOccurrence Can this option be passed multiple times?
 266          */
 267         protected function addOption( $name, $description, $required = false,
 268                 $withArg = false, $shortName = false, $multiOccurrence = false
 269         ) {
 270                 $this->mParams[$name] = [
 271                         'desc' => $description,
 272                         'require' => $required,
 273                         'withArg' => $withArg,
 274                         'shortName' => $shortName,
 275                         'multiOccurrence' => $multiOccurrence
 276                 ];
 277
 278                 if ( $shortName !== false ) {
 279                         $this->mShortParamsMap[$shortName] = $name;
 280                 }
 281         }
 282
 283         /**
 284          * Checks to see if a particular option exists.
 285          * @param string $name The name of the option
 286          * @return bool
 287          */
 288         protected function hasOption( $name ) {
 289                 return isset( $this->mOptions[$name] );
 290         }
 291
 292         /**
 293          * Get an option, or return the default.
 294          *
 295          * If the option was added to support multiple occurrences,
 296          * this will return an array.
 297          *
 298          * @param string $name The name of the param
 299          * @param mixed|null $default Anything you want, default null
 300          * @return mixed
 301          */
 302         protected function getOption( $name, $default = null ) {
 303                 if ( $this->hasOption( $name ) ) {
 304                         return $this->mOptions[$name];
 305                 } else {
 306                         // Set it so we don't have to provide the default again
 307                         $this->mOptions[$name] = $default;
 308
 309                         return $this->mOptions[$name];
 310                 }
 311         }
 312
 313         /**
 314          * Add some args that are needed
 315          * @param string $arg Name of the arg, like 'start'
 316          * @param string $description Short description of the arg
 317          * @param bool $required Is this required?
 318          */
 319         protected function addArg( $arg, $description, $required = true ) {
 320                 $this->mArgList[] = [
 321                         'name' => $arg,
 322                         'desc' => $description,
 323                         'require' => $required
 324                 ];
 325         }
 326
 327         /**
 328          * Remove an option.  Useful for removing options that won't be used in your script.
 329          * @param string $name The option to remove.
 330          */
 331         protected function deleteOption( $name ) {
 332                 unset( $this->mParams[$name] );
 333         }
 334
 335         /**
 336          * Sets whether to allow unregistered options, which are options passed to
 337          * a script that do not match an expected parameter.
 338          * @param bool $allow Should we allow?
 339          */
 340         protected function setAllowUnregisteredOptions( $allow ) {
 341                 $this->mAllowUnregisteredOptions = $allow;
 342         }
 343
 344         /**
 345          * Set the description text.
 346          * @param string $text The text of the description
 347          */
 348         protected function addDescription( $text ) {
 349                 $this->mDescription = $text;
 350         }
 351
 352         /**
 353          * Does a given argument exist?
 354          * @param int $argId The integer value (from zero) for the arg
 355          * @return bool
 356          */
 357         protected function hasArg( $argId = 0 ) {
 358                 if ( func_num_args() === 0 ) {
 359                         wfDeprecated( __METHOD__ . ' without an $argId', '1.33' );
 360                 }
 361
 362                 return isset( $this->mArgs[$argId] );
 363         }
 364
 365         /**
 366          * Get an argument.
 367          * @param int $argId The integer value (from zero) for the arg
 368          * @param mixed|null $default The default if it doesn't exist
 369          * @return mixed
 370          */
 371         protected function getArg( $argId = 0, $default = null ) {
 372                 if ( func_num_args() === 0 ) {
 373                         wfDeprecated( __METHOD__ . ' without an $argId', '1.33' );
 374                 }
 375
 376                 return $this->mArgs[$argId] ?? $default;
 377         }
 378
 379         /**
 380          * Returns batch size
 381          *
 382          * @since 1.31
 383          *
 384          * @return int|null
 385          */
 386         protected function getBatchSize() {
 387                 return $this->mBatchSize;
 388         }
 389
 390         /**
 391          * Set the batch size.
 392          * @param int $s The number of operations to do in a batch
 393          */
 394         protected function setBatchSize( $s = 0 ) {
 395                 $this->mBatchSize = $s;
 396
 397                 // If we support $mBatchSize, show the option.
 398                 // Used to be in addDefaultParams, but in order for that to
 399                 // work, subclasses would have to call this function in the constructor
 400                 // before they called parent::__construct which is just weird
 401                 // (and really wasn't done).
 402                 if ( $this->mBatchSize ) {
 403                         $this->addOption( 'batch-size', 'Run this many operations ' .
 404                                 'per batch, default: ' . $this->mBatchSize, false, true );
 405                         if ( isset( $this->mParams['batch-size'] ) ) {
 406                                 // This seems a little ugly...
 407                                 $this->mDependantParameters['batch-size'] = $this->mParams['batch-size'];
 408                         }
 409                 }
 410         }
 411
 412         /**
 413          * Get the script's name
 414          * @return string
 415          */
 416         public function getName() {
 417                 return $this->mSelf;
 418         }
 419
 420         /**
 421          * Return input from stdin.
 422          * @param int|null $len The number of bytes to read. If null, just return the handle.
 423          *   Maintenance::STDIN_ALL returns the full length
 424          * @return mixed
 425          */
 426         protected function getStdin( $len = null ) {
 427                 if ( $len == self::STDIN_ALL ) {
 428                         return file_get_contents( 'php://stdin' );
 429                 }
 430                 $f = fopen( 'php://stdin', 'rt' );
 431                 if ( !$len ) {
 432                         return $f;
 433                 }
 434                 $input = fgets( $f, $len );
 435                 fclose( $f );
 436
 437                 return rtrim( $input );
 438         }
 439
 440         /**
 441          * @return bool
 442          */
 443         public function isQuiet() {
 444                 return $this->mQuiet;
 445         }
 446
 447         /**
 448          * Throw some output to the user. Scripts can call this with no fears,
 449          * as we handle all --quiet stuff here
 450          * @param string $out The text to show to the user
 451          * @param mixed|null $channel Unique identifier for the channel. See function outputChanneled.
 452          */
 453         protected function output( $out, $channel = null ) {
 454                 // This is sometimes called very early, before Setup.php is included.
 455                 if ( class_exists( MediaWikiServices::class ) ) {
 456                         // Try to periodically flush buffered metrics to avoid OOMs
 457                         $stats = MediaWikiServices::getInstance()->getStatsdDataFactory();
 458                         if ( $stats->getDataCount() > 1000 ) {
 459                                 MediaWiki::emitBufferedStatsdData( $stats, $this->getConfig() );
 460                         }
 461                 }
 462
 463                 if ( $this->mQuiet ) {
 464                         return;
 465                 }
 466                 if ( $channel === null ) {
 467                         $this->cleanupChanneled();
 468                         print $out;
 469                 } else {
 470                         $out = preg_replace( '/\n\z/', '', $out );
 471                         $this->outputChanneled( $out, $channel );
 472                 }
 473         }
 474
 475         /**
 476          * Throw an error to the user. Doesn't respect --quiet, so don't use
 477          * this for non-error output
 478          * @param string $err The error to display
 479          * @param int $die Deprecated since 1.31, use Maintenance::fatalError() instead
 480          */
 481         protected function error( $err, $die = 0 ) {
 482                 if ( intval( $die ) !== 0 ) {
 483                         wfDeprecated( __METHOD__ . '( $err, $die )', '1.31' );
 484                         $this->fatalError( $err, intval( $die ) );
 485                 }
 486                 $this->outputChanneled( false );
 487                 if (
 488                         ( PHP_SAPI == 'cli' || PHP_SAPI == 'phpdbg' ) &&
 489                         !defined( 'MW_PHPUNIT_TEST' )
 490                 ) {
 491                         fwrite( STDERR, $err . "\n" );
 492                 } else {
 493                         print $err;
 494                 }
 495         }
 496
 497         /**
 498          * Output a message and terminate the current script.
 499          *
 500          * @param string $msg Error message
 501          * @param int $exitCode PHP exit status. Should be in range 1-254.
 502          * @since 1.31
 503          */
 504         protected function fatalError( $msg, $exitCode = 1 ) {
 505                 $this->error( $msg );
 506                 exit( $exitCode );
 507         }
 508
 509         private $atLineStart = true;
 510         private $lastChannel = null;
 511
 512         /**
 513          * Clean up channeled output.  Output a newline if necessary.
 514          */
 515         public function cleanupChanneled() {
 516                 if ( !$this->atLineStart ) {
 517                         print "\n";
 518                         $this->atLineStart = true;
 519                 }
 520         }
 521
 522         /**
 523          * Message outputter with channeled message support. Messages on the
 524          * same channel are concatenated, but any intervening messages in another
 525          * channel start a new line.
 526          * @param string $msg The message without trailing newline
 527          * @param string|null $channel Channel identifier or null for no
 528          *     channel. Channel comparison uses ===.
 529          */
 530         public function outputChanneled( $msg, $channel = null ) {
 531                 if ( $msg === false ) {
 532                         $this->cleanupChanneled();
 533
 534                         return;
 535                 }
 536
 537                 // End the current line if necessary
 538                 if ( !$this->atLineStart && $channel !== $this->lastChannel ) {
 539                         print "\n";
 540                 }
 541
 542                 print $msg;
 543
 544                 $this->atLineStart = false;
 545                 if ( $channel === null ) {
 546                         // For unchanneled messages, output trailing newline immediately
 547                         print "\n";
 548                         $this->atLineStart = true;
 549                 }
 550                 $this->lastChannel = $channel;
 551         }
 552
 553         /**
 554          * Does the script need different DB access? By default, we give Maintenance
 555          * scripts normal rights to the DB. Sometimes, a script needs admin rights
 556          * access for a reason and sometimes they want no access. Subclasses should
 557          * override and return one of the following values, as needed:
 558          *    Maintenance::DB_NONE  -  For no DB access at all
 559          *    Maintenance::DB_STD   -  For normal DB access, default
 560          *    Maintenance::DB_ADMIN -  For admin DB access
 561          * @return int
 562          */
 563         public function getDbType() {
 564                 return self::DB_STD;
 565         }
 566
 567         /**
 568          * Add the default parameters to the scripts
 569          */
 570         protected function addDefaultParams() {
 571                 # Generic (non script dependant) options:
 572
 573                 $this->addOption( 'help', 'Display this help message', false, false, 'h' );
 574                 $this->addOption( 'quiet', 'Whether to suppress non-error output', false, false, 'q' );
 575                 $this->addOption( 'conf', 'Location of LocalSettings.php, if not default', false, true );
 576                 $this->addOption( 'wiki', 'For specifying the wiki ID', false, true );
 577                 $this->addOption( 'globals', 'Output globals at the end of processing for debugging' );
 578                 $this->addOption(
 579                         'memory-limit',
 580                         'Set a specific memory limit for the script, '
 581                                 . '"max" for no limit or "default" to avoid changing it',
 582                         false,
 583                         true
 584                 );
 585                 $this->addOption( 'server', "The protocol and server name to use in URLs, e.g. " .
 586                         "http://en.wikipedia.org. This is sometimes necessary because " .
 587                         "server name detection may fail in command line scripts.", false, true );
 588                 $this->addOption( 'profiler', 'Profiler output format (usually "text")', false, true );
 589                 // This is named --mwdebug, because --debug would conflict in the phpunit.php CLI script.
 590                 $this->addOption( 'mwdebug', 'Enable built-in MediaWiki development settings', false, true );
 591
 592                 # Save generic options to display them separately in help
 593                 $this->mGenericParameters = $this->mParams;
 594
 595                 # Script dependant options:
 596
 597                 // If we support a DB, show the options
 598                 if ( $this->getDbType() > 0 ) {
 599                         $this->addOption( 'dbuser', 'The DB user to use for this script', false, true );
 600                         $this->addOption( 'dbpass', 'The password to use for this script', false, true );
 601                         $this->addOption( 'dbgroupdefault', 'The default DB group to use.', false, true );
 602                 }
 603
 604                 # Save additional script dependant options to display
 605                 #  them separately in help
 606                 $this->mDependantParameters = array_diff_key( $this->mParams, $this->mGenericParameters );
 607         }
 608
 609         /**
 610          * @since 1.24
 611          * @return Config
 612          */
 613         public function getConfig() {
 614                 if ( $this->config === null ) {
 615                         $this->config = MediaWikiServices::getInstance()->getMainConfig();
 616                 }
 617
 618                 return $this->config;
 619         }
 620
 621         /**
 622          * @since 1.24
 623          * @param Config $config
 624          */
 625         public function setConfig( Config $config ) {
 626                 $this->config = $config;
 627         }
 628
 629         /**
 630          * Indicate that the specified extension must be
 631          * loaded before the script can run.
 632          *
 633          * This *must* be called in the constructor.
 634          *
 635          * @since 1.28
 636          * @param string $name
 637          */
 638         protected function requireExtension( $name ) {
 639                 $this->requiredExtensions[] = $name;
 640         }
 641
 642         /**
 643          * Verify that the required extensions are installed
 644          *
 645          * @since 1.28
 646          */
 647         public function checkRequiredExtensions() {
 648                 $registry = ExtensionRegistry::getInstance();
 649                 $missing = [];
 650                 foreach ( $this->requiredExtensions as $name ) {
 651                         if ( !$registry->isLoaded( $name ) ) {
 652                                 $missing[] = $name;
 653                         }
 654                 }
 655
 656                 if ( $missing ) {
 657                         $joined = implode( ', ', $missing );
 658                         $msg = "The following extensions are required to be installed "
 659                                 . "for this script to run: $joined. Please enable them and then try again.";
 660                         $this->fatalError( $msg );
 661                 }
 662         }
 663
 664         /**
 665          * Set triggers like when to try to run deferred updates
 666          * @since 1.28
 667          */
 668         public function setAgentAndTriggers() {
 669                 if ( function_exists( 'posix_getpwuid' ) ) {
 670                         $agent = posix_getpwuid( posix_geteuid() )['name'];
 671                 } else {
 672                         $agent = 'sysadmin';
 673                 }
 674                 $agent .= '@' . wfHostname();
 675
 676                 $lbFactory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
 677                 // Add a comment for easy SHOW PROCESSLIST interpretation
 678                 $lbFactory->setAgentName(
 679                         mb_strlen( $agent ) > 15 ? mb_substr( $agent, 0, 15 ) . '...' : $agent
 680                 );
 681                 self::setLBFactoryTriggers( $lbFactory, $this->getConfig() );
 682         }
 683
 684         /**
 685          * @param LBFactory $LBFactory
 686          * @param Config $config
 687          * @since 1.28
 688          */
 689         public static function setLBFactoryTriggers( LBFactory $LBFactory, Config $config ) {
 690                 $services = MediaWikiServices::getInstance();
 691                 $stats = $services->getStatsdDataFactory();
 692                 // Hook into period lag checks which often happen in long-running scripts
 693                 $lbFactory = $services->getDBLoadBalancerFactory();
 694                 $lbFactory->setWaitForReplicationListener(
 695                         __METHOD__,
 696                         function () use ( $stats, $config ) {
 697                                 // Check config in case of JobRunner and unit tests
 698                                 if ( $config->get( 'CommandLineMode' ) ) {
 699                                         DeferredUpdates::tryOpportunisticExecute( 'run' );
 700                                 }
 701                                 // Try to periodically flush buffered metrics to avoid OOMs
 702                                 MediaWiki::emitBufferedStatsdData( $stats, $config );
 703                         }
 704                 );
 705                 // Check for other windows to run them. A script may read or do a few writes
 706                 // to the master but mostly be writing to something else, like a file store.
 707                 $lbFactory->getMainLB()->setTransactionListener(
 708                         __METHOD__,
 709                         function ( $trigger ) use ( $stats, $config ) {
 710                                 // Check config in case of JobRunner and unit tests
 711                                 if ( $config->get( 'CommandLineMode' ) && $trigger === IDatabase::TRIGGER_COMMIT ) {
 712                                         DeferredUpdates::tryOpportunisticExecute( 'run' );
 713                                 }
 714                                 // Try to periodically flush buffered metrics to avoid OOMs
 715                                 MediaWiki::emitBufferedStatsdData( $stats, $config );
 716                         }
 717                 );
 718         }
 719
 720         /**
 721          * Run a child maintenance script. Pass all of the current arguments
 722          * to it.
 723          * @param string $maintClass A name of a child maintenance class
 724          * @param string|null $classFile Full path of where the child is
 725          * @return Maintenance
 726          */
 727         public function runChild( $maintClass, $classFile = null ) {
 728                 // Make sure the class is loaded first
 729                 if ( !class_exists( $maintClass ) ) {
 730                         if ( $classFile ) {
 731                                 require_once $classFile;
 732                         }
 733                         if ( !class_exists( $maintClass ) ) {
 734                                 $this->error( "Cannot spawn child: $maintClass" );
 735                         }
 736                 }
 737
 738                 /**
 739                  * @var Maintenance $child
 740                  */
 741                 $child = new $maintClass();
 742                 $child->loadParamsAndArgs( $this->mSelf, $this->mOptions, $this->mArgs );
 743                 if ( !is_null( $this->mDb ) ) {
 744                         $child->setDB( $this->mDb );
 745                 }
 746
 747                 return $child;
 748         }
 749
 750         /**
 751          * Do some sanity checking and basic setup
 752          */
 753         public function setup() {
 754                 global $IP, $wgCommandLineMode;
 755
 756                 # Abort if called from a web server
 757                 # wfIsCLI() is not available yet
 758                 if ( PHP_SAPI !== 'cli' && PHP_SAPI !== 'phpdbg' ) {
 759                         $this->fatalError( 'This script must be run from the command line' );
 760                 }
 761
 762                 if ( $IP === null ) {
 763                         $this->fatalError( "\$IP not set, aborting!\n" .
 764                                 '(Did you forget to call parent::__construct() in your maintenance script?)' );
 765                 }
 766
 767                 # Make sure we can handle script parameters
 768                 if ( !defined( 'HPHP_VERSION' ) && !ini_get( 'register_argc_argv' ) ) {
 769                         $this->fatalError( 'Cannot get command line arguments, register_argc_argv is set to false' );
 770                 }
 771
 772                 // Send PHP warnings and errors to stderr instead of stdout.
 773                 // This aids in diagnosing problems, while keeping messages
 774                 // out of redirected output.
 775                 if ( ini_get( 'display_errors' ) ) {
 776                         ini_set( 'display_errors', 'stderr' );
 777                 }
 778
 779                 $this->loadParamsAndArgs();
 780
 781                 # Set the memory limit
 782                 # Note we need to set it again later in cache LocalSettings changed it
 783                 $this->adjustMemoryLimit();
 784
 785                 # Set max execution time to 0 (no limit). PHP.net says that
 786                 # "When running PHP from the command line the default setting is 0."
 787                 # But sometimes this doesn't seem to be the case.
 788                 ini_set( 'max_execution_time', 0 );
 789
 790                 # Define us as being in MediaWiki
 791                 define( 'MEDIAWIKI', true );
 792
 793                 $wgCommandLineMode = true;
 794
 795                 # Turn off output buffering if it's on
 796                 while ( ob_get_level() > 0 ) {
 797                         ob_end_flush();
 798                 }
 799         }
 800
 801         /**
 802          * Normally we disable the memory_limit when running admin scripts.
 803          * Some scripts may wish to actually set a limit, however, to avoid
 804          * blowing up unexpectedly. We also support a --memory-limit option,
 805          * to allow sysadmins to explicitly set one if they'd prefer to override
 806          * defaults (or for people using Suhosin which yells at you for trying
 807          * to disable the limits)
 808          * @return string
 809          */
 810         public function memoryLimit() {
 811                 $limit = $this->getOption( 'memory-limit', 'max' );
 812                 $limit = trim( $limit, "\" '" ); // trim quotes in case someone misunderstood
 813                 return $limit;
 814         }
 815
 816         /**
 817          * Adjusts PHP's memory limit to better suit our needs, if needed.
 818          */
 819         protected function adjustMemoryLimit() {
 820                 $limit = $this->memoryLimit();
 821                 if ( $limit == 'max' ) {
 822                         $limit = -1; // no memory limit
 823                 }
 824                 if ( $limit != 'default' ) {
 825                         ini_set( 'memory_limit', $limit );
 826                 }
 827         }
 828
 829         /**
 830          * Activate the profiler (assuming $wgProfiler is set)
 831          */
 832         protected function activateProfiler() {
 833                 global $wgProfiler, $wgProfileLimit, $wgTrxProfilerLimits;
 834
 835                 $output = $this->getOption( 'profiler' );
 836                 if ( !$output ) {
 837                         return;
 838                 }
 839
 840                 if ( is_array( $wgProfiler ) && isset( $wgProfiler['class'] ) ) {
 841                         $class = $wgProfiler['class'];
 842                         /** @var Profiler $profiler */
 843                         $profiler = new $class(
 844                                 [ 'sampling' => 1, 'output' => [ $output ] ]
 845                                         + $wgProfiler
 846                                         + [ 'threshold' => $wgProfileLimit ]
 847                         );
 848                         $profiler->setAllowOutput();
 849                         Profiler::replaceStubInstance( $profiler );
 850                 }
 851
 852                 $trxProfiler = Profiler::instance()->getTransactionProfiler();
 853                 $trxProfiler->setLogger( LoggerFactory::getInstance( 'DBPerformance' ) );
 854                 $trxProfiler->setExpectations( $wgTrxProfilerLimits['Maintenance'], __METHOD__ );
 855         }
 856
 857         /**
 858          * Clear all params and arguments.
 859          */
 860         public function clearParamsAndArgs() {
 861                 $this->mOptions = [];
 862                 $this->mArgs = [];
 863                 $this->mInputLoaded = false;
 864         }
 865
 866         /**
 867          * Load params and arguments from a given array
 868          * of command-line arguments
 869          *
 870          * @since 1.27
 871          * @param array $argv
 872          */
 873         public function loadWithArgv( $argv ) {
 874                 $options = [];
 875                 $args = [];
 876                 $this->orderedOptions = [];
 877
 878                 # Parse arguments
 879                 for ( $arg = reset( $argv ); $arg !== false; $arg = next( $argv ) ) {
 880                         if ( $arg == '--' ) {
 881                                 # End of options, remainder should be considered arguments
 882                                 $arg = next( $argv );
 883                                 while ( $arg !== false ) {
 884                                         $args[] = $arg;
 885                                         $arg = next( $argv );
 886                                 }
 887                                 break;
 888                         } elseif ( substr( $arg, 0, 2 ) == '--' ) {
 889                                 # Long options
 890                                 $option = substr( $arg, 2 );
 891                                 if ( isset( $this->mParams[$option] ) && $this->mParams[$option]['withArg'] ) {
 892                                         $param = next( $argv );
 893                                         if ( $param === false ) {
 894                                                 $this->error( "\nERROR: $option parameter needs a value after it\n" );
 895                                                 $this->maybeHelp( true );
 896                                         }
 897
 898                                         $this->setParam( $options, $option, $param );
 899                                 } else {
 900                                         $bits = explode( '=', $option, 2 );
 901                                         $this->setParam( $options, $bits[0], $bits[1] ?? 1 );
 902                                 }
 903                         } elseif ( $arg == '-' ) {
 904                                 # Lonely "-", often used to indicate stdin or stdout.
 905                                 $args[] = $arg;
 906                         } elseif ( substr( $arg, 0, 1 ) == '-' ) {
 907                                 # Short options
 908                                 $argLength = strlen( $arg );
 909                                 for ( $p = 1; $p < $argLength; $p++ ) {
 910                                         $option = $arg[$p];
 911                                         if ( !isset( $this->mParams[$option] ) && isset( $this->mShortParamsMap[$option] ) ) {
 912                                                 $option = $this->mShortParamsMap[$option];
 913                                         }
 914
 915                                         if ( isset( $this->mParams[$option]['withArg'] ) && $this->mParams[$option]['withArg'] ) {
 916                                                 $param = next( $argv );
 917                                                 if ( $param === false ) {
 918                                                         $this->error( "\nERROR: $option parameter needs a value after it\n" );
 919                                                         $this->maybeHelp( true );
 920                                                 }
 921                                                 $this->setParam( $options, $option, $param );
 922                                         } else {
 923                                                 $this->setParam( $options, $option, 1 );
 924                                         }
 925                                 }
 926                         } else {
 927                                 $args[] = $arg;
 928                         }
 929                 }
 930
 931                 $this->mOptions = $options;
 932                 $this->mArgs = $args;
 933                 $this->loadSpecialVars();
 934                 $this->mInputLoaded = true;
 935         }
 936
 937         /**
 938          * Helper function used solely by loadParamsAndArgs
 939          * to prevent code duplication
 940          *
 941          * This sets the param in the options array based on
 942          * whether or not it can be specified multiple times.
 943          *
 944          * @since 1.27
 945          * @param array $options
 946          * @param string $option
 947          * @param mixed $value
 948          */
 949         private function setParam( &$options, $option, $value ) {
 950                 $this->orderedOptions[] = [ $option, $value ];
 951
 952                 if ( isset( $this->mParams[$option] ) ) {
 953                         $multi = $this->mParams[$option]['multiOccurrence'];
 954                 } else {
 955                         $multi = false;
 956                 }
 957                 $exists = array_key_exists( $option, $options );
 958                 if ( $multi && $exists ) {
 959                         $options[$option][] = $value;
 960                 } elseif ( $multi ) {
 961                         $options[$option] = [ $value ];
 962                 } elseif ( !$exists ) {
 963                         $options[$option] = $value;
 964                 } else {
 965                         $this->error( "\nERROR: $option parameter given twice\n" );
 966                         $this->maybeHelp( true );
 967                 }
 968         }
 969
 970         /**
 971          * Process command line arguments
 972          * $mOptions becomes an array with keys set to the option names
 973          * $mArgs becomes a zero-based array containing the non-option arguments
 974          *
 975          * @param string|null $self The name of the script, if any
 976          * @param array|null $opts An array of options, in form of key=>value
 977          * @param array|null $args An array of command line arguments
 978          */
 979         public function loadParamsAndArgs( $self = null, $opts = null, $args = null ) {
 980                 # If we were given opts or args, set those and return early
 981                 if ( $self ) {
 982                         $this->mSelf = $self;
 983                         $this->mInputLoaded = true;
 984                 }
 985                 if ( $opts ) {
 986                         $this->mOptions = $opts;
 987                         $this->mInputLoaded = true;
 988                 }
 989                 if ( $args ) {
 990                         $this->mArgs = $args;
 991                         $this->mInputLoaded = true;
 992                 }
 993
 994                 # If we've already loaded input (either by user values or from $argv)
 995                 # skip on loading it again. The array_shift() will corrupt values if
 996                 # it's run again and again
 997                 if ( $this->mInputLoaded ) {
 998                         $this->loadSpecialVars();
 999
1000                         return;
1001                 }
1002
1003                 global $argv;
1004                 $this->mSelf = $argv[0];
1005                 $this->loadWithArgv( array_slice( $argv, 1 ) );
1006         }
1007
1008         /**
1009          * Run some validation checks on the params, etc
1010          */
1011         public function validateParamsAndArgs() {
1012                 $die = false;
1013                 # Check to make sure we've got all the required options
1014                 foreach ( $this->mParams as $opt => $info ) {
1015                         if ( $info['require'] && !$this->hasOption( $opt ) ) {
1016                                 $this->error( "Param $opt required!" );
1017                                 $die = true;
1018                         }
1019                 }
1020                 # Check arg list too
1021                 foreach ( $this->mArgList as $k => $info ) {
1022                         if ( $info['require'] && !$this->hasArg( $k ) ) {
1023                                 $this->error( 'Argument <' . $info['name'] . '> required!' );
1024                                 $die = true;
1025                         }
1026                 }
1027                 if ( !$this->mAllowUnregisteredOptions ) {
1028                         # Check for unexpected options
1029                         foreach ( $this->mOptions as $opt => $val ) {
1030                                 if ( !$this->supportsOption( $opt ) ) {
1031                                         $this->error( "Unexpected option $opt!" );
1032                                         $die = true;
1033                                 }
1034                         }
1035                 }
1036
1037                 $this->maybeHelp( $die );
1038         }
1039
1040         /**
1041          * Handle the special variables that are global to all scripts
1042          */
1043         protected function loadSpecialVars() {
1044                 if ( $this->hasOption( 'dbuser' ) ) {
1045                         $this->mDbUser = $this->getOption( 'dbuser' );
1046                 }
1047                 if ( $this->hasOption( 'dbpass' ) ) {
1048                         $this->mDbPass = $this->getOption( 'dbpass' );
1049                 }
1050                 if ( $this->hasOption( 'quiet' ) ) {
1051                         $this->mQuiet = true;
1052                 }
1053                 if ( $this->hasOption( 'batch-size' ) ) {
1054                         $this->mBatchSize = intval( $this->getOption( 'batch-size' ) );
1055                 }
1056         }
1057
1058         /**
1059          * Maybe show the help.
1060          * @param bool $force Whether to force the help to show, default false
1061          */
1062         protected function maybeHelp( $force = false ) {
1063                 if ( !$force && !$this->hasOption( 'help' ) ) {
1064                         return;
1065                 }
1066
1067                 $screenWidth = 80; // TODO: Calculate this!
1068                 $tab = "    ";
1069                 $descWidth = $screenWidth - ( 2 * strlen( $tab ) );
1070
1071                 ksort( $this->mParams );
1072                 $this->mQuiet = false;
1073
1074                 // Description ...
1075                 if ( $this->mDescription ) {
1076                         $this->output( "\n" . wordwrap( $this->mDescription, $screenWidth ) . "\n" );
1077                 }
1078                 $output = "\nUsage: php " . basename( $this->mSelf );
1079
1080                 // ... append parameters ...
1081                 if ( $this->mParams ) {
1082                         $output .= " [--" . implode( "|--", array_keys( $this->mParams ) ) . "]";
1083                 }
1084
1085                 // ... and append arguments.
1086                 if ( $this->mArgList ) {
1087                         $output .= ' ';
1088                         foreach ( $this->mArgList as $k => $arg ) {
1089                                 if ( $arg['require'] ) {
1090                                         $output .= '<' . $arg['name'] . '>';
1091                                 } else {
1092                                         $output .= '[' . $arg['name'] . ']';
1093                                 }
1094                                 if ( $k < count( $this->mArgList ) - 1 ) {
1095                                         $output .= ' ';
1096                                 }
1097                         }
1098                 }
1099                 $this->output( "$output\n\n" );
1100
1101                 # TODO abstract some repetitive code below
1102
1103                 // Generic parameters
1104                 $this->output( "Generic maintenance parameters:\n" );
1105                 foreach ( $this->mGenericParameters as $par => $info ) {
1106                         if ( $info['shortName'] !== false ) {
1107                                 $par .= " (-{$info['shortName']})";
1108                         }
1109                         $this->output(
1110                                 wordwrap( "$tab--$par: " . $info['desc'], $descWidth,
1111                                         "\n$tab$tab" ) . "\n"
1112                         );
1113                 }
1114                 $this->output( "\n" );
1115
1116                 $scriptDependantParams = $this->mDependantParameters;
1117                 if ( count( $scriptDependantParams ) > 0 ) {
1118                         $this->output( "Script dependant parameters:\n" );
1119                         // Parameters description
1120                         foreach ( $scriptDependantParams as $par => $info ) {
1121                                 if ( $info['shortName'] !== false ) {
1122                                         $par .= " (-{$info['shortName']})";
1123                                 }
1124                                 $this->output(
1125                                         wordwrap( "$tab--$par: " . $info['desc'], $descWidth,
1126                                                 "\n$tab$tab" ) . "\n"
1127                                 );
1128                         }
1129                         $this->output( "\n" );
1130                 }
1131
1132                 // Script specific parameters not defined on construction by
1133                 // Maintenance::addDefaultParams()
1134                 $scriptSpecificParams = array_diff_key(
1135                         # all script parameters:
1136                         $this->mParams,
1137                         # remove the Maintenance default parameters:
1138                         $this->mGenericParameters,
1139                         $this->mDependantParameters
1140                 );
1141                 if ( count( $scriptSpecificParams ) > 0 ) {
1142                         $this->output( "Script specific parameters:\n" );
1143                         // Parameters description
1144                         foreach ( $scriptSpecificParams as $par => $info ) {
1145                                 if ( $info['shortName'] !== false ) {
1146                                         $par .= " (-{$info['shortName']})";
1147                                 }
1148                                 $this->output(
1149                                         wordwrap( "$tab--$par: " . $info['desc'], $descWidth,
1150                                                 "\n$tab$tab" ) . "\n"
1151                                 );
1152                         }
1153                         $this->output( "\n" );
1154                 }
1155
1156                 // Print arguments
1157                 if ( count( $this->mArgList ) > 0 ) {
1158                         $this->output( "Arguments:\n" );
1159                         // Arguments description
1160                         foreach ( $this->mArgList as $info ) {
1161                                 $openChar = $info['require'] ? '<' : '[';
1162                                 $closeChar = $info['require'] ? '>' : ']';
1163                                 $this->output(
1164                                         wordwrap( "$tab$openChar" . $info['name'] . "$closeChar: " .
1165                                                 $info['desc'], $descWidth, "\n$tab$tab" ) . "\n"
1166                                 );
1167                         }
1168                         $this->output( "\n" );
1169                 }
1170
1171                 die( 1 );
1172         }
1173
1174         /**
1175          * Handle some last-minute setup here.
1176          */
1177         public function finalSetup() {
1178                 global $wgCommandLineMode, $wgServer, $wgShowExceptionDetails, $wgShowHostnames;
1179                 global $wgDBadminuser, $wgDBadminpassword, $wgDBDefaultGroup;
1180                 global $wgDBuser, $wgDBpassword, $wgDBservers, $wgLBFactoryConf;
1181
1182                 # Turn off output buffering again, it might have been turned on in the settings files
1183                 if ( ob_get_level() ) {
1184                         ob_end_flush();
1185                 }
1186                 # Same with these
1187                 $wgCommandLineMode = true;
1188
1189                 # Override $wgServer
1190                 if ( $this->hasOption( 'server' ) ) {
1191                         $wgServer = $this->getOption( 'server', $wgServer );
1192                 }
1193
1194                 # If these were passed, use them
1195                 if ( $this->mDbUser ) {
1196                         $wgDBadminuser = $this->mDbUser;
1197                 }
1198                 if ( $this->mDbPass ) {
1199                         $wgDBadminpassword = $this->mDbPass;
1200                 }
1201                 if ( $this->hasOption( 'dbgroupdefault' ) ) {
1202                         $wgDBDefaultGroup = $this->getOption( 'dbgroupdefault', null );
1203
1204                         MediaWikiServices::getInstance()->getDBLoadBalancerFactory()->destroy();
1205                 }
1206
1207                 if ( $this->getDbType() == self::DB_ADMIN && isset( $wgDBadminuser ) ) {
1208                         $wgDBuser = $wgDBadminuser;
1209                         $wgDBpassword = $wgDBadminpassword;
1210
1211                         if ( $wgDBservers ) {
1212                                 /**
1213                                  * @var array $wgDBservers
1214                                  */
1215                                 foreach ( $wgDBservers as $i => $server ) {
1216                                         $wgDBservers[$i]['user'] = $wgDBuser;
1217                                         $wgDBservers[$i]['password'] = $wgDBpassword;
1218                                 }
1219                         }
1220                         if ( isset( $wgLBFactoryConf['serverTemplate'] ) ) {
1221                                 $wgLBFactoryConf['serverTemplate']['user'] = $wgDBuser;
1222                                 $wgLBFactoryConf['serverTemplate']['password'] = $wgDBpassword;
1223                         }
1224                         MediaWikiServices::getInstance()->getDBLoadBalancerFactory()->destroy();
1225                 }
1226
1227                 # Apply debug settings
1228                 if ( $this->hasOption( 'mwdebug' ) ) {
1229                         require __DIR__ . '/../includes/DevelopmentSettings.php';
1230                 }
1231
1232                 // Per-script profiling; useful for debugging
1233                 $this->activateProfiler();
1234
1235                 $this->afterFinalSetup();
1236
1237                 $wgShowExceptionDetails = true;
1238                 $wgShowHostnames = true;
1239
1240                 Wikimedia\suppressWarnings();
1241                 set_time_limit( 0 );
1242                 Wikimedia\restoreWarnings();
1243
1244                 $this->adjustMemoryLimit();
1245         }
1246
1247         /**
1248          * Execute a callback function at the end of initialisation
1249          */
1250         protected function afterFinalSetup() {
1251                 if ( defined( 'MW_CMDLINE_CALLBACK' ) ) {
1252                         // @phan-suppress-next-line PhanUndeclaredConstant
1253                         call_user_func( MW_CMDLINE_CALLBACK );
1254                 }
1255         }
1256
1257         /**
1258          * Potentially debug globals. Originally a feature only
1259          * for refreshLinks
1260          */
1261         public function globals() {
1262                 if ( $this->hasOption( 'globals' ) ) {
1263                         print_r( $GLOBALS );
1264                 }
1265         }
1266
1267         /**
1268          * Generic setup for most installs. Returns the location of LocalSettings
1269          * @return string
1270          */
1271         public function loadSettings() {
1272                 global $wgCommandLineMode, $IP;
1273
1274                 if ( isset( $this->mOptions['conf'] ) ) {
1275                         $settingsFile = $this->mOptions['conf'];
1276                 } elseif ( defined( "MW_CONFIG_FILE" ) ) {
1277                         $settingsFile = MW_CONFIG_FILE;
1278                 } else {
1279                         $settingsFile = "$IP/LocalSettings.php";
1280                 }
1281                 if ( isset( $this->mOptions['wiki'] ) ) {
1282                         $bits = explode( '-', $this->mOptions['wiki'], 2 );
1283                         define( 'MW_DB', $bits[0] );
1284                         define( 'MW_PREFIX', $bits[1] ?? '' );
1285                 } elseif ( isset( $this->mOptions['server'] ) ) {
1286                         // Provide the option for site admins to detect and configure
1287                         // multiple wikis based on server names. This offers --server
1288                         // as alternative to --wiki.
1289                         // See https://www.mediawiki.org/wiki/Manual:Wiki_family
1290                         $_SERVER['SERVER_NAME'] = $this->mOptions['server'];
1291                 }
1292
1293                 if ( !is_readable( $settingsFile ) ) {
1294                         $this->fatalError( "A copy of your installation's LocalSettings.php\n" .
1295                                 "must exist and be readable in the source directory.\n" .
1296                                 "Use --conf to specify it." );
1297                 }
1298                 $wgCommandLineMode = true;
1299
1300                 return $settingsFile;
1301         }
1302
1303         /**
1304          * Support function for cleaning up redundant text records
1305          * @param bool $delete Whether or not to actually delete the records
1306          * @author Rob Church <robchur@gmail.com>
1307          */
1308         public function purgeRedundantText( $delete = true ) {
1309                 global $wgMultiContentRevisionSchemaMigrationStage;
1310
1311                 # Data should come off the master, wrapped in a transaction
1312                 $dbw = $this->getDB( DB_MASTER );
1313                 $this->beginTransaction( $dbw, __METHOD__ );
1314
1315                 if ( $wgMultiContentRevisionSchemaMigrationStage & SCHEMA_COMPAT_READ_OLD ) {
1316                         # Get "active" text records from the revisions table
1317                         $cur = [];
1318                         $this->output( 'Searching for active text records in revisions table...' );
1319                         $res = $dbw->select( 'revision', 'rev_text_id', [], __METHOD__, [ 'DISTINCT' ] );
1320                         foreach ( $res as $row ) {
1321                                 $cur[] = $row->rev_text_id;
1322                         }
1323                         $this->output( "done.\n" );
1324
1325                         # Get "active" text records from the archive table
1326                         $this->output( 'Searching for active text records in archive table...' );
1327                         $res = $dbw->select( 'archive', 'ar_text_id', [], __METHOD__, [ 'DISTINCT' ] );
1328                         foreach ( $res as $row ) {
1329                                 # old pre-MW 1.5 records can have null ar_text_id's.
1330                                 if ( $row->ar_text_id !== null ) {
1331                                         $cur[] = $row->ar_text_id;
1332                                 }
1333                         }
1334                         $this->output( "done.\n" );
1335                 } else {
1336                         # Get "active" text records via the content table
1337                         $cur = [];
1338                         $this->output( 'Searching for active text records via contents table...' );
1339                         $res = $dbw->select( 'content', 'content_address', [], __METHOD__, [ 'DISTINCT' ] );
1340                         $blobStore = MediaWikiServices::getInstance()->getBlobStore();
1341                         foreach ( $res as $row ) {
1342                                 // @phan-suppress-next-line PhanUndeclaredMethod
1343                                 $textId = $blobStore->getTextIdFromAddress( $row->content_address );
1344                                 if ( $textId ) {
1345                                         $cur[] = $textId;
1346                                 }
1347                         }
1348                         $this->output( "done.\n" );
1349                 }
1350                 $this->output( "done.\n" );
1351
1352                 # Get the IDs of all text records not in these sets
1353                 $this->output( 'Searching for inactive text records...' );
1354                 $cond = 'old_id NOT IN ( ' . $dbw->makeList( $cur ) . ' )';
1355                 $res = $dbw->select( 'text', 'old_id', [ $cond ], __METHOD__, [ 'DISTINCT' ] );
1356                 $old = [];
1357                 foreach ( $res as $row ) {
1358                         $old[] = $row->old_id;
1359                 }
1360                 $this->output( "done.\n" );
1361
1362                 # Inform the user of what we're going to do
1363                 $count = count( $old );
1364                 $this->output( "$count inactive items found.\n" );
1365
1366                 # Delete as appropriate
1367                 if ( $delete && $count ) {
1368                         $this->output( 'Deleting...' );
1369                         $dbw->delete( 'text', [ 'old_id' => $old ], __METHOD__ );
1370                         $this->output( "done.\n" );
1371                 }
1372
1373                 $this->commitTransaction( $dbw, __METHOD__ );
1374         }
1375
1376         /**
1377          * Get the maintenance directory.
1378          * @return string
1379          */
1380         protected function getDir() {
1381                 return __DIR__;
1382         }
1383
1384         /**
1385          * Returns a database to be used by current maintenance script.
1386          *
1387          * This uses the main LBFactory instance by default unless overriden via setDB().
1388          *
1389          * This function has the same parameters as LoadBalancer::getConnection().
1390          *
1391          * @param int $db DB index (DB_REPLICA/DB_MASTER)
1392          * @param string|string[] $groups default: empty array
1393          * @param string|bool $dbDomain default: current wiki
1394          * @return IMaintainableDatabase
1395          */
1396         protected function getDB( $db, $groups = [], $dbDomain = false ) {
1397                 if ( $this->mDb === null ) {
1398                         return MediaWikiServices::getInstance()
1399                                 ->getDBLoadBalancerFactory()
1400                                 ->getMainLB( $dbDomain )
1401                                 ->getMaintenanceConnectionRef( $db, $groups, $dbDomain );
1402                 }
1403
1404                 return $this->mDb;
1405         }
1406
1407         /**
1408          * Sets database object to be returned by getDB().
1409          *
1410          * @param IMaintainableDatabase $db
1411          */
1412         public function setDB( IMaintainableDatabase $db ) {
1413                 $this->mDb = $db;
1414         }
1415
1416         /**
1417          * Begin a transcation on a DB
1418          *
1419          * This method makes it clear that begin() is called from a maintenance script,
1420          * which has outermost scope. This is safe, unlike $dbw->begin() called in other places.
1421          *
1422          * @param IDatabase $dbw
1423          * @param string $fname Caller name
1424          * @since 1.27
1425          */
1426         protected function beginTransaction( IDatabase $dbw, $fname ) {
1427                 $dbw->begin( $fname );
1428         }
1429
1430         /**
1431          * Commit the transcation on a DB handle and wait for replica DBs to catch up
1432          *
1433          * This method makes it clear that commit() is called from a maintenance script,
1434          * which has outermost scope. This is safe, unlike $dbw->commit() called in other places.
1435          *
1436          * @param IDatabase $dbw
1437          * @param string $fname Caller name
1438          * @return bool Whether the replica DB wait succeeded
1439          * @since 1.27
1440          */
1441         protected function commitTransaction( IDatabase $dbw, $fname ) {
1442                 $dbw->commit( $fname );
1443                 $lbFactory = MediaWikiServices::getInstance()->getDBLoadBalancerFactory();
1444                 $waitSucceeded = $lbFactory->waitForReplication(
1445                         [ 'timeout' => 30, 'ifWritesSince' => $this->lastReplicationWait ]
1446                 );
1447                 $this->lastReplicationWait = microtime( true );
1448                 return $waitSucceeded;
1449         }
1450
1451         /**
1452          * Rollback the transcation on a DB handle
1453          *
1454          * This method makes it clear that rollback() is called from a maintenance script,
1455          * which has outermost scope. This is safe, unlike $dbw->rollback() called in other places.
1456          *
1457          * @param IDatabase $dbw
1458          * @param string $fname Caller name
1459          * @since 1.27
1460          */
1461         protected function rollbackTransaction( IDatabase $dbw, $fname ) {
1462                 $dbw->rollback( $fname );
1463         }
1464
1465         /**
1466          * Lock the search index
1467          * @param IMaintainableDatabase &$db
1468          */
1469         private function lockSearchindex( $db ) {
1470                 $write = [ 'searchindex' ];
1471                 $read = [
1472                         'page',
1473                         'revision',
1474                         'text',
1475                         'interwiki',
1476                         'l10n_cache',
1477                         'user',
1478                         'page_restrictions'
1479                 ];
1480                 $db->lockTables( $read, $write, __CLASS__ . '-searchIndexLock' );
1481         }
1482
1483         /**
1484          * Unlock the tables
1485          * @param IMaintainableDatabase &$db
1486          */
1487         private function unlockSearchindex( $db ) {
1488                 $db->unlockTables( __CLASS__ . '-searchIndexLock' );
1489         }
1490
1491         /**
1492          * Unlock and lock again
1493          * Since the lock is low-priority, queued reads will be able to complete
1494          * @param IMaintainableDatabase &$db
1495          */
1496         private function relockSearchindex( $db ) {
1497                 $this->unlockSearchindex( $db );
1498                 $this->lockSearchindex( $db );
1499         }
1500
1501         /**
1502          * Perform a search index update with locking
1503          * @param int $maxLockTime The maximum time to keep the search index locked.
1504          * @param callable $callback The function that will update the function.
1505          * @param IMaintainableDatabase $dbw
1506          * @param array|IResultWrapper $results
1507          */
1508         public function updateSearchIndex( $maxLockTime, $callback, $dbw, $results ) {
1509                 $lockTime = time();
1510
1511                 # Lock searchindex
1512                 if ( $maxLockTime ) {
1513                         $this->output( "   --- Waiting for lock ---" );
1514                         $this->lockSearchindex( $dbw );
1515                         $lockTime = time();
1516                         $this->output( "\n" );
1517                 }
1518
1519                 # Loop through the results and do a search update
1520                 foreach ( $results as $row ) {
1521                         # Allow reads to be processed
1522                         if ( $maxLockTime && time() > $lockTime + $maxLockTime ) {
1523                                 $this->output( "    --- Relocking ---" );
1524                                 $this->relockSearchindex( $dbw );
1525                                 $lockTime = time();
1526                                 $this->output( "\n" );
1527                         }
1528                         call_user_func( $callback, $dbw, $row );
1529                 }
1530
1531                 # Unlock searchindex
1532                 if ( $maxLockTime ) {
1533                         $this->output( "    --- Unlocking --" );
1534                         $this->unlockSearchindex( $dbw );
1535                         $this->output( "\n" );
1536                 }
1537         }
1538
1539         /**
1540          * Update the searchindex table for a given pageid
1541          * @param IDatabase $dbw A database write handle
1542          * @param int $pageId The page ID to update.
1543          * @return null|string
1544          */
1545         public function updateSearchIndexForPage( $dbw, $pageId ) {
1546                 // Get current revision
1547                 $rev = Revision::loadFromPageId( $dbw, $pageId );
1548                 $title = null;
1549                 if ( $rev ) {
1550                         $titleObj = $rev->getTitle();
1551                         $title = $titleObj->getPrefixedDBkey();
1552                         $this->output( "$title..." );
1553                         # Update searchindex
1554                         $u = new SearchUpdate( $pageId, $titleObj, $rev->getContent() );
1555                         $u->doUpdate();
1556                         $this->output( "\n" );
1557                 }
1558
1559                 return $title;
1560         }
1561
1562         /**
1563          * Count down from $seconds to zero on the terminal, with a one-second pause
1564          * between showing each number. If the maintenance script is in quiet mode,
1565          * this function does nothing.
1566          *
1567          * @since 1.31
1568          *
1569          * @codeCoverageIgnore
1570          * @param int $seconds
1571          */
1572         protected function countDown( $seconds ) {
1573                 if ( $this->isQuiet() ) {
1574                         return;
1575                 }
1576                 for ( $i = $seconds; $i >= 0; $i-- ) {
1577                         if ( $i != $seconds ) {
1578                                 $this->output( str_repeat( "\x08", strlen( $i + 1 ) ) );
1579                         }
1580                         $this->output( $i );
1581                         if ( $i ) {
1582                                 sleep( 1 );
1583                         }
1584                 }
1585                 $this->output( "\n" );
1586         }
1587
1588         /**
1589          * Wrapper for posix_isatty()
1590          * We default as considering stdin a tty (for nice readline methods)
1591          * but treating stout as not a tty to avoid color codes
1592          *
1593          * @param mixed $fd File descriptor
1594          * @return bool
1595          */
1596         public static function posix_isatty( $fd ) {
1597                 if ( !function_exists( 'posix_isatty' ) ) {
1598                         return !$fd;
1599                 } else {
1600                         return posix_isatty( $fd );
1601                 }
1602         }
1603
1604         /**
1605          * Prompt the console for input
1606          * @param string $prompt What to begin the line with, like '> '
1607          * @return string Response
1608          */
1609         public static function readconsole( $prompt = '> ' ) {
1610                 static $isatty = null;
1611                 if ( is_null( $isatty ) ) {
1612                         $isatty = self::posix_isatty( 0 /*STDIN*/ );
1613                 }
1614
1615                 if ( $isatty && function_exists( 'readline' ) ) {
1616                         return readline( $prompt );
1617                 } else {
1618                         if ( $isatty ) {
1619                                 $st = self::readlineEmulation( $prompt );
1620                         } else {
1621                                 if ( feof( STDIN ) ) {
1622                                         $st = false;
1623                                 } else {
1624                                         $st = fgets( STDIN, 1024 );
1625                                 }
1626                         }
1627                         if ( $st === false ) {
1628                                 return false;
1629                         }
1630                         $resp = trim( $st );
1631
1632                         return $resp;
1633                 }
1634         }
1635
1636         /**
1637          * Emulate readline()
1638          * @param string $prompt What to begin the line with, like '> '
1639          * @return string
1640          */
1641         private static function readlineEmulation( $prompt ) {
1642                 $bash = ExecutableFinder::findInDefaultPaths( 'bash' );
1643                 if ( !wfIsWindows() && $bash ) {
1644                         $retval = false;
1645                         $encPrompt = Shell::escape( $prompt );
1646                         $command = "read -er -p $encPrompt && echo \"\$REPLY\"";
1647                         $encCommand = Shell::escape( $command );
1648                         $line = Shell::escape( "$bash -c $encCommand", $retval, [], [ 'walltime' => 0 ] );
1649
1650                         if ( $retval == 0 ) {
1651                                 return $line;
1652                         } elseif ( $retval == 127 ) {
1653                                 // Couldn't execute bash even though we thought we saw it.
1654                                 // Shell probably spit out an error message, sorry :(
1655                                 // Fall through to fgets()...
1656                         } else {
1657                                 // EOF/ctrl+D
1658                                 return false;
1659                         }
1660                 }
1661
1662                 // Fallback... we'll have no editing controls, EWWW
1663                 if ( feof( STDIN ) ) {
1664                         return false;
1665                 }
1666                 print $prompt;
1667
1668                 return fgets( STDIN, 1024 );
1669         }
1670
1671         /**
1672          * Get the terminal size as a two-element array where the first element
1673          * is the width (number of columns) and the second element is the height
1674          * (number of rows).
1675          *
1676          * @return array
1677          */
1678         public static function getTermSize() {
1679                 $default = [ 80, 50 ];
1680                 if ( wfIsWindows() ) {
1681                         return $default;
1682                 }
1683                 if ( Shell::isDisabled() ) {
1684                         return $default;
1685                 }
1686                 // It's possible to get the screen size with VT-100 terminal escapes,
1687                 // but reading the responses is not possible without setting raw mode
1688                 // (unless you want to require the user to press enter), and that
1689                 // requires an ioctl(), which we can't do. So we have to shell out to
1690                 // something that can do the relevant syscalls. There are a few
1691                 // options. Linux and Mac OS X both have "stty size" which does the
1692                 // job directly.
1693                 $result = Shell::command( 'stty', 'size' )
1694                         ->execute();
1695                 if ( $result->getExitCode() !== 0 ) {
1696                         return $default;
1697                 }
1698                 if ( !preg_match( '/^(\d+) (\d+)$/', $result->getStdout(), $m ) ) {
1699                         return $default;
1700                 }
1701                 return [ intval( $m[2] ), intval( $m[1] ) ];
1702         }
1703
1704         /**
1705          * Call this to set up the autoloader to allow classes to be used from the
1706          * tests directory.
1707          */
1708         public static function requireTestsAutoloader() {
1709                 require_once __DIR__ . '/../tests/common/TestsAutoLoader.php';
1710         }
1711 }
1712
1713 /**
1714  * Fake maintenance wrapper, mostly used for the web installer/updater
1715  */
1716 class FakeMaintenance extends Maintenance {
1717         protected $mSelf = "FakeMaintenanceScript";
1718
1719         public function execute() {
1720         }
1721 }
1722
1723 /**
1724  * Class for scripts that perform database maintenance and want to log the
1725  * update in `updatelog` so we can later skip it
1726  */
1727 abstract class LoggedUpdateMaintenance extends Maintenance {
1728         public function __construct() {
1729                 parent::__construct();
1730                 $this->addOption( 'force', 'Run the update even if it was completed already' );
1731                 $this->setBatchSize( 200 );
1732         }
1733
1734         public function execute() {
1735                 $db = $this->getDB( DB_MASTER );
1736                 $key = $this->getUpdateKey();
1737
1738                 if ( !$this->hasOption( 'force' )
1739                         && $db->selectRow( 'updatelog', '1', [ 'ul_key' => $key ], __METHOD__ )
1740                 ) {
1741                         $this->output( "..." . $this->updateSkippedMessage() . "\n" );
1742
1743                         return true;
1744                 }
1745
1746                 if ( !$this->doDBUpdates() ) {
1747                         return false;
1748                 }
1749
1750                 $db->insert( 'updatelog', [ 'ul_key' => $key ], __METHOD__, [ 'IGNORE' ] );
1751
1752                 return true;
1753         }
1754
1755         /**
1756          * Message to show that the update was done already and was just skipped
1757          * @return string
1758          */
1759         protected function updateSkippedMessage() {
1760                 $key = $this->getUpdateKey();
1761
1762                 return "Update '{$key}' already logged as completed. Use --force to run it again.";
1763         }
1764
1765         /**
1766          * Do the actual work. All child classes will need to implement this.
1767          * Return true to log the update as done or false (usually on failure).
1768          * @return bool
1769          */
1770         abstract protected function doDBUpdates();
1771
1772         /**
1773          * Get the update key name to go in the update log table
1774          * @return string
1775          */
1776         abstract protected function getUpdateKey();
1777 }