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