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 // Make sure we're on PHP5.3.2 or better
  24 if ( !function_exists( 'version_compare' ) || version_compare( PHP_VERSION, '5.3.2' ) < 0 ) {
  25         // We need to use dirname( __FILE__ ) here cause __DIR__ is PHP5.3+
  26         require_once dirname( __FILE__ ) . '/../includes/PHPVersionError.php';
  27         wfPHPVersionError( 'cli' );
  28 }
  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 define( 'DO_MAINTENANCE', RUN_MAINTENANCE_IF_MAIN ); // original name, harmless
  38
  39 $maintClass = false;
  40
  41 /**
  42  * Abstract maintenance class for quickly writing and churning out
  43  * maintenance scripts with minimal effort. All that _must_ be defined
  44  * is the execute() method. See docs/maintenance.txt for more info
  45  * and a quick demo of how to use it.
  46  *
  47  * @author Chad Horohoe <chad@anyonecanedit.org>
  48  * @since 1.16
  49  * @ingroup Maintenance
  50  */
  51 abstract class Maintenance {
  52
  53         /**
  54          * Constants for DB access type
  55          * @see Maintenance::getDbType()
  56          */
  57         const DB_NONE = 0;
  58         const DB_STD = 1;
  59         const DB_ADMIN = 2;
  60
  61         // Const for getStdin()
  62         const STDIN_ALL = 'all';
  63
  64         // This is the desired params
  65         protected $mParams = array();
  66
  67         // Array of mapping short parameters to long ones
  68         protected $mShortParamsMap = array();
  69
  70         // Array of desired args
  71         protected $mArgList = array();
  72
  73         // This is the list of options that were actually passed
  74         protected $mOptions = array();
  75
  76         // This is the list of arguments that were actually passed
  77         protected $mArgs = array();
  78
  79         // Name of the script currently running
  80         protected $mSelf;
  81
  82         // Special vars for params that are always used
  83         protected $mQuiet = false;
  84         protected $mDbUser, $mDbPass;
  85
  86         // A description of the script, children should change this
  87         protected $mDescription = '';
  88
  89         // Have we already loaded our user input?
  90         protected $mInputLoaded = false;
  91
  92         /**
  93          * Batch size. If a script supports this, they should set
  94          * a default with setBatchSize()
  95          *
  96          * @var int
  97          */
  98         protected $mBatchSize = null;
  99
 100         // Generic options added by addDefaultParams()
 101         private $mGenericParameters = array();
 102         // Generic options which might or not be supported by the script
 103         private $mDependantParameters = array();
 104
 105         /**
 106          * Used by getDD() / setDB()
 107          * @var DatabaseBase
 108          */
 109         private $mDb = null;
 110
 111         /**
 112          * Used when creating separate schema files.
 113          * @var resource
 114          */
 115         public $fileHandle;
 116
 117         /**
 118          * List of all the core maintenance scripts. This is added
 119          * to scripts added by extensions in $wgMaintenanceScripts
 120          * and returned by getMaintenanceScripts()
 121          */
 122         protected static $mCoreScripts = null;
 123
 124         /**
 125          * Default constructor. Children should call this *first* if implementing
 126          * their own constructors
 127          */
 128         public function __construct() {
 129                 // Setup $IP, using MW_INSTALL_PATH if it exists
 130                 global $IP;
 131                 $IP = strval( getenv( 'MW_INSTALL_PATH' ) ) !== ''
 132                         ? getenv( 'MW_INSTALL_PATH' )
 133                         : realpath( __DIR__ . '/..' );
 134
 135                 $this->addDefaultParams();
 136                 register_shutdown_function( array( $this, 'outputChanneled' ), false );
 137         }
 138
 139         /**
 140          * Should we execute the maintenance script, or just allow it to be included
 141          * as a standalone class? It checks that the call stack only includes this
 142          * function and "requires" (meaning was called from the file scope)
 143          *
 144          * @return Boolean
 145          */
 146         public static function shouldExecute() {
 147                 $bt = debug_backtrace();
 148                 $count = count( $bt );
 149                 if ( $count < 2 ) {
 150                         return false; // sanity
 151                 }
 152                 if ( $bt[0]['class'] !== 'Maintenance' || $bt[0]['function'] !== 'shouldExecute' ) {
 153                         return false; // last call should be to this function
 154                 }
 155                 $includeFuncs = array( 'require_once', 'require', 'include', 'include_once' );
 156                 for ( $i = 1; $i < $count; $i++ ) {
 157                         if ( !in_array( $bt[$i]['function'], $includeFuncs ) ) {
 158                                 return false; // previous calls should all be "requires"
 159                         }
 160                 }
 161                 return true;
 162         }
 163
 164         /**
 165          * Do the actual work. All child classes will need to implement this
 166          */
 167         abstract public function execute();
 168
 169         /**
 170          * Add a parameter to the script. Will be displayed on --help
 171          * with the associated description
 172          *
 173          * @param $name String: the name of the param (help, version, etc)
 174          * @param $description String: the description of the param to show on --help
 175          * @param $required Boolean: is the param required?
 176          * @param $withArg Boolean: is an argument required with this option?
 177          * @param $shortName String: character to use as short name
 178          */
 179         protected function addOption( $name, $description, $required = false, $withArg = false, $shortName = false ) {
 180                 $this->mParams[$name] = array( 'desc' => $description, 'require' => $required, 'withArg' => $withArg, 'shortName' => $shortName );
 181                 if ( $shortName !== false ) {
 182                         $this->mShortParamsMap[$shortName] = $name;
 183                 }
 184         }
 185
 186         /**
 187          * Checks to see if a particular param exists.
 188          * @param $name String: the name of the param
 189          * @return Boolean
 190          */
 191         protected function hasOption( $name ) {
 192                 return isset( $this->mOptions[$name] );
 193         }
 194
 195         /**
 196          * Get an option, or return the default
 197          * @param $name String: the name of the param
 198          * @param $default Mixed: anything you want, default null
 199          * @return Mixed
 200          */
 201         protected function getOption( $name, $default = null ) {
 202                 if ( $this->hasOption( $name ) ) {
 203                         return $this->mOptions[$name];
 204                 } else {
 205                         // Set it so we don't have to provide the default again
 206                         $this->mOptions[$name] = $default;
 207                         return $this->mOptions[$name];
 208                 }
 209         }
 210
 211         /**
 212          * Add some args that are needed
 213          * @param $arg String: name of the arg, like 'start'
 214          * @param $description String: short description of the arg
 215          * @param $required Boolean: is this required?
 216          */
 217         protected function addArg( $arg, $description, $required = true ) {
 218                 $this->mArgList[] = array(
 219                         'name' => $arg,
 220                         'desc' => $description,
 221                         'require' => $required
 222                 );
 223         }
 224
 225         /**
 226          * Remove an option.  Useful for removing options that won't be used in your script.
 227          * @param $name String: the option to remove.
 228          */
 229         protected function deleteOption( $name ) {
 230                 unset( $this->mParams[$name] );
 231         }
 232
 233         /**
 234          * Set the description text.
 235          * @param $text String: the text of the description
 236          */
 237         protected function addDescription( $text ) {
 238                 $this->mDescription = $text;
 239         }
 240
 241         /**
 242          * Does a given argument exist?
 243          * @param $argId Integer: the integer value (from zero) for the arg
 244          * @return Boolean
 245          */
 246         protected function hasArg( $argId = 0 ) {
 247                 return isset( $this->mArgs[$argId] );
 248         }
 249
 250         /**
 251          * Get an argument.
 252          * @param $argId Integer: the integer value (from zero) for the arg
 253          * @param $default Mixed: the default if it doesn't exist
 254          * @return mixed
 255          */
 256         protected function getArg( $argId = 0, $default = null ) {
 257                 return $this->hasArg( $argId ) ? $this->mArgs[$argId] : $default;
 258         }
 259
 260         /**
 261          * Set the batch size.
 262          * @param $s Integer: the number of operations to do in a batch
 263          */
 264         protected function setBatchSize( $s = 0 ) {
 265                 $this->mBatchSize = $s;
 266
 267                 // If we support $mBatchSize, show the option.
 268                 // Used to be in addDefaultParams, but in order for that to
 269                 // work, subclasses would have to call this function in the constructor
 270                 // before they called parent::__construct which is just weird
 271                 // (and really wasn't done).
 272                 if ( $this->mBatchSize ) {
 273                         $this->addOption( 'batch-size', 'Run this many operations ' .
 274                                 'per batch, default: ' . $this->mBatchSize, false, true );
 275                         if ( isset( $this->mParams['batch-size'] ) ) {
 276                                 // This seems a little ugly...
 277                                 $this->mDependantParameters['batch-size'] = $this->mParams['batch-size'];
 278                         }
 279                 }
 280         }
 281
 282         /**
 283          * Get the script's name
 284          * @return String
 285          */
 286         public function getName() {
 287                 return $this->mSelf;
 288         }
 289
 290         /**
 291          * Return input from stdin.
 292          * @param $len Integer: the number of bytes to read. If null,
 293          *        just return the handle. Maintenance::STDIN_ALL returns
 294          *        the full length
 295          * @return Mixed
 296          */
 297         protected function getStdin( $len = null ) {
 298                 if ( $len == Maintenance::STDIN_ALL ) {
 299                         return file_get_contents( 'php://stdin' );
 300                 }
 301                 $f = fopen( 'php://stdin', 'rt' );
 302                 if ( !$len ) {
 303                         return $f;
 304                 }
 305                 $input = fgets( $f, $len );
 306                 fclose( $f );
 307                 return rtrim( $input );
 308         }
 309
 310         /**
 311          * @return bool
 312          */
 313         public function isQuiet() {
 314                 return $this->mQuiet;
 315         }
 316
 317         /**
 318          * Throw some output to the user. Scripts can call this with no fears,
 319          * as we handle all --quiet stuff here
 320          * @param $out String: the text to show to the user
 321          * @param $channel Mixed: unique identifier for the channel. See
 322          *     function outputChanneled.
 323          */
 324         protected function output( $out, $channel = null ) {
 325                 if ( $this->mQuiet ) {
 326                         return;
 327                 }
 328                 if ( $channel === null ) {
 329                         $this->cleanupChanneled();
 330                         print $out;
 331                 } else {
 332                         $out = preg_replace( '/\n\z/', '', $out );
 333                         $this->outputChanneled( $out, $channel );
 334                 }
 335         }
 336
 337         /**
 338          * Throw an error to the user. Doesn't respect --quiet, so don't use
 339          * this for non-error output
 340          * @param $err String: the error to display
 341          * @param $die Int: if > 0, go ahead and die out using this int as the code
 342          */
 343         protected function error( $err, $die = 0 ) {
 344                 $this->outputChanneled( false );
 345                 if ( PHP_SAPI == 'cli' ) {
 346                         fwrite( STDERR, $err . "\n" );
 347                 } else {
 348                         print $err;
 349                 }
 350                 $die = intval( $die );
 351                 if ( $die > 0 ) {
 352                         die( $die );
 353                 }
 354         }
 355
 356         private $atLineStart = true;
 357         private $lastChannel = null;
 358
 359         /**
 360          * Clean up channeled output.  Output a newline if necessary.
 361          */
 362         public function cleanupChanneled() {
 363                 if ( !$this->atLineStart ) {
 364                         print "\n";
 365                         $this->atLineStart = true;
 366                 }
 367         }
 368
 369         /**
 370          * Message outputter with channeled message support. Messages on the
 371          * same channel are concatenated, but any intervening messages in another
 372          * channel start a new line.
 373          * @param $msg String: the message without trailing newline
 374          * @param $channel string Channel identifier or null for no
 375          *     channel. Channel comparison uses ===.
 376          */
 377         public function outputChanneled( $msg, $channel = null ) {
 378                 if ( $msg === false ) {
 379                         $this->cleanupChanneled();
 380                         return;
 381                 }
 382
 383                 // End the current line if necessary
 384                 if ( !$this->atLineStart && $channel !== $this->lastChannel ) {
 385                         print "\n";
 386                 }
 387
 388                 print $msg;
 389
 390                 $this->atLineStart = false;
 391                 if ( $channel === null ) {
 392                         // For unchanneled messages, output trailing newline immediately
 393                         print "\n";
 394                         $this->atLineStart = true;
 395                 }
 396                 $this->lastChannel = $channel;
 397         }
 398
 399         /**
 400          * Does the script need different DB access? By default, we give Maintenance
 401          * scripts normal rights to the DB. Sometimes, a script needs admin rights
 402          * access for a reason and sometimes they want no access. Subclasses should
 403          * override and return one of the following values, as needed:
 404          *    Maintenance::DB_NONE  -  For no DB access at all
 405          *    Maintenance::DB_STD   -  For normal DB access, default
 406          *    Maintenance::DB_ADMIN -  For admin DB access
 407          * @return Integer
 408          */
 409         public function getDbType() {
 410                 return Maintenance::DB_STD;
 411         }
 412
 413         /**
 414          * Add the default parameters to the scripts
 415          */
 416         protected function addDefaultParams() {
 417
 418                 # Generic (non script dependant) options:
 419
 420                 $this->addOption( 'help', 'Display this help message', false, false, 'h' );
 421                 $this->addOption( 'quiet', 'Whether to supress non-error output', false, false, 'q' );
 422                 $this->addOption( 'conf', 'Location of LocalSettings.php, if not default', false, true );
 423                 $this->addOption( 'wiki', 'For specifying the wiki ID', false, true );
 424                 $this->addOption( 'globals', 'Output globals at the end of processing for debugging' );
 425                 $this->addOption( 'memory-limit', 'Set a specific memory limit for the script, "max" for no limit or "default" to avoid changing it' );
 426                 $this->addOption( 'server', "The protocol and server name to use in URLs, e.g. " .
 427                                 "http://en.wikipedia.org. This is sometimes necessary because " .
 428                                 "server name detection may fail in command line scripts.", false, true );
 429                 $this->addOption( 'profiler', 'Set to "text" or "trace" to show profiling output', false, true );
 430
 431                 # Save generic options to display them separately in help
 432                 $this->mGenericParameters = $this->mParams;
 433
 434                 # Script dependant options:
 435
 436                 // If we support a DB, show the options
 437                 if ( $this->getDbType() > 0 ) {
 438                         $this->addOption( 'dbuser', 'The DB user to use for this script', false, true );
 439                         $this->addOption( 'dbpass', 'The password to use for this script', false, true );
 440                 }
 441
 442                 # Save additional script dependant options to display
 443                 # them separately in help
 444                 $this->mDependantParameters = array_diff_key( $this->mParams, $this->mGenericParameters );
 445         }
 446
 447         /**
 448          * Run a child maintenance script. Pass all of the current arguments
 449          * to it.
 450          * @param $maintClass String: a name of a child maintenance class
 451          * @param $classFile String: full path of where the child is
 452          * @return Maintenance child
 453          */
 454         public function runChild( $maintClass, $classFile = null ) {
 455                 // Make sure the class is loaded first
 456                 if ( !MWInit::classExists( $maintClass ) ) {
 457                         if ( $classFile ) {
 458                                 require_once $classFile;
 459                         }
 460                         if ( !MWInit::classExists( $maintClass ) ) {
 461                                 $this->error( "Cannot spawn child: $maintClass" );
 462                         }
 463                 }
 464
 465                 /**
 466                  * @var $child Maintenance
 467                  */
 468                 $child = new $maintClass();
 469                 $child->loadParamsAndArgs( $this->mSelf, $this->mOptions, $this->mArgs );
 470                 if ( !is_null( $this->mDb ) ) {
 471                         $child->setDB( $this->mDb );
 472                 }
 473                 return $child;
 474         }
 475
 476         /**
 477          * Do some sanity checking and basic setup
 478          */
 479         public function setup() {
 480                 global $wgCommandLineMode, $wgRequestTime;
 481
 482                 # Abort if called from a web server
 483                 if ( isset( $_SERVER ) && isset( $_SERVER['REQUEST_METHOD'] ) ) {
 484                         $this->error( 'This script must be run from the command line', true );
 485                 }
 486
 487                 # Make sure we can handle script parameters
 488                 if ( !function_exists( 'hphp_thread_set_warmup_enabled' ) && !ini_get( 'register_argc_argv' ) ) {
 489                         $this->error( 'Cannot get command line arguments, register_argc_argv is set to false', true );
 490                 }
 491
 492                 // Send PHP warnings and errors to stderr instead of stdout.
 493                 // This aids in diagnosing problems, while keeping messages
 494                 // out of redirected output.
 495                 if ( ini_get( 'display_errors' ) ) {
 496                         ini_set( 'display_errors', 'stderr' );
 497                 }
 498
 499                 $this->loadParamsAndArgs();
 500                 $this->maybeHelp();
 501
 502                 # Set the memory limit
 503                 # Note we need to set it again later in cache LocalSettings changed it
 504                 $this->adjustMemoryLimit();
 505
 506                 # Set max execution time to 0 (no limit). PHP.net says that
 507                 # "When running PHP from the command line the default setting is 0."
 508                 # But sometimes this doesn't seem to be the case.
 509                 ini_set( 'max_execution_time', 0 );
 510
 511                 $wgRequestTime = microtime( true );
 512
 513                 # Define us as being in MediaWiki
 514                 define( 'MEDIAWIKI', true );
 515
 516                 $wgCommandLineMode = true;
 517
 518                 # Turn off output buffering if it's on
 519                 while ( ob_get_level() > 0 ) {
 520                         ob_end_flush();
 521                 }
 522
 523                 $this->validateParamsAndArgs();
 524         }
 525
 526         /**
 527          * Normally we disable the memory_limit when running admin scripts.
 528          * Some scripts may wish to actually set a limit, however, to avoid
 529          * blowing up unexpectedly. We also support a --memory-limit option,
 530          * to allow sysadmins to explicitly set one if they'd prefer to override
 531          * defaults (or for people using Suhosin which yells at you for trying
 532          * to disable the limits)
 533          * @return string
 534          */
 535         public function memoryLimit() {
 536                 $limit = $this->getOption( 'memory-limit', 'max' );
 537                 $limit = trim( $limit, "\" '" ); // trim quotes in case someone misunderstood
 538                 return $limit;
 539         }
 540
 541         /**
 542          * Adjusts PHP's memory limit to better suit our needs, if needed.
 543          */
 544         protected function adjustMemoryLimit() {
 545                 $limit = $this->memoryLimit();
 546                 if ( $limit == 'max' ) {
 547                         $limit = -1; // no memory limit
 548                 }
 549                 if ( $limit != 'default' ) {
 550                         ini_set( 'memory_limit', $limit );
 551                 }
 552         }
 553
 554         /**
 555          * Clear all params and arguments.
 556          */
 557         public function clearParamsAndArgs() {
 558                 $this->mOptions = array();
 559                 $this->mArgs = array();
 560                 $this->mInputLoaded = false;
 561         }
 562
 563         /**
 564          * Process command line arguments
 565          * $mOptions becomes an array with keys set to the option names
 566          * $mArgs becomes a zero-based array containing the non-option arguments
 567          *
 568          * @param $self String The name of the script, if any
 569          * @param $opts Array An array of options, in form of key=>value
 570          * @param $args Array An array of command line arguments
 571          */
 572         public function loadParamsAndArgs( $self = null, $opts = null, $args = null ) {
 573                 # If we were given opts or args, set those and return early
 574                 if ( $self ) {
 575                         $this->mSelf = $self;
 576                         $this->mInputLoaded = true;
 577                 }
 578                 if ( $opts ) {
 579                         $this->mOptions = $opts;
 580                         $this->mInputLoaded = true;
 581                 }
 582                 if ( $args ) {
 583                         $this->mArgs = $args;
 584                         $this->mInputLoaded = true;
 585                 }
 586
 587                 # If we've already loaded input (either by user values or from $argv)
 588                 # skip on loading it again. The array_shift() will corrupt values if
 589                 # it's run again and again
 590                 if ( $this->mInputLoaded ) {
 591                         $this->loadSpecialVars();
 592                         return;
 593                 }
 594
 595                 global $argv;
 596                 $this->mSelf = array_shift( $argv );
 597
 598                 $options = array();
 599                 $args = array();
 600
 601                 # Parse arguments
 602                 for ( $arg = reset( $argv ); $arg !== false; $arg = next( $argv ) ) {
 603                         if ( $arg == '--' ) {
 604                                 # End of options, remainder should be considered arguments
 605                                 $arg = next( $argv );
 606                                 while ( $arg !== false ) {
 607                                         $args[] = $arg;
 608                                         $arg = next( $argv );
 609                                 }
 610                                 break;
 611                         } elseif ( substr( $arg, 0, 2 ) == '--' ) {
 612                                 # Long options
 613                                 $option = substr( $arg, 2 );
 614                                 if ( array_key_exists( $option, $options ) ) {
 615                                         $this->error( "\nERROR: $option parameter given twice\n" );
 616                                         $this->maybeHelp( true );
 617                                 }
 618                                 if ( isset( $this->mParams[$option] ) && $this->mParams[$option]['withArg'] ) {
 619                                         $param = next( $argv );
 620                                         if ( $param === false ) {
 621                                                 $this->error( "\nERROR: $option parameter needs a value after it\n" );
 622                                                 $this->maybeHelp( true );
 623                                         }
 624                                         $options[$option] = $param;
 625                                 } else {
 626                                         $bits = explode( '=', $option, 2 );
 627                                         if ( count( $bits ) > 1 ) {
 628                                                 $option = $bits[0];
 629                                                 $param = $bits[1];
 630                                         } else {
 631                                                 $param = 1;
 632                                         }
 633                                         $options[$option] = $param;
 634                                 }
 635                         } elseif ( substr( $arg, 0, 1 ) == '-' ) {
 636                                 # Short options
 637                                 for ( $p = 1; $p < strlen( $arg ); $p++ ) {
 638                                         $option = $arg { $p };
 639                                         if ( !isset( $this->mParams[$option] ) && isset( $this->mShortParamsMap[$option] ) ) {
 640                                                 $option = $this->mShortParamsMap[$option];
 641                                         }
 642                                         if ( array_key_exists( $option, $options ) ) {
 643                                                 $this->error( "\nERROR: $option parameter given twice\n" );
 644                                                 $this->maybeHelp( true );
 645                                         }
 646                                         if ( isset( $this->mParams[$option]['withArg'] ) && $this->mParams[$option]['withArg'] ) {
 647                                                 $param = next( $argv );
 648                                                 if ( $param === false ) {
 649                                                         $this->error( "\nERROR: $option parameter needs a value after it\n" );
 650                                                         $this->maybeHelp( true );
 651                                                 }
 652                                                 $options[$option] = $param;
 653                                         } else {
 654                                                 $options[$option] = 1;
 655                                         }
 656                                 }
 657                         } else {
 658                                 $args[] = $arg;
 659                         }
 660                 }
 661
 662                 $this->mOptions = $options;
 663                 $this->mArgs = $args;
 664                 $this->loadSpecialVars();
 665                 $this->mInputLoaded = true;
 666         }
 667
 668         /**
 669          * Run some validation checks on the params, etc
 670          */
 671         protected function validateParamsAndArgs() {
 672                 $die = false;
 673                 # Check to make sure we've got all the required options
 674                 foreach ( $this->mParams as $opt => $info ) {
 675                         if ( $info['require'] && !$this->hasOption( $opt ) ) {
 676                                 $this->error( "Param $opt required!" );
 677                                 $die = true;
 678                         }
 679                 }
 680                 # Check arg list too
 681                 foreach ( $this->mArgList as $k => $info ) {
 682                         if ( $info['require'] && !$this->hasArg( $k ) ) {
 683                                 $this->error( 'Argument <' . $info['name'] . '> required!' );
 684                                 $die = true;
 685                         }
 686                 }
 687
 688                 if ( $die ) {
 689                         $this->maybeHelp( true );
 690                 }
 691         }
 692
 693         /**
 694          * Handle the special variables that are global to all scripts
 695          */
 696         protected function loadSpecialVars() {
 697                 if ( $this->hasOption( 'dbuser' ) ) {
 698                         $this->mDbUser = $this->getOption( 'dbuser' );
 699                 }
 700                 if ( $this->hasOption( 'dbpass' ) ) {
 701                         $this->mDbPass = $this->getOption( 'dbpass' );
 702                 }
 703                 if ( $this->hasOption( 'quiet' ) ) {
 704                         $this->mQuiet = true;
 705                 }
 706                 if ( $this->hasOption( 'batch-size' ) ) {
 707                         $this->mBatchSize = intval( $this->getOption( 'batch-size' ) );
 708                 }
 709         }
 710
 711         /**
 712          * Maybe show the help.
 713          * @param $force boolean Whether to force the help to show, default false
 714          */
 715         protected function maybeHelp( $force = false ) {
 716                 if ( !$force && !$this->hasOption( 'help' ) ) {
 717                         return;
 718                 }
 719
 720                 $screenWidth = 80; // TODO: Caculate this!
 721                 $tab = "    ";
 722                 $descWidth = $screenWidth - ( 2 * strlen( $tab ) );
 723
 724                 ksort( $this->mParams );
 725                 $this->mQuiet = false;
 726
 727                 // Description ...
 728                 if ( $this->mDescription ) {
 729                         $this->output( "\n" . $this->mDescription . "\n" );
 730                 }
 731                 $output = "\nUsage: php " . basename( $this->mSelf );
 732
 733                 // ... append parameters ...
 734                 if ( $this->mParams ) {
 735                         $output .= " [--" . implode( array_keys( $this->mParams ), "|--" ) . "]";
 736                 }
 737
 738                 // ... and append arguments.
 739                 if ( $this->mArgList ) {
 740                         $output .= ' ';
 741                         foreach ( $this->mArgList as $k => $arg ) {
 742                                 if ( $arg['require'] ) {
 743                                         $output .= '<' . $arg['name'] . '>';
 744                                 } else {
 745                                         $output .= '[' . $arg['name'] . ']';
 746                                 }
 747                                 if ( $k < count( $this->mArgList ) - 1 ) {
 748                                         $output .= ' ';
 749                                 }
 750                         }
 751                 }
 752                 $this->output( "$output\n\n" );
 753
 754                 # TODO abstract some repetitive code below
 755
 756                 // Generic parameters
 757                 $this->output( "Generic maintenance parameters:\n" );
 758                 foreach ( $this->mGenericParameters as $par => $info ) {
 759                         if ( $info['shortName'] !== false ) {
 760                                 $par .= " (-{$info['shortName']})";
 761                         }
 762                         $this->output(
 763                                 wordwrap( "$tab--$par: " . $info['desc'], $descWidth,
 764                                                 "\n$tab$tab" ) . "\n"
 765                         );
 766                 }
 767                 $this->output( "\n" );
 768
 769                 $scriptDependantParams = $this->mDependantParameters;
 770                 if ( count( $scriptDependantParams ) > 0 ) {
 771                         $this->output( "Script dependant parameters:\n" );
 772                         // Parameters description
 773                         foreach ( $scriptDependantParams as $par => $info ) {
 774                                 if ( $info['shortName'] !== false ) {
 775                                         $par .= " (-{$info['shortName']})";
 776                                 }
 777                                 $this->output(
 778                                         wordwrap( "$tab--$par: " . $info['desc'], $descWidth,
 779                                                         "\n$tab$tab" ) . "\n"
 780                                 );
 781                         }
 782                         $this->output( "\n" );
 783                 }
 784
 785
 786                 // Script specific parameters not defined on construction by
 787                 // Maintenance::addDefaultParams()
 788                 $scriptSpecificParams = array_diff_key(
 789                         # all script parameters:
 790                         $this->mParams,
 791                         # remove the Maintenance default parameters:
 792                         $this->mGenericParameters,
 793                         $this->mDependantParameters
 794                 );
 795                 if ( count( $scriptSpecificParams ) > 0 ) {
 796                         $this->output( "Script specific parameters:\n" );
 797                         // Parameters description
 798                         foreach ( $scriptSpecificParams as $par => $info ) {
 799                                 if ( $info['shortName'] !== false ) {
 800                                         $par .= " (-{$info['shortName']})";
 801                                 }
 802                                 $this->output(
 803                                         wordwrap( "$tab--$par: " . $info['desc'], $descWidth,
 804                                                         "\n$tab$tab" ) . "\n"
 805                                 );
 806                         }
 807                         $this->output( "\n" );
 808                 }
 809
 810                 // Print arguments
 811                 if ( count( $this->mArgList ) > 0 ) {
 812                         $this->output( "Arguments:\n" );
 813                         // Arguments description
 814                         foreach ( $this->mArgList as $info ) {
 815                                 $openChar = $info['require'] ? '<' : '[';
 816                                 $closeChar = $info['require'] ? '>' : ']';
 817                                 $this->output(
 818                                         wordwrap( "$tab$openChar" . $info['name'] . "$closeChar: " .
 819                                                 $info['desc'], $descWidth, "\n$tab$tab" ) . "\n"
 820                                 );
 821                         }
 822                         $this->output( "\n" );
 823                 }
 824
 825                 die( 1 );
 826         }
 827
 828         /**
 829          * Handle some last-minute setup here.
 830          */
 831         public function finalSetup() {
 832                 global $wgCommandLineMode, $wgShowSQLErrors, $wgServer;
 833                 global $wgDBadminuser, $wgDBadminpassword;
 834                 global $wgDBuser, $wgDBpassword, $wgDBservers, $wgLBFactoryConf;
 835
 836                 # Turn off output buffering again, it might have been turned on in the settings files
 837                 if ( ob_get_level() ) {
 838                         ob_end_flush();
 839                 }
 840                 # Same with these
 841                 $wgCommandLineMode = true;
 842
 843                 # Override $wgServer
 844                 if ( $this->hasOption( 'server' ) ) {
 845                         $wgServer = $this->getOption( 'server', $wgServer );
 846                 }
 847
 848                 # If these were passed, use them
 849                 if ( $this->mDbUser ) {
 850                         $wgDBadminuser = $this->mDbUser;
 851                 }
 852                 if ( $this->mDbPass ) {
 853                         $wgDBadminpassword = $this->mDbPass;
 854                 }
 855
 856                 if ( $this->getDbType() == self::DB_ADMIN && isset( $wgDBadminuser ) ) {
 857                         $wgDBuser = $wgDBadminuser;
 858                         $wgDBpassword = $wgDBadminpassword;
 859
 860                         if ( $wgDBservers ) {
 861                                 /**
 862                                  * @var $wgDBservers array
 863                                  */
 864                                 foreach ( $wgDBservers as $i => $server ) {
 865                                         $wgDBservers[$i]['user'] = $wgDBuser;
 866                                         $wgDBservers[$i]['password'] = $wgDBpassword;
 867                                 }
 868                         }
 869                         if ( isset( $wgLBFactoryConf['serverTemplate'] ) ) {
 870                                 $wgLBFactoryConf['serverTemplate']['user'] = $wgDBuser;
 871                                 $wgLBFactoryConf['serverTemplate']['password'] = $wgDBpassword;
 872                         }
 873                         LBFactory::destroyInstance();
 874                 }
 875
 876                 $this->afterFinalSetup();
 877
 878                 $wgShowSQLErrors = true;
 879                 @set_time_limit( 0 );
 880                 $this->adjustMemoryLimit();
 881
 882                 // Per-script profiling; useful for debugging
 883                 $forcedProfiler = $this->getOption( 'profiler' );
 884                 if ( $forcedProfiler === 'text' ) {
 885                         Profiler::setInstance( new ProfilerSimpleText( array() ) );
 886                         Profiler::instance()->setTemplated( true );
 887                 } elseif ( $forcedProfiler === 'trace' ) {
 888                         Profiler::setInstance( new ProfilerSimpleTrace( array() ) );
 889                         Profiler::instance()->setTemplated( true );
 890                 }
 891         }
 892
 893         /**
 894          * Execute a callback function at the end of initialisation
 895          */
 896         protected function afterFinalSetup() {
 897                 if ( defined( 'MW_CMDLINE_CALLBACK' ) ) {
 898                         call_user_func( MW_CMDLINE_CALLBACK );
 899                 }
 900         }
 901
 902         /**
 903          * Potentially debug globals. Originally a feature only
 904          * for refreshLinks
 905          */
 906         public function globals() {
 907                 if ( $this->hasOption( 'globals' ) ) {
 908                         print_r( $GLOBALS );
 909                 }
 910         }
 911
 912         /**
 913          * Generic setup for most installs. Returns the location of LocalSettings
 914          * @return String
 915          */
 916         public function loadSettings() {
 917                 global $wgCommandLineMode, $IP;
 918
 919                 if ( isset( $this->mOptions['conf'] ) ) {
 920                         $settingsFile = $this->mOptions['conf'];
 921                 } elseif ( defined( "MW_CONFIG_FILE" ) ) {
 922                         $settingsFile = MW_CONFIG_FILE;
 923                 } else {
 924                         $settingsFile = "$IP/LocalSettings.php";
 925                 }
 926                 if ( isset( $this->mOptions['wiki'] ) ) {
 927                         $bits = explode( '-', $this->mOptions['wiki'] );
 928                         if ( count( $bits ) == 1 ) {
 929                                 $bits[] = '';
 930                         }
 931                         define( 'MW_DB', $bits[0] );
 932                         define( 'MW_PREFIX', $bits[1] );
 933                 }
 934
 935                 if ( !is_readable( $settingsFile ) ) {
 936                         $this->error( "A copy of your installation's LocalSettings.php\n" .
 937                                                 "must exist and be readable in the source directory.\n" .
 938                                                 "Use --conf to specify it.", true );
 939                 }
 940                 $wgCommandLineMode = true;
 941                 return $settingsFile;
 942         }
 943
 944         /**
 945          * Support function for cleaning up redundant text records
 946          * @param $delete Boolean: whether or not to actually delete the records
 947          * @author Rob Church <robchur@gmail.com>
 948          */
 949         public function purgeRedundantText( $delete = true ) {
 950                 # Data should come off the master, wrapped in a transaction
 951                 $dbw = $this->getDB( DB_MASTER );
 952                 $dbw->begin( __METHOD__ );
 953
 954                 # Get "active" text records from the revisions table
 955                 $this->output( 'Searching for active text records in revisions table...' );
 956                 $res = $dbw->select( 'revision', 'rev_text_id', array(), __METHOD__, array( 'DISTINCT' ) );
 957                 foreach ( $res as $row ) {
 958                         $cur[] = $row->rev_text_id;
 959                 }
 960                 $this->output( "done.\n" );
 961
 962                 # Get "active" text records from the archive table
 963                 $this->output( 'Searching for active text records in archive table...' );
 964                 $res = $dbw->select( 'archive', 'ar_text_id', array(), __METHOD__, array( 'DISTINCT' ) );
 965                 foreach ( $res as $row ) {
 966                         # old pre-MW 1.5 records can have null ar_text_id's.
 967                         if ( $row->ar_text_id !== null ) {
 968                                 $cur[] = $row->ar_text_id;
 969                         }
 970                 }
 971                 $this->output( "done.\n" );
 972
 973                 # Get the IDs of all text records not in these sets
 974                 $this->output( 'Searching for inactive text records...' );
 975                 $cond = 'old_id NOT IN ( ' . $dbw->makeList( $cur ) . ' )';
 976                 $res = $dbw->select( 'text', 'old_id', array( $cond ), __METHOD__, array( 'DISTINCT' ) );
 977                 $old = array();
 978                 foreach ( $res as $row ) {
 979                         $old[] = $row->old_id;
 980                 }
 981                 $this->output( "done.\n" );
 982
 983                 # Inform the user of what we're going to do
 984                 $count = count( $old );
 985                 $this->output( "$count inactive items found.\n" );
 986
 987                 # Delete as appropriate
 988                 if ( $delete && $count ) {
 989                         $this->output( 'Deleting...' );
 990                         $dbw->delete( 'text', array( 'old_id' => $old ), __METHOD__ );
 991                         $this->output( "done.\n" );
 992                 }
 993
 994                 # Done
 995                 $dbw->commit( __METHOD__ );
 996         }
 997
 998         /**
 999          * Get the maintenance directory.
1000          * @return string
1001          */
1002         protected function getDir() {
1003                 return __DIR__;
1004         }
1005
1006         /**
1007          * Get the list of available maintenance scripts. Note
1008          * that if you call this _before_ calling doMaintenance
1009          * you won't have any extensions in it yet
1010          * @return Array
1011          */
1012         public static function getMaintenanceScripts() {
1013                 global $wgMaintenanceScripts;
1014                 return $wgMaintenanceScripts + self::getCoreScripts();
1015         }
1016
1017         /**
1018          * Return all of the core maintenance scripts
1019          * @return array
1020          */
1021         protected static function getCoreScripts() {
1022                 if ( !self::$mCoreScripts ) {
1023                         $paths = array(
1024                                 __DIR__,
1025                                 __DIR__ . '/language',
1026                                 __DIR__ . '/storage',
1027                         );
1028                         self::$mCoreScripts = array();
1029                         foreach ( $paths as $p ) {
1030                                 $handle = opendir( $p );
1031                                 while ( ( $file = readdir( $handle ) ) !== false ) {
1032                                         if ( $file == 'Maintenance.php' ) {
1033                                                 continue;
1034                                         }
1035                                         $file = $p . '/' . $file;
1036                                         if ( is_dir( $file ) || !strpos( $file, '.php' ) ||
1037                                                 ( strpos( file_get_contents( $file ), '$maintClass' ) === false ) ) {
1038                                                 continue;
1039                                         }
1040                                         require $file;
1041                                         $vars = get_defined_vars();
1042                                         if ( array_key_exists( 'maintClass', $vars ) ) {
1043                                                 self::$mCoreScripts[$vars['maintClass']] = $file;
1044                                         }
1045                                 }
1046                                 closedir( $handle );
1047                         }
1048                 }
1049                 return self::$mCoreScripts;
1050         }
1051
1052         /**
1053          * Returns a database to be used by current maintenance script. It can be set by setDB().
1054          * If not set, wfGetDB() will be used.
1055          * This function has the same parameters as wfGetDB()
1056          *
1057          * @return DatabaseBase
1058          */
1059         protected function &getDB( $db, $groups = array(), $wiki = false ) {
1060                 if ( is_null( $this->mDb ) ) {
1061                         return wfGetDB( $db, $groups, $wiki );
1062                 } else {
1063                         return $this->mDb;
1064                 }
1065         }
1066
1067         /**
1068          * Sets database object to be returned by getDB().
1069          *
1070          * @param $db DatabaseBase: Database object to be used
1071          */
1072         public function setDB( &$db ) {
1073                 $this->mDb = $db;
1074         }
1075
1076         /**
1077          * Lock the search index
1078          * @param &$db DatabaseBase object
1079          */
1080         private function lockSearchindex( &$db ) {
1081                 $write = array( 'searchindex' );
1082                 $read = array( 'page', 'revision', 'text', 'interwiki', 'l10n_cache', 'user' );
1083                 $db->lockTables( $read, $write, __CLASS__ . '::' . __METHOD__ );
1084         }
1085
1086         /**
1087          * Unlock the tables
1088          * @param &$db DatabaseBase object
1089          */
1090         private function unlockSearchindex( &$db ) {
1091                 $db->unlockTables( __CLASS__ . '::' . __METHOD__ );
1092         }
1093
1094         /**
1095          * Unlock and lock again
1096          * Since the lock is low-priority, queued reads will be able to complete
1097          * @param &$db DatabaseBase object
1098          */
1099         private function relockSearchindex( &$db ) {
1100                 $this->unlockSearchindex( $db );
1101                 $this->lockSearchindex( $db );
1102         }
1103
1104         /**
1105          * Perform a search index update with locking
1106          * @param $maxLockTime Integer: the maximum time to keep the search index locked.
1107          * @param $callback callback String: the function that will update the function.
1108          * @param $dbw DatabaseBase object
1109          * @param $results
1110          */
1111         public function updateSearchIndex( $maxLockTime, $callback, $dbw, $results ) {
1112                 $lockTime = time();
1113
1114                 # Lock searchindex
1115                 if ( $maxLockTime ) {
1116                         $this->output( "   --- Waiting for lock ---" );
1117                         $this->lockSearchindex( $dbw );
1118                         $lockTime = time();
1119                         $this->output( "\n" );
1120                 }
1121
1122                 # Loop through the results and do a search update
1123                 foreach ( $results as $row ) {
1124                         # Allow reads to be processed
1125                         if ( $maxLockTime && time() > $lockTime + $maxLockTime ) {
1126                                 $this->output( "    --- Relocking ---" );
1127                                 $this->relockSearchindex( $dbw );
1128                                 $lockTime = time();
1129                                 $this->output( "\n" );
1130                         }
1131                         call_user_func( $callback, $dbw, $row );
1132                 }
1133
1134                 # Unlock searchindex
1135                 if ( $maxLockTime ) {
1136                         $this->output( "    --- Unlocking --" );
1137                         $this->unlockSearchindex( $dbw );
1138                         $this->output( "\n" );
1139                 }
1140
1141         }
1142
1143         /**
1144          * Update the searchindex table for a given pageid
1145          * @param $dbw DatabaseBase a database write handle
1146          * @param $pageId Integer: the page ID to update.
1147          * @return null|string
1148          */
1149         public function updateSearchIndexForPage( $dbw, $pageId ) {
1150                 // Get current revision
1151                 $rev = Revision::loadFromPageId( $dbw, $pageId );
1152                 $title = null;
1153                 if ( $rev ) {
1154                         $titleObj = $rev->getTitle();
1155                         $title = $titleObj->getPrefixedDBkey();
1156                         $this->output( "$title..." );
1157                         # Update searchindex
1158                         $u = new SearchUpdate( $pageId, $titleObj->getText(), $rev->getContent() );
1159                         $u->doUpdate();
1160                         $this->output( "\n" );
1161                 }
1162                 return $title;
1163         }
1164
1165         /**
1166          * Wrapper for posix_isatty()
1167          * We default as considering stdin a tty (for nice readline methods)
1168          * but treating stout as not a tty to avoid color codes
1169          *
1170          * @param $fd int File descriptor
1171          * @return bool
1172          */
1173         public static function posix_isatty( $fd ) {
1174                 if ( !MWInit::functionExists( 'posix_isatty' ) ) {
1175                         return !$fd;
1176                 } else {
1177                         return posix_isatty( $fd );
1178                 }
1179         }
1180
1181         /**
1182          * Prompt the console for input
1183          * @param $prompt String what to begin the line with, like '> '
1184          * @return String response
1185          */
1186         public static function readconsole( $prompt = '> ' ) {
1187                 static $isatty = null;
1188                 if ( is_null( $isatty ) ) {
1189                         $isatty = self::posix_isatty( 0 /*STDIN*/ );
1190                 }
1191
1192                 if ( $isatty && function_exists( 'readline' ) ) {
1193                         return readline( $prompt );
1194                 } else {
1195                         if ( $isatty ) {
1196                                 $st = self::readlineEmulation( $prompt );
1197                         } else {
1198                                 if ( feof( STDIN ) ) {
1199                                         $st = false;
1200                                 } else {
1201                                         $st = fgets( STDIN, 1024 );
1202                                 }
1203                         }
1204                         if ( $st === false ) {
1205                                 return false;
1206                         }
1207                         $resp = trim( $st );
1208                         return $resp;
1209                 }
1210         }
1211
1212         /**
1213          * Emulate readline()
1214          * @param $prompt String what to begin the line with, like '> '
1215          * @return String
1216          */
1217         private static function readlineEmulation( $prompt ) {
1218                 $bash = Installer::locateExecutableInDefaultPaths( array( 'bash' ) );
1219                 if ( !wfIsWindows() && $bash ) {
1220                         $retval = false;
1221                         $encPrompt = wfEscapeShellArg( $prompt );
1222                         $command = "read -er -p $encPrompt && echo \"\$REPLY\"";
1223                         $encCommand = wfEscapeShellArg( $command );
1224                         $line = wfShellExec( "$bash -c $encCommand", $retval, array(), array( 'walltime' => 0 ) );
1225
1226                         if ( $retval == 0 ) {
1227                                 return $line;
1228                         } elseif ( $retval == 127 ) {
1229                                 // Couldn't execute bash even though we thought we saw it.
1230                                 // Shell probably spit out an error message, sorry :(
1231                                 // Fall through to fgets()...
1232                         } else {
1233                                 // EOF/ctrl+D
1234                                 return false;
1235                         }
1236                 }
1237
1238                 // Fallback... we'll have no editing controls, EWWW
1239                 if ( feof( STDIN ) ) {
1240                         return false;
1241                 }
1242                 print $prompt;
1243                 return fgets( STDIN, 1024 );
1244         }
1245 }
1246
1247 /**
1248  * Fake maintenance wrapper, mostly used for the web installer/updater
1249  */
1250 class FakeMaintenance extends Maintenance {
1251         protected $mSelf = "FakeMaintenanceScript";
1252         public function execute() {
1253                 return;
1254         }
1255 }
1256
1257 /**
1258  * Class for scripts that perform database maintenance and want to log the
1259  * update in `updatelog` so we can later skip it
1260  */
1261 abstract class LoggedUpdateMaintenance extends Maintenance {
1262         public function __construct() {
1263                 parent::__construct();
1264                 $this->addOption( 'force', 'Run the update even if it was completed already' );
1265                 $this->setBatchSize( 200 );
1266         }
1267
1268         public function execute() {
1269                 $db = $this->getDB( DB_MASTER );
1270                 $key = $this->getUpdateKey();
1271
1272                 if ( !$this->hasOption( 'force' ) &&
1273                         $db->selectRow( 'updatelog', '1', array( 'ul_key' => $key ), __METHOD__ ) )
1274                 {
1275                         $this->output( "..." . $this->updateSkippedMessage() . "\n" );
1276                         return true;
1277                 }
1278
1279                 if ( !$this->doDBUpdates() ) {
1280                         return false;
1281                 }
1282
1283                 if (
1284                         $db->insert( 'updatelog', array( 'ul_key' => $key ), __METHOD__, 'IGNORE' ) )
1285                 {
1286                         return true;
1287                 } else {
1288                         $this->output( $this->updatelogFailedMessage() . "\n" );
1289                         return false;
1290                 }
1291         }
1292
1293         /**
1294          * Message to show that the update was done already and was just skipped
1295          * @return String
1296          */
1297         protected function updateSkippedMessage() {
1298                 $key = $this->getUpdateKey();
1299                 return "Update '{$key}' already logged as completed.";
1300         }
1301
1302         /**
1303          * Message to show the the update log was unable to log the completion of this update
1304          * @return String
1305          */
1306         protected function updatelogFailedMessage() {
1307                 $key = $this->getUpdateKey();
1308                 return "Unable to log update '{$key}' as completed.";
1309         }
1310
1311         /**
1312          * Do the actual work. All child classes will need to implement this.
1313          * Return true to log the update as done or false (usually on failure).
1314          * @return Bool
1315          */
1316         abstract protected function doDBUpdates();
1317
1318         /**
1319          * Get the update key name to go in the update log table
1320          * @return String
1321          */
1322         abstract protected function getUpdateKey();
1323 }