includes/filerepo/backend/FileBackend.php

   1 <?php
   2 /**
   3  * @defgroup FileBackend File backend
   4  * @ingroup  FileRepo
   5  *
   6  * This module regroup classes meant for MediaWiki to interacts with
   7  */
   8
   9 /**
  10  * @file
  11  * @ingroup FileBackend
  12  * @author Aaron Schulz
  13  */
  14
  15 /**
  16  * Base class for all file backend classes (including multi-write backends).
  17  *
  18  * This class defines the methods as abstract that subclasses must implement.
  19  * Outside callers can assume that all backends will have these functions.
  20  *
  21  * All "storage paths" are of the format "mwstore://backend/container/path".
  22  * The paths use UNIX file system (FS) notation, though any particular backend may
  23  * not actually be using a local filesystem. Therefore, the paths are only virtual.
  24  *
  25  * Backend contents are stored under wiki-specific container names by default.
  26  * For legacy reasons, this has no effect for the FS backend class, and per-wiki
  27  * segregation must be done by setting the container paths appropriately.
  28  *
  29  * FS-based backends are somewhat more restrictive due to the existence of real
  30  * directory files; a regular file cannot have the same name as a directory. Other
  31  * backends with virtual directories may not have this limitation. Callers should
  32  * store files in such a way that no files and directories are under the same path.
  33  *
  34  * Methods should avoid throwing exceptions at all costs.
  35  * As a corollary, external dependencies should be kept to a minimum.
  36  *
  37  * @ingroup FileBackend
  38  * @since 1.19
  39  */
  40 abstract class FileBackend {
  41         protected $name; // string; unique backend name
  42         protected $wikiId; // string; unique wiki name
  43         protected $readOnly; // string; read-only explanation message
  44         /** @var LockManager */
  45         protected $lockManager;
  46
  47         /**
  48          * Create a new backend instance from configuration.
  49          * This should only be called from within FileBackendGroup.
  50          *
  51          * $config includes:
  52          *     'name'        : The unique name of this backend.
  53          *                     This should consist of alphanumberic, '-', and '_' characters.
  54          *     'wikiId'      : Prefix to container names that is unique to this wiki.
  55          *                     This should consist of alphanumberic, '-', and '_' characters.
  56          *     'lockManager' : Registered name of a file lock manager to use.
  57          *     'readOnly'    : Write operations are disallowed if this is a non-empty string.
  58          *                     It should be an explanation for the backend being read-only.
  59          *
  60          * @param $config Array
  61          */
  62         public function __construct( array $config ) {
  63                 $this->name = $config['name'];
  64                 $this->wikiId = isset( $config['wikiId'] )
  65                         ? $config['wikiId']
  66                         : wfWikiID(); // e.g. "my_wiki-en_"
  67                 $this->lockManager = ( $config['lockManager'] instanceof LockManager )
  68                         ? $config['lockManager']
  69                         : LockManagerGroup::singleton()->get( $config['lockManager'] );
  70                 $this->readOnly = isset( $config['readOnly'] )
  71                         ? (string)$config['readOnly']
  72                         : '';
  73         }
  74
  75         /**
  76          * Get the unique backend name.
  77          * We may have multiple different backends of the same type.
  78          * For example, we can have two Swift backends using different proxies.
  79          *
  80          * @return string
  81          */
  82         final public function getName() {
  83                 return $this->name;
  84         }
  85
  86         /**
  87          * Check if this backend is read-only
  88          *
  89          * @return bool
  90          */
  91         final public function isReadOnly() {
  92                 return ( $this->readOnly != '' );
  93         }
  94
  95         /**
  96          * Get an explanatory message if this backend is read-only
  97          *
  98          * @return string|false Returns falls if the backend is not read-only
  99          */
 100         final public function getReadOnlyReason() {
 101                 return ( $this->readOnly != '' ) ? $this->readOnly : false;
 102         }
 103
 104         /**
 105          * This is the main entry point into the backend for write operations.
 106          * Callers supply an ordered list of operations to perform as a transaction.
 107          * Files will be locked, the stat cache cleared, and then the operations attempted.
 108          * If any serious errors occur, all attempted operations will be rolled back.
 109          *
 110          * $ops is an array of arrays. The outer array holds a list of operations.
 111          * Each inner array is a set of key value pairs that specify an operation.
 112          *
 113          * Supported operations and their parameters:
 114          * a) Create a new file in storage with the contents of a string
 115          *     array(
 116          *         'op'                  => 'create',
 117          *         'dst'                 => <storage path>,
 118          *         'content'             => <string of new file contents>,
 119          *         'overwrite'           => <boolean>,
 120          *         'overwriteSame'       => <boolean>
 121          *     )
 122          * b) Copy a file system file into storage
 123          *     array(
 124          *         'op'                  => 'store',
 125          *         'src'                 => <file system path>,
 126          *         'dst'                 => <storage path>,
 127          *         'overwrite'           => <boolean>,
 128          *         'overwriteSame'       => <boolean>
 129          *     )
 130          * c) Copy a file within storage
 131          *     array(
 132          *         'op'                  => 'copy',
 133          *         'src'                 => <storage path>,
 134          *         'dst'                 => <storage path>,
 135          *         'overwrite'           => <boolean>,
 136          *         'overwriteSame'       => <boolean>
 137          *     )
 138          * d) Move a file within storage
 139          *     array(
 140          *         'op'                  => 'move',
 141          *         'src'                 => <storage path>,
 142          *         'dst'                 => <storage path>,
 143          *         'overwrite'           => <boolean>,
 144          *         'overwriteSame'       => <boolean>
 145          *     )
 146          * e) Delete a file within storage
 147          *     array(
 148          *         'op'                  => 'delete',
 149          *         'src'                 => <storage path>,
 150          *         'ignoreMissingSource' => <boolean>
 151          *     )
 152          * f) Do nothing (no-op)
 153          *     array(
 154          *         'op'                  => 'null',
 155          *     )
 156          *
 157          * Boolean flags for operations (operation-specific):
 158          * 'ignoreMissingSource' : The operation will simply succeed and do
 159          *                         nothing if the source file does not exist.
 160          * 'overwrite'           : Any destination file will be overwritten.
 161          * 'overwriteSame'       : An error will not be given if a file already
 162          *                         exists at the destination that has the same
 163          *                         contents as the new contents to be written there.
 164          *
 165          * $opts is an associative of boolean flags, including:
 166          * 'force'               : Errors that would normally cause a rollback do not.
 167          *                         The remaining operations are still attempted if any fail.
 168          * 'nonLocking'          : No locks are acquired for the operations.
 169          *                         This can increase performance for non-critical writes.
 170          *                         This has no effect unless the 'force' flag is set.
 171          * 'allowStale'          : Don't require the latest available data.
 172          *                         This can increase performance for non-critical writes.
 173          *                         This has no effect unless the 'force' flag is set.
 174          *
 175          * Remarks on locking:
 176          * File system paths given to operations should refer to files that are
 177          * already locked or otherwise safe from modification from other processes.
 178          * Normally these files will be new temp files, which should be adequate.
 179          *
 180          * Return value:
 181          * This returns a Status, which contains all warnings and fatals that occured
 182          * during the operation. The 'failCount', 'successCount', and 'success' members
 183          * will reflect each operation attempted. The status will be "OK" unless any
 184          * of the operations failed and the 'force' parameter was not set.
 185          *
 186          * @param $ops Array List of operations to execute in order
 187          * @param $opts Array Batch operation options
 188          * @return Status
 189          */
 190         final public function doOperations( array $ops, array $opts = array() ) {
 191                 if ( $this->isReadOnly() ) {
 192                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 193                 }
 194                 if ( empty( $opts['force'] ) ) { // sanity
 195                         unset( $opts['nonLocking'] );
 196                         unset( $opts['allowStale'] );
 197                 }
 198                 return $this->doOperationsInternal( $ops, $opts );
 199         }
 200
 201         /**
 202          * @see FileBackend::doOperations()
 203          */
 204         abstract protected function doOperationsInternal( array $ops, array $opts );
 205
 206         /**
 207          * Same as doOperations() except it takes a single operation.
 208          * If you are doing a batch of operations that should either
 209          * all succeed or all fail, then use that function instead.
 210          *
 211          * @see FileBackend::doOperations()
 212          *
 213          * @param $op Array Operation
 214          * @param $opts Array Operation options
 215          * @return Status
 216          */
 217         final public function doOperation( array $op, array $opts = array() ) {
 218                 return $this->doOperations( array( $op ), $opts );
 219         }
 220
 221         /**
 222          * Performs a single create operation.
 223          * This sets $params['op'] to 'create' and passes it to doOperation().
 224          *
 225          * @see FileBackend::doOperation()
 226          *
 227          * @param $params Array Operation parameters
 228          * @param $opts Array Operation options
 229          * @return Status
 230          */
 231         final public function create( array $params, array $opts = array() ) {
 232                 $params['op'] = 'create';
 233                 return $this->doOperation( $params, $opts );
 234         }
 235
 236         /**
 237          * Performs a single store operation.
 238          * This sets $params['op'] to 'store' and passes it to doOperation().
 239          *
 240          * @see FileBackend::doOperation()
 241          *
 242          * @param $params Array Operation parameters
 243          * @param $opts Array Operation options
 244          * @return Status
 245          */
 246         final public function store( array $params, array $opts = array() ) {
 247                 $params['op'] = 'store';
 248                 return $this->doOperation( $params, $opts );
 249         }
 250
 251         /**
 252          * Performs a single copy operation.
 253          * This sets $params['op'] to 'copy' and passes it to doOperation().
 254          *
 255          * @see FileBackend::doOperation()
 256          *
 257          * @param $params Array Operation parameters
 258          * @param $opts Array Operation options
 259          * @return Status
 260          */
 261         final public function copy( array $params, array $opts = array() ) {
 262                 $params['op'] = 'copy';
 263                 return $this->doOperation( $params, $opts );
 264         }
 265
 266         /**
 267          * Performs a single move operation.
 268          * This sets $params['op'] to 'move' and passes it to doOperation().
 269          *
 270          * @see FileBackend::doOperation()
 271          *
 272          * @param $params Array Operation parameters
 273          * @param $opts Array Operation options
 274          * @return Status
 275          */
 276         final public function move( array $params, array $opts = array() ) {
 277                 $params['op'] = 'move';
 278                 return $this->doOperation( $params, $opts );
 279         }
 280
 281         /**
 282          * Performs a single delete operation.
 283          * This sets $params['op'] to 'delete' and passes it to doOperation().
 284          *
 285          * @see FileBackend::doOperation()
 286          *
 287          * @param $params Array Operation parameters
 288          * @param $opts Array Operation options
 289          * @return Status
 290          */
 291         final public function delete( array $params, array $opts = array() ) {
 292                 $params['op'] = 'delete';
 293                 return $this->doOperation( $params, $opts );
 294         }
 295
 296         /**
 297          * Concatenate a list of storage files into a single file system file.
 298          * The target path should refer to a file that is already locked or
 299          * otherwise safe from modification from other processes. Normally,
 300          * the file will be a new temp file, which should be adequate.
 301          * $params include:
 302          *     srcs          : ordered source storage paths (e.g. chunk1, chunk2, ...)
 303          *     dst           : file system path to 0-byte temp file
 304          *
 305          * @param $params Array Operation parameters
 306          * @return Status
 307          */
 308         abstract public function concatenate( array $params );
 309
 310         /**
 311          * Prepare a storage directory for usage.
 312          * This will create any required containers and parent directories.
 313          * Backends using key/value stores only need to create the container.
 314          *
 315          * $params include:
 316          *     dir : storage directory
 317          *
 318          * @param $params Array
 319          * @return Status
 320          */
 321         final public function prepare( array $params ) {
 322                 if ( $this->isReadOnly() ) {
 323                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 324                 }
 325                 return $this->doPrepare( $params );
 326         }
 327
 328         /**
 329          * @see FileBackend::prepare()
 330          */
 331         abstract protected function doPrepare( array $params );
 332
 333         /**
 334          * Take measures to block web access to a storage directory and
 335          * the container it belongs to. FS backends might add .htaccess
 336          * files whereas key/value store backends might restrict container
 337          * access to the auth user that represents end-users in web request.
 338          * This is not guaranteed to actually do anything.
 339          *
 340          * $params include:
 341          *     dir       : storage directory
 342          *     noAccess  : try to deny file access
 343          *     noListing : try to deny file listing
 344          *
 345          * @param $params Array
 346          * @return Status
 347          */
 348         final public function secure( array $params ) {
 349                 if ( $this->isReadOnly() ) {
 350                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 351                 }
 352                 $status = $this->doPrepare( $params ); // dir must exist to restrict it
 353                 if ( $status->isOK() ) {
 354                         $status->merge( $this->doSecure( $params ) );
 355                 }
 356                 return $status;
 357         }
 358
 359         /**
 360          * @see FileBackend::secure()
 361          */
 362         abstract protected function doSecure( array $params );
 363
 364         /**
 365          * Delete a storage directory if it is empty.
 366          * Backends using key/value stores may do nothing unless the directory
 367          * is that of an empty container, in which case it should be deleted.
 368          *
 369          * $params include:
 370          *     dir : storage directory
 371          *
 372          * @param $params Array
 373          * @return Status
 374          */
 375         final public function clean( array $params ) {
 376                 if ( $this->isReadOnly() ) {
 377                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 378                 }
 379                 return $this->doClean( $params );
 380         }
 381
 382         /**
 383          * @see FileBackend::clean()
 384          */
 385         abstract protected function doClean( array $params );
 386
 387         /**
 388          * Check if a file exists at a storage path in the backend.
 389          * This returns false if only a directory exists at the path.
 390          *
 391          * $params include:
 392          *     src    : source storage path
 393          *     latest : use the latest available data
 394          *
 395          * @param $params Array
 396          * @return bool|null Returns null on failure
 397          */
 398         abstract public function fileExists( array $params );
 399
 400         /**
 401          * Get the last-modified timestamp of the file at a storage path.
 402          *
 403          * $params include:
 404          *     src    : source storage path
 405          *     latest : use the latest available data
 406          *
 407          * @param $params Array
 408          * @return string|false TS_MW timestamp or false on failure
 409          */
 410         abstract public function getFileTimestamp( array $params );
 411
 412         /**
 413          * Get the contents of a file at a storage path in the backend.
 414          * This should be avoided for potentially large files.
 415          *
 416          * $params include:
 417          *     src    : source storage path
 418          *     latest : use the latest available data
 419          *
 420          * @param $params Array
 421          * @return string|false Returns false on failure
 422          */
 423         abstract public function getFileContents( array $params );
 424
 425         /**
 426          * Get the size (bytes) of a file at a storage path in the backend.
 427          *
 428          * $params include:
 429          *     src    : source storage path
 430          *     latest : use the latest available data
 431          *
 432          * @param $params Array
 433          * @return integer|false Returns false on failure
 434          */
 435         abstract public function getFileSize( array $params );
 436
 437         /**
 438          * Get quick information about a file at a storage path in the backend.
 439          * If the file does not exist, then this returns false.
 440          * Otherwise, the result is an associative array that includes:
 441          *     mtime  : the last-modified timestamp (TS_MW)
 442          *     size   : the file size (bytes)
 443          * Additional values may be included for internal use only.
 444          *
 445          * $params include:
 446          *     src    : source storage path
 447          *     latest : use the latest available data
 448          *
 449          * @param $params Array
 450          * @return Array|false|null Returns null on failure
 451          */
 452         abstract public function getFileStat( array $params );
 453
 454         /**
 455          * Get a SHA-1 hash of the file at a storage path in the backend.
 456          *
 457          * $params include:
 458          *     src    : source storage path
 459          *     latest : use the latest available data
 460          *
 461          * @param $params Array
 462          * @return string|false Hash string or false on failure
 463          */
 464         abstract public function getFileSha1Base36( array $params );
 465
 466         /**
 467          * Get the properties of the file at a storage path in the backend.
 468          * Returns FSFile::placeholderProps() on failure.
 469          *
 470          * $params include:
 471          *     src    : source storage path
 472          *     latest : use the latest available data
 473          *
 474          * @param $params Array
 475          * @return Array
 476          */
 477         abstract public function getFileProps( array $params );
 478
 479         /**
 480          * Stream the file at a storage path in the backend.
 481          * If the file does not exists, a 404 error will be given.
 482          * Appropriate HTTP headers (Status, Content-Type, Content-Length)
 483          * must be sent if streaming began, while none should be sent otherwise.
 484          * Implementations should flush the output buffer before sending data.
 485          *
 486          * $params include:
 487          *     src     : source storage path
 488          *     headers : additional HTTP headers to send on success
 489          *     latest  : use the latest available data
 490          *
 491          * @param $params Array
 492          * @return Status
 493          */
 494         abstract public function streamFile( array $params );
 495
 496         /**
 497          * Returns a file system file, identical to the file at a storage path.
 498          * The file returned is either:
 499          * a) A local copy of the file at a storage path in the backend.
 500          *    The temporary copy will have the same extension as the source.
 501          * b) An original of the file at a storage path in the backend.
 502          * Temporary files may be purged when the file object falls out of scope.
 503          *
 504          * Write operations should *never* be done on this file as some backends
 505          * may do internal tracking or may be instances of FileBackendMultiWrite.
 506          * In that later case, there are copies of the file that must stay in sync.
 507          * Additionally, further calls to this function may return the same file.
 508          *
 509          * $params include:
 510          *     src    : source storage path
 511          *     latest : use the latest available data
 512          *
 513          * @param $params Array
 514          * @return FSFile|null Returns null on failure
 515          */
 516         abstract public function getLocalReference( array $params );
 517
 518         /**
 519          * Get a local copy on disk of the file at a storage path in the backend.
 520          * The temporary copy will have the same file extension as the source.
 521          * Temporary files may be purged when the file object falls out of scope.
 522          *
 523          * $params include:
 524          *     src    : source storage path
 525          *     latest : use the latest available data
 526          *
 527          * @param $params Array
 528          * @return TempFSFile|null Returns null on failure
 529          */
 530         abstract public function getLocalCopy( array $params );
 531
 532         /**
 533          * Get an iterator to list out all stored files under a storage directory.
 534          * If the directory is of the form "mwstore://backend/container",
 535          * then all files in the container should be listed.
 536          * If the directory is of form "mwstore://backend/container/dir",
 537          * then all files under that container directory should be listed.
 538          * Results should be storage paths relative to the given directory.
 539          *
 540          * Storage backends with eventual consistency might return stale data.
 541          *
 542          * $params include:
 543          *     dir : storage path directory
 544          *
 545          * @return Traversable|Array|null Returns null on failure
 546          */
 547         abstract public function getFileList( array $params );
 548
 549         /**
 550          * Invalidate any in-process file existence and property cache.
 551          * If $paths is given, then only the cache for those files will be cleared.
 552          *
 553          * @param $paths Array Storage paths (optional)
 554          * @return void
 555          */
 556         public function clearCache( array $paths = null ) {}
 557
 558         /**
 559          * Lock the files at the given storage paths in the backend.
 560          * This will either lock all the files or none (on failure).
 561          *
 562          * Callers should consider using getScopedFileLocks() instead.
 563          *
 564          * @param $paths Array Storage paths
 565          * @param $type integer LockManager::LOCK_* constant
 566          * @return Status
 567          */
 568         final public function lockFiles( array $paths, $type ) {
 569                 return $this->lockManager->lock( $paths, $type );
 570         }
 571
 572         /**
 573          * Unlock the files at the given storage paths in the backend.
 574          *
 575          * @param $paths Array Storage paths
 576          * @param $type integer LockManager::LOCK_* constant
 577          * @return Status
 578          */
 579         final public function unlockFiles( array $paths, $type ) {
 580                 return $this->lockManager->unlock( $paths, $type );
 581         }
 582
 583         /**
 584          * Lock the files at the given storage paths in the backend.
 585          * This will either lock all the files or none (on failure).
 586          * On failure, the status object will be updated with errors.
 587          *
 588          * Once the return value goes out scope, the locks will be released and
 589          * the status updated. Unlock fatals will not change the status "OK" value.
 590          *
 591          * @param $paths Array Storage paths
 592          * @param $type integer LockManager::LOCK_* constant
 593          * @param $status Status Status to update on lock/unlock
 594          * @return ScopedLock|null Returns null on failure
 595          */
 596         final public function getScopedFileLocks( array $paths, $type, Status $status ) {
 597                 return ScopedLock::factory( $this->lockManager, $paths, $type, $status );
 598         }
 599
 600         /**
 601          * Check if a given path is a "mwstore://" path.
 602          * This does not do any further validation or any existence checks.
 603          *
 604          * @param $path string
 605          * @return bool
 606          */
 607         final public static function isStoragePath( $path ) {
 608                 return ( strpos( $path, 'mwstore://' ) === 0 );
 609         }
 610
 611         /**
 612          * Split a storage path into a backend name, a container name,
 613          * and a relative file path. The relative path may be the empty string.
 614          * This does not do any path normalization or traversal checks.
 615          *
 616          * @param $storagePath string
 617          * @return Array (backend, container, rel object) or (null, null, null)
 618          */
 619         final public static function splitStoragePath( $storagePath ) {
 620                 if ( self::isStoragePath( $storagePath ) ) {
 621                         // Remove the "mwstore://" prefix and split the path
 622                         $parts = explode( '/', substr( $storagePath, 10 ), 3 );
 623                         if ( count( $parts ) >= 2 && $parts[0] != '' && $parts[1] != '' ) {
 624                                 if ( count( $parts ) == 3 ) {
 625                                         return $parts; // e.g. "backend/container/path"
 626                                 } else {
 627                                         return array( $parts[0], $parts[1], '' ); // e.g. "backend/container"
 628                                 }
 629                         }
 630                 }
 631                 return array( null, null, null );
 632         }
 633
 634         /**
 635          * Normalize a storage path by cleaning up directory separators.
 636          * Returns null if the path is not of the format of a valid storage path.
 637          *
 638          * @param $storagePath string
 639          * @return string|null
 640          */
 641         final public static function normalizeStoragePath( $storagePath ) {
 642                 list( $backend, $container, $relPath ) = self::splitStoragePath( $storagePath );
 643                 if ( $relPath !== null ) { // must be for this backend
 644                         $relPath = self::normalizeContainerPath( $relPath );
 645                         if ( $relPath !== null ) {
 646                                 return ( $relPath != '' )
 647                                         ? "mwstore://{$backend}/{$container}/{$relPath}"
 648                                         : "mwstore://{$backend}/{$container}";
 649                         }
 650                 }
 651                 return null;
 652         }
 653
 654         /**
 655          * Validate and normalize a relative storage path.
 656          * Null is returned if the path involves directory traversal.
 657          * Traversal is insecure for FS backends and broken for others.
 658          *
 659          * @param $path string Storage path relative to a container
 660          * @return string|null
 661          */
 662         final protected static function normalizeContainerPath( $path ) {
 663                 // Normalize directory separators
 664                 $path = strtr( $path, '\\', '/' );
 665                 // Collapse any consecutive directory separators
 666                 $path = preg_replace( '![/]{2,}!', '/', $path );
 667                 // Remove any leading directory separator
 668                 $path = ltrim( $path, '/' );
 669                 // Use the same traversal protection as Title::secureAndSplit()
 670                 if ( strpos( $path, '.' ) !== false ) {
 671                         if (
 672                                 $path === '.' ||
 673                                 $path === '..' ||
 674                                 strpos( $path, './' ) === 0 ||
 675                                 strpos( $path, '../' ) === 0 ||
 676                                 strpos( $path, '/./' ) !== false ||
 677                                 strpos( $path, '/../' ) !== false
 678                         ) {
 679                                 return null;
 680                         }
 681                 }
 682                 return $path;
 683         }
 684
 685         /**
 686          * Get the parent storage directory of a storage path.
 687          * This returns a path like "mwstore://backend/container",
 688          * "mwstore://backend/container/...", or null if there is no parent.
 689          *
 690          * @param $storagePath string
 691          * @return string|null
 692          */
 693         final public static function parentStoragePath( $storagePath ) {
 694                 $storagePath = dirname( $storagePath );
 695                 list( $b, $cont, $rel ) = self::splitStoragePath( $storagePath );
 696                 return ( $rel === null ) ? null : $storagePath;
 697         }
 698
 699         /**
 700          * Get the final extension from a storage or FS path
 701          *
 702          * @param $path string
 703          * @return string
 704          */
 705         final public static function extensionFromPath( $path ) {
 706                 $i = strrpos( $path, '.' );
 707                 return strtolower( $i ? substr( $path, $i + 1 ) : '' );
 708         }
 709 }
 710
 711 /**
 712  * @brief Base class for all backends associated with a particular storage medium.
 713  *
 714  * This class defines the methods as abstract that subclasses must implement.
 715  * Outside callers should *not* use functions with "Internal" in the name.
 716  *
 717  * The FileBackend operations are implemented using basic functions
 718  * such as storeInternal(), copyInternal(), deleteInternal() and the like.
 719  * This class is also responsible for path resolution and sanitization.
 720  *
 721  * @ingroup FileBackend
 722  * @since 1.19
 723  */
 724 abstract class FileBackendStore extends FileBackend {
 725         /** @var Array Map of paths to small (RAM/disk) cache items */
 726         protected $cache = array(); // (storage path => key => value)
 727         protected $maxCacheSize = 100; // integer; max paths with entries
 728         /** @var Array Map of paths to large (RAM/disk) cache items */
 729         protected $expensiveCache = array(); // (storage path => key => value)
 730         protected $maxExpensiveCacheSize = 10; // integer; max paths with entries
 731
 732         /** @var Array Map of container names to sharding settings */
 733         protected $shardViaHashLevels = array(); // (container name => config array)
 734
 735         protected $maxFileSize = 1000000000; // integer bytes (1GB)
 736
 737         /**
 738          * Get the maximum allowable file size given backend
 739          * medium restrictions and basic performance constraints.
 740          * Do not call this function from places outside FileBackend and FileOp.
 741          *
 742          * @return integer Bytes
 743          */
 744         final public function maxFileSizeInternal() {
 745                 return $this->maxFileSize;
 746         }
 747
 748         /**
 749          * Check if a file can be created at a given storage path.
 750          * FS backends should check if the parent directory exists and the file is writable.
 751          * Backends using key/value stores should check if the container exists.
 752          *
 753          * @param $storagePath string
 754          * @return bool
 755          */
 756         abstract public function isPathUsableInternal( $storagePath );
 757
 758         /**
 759          * Create a file in the backend with the given contents.
 760          * Do not call this function from places outside FileBackend and FileOp.
 761          *
 762          * $params include:
 763          *     content       : the raw file contents
 764          *     dst           : destination storage path
 765          *     overwrite     : overwrite any file that exists at the destination
 766          *
 767          * @param $params Array
 768          * @return Status
 769          */
 770         final public function createInternal( array $params ) {
 771                 wfProfileIn( __METHOD__ );
 772                 if ( strlen( $params['content'] ) > $this->maxFileSizeInternal() ) {
 773                         $status = Status::newFatal( 'backend-fail-create', $params['dst'] );
 774                 } else {
 775                         $status = $this->doCreateInternal( $params );
 776                         $this->clearCache( array( $params['dst'] ) );
 777                 }
 778                 wfProfileOut( __METHOD__ );
 779                 return $status;
 780         }
 781
 782         /**
 783          * @see FileBackendStore::createInternal()
 784          */
 785         abstract protected function doCreateInternal( array $params );
 786
 787         /**
 788          * Store a file into the backend from a file on disk.
 789          * Do not call this function from places outside FileBackend and FileOp.
 790          *
 791          * $params include:
 792          *     src           : source path on disk
 793          *     dst           : destination storage path
 794          *     overwrite     : overwrite any file that exists at the destination
 795          *
 796          * @param $params Array
 797          * @return Status
 798          */
 799         final public function storeInternal( array $params ) {
 800                 wfProfileIn( __METHOD__ );
 801                 if ( filesize( $params['src'] ) > $this->maxFileSizeInternal() ) {
 802                         $status = Status::newFatal( 'backend-fail-store', $params['dst'] );
 803                 } else {
 804                         $status = $this->doStoreInternal( $params );
 805                         $this->clearCache( array( $params['dst'] ) );
 806                 }
 807                 wfProfileOut( __METHOD__ );
 808                 return $status;
 809         }
 810
 811         /**
 812          * @see FileBackendStore::storeInternal()
 813          */
 814         abstract protected function doStoreInternal( array $params );
 815
 816         /**
 817          * Copy a file from one storage path to another in the backend.
 818          * Do not call this function from places outside FileBackend and FileOp.
 819          *
 820          * $params include:
 821          *     src           : source storage path
 822          *     dst           : destination storage path
 823          *     overwrite     : overwrite any file that exists at the destination
 824          *
 825          * @param $params Array
 826          * @return Status
 827          */
 828         final public function copyInternal( array $params ) {
 829                 wfProfileIn( __METHOD__ );
 830                 $status = $this->doCopyInternal( $params );
 831                 $this->clearCache( array( $params['dst'] ) );
 832                 wfProfileOut( __METHOD__ );
 833                 return $status;
 834         }
 835
 836         /**
 837          * @see FileBackendStore::copyInternal()
 838          */
 839         abstract protected function doCopyInternal( array $params );
 840
 841         /**
 842          * Delete a file at the storage path.
 843          * Do not call this function from places outside FileBackend and FileOp.
 844          *
 845          * $params include:
 846          *     src                 : source storage path
 847          *     ignoreMissingSource : do nothing if the source file does not exist
 848          *
 849          * @param $params Array
 850          * @return Status
 851          */
 852         final public function deleteInternal( array $params ) {
 853                 wfProfileIn( __METHOD__ );
 854                 $status = $this->doDeleteInternal( $params );
 855                 $this->clearCache( array( $params['src'] ) );
 856                 wfProfileOut( __METHOD__ );
 857                 return $status;
 858         }
 859
 860         /**
 861          * @see FileBackendStore::deleteInternal()
 862          */
 863         abstract protected function doDeleteInternal( array $params );
 864
 865         /**
 866          * Move a file from one storage path to another in the backend.
 867          * Do not call this function from places outside FileBackend and FileOp.
 868          *
 869          * $params include:
 870          *     src           : source storage path
 871          *     dst           : destination storage path
 872          *     overwrite     : overwrite any file that exists at the destination
 873          *
 874          * @param $params Array
 875          * @return Status
 876          */
 877         final public function moveInternal( array $params ) {
 878                 wfProfileIn( __METHOD__ );
 879                 $status = $this->doMoveInternal( $params );
 880                 $this->clearCache( array( $params['src'], $params['dst'] ) );
 881                 wfProfileOut( __METHOD__ );
 882                 return $status;
 883         }
 884
 885         /**
 886          * @see FileBackendStore::moveInternal()
 887          */
 888         protected function doMoveInternal( array $params ) {
 889                 // Copy source to dest
 890                 $status = $this->copyInternal( $params );
 891                 if ( $status->isOK() ) {
 892                         // Delete source (only fails due to races or medium going down)
 893                         $status->merge( $this->deleteInternal( array( 'src' => $params['src'] ) ) );
 894                         $status->setResult( true, $status->value ); // ignore delete() errors
 895                 }
 896                 return $status;
 897         }
 898
 899         /**
 900          * @see FileBackend::concatenate()
 901          */
 902         final public function concatenate( array $params ) {
 903                 wfProfileIn( __METHOD__ );
 904                 $status = Status::newGood();
 905
 906                 // Try to lock the source files for the scope of this function
 907                 $scopeLockS = $this->getScopedFileLocks( $params['srcs'], LockManager::LOCK_UW, $status );
 908                 if ( $status->isOK() ) {
 909                         // Actually do the concatenation
 910                         $status->merge( $this->doConcatenate( $params ) );
 911                 }
 912
 913                 wfProfileOut( __METHOD__ );
 914                 return $status;
 915         }
 916
 917         /**
 918          * @see FileBackendStore::concatenate()
 919          */
 920         protected function doConcatenate( array $params ) {
 921                 $status = Status::newGood();
 922                 $tmpPath = $params['dst']; // convenience
 923
 924                 // Check that the specified temp file is valid...
 925                 wfSuppressWarnings();
 926                 $ok = ( is_file( $tmpPath ) && !filesize( $tmpPath ) );
 927                 wfRestoreWarnings();
 928                 if ( !$ok ) { // not present or not empty
 929                         $status->fatal( 'backend-fail-opentemp', $tmpPath );
 930                         return $status;
 931                 }
 932
 933                 // Build up the temp file using the source chunks (in order)...
 934                 $tmpHandle = fopen( $tmpPath, 'ab' );
 935                 if ( $tmpHandle === false ) {
 936                         $status->fatal( 'backend-fail-opentemp', $tmpPath );
 937                         return $status;
 938                 }
 939                 foreach ( $params['srcs'] as $virtualSource ) {
 940                         // Get a local FS version of the chunk
 941                         $tmpFile = $this->getLocalReference( array( 'src' => $virtualSource ) );
 942                         if ( !$tmpFile ) {
 943                                 $status->fatal( 'backend-fail-read', $virtualSource );
 944                                 return $status;
 945                         }
 946                         // Get a handle to the local FS version
 947                         $sourceHandle = fopen( $tmpFile->getPath(), 'r' );
 948                         if ( $sourceHandle === false ) {
 949                                 fclose( $tmpHandle );
 950                                 $status->fatal( 'backend-fail-read', $virtualSource );
 951                                 return $status;
 952                         }
 953                         // Append chunk to file (pass chunk size to avoid magic quotes)
 954                         if ( !stream_copy_to_stream( $sourceHandle, $tmpHandle ) ) {
 955                                 fclose( $sourceHandle );
 956                                 fclose( $tmpHandle );
 957                                 $status->fatal( 'backend-fail-writetemp', $tmpPath );
 958                                 return $status;
 959                         }
 960                         fclose( $sourceHandle );
 961                 }
 962                 if ( !fclose( $tmpHandle ) ) {
 963                         $status->fatal( 'backend-fail-closetemp', $tmpPath );
 964                         return $status;
 965                 }
 966
 967                 clearstatcache(); // temp file changed
 968
 969                 return $status;
 970         }
 971
 972         /**
 973          * @see FileBackend::doPrepare()
 974          */
 975         final protected function doPrepare( array $params ) {
 976                 wfProfileIn( __METHOD__ );
 977
 978                 $status = Status::newGood();
 979                 list( $fullCont, $dir, $shard ) = $this->resolveStoragePath( $params['dir'] );
 980                 if ( $dir === null ) {
 981                         $status->fatal( 'backend-fail-invalidpath', $params['dir'] );
 982                         wfProfileOut( __METHOD__ );
 983                         return $status; // invalid storage path
 984                 }
 985
 986                 if ( $shard !== null ) { // confined to a single container/shard
 987                         $status->merge( $this->doPrepareInternal( $fullCont, $dir, $params ) );
 988                 } else { // directory is on several shards
 989                         wfDebug( __METHOD__ . ": iterating over all container shards.\n" );
 990                         list( $b, $shortCont, $r ) = self::splitStoragePath( $params['dir'] );
 991                         foreach ( $this->getContainerSuffixes( $shortCont ) as $suffix ) {
 992                                 $status->merge( $this->doPrepareInternal( "{$fullCont}{$suffix}", $dir, $params ) );
 993                         }
 994                 }
 995
 996                 wfProfileOut( __METHOD__ );
 997                 return $status;
 998         }
 999
1000         /**
1001          * @see FileBackendStore::doPrepare()
1002          */
1003         protected function doPrepareInternal( $container, $dir, array $params ) {
1004                 return Status::newGood();
1005         }
1006
1007         /**
1008          * @see FileBackend::doSecure()
1009          */
1010         final protected function doSecure( array $params ) {
1011                 wfProfileIn( __METHOD__ );
1012                 $status = Status::newGood();
1013
1014                 list( $fullCont, $dir, $shard ) = $this->resolveStoragePath( $params['dir'] );
1015                 if ( $dir === null ) {
1016                         $status->fatal( 'backend-fail-invalidpath', $params['dir'] );
1017                         wfProfileOut( __METHOD__ );
1018                         return $status; // invalid storage path
1019                 }
1020
1021                 if ( $shard !== null ) { // confined to a single container/shard
1022                         $status->merge( $this->doSecureInternal( $fullCont, $dir, $params ) );
1023                 } else { // directory is on several shards
1024                         wfDebug( __METHOD__ . ": iterating over all container shards.\n" );
1025                         list( $b, $shortCont, $r ) = self::splitStoragePath( $params['dir'] );
1026                         foreach ( $this->getContainerSuffixes( $shortCont ) as $suffix ) {
1027                                 $status->merge( $this->doSecureInternal( "{$fullCont}{$suffix}", $dir, $params ) );
1028                         }
1029                 }
1030
1031                 wfProfileOut( __METHOD__ );
1032                 return $status;
1033         }
1034
1035         /**
1036          * @see FileBackendStore::doSecure()
1037          */
1038         protected function doSecureInternal( $container, $dir, array $params ) {
1039                 return Status::newGood();
1040         }
1041
1042         /**
1043          * @see FileBackend::doClean()
1044          */
1045         final protected function doClean( array $params ) {
1046                 wfProfileIn( __METHOD__ );
1047                 $status = Status::newGood();
1048
1049                 list( $fullCont, $dir, $shard ) = $this->resolveStoragePath( $params['dir'] );
1050                 if ( $dir === null ) {
1051                         $status->fatal( 'backend-fail-invalidpath', $params['dir'] );
1052                         wfProfileOut( __METHOD__ );
1053                         return $status; // invalid storage path
1054                 }
1055
1056                 // Attempt to lock this directory...
1057                 $filesLockEx = array( $params['dir'] );
1058                 $scopedLockE = $this->getScopedFileLocks( $filesLockEx, LockManager::LOCK_EX, $status );
1059                 if ( !$status->isOK() ) {
1060                         wfProfileOut( __METHOD__ );
1061                         return $status; // abort
1062                 }
1063
1064                 if ( $shard !== null ) { // confined to a single container/shard
1065                         $status->merge( $this->doCleanInternal( $fullCont, $dir, $params ) );
1066                 } else { // directory is on several shards
1067                         wfDebug( __METHOD__ . ": iterating over all container shards.\n" );
1068                         list( $b, $shortCont, $r ) = self::splitStoragePath( $params['dir'] );
1069                         foreach ( $this->getContainerSuffixes( $shortCont ) as $suffix ) {
1070                                 $status->merge( $this->doCleanInternal( "{$fullCont}{$suffix}", $dir, $params ) );
1071                         }
1072                 }
1073
1074                 wfProfileOut( __METHOD__ );
1075                 return $status;
1076         }
1077
1078         /**
1079          * @see FileBackendStore::doClean()
1080          */
1081         protected function doCleanInternal( $container, $dir, array $params ) {
1082                 return Status::newGood();
1083         }
1084
1085         /**
1086          * @see FileBackend::fileExists()
1087          */
1088         final public function fileExists( array $params ) {
1089                 wfProfileIn( __METHOD__ );
1090                 $stat = $this->getFileStat( $params );
1091                 wfProfileOut( __METHOD__ );
1092                 return ( $stat === null ) ? null : (bool)$stat; // null => failure
1093         }
1094
1095         /**
1096          * @see FileBackend::getFileTimestamp()
1097          */
1098         final public function getFileTimestamp( array $params ) {
1099                 wfProfileIn( __METHOD__ );
1100                 $stat = $this->getFileStat( $params );
1101                 wfProfileOut( __METHOD__ );
1102                 return $stat ? $stat['mtime'] : false;
1103         }
1104
1105         /**
1106          * @see FileBackend::getFileSize()
1107          */
1108         final public function getFileSize( array $params ) {
1109                 wfProfileIn( __METHOD__ );
1110                 $stat = $this->getFileStat( $params );
1111                 wfProfileOut( __METHOD__ );
1112                 return $stat ? $stat['size'] : false;
1113         }
1114
1115         /**
1116          * @see FileBackend::getFileStat()
1117          */
1118         final public function getFileStat( array $params ) {
1119                 wfProfileIn( __METHOD__ );
1120                 $path = self::normalizeStoragePath( $params['src'] );
1121                 if ( $path === null ) {
1122                         return false; // invalid storage path
1123                 }
1124                 $latest = !empty( $params['latest'] );
1125                 if ( isset( $this->cache[$path]['stat'] ) ) {
1126                         // If we want the latest data, check that this cached
1127                         // value was in fact fetched with the latest available data.
1128                         if ( !$latest || $this->cache[$path]['stat']['latest'] ) {
1129                                 wfProfileOut( __METHOD__ );
1130                                 return $this->cache[$path]['stat'];
1131                         }
1132                 }
1133                 $stat = $this->doGetFileStat( $params );
1134                 if ( is_array( $stat ) ) { // don't cache negatives
1135                         $this->trimCache(); // limit memory
1136                         $this->cache[$path]['stat'] = $stat;
1137                         $this->cache[$path]['stat']['latest'] = $latest;
1138                 }
1139                 wfProfileOut( __METHOD__ );
1140                 return $stat;
1141         }
1142
1143         /**
1144          * @see FileBackendStore::getFileStat()
1145          */
1146         abstract protected function doGetFileStat( array $params );
1147
1148         /**
1149          * @see FileBackend::getFileContents()
1150          */
1151         public function getFileContents( array $params ) {
1152                 wfProfileIn( __METHOD__ );
1153                 $tmpFile = $this->getLocalReference( $params );
1154                 if ( !$tmpFile ) {
1155                         wfProfileOut( __METHOD__ );
1156                         return false;
1157                 }
1158                 wfSuppressWarnings();
1159                 $data = file_get_contents( $tmpFile->getPath() );
1160                 wfRestoreWarnings();
1161                 wfProfileOut( __METHOD__ );
1162                 return $data;
1163         }
1164
1165         /**
1166          * @see FileBackend::getFileSha1Base36()
1167          */
1168         final public function getFileSha1Base36( array $params ) {
1169                 wfProfileIn( __METHOD__ );
1170                 $path = $params['src'];
1171                 if ( isset( $this->cache[$path]['sha1'] ) ) {
1172                         wfProfileOut( __METHOD__ );
1173                         return $this->cache[$path]['sha1'];
1174                 }
1175                 $hash = $this->doGetFileSha1Base36( $params );
1176                 if ( $hash ) { // don't cache negatives
1177                         $this->trimCache(); // limit memory
1178                         $this->cache[$path]['sha1'] = $hash;
1179                 }
1180                 wfProfileOut( __METHOD__ );
1181                 return $hash;
1182         }
1183
1184         /**
1185          * @see FileBackendStore::getFileSha1Base36()
1186          */
1187         protected function doGetFileSha1Base36( array $params ) {
1188                 $fsFile = $this->getLocalReference( $params );
1189                 if ( !$fsFile ) {
1190                         return false;
1191                 } else {
1192                         return $fsFile->getSha1Base36();
1193                 }
1194         }
1195
1196         /**
1197          * @see FileBackend::getFileProps()
1198          */
1199         final public function getFileProps( array $params ) {
1200                 wfProfileIn( __METHOD__ );
1201                 $fsFile = $this->getLocalReference( $params );
1202                 $props = $fsFile ? $fsFile->getProps() : FSFile::placeholderProps();
1203                 wfProfileOut( __METHOD__ );
1204                 return $props;
1205         }
1206
1207         /**
1208          * @see FileBackend::getLocalReference()
1209          */
1210         public function getLocalReference( array $params ) {
1211                 wfProfileIn( __METHOD__ );
1212                 $path = $params['src'];
1213                 if ( isset( $this->expensiveCache[$path]['localRef'] ) ) {
1214                         wfProfileOut( __METHOD__ );
1215                         return $this->expensiveCache[$path]['localRef'];
1216                 }
1217                 $tmpFile = $this->getLocalCopy( $params );
1218                 if ( $tmpFile ) { // don't cache negatives
1219                         $this->trimExpensiveCache(); // limit memory
1220                         $this->expensiveCache[$path]['localRef'] = $tmpFile;
1221                 }
1222                 wfProfileOut( __METHOD__ );
1223                 return $tmpFile;
1224         }
1225
1226         /**
1227          * @see FileBackend::streamFile()
1228          */
1229         final public function streamFile( array $params ) {
1230                 wfProfileIn( __METHOD__ );
1231                 $status = Status::newGood();
1232
1233                 $info = $this->getFileStat( $params );
1234                 if ( !$info ) { // let StreamFile handle the 404
1235                         $status->fatal( 'backend-fail-notexists', $params['src'] );
1236                 }
1237
1238                 // Set output buffer and HTTP headers for stream
1239                 $extraHeaders = $params['headers'] ? $params['headers'] : array();
1240                 $res = StreamFile::prepareForStream( $params['src'], $info, $extraHeaders );
1241                 if ( $res == StreamFile::NOT_MODIFIED ) {
1242                         // do nothing; client cache is up to date
1243                 } elseif ( $res == StreamFile::READY_STREAM ) {
1244                         $status = $this->doStreamFile( $params );
1245                 } else {
1246                         $status->fatal( 'backend-fail-stream', $params['src'] );
1247                 }
1248
1249                 wfProfileOut( __METHOD__ );
1250                 return $status;
1251         }
1252
1253         /**
1254          * @see FileBackendStore::streamFile()
1255          */
1256         protected function doStreamFile( array $params ) {
1257                 $status = Status::newGood();
1258
1259                 $fsFile = $this->getLocalReference( $params );
1260                 if ( !$fsFile ) {
1261                         $status->fatal( 'backend-fail-stream', $params['src'] );
1262                 } elseif ( !readfile( $fsFile->getPath() ) ) {
1263                         $status->fatal( 'backend-fail-stream', $params['src'] );
1264                 }
1265
1266                 return $status;
1267         }
1268
1269         /**
1270          * @copydoc FileBackend::getFileList()
1271          */
1272         final public function getFileList( array $params ) {
1273                 list( $fullCont, $dir, $shard ) = $this->resolveStoragePath( $params['dir'] );
1274                 if ( $dir === null ) { // invalid storage path
1275                         return null;
1276                 }
1277                 if ( $shard !== null ) {
1278                         // File listing is confined to a single container/shard
1279                         return $this->getFileListInternal( $fullCont, $dir, $params );
1280                 } else {
1281                         wfDebug( __METHOD__ . ": iterating over all container shards.\n" );
1282                         // File listing spans multiple containers/shards
1283                         list( $b, $shortCont, $r ) = self::splitStoragePath( $params['dir'] );
1284                         return new FileBackendStoreShardListIterator( $this,
1285                                 $fullCont, $dir, $this->getContainerSuffixes( $shortCont ), $params );
1286                 }
1287         }
1288
1289         /**
1290          * Do not call this function from places outside FileBackend
1291          *
1292          * @see FileBackendStore::getFileList()
1293          *
1294          * @param $container string Resolved container name
1295          * @param $dir string Resolved path relative to container
1296          * @param $params Array
1297          * @return Traversable|Array|null
1298          */
1299         abstract public function getFileListInternal( $container, $dir, array $params );
1300
1301         /**
1302          * Get the list of supported operations and their corresponding FileOp classes.
1303          *
1304          * @return Array
1305          */
1306         protected function supportedOperations() {
1307                 return array(
1308                         'store'       => 'StoreFileOp',
1309                         'copy'        => 'CopyFileOp',
1310                         'move'        => 'MoveFileOp',
1311                         'delete'      => 'DeleteFileOp',
1312                         'create'      => 'CreateFileOp',
1313                         'null'        => 'NullFileOp'
1314                 );
1315         }
1316
1317         /**
1318          * Return a list of FileOp objects from a list of operations.
1319          * Do not call this function from places outside FileBackend.
1320          *
1321          * The result must have the same number of items as the input.
1322          * An exception is thrown if an unsupported operation is requested.
1323          *
1324          * @param $ops Array Same format as doOperations()
1325          * @return Array List of FileOp objects
1326          * @throws MWException
1327          */
1328         final public function getOperations( array $ops ) {
1329                 $supportedOps = $this->supportedOperations();
1330
1331                 $performOps = array(); // array of FileOp objects
1332                 // Build up ordered array of FileOps...
1333                 foreach ( $ops as $operation ) {
1334                         $opName = $operation['op'];
1335                         if ( isset( $supportedOps[$opName] ) ) {
1336                                 $class = $supportedOps[$opName];
1337                                 // Get params for this operation
1338                                 $params = $operation;
1339                                 // Append the FileOp class
1340                                 $performOps[] = new $class( $this, $params );
1341                         } else {
1342                                 throw new MWException( "Operation `$opName` is not supported." );
1343                         }
1344                 }
1345
1346                 return $performOps;
1347         }
1348
1349         /**
1350          * @see FileBackend::doOperationsInternal()
1351          */
1352         protected function doOperationsInternal( array $ops, array $opts ) {
1353                 wfProfileIn( __METHOD__ );
1354                 $status = Status::newGood();
1355
1356                 // Build up a list of FileOps...
1357                 $performOps = $this->getOperations( $ops );
1358
1359                 // Acquire any locks as needed...
1360                 if ( empty( $opts['nonLocking'] ) ) {
1361                         // Build up a list of files to lock...
1362                         $filesLockEx = $filesLockSh = array();
1363                         foreach ( $performOps as $fileOp ) {
1364                                 $filesLockSh = array_merge( $filesLockSh, $fileOp->storagePathsRead() );
1365                                 $filesLockEx = array_merge( $filesLockEx, $fileOp->storagePathsChanged() );
1366                         }
1367                         // Optimization: if doing an EX lock anyway, don't also set an SH one
1368                         $filesLockSh = array_diff( $filesLockSh, $filesLockEx );
1369                         // Get a shared lock on the parent directory of each path changed
1370                         $filesLockSh = array_merge( $filesLockSh, array_map( 'dirname', $filesLockEx ) );
1371                         // Try to lock those files for the scope of this function...
1372                         $scopeLockS = $this->getScopedFileLocks( $filesLockSh, LockManager::LOCK_UW, $status );
1373                         $scopeLockE = $this->getScopedFileLocks( $filesLockEx, LockManager::LOCK_EX, $status );
1374                         if ( !$status->isOK() ) {
1375                                 wfProfileOut( __METHOD__ );
1376                                 return $status; // abort
1377                         }
1378                 }
1379
1380                 // Clear any cache entries (after locks acquired)
1381                 $this->clearCache();
1382
1383                 // Actually attempt the operation batch...
1384                 $subStatus = FileOp::attemptBatch( $performOps, $opts );
1385
1386                 // Merge errors into status fields
1387                 $status->merge( $subStatus );
1388                 $status->success = $subStatus->success; // not done in merge()
1389
1390                 wfProfileOut( __METHOD__ );
1391                 return $status;
1392         }
1393
1394         /**
1395          * @see FileBackend::clearCache()
1396          */
1397         final public function clearCache( array $paths = null ) {
1398                 if ( is_array( $paths ) ) {
1399                         $paths = array_map( 'FileBackend::normalizeStoragePath', $paths );
1400                         $paths = array_filter( $paths, 'strlen' ); // remove nulls
1401                 }
1402                 if ( $paths === null ) {
1403                         $this->cache = array();
1404                         $this->expensiveCache = array();
1405                 } else {
1406                         foreach ( $paths as $path ) {
1407                                 unset( $this->cache[$path] );
1408                                 unset( $this->expensiveCache[$path] );
1409                         }
1410                 }
1411                 $this->doClearCache( $paths );
1412         }
1413
1414         /**
1415          * Clears any additional stat caches for storage paths
1416          *
1417          * @see FileBackend::clearCache()
1418          *
1419          * @param $paths Array Storage paths (optional)
1420          * @return void
1421          */
1422         protected function doClearCache( array $paths = null ) {}
1423
1424         /**
1425          * Prune the inexpensive cache if it is too big to add an item
1426          *
1427          * @return void
1428          */
1429         protected function trimCache() {
1430                 if ( count( $this->cache ) >= $this->maxCacheSize ) {
1431                         reset( $this->cache );
1432                         unset( $this->cache[key( $this->cache )] );
1433                 }
1434         }
1435
1436         /**
1437          * Prune the expensive cache if it is too big to add an item
1438          *
1439          * @return void
1440          */
1441         protected function trimExpensiveCache() {
1442                 if ( count( $this->expensiveCache ) >= $this->maxExpensiveCacheSize ) {
1443                         reset( $this->expensiveCache );
1444                         unset( $this->expensiveCache[key( $this->expensiveCache )] );
1445                 }
1446         }
1447
1448         /**
1449          * Check if a container name is valid.
1450          * This checks for for length and illegal characters.
1451          *
1452          * @param $container string
1453          * @return bool
1454          */
1455         final protected static function isValidContainerName( $container ) {
1456                 // This accounts for Swift and S3 restrictions while leaving room
1457                 // for things like '.xxx' (hex shard chars) or '.seg' (segments).
1458                 // This disallows directory separators or traversal characters.
1459                 // Note that matching strings URL encode to the same string;
1460                 // in Swift, the length restriction is *after* URL encoding.
1461                 return preg_match( '/^[a-z0-9][a-z0-9-_]{0,199}$/i', $container );
1462         }
1463
1464         /**
1465          * Splits a storage path into an internal container name,
1466          * an internal relative file name, and a container shard suffix.
1467          * Any shard suffix is already appended to the internal container name.
1468          * This also checks that the storage path is valid and within this backend.
1469          *
1470          * If the container is sharded but a suffix could not be determined,
1471          * this means that the path can only refer to a directory and can only
1472          * be scanned by looking in all the container shards.
1473          *
1474          * @param $storagePath string
1475          * @return Array (container, path, container suffix) or (null, null, null) if invalid
1476          */
1477         final protected function resolveStoragePath( $storagePath ) {
1478                 list( $backend, $container, $relPath ) = self::splitStoragePath( $storagePath );
1479                 if ( $backend === $this->name ) { // must be for this backend
1480                         $relPath = self::normalizeContainerPath( $relPath );
1481                         if ( $relPath !== null ) {
1482                                 // Get shard for the normalized path if this container is sharded
1483                                 $cShard = $this->getContainerShard( $container, $relPath );
1484                                 // Validate and sanitize the relative path (backend-specific)
1485                                 $relPath = $this->resolveContainerPath( $container, $relPath );
1486                                 if ( $relPath !== null ) {
1487                                         // Prepend any wiki ID prefix to the container name
1488                                         $container = $this->fullContainerName( $container );
1489                                         if ( self::isValidContainerName( $container ) ) {
1490                                                 // Validate and sanitize the container name (backend-specific)
1491                                                 $container = $this->resolveContainerName( "{$container}{$cShard}" );
1492                                                 if ( $container !== null ) {
1493                                                         return array( $container, $relPath, $cShard );
1494                                                 }
1495                                         }
1496                                 }
1497                         }
1498                 }
1499                 return array( null, null, null );
1500         }
1501
1502         /**
1503          * Like resolveStoragePath() except null values are returned if
1504          * the container is sharded and the shard could not be determined.
1505          *
1506          * @see FileBackendStore::resolveStoragePath()
1507          *
1508          * @param $storagePath string
1509          * @return Array (container, path) or (null, null) if invalid
1510          */
1511         final protected function resolveStoragePathReal( $storagePath ) {
1512                 list( $container, $relPath, $cShard ) = $this->resolveStoragePath( $storagePath );
1513                 if ( $cShard !== null ) {
1514                         return array( $container, $relPath );
1515                 }
1516                 return array( null, null );
1517         }
1518
1519         /**
1520          * Get the container name shard suffix for a given path.
1521          * Any empty suffix means the container is not sharded.
1522          *
1523          * @param $container string Container name
1524          * @param $relStoragePath string Storage path relative to the container
1525          * @return string|null Returns null if shard could not be determined
1526          */
1527         final protected function getContainerShard( $container, $relPath ) {
1528                 list( $levels, $base, $repeat ) = $this->getContainerHashLevels( $container );
1529                 if ( $levels == 1 || $levels == 2 ) {
1530                         // Hash characters are either base 16 or 36
1531                         $char = ( $base == 36 ) ? '[0-9a-z]' : '[0-9a-f]';
1532                         // Get a regex that represents the shard portion of paths.
1533                         // The concatenation of the captures gives us the shard.
1534                         if ( $levels === 1 ) { // 16 or 36 shards per container
1535                                 $hashDirRegex = '(' . $char . ')';
1536                         } else { // 256 or 1296 shards per container
1537                                 if ( $repeat ) { // verbose hash dir format (e.g. "a/ab/abc")
1538                                         $hashDirRegex = $char . '/(' . $char . '{2})';
1539                                 } else { // short hash dir format (e.g. "a/b/c")
1540                                         $hashDirRegex = '(' . $char . ')/(' . $char . ')';
1541                                 }
1542                         }
1543                         // Allow certain directories to be above the hash dirs so as
1544                         // to work with FileRepo (e.g. "archive/a/ab" or "temp/a/ab").
1545                         // They must be 2+ chars to avoid any hash directory ambiguity.
1546                         $m = array();
1547                         if ( preg_match( "!^(?:[^/]{2,}/)*$hashDirRegex(?:/|$)!", $relPath, $m ) ) {
1548                                 return '.' . implode( '', array_slice( $m, 1 ) );
1549                         }
1550                         return null; // failed to match
1551                 }
1552                 return ''; // no sharding
1553         }
1554
1555         /**
1556          * Get the sharding config for a container.
1557          * If greater than 0, then all file storage paths within
1558          * the container are required to be hashed accordingly.
1559          *
1560          * @param $container string
1561          * @return Array (integer levels, integer base, repeat flag) or (0, 0, false)
1562          */
1563         final protected function getContainerHashLevels( $container ) {
1564                 if ( isset( $this->shardViaHashLevels[$container] ) ) {
1565                         $config = $this->shardViaHashLevels[$container];
1566                         $hashLevels = (int)$config['levels'];
1567                         if ( $hashLevels == 1 || $hashLevels == 2 ) {
1568                                 $hashBase = (int)$config['base'];
1569                                 if ( $hashBase == 16 || $hashBase == 36 ) {
1570                                         return array( $hashLevels, $hashBase, $config['repeat'] );
1571                                 }
1572                         }
1573                 }
1574                 return array( 0, 0, false ); // no sharding
1575         }
1576
1577         /**
1578          * Get a list of full container shard suffixes for a container
1579          *
1580          * @param $container string
1581          * @return Array
1582          */
1583         final protected function getContainerSuffixes( $container ) {
1584                 $shards = array();
1585                 list( $digits, $base ) = $this->getContainerHashLevels( $container );
1586                 if ( $digits > 0 ) {
1587                         $numShards = pow( $base, $digits );
1588                         for ( $index = 0; $index < $numShards; $index++ ) {
1589                                 $shards[] = '.' . wfBaseConvert( $index, 10, $base, $digits );
1590                         }
1591                 }
1592                 return $shards;
1593         }
1594
1595         /**
1596          * Get the full container name, including the wiki ID prefix
1597          *
1598          * @param $container string
1599          * @return string
1600          */
1601         final protected function fullContainerName( $container ) {
1602                 if ( $this->wikiId != '' ) {
1603                         return "{$this->wikiId}-$container";
1604                 } else {
1605                         return $container;
1606                 }
1607         }
1608
1609         /**
1610          * Resolve a container name, checking if it's allowed by the backend.
1611          * This is intended for internal use, such as encoding illegal chars.
1612          * Subclasses can override this to be more restrictive.
1613          *
1614          * @param $container string
1615          * @return string|null
1616          */
1617         protected function resolveContainerName( $container ) {
1618                 return $container;
1619         }
1620
1621         /**
1622          * Resolve a relative storage path, checking if it's allowed by the backend.
1623          * This is intended for internal use, such as encoding illegal chars or perhaps
1624          * getting absolute paths (e.g. FS based backends). Note that the relative path
1625          * may be the empty string (e.g. the path is simply to the container).
1626          *
1627          * @param $container string Container name
1628          * @param $relStoragePath string Storage path relative to the container
1629          * @return string|null Path or null if not valid
1630          */
1631         protected function resolveContainerPath( $container, $relStoragePath ) {
1632                 return $relStoragePath;
1633         }
1634 }
1635
1636 /**
1637  * FileBackendStore helper function to handle file listings that span container shards.
1638  * Do not use this class from places outside of FileBackendStore.
1639  *
1640  * @ingroup FileBackend
1641  */
1642 class FileBackendStoreShardListIterator implements Iterator {
1643         /* @var FileBackendStore */
1644         protected $backend;
1645         /* @var Array */
1646         protected $params;
1647         /* @var Array */
1648         protected $shardSuffixes;
1649         protected $container; // string
1650         protected $directory; // string
1651
1652         /* @var Traversable */
1653         protected $iter;
1654         protected $curShard = 0; // integer
1655         protected $pos = 0; // integer
1656
1657         /**
1658          * @param $backend FileBackendStore
1659          * @param $container string Full storage container name
1660          * @param $dir string Storage directory relative to container
1661          * @param $suffixes Array List of container shard suffixes
1662          * @param $params Array
1663          */
1664         public function __construct(
1665                 FileBackendStore $backend, $container, $dir, array $suffixes, array $params
1666         ) {
1667                 $this->backend = $backend;
1668                 $this->container = $container;
1669                 $this->directory = $dir;
1670                 $this->shardSuffixes = $suffixes;
1671                 $this->params = $params;
1672         }
1673
1674         public function current() {
1675                 if ( is_array( $this->iter ) ) {
1676                         return current( $this->iter );
1677                 } else {
1678                         return $this->iter->current();
1679                 }
1680         }
1681
1682         public function key() {
1683                 return $this->pos;
1684         }
1685
1686         public function next() {
1687                 ++$this->pos;
1688                 if ( is_array( $this->iter ) ) {
1689                         next( $this->iter );
1690                 } else {
1691                         $this->iter->next();
1692                 }
1693                 // Find the next non-empty shard if no elements are left
1694                 $this->nextShardIteratorIfNotValid();
1695         }
1696
1697         /**
1698          * If the iterator for this container shard is out of items,
1699          * then move on to the next container that has items.
1700          * If there are none, then it advances to the last container.
1701          */
1702         protected function nextShardIteratorIfNotValid() {
1703                 while ( !$this->valid() ) {
1704                         if ( ++$this->curShard >= count( $this->shardSuffixes ) ) {
1705                                 break; // no more container shards
1706                         }
1707                         $this->setIteratorFromCurrentShard();
1708                 }
1709         }
1710
1711         protected function setIteratorFromCurrentShard() {
1712                 $suffix = $this->shardSuffixes[$this->curShard];
1713                 $this->iter = $this->backend->getFileListInternal(
1714                         "{$this->container}{$suffix}", $this->directory, $this->params );
1715         }
1716
1717         public function rewind() {
1718                 $this->pos = 0;
1719                 $this->curShard = 0;
1720                 $this->setIteratorFromCurrentShard();
1721                 // Find the next non-empty shard if this one has no elements
1722                 $this->nextShardIteratorIfNotValid();
1723         }
1724
1725         public function valid() {
1726                 if ( $this->iter == null ) {
1727                         return false; // some failure?
1728                 } elseif ( is_array( $this->iter ) ) {
1729                         return ( current( $this->iter ) !== false ); // no paths can have this value
1730                 } else {
1731                         return $this->iter->valid();
1732                 }
1733         }
1734 }