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://container", then all items in
 535          * the container should be listed. If of the form "mwstore://container/dir",
 536          * then all items under that container directory should be listed.
 537          * Results should be storage paths relative to the given directory.
 538          *
 539          * $params include:
 540          *     dir : storage path directory
 541          *
 542          * @return Traversable|Array|null Returns null on failure
 543          */
 544         abstract public function getFileList( array $params );
 545
 546         /**
 547          * Invalidate any in-process file existence and property cache.
 548          * If $paths is given, then only the cache for those files will be cleared.
 549          *
 550          * @param $paths Array Storage paths (optional)
 551          * @return void
 552          */
 553         public function clearCache( array $paths = null ) {}
 554
 555         /**
 556          * Lock the files at the given storage paths in the backend.
 557          * This will either lock all the files or none (on failure).
 558          *
 559          * Callers should consider using getScopedFileLocks() instead.
 560          *
 561          * @param $paths Array Storage paths
 562          * @param $type integer LockManager::LOCK_* constant
 563          * @return Status
 564          */
 565         final public function lockFiles( array $paths, $type ) {
 566                 return $this->lockManager->lock( $paths, $type );
 567         }
 568
 569         /**
 570          * Unlock the files at the given storage paths in the backend.
 571          *
 572          * @param $paths Array Storage paths
 573          * @param $type integer LockManager::LOCK_* constant
 574          * @return Status
 575          */
 576         final public function unlockFiles( array $paths, $type ) {
 577                 return $this->lockManager->unlock( $paths, $type );
 578         }
 579
 580         /**
 581          * Lock the files at the given storage paths in the backend.
 582          * This will either lock all the files or none (on failure).
 583          * On failure, the status object will be updated with errors.
 584          *
 585          * Once the return value goes out scope, the locks will be released and
 586          * the status updated. Unlock fatals will not change the status "OK" value.
 587          *
 588          * @param $paths Array Storage paths
 589          * @param $type integer LockManager::LOCK_* constant
 590          * @param $status Status Status to update on lock/unlock
 591          * @return ScopedLock|null Returns null on failure
 592          */
 593         final public function getScopedFileLocks( array $paths, $type, Status $status ) {
 594                 return ScopedLock::factory( $this->lockManager, $paths, $type, $status );
 595         }
 596
 597         /**
 598          * Check if a given path is a "mwstore://" path.
 599          * This does not do any further validation or any existence checks.
 600          *
 601          * @param $path string
 602          * @return bool
 603          */
 604         final public static function isStoragePath( $path ) {
 605                 return ( strpos( $path, 'mwstore://' ) === 0 );
 606         }
 607
 608         /**
 609          * Split a storage path into a backend name, a container name,
 610          * and a relative file path. The relative path may be the empty string.
 611          * This does not do any path normalization or traversal checks.
 612          *
 613          * @param $storagePath string
 614          * @return Array (backend, container, rel object) or (null, null, null)
 615          */
 616         final public static function splitStoragePath( $storagePath ) {
 617                 if ( self::isStoragePath( $storagePath ) ) {
 618                         // Remove the "mwstore://" prefix and split the path
 619                         $parts = explode( '/', substr( $storagePath, 10 ), 3 );
 620                         if ( count( $parts ) >= 2 && $parts[0] != '' && $parts[1] != '' ) {
 621                                 if ( count( $parts ) == 3 ) {
 622                                         return $parts; // e.g. "backend/container/path"
 623                                 } else {
 624                                         return array( $parts[0], $parts[1], '' ); // e.g. "backend/container"
 625                                 }
 626                         }
 627                 }
 628                 return array( null, null, null );
 629         }
 630
 631         /**
 632          * Normalize a storage path by cleaning up directory separators.
 633          * Returns null if the path is not of the format of a valid storage path.
 634          *
 635          * @param $storagePath string
 636          * @return string|null
 637          */
 638         final public static function normalizeStoragePath( $storagePath ) {
 639                 list( $backend, $container, $relPath ) = self::splitStoragePath( $storagePath );
 640                 if ( $relPath !== null ) { // must be for this backend
 641                         $relPath = self::normalizeContainerPath( $relPath );
 642                         if ( $relPath !== null ) {
 643                                 return ( $relPath != '' )
 644                                         ? "mwstore://{$backend}/{$container}/{$relPath}"
 645                                         : "mwstore://{$backend}/{$container}";
 646                         }
 647                 }
 648                 return null;
 649         }
 650
 651         /**
 652          * Validate and normalize a relative storage path.
 653          * Null is returned if the path involves directory traversal.
 654          * Traversal is insecure for FS backends and broken for others.
 655          *
 656          * @param $path string Storage path relative to a container
 657          * @return string|null
 658          */
 659         final protected static function normalizeContainerPath( $path ) {
 660                 // Normalize directory separators
 661                 $path = strtr( $path, '\\', '/' );
 662                 // Collapse any consecutive directory separators
 663                 $path = preg_replace( '![/]{2,}!', '/', $path );
 664                 // Remove any leading directory separator
 665                 $path = ltrim( $path, '/' );
 666                 // Use the same traversal protection as Title::secureAndSplit()
 667                 if ( strpos( $path, '.' ) !== false ) {
 668                         if (
 669                                 $path === '.' ||
 670                                 $path === '..' ||
 671                                 strpos( $path, './' ) === 0 ||
 672                                 strpos( $path, '../' ) === 0 ||
 673                                 strpos( $path, '/./' ) !== false ||
 674                                 strpos( $path, '/../' ) !== false
 675                         ) {
 676                                 return null;
 677                         }
 678                 }
 679                 return $path;
 680         }
 681
 682         /**
 683          * Get the parent storage directory of a storage path.
 684          * This returns a path like "mwstore://backend/container",
 685          * "mwstore://backend/container/...", or null if there is no parent.
 686          *
 687          * @param $storagePath string
 688          * @return string|null
 689          */
 690         final public static function parentStoragePath( $storagePath ) {
 691                 $storagePath = dirname( $storagePath );
 692                 list( $b, $cont, $rel ) = self::splitStoragePath( $storagePath );
 693                 return ( $rel === null ) ? null : $storagePath;
 694         }
 695
 696         /**
 697          * Get the final extension from a storage or FS path
 698          *
 699          * @param $path string
 700          * @return string
 701          */
 702         final public static function extensionFromPath( $path ) {
 703                 $i = strrpos( $path, '.' );
 704                 return strtolower( $i ? substr( $path, $i + 1 ) : '' );
 705         }
 706 }
 707
 708 /**
 709  * @brief Base class for all backends associated with a particular storage medium.
 710  *
 711  * This class defines the methods as abstract that subclasses must implement.
 712  * Outside callers should *not* use functions with "Internal" in the name.
 713  *
 714  * The FileBackend operations are implemented using basic functions
 715  * such as storeInternal(), copyInternal(), deleteInternal() and the like.
 716  * This class is also responsible for path resolution and sanitization.
 717  *
 718  * @ingroup FileBackend
 719  * @since 1.19
 720  */
 721 abstract class FileBackendStore extends FileBackend {
 722         /** @var Array Map of paths to small (RAM/disk) cache items */
 723         protected $cache = array(); // (storage path => key => value)
 724         protected $maxCacheSize = 100; // integer; max paths with entries
 725         /** @var Array Map of paths to large (RAM/disk) cache items */
 726         protected $expensiveCache = array(); // (storage path => key => value)
 727         protected $maxExpensiveCacheSize = 10; // integer; max paths with entries
 728
 729         /** @var Array Map of container names to sharding settings */
 730         protected $shardViaHashLevels = array(); // (container name => config array)
 731
 732         protected $maxFileSize = 1000000000; // integer bytes (1GB)
 733
 734         /**
 735          * Get the maximum allowable file size given backend
 736          * medium restrictions and basic performance constraints.
 737          * Do not call this function from places outside FileBackend and FileOp.
 738          *
 739          * @return integer Bytes
 740          */
 741         final public function maxFileSizeInternal() {
 742                 return $this->maxFileSize;
 743         }
 744
 745         /**
 746          * Check if a file can be created at a given storage path.
 747          * FS backends should check if the parent directory exists and the file is writable.
 748          * Backends using key/value stores should check if the container exists.
 749          *
 750          * @param $storagePath string
 751          * @return bool
 752          */
 753         abstract public function isPathUsableInternal( $storagePath );
 754
 755         /**
 756          * Create a file in the backend with the given contents.
 757          * Do not call this function from places outside FileBackend and FileOp.
 758          *
 759          * $params include:
 760          *     content       : the raw file contents
 761          *     dst           : destination storage path
 762          *     overwrite     : overwrite any file that exists at the destination
 763          *
 764          * @param $params Array
 765          * @return Status
 766          */
 767         final public function createInternal( array $params ) {
 768                 wfProfileIn( __METHOD__ );
 769                 if ( strlen( $params['content'] ) > $this->maxFileSizeInternal() ) {
 770                         $status = Status::newFatal( 'backend-fail-create', $params['dst'] );
 771                 } else {
 772                         $status = $this->doCreateInternal( $params );
 773                         $this->clearCache( array( $params['dst'] ) );
 774                 }
 775                 wfProfileOut( __METHOD__ );
 776                 return $status;
 777         }
 778
 779         /**
 780          * @see FileBackendStore::createInternal()
 781          */
 782         abstract protected function doCreateInternal( array $params );
 783
 784         /**
 785          * Store a file into the backend from a file on disk.
 786          * Do not call this function from places outside FileBackend and FileOp.
 787          *
 788          * $params include:
 789          *     src           : source path on disk
 790          *     dst           : destination storage path
 791          *     overwrite     : overwrite any file that exists at the destination
 792          *
 793          * @param $params Array
 794          * @return Status
 795          */
 796         final public function storeInternal( array $params ) {
 797                 wfProfileIn( __METHOD__ );
 798                 if ( filesize( $params['src'] ) > $this->maxFileSizeInternal() ) {
 799                         $status = Status::newFatal( 'backend-fail-store', $params['dst'] );
 800                 } else {
 801                         $status = $this->doStoreInternal( $params );
 802                         $this->clearCache( array( $params['dst'] ) );
 803                 }
 804                 wfProfileOut( __METHOD__ );
 805                 return $status;
 806         }
 807
 808         /**
 809          * @see FileBackendStore::storeInternal()
 810          */
 811         abstract protected function doStoreInternal( array $params );
 812
 813         /**
 814          * Copy a file from one storage path to another in the backend.
 815          * Do not call this function from places outside FileBackend and FileOp.
 816          *
 817          * $params include:
 818          *     src           : source storage path
 819          *     dst           : destination storage path
 820          *     overwrite     : overwrite any file that exists at the destination
 821          *
 822          * @param $params Array
 823          * @return Status
 824          */
 825         final public function copyInternal( array $params ) {
 826                 wfProfileIn( __METHOD__ );
 827                 $status = $this->doCopyInternal( $params );
 828                 $this->clearCache( array( $params['dst'] ) );
 829                 wfProfileOut( __METHOD__ );
 830                 return $status;
 831         }
 832
 833         /**
 834          * @see FileBackendStore::copyInternal()
 835          */
 836         abstract protected function doCopyInternal( array $params );
 837
 838         /**
 839          * Delete a file at the storage path.
 840          * Do not call this function from places outside FileBackend and FileOp.
 841          *
 842          * $params include:
 843          *     src                 : source storage path
 844          *     ignoreMissingSource : do nothing if the source file does not exist
 845          *
 846          * @param $params Array
 847          * @return Status
 848          */
 849         final public function deleteInternal( array $params ) {
 850                 wfProfileIn( __METHOD__ );
 851                 $status = $this->doDeleteInternal( $params );
 852                 $this->clearCache( array( $params['src'] ) );
 853                 wfProfileOut( __METHOD__ );
 854                 return $status;
 855         }
 856
 857         /**
 858          * @see FileBackendStore::deleteInternal()
 859          */
 860         abstract protected function doDeleteInternal( array $params );
 861
 862         /**
 863          * Move a file from one storage path to another in the backend.
 864          * Do not call this function from places outside FileBackend and FileOp.
 865          *
 866          * $params include:
 867          *     src           : source storage path
 868          *     dst           : destination storage path
 869          *     overwrite     : overwrite any file that exists at the destination
 870          *
 871          * @param $params Array
 872          * @return Status
 873          */
 874         final public function moveInternal( array $params ) {
 875                 wfProfileIn( __METHOD__ );
 876                 $status = $this->doMoveInternal( $params );
 877                 $this->clearCache( array( $params['src'], $params['dst'] ) );
 878                 wfProfileOut( __METHOD__ );
 879                 return $status;
 880         }
 881
 882         /**
 883          * @see FileBackendStore::moveInternal()
 884          */
 885         protected function doMoveInternal( array $params ) {
 886                 // Copy source to dest
 887                 $status = $this->copyInternal( $params );
 888                 if ( $status->isOK() ) {
 889                         // Delete source (only fails due to races or medium going down)
 890                         $status->merge( $this->deleteInternal( array( 'src' => $params['src'] ) ) );
 891                         $status->setResult( true, $status->value ); // ignore delete() errors
 892                 }
 893                 return $status;
 894         }
 895
 896         /**
 897          * @see FileBackend::concatenate()
 898          */
 899         final public function concatenate( array $params ) {
 900                 wfProfileIn( __METHOD__ );
 901                 $status = Status::newGood();
 902
 903                 // Try to lock the source files for the scope of this function
 904                 $scopeLockS = $this->getScopedFileLocks( $params['srcs'], LockManager::LOCK_UW, $status );
 905                 if ( $status->isOK() ) {
 906                         // Actually do the concatenation
 907                         $status->merge( $this->doConcatenate( $params ) );
 908                 }
 909
 910                 wfProfileOut( __METHOD__ );
 911                 return $status;
 912         }
 913
 914         /**
 915          * @see FileBackendStore::concatenate()
 916          */
 917         protected function doConcatenate( array $params ) {
 918                 $status = Status::newGood();
 919                 $tmpPath = $params['dst']; // convenience
 920
 921                 // Check that the specified temp file is valid...
 922                 wfSuppressWarnings();
 923                 $ok = ( is_file( $tmpPath ) && !filesize( $tmpPath ) );
 924                 wfRestoreWarnings();
 925                 if ( !$ok ) { // not present or not empty
 926                         $status->fatal( 'backend-fail-opentemp', $tmpPath );
 927                         return $status;
 928                 }
 929
 930                 // Build up the temp file using the source chunks (in order)...
 931                 $tmpHandle = fopen( $tmpPath, 'ab' );
 932                 if ( $tmpHandle === false ) {
 933                         $status->fatal( 'backend-fail-opentemp', $tmpPath );
 934                         return $status;
 935                 }
 936                 foreach ( $params['srcs'] as $virtualSource ) {
 937                         // Get a local FS version of the chunk
 938                         $tmpFile = $this->getLocalReference( array( 'src' => $virtualSource ) );
 939                         if ( !$tmpFile ) {
 940                                 $status->fatal( 'backend-fail-read', $virtualSource );
 941                                 return $status;
 942                         }
 943                         // Get a handle to the local FS version
 944                         $sourceHandle = fopen( $tmpFile->getPath(), 'r' );
 945                         if ( $sourceHandle === false ) {
 946                                 fclose( $tmpHandle );
 947                                 $status->fatal( 'backend-fail-read', $virtualSource );
 948                                 return $status;
 949                         }
 950                         // Append chunk to file (pass chunk size to avoid magic quotes)
 951                         if ( !stream_copy_to_stream( $sourceHandle, $tmpHandle ) ) {
 952                                 fclose( $sourceHandle );
 953                                 fclose( $tmpHandle );
 954                                 $status->fatal( 'backend-fail-writetemp', $tmpPath );
 955                                 return $status;
 956                         }
 957                         fclose( $sourceHandle );
 958                 }
 959                 if ( !fclose( $tmpHandle ) ) {
 960                         $status->fatal( 'backend-fail-closetemp', $tmpPath );
 961                         return $status;
 962                 }
 963
 964                 clearstatcache(); // temp file changed
 965
 966                 return $status;
 967         }
 968
 969         /**
 970          * @see FileBackend::doPrepare()
 971          */
 972         final protected function doPrepare( array $params ) {
 973                 wfProfileIn( __METHOD__ );
 974
 975                 $status = Status::newGood();
 976                 list( $fullCont, $dir, $shard ) = $this->resolveStoragePath( $params['dir'] );
 977                 if ( $dir === null ) {
 978                         $status->fatal( 'backend-fail-invalidpath', $params['dir'] );
 979                         wfProfileOut( __METHOD__ );
 980                         return $status; // invalid storage path
 981                 }
 982
 983                 if ( $shard !== null ) { // confined to a single container/shard
 984                         $status->merge( $this->doPrepareInternal( $fullCont, $dir, $params ) );
 985                 } else { // directory is on several shards
 986                         wfDebug( __METHOD__ . ": iterating over all container shards.\n" );
 987                         list( $b, $shortCont, $r ) = self::splitStoragePath( $params['dir'] );
 988                         foreach ( $this->getContainerSuffixes( $shortCont ) as $suffix ) {
 989                                 $status->merge( $this->doPrepareInternal( "{$fullCont}{$suffix}", $dir, $params ) );
 990                         }
 991                 }
 992
 993                 wfProfileOut( __METHOD__ );
 994                 return $status;
 995         }
 996
 997         /**
 998          * @see FileBackendStore::doPrepare()
 999          */
1000         protected function doPrepareInternal( $container, $dir, array $params ) {
1001                 return Status::newGood();
1002         }
1003
1004         /**
1005          * @see FileBackend::doSecure()
1006          */
1007         final protected function doSecure( array $params ) {
1008                 wfProfileIn( __METHOD__ );
1009                 $status = Status::newGood();
1010
1011                 list( $fullCont, $dir, $shard ) = $this->resolveStoragePath( $params['dir'] );
1012                 if ( $dir === null ) {
1013                         $status->fatal( 'backend-fail-invalidpath', $params['dir'] );
1014                         wfProfileOut( __METHOD__ );
1015                         return $status; // invalid storage path
1016                 }
1017
1018                 if ( $shard !== null ) { // confined to a single container/shard
1019                         $status->merge( $this->doSecureInternal( $fullCont, $dir, $params ) );
1020                 } else { // directory is on several shards
1021                         wfDebug( __METHOD__ . ": iterating over all container shards.\n" );
1022                         list( $b, $shortCont, $r ) = self::splitStoragePath( $params['dir'] );
1023                         foreach ( $this->getContainerSuffixes( $shortCont ) as $suffix ) {
1024                                 $status->merge( $this->doSecureInternal( "{$fullCont}{$suffix}", $dir, $params ) );
1025                         }
1026                 }
1027
1028                 wfProfileOut( __METHOD__ );
1029                 return $status;
1030         }
1031
1032         /**
1033          * @see FileBackendStore::doSecure()
1034          */
1035         protected function doSecureInternal( $container, $dir, array $params ) {
1036                 return Status::newGood();
1037         }
1038
1039         /**
1040          * @see FileBackend::doClean()
1041          */
1042         final protected function doClean( array $params ) {
1043                 wfProfileIn( __METHOD__ );
1044                 $status = Status::newGood();
1045
1046                 list( $fullCont, $dir, $shard ) = $this->resolveStoragePath( $params['dir'] );
1047                 if ( $dir === null ) {
1048                         $status->fatal( 'backend-fail-invalidpath', $params['dir'] );
1049                         wfProfileOut( __METHOD__ );
1050                         return $status; // invalid storage path
1051                 }
1052
1053                 // Attempt to lock this directory...
1054                 $filesLockEx = array( $params['dir'] );
1055                 $scopedLockE = $this->getScopedFileLocks( $filesLockEx, LockManager::LOCK_EX, $status );
1056                 if ( !$status->isOK() ) {
1057                         wfProfileOut( __METHOD__ );
1058                         return $status; // abort
1059                 }
1060
1061                 if ( $shard !== null ) { // confined to a single container/shard
1062                         $status->merge( $this->doCleanInternal( $fullCont, $dir, $params ) );
1063                 } else { // directory is on several shards
1064                         wfDebug( __METHOD__ . ": iterating over all container shards.\n" );
1065                         list( $b, $shortCont, $r ) = self::splitStoragePath( $params['dir'] );
1066                         foreach ( $this->getContainerSuffixes( $shortCont ) as $suffix ) {
1067                                 $status->merge( $this->doCleanInternal( "{$fullCont}{$suffix}", $dir, $params ) );
1068                         }
1069                 }
1070
1071                 wfProfileOut( __METHOD__ );
1072                 return $status;
1073         }
1074
1075         /**
1076          * @see FileBackendStore::doClean()
1077          */
1078         protected function doCleanInternal( $container, $dir, array $params ) {
1079                 return Status::newGood();
1080         }
1081
1082         /**
1083          * @see FileBackend::fileExists()
1084          */
1085         final public function fileExists( array $params ) {
1086                 wfProfileIn( __METHOD__ );
1087                 $stat = $this->getFileStat( $params );
1088                 wfProfileOut( __METHOD__ );
1089                 return ( $stat === null ) ? null : (bool)$stat; // null => failure
1090         }
1091
1092         /**
1093          * @see FileBackend::getFileTimestamp()
1094          */
1095         final public function getFileTimestamp( array $params ) {
1096                 wfProfileIn( __METHOD__ );
1097                 $stat = $this->getFileStat( $params );
1098                 wfProfileOut( __METHOD__ );
1099                 return $stat ? $stat['mtime'] : false;
1100         }
1101
1102         /**
1103          * @see FileBackend::getFileSize()
1104          */
1105         final public function getFileSize( array $params ) {
1106                 wfProfileIn( __METHOD__ );
1107                 $stat = $this->getFileStat( $params );
1108                 wfProfileOut( __METHOD__ );
1109                 return $stat ? $stat['size'] : false;
1110         }
1111
1112         /**
1113          * @see FileBackend::getFileStat()
1114          */
1115         final public function getFileStat( array $params ) {
1116                 wfProfileIn( __METHOD__ );
1117                 $path = self::normalizeStoragePath( $params['src'] );
1118                 if ( $path === null ) {
1119                         return false; // invalid storage path
1120                 }
1121                 $latest = !empty( $params['latest'] );
1122                 if ( isset( $this->cache[$path]['stat'] ) ) {
1123                         // If we want the latest data, check that this cached
1124                         // value was in fact fetched with the latest available data.
1125                         if ( !$latest || $this->cache[$path]['stat']['latest'] ) {
1126                                 wfProfileOut( __METHOD__ );
1127                                 return $this->cache[$path]['stat'];
1128                         }
1129                 }
1130                 $stat = $this->doGetFileStat( $params );
1131                 if ( is_array( $stat ) ) { // don't cache negatives
1132                         $this->trimCache(); // limit memory
1133                         $this->cache[$path]['stat'] = $stat;
1134                         $this->cache[$path]['stat']['latest'] = $latest;
1135                 }
1136                 wfProfileOut( __METHOD__ );
1137                 return $stat;
1138         }
1139
1140         /**
1141          * @see FileBackendStore::getFileStat()
1142          */
1143         abstract protected function doGetFileStat( array $params );
1144
1145         /**
1146          * @see FileBackend::getFileContents()
1147          */
1148         public function getFileContents( array $params ) {
1149                 wfProfileIn( __METHOD__ );
1150                 $tmpFile = $this->getLocalReference( $params );
1151                 if ( !$tmpFile ) {
1152                         wfProfileOut( __METHOD__ );
1153                         return false;
1154                 }
1155                 wfSuppressWarnings();
1156                 $data = file_get_contents( $tmpFile->getPath() );
1157                 wfRestoreWarnings();
1158                 wfProfileOut( __METHOD__ );
1159                 return $data;
1160         }
1161
1162         /**
1163          * @see FileBackend::getFileSha1Base36()
1164          */
1165         final public function getFileSha1Base36( array $params ) {
1166                 wfProfileIn( __METHOD__ );
1167                 $path = $params['src'];
1168                 if ( isset( $this->cache[$path]['sha1'] ) ) {
1169                         wfProfileOut( __METHOD__ );
1170                         return $this->cache[$path]['sha1'];
1171                 }
1172                 $hash = $this->doGetFileSha1Base36( $params );
1173                 if ( $hash ) { // don't cache negatives
1174                         $this->trimCache(); // limit memory
1175                         $this->cache[$path]['sha1'] = $hash;
1176                 }
1177                 wfProfileOut( __METHOD__ );
1178                 return $hash;
1179         }
1180
1181         /**
1182          * @see FileBackendStore::getFileSha1Base36()
1183          */
1184         protected function doGetFileSha1Base36( array $params ) {
1185                 $fsFile = $this->getLocalReference( $params );
1186                 if ( !$fsFile ) {
1187                         return false;
1188                 } else {
1189                         return $fsFile->getSha1Base36();
1190                 }
1191         }
1192
1193         /**
1194          * @see FileBackend::getFileProps()
1195          */
1196         final public function getFileProps( array $params ) {
1197                 wfProfileIn( __METHOD__ );
1198                 $fsFile = $this->getLocalReference( $params );
1199                 $props = $fsFile ? $fsFile->getProps() : FSFile::placeholderProps();
1200                 wfProfileOut( __METHOD__ );
1201                 return $props;
1202         }
1203
1204         /**
1205          * @see FileBackend::getLocalReference()
1206          */
1207         public function getLocalReference( array $params ) {
1208                 wfProfileIn( __METHOD__ );
1209                 $path = $params['src'];
1210                 if ( isset( $this->expensiveCache[$path]['localRef'] ) ) {
1211                         wfProfileOut( __METHOD__ );
1212                         return $this->expensiveCache[$path]['localRef'];
1213                 }
1214                 $tmpFile = $this->getLocalCopy( $params );
1215                 if ( $tmpFile ) { // don't cache negatives
1216                         $this->trimExpensiveCache(); // limit memory
1217                         $this->expensiveCache[$path]['localRef'] = $tmpFile;
1218                 }
1219                 wfProfileOut( __METHOD__ );
1220                 return $tmpFile;
1221         }
1222
1223         /**
1224          * @see FileBackend::streamFile()
1225          */
1226         final public function streamFile( array $params ) {
1227                 wfProfileIn( __METHOD__ );
1228                 $status = Status::newGood();
1229
1230                 $info = $this->getFileStat( $params );
1231                 if ( !$info ) { // let StreamFile handle the 404
1232                         $status->fatal( 'backend-fail-notexists', $params['src'] );
1233                 }
1234
1235                 // Set output buffer and HTTP headers for stream
1236                 $extraHeaders = $params['headers'] ? $params['headers'] : array();
1237                 $res = StreamFile::prepareForStream( $params['src'], $info, $extraHeaders );
1238                 if ( $res == StreamFile::NOT_MODIFIED ) {
1239                         // do nothing; client cache is up to date
1240                 } elseif ( $res == StreamFile::READY_STREAM ) {
1241                         $status = $this->doStreamFile( $params );
1242                 } else {
1243                         $status->fatal( 'backend-fail-stream', $params['src'] );
1244                 }
1245
1246                 wfProfileOut( __METHOD__ );
1247                 return $status;
1248         }
1249
1250         /**
1251          * @see FileBackendStore::streamFile()
1252          */
1253         protected function doStreamFile( array $params ) {
1254                 $status = Status::newGood();
1255
1256                 $fsFile = $this->getLocalReference( $params );
1257                 if ( !$fsFile ) {
1258                         $status->fatal( 'backend-fail-stream', $params['src'] );
1259                 } elseif ( !readfile( $fsFile->getPath() ) ) {
1260                         $status->fatal( 'backend-fail-stream', $params['src'] );
1261                 }
1262
1263                 return $status;
1264         }
1265
1266         /**
1267          * @copydoc FileBackend::getFileList()
1268          */
1269         final public function getFileList( array $params ) {
1270                 list( $fullCont, $dir, $shard ) = $this->resolveStoragePath( $params['dir'] );
1271                 if ( $dir === null ) { // invalid storage path
1272                         return null;
1273                 }
1274                 if ( $shard !== null ) {
1275                         // File listing is confined to a single container/shard
1276                         return $this->getFileListInternal( $fullCont, $dir, $params );
1277                 } else {
1278                         wfDebug( __METHOD__ . ": iterating over all container shards.\n" );
1279                         // File listing spans multiple containers/shards
1280                         list( $b, $shortCont, $r ) = self::splitStoragePath( $params['dir'] );
1281                         return new FileBackendStoreShardListIterator( $this,
1282                                 $fullCont, $dir, $this->getContainerSuffixes( $shortCont ), $params );
1283                 }
1284         }
1285
1286         /**
1287          * Do not call this function from places outside FileBackend
1288          *
1289          * @see FileBackendStore::getFileList()
1290          *
1291          * @param $container string Resolved container name
1292          * @param $dir string Resolved path relative to container
1293          * @param $params Array
1294          * @return Traversable|Array|null
1295          */
1296         abstract public function getFileListInternal( $container, $dir, array $params );
1297
1298         /**
1299          * Get the list of supported operations and their corresponding FileOp classes.
1300          *
1301          * @return Array
1302          */
1303         protected function supportedOperations() {
1304                 return array(
1305                         'store'       => 'StoreFileOp',
1306                         'copy'        => 'CopyFileOp',
1307                         'move'        => 'MoveFileOp',
1308                         'delete'      => 'DeleteFileOp',
1309                         'create'      => 'CreateFileOp',
1310                         'null'        => 'NullFileOp'
1311                 );
1312         }
1313
1314         /**
1315          * Return a list of FileOp objects from a list of operations.
1316          * Do not call this function from places outside FileBackend.
1317          *
1318          * The result must have the same number of items as the input.
1319          * An exception is thrown if an unsupported operation is requested.
1320          *
1321          * @param $ops Array Same format as doOperations()
1322          * @return Array List of FileOp objects
1323          * @throws MWException
1324          */
1325         final public function getOperations( array $ops ) {
1326                 $supportedOps = $this->supportedOperations();
1327
1328                 $performOps = array(); // array of FileOp objects
1329                 // Build up ordered array of FileOps...
1330                 foreach ( $ops as $operation ) {
1331                         $opName = $operation['op'];
1332                         if ( isset( $supportedOps[$opName] ) ) {
1333                                 $class = $supportedOps[$opName];
1334                                 // Get params for this operation
1335                                 $params = $operation;
1336                                 // Append the FileOp class
1337                                 $performOps[] = new $class( $this, $params );
1338                         } else {
1339                                 throw new MWException( "Operation `$opName` is not supported." );
1340                         }
1341                 }
1342
1343                 return $performOps;
1344         }
1345
1346         /**
1347          * @see FileBackend::doOperationsInternal()
1348          */
1349         protected function doOperationsInternal( array $ops, array $opts ) {
1350                 wfProfileIn( __METHOD__ );
1351                 $status = Status::newGood();
1352
1353                 // Build up a list of FileOps...
1354                 $performOps = $this->getOperations( $ops );
1355
1356                 // Acquire any locks as needed...
1357                 if ( empty( $opts['nonLocking'] ) ) {
1358                         // Build up a list of files to lock...
1359                         $filesLockEx = $filesLockSh = array();
1360                         foreach ( $performOps as $fileOp ) {
1361                                 $filesLockSh = array_merge( $filesLockSh, $fileOp->storagePathsRead() );
1362                                 $filesLockEx = array_merge( $filesLockEx, $fileOp->storagePathsChanged() );
1363                         }
1364                         // Optimization: if doing an EX lock anyway, don't also set an SH one
1365                         $filesLockSh = array_diff( $filesLockSh, $filesLockEx );
1366                         // Get a shared lock on the parent directory of each path changed
1367                         $filesLockSh = array_merge( $filesLockSh, array_map( 'dirname', $filesLockEx ) );
1368                         // Try to lock those files for the scope of this function...
1369                         $scopeLockS = $this->getScopedFileLocks( $filesLockSh, LockManager::LOCK_UW, $status );
1370                         $scopeLockE = $this->getScopedFileLocks( $filesLockEx, LockManager::LOCK_EX, $status );
1371                         if ( !$status->isOK() ) {
1372                                 wfProfileOut( __METHOD__ );
1373                                 return $status; // abort
1374                         }
1375                 }
1376
1377                 // Clear any cache entries (after locks acquired)
1378                 $this->clearCache();
1379
1380                 // Actually attempt the operation batch...
1381                 $subStatus = FileOp::attemptBatch( $performOps, $opts );
1382
1383                 // Merge errors into status fields
1384                 $status->merge( $subStatus );
1385                 $status->success = $subStatus->success; // not done in merge()
1386
1387                 wfProfileOut( __METHOD__ );
1388                 return $status;
1389         }
1390
1391         /**
1392          * @see FileBackend::clearCache()
1393          */
1394         final public function clearCache( array $paths = null ) {
1395                 if ( is_array( $paths ) ) {
1396                         $paths = array_map( 'FileBackend::normalizeStoragePath', $paths );
1397                         $paths = array_filter( $paths, 'strlen' ); // remove nulls
1398                 }
1399                 if ( $paths === null ) {
1400                         $this->cache = array();
1401                         $this->expensiveCache = array();
1402                 } else {
1403                         foreach ( $paths as $path ) {
1404                                 unset( $this->cache[$path] );
1405                                 unset( $this->expensiveCache[$path] );
1406                         }
1407                 }
1408                 $this->doClearCache( $paths );
1409         }
1410
1411         /**
1412          * Clears any additional stat caches for storage paths
1413          *
1414          * @see FileBackend::clearCache()
1415          *
1416          * @param $paths Array Storage paths (optional)
1417          * @return void
1418          */
1419         protected function doClearCache( array $paths = null ) {}
1420
1421         /**
1422          * Prune the inexpensive cache if it is too big to add an item
1423          *
1424          * @return void
1425          */
1426         protected function trimCache() {
1427                 if ( count( $this->cache ) >= $this->maxCacheSize ) {
1428                         reset( $this->cache );
1429                         unset( $this->cache[key( $this->cache )] );
1430                 }
1431         }
1432
1433         /**
1434          * Prune the expensive cache if it is too big to add an item
1435          *
1436          * @return void
1437          */
1438         protected function trimExpensiveCache() {
1439                 if ( count( $this->expensiveCache ) >= $this->maxExpensiveCacheSize ) {
1440                         reset( $this->expensiveCache );
1441                         unset( $this->expensiveCache[key( $this->expensiveCache )] );
1442                 }
1443         }
1444
1445         /**
1446          * Check if a container name is valid.
1447          * This checks for for length and illegal characters.
1448          *
1449          * @param $container string
1450          * @return bool
1451          */
1452         final protected static function isValidContainerName( $container ) {
1453                 // This accounts for Swift and S3 restrictions while leaving room
1454                 // for things like '.xxx' (hex shard chars) or '.seg' (segments).
1455                 // This disallows directory separators or traversal characters.
1456                 // Note that matching strings URL encode to the same string;
1457                 // in Swift, the length restriction is *after* URL encoding.
1458                 return preg_match( '/^[a-z0-9][a-z0-9-_]{0,199}$/i', $container );
1459         }
1460
1461         /**
1462          * Splits a storage path into an internal container name,
1463          * an internal relative file name, and a container shard suffix.
1464          * Any shard suffix is already appended to the internal container name.
1465          * This also checks that the storage path is valid and within this backend.
1466          *
1467          * If the container is sharded but a suffix could not be determined,
1468          * this means that the path can only refer to a directory and can only
1469          * be scanned by looking in all the container shards.
1470          *
1471          * @param $storagePath string
1472          * @return Array (container, path, container suffix) or (null, null, null) if invalid
1473          */
1474         final protected function resolveStoragePath( $storagePath ) {
1475                 list( $backend, $container, $relPath ) = self::splitStoragePath( $storagePath );
1476                 if ( $backend === $this->name ) { // must be for this backend
1477                         $relPath = self::normalizeContainerPath( $relPath );
1478                         if ( $relPath !== null ) {
1479                                 // Get shard for the normalized path if this container is sharded
1480                                 $cShard = $this->getContainerShard( $container, $relPath );
1481                                 // Validate and sanitize the relative path (backend-specific)
1482                                 $relPath = $this->resolveContainerPath( $container, $relPath );
1483                                 if ( $relPath !== null ) {
1484                                         // Prepend any wiki ID prefix to the container name
1485                                         $container = $this->fullContainerName( $container );
1486                                         if ( self::isValidContainerName( $container ) ) {
1487                                                 // Validate and sanitize the container name (backend-specific)
1488                                                 $container = $this->resolveContainerName( "{$container}{$cShard}" );
1489                                                 if ( $container !== null ) {
1490                                                         return array( $container, $relPath, $cShard );
1491                                                 }
1492                                         }
1493                                 }
1494                         }
1495                 }
1496                 return array( null, null, null );
1497         }
1498
1499         /**
1500          * Like resolveStoragePath() except null values are returned if
1501          * the container is sharded and the shard could not be determined.
1502          *
1503          * @see FileBackendStore::resolveStoragePath()
1504          *
1505          * @param $storagePath string
1506          * @return Array (container, path) or (null, null) if invalid
1507          */
1508         final protected function resolveStoragePathReal( $storagePath ) {
1509                 list( $container, $relPath, $cShard ) = $this->resolveStoragePath( $storagePath );
1510                 if ( $cShard !== null ) {
1511                         return array( $container, $relPath );
1512                 }
1513                 return array( null, null );
1514         }
1515
1516         /**
1517          * Get the container name shard suffix for a given path.
1518          * Any empty suffix means the container is not sharded.
1519          *
1520          * @param $container string Container name
1521          * @param $relStoragePath string Storage path relative to the container
1522          * @return string|null Returns null if shard could not be determined
1523          */
1524         final protected function getContainerShard( $container, $relPath ) {
1525                 list( $levels, $base, $repeat ) = $this->getContainerHashLevels( $container );
1526                 if ( $levels == 1 || $levels == 2 ) {
1527                         // Hash characters are either base 16 or 36
1528                         $char = ( $base == 36 ) ? '[0-9a-z]' : '[0-9a-f]';
1529                         // Get a regex that represents the shard portion of paths.
1530                         // The concatenation of the captures gives us the shard.
1531                         if ( $levels === 1 ) { // 16 or 36 shards per container
1532                                 $hashDirRegex = '(' . $char . ')';
1533                         } else { // 256 or 1296 shards per container
1534                                 if ( $repeat ) { // verbose hash dir format (e.g. "a/ab/abc")
1535                                         $hashDirRegex = $char . '/(' . $char . '{2})';
1536                                 } else { // short hash dir format (e.g. "a/b/c")
1537                                         $hashDirRegex = '(' . $char . ')/(' . $char . ')';
1538                                 }
1539                         }
1540                         // Allow certain directories to be above the hash dirs so as
1541                         // to work with FileRepo (e.g. "archive/a/ab" or "temp/a/ab").
1542                         // They must be 2+ chars to avoid any hash directory ambiguity.
1543                         $m = array();
1544                         if ( preg_match( "!^(?:[^/]{2,}/)*$hashDirRegex(?:/|$)!", $relPath, $m ) ) {
1545                                 return '.' . implode( '', array_slice( $m, 1 ) );
1546                         }
1547                         return null; // failed to match
1548                 }
1549                 return ''; // no sharding
1550         }
1551
1552         /**
1553          * Get the sharding config for a container.
1554          * If greater than 0, then all file storage paths within
1555          * the container are required to be hashed accordingly.
1556          *
1557          * @param $container string
1558          * @return Array (integer levels, integer base, repeat flag) or (0, 0, false)
1559          */
1560         final protected function getContainerHashLevels( $container ) {
1561                 if ( isset( $this->shardViaHashLevels[$container] ) ) {
1562                         $config = $this->shardViaHashLevels[$container];
1563                         $hashLevels = (int)$config['levels'];
1564                         if ( $hashLevels == 1 || $hashLevels == 2 ) {
1565                                 $hashBase = (int)$config['base'];
1566                                 if ( $hashBase == 16 || $hashBase == 36 ) {
1567                                         return array( $hashLevels, $hashBase, $config['repeat'] );
1568                                 }
1569                         }
1570                 }
1571                 return array( 0, 0, false ); // no sharding
1572         }
1573
1574         /**
1575          * Get a list of full container shard suffixes for a container
1576          *
1577          * @param $container string
1578          * @return Array
1579          */
1580         final protected function getContainerSuffixes( $container ) {
1581                 $shards = array();
1582                 list( $digits, $base ) = $this->getContainerHashLevels( $container );
1583                 if ( $digits > 0 ) {
1584                         $numShards = pow( $base, $digits );
1585                         for ( $index = 0; $index < $numShards; $index++ ) {
1586                                 $shards[] = '.' . wfBaseConvert( $index, 10, $base, $digits );
1587                         }
1588                 }
1589                 return $shards;
1590         }
1591
1592         /**
1593          * Get the full container name, including the wiki ID prefix
1594          *
1595          * @param $container string
1596          * @return string
1597          */
1598         final protected function fullContainerName( $container ) {
1599                 if ( $this->wikiId != '' ) {
1600                         return "{$this->wikiId}-$container";
1601                 } else {
1602                         return $container;
1603                 }
1604         }
1605
1606         /**
1607          * Resolve a container name, checking if it's allowed by the backend.
1608          * This is intended for internal use, such as encoding illegal chars.
1609          * Subclasses can override this to be more restrictive.
1610          *
1611          * @param $container string
1612          * @return string|null
1613          */
1614         protected function resolveContainerName( $container ) {
1615                 return $container;
1616         }
1617
1618         /**
1619          * Resolve a relative storage path, checking if it's allowed by the backend.
1620          * This is intended for internal use, such as encoding illegal chars or perhaps
1621          * getting absolute paths (e.g. FS based backends). Note that the relative path
1622          * may be the empty string (e.g. the path is simply to the container).
1623          *
1624          * @param $container string Container name
1625          * @param $relStoragePath string Storage path relative to the container
1626          * @return string|null Path or null if not valid
1627          */
1628         protected function resolveContainerPath( $container, $relStoragePath ) {
1629                 return $relStoragePath;
1630         }
1631 }
1632
1633 /**
1634  * FileBackendStore helper function to handle file listings that span container shards.
1635  * Do not use this class from places outside of FileBackendStore.
1636  *
1637  * @ingroup FileBackend
1638  */
1639 class FileBackendStoreShardListIterator implements Iterator {
1640         /* @var FileBackendStore */
1641         protected $backend;
1642         /* @var Array */
1643         protected $params;
1644         /* @var Array */
1645         protected $shardSuffixes;
1646         protected $container; // string
1647         protected $directory; // string
1648
1649         /* @var Traversable */
1650         protected $iter;
1651         protected $curShard = 0; // integer
1652         protected $pos = 0; // integer
1653
1654         /**
1655          * @param $backend FileBackendStore
1656          * @param $container string Full storage container name
1657          * @param $dir string Storage directory relative to container
1658          * @param $suffixes Array List of container shard suffixes
1659          * @param $params Array
1660          */
1661         public function __construct(
1662                 FileBackendStore $backend, $container, $dir, array $suffixes, array $params
1663         ) {
1664                 $this->backend = $backend;
1665                 $this->container = $container;
1666                 $this->directory = $dir;
1667                 $this->shardSuffixes = $suffixes;
1668                 $this->params = $params;
1669         }
1670
1671         public function current() {
1672                 if ( is_array( $this->iter ) ) {
1673                         return current( $this->iter );
1674                 } else {
1675                         return $this->iter->current();
1676                 }
1677         }
1678
1679         public function key() {
1680                 return $this->pos;
1681         }
1682
1683         public function next() {
1684                 ++$this->pos;
1685                 if ( is_array( $this->iter ) ) {
1686                         next( $this->iter );
1687                 } else {
1688                         $this->iter->next();
1689                 }
1690                 // Find the next non-empty shard if no elements are left
1691                 $this->nextShardIteratorIfNotValid();
1692         }
1693
1694         /**
1695          * If the iterator for this container shard is out of items,
1696          * then move on to the next container that has items.
1697          * If there are none, then it advances to the last container.
1698          */
1699         protected function nextShardIteratorIfNotValid() {
1700                 while ( !$this->valid() ) {
1701                         if ( ++$this->curShard >= count( $this->shardSuffixes ) ) {
1702                                 break; // no more container shards
1703                         }
1704                         $this->setIteratorFromCurrentShard();
1705                 }
1706         }
1707
1708         protected function setIteratorFromCurrentShard() {
1709                 $suffix = $this->shardSuffixes[$this->curShard];
1710                 $this->iter = $this->backend->getFileListInternal(
1711                         "{$this->container}{$suffix}", $this->directory, $this->params );
1712         }
1713
1714         public function rewind() {
1715                 $this->pos = 0;
1716                 $this->curShard = 0;
1717                 $this->setIteratorFromCurrentShard();
1718                 // Find the next non-empty shard if this one has no elements
1719                 $this->nextShardIteratorIfNotValid();
1720         }
1721
1722         public function valid() {
1723                 if ( $this->iter == null ) {
1724                         return false; // some failure?
1725                 } elseif ( is_array( $this->iter ) ) {
1726                         return ( current( $this->iter ) !== false ); // no paths can have this value
1727                 } else {
1728                         return $this->iter->valid();
1729                 }
1730         }
1731 }