includes/filerepo/backend/FileBackend.php

   1 <?php
   2 /**
   3  * @file
   4  * @ingroup FileBackend
   5  * @author Aaron Schulz
   6  */
   7
   8 /**
   9  * Base class for all file backend classes (including multi-write backends).
  10  * This class defines the methods as abstract that subclasses must implement.
  11  * Outside callers can assume that all backends will have these functions.
  12  *
  13  * All "storage paths" are of the format "mwstore://backend/container/path".
  14  * The paths use typical file system (FS) notation, though any particular backend may
  15  * not actually be using a local filesystem. Therefore, the paths are only virtual.
  16  *
  17  * FS-based backends are somewhat more restrictive due to the existence of real
  18  * directory files; a regular file cannot have the same name as a directory. Other
  19  * backends with virtual directories may not have this limitation. Callers should
  20  * store files in such a way that no files and directories under the same path.
  21  *
  22  * Methods should avoid throwing exceptions at all costs.
  23  * As a corollary, external dependencies should be kept to a minimum.
  24  *
  25  * @ingroup FileBackend
  26  * @since 1.19
  27  */
  28 abstract class FileBackendBase {
  29         protected $name; // unique backend name
  30         protected $wikiId; // unique wiki name
  31         protected $readOnly; // string
  32         /** @var LockManager */
  33         protected $lockManager;
  34
  35         /**
  36          * Build a new object from configuration.
  37          * This should only be called from within FileBackendGroup.
  38          *
  39          * $config includes:
  40          *     'name'        : The unique name of this backend
  41          *     'wikiId'      : Prefix to container names that is unique to this wiki
  42          *     'lockManager' : Registered name of a file lock manager to use
  43          *     'readOnly'    : Write operations are disallowed if this is a non-empty string.
  44          *                     It should be an explanation for the backend being read-only.
  45          *
  46          * @param $config Array
  47          */
  48         public function __construct( array $config ) {
  49                 $this->name = $config['name'];
  50                 $this->wikiId = isset( $config['wikiId'] )
  51                         ? $config['wikiId']
  52                         : wfWikiID();
  53                 $this->lockManager = LockManagerGroup::singleton()->get( $config['lockManager'] );
  54                 $this->readOnly = isset( $config['readOnly'] )
  55                         ? (string)$config['readOnly']
  56                         : '';
  57         }
  58
  59         /**
  60          * Get the unique backend name.
  61          *
  62          * We may have multiple different backends of the same type.
  63          * For example, we can have two Swift backends using different proxies.
  64          *
  65          * @return string
  66          */
  67         final public function getName() {
  68                 return $this->name;
  69         }
  70
  71         /**
  72          * This is the main entry point into the backend for write operations.
  73          * Callers supply an ordered list of operations to perform as a transaction.
  74          * If any serious errors occur, all attempted operations will be rolled back.
  75          *
  76          * $ops is an array of arrays. The outer array holds a list of operations.
  77          * Each inner array is a set of key value pairs that specify an operation.
  78          *
  79          * Supported operations and their parameters:
  80          * a) Create a new file in storage with the contents of a string
  81          *     array(
  82          *         'op'                  => 'create',
  83          *         'dst'                 => <storage path>,
  84          *         'content'             => <string of new file contents>,
  85          *         'overwriteDest'       => <boolean>,
  86          *         'overwriteSame'       => <boolean>
  87          *     )
  88          * b) Copy a file system file into storage
  89          *     array(
  90          *         'op'                  => 'store',
  91          *         'src'                 => <file system path>,
  92          *         'dst'                 => <storage path>,
  93          *         'overwriteDest'       => <boolean>,
  94          *         'overwriteSame'       => <boolean>
  95          *     )
  96          * c) Copy a file within storage
  97          *     array(
  98          *         'op'                  => 'copy',
  99          *         'src'                 => <storage path>,
 100          *         'dst'                 => <storage path>,
 101          *         'overwriteDest'       => <boolean>,
 102          *         'overwriteSame'       => <boolean>
 103          *     )
 104          * d) Move a file within storage
 105          *     array(
 106          *         'op'                  => 'move',
 107          *         'src'                 => <storage path>,
 108          *         'dst'                 => <storage path>,
 109          *         'overwriteDest'       => <boolean>,
 110          *         'overwriteSame'       => <boolean>
 111          *     )
 112          * e) Delete a file within storage
 113          *     array(
 114          *         'op'                  => 'delete',
 115          *         'src'                 => <storage path>,
 116          *         'ignoreMissingSource' => <boolean>
 117          *     )
 118          * f) Concatenate a list of files within storage into a single temp file
 119          *     array(
 120          *         'op'                  => 'concatenate',
 121          *         'srcs'                => <ordered array of storage paths>,
 122          *         'dst'                 => <file system path to 0-byte temp file>
 123          *     )
 124          * g) Do nothing (no-op)
 125          *     array(
 126          *         'op'                  => 'null',
 127          *     )
 128          *
 129          * Boolean flags for operations (operation-specific):
 130          * 'ignoreMissingSource' : The operation will simply succeed and do
 131          *                         nothing if the source file does not exist.
 132          * 'overwriteDest'       : Any destination file will be overwritten.
 133          * 'overwriteSame'       : An error will not be given if a file already
 134          *                         exists at the destination that has the same
 135          *                         contents as the new contents to be written there.
 136          *
 137          * $opts is an associative of boolean flags, including:
 138          * 'ignoreErrors'        : Errors that would normally cause a rollback do not.
 139          *                         The remaining operations are still attempted if any fail.
 140          * 'nonLocking'          : No locks are acquired for the operations.
 141          *                         This can increase performance for non-critical writes.
 142          *                         This has no effect unless the 'ignoreErrors' flag is set.
 143          * 'allowStale'          : Don't require the latest available data.
 144          *                         This can increase performance for non-critical writes.
 145          *                         This has no effect unless the 'ignoreErrors' flag is set.
 146          *
 147          * Return value:
 148          * This returns a Status, which contains all warnings and fatals that occured
 149          * during the operation. The 'failCount', 'successCount', and 'success' members
 150          * will reflect each operation attempted. The status will be "OK" unless any
 151          * of the operations failed and the 'ignoreErrors' parameter was not set.
 152          *
 153          * @param $ops Array List of operations to execute in order
 154          * @param $opts Array Batch operation options
 155          * @return Status
 156          */
 157         final public function doOperations( array $ops, array $opts = array() ) {
 158                 if ( $this->readOnly != '' ) {
 159                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 160                 }
 161                 if ( empty( $opts['ignoreErrors'] ) ) { // sanity
 162                         unset( $opts['nonLocking'] );
 163                         unset( $opts['allowStale'] );
 164                 }
 165                 return $this->doOperationsInternal( $ops, $opts );
 166         }
 167
 168         /**
 169          * @see FileBackendBase::doOperations()
 170          */
 171         abstract protected function doOperationsInternal( array $ops, array $opts );
 172
 173         /**
 174          * Same as doOperations() except it takes a single operation.
 175          * If you are doing a batch of operations that should either
 176          * all succeed or all fail, then use that function instead.
 177          *
 178          * @see FileBackendBase::doOperations()
 179          *
 180          * @param $op Array Operation
 181          * @param $opts Array Operation options
 182          * @return Status
 183          */
 184         final public function doOperation( array $op, array $opts = array() ) {
 185                 return $this->doOperations( array( $op ), $opts );
 186         }
 187
 188         /**
 189          * Performs a single store operation.
 190          * This sets $params['op'] to 'store' and passes it to doOperation().
 191          *
 192          * @see FileBackendBase::doOperation()
 193          *
 194          * @param $params Array Operation parameters
 195          * @param $opts Array Operation options
 196          * @return Status
 197          */
 198         final public function store( array $params, array $opts = array() ) {
 199                 $params['op'] = 'store';
 200                 return $this->doOperation( $params, $opts );
 201         }
 202
 203         /**
 204          * Performs a single copy operation.
 205          * This sets $params['op'] to 'copy' and passes it to doOperation().
 206          *
 207          * @see FileBackendBase::doOperation()
 208          *
 209          * @param $params Array Operation parameters
 210          * @param $opts Array Operation options
 211          * @return Status
 212          */
 213         final public function copy( array $params, array $opts = array() ) {
 214                 $params['op'] = 'copy';
 215                 return $this->doOperation( $params, $opts );
 216         }
 217
 218         /**
 219          * Performs a single move operation.
 220          * This sets $params['op'] to 'move' and passes it to doOperation().
 221          *
 222          * @see FileBackendBase::doOperation()
 223          *
 224          * @param $params Array Operation parameters
 225          * @param $opts Array Operation options
 226          * @return Status
 227          */
 228         final public function move( array $params, array $opts = array() ) {
 229                 $params['op'] = 'move';
 230                 return $this->doOperation( $params, $opts );
 231         }
 232
 233         /**
 234          * Performs a single delete operation.
 235          * This sets $params['op'] to 'delete' and passes it to doOperation().
 236          *
 237          * @see FileBackendBase::doOperation()
 238          *
 239          * @param $params Array Operation parameters
 240          * @param $opts Array Operation options
 241          * @return Status
 242          */
 243         final public function delete( array $params, array $opts = array() ) {
 244                 $params['op'] = 'delete';
 245                 return $this->doOperation( $params, $opts );
 246         }
 247
 248         /**
 249          * Performs a single create operation.
 250          * This sets $params['op'] to 'create' and passes it to doOperation().
 251          *
 252          * @see FileBackendBase::doOperation()
 253          *
 254          * @param $params Array Operation parameters
 255          * @param $opts Array Operation options
 256          * @return Status
 257          */
 258         final public function create( array $params, array $opts = array() ) {
 259                 $params['op'] = 'create';
 260                 return $this->doOperation( $params, $opts );
 261         }
 262
 263         /**
 264          * Performs a single concatenate operation.
 265          * This sets $params['op'] to 'concatenate' and passes it to doOperation().
 266          *
 267          * @see FileBackendBase::doOperation()
 268          *
 269          * @param $params Array Operation parameters
 270          * @param $opts Array Operation options
 271          * @return Status
 272          */
 273         final public function concatenate( array $params, array $opts = array() ) {
 274                 $params['op'] = 'concatenate';
 275                 return $this->doOperation( $params, $opts );
 276         }
 277
 278         /**
 279          * Prepare a storage path for usage. This will create containers
 280          * that don't yet exist or, on FS backends, create parent directories.
 281          *
 282          * $params include:
 283          *     dir : storage directory
 284          *
 285          * @param $params Array
 286          * @return Status
 287          */
 288         abstract public function prepare( array $params );
 289
 290         /**
 291          * Take measures to block web access to a directory and
 292          * the container it belongs to. FS backends might add .htaccess
 293          * files wheras backends like Swift this might restrict container
 294          * access to backend user that represents end-users in web request.
 295          * This is not guaranteed to actually do anything.
 296          *
 297          * $params include:
 298          *     dir       : storage directory
 299          *     noAccess  : try to deny file access
 300          *     noListing : try to deny file listing
 301          *
 302          * @param $params Array
 303          * @return Status
 304          */
 305         abstract public function secure( array $params );
 306
 307         /**
 308          * Clean up an empty storage directory.
 309          * On FS backends, the directory will be deleted. Others may do nothing.
 310          *
 311          * $params include:
 312          *     dir : storage directory
 313          *
 314          * @param $params Array
 315          * @return Status
 316          */
 317         abstract public function clean( array $params );
 318
 319         /**
 320          * Check if a file exists at a storage path in the backend.
 321          *
 322          * $params include:
 323          *     src    : source storage path
 324          *     latest : use the latest available data
 325          *
 326          * @param $params Array
 327          * @return bool|null Returns null on failure
 328          */
 329         abstract public function fileExists( array $params );
 330
 331         /**
 332          * Get the last-modified timestamp of the file at a storage path.
 333          *
 334          * $params include:
 335          *     src    : source storage path
 336          *     latest : use the latest available data
 337          *
 338          * @param $params Array
 339          * @return string|false TS_MW timestamp or false on failure
 340          */
 341         abstract public function getFileTimestamp( array $params );
 342
 343         /**
 344          * Get the contents of a file at a storage path in the backend.
 345          * This should be avoided for potentially large files.
 346          *
 347          * $params include:
 348          *     src    : source storage path
 349          *     latest : use the latest available data
 350          *
 351          * @param $params Array
 352          * @return string|false Returns false on failure
 353          */
 354         abstract public function getFileContents( array $params );
 355
 356         /**
 357          * Get a SHA-1 hash of the file at a storage path in the backend.
 358          *
 359          * $params include:
 360          *     src    : source storage path
 361          *     latest : use the latest available data
 362          *
 363          * @param $params Array
 364          * @return string|false Hash string or false on failure
 365          */
 366         abstract public function getFileSha1Base36( array $params );
 367
 368         /**
 369          * Get the properties of the file at a storage path in the backend.
 370          * Returns FSFile::placeholderProps() on failure.
 371          *
 372          * $params include:
 373          *     src    : source storage path
 374          *     latest : use the latest available data
 375          *
 376          * @param $params Array
 377          * @return Array
 378          */
 379         abstract public function getFileProps( array $params );
 380
 381         /**
 382          * Stream the file at a storage path in the backend.
 383          * Appropriate HTTP headers (Status, Content-Type, Content-Length)
 384          * must be sent if streaming began, while none should be sent otherwise.
 385          * Implementations should flush the output buffer before sending data.
 386          *
 387          * $params include:
 388          *     src     : source storage path
 389          *     headers : additional HTTP headers to send on success
 390          *     latest  : use the latest available data
 391          *
 392          * @param $params Array
 393          * @return Status
 394          */
 395         abstract public function streamFile( array $params );
 396
 397         /**
 398          * Returns a file system file, identical to the file at a storage path.
 399          * The file returned is either:
 400          * a) A local copy of the file at a storage path in the backend.
 401          *    The temporary copy will have the same extension as the source.
 402          * b) An original of the file at a storage path in the backend.
 403          * Temporary files may be purged when the file object falls out of scope.
 404          *
 405          * Write operations should *never* be done on this file as some backends
 406          * may do internal tracking or may be instances of FileBackendMultiWrite.
 407          * In that later case, there are copies of the file that must stay in sync.
 408          *
 409          * $params include:
 410          *     src    : source storage path
 411          *     latest : use the latest available data
 412          *
 413          * @param $params Array
 414          * @return FSFile|null Returns null on failure
 415          */
 416         abstract public function getLocalReference( array $params );
 417
 418         /**
 419          * Get a local copy on disk of the file at a storage path in the backend.
 420          * The temporary copy will have the same file extension as the source.
 421          * Temporary files may be purged when the file object falls out of scope.
 422          *
 423          * $params include:
 424          *     src    : source storage path
 425          *     latest : use the latest available data
 426          *
 427          * @param $params Array
 428          * @return TempFSFile|null Returns null on failure
 429          */
 430         abstract public function getLocalCopy( array $params );
 431
 432         /**
 433          * Get an iterator to list out all object files under a storage directory.
 434          * If the directory is of the form "mwstore://container", then all items in
 435          * the container should be listed. If of the form "mwstore://container/dir",
 436          * then all items under that container directory should be listed.
 437          * Results should be storage paths relative to the given directory.
 438          *
 439          * $params include:
 440          *     dir : storage path directory
 441          *
 442          * @return Traversable|Array|null Returns null on failure
 443          */
 444         abstract public function getFileList( array $params );
 445
 446         /**
 447          * Lock the files at the given storage paths in the backend.
 448          * This will either lock all the files or none (on failure).
 449          *
 450          * Callers should consider using getScopedFileLocks() instead.
 451          *
 452          * @param $paths Array Storage paths
 453          * @param $type integer LockManager::LOCK_* constant
 454          * @return Status
 455          */
 456         final public function lockFiles( array $paths, $type ) {
 457                 return $this->lockManager->lock( $paths, $type );
 458         }
 459
 460         /**
 461          * Unlock the files at the given storage paths in the backend.
 462          *
 463          * @param $paths Array Storage paths
 464          * @param $type integer LockManager::LOCK_* constant
 465          * @return Status
 466          */
 467         final public function unlockFiles( array $paths, $type ) {
 468                 return $this->lockManager->unlock( $paths, $type );
 469         }
 470
 471         /**
 472          * Lock the files at the given storage paths in the backend.
 473          * This will either lock all the files or none (on failure).
 474          * On failure, the status object will be updated with errors.
 475          *
 476          * Once the return value goes out scope, the locks will be released and
 477          * the status updated. Unlock fatals will not change the status "OK" value.
 478          *
 479          * @param $paths Array Storage paths
 480          * @param $type integer LockManager::LOCK_* constant
 481          * @param $status Status Status to update on lock/unlock
 482          * @return ScopedLock|null Returns null on failure
 483          */
 484         final public function getScopedFileLocks( array $paths, $type, Status $status ) {
 485                 return ScopedLock::factory( $this->lockManager, $paths, $type, $status );
 486         }
 487 }
 488
 489 /**
 490  * Base class for all single-write backends.
 491  * This class defines the methods as abstract that subclasses must implement.
 492  *
 493  * @ingroup FileBackend
 494  * @since 1.19
 495  */
 496 abstract class FileBackend extends FileBackendBase {
 497         /** @var Array */
 498         protected $cache = array(); // (storage path => key => value)
 499         protected $maxCacheSize = 50; // integer; max paths with entries
 500         /** @var Array */
 501         protected $shardViaHashLevels = array(); // (container name => integer)
 502
 503         /**
 504          * Create a file in the backend with the given contents.
 505          * Do not call this function from places outside FileBackend and FileOp.
 506          * $params include:
 507          *     content       : the raw file contents
 508          *     dst           : destination storage path
 509          *     overwriteDest : overwrite any file that exists at the destination
 510          *
 511          * @param $params Array
 512          * @return Status
 513          */
 514         final public function createInternal( array $params ) {
 515                 $status = $this->doCreateInternal( $params );
 516                 $this->clearCache( array( $params['dst'] ) );
 517                 return $status;
 518         }
 519
 520         /**
 521          * @see FileBackend::createInternal()
 522          */
 523         abstract protected function doCreateInternal( array $params );
 524
 525         /**
 526          * Store a file into the backend from a file on disk.
 527          * Do not call this function from places outside FileBackend and FileOp.
 528          * $params include:
 529          *     src           : source path on disk
 530          *     dst           : destination storage path
 531          *     overwriteDest : overwrite any file that exists at the destination
 532          *
 533          * @param $params Array
 534          * @return Status
 535          */
 536         final public function storeInternal( array $params ) {
 537                 $status = $this->doStoreInternal( $params );
 538                 $this->clearCache( array( $params['dst'] ) );
 539                 return $status;
 540         }
 541
 542         /**
 543          * @see FileBackend::storeInternal()
 544          */
 545         abstract protected function doStoreInternal( array $params );
 546
 547         /**
 548          * Copy a file from one storage path to another in the backend.
 549          * Do not call this function from places outside FileBackend and FileOp.
 550          * $params include:
 551          *     src           : source storage path
 552          *     dst           : destination storage path
 553          *     overwriteDest : overwrite any file that exists at the destination
 554          *
 555          * @param $params Array
 556          * @return Status
 557          */
 558         final public function copyInternal( array $params ) {
 559                 $status = $this->doCopyInternal( $params );
 560                 $this->clearCache( array( $params['dst'] ) );
 561                 return $status;
 562         }
 563
 564         /**
 565          * @see FileBackend::copyInternal()
 566          */
 567         abstract protected function doCopyInternal( array $params );
 568
 569         /**
 570          * Delete a file at the storage path.
 571          * Do not call this function from places outside FileBackend and FileOp.
 572          * $params include:
 573          *     src                 : source storage path
 574          *     ignoreMissingSource : do nothing if the source file does not exist
 575          *
 576          * @param $params Array
 577          * @return Status
 578          */
 579         final public function deleteInternal( array $params ) {
 580                 $status = $this->doDeleteInternal( $params );
 581                 $this->clearCache( array( $params['src'] ) );
 582                 return $status;
 583         }
 584
 585         /**
 586          * @see FileBackend::deleteInternal()
 587          */
 588         abstract protected function doDeleteInternal( array $params );
 589
 590         /**
 591          * Move a file from one storage path to another in the backend.
 592          * Do not call this function from places outside FileBackend and FileOp.
 593          * $params include:
 594          *     src           : source storage path
 595          *     dst           : destination storage path
 596          *     overwriteDest : overwrite any file that exists at the destination
 597          *
 598          * @param $params Array
 599          * @return Status
 600          */
 601         final public function moveInternal( array $params ) {
 602                 $status = $this->doMoveInternal( $params );
 603                 $this->clearCache( array( $params['src'], $params['dst'] ) );
 604                 return $status;
 605         }
 606
 607         /**
 608          * @see FileBackend::moveInternal()
 609          */
 610         protected function doMoveInternal( array $params ) {
 611                 // Copy source to dest
 612                 $status = $this->copyInternal( $params );
 613                 if ( !$status->isOK() ) {
 614                         return $status;
 615                 }
 616                 // Delete source (only fails due to races or medium going down)
 617                 $status->merge( $this->deleteInternal( array( 'src' => $params['src'] ) ) );
 618                 $status->setResult( true, $status->value ); // ignore delete() errors
 619                 return $status;
 620         }
 621
 622         /**
 623          * Combines files from several storage paths into a new file in the backend.
 624          * Do not call this function from places outside FileBackend and FileOp.
 625          * $params include:
 626          *     srcs          : ordered source storage paths (e.g. chunk1, chunk2, ...)
 627          *     dst           : file system path to 0-byte temp file
 628          *     overwriteDest : overwrite any file that exists at the destination
 629          *
 630          * @param $params Array
 631          * @return Status
 632          */
 633         final public function concatenateInternal( array $params ) {
 634                 $status = $this->doConcatenateInternal( $params );
 635                 return $status;
 636         }
 637
 638         /**
 639          * @see FileBackend::concatenateInternal()
 640          */
 641         protected function doConcatenateInternal( array $params ) {
 642                 $status = Status::newGood();
 643                 $tmpPath = $params['dst']; // convenience
 644
 645                 // Check that the specified temp file is valid...
 646                 wfSuppressWarnings();
 647                 $ok = ( is_file( $tmpPath ) && !filesize( $tmpPath ) );
 648                 wfRestoreWarnings();
 649                 if ( !$ok ) { // not present or not empty
 650                         $status->fatal( 'backend-fail-opentemp', $tmpPath );
 651                         return $status;
 652                 }
 653
 654                 // Build up the temp file using the source chunks (in order)...
 655                 $tmpHandle = fopen( $tmpPath, 'a' );
 656                 if ( $tmpHandle === false ) {
 657                         $status->fatal( 'backend-fail-opentemp', $tmpPath );
 658                         return $status;
 659                 }
 660                 foreach ( $params['srcs'] as $virtualSource ) {
 661                         // Get a local FS version of the chunk
 662                         $tmpFile = $this->getLocalReference( array( 'src' => $virtualSource ) );
 663                         if ( !$tmpFile ) {
 664                                 $status->fatal( 'backend-fail-read', $virtualSource );
 665                                 return $status;
 666                         }
 667                         // Get a handle to the local FS version
 668                         $sourceHandle = fopen( $tmpFile->getPath(), 'r' );
 669                         if ( $sourceHandle === false ) {
 670                                 fclose( $tmpHandle );
 671                                 $status->fatal( 'backend-fail-read', $virtualSource );
 672                                 return $status;
 673                         }
 674                         // Append chunk to file (pass chunk size to avoid magic quotes)
 675                         if ( !stream_copy_to_stream( $sourceHandle, $tmpHandle ) ) {
 676                                 fclose( $sourceHandle );
 677                                 fclose( $tmpHandle );
 678                                 $status->fatal( 'backend-fail-writetemp', $tmpPath );
 679                                 return $status;
 680                         }
 681                         fclose( $sourceHandle );
 682                 }
 683                 if ( !fclose( $tmpHandle ) ) {
 684                         $status->fatal( 'backend-fail-closetemp', $tmpPath );
 685                         return $status;
 686                 }
 687
 688                 return $status;
 689         }
 690
 691         /**
 692          * @see FileBackendBase::prepare()
 693          */
 694         final public function prepare( array $params ) {
 695                 $status = Status::newGood();
 696                 list( $fullCont, $dir, $shard ) = $this->resolveStoragePath( $params['dir'] );
 697                 if ( $dir === null ) {
 698                         $status->fatal( 'backend-fail-invalidpath', $params['dir'] );
 699                         return $status; // invalid storage path
 700                 }
 701                 if ( $shard !== null ) { // confined to a single container/shard
 702                         $status->merge( $this->doPrepare( $fullCont, $dir, $params ) );
 703                 } else { // directory is on several shards
 704                         wfDebug( __METHOD__ . ": iterating over all container shards.\n" );
 705                         list( $b, $shortCont, $r ) = self::splitStoragePath( $params['dir'] );
 706                         foreach ( $this->getContainerSuffixes( $shortCont ) as $suffix ) {
 707                                 $status->merge( $this->doPrepare( "{$fullCont}{$suffix}", $dir, $params ) );
 708                         }
 709                 }
 710                 return $status;
 711         }
 712
 713         /**
 714          * @see FileBackend::prepare()
 715          */
 716         protected function doPrepare( $container, $dir, array $params ) {
 717                 return Status::newGood();
 718         }
 719
 720         /**
 721          * @see FileBackendBase::secure()
 722          */
 723         final public function secure( array $params ) {
 724                 $status = Status::newGood();
 725                 list( $fullCont, $dir, $shard ) = $this->resolveStoragePath( $params['dir'] );
 726                 if ( $dir === null ) {
 727                         $status->fatal( 'backend-fail-invalidpath', $params['dir'] );
 728                         return $status; // invalid storage path
 729                 }
 730                 if ( $shard !== null ) { // confined to a single container/shard
 731                         $status->merge( $this->doSecure( $fullCont, $dir, $params ) );
 732                 } else { // directory is on several shards
 733                         wfDebug( __METHOD__ . ": iterating over all container shards.\n" );
 734                         list( $b, $shortCont, $r ) = self::splitStoragePath( $params['dir'] );
 735                         foreach ( $this->getContainerSuffixes( $shortCont ) as $suffix ) {
 736                                 $status->merge( $this->doSecure( "{$fullCont}{$suffix}", $dir, $params ) );
 737                         }
 738                 }
 739                 return $status;
 740         }
 741
 742         /**
 743          * @see FileBackend::secure()
 744          */
 745         protected function doSecure( $container, $dir, array $params ) {
 746                 return Status::newGood();
 747         }
 748
 749         /**
 750          * @see FileBackendBase::clean()
 751          */
 752         final public function clean( array $params ) {
 753                 $status = Status::newGood();
 754                 list( $fullCont, $dir, $shard ) = $this->resolveStoragePath( $params['dir'] );
 755                 if ( $dir === null ) {
 756                         $status->fatal( 'backend-fail-invalidpath', $params['dir'] );
 757                         return $status; // invalid storage path
 758                 }
 759                 if ( $shard !== null ) { // confined to a single container/shard
 760                         $status->merge( $this->doClean( $fullCont, $dir, $params ) );
 761                 } else { // directory is on several shards
 762                         wfDebug( __METHOD__ . ": iterating over all container shards.\n" );
 763                         list( $b, $shortCont, $r ) = self::splitStoragePath( $params['dir'] );
 764                         foreach ( $this->getContainerSuffixes( $shortCont ) as $suffix ) {
 765                                 $status->merge( $this->doClean( "{$fullCont}{$suffix}", $dir, $params ) );
 766                         }
 767                 }
 768                 return $status;
 769         }
 770
 771         /**
 772          * @see FileBackend::clean()
 773          */
 774         protected function doClean( $container, $dir, array $params ) {
 775                 return Status::newGood();
 776         }
 777
 778         /**
 779          * @see FileBackendBase::getFileContents()
 780          */
 781         public function getFileContents( array $params ) {
 782                 $tmpFile = $this->getLocalReference( $params );
 783                 if ( !$tmpFile ) {
 784                         return false;
 785                 }
 786                 wfSuppressWarnings();
 787                 $data = file_get_contents( $tmpFile->getPath() );
 788                 wfRestoreWarnings();
 789                 return $data;
 790         }
 791
 792         /**
 793          * @see FileBackendBase::getFileSha1Base36()
 794          */
 795         public function getFileSha1Base36( array $params ) {
 796                 $path = $params['src'];
 797                 if ( isset( $this->cache[$path]['sha1'] ) ) {
 798                         return $this->cache[$path]['sha1'];
 799                 }
 800                 $fsFile = $this->getLocalReference( $params );
 801                 if ( !$fsFile ) {
 802                         return false;
 803                 } else {
 804                         $sha1 = $fsFile->getSha1Base36();
 805                         if ( $sha1 !== false ) { // don't cache negatives
 806                                 $this->trimCache(); // limit memory
 807                                 $this->cache[$path]['sha1'] = $sha1;
 808                         }
 809                         return $sha1;
 810                 }
 811         }
 812
 813         /**
 814          * @see FileBackendBase::getFileProps()
 815          */
 816         public function getFileProps( array $params ) {
 817                 $fsFile = $this->getLocalReference( $params );
 818                 if ( !$fsFile ) {
 819                         return FSFile::placeholderProps();
 820                 } else {
 821                         return $fsFile->getProps();
 822                 }
 823         }
 824
 825         /**
 826          * @see FileBackendBase::getLocalReference()
 827          */
 828         public function getLocalReference( array $params ) {
 829                 return $this->getLocalCopy( $params );
 830         }
 831
 832         /**
 833          * @see FileBackendBase::streamFile()
 834          */
 835         public function streamFile( array $params ) {
 836                 $status = Status::newGood();
 837
 838                 $fsFile = $this->getLocalReference( $params );
 839                 if ( !$fsFile ) {
 840                         $status->fatal( 'backend-fail-stream', $params['src'] );
 841                         return $status;
 842                 }
 843
 844                 $extraHeaders = isset( $params['headers'] )
 845                         ? $params['headers']
 846                         : array();
 847
 848                 $ok = StreamFile::stream( $fsFile->getPath(), $extraHeaders, false );
 849                 if ( !$ok ) {
 850                         $status->fatal( 'backend-fail-stream', $params['src'] );
 851                         return $status;
 852                 }
 853
 854                 return $status;
 855         }
 856
 857         /**
 858          * @see FileBackendBase::getFileList()
 859          */
 860         final public function getFileList( array $params ) {
 861                 list( $fullCont, $dir, $shard ) = $this->resolveStoragePath( $params['dir'] );
 862                 if ( $dir === null ) { // invalid storage path
 863                         return null;
 864                 }
 865                 if ( $shard !== null ) {
 866                         // File listing is confined to a single container/shard
 867                         return $this->getFileListInternal( $fullCont, $dir, $params );
 868                 } else {
 869                         wfDebug( __METHOD__ . ": iterating over all container shards.\n" );
 870                         // File listing spans multiple containers/shards
 871                         list( $b, $shortCont, $r ) = self::splitStoragePath( $params['dir'] );
 872                         return new ContainerShardListIterator( $this,
 873                                 $fullCont, $this->getContainerSuffixes( $shortCont ), $params );
 874                 }
 875         }
 876
 877         /**
 878          * Do not call this function from places outside FileBackend and ContainerFileListIterator
 879          *
 880          * @param $container string Resolved container name
 881          * @param $dir string Resolved path relative to container
 882          * @param $params Array
 883          * @see FileBackend::getFileList()
 884          */
 885         abstract public function getFileListInternal( $container, $dir, array $params );
 886
 887         /**
 888          * Get the list of supported operations and their corresponding FileOp classes.
 889          *
 890          * @return Array
 891          */
 892         protected function supportedOperations() {
 893                 return array(
 894                         'store'       => 'StoreFileOp',
 895                         'copy'        => 'CopyFileOp',
 896                         'move'        => 'MoveFileOp',
 897                         'delete'      => 'DeleteFileOp',
 898                         'concatenate' => 'ConcatenateFileOp',
 899                         'create'      => 'CreateFileOp',
 900                         'null'        => 'NullFileOp'
 901                 );
 902         }
 903
 904         /**
 905          * Return a list of FileOp objects from a list of operations.
 906          * Do not call this function from places outside FileBackend.
 907          *
 908          * The result must have the same number of items as the input.
 909          * An exception is thrown if an unsupported operation is requested.
 910          *
 911          * @param $ops Array Same format as doOperations()
 912          * @return Array List of FileOp objects
 913          * @throws MWException
 914          */
 915         final public function getOperations( array $ops ) {
 916                 $supportedOps = $this->supportedOperations();
 917
 918                 $performOps = array(); // array of FileOp objects
 919                 // Build up ordered array of FileOps...
 920                 foreach ( $ops as $operation ) {
 921                         $opName = $operation['op'];
 922                         if ( isset( $supportedOps[$opName] ) ) {
 923                                 $class = $supportedOps[$opName];
 924                                 // Get params for this operation
 925                                 $params = $operation;
 926                                 // Append the FileOp class
 927                                 $performOps[] = new $class( $this, $params );
 928                         } else {
 929                                 throw new MWException( "Operation `$opName` is not supported." );
 930                         }
 931                 }
 932
 933                 return $performOps;
 934         }
 935
 936         /**
 937          * @see FileBackendBase::doOperationsInternal()
 938          */
 939         protected function doOperationsInternal( array $ops, array $opts ) {
 940                 $status = Status::newGood();
 941
 942                 // Build up a list of FileOps...
 943                 $performOps = $this->getOperations( $ops );
 944
 945                 // Acquire any locks as needed...
 946                 if ( empty( $opts['nonLocking'] ) ) {
 947                         // Build up a list of files to lock...
 948                         $filesLockEx = $filesLockSh = array();
 949                         foreach ( $performOps as $fileOp ) {
 950                                 $filesLockSh = array_merge( $filesLockSh, $fileOp->storagePathsRead() );
 951                                 $filesLockEx = array_merge( $filesLockEx, $fileOp->storagePathsChanged() );
 952                         }
 953                         // Optimization: if doing an EX lock anyway, don't also set an SH one
 954                         $filesLockSh = array_diff( $filesLockSh, $filesLockEx );
 955                         // Try to lock those files for the scope of this function...
 956                         $scopeLockS = $this->getScopedFileLocks( $filesLockSh, LockManager::LOCK_UW, $status );
 957                         $scopeLockE = $this->getScopedFileLocks( $filesLockEx, LockManager::LOCK_EX, $status );
 958                         if ( !$status->isOK() ) {
 959                                 return $status; // abort
 960                         }
 961                 }
 962
 963                 // Clear any cache entries (after locks acquired)
 964                 $this->clearCache();
 965                 // Actually attempt the operation batch...
 966                 $status->merge( FileOp::attemptBatch( $performOps, $opts ) );
 967
 968                 return $status;
 969         }
 970
 971         /**
 972          * Invalidate the file existence and property cache
 973          *
 974          * @param $paths Array Clear cache for specific files
 975          * @return void
 976          */
 977         final public function clearCache( array $paths = null ) {
 978                 if ( $paths === null ) {
 979                         $this->cache = array();
 980                 } else {
 981                         foreach ( $paths as $path ) {
 982                                 unset( $this->cache[$path] );
 983                         }
 984                 }
 985         }
 986
 987         /**
 988          * Prune the cache if it is too big to add an item
 989          *
 990          * @return void
 991          */
 992         protected function trimCache() {
 993                 if ( count( $this->cache ) >= $this->maxCacheSize ) {
 994                         reset( $this->cache );
 995                         $key = key( $this->cache );
 996                         unset( $this->cache[$key] );
 997                 }
 998         }
 999
1000         /**
1001          * Check if a given path is a mwstore:// path.
1002          * This does not do any actual validation or existence checks.
1003          *
1004          * @param $path string
1005          * @return bool
1006          */
1007         final public static function isStoragePath( $path ) {
1008                 return ( strpos( $path, 'mwstore://' ) === 0 );
1009         }
1010
1011         /**
1012          * Split a storage path (e.g. "mwstore://backend/container/path/to/object")
1013          * into a backend name, a container name, and a relative object path.
1014          *
1015          * @param $storagePath string
1016          * @return Array (backend, container, rel object) or (null, null, null)
1017          */
1018         final public static function splitStoragePath( $storagePath ) {
1019                 if ( self::isStoragePath( $storagePath ) ) {
1020                         // Note: strlen( 'mwstore://' ) = 10
1021                         $parts = explode( '/', substr( $storagePath, 10 ), 3 );
1022                         if ( count( $parts ) == 3 ) {
1023                                 return $parts; // e.g. "backend/container/path"
1024                         } elseif ( count( $parts ) == 2 ) {
1025                                 return array( $parts[0], $parts[1], '' ); // e.g. "backend/container"
1026                         }
1027                 }
1028                 return array( null, null, null );
1029         }
1030
1031         /**
1032          * Check if a container name is valid.
1033          * This checks for for length and illegal characters.
1034          *
1035          * @param $container string
1036          * @return bool
1037          */
1038         final protected static function isValidContainerName( $container ) {
1039                 // This accounts for Swift and S3 restrictions. Also note
1040                 // that these urlencode to the same string, which is useful
1041                 // since the Swift size limit is *after* URL encoding.
1042                 // Limit to 200 to leave room for '.shard-XX' or '.segment'.
1043                 return preg_match( '/^[a-zA-Z0-9._-]{1,200}$/u', $container );
1044         }
1045
1046         /**
1047          * Validate and normalize a relative storage path.
1048          * Null is returned if the path involves directory traversal.
1049          * Traversal is insecure for FS backends and broken for others.
1050          *
1051          * @param $path string
1052          * @return string|null
1053          */
1054         final protected static function normalizeStoragePath( $path ) {
1055                 // Normalize directory separators
1056                 $path = strtr( $path, '\\', '/' );
1057                 // Use the same traversal protection as Title::secureAndSplit()
1058                 if ( strpos( $path, '.' ) !== false ) {
1059                         if (
1060                                 $path === '.' ||
1061                                 $path === '..' ||
1062                                 strpos( $path, './' ) === 0 ||
1063                                 strpos( $path, '../' ) === 0 ||
1064                                 strpos( $path, '/./' ) !== false ||
1065                                 strpos( $path, '/../' ) !== false
1066                         ) {
1067                                 return null;
1068                         }
1069                 }
1070                 return $path;
1071         }
1072
1073         /**
1074          * Splits a storage path into an internal container name,
1075          * an internal relative object name, and a container shard suffix.
1076          * Any shard suffix is already appended to the internal container name.
1077          * This also checks that the storage path is valid and within this backend.
1078          *
1079          * If the container is sharded but a suffix could not be determined,
1080          * this means that the path can only refer to a directory and can only
1081          * be scanned by looking in all the container shards.
1082          *
1083          * @param $storagePath string
1084          * @return Array (container, path, container suffix) or (null, null, null) if invalid
1085          */
1086         final protected function resolveStoragePath( $storagePath ) {
1087                 list( $backend, $container, $relPath ) = self::splitStoragePath( $storagePath );
1088                 if ( $backend === $this->name ) { // must be for this backend
1089                         $relPath = self::normalizeStoragePath( $relPath );
1090                         if ( $relPath !== null ) {
1091                                 // Get shard for the normalized path if this container is sharded
1092                                 $cShard = $this->getContainerShard( $container, $relPath );
1093                                 // Validate and sanitize the relative path (backend-specific)
1094                                 $relPath = $this->resolveContainerPath( $container, $relPath );
1095                                 if ( $relPath !== null ) {
1096                                         // Prepend any wiki ID prefix to the container name
1097                                         $container = $this->fullContainerName( $container );
1098                                         if ( self::isValidContainerName( $container ) ) {
1099                                                 // Validate and sanitize the container name (backend-specific)
1100                                                 $container = $this->resolveContainerName( "{$container}{$cShard}" );
1101                                                 if ( $container !== null ) {
1102                                                         return array( $container, $relPath, $cShard );
1103                                                 }
1104                                         }
1105                                 }
1106                         }
1107                 }
1108                 return array( null, null, null );
1109         }
1110
1111         /**
1112          * Like resolveStoragePath() except null values are returned if
1113          * the container is sharded and the shard could not be determined.
1114          *
1115          * @see FileBackend::resolveStoragePath()
1116          *
1117          * @param $storagePath string
1118          * @return Array (container, path) or (null, null) if invalid
1119          */
1120         final protected function resolveStoragePathReal( $storagePath ) {
1121                 list( $container, $relPath, $cShard ) = $this->resolveStoragePath( $storagePath );
1122                 if ( $cShard !== null ) {
1123                         return array( $container, $relPath );
1124                 }
1125                 return array( null, null );
1126         }
1127
1128         /**
1129          * Get the container name shard suffix for a given path.
1130          * Any empty suffix means the container is not sharded.
1131          *
1132          * @param $container string Container name
1133          * @param $relStoragePath string Storage path relative to the container
1134          * @return string|null Returns null if shard could not be determined
1135          */
1136         final protected function getContainerShard( $container, $relPath ) {
1137                 $hashLevels = $this->getContainerHashLevels( $container );
1138                 if ( $hashLevels === 1 ) { // 16 shards per container
1139                         $hashDirRegex = '(?P<shard>[0-9a-f])';
1140                 } elseif ( $hashLevels === 2 ) { // 256 shards per container
1141                         $hashDirRegex = '[0-9a-f]/(?P<shard>[0-9a-f]{2})';
1142                 } else {
1143                         return ''; // no sharding
1144                 }
1145                 // Allow certain directories to be above the hash dirs so as
1146                 // to work with FileRepo (e.g. "archive/a/ab" or "temp/a/ab").
1147                 // They must be 2+ chars to avoid any hash directory ambiguity.
1148                 if ( preg_match( "!^(?:[^/]{2,}/)*$hashDirRegex(?:/|$)!", $relPath, $m ) ) {
1149                         return '.shard-' . str_pad( $m['shard'], $hashLevels, '0', STR_PAD_LEFT );
1150                 }
1151                 return null; // failed to match
1152         }
1153
1154         /**
1155          * Get the number of hash levels for a container.
1156          * If greater than 0, then all file storage paths within
1157          * the container are required to be hashed accordingly.
1158          *
1159          * @param $container string
1160          * @return integer
1161          */
1162         final protected function getContainerHashLevels( $container ) {
1163                 if ( isset( $this->shardViaHashLevels[$container] ) ) {
1164                         $hashLevels = (int)$this->shardViaHashLevels[$container];
1165                         if ( $hashLevels >= 0 && $hashLevels <= 2 ) {
1166                                 return $hashLevels;
1167                         }
1168                 }
1169                 return 0; // no sharding
1170         }
1171
1172         /**
1173          * Get a list of full container shard suffixes for a container
1174          *
1175          * @param $container string
1176          * @return Array
1177          */
1178         final protected function getContainerSuffixes( $container ) {
1179                 $shards = array();
1180                 $digits = $this->getContainerHashLevels( $container );
1181                 if ( $digits > 0 ) {
1182                         $numShards = 1 << ( $digits * 4 );
1183                         for ( $index = 0; $index < $numShards; $index++ ) {
1184                                 $shards[] = '.shard-' . str_pad( dechex( $index ), $digits, '0', STR_PAD_LEFT );
1185                         }
1186                 }
1187                 return $shards;
1188         }
1189
1190         /**
1191          * Get the full container name, including the wiki ID prefix
1192          *
1193          * @param $container string
1194          * @return string
1195          */
1196         final protected function fullContainerName( $container ) {
1197                 if ( $this->wikiId != '' ) {
1198                         return "{$this->wikiId}-$container";
1199                 } else {
1200                         return $container;
1201                 }
1202         }
1203
1204         /**
1205          * Resolve a container name, checking if it's allowed by the backend.
1206          * This is intended for internal use, such as encoding illegal chars.
1207          * Subclasses can override this to be more restrictive.
1208          *
1209          * @param $container string
1210          * @return string|null
1211          */
1212         protected function resolveContainerName( $container ) {
1213                 return $container;
1214         }
1215
1216         /**
1217          * Resolve a relative storage path, checking if it's allowed by the backend.
1218          * This is intended for internal use, such as encoding illegal chars or perhaps
1219          * getting absolute paths (e.g. FS based backends). Note that the relative path
1220          * may be the empty string (e.g. the path is simply to the container).
1221          *
1222          * @param $container string Container name
1223          * @param $relStoragePath string Storage path relative to the container
1224          * @return string|null Path or null if not valid
1225          */
1226         protected function resolveContainerPath( $container, $relStoragePath ) {
1227                 return $relStoragePath;
1228         }
1229
1230         /**
1231          * Get the final extension from a storage or FS path
1232          *
1233          * @param $path string
1234          * @return string
1235          */
1236         final public static function extensionFromPath( $path ) {
1237                 $i = strrpos( $path, '.' );
1238                 return strtolower( $i ? substr( $path, $i + 1 ) : '' );
1239         }
1240 }
1241
1242 /**
1243  * FileBackend helper function to handle file listings that span container shards.
1244  * Do not use this class from places outside of FileBackend.
1245  *
1246  * @ingroup FileBackend
1247  */
1248 class ContainerShardListIterator implements Iterator {
1249         /* @var FileBackend */
1250         protected $backend;
1251         /* @var Array */
1252         protected $params;
1253         /* @var Array */
1254         protected $shardSuffixes;
1255         protected $container; // string
1256         protected $directory; // string
1257
1258         /* @var Traversable */
1259         protected $iter;
1260         protected $curShard = 0; // integer
1261         protected $pos = 0; // integer
1262
1263         /**
1264          * @param $backend FileBackend
1265          * @param $container string Full storage container name
1266          * @param $dir string Storage directory relative to container
1267          * @param $suffixes Array List of container shard suffixes
1268          * @param $params Array
1269          */
1270         public function __construct(
1271                 FileBackend $backend, $container, $dir, array $suffixes, array $params
1272         ) {
1273                 $this->backend = $backend;
1274                 $this->container = $container;
1275                 $this->directory = $dir;
1276                 $this->shardSuffixes = $suffixes;
1277                 $this->params = $params;
1278         }
1279
1280         public function current() {
1281                 if ( is_array( $this->iter ) ) {
1282                         return current( $this->iter );
1283                 } else {
1284                         return $this->iter->current();
1285                 }
1286         }
1287
1288         public function key() {
1289                 return $this->pos;
1290         }
1291
1292         public function next() {
1293                 ++$this->pos;
1294                 if ( is_array( $this->iter ) ) {
1295                         next( $this->iter );
1296                 } else {
1297                         $this->iter->next();
1298                 }
1299                 // Find the next non-empty shard if no elements are left
1300                 $this->nextShardIteratorIfNotValid();
1301         }
1302
1303         /**
1304          * If the iterator for this container shard is out of items,
1305          * then move on to the next container that has items.
1306          */
1307         protected function nextShardIteratorIfNotValid() {
1308                 while ( !$this->valid() ) {
1309                         if ( ++$this->curShard >= count( $this->shardSuffixes ) ) {
1310                                 break; // no more container shards
1311                         }
1312                         $this->setIteratorFromCurrentShard();
1313                 }
1314         }
1315
1316         protected function setIteratorFromCurrentShard() {
1317                 $suffix = $this->shardSuffixes[$this->curShard];
1318                 $this->iter = $this->backend->getFileListInternal(
1319                         "{$this->container}{$suffix}", $this->directory, $this->params );
1320         }
1321
1322         public function rewind() {
1323                 $this->pos = 0;
1324                 $this->curShard = 0;
1325                 $this->setIteratorFromCurrentShard();
1326                 // Find the next non-empty shard if this one has no elements
1327                 $this->nextShardIteratorIfNotValid();
1328         }
1329
1330         public function valid() {
1331                 if ( $this->iter == null ) {
1332                         return false; // some failure?
1333                 } elseif ( is_array( $this->iter ) ) {
1334                         return ( current( $this->iter ) !== false ); // no paths can have this value
1335                 } else {
1336                         return $this->iter->valid();
1337                 }
1338         }
1339 }