includes/filebackend/FileBackend.php

   1 <?php
   2 /**
   3  * @defgroup FileBackend File backend
   4  * @ingroup  FileRepo
   5  *
   6  * File backend is used to interact with file storage systems,
   7  * such as the local file system, NFS, or cloud storage systems.
   8  */
   9
  10 /**
  11  * Base class for all file backends.
  12  *
  13  * This program is free software; you can redistribute it and/or modify
  14  * it under the terms of the GNU General Public License as published by
  15  * the Free Software Foundation; either version 2 of the License, or
  16  * (at your option) any later version.
  17  *
  18  * This program is distributed in the hope that it will be useful,
  19  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  20  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  21  * GNU General Public License for more details.
  22  *
  23  * You should have received a copy of the GNU General Public License along
  24  * with this program; if not, write to the Free Software Foundation, Inc.,
  25  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  26  * http://www.gnu.org/copyleft/gpl.html
  27  *
  28  * @file
  29  * @ingroup FileBackend
  30  * @author Aaron Schulz
  31  */
  32
  33 /**
  34  * @brief Base class for all file backend classes (including multi-write backends).
  35  *
  36  * This class defines the methods as abstract that subclasses must implement.
  37  * Outside callers can assume that all backends will have these functions.
  38  *
  39  * All "storage paths" are of the format "mwstore://<backend>/<container>/<path>".
  40  * The "<path>" portion is a relative path that uses UNIX file system (FS)
  41  * notation, though any particular backend may not actually be using a local
  42  * filesystem.
  43  * Therefore, the relative paths are only virtual.
  44  *
  45  * Backend contents are stored under wiki-specific container names by default.
  46  * For legacy reasons, this has no effect for the FS backend class, and per-wiki
  47  * segregation must be done by setting the container paths appropriately.
  48  *
  49  * FS-based backends are somewhat more restrictive due to the existence of real
  50  * directory files; a regular file cannot have the same name as a directory. Other
  51  * backends with virtual directories may not have this limitation. Callers should
  52  * store files in such a way that no files and directories are under the same path.
  53  *
  54  * Methods should avoid throwing exceptions at all costs.
  55  * As a corollary, external dependencies should be kept to a minimum.
  56  *
  57  * @ingroup FileBackend
  58  * @since 1.19
  59  */
  60 abstract class FileBackend {
  61         protected $name; // string; unique backend name
  62         protected $wikiId; // string; unique wiki name
  63         protected $readOnly; // string; read-only explanation message
  64         protected $parallelize; // string; when to do operations in parallel
  65         protected $concurrency; // integer; how many operations can be done in parallel
  66
  67         /** @var LockManager */
  68         protected $lockManager;
  69         /** @var FileJournal */
  70         protected $fileJournal;
  71
  72         /**
  73          * Create a new backend instance from configuration.
  74          * This should only be called from within FileBackendGroup.
  75          *
  76          * $config includes:
  77          *   - name        : The unique name of this backend.
  78          *                   This should consist of alphanumberic, '-', and '_' characters.
  79          *                   This name should not be changed after use.
  80          *   - wikiId      : Prefix to container names that is unique to this wiki.
  81          *                   It should only consist of alphanumberic, '-', and '_' characters.
  82          *   - lockManager : Registered name of a file lock manager to use.
  83          *   - fileJournal : File journal configuration; see FileJournal::factory().
  84          *                   Journals simply log changes to files stored in the backend.
  85          *   - readOnly    : Write operations are disallowed if this is a non-empty string.
  86          *                   It should be an explanation for the backend being read-only.
  87          *   - parallelize : When to do file operations in parallel (when possible).
  88          *                   Allowed values are "implicit", "explicit" and "off".
  89          *   - concurrency : How many file operations can be done in parallel.
  90          *
  91          * @param $config Array
  92          * @throws MWException
  93          */
  94         public function __construct( array $config ) {
  95                 $this->name = $config['name'];
  96                 if ( !preg_match( '!^[a-zA-Z0-9-_]{1,255}$!', $this->name ) ) {
  97                         throw new MWException( "Backend name `{$this->name}` is invalid." );
  98                 }
  99                 $this->wikiId = isset( $config['wikiId'] )
 100                         ? $config['wikiId']
 101                         : wfWikiID(); // e.g. "my_wiki-en_"
 102                 $this->lockManager = ( $config['lockManager'] instanceof LockManager )
 103                         ? $config['lockManager']
 104                         : LockManagerGroup::singleton()->get( $config['lockManager'] );
 105                 $this->fileJournal = isset( $config['fileJournal'] )
 106                         ? ( ( $config['fileJournal'] instanceof FileJournal )
 107                                 ? $config['fileJournal']
 108                                 : FileJournal::factory( $config['fileJournal'], $this->name ) )
 109                         : FileJournal::factory( array( 'class' => 'NullFileJournal' ), $this->name );
 110                 $this->readOnly = isset( $config['readOnly'] )
 111                         ? (string)$config['readOnly']
 112                         : '';
 113                 $this->parallelize = isset( $config['parallelize'] )
 114                         ? (string)$config['parallelize']
 115                         : 'off';
 116                 $this->concurrency = isset( $config['concurrency'] )
 117                         ? (int)$config['concurrency']
 118                         : 50;
 119         }
 120
 121         /**
 122          * Get the unique backend name.
 123          * We may have multiple different backends of the same type.
 124          * For example, we can have two Swift backends using different proxies.
 125          *
 126          * @return string
 127          */
 128         final public function getName() {
 129                 return $this->name;
 130         }
 131
 132         /**
 133          * Check if this backend is read-only
 134          *
 135          * @return bool
 136          */
 137         final public function isReadOnly() {
 138                 return ( $this->readOnly != '' );
 139         }
 140
 141         /**
 142          * Get an explanatory message if this backend is read-only
 143          *
 144          * @return string|bool Returns false if the backend is not read-only
 145          */
 146         final public function getReadOnlyReason() {
 147                 return ( $this->readOnly != '' ) ? $this->readOnly : false;
 148         }
 149
 150         /**
 151          * This is the main entry point into the backend for write operations.
 152          * Callers supply an ordered list of operations to perform as a transaction.
 153          * Files will be locked, the stat cache cleared, and then the operations attempted.
 154          * If any serious errors occur, all attempted operations will be rolled back.
 155          *
 156          * $ops is an array of arrays. The outer array holds a list of operations.
 157          * Each inner array is a set of key value pairs that specify an operation.
 158          *
 159          * Supported operations and their parameters. The supported actions are:
 160          *  - create
 161          *  - store
 162          *  - copy
 163          *  - move
 164          *  - delete
 165          *  - null
 166          *
 167          * a) Create a new file in storage with the contents of a string
 168          * @code
 169          *     array(
 170          *         'op'                  => 'create',
 171          *         'dst'                 => <storage path>,
 172          *         'content'             => <string of new file contents>,
 173          *         'overwrite'           => <boolean>,
 174          *         'overwriteSame'       => <boolean>
 175          *     );
 176          * @endcode
 177          *
 178          * b) Copy a file system file into storage
 179          * @code
 180          *     array(
 181          *         'op'                  => 'store',
 182          *         'src'                 => <file system path>,
 183          *         'dst'                 => <storage path>,
 184          *         'overwrite'           => <boolean>,
 185          *         'overwriteSame'       => <boolean>
 186          *     )
 187          * @endcode
 188          *
 189          * c) Copy a file within storage
 190          * @code
 191          *     array(
 192          *         'op'                  => 'copy',
 193          *         'src'                 => <storage path>,
 194          *         'dst'                 => <storage path>,
 195          *         'overwrite'           => <boolean>,
 196          *         'overwriteSame'       => <boolean>
 197          *     )
 198          * @endcode
 199          *
 200          * d) Move a file within storage
 201          * @code
 202          *     array(
 203          *         'op'                  => 'move',
 204          *         'src'                 => <storage path>,
 205          *         'dst'                 => <storage path>,
 206          *         'overwrite'           => <boolean>,
 207          *         'overwriteSame'       => <boolean>
 208          *     )
 209          * @endcode
 210          *
 211          * e) Delete a file within storage
 212          * @code
 213          *     array(
 214          *         'op'                  => 'delete',
 215          *         'src'                 => <storage path>,
 216          *         'ignoreMissingSource' => <boolean>
 217          *     )
 218          * @endcode
 219          *
 220          * f) Do nothing (no-op)
 221          * @code
 222          *     array(
 223          *         'op'                  => 'null',
 224          *     )
 225          * @endcode
 226          *
 227          * Boolean flags for operations (operation-specific):
 228          *   - ignoreMissingSource : The operation will simply succeed and do
 229          *                           nothing if the source file does not exist.
 230          *   - overwrite           : Any destination file will be overwritten.
 231          *   - overwriteSame       : An error will not be given if a file already
 232          *                           exists at the destination that has the same
 233          *                           contents as the new contents to be written there.
 234          *
 235          * $opts is an associative of boolean flags, including:
 236          *   - force               : Operation precondition errors no longer trigger an abort.
 237          *                           Any remaining operations are still attempted. Unexpected
 238          *                           failures may still cause remaning operations to be aborted.
 239          *   - nonLocking          : No locks are acquired for the operations.
 240          *                           This can increase performance for non-critical writes.
 241          *                           This has no effect unless the 'force' flag is set.
 242          *   - allowStale          : Don't require the latest available data.
 243          *                           This can increase performance for non-critical writes.
 244          *                           This has no effect unless the 'force' flag is set.
 245          *   - preserveCache       : Don't clear the process cache before checking files.
 246          *                           This should only be used if all entries in the process
 247          *                           cache were added after the files were already locked.
 248          *   - nonJournaled        : Don't log this operation batch in the file journal.
 249          *                           This limits the ability of recovery scripts.
 250          *   - parallelize         : Try to do operations in parallel when possible.
 251          *   - bypassReadOnly      : Allow writes in read-only mode (since 1.20).
 252          *
 253          * @remarks Remarks on locking:
 254          * File system paths given to operations should refer to files that are
 255          * already locked or otherwise safe from modification from other processes.
 256          * Normally these files will be new temp files, which should be adequate.
 257          *
 258          * @par Return value:
 259          *
 260          * This returns a Status, which contains all warnings and fatals that occurred
 261          * during the operation. The 'failCount', 'successCount', and 'success' members
 262          * will reflect each operation attempted.
 263          *
 264          * The status will be "OK" unless:
 265          *   - a) unexpected operation errors occurred (network partitions, disk full...)
 266          *   - b) significant operation errors occurred and 'force' was not set
 267          *
 268          * @param $ops Array List of operations to execute in order
 269          * @param $opts Array Batch operation options
 270          * @return Status
 271          */
 272         final public function doOperations( array $ops, array $opts = array() ) {
 273                 if ( empty( $opts['bypassReadOnly'] ) && $this->isReadOnly() ) {
 274                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 275                 }
 276                 if ( empty( $opts['force'] ) ) { // sanity
 277                         unset( $opts['nonLocking'] );
 278                         unset( $opts['allowStale'] );
 279                 }
 280                 $opts['concurrency'] = 1; // off
 281                 if ( $this->parallelize === 'implicit' ) {
 282                         if ( !isset( $opts['parallelize'] ) || $opts['parallelize'] ) {
 283                                 $opts['concurrency'] = $this->concurrency;
 284                         }
 285                 } elseif ( $this->parallelize === 'explicit' ) {
 286                         if ( !empty( $opts['parallelize'] ) ) {
 287                                 $opts['concurrency'] = $this->concurrency;
 288                         }
 289                 }
 290                 return $this->doOperationsInternal( $ops, $opts );
 291         }
 292
 293         /**
 294          * @see FileBackend::doOperations()
 295          */
 296         abstract protected function doOperationsInternal( array $ops, array $opts );
 297
 298         /**
 299          * Same as doOperations() except it takes a single operation.
 300          * If you are doing a batch of operations that should either
 301          * all succeed or all fail, then use that function instead.
 302          *
 303          * @see FileBackend::doOperations()
 304          *
 305          * @param $op Array Operation
 306          * @param $opts Array Operation options
 307          * @return Status
 308          */
 309         final public function doOperation( array $op, array $opts = array() ) {
 310                 return $this->doOperations( array( $op ), $opts );
 311         }
 312
 313         /**
 314          * Performs a single create operation.
 315          * This sets $params['op'] to 'create' and passes it to doOperation().
 316          *
 317          * @see FileBackend::doOperation()
 318          *
 319          * @param $params Array Operation parameters
 320          * @param $opts Array Operation options
 321          * @return Status
 322          */
 323         final public function create( array $params, array $opts = array() ) {
 324                 return $this->doOperation( array( 'op' => 'create' ) + $params, $opts );
 325         }
 326
 327         /**
 328          * Performs a single store operation.
 329          * This sets $params['op'] to 'store' and passes it to doOperation().
 330          *
 331          * @see FileBackend::doOperation()
 332          *
 333          * @param $params Array Operation parameters
 334          * @param $opts Array Operation options
 335          * @return Status
 336          */
 337         final public function store( array $params, array $opts = array() ) {
 338                 return $this->doOperation( array( 'op' => 'store' ) + $params, $opts );
 339         }
 340
 341         /**
 342          * Performs a single copy operation.
 343          * This sets $params['op'] to 'copy' and passes it to doOperation().
 344          *
 345          * @see FileBackend::doOperation()
 346          *
 347          * @param $params Array Operation parameters
 348          * @param $opts Array Operation options
 349          * @return Status
 350          */
 351         final public function copy( array $params, array $opts = array() ) {
 352                 return $this->doOperation( array( 'op' => 'copy' ) + $params, $opts );
 353         }
 354
 355         /**
 356          * Performs a single move operation.
 357          * This sets $params['op'] to 'move' and passes it to doOperation().
 358          *
 359          * @see FileBackend::doOperation()
 360          *
 361          * @param $params Array Operation parameters
 362          * @param $opts Array Operation options
 363          * @return Status
 364          */
 365         final public function move( array $params, array $opts = array() ) {
 366                 return $this->doOperation( array( 'op' => 'move' ) + $params, $opts );
 367         }
 368
 369         /**
 370          * Performs a single delete operation.
 371          * This sets $params['op'] to 'delete' and passes it to doOperation().
 372          *
 373          * @see FileBackend::doOperation()
 374          *
 375          * @param $params Array Operation parameters
 376          * @param $opts Array Operation options
 377          * @return Status
 378          */
 379         final public function delete( array $params, array $opts = array() ) {
 380                 return $this->doOperation( array( 'op' => 'delete' ) + $params, $opts );
 381         }
 382
 383         /**
 384          * Perform a set of independent file operations on some files.
 385          *
 386          * This does no locking, nor journaling, and possibly no stat calls.
 387          * Any destination files that already exist will be overwritten.
 388          * This should *only* be used on non-original files, like cache files.
 389          *
 390          * Supported operations and their parameters:
 391          *  - create
 392          *  - store
 393          *  - copy
 394          *  - move
 395          *  - delete
 396          *  - null
 397          *
 398          * a) Create a new file in storage with the contents of a string
 399          * @code
 400          *     array(
 401          *         'op'                  => 'create',
 402          *         'dst'                 => <storage path>,
 403          *         'content'             => <string of new file contents>
 404          *     )
 405          * @endcode
 406          * b) Copy a file system file into storage
 407          * @code
 408          *     array(
 409          *         'op'                  => 'store',
 410          *         'src'                 => <file system path>,
 411          *         'dst'                 => <storage path>
 412          *     )
 413          * @endcode
 414          * c) Copy a file within storage
 415          * @code
 416          *     array(
 417          *         'op'                  => 'copy',
 418          *         'src'                 => <storage path>,
 419          *         'dst'                 => <storage path>
 420          *     )
 421          * @endcode
 422          * d) Move a file within storage
 423          * @code
 424          *     array(
 425          *         'op'                  => 'move',
 426          *         'src'                 => <storage path>,
 427          *         'dst'                 => <storage path>
 428          *     )
 429          * @endcode
 430          * e) Delete a file within storage
 431          * @code
 432          *     array(
 433          *         'op'                  => 'delete',
 434          *         'src'                 => <storage path>,
 435          *         'ignoreMissingSource' => <boolean>
 436          *     )
 437          * @endcode
 438          * f) Do nothing (no-op)
 439          * @code
 440          *     array(
 441          *         'op'                  => 'null',
 442          *     )
 443          * @endcode
 444          *
 445          * @par Boolean flags for operations (operation-specific):
 446          *   - ignoreMissingSource : The operation will simply succeed and do
 447          *                           nothing if the source file does not exist.
 448          *
 449          * $opts is an associative of boolean flags, including:
 450          *   - bypassReadOnly      : Allow writes in read-only mode (since 1.20)
 451          *
 452          * @par Return value:
 453          * This returns a Status, which contains all warnings and fatals that occurred
 454          * during the operation. The 'failCount', 'successCount', and 'success' members
 455          * will reflect each operation attempted for the given files. The status will be
 456          * considered "OK" as long as no fatal errors occurred.
 457          *
 458          * @param $ops Array Set of operations to execute
 459          * @param $opts Array Batch operation options
 460          * @return Status
 461          * @since 1.20
 462          */
 463         final public function doQuickOperations( array $ops, array $opts = array() ) {
 464                 if ( empty( $opts['bypassReadOnly'] ) && $this->isReadOnly() ) {
 465                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 466                 }
 467                 foreach ( $ops as &$op ) {
 468                         $op['overwrite'] = true; // avoids RTTs in key/value stores
 469                 }
 470                 return $this->doQuickOperationsInternal( $ops );
 471         }
 472
 473         /**
 474          * @see FileBackend::doQuickOperations()
 475          * @since 1.20
 476          */
 477         abstract protected function doQuickOperationsInternal( array $ops );
 478
 479         /**
 480          * Same as doQuickOperations() except it takes a single operation.
 481          * If you are doing a batch of operations, then use that function instead.
 482          *
 483          * @see FileBackend::doQuickOperations()
 484          *
 485          * @param $op Array Operation
 486          * @return Status
 487          * @since 1.20
 488          */
 489         final public function doQuickOperation( array $op ) {
 490                 return $this->doQuickOperations( array( $op ) );
 491         }
 492
 493         /**
 494          * Performs a single quick create operation.
 495          * This sets $params['op'] to 'create' and passes it to doQuickOperation().
 496          *
 497          * @see FileBackend::doQuickOperation()
 498          *
 499          * @param $params Array Operation parameters
 500          * @return Status
 501          * @since 1.20
 502          */
 503         final public function quickCreate( array $params ) {
 504                 return $this->doQuickOperation( array( 'op' => 'create' ) + $params );
 505         }
 506
 507         /**
 508          * Performs a single quick store operation.
 509          * This sets $params['op'] to 'store' and passes it to doQuickOperation().
 510          *
 511          * @see FileBackend::doQuickOperation()
 512          *
 513          * @param $params Array Operation parameters
 514          * @return Status
 515          * @since 1.20
 516          */
 517         final public function quickStore( array $params ) {
 518                 return $this->doQuickOperation( array( 'op' => 'store' ) + $params );
 519         }
 520
 521         /**
 522          * Performs a single quick copy operation.
 523          * This sets $params['op'] to 'copy' and passes it to doQuickOperation().
 524          *
 525          * @see FileBackend::doQuickOperation()
 526          *
 527          * @param $params Array Operation parameters
 528          * @return Status
 529          * @since 1.20
 530          */
 531         final public function quickCopy( array $params ) {
 532                 return $this->doQuickOperation( array( 'op' => 'copy' ) + $params );
 533         }
 534
 535         /**
 536          * Performs a single quick move operation.
 537          * This sets $params['op'] to 'move' and passes it to doQuickOperation().
 538          *
 539          * @see FileBackend::doQuickOperation()
 540          *
 541          * @param $params Array Operation parameters
 542          * @return Status
 543          * @since 1.20
 544          */
 545         final public function quickMove( array $params ) {
 546                 return $this->doQuickOperation( array( 'op' => 'move' ) + $params );
 547         }
 548
 549         /**
 550          * Performs a single quick delete operation.
 551          * This sets $params['op'] to 'delete' and passes it to doQuickOperation().
 552          *
 553          * @see FileBackend::doQuickOperation()
 554          *
 555          * @param $params Array Operation parameters
 556          * @return Status
 557          * @since 1.20
 558          */
 559         final public function quickDelete( array $params ) {
 560                 return $this->doQuickOperation( array( 'op' => 'delete' ) + $params );
 561         }
 562
 563         /**
 564          * Concatenate a list of storage files into a single file system file.
 565          * The target path should refer to a file that is already locked or
 566          * otherwise safe from modification from other processes. Normally,
 567          * the file will be a new temp file, which should be adequate.
 568          *
 569          * @param $params Array Operation parameters
 570          * $params include:
 571          *   - srcs : ordered source storage paths (e.g. chunk1, chunk2, ...)
 572          *   - dst  : file system path to 0-byte temp file
 573          * @return Status
 574          */
 575         abstract public function concatenate( array $params );
 576
 577         /**
 578          * Prepare a storage directory for usage.
 579          * This will create any required containers and parent directories.
 580          * Backends using key/value stores only need to create the container.
 581          *
 582          * The 'noAccess' and 'noListing' parameters works the same as in secure(),
 583          * except they are only applied *if* the directory/container had to be created.
 584          * These flags should always be set for directories that have private files.
 585          *
 586          * @param $params Array
 587          * $params include:
 588          *   - dir            : storage directory
 589          *   - noAccess       : try to deny file access (since 1.20)
 590          *   - noListing      : try to deny file listing (since 1.20)
 591          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 592          * @return Status
 593          */
 594         final public function prepare( array $params ) {
 595                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 596                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 597                 }
 598                 return $this->doPrepare( $params );
 599         }
 600
 601         /**
 602          * @see FileBackend::prepare()
 603          */
 604         abstract protected function doPrepare( array $params );
 605
 606         /**
 607          * Take measures to block web access to a storage directory and
 608          * the container it belongs to. FS backends might add .htaccess
 609          * files whereas key/value store backends might revoke container
 610          * access to the storage user representing end-users in web requests.
 611          * This is not guaranteed to actually do anything.
 612          *
 613          * @param $params Array
 614          * $params include:
 615          *   - dir            : storage directory
 616          *   - noAccess       : try to deny file access
 617          *   - noListing      : try to deny file listing
 618          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 619          * @return Status
 620          */
 621         final public function secure( array $params ) {
 622                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 623                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 624                 }
 625                 return $this->doSecure( $params );
 626         }
 627
 628         /**
 629          * @see FileBackend::secure()
 630          */
 631         abstract protected function doSecure( array $params );
 632
 633         /**
 634          * Remove measures to block web access to a storage directory and
 635          * the container it belongs to. FS backends might remove .htaccess
 636          * files whereas key/value store backends might grant container
 637          * access to the storage user representing end-users in web requests.
 638          * This essentially can undo the result of secure() calls.
 639          *
 640          * @param $params Array
 641          * $params include:
 642          *   - dir            : storage directory
 643          *   - access         : try to allow file access
 644          *   - listing        : try to allow file listing
 645          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 646          * @return Status
 647          * @since 1.20
 648          */
 649         final public function publish( array $params ) {
 650                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 651                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 652                 }
 653                 return $this->doPublish( $params );
 654         }
 655
 656         /**
 657          * @see FileBackend::publish()
 658          */
 659         abstract protected function doPublish( array $params );
 660
 661         /**
 662          * Delete a storage directory if it is empty.
 663          * Backends using key/value stores may do nothing unless the directory
 664          * is that of an empty container, in which case it should be deleted.
 665          *
 666          * @param $params Array
 667          * $params include:
 668          *   - dir            : storage directory
 669          *   - recursive      : recursively delete empty subdirectories first (since 1.20)
 670          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 671          * @return Status
 672          */
 673         final public function clean( array $params ) {
 674                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 675                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 676                 }
 677                 return $this->doClean( $params );
 678         }
 679
 680         /**
 681          * @see FileBackend::clean()
 682          */
 683         abstract protected function doClean( array $params );
 684
 685         /**
 686          * Check if a file exists at a storage path in the backend.
 687          * This returns false if only a directory exists at the path.
 688          *
 689          * @param $params Array
 690          * $params include:
 691          *   - src    : source storage path
 692          *   - latest : use the latest available data
 693          * @return bool|null Returns null on failure
 694          */
 695         abstract public function fileExists( array $params );
 696
 697         /**
 698          * Get the last-modified timestamp of the file at a storage path.
 699          *
 700          * @param $params Array
 701          * $params include:
 702          *   - src    : source storage path
 703          *   - latest : use the latest available data
 704          * @return string|bool TS_MW timestamp or false on failure
 705          */
 706         abstract public function getFileTimestamp( array $params );
 707
 708         /**
 709          * Get the contents of a file at a storage path in the backend.
 710          * This should be avoided for potentially large files.
 711          *
 712          * @param $params Array
 713          * $params include:
 714          *   - src    : source storage path
 715          *   - latest : use the latest available data
 716          * @return string|bool Returns false on failure
 717          */
 718         abstract public function getFileContents( array $params );
 719
 720         /**
 721          * Get the size (bytes) of a file at a storage path in the backend.
 722          *
 723          * @param $params Array
 724          * $params include:
 725          *   - src    : source storage path
 726          *   - latest : use the latest available data
 727          * @return integer|bool Returns false on failure
 728          */
 729         abstract public function getFileSize( array $params );
 730
 731         /**
 732          * Get quick information about a file at a storage path in the backend.
 733          * If the file does not exist, then this returns false.
 734          * Otherwise, the result is an associative array that includes:
 735          *   - mtime  : the last-modified timestamp (TS_MW)
 736          *   - size   : the file size (bytes)
 737          * Additional values may be included for internal use only.
 738          *
 739          * @param $params Array
 740          * $params include:
 741          *   - src    : source storage path
 742          *   - latest : use the latest available data
 743          * @return Array|bool|null Returns null on failure
 744          */
 745         abstract public function getFileStat( array $params );
 746
 747         /**
 748          * Get a SHA-1 hash of the file at a storage path in the backend.
 749          *
 750          * @param $params Array
 751          * $params include:
 752          *   - src    : source storage path
 753          *   - latest : use the latest available data
 754          * @return string|bool Hash string or false on failure
 755          */
 756         abstract public function getFileSha1Base36( array $params );
 757
 758         /**
 759          * Get the properties of the file at a storage path in the backend.
 760          * Returns FSFile::placeholderProps() on failure.
 761          *
 762          * @param $params Array
 763          * $params include:
 764          *   - src    : source storage path
 765          *   - latest : use the latest available data
 766          * @return Array
 767          */
 768         abstract public function getFileProps( array $params );
 769
 770         /**
 771          * Stream the file at a storage path in the backend.
 772          * If the file does not exists, a 404 error will be given.
 773          * Appropriate HTTP headers (Status, Content-Type, Content-Length)
 774          * must be sent if streaming began, while none should be sent otherwise.
 775          * Implementations should flush the output buffer before sending data.
 776          *
 777          * @param $params Array
 778          * $params include:
 779          *   - src     : source storage path
 780          *   - headers : additional HTTP headers to send on success
 781          *   - latest  : use the latest available data
 782          * @return Status
 783          */
 784         abstract public function streamFile( array $params );
 785
 786         /**
 787          * Returns a file system file, identical to the file at a storage path.
 788          * The file returned is either:
 789          *   - a) A local copy of the file at a storage path in the backend.
 790          *        The temporary copy will have the same extension as the source.
 791          *   - b) An original of the file at a storage path in the backend.
 792          * Temporary files may be purged when the file object falls out of scope.
 793          *
 794          * Write operations should *never* be done on this file as some backends
 795          * may do internal tracking or may be instances of FileBackendMultiWrite.
 796          * In that later case, there are copies of the file that must stay in sync.
 797          * Additionally, further calls to this function may return the same file.
 798          *
 799          * @param $params Array
 800          * $params include:
 801          *   - src    : source storage path
 802          *   - latest : use the latest available data
 803          * @return FSFile|null Returns null on failure
 804          */
 805         abstract public function getLocalReference( array $params );
 806
 807         /**
 808          * Get a local copy on disk of the file at a storage path in the backend.
 809          * The temporary copy will have the same file extension as the source.
 810          * Temporary files may be purged when the file object falls out of scope.
 811          *
 812          * @param $params Array
 813          * $params include:
 814          *   - src    : source storage path
 815          *   - latest : use the latest available data
 816          * @return TempFSFile|null Returns null on failure
 817          */
 818         abstract public function getLocalCopy( array $params );
 819
 820         /**
 821          * Check if a directory exists at a given storage path.
 822          * Backends using key/value stores will check if the path is a
 823          * virtual directory, meaning there are files under the given directory.
 824          *
 825          * Storage backends with eventual consistency might return stale data.
 826          *
 827          * @param $params array
 828          * $params include:
 829          *   - dir : storage directory
 830          * @return bool|null Returns null on failure
 831          * @since 1.20
 832          */
 833         abstract public function directoryExists( array $params );
 834
 835         /**
 836          * Get an iterator to list *all* directories under a storage directory.
 837          * If the directory is of the form "mwstore://backend/container",
 838          * then all directories in the container should be listed.
 839          * If the directory is of form "mwstore://backend/container/dir",
 840          * then all directories directly under that directory should be listed.
 841          * Results should be storage directories relative to the given directory.
 842          *
 843          * Storage backends with eventual consistency might return stale data.
 844          *
 845          * @param $params array
 846          * $params include:
 847          *   - dir     : storage directory
 848          *   - topOnly : only return direct child dirs of the directory
 849          * @return Traversable|Array|null Returns null on failure
 850          * @since 1.20
 851          */
 852         abstract public function getDirectoryList( array $params );
 853
 854         /**
 855          * Same as FileBackend::getDirectoryList() except only lists
 856          * directories that are immediately under the given directory.
 857          *
 858          * Storage backends with eventual consistency might return stale data.
 859          *
 860          * @param $params array
 861          * $params include:
 862          *   - dir : storage directory
 863          * @return Traversable|Array|null Returns null on failure
 864          * @since 1.20
 865          */
 866         final public function getTopDirectoryList( array $params ) {
 867                 return $this->getDirectoryList( array( 'topOnly' => true ) + $params );
 868         }
 869
 870         /**
 871          * Get an iterator to list *all* stored files under a storage directory.
 872          * If the directory is of the form "mwstore://backend/container",
 873          * then all files in the container should be listed.
 874          * If the directory is of form "mwstore://backend/container/dir",
 875          * then all files under that directory should be listed.
 876          * Results should be storage paths relative to the given directory.
 877          *
 878          * Storage backends with eventual consistency might return stale data.
 879          *
 880          * @param $params array
 881          * $params include:
 882          *   - dir     : storage directory
 883          *   - topOnly : only return direct child files of the directory (since 1.20)
 884          * @return Traversable|Array|null Returns null on failure
 885          */
 886         abstract public function getFileList( array $params );
 887
 888         /**
 889          * Same as FileBackend::getFileList() except only lists
 890          * files that are immediately under the given directory.
 891          *
 892          * Storage backends with eventual consistency might return stale data.
 893          *
 894          * @param $params array
 895          * $params include:
 896          *   - dir : storage directory
 897          * @return Traversable|Array|null Returns null on failure
 898          * @since 1.20
 899          */
 900         final public function getTopFileList( array $params ) {
 901                 return $this->getFileList( array( 'topOnly' => true ) + $params );
 902         }
 903
 904         /**
 905          * Preload persistent file stat and property cache into in-process cache.
 906          * This should be used when stat calls will be made on a known list of a many files.
 907          *
 908          * @param $paths Array Storage paths
 909          * @return void
 910          */
 911         public function preloadCache( array $paths ) {}
 912
 913         /**
 914          * Invalidate any in-process file stat and property cache.
 915          * If $paths is given, then only the cache for those files will be cleared.
 916          *
 917          * @param $paths Array Storage paths (optional)
 918          * @return void
 919          */
 920         public function clearCache( array $paths = null ) {}
 921
 922         /**
 923          * Lock the files at the given storage paths in the backend.
 924          * This will either lock all the files or none (on failure).
 925          *
 926          * Callers should consider using getScopedFileLocks() instead.
 927          *
 928          * @param $paths Array Storage paths
 929          * @param $type integer LockManager::LOCK_* constant
 930          * @return Status
 931          */
 932         final public function lockFiles( array $paths, $type ) {
 933                 return $this->lockManager->lock( $paths, $type );
 934         }
 935
 936         /**
 937          * Unlock the files at the given storage paths in the backend.
 938          *
 939          * @param $paths Array Storage paths
 940          * @param $type integer LockManager::LOCK_* constant
 941          * @return Status
 942          */
 943         final public function unlockFiles( array $paths, $type ) {
 944                 return $this->lockManager->unlock( $paths, $type );
 945         }
 946
 947         /**
 948          * Lock the files at the given storage paths in the backend.
 949          * This will either lock all the files or none (on failure).
 950          * On failure, the status object will be updated with errors.
 951          *
 952          * Once the return value goes out scope, the locks will be released and
 953          * the status updated. Unlock fatals will not change the status "OK" value.
 954          *
 955          * @param $paths Array Storage paths
 956          * @param $type integer LockManager::LOCK_* constant
 957          * @param $status Status Status to update on lock/unlock
 958          * @return ScopedLock|null Returns null on failure
 959          */
 960         final public function getScopedFileLocks( array $paths, $type, Status $status ) {
 961                 return ScopedLock::factory( $this->lockManager, $paths, $type, $status );
 962         }
 963
 964         /**
 965          * Get an array of scoped locks needed for a batch of file operations.
 966          *
 967          * Normally, FileBackend::doOperations() handles locking, unless
 968          * the 'nonLocking' param is passed in. This function is useful if you
 969          * want the files to be locked for a broader scope than just when the
 970          * files are changing. For example, if you need to update DB metadata,
 971          * you may want to keep the files locked until finished.
 972          *
 973          * @see FileBackend::doOperations()
 974          *
 975          * @param $ops Array List of file operations to FileBackend::doOperations()
 976          * @param $status Status Status to update on lock/unlock
 977          * @return Array List of ScopedFileLocks or null values
 978          * @since 1.20
 979          */
 980         abstract public function getScopedLocksForOps( array $ops, Status $status );
 981
 982         /**
 983          * Get the root storage path of this backend.
 984          * All container paths are "subdirectories" of this path.
 985          *
 986          * @return string Storage path
 987          * @since 1.20
 988          */
 989         final public function getRootStoragePath() {
 990                 return "mwstore://{$this->name}";
 991         }
 992
 993         /**
 994          * Get the file journal object for this backend
 995          *
 996          * @return FileJournal
 997          */
 998         final public function getJournal() {
 999                 return $this->fileJournal;
1000         }
1001
1002         /**
1003          * Check if a given path is a "mwstore://" path.
1004          * This does not do any further validation or any existence checks.
1005          *
1006          * @param $path string
1007          * @return bool
1008          */
1009         final public static function isStoragePath( $path ) {
1010                 return ( strpos( $path, 'mwstore://' ) === 0 );
1011         }
1012
1013         /**
1014          * Split a storage path into a backend name, a container name,
1015          * and a relative file path. The relative path may be the empty string.
1016          * This does not do any path normalization or traversal checks.
1017          *
1018          * @param $storagePath string
1019          * @return Array (backend, container, rel object) or (null, null, null)
1020          */
1021         final public static function splitStoragePath( $storagePath ) {
1022                 if ( self::isStoragePath( $storagePath ) ) {
1023                         // Remove the "mwstore://" prefix and split the path
1024                         $parts = explode( '/', substr( $storagePath, 10 ), 3 );
1025                         if ( count( $parts ) >= 2 && $parts[0] != '' && $parts[1] != '' ) {
1026                                 if ( count( $parts ) == 3 ) {
1027                                         return $parts; // e.g. "backend/container/path"
1028                                 } else {
1029                                         return array( $parts[0], $parts[1], '' ); // e.g. "backend/container"
1030                                 }
1031                         }
1032                 }
1033                 return array( null, null, null );
1034         }
1035
1036         /**
1037          * Normalize a storage path by cleaning up directory separators.
1038          * Returns null if the path is not of the format of a valid storage path.
1039          *
1040          * @param $storagePath string
1041          * @return string|null
1042          */
1043         final public static function normalizeStoragePath( $storagePath ) {
1044                 list( $backend, $container, $relPath ) = self::splitStoragePath( $storagePath );
1045                 if ( $relPath !== null ) { // must be for this backend
1046                         $relPath = self::normalizeContainerPath( $relPath );
1047                         if ( $relPath !== null ) {
1048                                 return ( $relPath != '' )
1049                                         ? "mwstore://{$backend}/{$container}/{$relPath}"
1050                                         : "mwstore://{$backend}/{$container}";
1051                         }
1052                 }
1053                 return null;
1054         }
1055
1056         /**
1057          * Get the parent storage directory of a storage path.
1058          * This returns a path like "mwstore://backend/container",
1059          * "mwstore://backend/container/...", or null if there is no parent.
1060          *
1061          * @param $storagePath string
1062          * @return string|null
1063          */
1064         final public static function parentStoragePath( $storagePath ) {
1065                 $storagePath = dirname( $storagePath );
1066                 list( $b, $cont, $rel ) = self::splitStoragePath( $storagePath );
1067                 return ( $rel === null ) ? null : $storagePath;
1068         }
1069
1070         /**
1071          * Get the final extension from a storage or FS path
1072          *
1073          * @param $path string
1074          * @return string
1075          */
1076         final public static function extensionFromPath( $path ) {
1077                 $i = strrpos( $path, '.' );
1078                 return strtolower( $i ? substr( $path, $i + 1 ) : '' );
1079         }
1080
1081         /**
1082          * Check if a relative path has no directory traversals
1083          *
1084          * @param $path string
1085          * @return bool
1086          * @since 1.20
1087          */
1088         final public static function isPathTraversalFree( $path ) {
1089                 return ( self::normalizeContainerPath( $path ) !== null );
1090         }
1091
1092         /**
1093          * Validate and normalize a relative storage path.
1094          * Null is returned if the path involves directory traversal.
1095          * Traversal is insecure for FS backends and broken for others.
1096          *
1097          * This uses the same traversal protection as Title::secureAndSplit().
1098          *
1099          * @param $path string Storage path relative to a container
1100          * @return string|null
1101          */
1102         final protected static function normalizeContainerPath( $path ) {
1103                 // Normalize directory separators
1104                 $path = strtr( $path, '\\', '/' );
1105                 // Collapse any consecutive directory separators
1106                 $path = preg_replace( '![/]{2,}!', '/', $path );
1107                 // Remove any leading directory separator
1108                 $path = ltrim( $path, '/' );
1109                 // Use the same traversal protection as Title::secureAndSplit()
1110                 if ( strpos( $path, '.' ) !== false ) {
1111                         if (
1112                                 $path === '.' ||
1113                                 $path === '..' ||
1114                                 strpos( $path, './' ) === 0 ||
1115                                 strpos( $path, '../' ) === 0 ||
1116                                 strpos( $path, '/./' ) !== false ||
1117                                 strpos( $path, '/../' ) !== false
1118                         ) {
1119                                 return null;
1120                         }
1121                 }
1122                 return $path;
1123         }
1124 }