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. Therefore, the relative paths are only virtual.
  43  *
  44  * Backend contents are stored under wiki-specific container names by default.
  45  * For legacy reasons, this has no effect for the FS backend class, and per-wiki
  46  * segregation must be done by setting the container paths appropriately.
  47  *
  48  * FS-based backends are somewhat more restrictive due to the existence of real
  49  * directory files; a regular file cannot have the same name as a directory. Other
  50  * backends with virtual directories may not have this limitation. Callers should
  51  * store files in such a way that no files and directories are under the same path.
  52  *
  53  * Methods of subclasses should avoid throwing exceptions at all costs.
  54  * As a corollary, external dependencies should be kept to a minimum.
  55  *
  56  * @ingroup FileBackend
  57  * @since 1.19
  58  */
  59 abstract class FileBackend {
  60         protected $name; // string; unique backend name
  61         protected $wikiId; // string; unique wiki name
  62         protected $readOnly; // string; read-only explanation message
  63         protected $parallelize; // string; when to do operations in parallel
  64         protected $concurrency; // integer; how many operations can be done in parallel
  65
  66         /** @var LockManager */
  67         protected $lockManager;
  68         /** @var FileJournal */
  69         protected $fileJournal;
  70
  71         /**
  72          * Create a new backend instance from configuration.
  73          * This should only be called from within FileBackendGroup.
  74          *
  75          * $config includes:
  76          *   - name        : The unique name of this backend.
  77          *                   This should consist of alphanumberic, '-', and '_' characters.
  78          *                   This name should not be changed after use.
  79          *   - wikiId      : Prefix to container names that is unique to this wiki.
  80          *                   It should only consist of alphanumberic, '-', and '_' characters.
  81          *   - lockManager : Registered name of a file lock manager to use.
  82          *   - fileJournal : File journal configuration; see FileJournal::factory().
  83          *                   Journals simply log changes to files stored in the backend.
  84          *   - readOnly    : Write operations are disallowed if this is a non-empty string.
  85          *                   It should be an explanation for the backend being read-only.
  86          *   - parallelize : When to do file operations in parallel (when possible).
  87          *                   Allowed values are "implicit", "explicit" and "off".
  88          *   - concurrency : How many file operations can be done in parallel.
  89          *
  90          * @param $config Array
  91          * @throws MWException
  92          */
  93         public function __construct( array $config ) {
  94                 $this->name = $config['name'];
  95                 if ( !preg_match( '!^[a-zA-Z0-9-_]{1,255}$!', $this->name ) ) {
  96                         throw new MWException( "Backend name `{$this->name}` is invalid." );
  97                 }
  98                 $this->wikiId = isset( $config['wikiId'] )
  99                         ? $config['wikiId']
 100                         : wfWikiID(); // e.g. "my_wiki-en_"
 101                 $this->lockManager = ( $config['lockManager'] instanceof LockManager )
 102                         ? $config['lockManager']
 103                         : LockManagerGroup::singleton()->get( $config['lockManager'] );
 104                 $this->fileJournal = isset( $config['fileJournal'] )
 105                         ? ( ( $config['fileJournal'] instanceof FileJournal )
 106                                 ? $config['fileJournal']
 107                                 : FileJournal::factory( $config['fileJournal'], $this->name ) )
 108                         : FileJournal::factory( array( 'class' => 'NullFileJournal' ), $this->name );
 109                 $this->readOnly = isset( $config['readOnly'] )
 110                         ? (string)$config['readOnly']
 111                         : '';
 112                 $this->parallelize = isset( $config['parallelize'] )
 113                         ? (string)$config['parallelize']
 114                         : 'off';
 115                 $this->concurrency = isset( $config['concurrency'] )
 116                         ? (int)$config['concurrency']
 117                         : 50;
 118         }
 119
 120         /**
 121          * Get the unique backend name.
 122          * We may have multiple different backends of the same type.
 123          * For example, we can have two Swift backends using different proxies.
 124          *
 125          * @return string
 126          */
 127         final public function getName() {
 128                 return $this->name;
 129         }
 130
 131         /**
 132          * Get the wiki identifier used for this backend (possibly empty)
 133          *
 134          * @return string
 135          * @since 1.20
 136          */
 137         final public function getWikiId() {
 138                 return $this->wikiId;
 139         }
 140
 141         /**
 142          * Check if this backend is read-only
 143          *
 144          * @return bool
 145          */
 146         final public function isReadOnly() {
 147                 return ( $this->readOnly != '' );
 148         }
 149
 150         /**
 151          * Get an explanatory message if this backend is read-only
 152          *
 153          * @return string|bool Returns false if the backend is not read-only
 154          */
 155         final public function getReadOnlyReason() {
 156                 return ( $this->readOnly != '' ) ? $this->readOnly : false;
 157         }
 158
 159         /**
 160          * This is the main entry point into the backend for write operations.
 161          * Callers supply an ordered list of operations to perform as a transaction.
 162          * Files will be locked, the stat cache cleared, and then the operations attempted.
 163          * If any serious errors occur, all attempted operations will be rolled back.
 164          *
 165          * $ops is an array of arrays. The outer array holds a list of operations.
 166          * Each inner array is a set of key value pairs that specify an operation.
 167          *
 168          * Supported operations and their parameters. The supported actions are:
 169          *  - create
 170          *  - store
 171          *  - copy
 172          *  - move
 173          *  - delete
 174          *  - null
 175          *
 176          * a) Create a new file in storage with the contents of a string
 177          * @code
 178          *     array(
 179          *         'op'                  => 'create',
 180          *         'dst'                 => <storage path>,
 181          *         'content'             => <string of new file contents>,
 182          *         'overwrite'           => <boolean>,
 183          *         'overwriteSame'       => <boolean>,
 184          *         'disposition'         => <Content-Disposition header value>
 185          *     );
 186          * @endcode
 187          *
 188          * b) Copy a file system file into storage
 189          * @code
 190          *     array(
 191          *         'op'                  => 'store',
 192          *         'src'                 => <file system path>,
 193          *         'dst'                 => <storage path>,
 194          *         'overwrite'           => <boolean>,
 195          *         'overwriteSame'       => <boolean>,
 196          *         'disposition'         => <Content-Disposition header value>
 197          *     )
 198          * @endcode
 199          *
 200          * c) Copy a file within storage
 201          * @code
 202          *     array(
 203          *         'op'                  => 'copy',
 204          *         'src'                 => <storage path>,
 205          *         'dst'                 => <storage path>,
 206          *         'overwrite'           => <boolean>,
 207          *         'overwriteSame'       => <boolean>,
 208          *         'disposition'         => <Content-Disposition header value>
 209          *     )
 210          * @endcode
 211          *
 212          * d) Move a file within storage
 213          * @code
 214          *     array(
 215          *         'op'                  => 'move',
 216          *         'src'                 => <storage path>,
 217          *         'dst'                 => <storage path>,
 218          *         'overwrite'           => <boolean>,
 219          *         'overwriteSame'       => <boolean>,
 220          *         'disposition'         => <Content-Disposition header value>
 221          *     )
 222          * @endcode
 223          *
 224          * e) Delete a file within storage
 225          * @code
 226          *     array(
 227          *         'op'                  => 'delete',
 228          *         'src'                 => <storage path>,
 229          *         'ignoreMissingSource' => <boolean>
 230          *     )
 231          * @endcode
 232          *
 233          * f) Do nothing (no-op)
 234          * @code
 235          *     array(
 236          *         'op'                  => 'null',
 237          *     )
 238          * @endcode
 239          *
 240          * Boolean flags for operations (operation-specific):
 241          *   - ignoreMissingSource : The operation will simply succeed and do
 242          *                           nothing if the source file does not exist.
 243          *   - overwrite           : Any destination file will be overwritten.
 244          *   - overwriteSame       : An error will not be given if a file already
 245          *                           exists at the destination that has the same
 246          *                           contents as the new contents to be written there.
 247          *   - disposition         : When supplied, the backend will add a Content-Disposition
 248          *                           header when GETs/HEADs of the destination file are made.
 249          *                           Backends that don't support file metadata will ignore this.
 250          *                           See http://tools.ietf.org/html/rfc6266 (since 1.20).
 251          *
 252          * $opts is an associative of boolean flags, including:
 253          *   - force               : Operation precondition errors no longer trigger an abort.
 254          *                           Any remaining operations are still attempted. Unexpected
 255          *                           failures may still cause remaning operations to be aborted.
 256          *   - nonLocking          : No locks are acquired for the operations.
 257          *                           This can increase performance for non-critical writes.
 258          *                           This has no effect unless the 'force' flag is set.
 259          *   - allowStale          : Don't require the latest available data.
 260          *                           This can increase performance for non-critical writes.
 261          *                           This has no effect unless the 'force' flag is set.
 262          *   - nonJournaled        : Don't log this operation batch in the file journal.
 263          *                           This limits the ability of recovery scripts.
 264          *   - parallelize         : Try to do operations in parallel when possible.
 265          *   - bypassReadOnly      : Allow writes in read-only mode (since 1.20).
 266          *   - preserveCache       : Don't clear the process cache before checking files.
 267          *                           This should only be used if all entries in the process
 268          *                           cache were added after the files were already locked (since 1.20).
 269          *
 270          * @remarks Remarks on locking:
 271          * File system paths given to operations should refer to files that are
 272          * already locked or otherwise safe from modification from other processes.
 273          * Normally these files will be new temp files, which should be adequate.
 274          *
 275          * @par Return value:
 276          *
 277          * This returns a Status, which contains all warnings and fatals that occurred
 278          * during the operation. The 'failCount', 'successCount', and 'success' members
 279          * will reflect each operation attempted.
 280          *
 281          * The status will be "OK" unless:
 282          *   - a) unexpected operation errors occurred (network partitions, disk full...)
 283          *   - b) significant operation errors occurred and 'force' was not set
 284          *
 285          * @param $ops Array List of operations to execute in order
 286          * @param $opts Array Batch operation options
 287          * @return Status
 288          */
 289         final public function doOperations( array $ops, array $opts = array() ) {
 290                 if ( empty( $opts['bypassReadOnly'] ) && $this->isReadOnly() ) {
 291                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 292                 }
 293                 if ( empty( $opts['force'] ) ) { // sanity
 294                         unset( $opts['nonLocking'] );
 295                         unset( $opts['allowStale'] );
 296                 }
 297                 return $this->doOperationsInternal( $ops, $opts );
 298         }
 299
 300         /**
 301          * @see FileBackend::doOperations()
 302          */
 303         abstract protected function doOperationsInternal( array $ops, array $opts );
 304
 305         /**
 306          * Same as doOperations() except it takes a single operation.
 307          * If you are doing a batch of operations that should either
 308          * all succeed or all fail, then use that function instead.
 309          *
 310          * @see FileBackend::doOperations()
 311          *
 312          * @param $op Array Operation
 313          * @param $opts Array Operation options
 314          * @return Status
 315          */
 316         final public function doOperation( array $op, array $opts = array() ) {
 317                 return $this->doOperations( array( $op ), $opts );
 318         }
 319
 320         /**
 321          * Performs a single create operation.
 322          * This sets $params['op'] to 'create' and passes it to doOperation().
 323          *
 324          * @see FileBackend::doOperation()
 325          *
 326          * @param $params Array Operation parameters
 327          * @param $opts Array Operation options
 328          * @return Status
 329          */
 330         final public function create( array $params, array $opts = array() ) {
 331                 return $this->doOperation( array( 'op' => 'create' ) + $params, $opts );
 332         }
 333
 334         /**
 335          * Performs a single store operation.
 336          * This sets $params['op'] to 'store' and passes it to doOperation().
 337          *
 338          * @see FileBackend::doOperation()
 339          *
 340          * @param $params Array Operation parameters
 341          * @param $opts Array Operation options
 342          * @return Status
 343          */
 344         final public function store( array $params, array $opts = array() ) {
 345                 return $this->doOperation( array( 'op' => 'store' ) + $params, $opts );
 346         }
 347
 348         /**
 349          * Performs a single copy operation.
 350          * This sets $params['op'] to 'copy' and passes it to doOperation().
 351          *
 352          * @see FileBackend::doOperation()
 353          *
 354          * @param $params Array Operation parameters
 355          * @param $opts Array Operation options
 356          * @return Status
 357          */
 358         final public function copy( array $params, array $opts = array() ) {
 359                 return $this->doOperation( array( 'op' => 'copy' ) + $params, $opts );
 360         }
 361
 362         /**
 363          * Performs a single move operation.
 364          * This sets $params['op'] to 'move' and passes it to doOperation().
 365          *
 366          * @see FileBackend::doOperation()
 367          *
 368          * @param $params Array Operation parameters
 369          * @param $opts Array Operation options
 370          * @return Status
 371          */
 372         final public function move( array $params, array $opts = array() ) {
 373                 return $this->doOperation( array( 'op' => 'move' ) + $params, $opts );
 374         }
 375
 376         /**
 377          * Performs a single delete operation.
 378          * This sets $params['op'] to 'delete' and passes it to doOperation().
 379          *
 380          * @see FileBackend::doOperation()
 381          *
 382          * @param $params Array Operation parameters
 383          * @param $opts Array Operation options
 384          * @return Status
 385          */
 386         final public function delete( array $params, array $opts = array() ) {
 387                 return $this->doOperation( array( 'op' => 'delete' ) + $params, $opts );
 388         }
 389
 390         /**
 391          * Perform a set of independent file operations on some files.
 392          *
 393          * This does no locking, nor journaling, and possibly no stat calls.
 394          * Any destination files that already exist will be overwritten.
 395          * This should *only* be used on non-original files, like cache files.
 396          *
 397          * Supported operations and their parameters:
 398          *  - create
 399          *  - store
 400          *  - copy
 401          *  - move
 402          *  - delete
 403          *  - null
 404          *
 405          * a) Create a new file in storage with the contents of a string
 406          * @code
 407          *     array(
 408          *         'op'                  => 'create',
 409          *         'dst'                 => <storage path>,
 410          *         'content'             => <string of new file contents>,
 411          *         'disposition'         => <Content-Disposition header value>
 412          *     )
 413          * @endcode
 414          * b) Copy a file system file into storage
 415          * @code
 416          *     array(
 417          *         'op'                  => 'store',
 418          *         'src'                 => <file system path>,
 419          *         'dst'                 => <storage path>,
 420          *         'disposition'         => <Content-Disposition header value>
 421          *     )
 422          * @endcode
 423          * c) Copy a file within storage
 424          * @code
 425          *     array(
 426          *         'op'                  => 'copy',
 427          *         'src'                 => <storage path>,
 428          *         'dst'                 => <storage path>,
 429          *         'disposition'         => <Content-Disposition header value>
 430          *     )
 431          * @endcode
 432          * d) Move a file within storage
 433          * @code
 434          *     array(
 435          *         'op'                  => 'move',
 436          *         'src'                 => <storage path>,
 437          *         'dst'                 => <storage path>,
 438          *         'disposition'         => <Content-Disposition header value>
 439          *     )
 440          * @endcode
 441          * e) Delete a file within storage
 442          * @code
 443          *     array(
 444          *         'op'                  => 'delete',
 445          *         'src'                 => <storage path>,
 446          *         'ignoreMissingSource' => <boolean>
 447          *     )
 448          * @endcode
 449          * f) Do nothing (no-op)
 450          * @code
 451          *     array(
 452          *         'op'                  => 'null',
 453          *     )
 454          * @endcode
 455          *
 456          * @par Boolean flags for operations (operation-specific):
 457          *   - ignoreMissingSource : The operation will simply succeed and do
 458          *                           nothing if the source file does not exist.
 459          *   - disposition         : When supplied, the backend will add a Content-Disposition
 460          *                           header when GETs/HEADs of the destination file are made.
 461          *                           Backends that don't support file metadata will ignore this.
 462          *                           See http://tools.ietf.org/html/rfc6266 (since 1.20).
 463          *
 464          * $opts is an associative of boolean flags, including:
 465          *   - bypassReadOnly      : Allow writes in read-only mode (since 1.20)
 466          *
 467          * @par Return value:
 468          * This returns a Status, which contains all warnings and fatals that occurred
 469          * during the operation. The 'failCount', 'successCount', and 'success' members
 470          * will reflect each operation attempted for the given files. The status will be
 471          * considered "OK" as long as no fatal errors occurred.
 472          *
 473          * @param $ops Array Set of operations to execute
 474          * @param $opts Array Batch operation options
 475          * @return Status
 476          * @since 1.20
 477          */
 478         final public function doQuickOperations( array $ops, array $opts = array() ) {
 479                 if ( empty( $opts['bypassReadOnly'] ) && $this->isReadOnly() ) {
 480                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 481                 }
 482                 foreach ( $ops as &$op ) {
 483                         $op['overwrite'] = true; // avoids RTTs in key/value stores
 484                 }
 485                 return $this->doQuickOperationsInternal( $ops );
 486         }
 487
 488         /**
 489          * @see FileBackend::doQuickOperations()
 490          * @since 1.20
 491          */
 492         abstract protected function doQuickOperationsInternal( array $ops );
 493
 494         /**
 495          * Same as doQuickOperations() except it takes a single operation.
 496          * If you are doing a batch of operations, then use that function instead.
 497          *
 498          * @see FileBackend::doQuickOperations()
 499          *
 500          * @param $op Array Operation
 501          * @return Status
 502          * @since 1.20
 503          */
 504         final public function doQuickOperation( array $op ) {
 505                 return $this->doQuickOperations( array( $op ) );
 506         }
 507
 508         /**
 509          * Performs a single quick create operation.
 510          * This sets $params['op'] to 'create' and passes it to doQuickOperation().
 511          *
 512          * @see FileBackend::doQuickOperation()
 513          *
 514          * @param $params Array Operation parameters
 515          * @return Status
 516          * @since 1.20
 517          */
 518         final public function quickCreate( array $params ) {
 519                 return $this->doQuickOperation( array( 'op' => 'create' ) + $params );
 520         }
 521
 522         /**
 523          * Performs a single quick store operation.
 524          * This sets $params['op'] to 'store' and passes it to doQuickOperation().
 525          *
 526          * @see FileBackend::doQuickOperation()
 527          *
 528          * @param $params Array Operation parameters
 529          * @return Status
 530          * @since 1.20
 531          */
 532         final public function quickStore( array $params ) {
 533                 return $this->doQuickOperation( array( 'op' => 'store' ) + $params );
 534         }
 535
 536         /**
 537          * Performs a single quick copy operation.
 538          * This sets $params['op'] to 'copy' and passes it to doQuickOperation().
 539          *
 540          * @see FileBackend::doQuickOperation()
 541          *
 542          * @param $params Array Operation parameters
 543          * @return Status
 544          * @since 1.20
 545          */
 546         final public function quickCopy( array $params ) {
 547                 return $this->doQuickOperation( array( 'op' => 'copy' ) + $params );
 548         }
 549
 550         /**
 551          * Performs a single quick move operation.
 552          * This sets $params['op'] to 'move' and passes it to doQuickOperation().
 553          *
 554          * @see FileBackend::doQuickOperation()
 555          *
 556          * @param $params Array Operation parameters
 557          * @return Status
 558          * @since 1.20
 559          */
 560         final public function quickMove( array $params ) {
 561                 return $this->doQuickOperation( array( 'op' => 'move' ) + $params );
 562         }
 563
 564         /**
 565          * Performs a single quick delete operation.
 566          * This sets $params['op'] to 'delete' and passes it to doQuickOperation().
 567          *
 568          * @see FileBackend::doQuickOperation()
 569          *
 570          * @param $params Array Operation parameters
 571          * @return Status
 572          * @since 1.20
 573          */
 574         final public function quickDelete( array $params ) {
 575                 return $this->doQuickOperation( array( 'op' => 'delete' ) + $params );
 576         }
 577
 578         /**
 579          * Concatenate a list of storage files into a single file system file.
 580          * The target path should refer to a file that is already locked or
 581          * otherwise safe from modification from other processes. Normally,
 582          * the file will be a new temp file, which should be adequate.
 583          *
 584          * @param $params Array Operation parameters
 585          * $params include:
 586          *   - srcs : ordered source storage paths (e.g. chunk1, chunk2, ...)
 587          *   - dst  : file system path to 0-byte temp file
 588          * @return Status
 589          */
 590         abstract public function concatenate( array $params );
 591
 592         /**
 593          * Prepare a storage directory for usage.
 594          * This will create any required containers and parent directories.
 595          * Backends using key/value stores only need to create the container.
 596          *
 597          * The 'noAccess' and 'noListing' parameters works the same as in secure(),
 598          * except they are only applied *if* the directory/container had to be created.
 599          * These flags should always be set for directories that have private files.
 600          *
 601          * @param $params Array
 602          * $params include:
 603          *   - dir            : storage directory
 604          *   - noAccess       : try to deny file access (since 1.20)
 605          *   - noListing      : try to deny file listing (since 1.20)
 606          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 607          * @return Status
 608          */
 609         final public function prepare( array $params ) {
 610                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 611                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 612                 }
 613                 return $this->doPrepare( $params );
 614         }
 615
 616         /**
 617          * @see FileBackend::prepare()
 618          */
 619         abstract protected function doPrepare( array $params );
 620
 621         /**
 622          * Take measures to block web access to a storage directory and
 623          * the container it belongs to. FS backends might add .htaccess
 624          * files whereas key/value store backends might revoke container
 625          * access to the storage user representing end-users in web requests.
 626          * This is not guaranteed to actually do anything.
 627          *
 628          * @param $params Array
 629          * $params include:
 630          *   - dir            : storage directory
 631          *   - noAccess       : try to deny file access
 632          *   - noListing      : try to deny file listing
 633          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 634          * @return Status
 635          */
 636         final public function secure( array $params ) {
 637                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 638                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 639                 }
 640                 return $this->doSecure( $params );
 641         }
 642
 643         /**
 644          * @see FileBackend::secure()
 645          */
 646         abstract protected function doSecure( array $params );
 647
 648         /**
 649          * Remove measures to block web access to a storage directory and
 650          * the container it belongs to. FS backends might remove .htaccess
 651          * files whereas key/value store backends might grant container
 652          * access to the storage user representing end-users in web requests.
 653          * This essentially can undo the result of secure() calls.
 654          *
 655          * @param $params Array
 656          * $params include:
 657          *   - dir            : storage directory
 658          *   - access         : try to allow file access
 659          *   - listing        : try to allow file listing
 660          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 661          * @return Status
 662          * @since 1.20
 663          */
 664         final public function publish( array $params ) {
 665                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 666                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 667                 }
 668                 return $this->doPublish( $params );
 669         }
 670
 671         /**
 672          * @see FileBackend::publish()
 673          */
 674         abstract protected function doPublish( array $params );
 675
 676         /**
 677          * Delete a storage directory if it is empty.
 678          * Backends using key/value stores may do nothing unless the directory
 679          * is that of an empty container, in which case it will be deleted.
 680          *
 681          * @param $params Array
 682          * $params include:
 683          *   - dir            : storage directory
 684          *   - recursive      : recursively delete empty subdirectories first (since 1.20)
 685          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 686          * @return Status
 687          */
 688         final public function clean( array $params ) {
 689                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 690                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 691                 }
 692                 return $this->doClean( $params );
 693         }
 694
 695         /**
 696          * @see FileBackend::clean()
 697          */
 698         abstract protected function doClean( array $params );
 699
 700         /**
 701          * Check if a file exists at a storage path in the backend.
 702          * This returns false if only a directory exists at the path.
 703          *
 704          * @param $params Array
 705          * $params include:
 706          *   - src    : source storage path
 707          *   - latest : use the latest available data
 708          * @return bool|null Returns null on failure
 709          */
 710         abstract public function fileExists( array $params );
 711
 712         /**
 713          * Get the last-modified timestamp of the file at a storage path.
 714          *
 715          * @param $params Array
 716          * $params include:
 717          *   - src    : source storage path
 718          *   - latest : use the latest available data
 719          * @return string|bool TS_MW timestamp or false on failure
 720          */
 721         abstract public function getFileTimestamp( array $params );
 722
 723         /**
 724          * Get the contents of a file at a storage path in the backend.
 725          * This should be avoided for potentially large files.
 726          *
 727          * @param $params Array
 728          * $params include:
 729          *   - src    : source storage path
 730          *   - latest : use the latest available data
 731          * @return string|bool Returns false on failure
 732          */
 733         abstract public function getFileContents( array $params );
 734
 735         /**
 736          * Get the size (bytes) of a file at a storage path in the backend.
 737          *
 738          * @param $params Array
 739          * $params include:
 740          *   - src    : source storage path
 741          *   - latest : use the latest available data
 742          * @return integer|bool Returns false on failure
 743          */
 744         abstract public function getFileSize( array $params );
 745
 746         /**
 747          * Get quick information about a file at a storage path in the backend.
 748          * If the file does not exist, then this returns false.
 749          * Otherwise, the result is an associative array that includes:
 750          *   - mtime  : the last-modified timestamp (TS_MW)
 751          *   - size   : the file size (bytes)
 752          * Additional values may be included for internal use only.
 753          *
 754          * @param $params Array
 755          * $params include:
 756          *   - src    : source storage path
 757          *   - latest : use the latest available data
 758          * @return Array|bool|null Returns null on failure
 759          */
 760         abstract public function getFileStat( array $params );
 761
 762         /**
 763          * Get a SHA-1 hash of the file at a storage path in the backend.
 764          *
 765          * @param $params Array
 766          * $params include:
 767          *   - src    : source storage path
 768          *   - latest : use the latest available data
 769          * @return string|bool Hash string or false on failure
 770          */
 771         abstract public function getFileSha1Base36( array $params );
 772
 773         /**
 774          * Get the properties of the file at a storage path in the backend.
 775          * Returns FSFile::placeholderProps() on failure.
 776          *
 777          * @param $params Array
 778          * $params include:
 779          *   - src    : source storage path
 780          *   - latest : use the latest available data
 781          * @return Array
 782          */
 783         abstract public function getFileProps( array $params );
 784
 785         /**
 786          * Stream the file at a storage path in the backend.
 787          * If the file does not exists, an HTTP 404 error will be given.
 788          * Appropriate HTTP headers (Status, Content-Type, Content-Length)
 789          * will be sent if streaming began, while none will be sent otherwise.
 790          * Implementations should flush the output buffer before sending data.
 791          *
 792          * @param $params Array
 793          * $params include:
 794          *   - src     : source storage path
 795          *   - headers : list of additional HTTP headers to send on success
 796          *   - latest  : use the latest available data
 797          * @return Status
 798          */
 799         abstract public function streamFile( array $params );
 800
 801         /**
 802          * Returns a file system file, identical to the file at a storage path.
 803          * The file returned is either:
 804          *   - a) A local copy of the file at a storage path in the backend.
 805          *        The temporary copy will have the same extension as the source.
 806          *   - b) An original of the file at a storage path in the backend.
 807          * Temporary files may be purged when the file object falls out of scope.
 808          *
 809          * Write operations should *never* be done on this file as some backends
 810          * may do internal tracking or may be instances of FileBackendMultiWrite.
 811          * In that later case, there are copies of the file that must stay in sync.
 812          * Additionally, further calls to this function may return the same file.
 813          *
 814          * @param $params Array
 815          * $params include:
 816          *   - src    : source storage path
 817          *   - latest : use the latest available data
 818          * @return FSFile|null Returns null on failure
 819          */
 820         abstract public function getLocalReference( array $params );
 821
 822         /**
 823          * Get a local copy on disk of the file at a storage path in the backend.
 824          * The temporary copy will have the same file extension as the source.
 825          * Temporary files may be purged when the file object falls out of scope.
 826          *
 827          * @param $params Array
 828          * $params include:
 829          *   - src    : source storage path
 830          *   - latest : use the latest available data
 831          * @return TempFSFile|null Returns null on failure
 832          */
 833         abstract public function getLocalCopy( array $params );
 834
 835         /**
 836          * Check if a directory exists at a given storage path.
 837          * Backends using key/value stores will check if the path is a
 838          * virtual directory, meaning there are files under the given directory.
 839          *
 840          * Storage backends with eventual consistency might return stale data.
 841          *
 842          * @param $params array
 843          * $params include:
 844          *   - dir : storage directory
 845          * @return bool|null Returns null on failure
 846          * @since 1.20
 847          */
 848         abstract public function directoryExists( array $params );
 849
 850         /**
 851          * Get an iterator to list *all* directories under a storage directory.
 852          * If the directory is of the form "mwstore://backend/container",
 853          * then all directories in the container will be listed.
 854          * If the directory is of form "mwstore://backend/container/dir",
 855          * then all directories directly under that directory will be listed.
 856          * Results will be storage directories relative to 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          *   - topOnly : only return direct child dirs of the directory
 864          * @return Traversable|Array|null Returns null on failure
 865          * @since 1.20
 866          */
 867         abstract public function getDirectoryList( array $params );
 868
 869         /**
 870          * Same as FileBackend::getDirectoryList() except only lists
 871          * directories that are immediately under the given directory.
 872          *
 873          * Storage backends with eventual consistency might return stale data.
 874          *
 875          * @param $params array
 876          * $params include:
 877          *   - dir : storage directory
 878          * @return Traversable|Array|null Returns null on failure
 879          * @since 1.20
 880          */
 881         final public function getTopDirectoryList( array $params ) {
 882                 return $this->getDirectoryList( array( 'topOnly' => true ) + $params );
 883         }
 884
 885         /**
 886          * Get an iterator to list *all* stored files under a storage directory.
 887          * If the directory is of the form "mwstore://backend/container",
 888          * then all files in the container will be listed.
 889          * If the directory is of form "mwstore://backend/container/dir",
 890          * then all files under that directory will be listed.
 891          * Results will be storage paths relative to the given directory.
 892          *
 893          * Storage backends with eventual consistency might return stale data.
 894          *
 895          * @param $params array
 896          * $params include:
 897          *   - dir     : storage directory
 898          *   - topOnly : only return direct child files of the directory (since 1.20)
 899          * @return Traversable|Array|null Returns null on failure
 900          */
 901         abstract public function getFileList( array $params );
 902
 903         /**
 904          * Same as FileBackend::getFileList() except only lists
 905          * files that are immediately under the given directory.
 906          *
 907          * Storage backends with eventual consistency might return stale data.
 908          *
 909          * @param $params array
 910          * $params include:
 911          *   - dir : storage directory
 912          * @return Traversable|Array|null Returns null on failure
 913          * @since 1.20
 914          */
 915         final public function getTopFileList( array $params ) {
 916                 return $this->getFileList( array( 'topOnly' => true ) + $params );
 917         }
 918
 919         /**
 920          * Preload persistent file stat and property cache into in-process cache.
 921          * This should be used when stat calls will be made on a known list of a many files.
 922          *
 923          * @param $paths Array Storage paths
 924          * @return void
 925          */
 926         public function preloadCache( array $paths ) {}
 927
 928         /**
 929          * Invalidate any in-process file stat and property cache.
 930          * If $paths is given, then only the cache for those files will be cleared.
 931          *
 932          * @param $paths Array Storage paths (optional)
 933          * @return void
 934          */
 935         public function clearCache( array $paths = null ) {}
 936
 937         /**
 938          * Lock the files at the given storage paths in the backend.
 939          * This will either lock all the files or none (on failure).
 940          *
 941          * Callers should consider using getScopedFileLocks() instead.
 942          *
 943          * @param $paths Array Storage paths
 944          * @param $type integer LockManager::LOCK_* constant
 945          * @return Status
 946          */
 947         final public function lockFiles( array $paths, $type ) {
 948                 return $this->lockManager->lock( $paths, $type );
 949         }
 950
 951         /**
 952          * Unlock the files at the given storage paths in the backend.
 953          *
 954          * @param $paths Array Storage paths
 955          * @param $type integer LockManager::LOCK_* constant
 956          * @return Status
 957          */
 958         final public function unlockFiles( array $paths, $type ) {
 959                 return $this->lockManager->unlock( $paths, $type );
 960         }
 961
 962         /**
 963          * Lock the files at the given storage paths in the backend.
 964          * This will either lock all the files or none (on failure).
 965          * On failure, the status object will be updated with errors.
 966          *
 967          * Once the return value goes out scope, the locks will be released and
 968          * the status updated. Unlock fatals will not change the status "OK" value.
 969          *
 970          * @param $paths Array Storage paths
 971          * @param $type integer LockManager::LOCK_* constant
 972          * @param $status Status Status to update on lock/unlock
 973          * @return ScopedLock|null Returns null on failure
 974          */
 975         final public function getScopedFileLocks( array $paths, $type, Status $status ) {
 976                 return ScopedLock::factory( $this->lockManager, $paths, $type, $status );
 977         }
 978
 979         /**
 980          * Get an array of scoped locks needed for a batch of file operations.
 981          *
 982          * Normally, FileBackend::doOperations() handles locking, unless
 983          * the 'nonLocking' param is passed in. This function is useful if you
 984          * want the files to be locked for a broader scope than just when the
 985          * files are changing. For example, if you need to update DB metadata,
 986          * you may want to keep the files locked until finished.
 987          *
 988          * @see FileBackend::doOperations()
 989          *
 990          * @param $ops Array List of file operations to FileBackend::doOperations()
 991          * @param $status Status Status to update on lock/unlock
 992          * @return Array List of ScopedFileLocks or null values
 993          * @since 1.20
 994          */
 995         abstract public function getScopedLocksForOps( array $ops, Status $status );
 996
 997         /**
 998          * Get the root storage path of this backend.
 999          * All container paths are "subdirectories" of this path.
1000          *
1001          * @return string Storage path
1002          * @since 1.20
1003          */
1004         final public function getRootStoragePath() {
1005                 return "mwstore://{$this->name}";
1006         }
1007
1008         /**
1009          * Get the file journal object for this backend
1010          *
1011          * @return FileJournal
1012          */
1013         final public function getJournal() {
1014                 return $this->fileJournal;
1015         }
1016
1017         /**
1018          * Check if a given path is a "mwstore://" path.
1019          * This does not do any further validation or any existence checks.
1020          *
1021          * @param $path string
1022          * @return bool
1023          */
1024         final public static function isStoragePath( $path ) {
1025                 return ( strpos( $path, 'mwstore://' ) === 0 );
1026         }
1027
1028         /**
1029          * Split a storage path into a backend name, a container name,
1030          * and a relative file path. The relative path may be the empty string.
1031          * This does not do any path normalization or traversal checks.
1032          *
1033          * @param $storagePath string
1034          * @return Array (backend, container, rel object) or (null, null, null)
1035          */
1036         final public static function splitStoragePath( $storagePath ) {
1037                 if ( self::isStoragePath( $storagePath ) ) {
1038                         // Remove the "mwstore://" prefix and split the path
1039                         $parts = explode( '/', substr( $storagePath, 10 ), 3 );
1040                         if ( count( $parts ) >= 2 && $parts[0] != '' && $parts[1] != '' ) {
1041                                 if ( count( $parts ) == 3 ) {
1042                                         return $parts; // e.g. "backend/container/path"
1043                                 } else {
1044                                         return array( $parts[0], $parts[1], '' ); // e.g. "backend/container"
1045                                 }
1046                         }
1047                 }
1048                 return array( null, null, null );
1049         }
1050
1051         /**
1052          * Normalize a storage path by cleaning up directory separators.
1053          * Returns null if the path is not of the format of a valid storage path.
1054          *
1055          * @param $storagePath string
1056          * @return string|null
1057          */
1058         final public static function normalizeStoragePath( $storagePath ) {
1059                 list( $backend, $container, $relPath ) = self::splitStoragePath( $storagePath );
1060                 if ( $relPath !== null ) { // must be for this backend
1061                         $relPath = self::normalizeContainerPath( $relPath );
1062                         if ( $relPath !== null ) {
1063                                 return ( $relPath != '' )
1064                                         ? "mwstore://{$backend}/{$container}/{$relPath}"
1065                                         : "mwstore://{$backend}/{$container}";
1066                         }
1067                 }
1068                 return null;
1069         }
1070
1071         /**
1072          * Get the parent storage directory of a storage path.
1073          * This returns a path like "mwstore://backend/container",
1074          * "mwstore://backend/container/...", or null if there is no parent.
1075          *
1076          * @param $storagePath string
1077          * @return string|null
1078          */
1079         final public static function parentStoragePath( $storagePath ) {
1080                 $storagePath = dirname( $storagePath );
1081                 list( $b, $cont, $rel ) = self::splitStoragePath( $storagePath );
1082                 return ( $rel === null ) ? null : $storagePath;
1083         }
1084
1085         /**
1086          * Get the final extension from a storage or FS path
1087          *
1088          * @param $path string
1089          * @return string
1090          */
1091         final public static function extensionFromPath( $path ) {
1092                 $i = strrpos( $path, '.' );
1093                 return strtolower( $i ? substr( $path, $i + 1 ) : '' );
1094         }
1095
1096         /**
1097          * Check if a relative path has no directory traversals
1098          *
1099          * @param $path string
1100          * @return bool
1101          * @since 1.20
1102          */
1103         final public static function isPathTraversalFree( $path ) {
1104                 return ( self::normalizeContainerPath( $path ) !== null );
1105         }
1106
1107         /**
1108          * Build a Content-Disposition header value per RFC 6266.
1109          *
1110          * @param $type string One of (attachment, inline)
1111          * @param $filename string Suggested file name (should not contain slashes)
1112          * @return string
1113          * @since 1.20
1114          */
1115         final public static function makeContentDisposition( $type, $filename = '' ) {
1116                 $parts = array();
1117
1118                 $type = strtolower( $type );
1119                 if ( !in_array( $type, array( 'inline', 'attachment' ) ) ) {
1120                         throw new MWException( "Invalid Content-Disposition type '$type'." );
1121                 }
1122                 $parts[] = $type;
1123
1124                 if ( strlen( $filename ) ) {
1125                         $parts[] = "filename*=UTF-8''" . rawurlencode( basename( $filename ) );
1126                 }
1127
1128                 return implode( ';', $parts );
1129         }
1130
1131         /**
1132          * Validate and normalize a relative storage path.
1133          * Null is returned if the path involves directory traversal.
1134          * Traversal is insecure for FS backends and broken for others.
1135          *
1136          * This uses the same traversal protection as Title::secureAndSplit().
1137          *
1138          * @param $path string Storage path relative to a container
1139          * @return string|null
1140          */
1141         final protected static function normalizeContainerPath( $path ) {
1142                 // Normalize directory separators
1143                 $path = strtr( $path, '\\', '/' );
1144                 // Collapse any consecutive directory separators
1145                 $path = preg_replace( '![/]{2,}!', '/', $path );
1146                 // Remove any leading directory separator
1147                 $path = ltrim( $path, '/' );
1148                 // Use the same traversal protection as Title::secureAndSplit()
1149                 if ( strpos( $path, '.' ) !== false ) {
1150                         if (
1151                                 $path === '.' ||
1152                                 $path === '..' ||
1153                                 strpos( $path, './' ) === 0 ||
1154                                 strpos( $path, '../' ) === 0 ||
1155                                 strpos( $path, '/./' ) !== false ||
1156                                 strpos( $path, '/../' ) !== false
1157                         ) {
1158                                 return null;
1159                         }
1160                 }
1161                 return $path;
1162         }
1163 }