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          * Note that this might *not* be in the same format as wfWikiID().
 134          *
 135          * @return string
 136          * @since 1.20
 137          */
 138         final public function getWikiId() {
 139                 return $this->wikiId;
 140         }
 141
 142         /**
 143          * Check if this backend is read-only
 144          *
 145          * @return bool
 146          */
 147         final public function isReadOnly() {
 148                 return ( $this->readOnly != '' );
 149         }
 150
 151         /**
 152          * Get an explanatory message if this backend is read-only
 153          *
 154          * @return string|bool Returns false if the backend is not read-only
 155          */
 156         final public function getReadOnlyReason() {
 157                 return ( $this->readOnly != '' ) ? $this->readOnly : false;
 158         }
 159
 160         /**
 161          * This is the main entry point into the backend for write operations.
 162          * Callers supply an ordered list of operations to perform as a transaction.
 163          * Files will be locked, the stat cache cleared, and then the operations attempted.
 164          * If any serious errors occur, all attempted operations will be rolled back.
 165          *
 166          * $ops is an array of arrays. The outer array holds a list of operations.
 167          * Each inner array is a set of key value pairs that specify an operation.
 168          *
 169          * Supported operations and their parameters. The supported actions are:
 170          *  - create
 171          *  - store
 172          *  - copy
 173          *  - move
 174          *  - delete
 175          *  - null
 176          *
 177          * a) Create a new file in storage with the contents of a string
 178          * @code
 179          *     array(
 180          *         'op'                  => 'create',
 181          *         'dst'                 => <storage path>,
 182          *         'content'             => <string of new file contents>,
 183          *         'overwrite'           => <boolean>,
 184          *         'overwriteSame'       => <boolean>,
 185          *         'disposition'         => <Content-Disposition header value>
 186          *     );
 187          * @endcode
 188          *
 189          * b) Copy a file system file into storage
 190          * @code
 191          *     array(
 192          *         'op'                  => 'store',
 193          *         'src'                 => <file system path>,
 194          *         'dst'                 => <storage path>,
 195          *         'overwrite'           => <boolean>,
 196          *         'overwriteSame'       => <boolean>,
 197          *         'disposition'         => <Content-Disposition header value>
 198          *     )
 199          * @endcode
 200          *
 201          * c) Copy a file within storage
 202          * @code
 203          *     array(
 204          *         'op'                  => 'copy',
 205          *         'src'                 => <storage path>,
 206          *         'dst'                 => <storage path>,
 207          *         'overwrite'           => <boolean>,
 208          *         'overwriteSame'       => <boolean>,
 209          *         'disposition'         => <Content-Disposition header value>
 210          *     )
 211          * @endcode
 212          *
 213          * d) Move a file within storage
 214          * @code
 215          *     array(
 216          *         'op'                  => 'move',
 217          *         'src'                 => <storage path>,
 218          *         'dst'                 => <storage path>,
 219          *         'overwrite'           => <boolean>,
 220          *         'overwriteSame'       => <boolean>,
 221          *         'disposition'         => <Content-Disposition header value>
 222          *     )
 223          * @endcode
 224          *
 225          * e) Delete a file within storage
 226          * @code
 227          *     array(
 228          *         'op'                  => 'delete',
 229          *         'src'                 => <storage path>,
 230          *         'ignoreMissingSource' => <boolean>
 231          *     )
 232          * @endcode
 233          *
 234          * f) Do nothing (no-op)
 235          * @code
 236          *     array(
 237          *         'op'                  => 'null',
 238          *     )
 239          * @endcode
 240          *
 241          * Boolean flags for operations (operation-specific):
 242          *   - ignoreMissingSource : The operation will simply succeed and do
 243          *                           nothing if the source file does not exist.
 244          *   - overwrite           : Any destination file will be overwritten.
 245          *   - overwriteSame       : An error will not be given if a file already
 246          *                           exists at the destination that has the same
 247          *                           contents as the new contents to be written there.
 248          *   - disposition         : When supplied, the backend will add a Content-Disposition
 249          *                           header when GETs/HEADs of the destination file are made.
 250          *                           Backends that don't support file metadata will ignore this.
 251          *                           See http://tools.ietf.org/html/rfc6266 (since 1.20).
 252          *
 253          * $opts is an associative of boolean flags, including:
 254          *   - force               : Operation precondition errors no longer trigger an abort.
 255          *                           Any remaining operations are still attempted. Unexpected
 256          *                           failures may still cause remaning operations to be aborted.
 257          *   - nonLocking          : No locks are acquired for the operations.
 258          *                           This can increase performance for non-critical writes.
 259          *                           This has no effect unless the 'force' flag is set.
 260          *   - allowStale          : Don't require the latest available data.
 261          *                           This can increase performance for non-critical writes.
 262          *                           This has no effect unless the 'force' flag is set.
 263          *   - nonJournaled        : Don't log this operation batch in the file journal.
 264          *                           This limits the ability of recovery scripts.
 265          *   - parallelize         : Try to do operations in parallel when possible.
 266          *   - bypassReadOnly      : Allow writes in read-only mode (since 1.20).
 267          *   - preserveCache       : Don't clear the process cache before checking files.
 268          *                           This should only be used if all entries in the process
 269          *                           cache were added after the files were already locked (since 1.20).
 270          *
 271          * @remarks Remarks on locking:
 272          * File system paths given to operations should refer to files that are
 273          * already locked or otherwise safe from modification from other processes.
 274          * Normally these files will be new temp files, which should be adequate.
 275          *
 276          * @par Return value:
 277          *
 278          * This returns a Status, which contains all warnings and fatals that occurred
 279          * during the operation. The 'failCount', 'successCount', and 'success' members
 280          * will reflect each operation attempted.
 281          *
 282          * The status will be "OK" unless:
 283          *   - a) unexpected operation errors occurred (network partitions, disk full...)
 284          *   - b) significant operation errors occurred and 'force' was not set
 285          *
 286          * @param $ops Array List of operations to execute in order
 287          * @param $opts Array Batch operation options
 288          * @return Status
 289          */
 290         final public function doOperations( array $ops, array $opts = array() ) {
 291                 if ( empty( $opts['bypassReadOnly'] ) && $this->isReadOnly() ) {
 292                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 293                 }
 294                 if ( empty( $opts['force'] ) ) { // sanity
 295                         unset( $opts['nonLocking'] );
 296                         unset( $opts['allowStale'] );
 297                 }
 298                 return $this->doOperationsInternal( $ops, $opts );
 299         }
 300
 301         /**
 302          * @see FileBackend::doOperations()
 303          */
 304         abstract protected function doOperationsInternal( array $ops, array $opts );
 305
 306         /**
 307          * Same as doOperations() except it takes a single operation.
 308          * If you are doing a batch of operations that should either
 309          * all succeed or all fail, then use that function instead.
 310          *
 311          * @see FileBackend::doOperations()
 312          *
 313          * @param $op Array Operation
 314          * @param $opts Array Operation options
 315          * @return Status
 316          */
 317         final public function doOperation( array $op, array $opts = array() ) {
 318                 return $this->doOperations( array( $op ), $opts );
 319         }
 320
 321         /**
 322          * Performs a single create operation.
 323          * This sets $params['op'] to 'create' and passes it to doOperation().
 324          *
 325          * @see FileBackend::doOperation()
 326          *
 327          * @param $params Array Operation parameters
 328          * @param $opts Array Operation options
 329          * @return Status
 330          */
 331         final public function create( array $params, array $opts = array() ) {
 332                 return $this->doOperation( array( 'op' => 'create' ) + $params, $opts );
 333         }
 334
 335         /**
 336          * Performs a single store operation.
 337          * This sets $params['op'] to 'store' and passes it to doOperation().
 338          *
 339          * @see FileBackend::doOperation()
 340          *
 341          * @param $params Array Operation parameters
 342          * @param $opts Array Operation options
 343          * @return Status
 344          */
 345         final public function store( array $params, array $opts = array() ) {
 346                 return $this->doOperation( array( 'op' => 'store' ) + $params, $opts );
 347         }
 348
 349         /**
 350          * Performs a single copy operation.
 351          * This sets $params['op'] to 'copy' and passes it to doOperation().
 352          *
 353          * @see FileBackend::doOperation()
 354          *
 355          * @param $params Array Operation parameters
 356          * @param $opts Array Operation options
 357          * @return Status
 358          */
 359         final public function copy( array $params, array $opts = array() ) {
 360                 return $this->doOperation( array( 'op' => 'copy' ) + $params, $opts );
 361         }
 362
 363         /**
 364          * Performs a single move operation.
 365          * This sets $params['op'] to 'move' and passes it to doOperation().
 366          *
 367          * @see FileBackend::doOperation()
 368          *
 369          * @param $params Array Operation parameters
 370          * @param $opts Array Operation options
 371          * @return Status
 372          */
 373         final public function move( array $params, array $opts = array() ) {
 374                 return $this->doOperation( array( 'op' => 'move' ) + $params, $opts );
 375         }
 376
 377         /**
 378          * Performs a single delete operation.
 379          * This sets $params['op'] to 'delete' and passes it to doOperation().
 380          *
 381          * @see FileBackend::doOperation()
 382          *
 383          * @param $params Array Operation parameters
 384          * @param $opts Array Operation options
 385          * @return Status
 386          */
 387         final public function delete( array $params, array $opts = array() ) {
 388                 return $this->doOperation( array( 'op' => 'delete' ) + $params, $opts );
 389         }
 390
 391         /**
 392          * Perform a set of independent file operations on some files.
 393          *
 394          * This does no locking, nor journaling, and possibly no stat calls.
 395          * Any destination files that already exist will be overwritten.
 396          * This should *only* be used on non-original files, like cache files.
 397          *
 398          * Supported operations and their parameters:
 399          *  - create
 400          *  - store
 401          *  - copy
 402          *  - move
 403          *  - delete
 404          *  - null
 405          *
 406          * a) Create a new file in storage with the contents of a string
 407          * @code
 408          *     array(
 409          *         'op'                  => 'create',
 410          *         'dst'                 => <storage path>,
 411          *         'content'             => <string of new file contents>,
 412          *         'disposition'         => <Content-Disposition header value>
 413          *     )
 414          * @endcode
 415          * b) Copy a file system file into storage
 416          * @code
 417          *     array(
 418          *         'op'                  => 'store',
 419          *         'src'                 => <file system path>,
 420          *         'dst'                 => <storage path>,
 421          *         'disposition'         => <Content-Disposition header value>
 422          *     )
 423          * @endcode
 424          * c) Copy a file within storage
 425          * @code
 426          *     array(
 427          *         'op'                  => 'copy',
 428          *         'src'                 => <storage path>,
 429          *         'dst'                 => <storage path>,
 430          *         'disposition'         => <Content-Disposition header value>
 431          *     )
 432          * @endcode
 433          * d) Move a file within storage
 434          * @code
 435          *     array(
 436          *         'op'                  => 'move',
 437          *         'src'                 => <storage path>,
 438          *         'dst'                 => <storage path>,
 439          *         'disposition'         => <Content-Disposition header value>
 440          *     )
 441          * @endcode
 442          * e) Delete a file within storage
 443          * @code
 444          *     array(
 445          *         'op'                  => 'delete',
 446          *         'src'                 => <storage path>,
 447          *         'ignoreMissingSource' => <boolean>
 448          *     )
 449          * @endcode
 450          * f) Do nothing (no-op)
 451          * @code
 452          *     array(
 453          *         'op'                  => 'null',
 454          *     )
 455          * @endcode
 456          *
 457          * @par Boolean flags for operations (operation-specific):
 458          *   - ignoreMissingSource : The operation will simply succeed and do
 459          *                           nothing if the source file does not exist.
 460          *   - disposition         : When supplied, the backend will add a Content-Disposition
 461          *                           header when GETs/HEADs of the destination file are made.
 462          *                           Backends that don't support file metadata will ignore this.
 463          *                           See http://tools.ietf.org/html/rfc6266 (since 1.20).
 464          *
 465          * $opts is an associative of boolean flags, including:
 466          *   - bypassReadOnly      : Allow writes in read-only mode (since 1.20)
 467          *
 468          * @par Return value:
 469          * This returns a Status, which contains all warnings and fatals that occurred
 470          * during the operation. The 'failCount', 'successCount', and 'success' members
 471          * will reflect each operation attempted for the given files. The status will be
 472          * considered "OK" as long as no fatal errors occurred.
 473          *
 474          * @param $ops Array Set of operations to execute
 475          * @param $opts Array Batch operation options
 476          * @return Status
 477          * @since 1.20
 478          */
 479         final public function doQuickOperations( array $ops, array $opts = array() ) {
 480                 if ( empty( $opts['bypassReadOnly'] ) && $this->isReadOnly() ) {
 481                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 482                 }
 483                 foreach ( $ops as &$op ) {
 484                         $op['overwrite'] = true; // avoids RTTs in key/value stores
 485                 }
 486                 return $this->doQuickOperationsInternal( $ops );
 487         }
 488
 489         /**
 490          * @see FileBackend::doQuickOperations()
 491          * @since 1.20
 492          */
 493         abstract protected function doQuickOperationsInternal( array $ops );
 494
 495         /**
 496          * Same as doQuickOperations() except it takes a single operation.
 497          * If you are doing a batch of operations, then use that function instead.
 498          *
 499          * @see FileBackend::doQuickOperations()
 500          *
 501          * @param $op Array Operation
 502          * @return Status
 503          * @since 1.20
 504          */
 505         final public function doQuickOperation( array $op ) {
 506                 return $this->doQuickOperations( array( $op ) );
 507         }
 508
 509         /**
 510          * Performs a single quick create operation.
 511          * This sets $params['op'] to 'create' and passes it to doQuickOperation().
 512          *
 513          * @see FileBackend::doQuickOperation()
 514          *
 515          * @param $params Array Operation parameters
 516          * @return Status
 517          * @since 1.20
 518          */
 519         final public function quickCreate( array $params ) {
 520                 return $this->doQuickOperation( array( 'op' => 'create' ) + $params );
 521         }
 522
 523         /**
 524          * Performs a single quick store operation.
 525          * This sets $params['op'] to 'store' and passes it to doQuickOperation().
 526          *
 527          * @see FileBackend::doQuickOperation()
 528          *
 529          * @param $params Array Operation parameters
 530          * @return Status
 531          * @since 1.20
 532          */
 533         final public function quickStore( array $params ) {
 534                 return $this->doQuickOperation( array( 'op' => 'store' ) + $params );
 535         }
 536
 537         /**
 538          * Performs a single quick copy operation.
 539          * This sets $params['op'] to 'copy' and passes it to doQuickOperation().
 540          *
 541          * @see FileBackend::doQuickOperation()
 542          *
 543          * @param $params Array Operation parameters
 544          * @return Status
 545          * @since 1.20
 546          */
 547         final public function quickCopy( array $params ) {
 548                 return $this->doQuickOperation( array( 'op' => 'copy' ) + $params );
 549         }
 550
 551         /**
 552          * Performs a single quick move operation.
 553          * This sets $params['op'] to 'move' and passes it to doQuickOperation().
 554          *
 555          * @see FileBackend::doQuickOperation()
 556          *
 557          * @param $params Array Operation parameters
 558          * @return Status
 559          * @since 1.20
 560          */
 561         final public function quickMove( array $params ) {
 562                 return $this->doQuickOperation( array( 'op' => 'move' ) + $params );
 563         }
 564
 565         /**
 566          * Performs a single quick delete operation.
 567          * This sets $params['op'] to 'delete' and passes it to doQuickOperation().
 568          *
 569          * @see FileBackend::doQuickOperation()
 570          *
 571          * @param $params Array Operation parameters
 572          * @return Status
 573          * @since 1.20
 574          */
 575         final public function quickDelete( array $params ) {
 576                 return $this->doQuickOperation( array( 'op' => 'delete' ) + $params );
 577         }
 578
 579         /**
 580          * Concatenate a list of storage files into a single file system file.
 581          * The target path should refer to a file that is already locked or
 582          * otherwise safe from modification from other processes. Normally,
 583          * the file will be a new temp file, which should be adequate.
 584          *
 585          * @param $params Array Operation parameters
 586          * $params include:
 587          *   - srcs : ordered source storage paths (e.g. chunk1, chunk2, ...)
 588          *   - dst  : file system path to 0-byte temp file
 589          * @return Status
 590          */
 591         abstract public function concatenate( array $params );
 592
 593         /**
 594          * Prepare a storage directory for usage.
 595          * This will create any required containers and parent directories.
 596          * Backends using key/value stores only need to create the container.
 597          *
 598          * The 'noAccess' and 'noListing' parameters works the same as in secure(),
 599          * except they are only applied *if* the directory/container had to be created.
 600          * These flags should always be set for directories that have private files.
 601          *
 602          * @param $params Array
 603          * $params include:
 604          *   - dir            : storage directory
 605          *   - noAccess       : try to deny file access (since 1.20)
 606          *   - noListing      : try to deny file listing (since 1.20)
 607          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 608          * @return Status
 609          */
 610         final public function prepare( array $params ) {
 611                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 612                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 613                 }
 614                 return $this->doPrepare( $params );
 615         }
 616
 617         /**
 618          * @see FileBackend::prepare()
 619          */
 620         abstract protected function doPrepare( array $params );
 621
 622         /**
 623          * Take measures to block web access to a storage directory and
 624          * the container it belongs to. FS backends might add .htaccess
 625          * files whereas key/value store backends might revoke container
 626          * access to the storage user representing end-users in web requests.
 627          * This is not guaranteed to actually do anything.
 628          *
 629          * @param $params Array
 630          * $params include:
 631          *   - dir            : storage directory
 632          *   - noAccess       : try to deny file access
 633          *   - noListing      : try to deny file listing
 634          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 635          * @return Status
 636          */
 637         final public function secure( array $params ) {
 638                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 639                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 640                 }
 641                 return $this->doSecure( $params );
 642         }
 643
 644         /**
 645          * @see FileBackend::secure()
 646          */
 647         abstract protected function doSecure( array $params );
 648
 649         /**
 650          * Remove measures to block web access to a storage directory and
 651          * the container it belongs to. FS backends might remove .htaccess
 652          * files whereas key/value store backends might grant container
 653          * access to the storage user representing end-users in web requests.
 654          * This essentially can undo the result of secure() calls.
 655          *
 656          * @param $params Array
 657          * $params include:
 658          *   - dir            : storage directory
 659          *   - access         : try to allow file access
 660          *   - listing        : try to allow file listing
 661          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 662          * @return Status
 663          * @since 1.20
 664          */
 665         final public function publish( array $params ) {
 666                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 667                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 668                 }
 669                 return $this->doPublish( $params );
 670         }
 671
 672         /**
 673          * @see FileBackend::publish()
 674          */
 675         abstract protected function doPublish( array $params );
 676
 677         /**
 678          * Delete a storage directory if it is empty.
 679          * Backends using key/value stores may do nothing unless the directory
 680          * is that of an empty container, in which case it will be deleted.
 681          *
 682          * @param $params Array
 683          * $params include:
 684          *   - dir            : storage directory
 685          *   - recursive      : recursively delete empty subdirectories first (since 1.20)
 686          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 687          * @return Status
 688          */
 689         final public function clean( array $params ) {
 690                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 691                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 692                 }
 693                 return $this->doClean( $params );
 694         }
 695
 696         /**
 697          * @see FileBackend::clean()
 698          */
 699         abstract protected function doClean( array $params );
 700
 701         /**
 702          * Check if a file exists at a storage path in the backend.
 703          * This returns false if only a directory exists at the path.
 704          *
 705          * @param $params Array
 706          * $params include:
 707          *   - src    : source storage path
 708          *   - latest : use the latest available data
 709          * @return bool|null Returns null on failure
 710          */
 711         abstract public function fileExists( array $params );
 712
 713         /**
 714          * Get the last-modified timestamp of the file at a storage path.
 715          *
 716          * @param $params Array
 717          * $params include:
 718          *   - src    : source storage path
 719          *   - latest : use the latest available data
 720          * @return string|bool TS_MW timestamp or false on failure
 721          */
 722         abstract public function getFileTimestamp( array $params );
 723
 724         /**
 725          * Get the contents of a file at a storage path in the backend.
 726          * This should be avoided for potentially large files.
 727          *
 728          * @param $params Array
 729          * $params include:
 730          *   - src    : source storage path
 731          *   - latest : use the latest available data
 732          * @return string|bool Returns false on failure
 733          */
 734         abstract public function getFileContents( array $params );
 735
 736         /**
 737          * Get the size (bytes) of a file at a storage path in the backend.
 738          *
 739          * @param $params Array
 740          * $params include:
 741          *   - src    : source storage path
 742          *   - latest : use the latest available data
 743          * @return integer|bool Returns false on failure
 744          */
 745         abstract public function getFileSize( array $params );
 746
 747         /**
 748          * Get quick information about a file at a storage path in the backend.
 749          * If the file does not exist, then this returns false.
 750          * Otherwise, the result is an associative array that includes:
 751          *   - mtime  : the last-modified timestamp (TS_MW)
 752          *   - size   : the file size (bytes)
 753          * Additional values may be included for internal use only.
 754          *
 755          * @param $params Array
 756          * $params include:
 757          *   - src    : source storage path
 758          *   - latest : use the latest available data
 759          * @return Array|bool|null Returns null on failure
 760          */
 761         abstract public function getFileStat( array $params );
 762
 763         /**
 764          * Get a SHA-1 hash of the file at a storage path in the backend.
 765          *
 766          * @param $params Array
 767          * $params include:
 768          *   - src    : source storage path
 769          *   - latest : use the latest available data
 770          * @return string|bool Hash string or false on failure
 771          */
 772         abstract public function getFileSha1Base36( array $params );
 773
 774         /**
 775          * Get the properties of the file at a storage path in the backend.
 776          * Returns FSFile::placeholderProps() on failure.
 777          *
 778          * @param $params Array
 779          * $params include:
 780          *   - src    : source storage path
 781          *   - latest : use the latest available data
 782          * @return Array
 783          */
 784         abstract public function getFileProps( array $params );
 785
 786         /**
 787          * Stream the file at a storage path in the backend.
 788          * If the file does not exists, an HTTP 404 error will be given.
 789          * Appropriate HTTP headers (Status, Content-Type, Content-Length)
 790          * will be sent if streaming began, while none will be sent otherwise.
 791          * Implementations should flush the output buffer before sending data.
 792          *
 793          * @param $params Array
 794          * $params include:
 795          *   - src     : source storage path
 796          *   - headers : list of additional HTTP headers to send on success
 797          *   - latest  : use the latest available data
 798          * @return Status
 799          */
 800         abstract public function streamFile( array $params );
 801
 802         /**
 803          * Returns a file system file, identical to the file at a storage path.
 804          * The file returned is either:
 805          *   - a) A local copy of the file at a storage path in the backend.
 806          *        The temporary copy will have the same extension as the source.
 807          *   - b) An original of the file at a storage path in the backend.
 808          * Temporary files may be purged when the file object falls out of scope.
 809          *
 810          * Write operations should *never* be done on this file as some backends
 811          * may do internal tracking or may be instances of FileBackendMultiWrite.
 812          * In that later case, there are copies of the file that must stay in sync.
 813          * Additionally, further calls to this function may return the same file.
 814          *
 815          * @param $params Array
 816          * $params include:
 817          *   - src    : source storage path
 818          *   - latest : use the latest available data
 819          * @return FSFile|null Returns null on failure
 820          */
 821         abstract public function getLocalReference( array $params );
 822
 823         /**
 824          * Get a local copy on disk of the file at a storage path in the backend.
 825          * The temporary copy will have the same file extension as the source.
 826          * Temporary files may be purged when the file object falls out of scope.
 827          *
 828          * @param $params Array
 829          * $params include:
 830          *   - src    : source storage path
 831          *   - latest : use the latest available data
 832          * @return TempFSFile|null Returns null on failure
 833          */
 834         abstract public function getLocalCopy( array $params );
 835
 836         /**
 837          * Check if a directory exists at a given storage path.
 838          * Backends using key/value stores will check if the path is a
 839          * virtual directory, meaning there are files under the given directory.
 840          *
 841          * Storage backends with eventual consistency might return stale data.
 842          *
 843          * @param $params array
 844          * $params include:
 845          *   - dir : storage directory
 846          * @return bool|null Returns null on failure
 847          * @since 1.20
 848          */
 849         abstract public function directoryExists( array $params );
 850
 851         /**
 852          * Get an iterator to list *all* directories under a storage directory.
 853          * If the directory is of the form "mwstore://backend/container",
 854          * then all directories in the container will be listed.
 855          * If the directory is of form "mwstore://backend/container/dir",
 856          * then all directories directly under that directory will be listed.
 857          * Results will be storage directories relative to the given directory.
 858          *
 859          * Storage backends with eventual consistency might return stale data.
 860          *
 861          * @param $params array
 862          * $params include:
 863          *   - dir     : storage directory
 864          *   - topOnly : only return direct child dirs of the directory
 865          * @return Traversable|Array|null Returns null on failure
 866          * @since 1.20
 867          */
 868         abstract public function getDirectoryList( array $params );
 869
 870         /**
 871          * Same as FileBackend::getDirectoryList() except only lists
 872          * directories that are immediately under the given directory.
 873          *
 874          * Storage backends with eventual consistency might return stale data.
 875          *
 876          * @param $params array
 877          * $params include:
 878          *   - dir : storage directory
 879          * @return Traversable|Array|null Returns null on failure
 880          * @since 1.20
 881          */
 882         final public function getTopDirectoryList( array $params ) {
 883                 return $this->getDirectoryList( array( 'topOnly' => true ) + $params );
 884         }
 885
 886         /**
 887          * Get an iterator to list *all* stored files under a storage directory.
 888          * If the directory is of the form "mwstore://backend/container",
 889          * then all files in the container will be listed.
 890          * If the directory is of form "mwstore://backend/container/dir",
 891          * then all files under that directory will be listed.
 892          * Results will be storage paths relative to the given directory.
 893          *
 894          * Storage backends with eventual consistency might return stale data.
 895          *
 896          * @param $params array
 897          * $params include:
 898          *   - dir     : storage directory
 899          *   - topOnly : only return direct child files of the directory (since 1.20)
 900          * @return Traversable|Array|null Returns null on failure
 901          */
 902         abstract public function getFileList( array $params );
 903
 904         /**
 905          * Same as FileBackend::getFileList() except only lists
 906          * files that are immediately under the given directory.
 907          *
 908          * Storage backends with eventual consistency might return stale data.
 909          *
 910          * @param $params array
 911          * $params include:
 912          *   - dir : storage directory
 913          * @return Traversable|Array|null Returns null on failure
 914          * @since 1.20
 915          */
 916         final public function getTopFileList( array $params ) {
 917                 return $this->getFileList( array( 'topOnly' => true ) + $params );
 918         }
 919
 920         /**
 921          * Preload persistent file stat and property cache into in-process cache.
 922          * This should be used when stat calls will be made on a known list of a many files.
 923          *
 924          * @param $paths Array Storage paths
 925          * @return void
 926          */
 927         public function preloadCache( array $paths ) {}
 928
 929         /**
 930          * Invalidate any in-process file stat and property cache.
 931          * If $paths is given, then only the cache for those files will be cleared.
 932          *
 933          * @param $paths Array Storage paths (optional)
 934          * @return void
 935          */
 936         public function clearCache( array $paths = null ) {}
 937
 938         /**
 939          * Lock the files at the given storage paths in the backend.
 940          * This will either lock all the files or none (on failure).
 941          *
 942          * Callers should consider using getScopedFileLocks() instead.
 943          *
 944          * @param $paths Array Storage paths
 945          * @param $type integer LockManager::LOCK_* constant
 946          * @return Status
 947          */
 948         final public function lockFiles( array $paths, $type ) {
 949                 return $this->lockManager->lock( $paths, $type );
 950         }
 951
 952         /**
 953          * Unlock the files at the given storage paths in the backend.
 954          *
 955          * @param $paths Array Storage paths
 956          * @param $type integer LockManager::LOCK_* constant
 957          * @return Status
 958          */
 959         final public function unlockFiles( array $paths, $type ) {
 960                 return $this->lockManager->unlock( $paths, $type );
 961         }
 962
 963         /**
 964          * Lock the files at the given storage paths in the backend.
 965          * This will either lock all the files or none (on failure).
 966          * On failure, the status object will be updated with errors.
 967          *
 968          * Once the return value goes out scope, the locks will be released and
 969          * the status updated. Unlock fatals will not change the status "OK" value.
 970          *
 971          * @param $paths Array Storage paths
 972          * @param $type integer LockManager::LOCK_* constant
 973          * @param $status Status Status to update on lock/unlock
 974          * @return ScopedLock|null Returns null on failure
 975          */
 976         final public function getScopedFileLocks( array $paths, $type, Status $status ) {
 977                 return ScopedLock::factory( $this->lockManager, $paths, $type, $status );
 978         }
 979
 980         /**
 981          * Get an array of scoped locks needed for a batch of file operations.
 982          *
 983          * Normally, FileBackend::doOperations() handles locking, unless
 984          * the 'nonLocking' param is passed in. This function is useful if you
 985          * want the files to be locked for a broader scope than just when the
 986          * files are changing. For example, if you need to update DB metadata,
 987          * you may want to keep the files locked until finished.
 988          *
 989          * @see FileBackend::doOperations()
 990          *
 991          * @param $ops Array List of file operations to FileBackend::doOperations()
 992          * @param $status Status Status to update on lock/unlock
 993          * @return Array List of ScopedFileLocks or null values
 994          * @since 1.20
 995          */
 996         abstract public function getScopedLocksForOps( array $ops, Status $status );
 997
 998         /**
 999          * Get the root storage path of this backend.
1000          * All container paths are "subdirectories" of this path.
1001          *
1002          * @return string Storage path
1003          * @since 1.20
1004          */
1005         final public function getRootStoragePath() {
1006                 return "mwstore://{$this->name}";
1007         }
1008
1009         /**
1010          * Get the file journal object for this backend
1011          *
1012          * @return FileJournal
1013          */
1014         final public function getJournal() {
1015                 return $this->fileJournal;
1016         }
1017
1018         /**
1019          * Check if a given path is a "mwstore://" path.
1020          * This does not do any further validation or any existence checks.
1021          *
1022          * @param $path string
1023          * @return bool
1024          */
1025         final public static function isStoragePath( $path ) {
1026                 return ( strpos( $path, 'mwstore://' ) === 0 );
1027         }
1028
1029         /**
1030          * Split a storage path into a backend name, a container name,
1031          * and a relative file path. The relative path may be the empty string.
1032          * This does not do any path normalization or traversal checks.
1033          *
1034          * @param $storagePath string
1035          * @return Array (backend, container, rel object) or (null, null, null)
1036          */
1037         final public static function splitStoragePath( $storagePath ) {
1038                 if ( self::isStoragePath( $storagePath ) ) {
1039                         // Remove the "mwstore://" prefix and split the path
1040                         $parts = explode( '/', substr( $storagePath, 10 ), 3 );
1041                         if ( count( $parts ) >= 2 && $parts[0] != '' && $parts[1] != '' ) {
1042                                 if ( count( $parts ) == 3 ) {
1043                                         return $parts; // e.g. "backend/container/path"
1044                                 } else {
1045                                         return array( $parts[0], $parts[1], '' ); // e.g. "backend/container"
1046                                 }
1047                         }
1048                 }
1049                 return array( null, null, null );
1050         }
1051
1052         /**
1053          * Normalize a storage path by cleaning up directory separators.
1054          * Returns null if the path is not of the format of a valid storage path.
1055          *
1056          * @param $storagePath string
1057          * @return string|null
1058          */
1059         final public static function normalizeStoragePath( $storagePath ) {
1060                 list( $backend, $container, $relPath ) = self::splitStoragePath( $storagePath );
1061                 if ( $relPath !== null ) { // must be for this backend
1062                         $relPath = self::normalizeContainerPath( $relPath );
1063                         if ( $relPath !== null ) {
1064                                 return ( $relPath != '' )
1065                                         ? "mwstore://{$backend}/{$container}/{$relPath}"
1066                                         : "mwstore://{$backend}/{$container}";
1067                         }
1068                 }
1069                 return null;
1070         }
1071
1072         /**
1073          * Get the parent storage directory of a storage path.
1074          * This returns a path like "mwstore://backend/container",
1075          * "mwstore://backend/container/...", or null if there is no parent.
1076          *
1077          * @param $storagePath string
1078          * @return string|null
1079          */
1080         final public static function parentStoragePath( $storagePath ) {
1081                 $storagePath = dirname( $storagePath );
1082                 list( $b, $cont, $rel ) = self::splitStoragePath( $storagePath );
1083                 return ( $rel === null ) ? null : $storagePath;
1084         }
1085
1086         /**
1087          * Get the final extension from a storage or FS path
1088          *
1089          * @param $path string
1090          * @return string
1091          */
1092         final public static function extensionFromPath( $path ) {
1093                 $i = strrpos( $path, '.' );
1094                 return strtolower( $i ? substr( $path, $i + 1 ) : '' );
1095         }
1096
1097         /**
1098          * Check if a relative path has no directory traversals
1099          *
1100          * @param $path string
1101          * @return bool
1102          * @since 1.20
1103          */
1104         final public static function isPathTraversalFree( $path ) {
1105                 return ( self::normalizeContainerPath( $path ) !== null );
1106         }
1107
1108         /**
1109          * Build a Content-Disposition header value per RFC 6266.
1110          *
1111          * @param $type string One of (attachment, inline)
1112          * @param $filename string Suggested file name (should not contain slashes)
1113          * @return string
1114          * @since 1.20
1115          */
1116         final public static function makeContentDisposition( $type, $filename = '' ) {
1117                 $parts = array();
1118
1119                 $type = strtolower( $type );
1120                 if ( !in_array( $type, array( 'inline', 'attachment' ) ) ) {
1121                         throw new MWException( "Invalid Content-Disposition type '$type'." );
1122                 }
1123                 $parts[] = $type;
1124
1125                 if ( strlen( $filename ) ) {
1126                         $parts[] = "filename*=UTF-8''" . rawurlencode( basename( $filename ) );
1127                 }
1128
1129                 return implode( ';', $parts );
1130         }
1131
1132         /**
1133          * Validate and normalize a relative storage path.
1134          * Null is returned if the path involves directory traversal.
1135          * Traversal is insecure for FS backends and broken for others.
1136          *
1137          * This uses the same traversal protection as Title::secureAndSplit().
1138          *
1139          * @param $path string Storage path relative to a container
1140          * @return string|null
1141          */
1142         final protected static function normalizeContainerPath( $path ) {
1143                 // Normalize directory separators
1144                 $path = strtr( $path, '\\', '/' );
1145                 // Collapse any consecutive directory separators
1146                 $path = preg_replace( '![/]{2,}!', '/', $path );
1147                 // Remove any leading directory separator
1148                 $path = ltrim( $path, '/' );
1149                 // Use the same traversal protection as Title::secureAndSplit()
1150                 if ( strpos( $path, '.' ) !== false ) {
1151                         if (
1152                                 $path === '.' ||
1153                                 $path === '..' ||
1154                                 strpos( $path, './' ) === 0 ||
1155                                 strpos( $path, '../' ) === 0 ||
1156                                 strpos( $path, '/./' ) !== false ||
1157                                 strpos( $path, '/../' ) !== false
1158                         ) {
1159                                 return null;
1160                         }
1161                 }
1162                 return $path;
1163         }
1164 }