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          * Check if this backend is read-only
 133          *
 134          * @return bool
 135          */
 136         final public function isReadOnly() {
 137                 return ( $this->readOnly != '' );
 138         }
 139
 140         /**
 141          * Get an explanatory message if this backend is read-only
 142          *
 143          * @return string|bool Returns false if the backend is not read-only
 144          */
 145         final public function getReadOnlyReason() {
 146                 return ( $this->readOnly != '' ) ? $this->readOnly : false;
 147         }
 148
 149         /**
 150          * This is the main entry point into the backend for write operations.
 151          * Callers supply an ordered list of operations to perform as a transaction.
 152          * Files will be locked, the stat cache cleared, and then the operations attempted.
 153          * If any serious errors occur, all attempted operations will be rolled back.
 154          *
 155          * $ops is an array of arrays. The outer array holds a list of operations.
 156          * Each inner array is a set of key value pairs that specify an operation.
 157          *
 158          * Supported operations and their parameters. The supported actions are:
 159          *  - create
 160          *  - store
 161          *  - copy
 162          *  - move
 163          *  - delete
 164          *  - null
 165          *
 166          * a) Create a new file in storage with the contents of a string
 167          * @code
 168          *     array(
 169          *         'op'                  => 'create',
 170          *         'dst'                 => <storage path>,
 171          *         'content'             => <string of new file contents>,
 172          *         'overwrite'           => <boolean>,
 173          *         'overwriteSame'       => <boolean>,
 174          *         'disposition'         => <Content-Disposition header value>
 175          *     );
 176          * @endcode
 177          *
 178          * b) Copy a file system file into storage
 179          * @code
 180          *     array(
 181          *         'op'                  => 'store',
 182          *         'src'                 => <file system path>,
 183          *         'dst'                 => <storage path>,
 184          *         'overwrite'           => <boolean>,
 185          *         'overwriteSame'       => <boolean>,
 186          *         'disposition'         => <Content-Disposition header value>
 187          *     )
 188          * @endcode
 189          *
 190          * c) Copy a file within storage
 191          * @code
 192          *     array(
 193          *         'op'                  => 'copy',
 194          *         'src'                 => <storage path>,
 195          *         'dst'                 => <storage path>,
 196          *         'overwrite'           => <boolean>,
 197          *         'overwriteSame'       => <boolean>,
 198          *         'disposition'         => <Content-Disposition header value>
 199          *     )
 200          * @endcode
 201          *
 202          * d) Move a file within storage
 203          * @code
 204          *     array(
 205          *         'op'                  => 'move',
 206          *         'src'                 => <storage path>,
 207          *         'dst'                 => <storage path>,
 208          *         'overwrite'           => <boolean>,
 209          *         'overwriteSame'       => <boolean>,
 210          *         'disposition'         => <Content-Disposition header value>
 211          *     )
 212          * @endcode
 213          *
 214          * e) Delete a file within storage
 215          * @code
 216          *     array(
 217          *         'op'                  => 'delete',
 218          *         'src'                 => <storage path>,
 219          *         'ignoreMissingSource' => <boolean>
 220          *     )
 221          * @endcode
 222          *
 223          * f) Do nothing (no-op)
 224          * @code
 225          *     array(
 226          *         'op'                  => 'null',
 227          *     )
 228          * @endcode
 229          *
 230          * Boolean flags for operations (operation-specific):
 231          *   - ignoreMissingSource : The operation will simply succeed and do
 232          *                           nothing if the source file does not exist.
 233          *   - overwrite           : Any destination file will be overwritten.
 234          *   - overwriteSame       : An error will not be given if a file already
 235          *                           exists at the destination that has the same
 236          *                           contents as the new contents to be written there.
 237          *   - disposition         : When supplied, the backend will add a Content-Disposition
 238          *                           header when GETs/HEADs of the destination file are made.
 239          *                           Backends that don't support file metadata will ignore this.
 240          *                           See http://tools.ietf.org/html/rfc6266 (since 1.20).
 241          *
 242          * $opts is an associative of boolean flags, including:
 243          *   - force               : Operation precondition errors no longer trigger an abort.
 244          *                           Any remaining operations are still attempted. Unexpected
 245          *                           failures may still cause remaning operations to be aborted.
 246          *   - nonLocking          : No locks are acquired for the operations.
 247          *                           This can increase performance for non-critical writes.
 248          *                           This has no effect unless the 'force' flag is set.
 249          *   - allowStale          : Don't require the latest available data.
 250          *                           This can increase performance for non-critical writes.
 251          *                           This has no effect unless the 'force' flag is set.
 252          *   - nonJournaled        : Don't log this operation batch in the file journal.
 253          *                           This limits the ability of recovery scripts.
 254          *   - parallelize         : Try to do operations in parallel when possible.
 255          *   - bypassReadOnly      : Allow writes in read-only mode (since 1.20).
 256          *   - preserveCache       : Don't clear the process cache before checking files.
 257          *                           This should only be used if all entries in the process
 258          *                           cache were added after the files were already locked (since 1.20).
 259          *
 260          * @remarks Remarks on locking:
 261          * File system paths given to operations should refer to files that are
 262          * already locked or otherwise safe from modification from other processes.
 263          * Normally these files will be new temp files, which should be adequate.
 264          *
 265          * @par Return value:
 266          *
 267          * This returns a Status, which contains all warnings and fatals that occurred
 268          * during the operation. The 'failCount', 'successCount', and 'success' members
 269          * will reflect each operation attempted.
 270          *
 271          * The status will be "OK" unless:
 272          *   - a) unexpected operation errors occurred (network partitions, disk full...)
 273          *   - b) significant operation errors occurred and 'force' was not set
 274          *
 275          * @param $ops Array List of operations to execute in order
 276          * @param $opts Array Batch operation options
 277          * @return Status
 278          */
 279         final public function doOperations( array $ops, array $opts = array() ) {
 280                 if ( empty( $opts['bypassReadOnly'] ) && $this->isReadOnly() ) {
 281                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 282                 }
 283                 if ( empty( $opts['force'] ) ) { // sanity
 284                         unset( $opts['nonLocking'] );
 285                         unset( $opts['allowStale'] );
 286                 }
 287                 return $this->doOperationsInternal( $ops, $opts );
 288         }
 289
 290         /**
 291          * @see FileBackend::doOperations()
 292          */
 293         abstract protected function doOperationsInternal( array $ops, array $opts );
 294
 295         /**
 296          * Same as doOperations() except it takes a single operation.
 297          * If you are doing a batch of operations that should either
 298          * all succeed or all fail, then use that function instead.
 299          *
 300          * @see FileBackend::doOperations()
 301          *
 302          * @param $op Array Operation
 303          * @param $opts Array Operation options
 304          * @return Status
 305          */
 306         final public function doOperation( array $op, array $opts = array() ) {
 307                 return $this->doOperations( array( $op ), $opts );
 308         }
 309
 310         /**
 311          * Performs a single create operation.
 312          * This sets $params['op'] to 'create' and passes it to doOperation().
 313          *
 314          * @see FileBackend::doOperation()
 315          *
 316          * @param $params Array Operation parameters
 317          * @param $opts Array Operation options
 318          * @return Status
 319          */
 320         final public function create( array $params, array $opts = array() ) {
 321                 return $this->doOperation( array( 'op' => 'create' ) + $params, $opts );
 322         }
 323
 324         /**
 325          * Performs a single store operation.
 326          * This sets $params['op'] to 'store' and passes it to doOperation().
 327          *
 328          * @see FileBackend::doOperation()
 329          *
 330          * @param $params Array Operation parameters
 331          * @param $opts Array Operation options
 332          * @return Status
 333          */
 334         final public function store( array $params, array $opts = array() ) {
 335                 return $this->doOperation( array( 'op' => 'store' ) + $params, $opts );
 336         }
 337
 338         /**
 339          * Performs a single copy operation.
 340          * This sets $params['op'] to 'copy' and passes it to doOperation().
 341          *
 342          * @see FileBackend::doOperation()
 343          *
 344          * @param $params Array Operation parameters
 345          * @param $opts Array Operation options
 346          * @return Status
 347          */
 348         final public function copy( array $params, array $opts = array() ) {
 349                 return $this->doOperation( array( 'op' => 'copy' ) + $params, $opts );
 350         }
 351
 352         /**
 353          * Performs a single move operation.
 354          * This sets $params['op'] to 'move' and passes it to doOperation().
 355          *
 356          * @see FileBackend::doOperation()
 357          *
 358          * @param $params Array Operation parameters
 359          * @param $opts Array Operation options
 360          * @return Status
 361          */
 362         final public function move( array $params, array $opts = array() ) {
 363                 return $this->doOperation( array( 'op' => 'move' ) + $params, $opts );
 364         }
 365
 366         /**
 367          * Performs a single delete operation.
 368          * This sets $params['op'] to 'delete' and passes it to doOperation().
 369          *
 370          * @see FileBackend::doOperation()
 371          *
 372          * @param $params Array Operation parameters
 373          * @param $opts Array Operation options
 374          * @return Status
 375          */
 376         final public function delete( array $params, array $opts = array() ) {
 377                 return $this->doOperation( array( 'op' => 'delete' ) + $params, $opts );
 378         }
 379
 380         /**
 381          * Perform a set of independent file operations on some files.
 382          *
 383          * This does no locking, nor journaling, and possibly no stat calls.
 384          * Any destination files that already exist will be overwritten.
 385          * This should *only* be used on non-original files, like cache files.
 386          *
 387          * Supported operations and their parameters:
 388          *  - create
 389          *  - store
 390          *  - copy
 391          *  - move
 392          *  - delete
 393          *  - null
 394          *
 395          * a) Create a new file in storage with the contents of a string
 396          * @code
 397          *     array(
 398          *         'op'                  => 'create',
 399          *         'dst'                 => <storage path>,
 400          *         'content'             => <string of new file contents>,
 401          *         'disposition'         => <Content-Disposition header value>
 402          *     )
 403          * @endcode
 404          * b) Copy a file system file into storage
 405          * @code
 406          *     array(
 407          *         'op'                  => 'store',
 408          *         'src'                 => <file system path>,
 409          *         'dst'                 => <storage path>,
 410          *         'disposition'         => <Content-Disposition header value>
 411          *     )
 412          * @endcode
 413          * c) Copy a file within storage
 414          * @code
 415          *     array(
 416          *         'op'                  => 'copy',
 417          *         'src'                 => <storage path>,
 418          *         'dst'                 => <storage path>,
 419          *         'disposition'         => <Content-Disposition header value>
 420          *     )
 421          * @endcode
 422          * d) Move a file within storage
 423          * @code
 424          *     array(
 425          *         'op'                  => 'move',
 426          *         'src'                 => <storage path>,
 427          *         'dst'                 => <storage path>,
 428          *         'disposition'         => <Content-Disposition header value>
 429          *     )
 430          * @endcode
 431          * e) Delete a file within storage
 432          * @code
 433          *     array(
 434          *         'op'                  => 'delete',
 435          *         'src'                 => <storage path>,
 436          *         'ignoreMissingSource' => <boolean>
 437          *     )
 438          * @endcode
 439          * f) Do nothing (no-op)
 440          * @code
 441          *     array(
 442          *         'op'                  => 'null',
 443          *     )
 444          * @endcode
 445          *
 446          * @par Boolean flags for operations (operation-specific):
 447          *   - ignoreMissingSource : The operation will simply succeed and do
 448          *                           nothing if the source file does not exist.
 449          *   - disposition         : When supplied, the backend will add a Content-Disposition
 450          *                           header when GETs/HEADs of the destination file are made.
 451          *                           Backends that don't support file metadata will ignore this.
 452          *                           See http://tools.ietf.org/html/rfc6266 (since 1.20).
 453          *
 454          * $opts is an associative of boolean flags, including:
 455          *   - bypassReadOnly      : Allow writes in read-only mode (since 1.20)
 456          *
 457          * @par Return value:
 458          * This returns a Status, which contains all warnings and fatals that occurred
 459          * during the operation. The 'failCount', 'successCount', and 'success' members
 460          * will reflect each operation attempted for the given files. The status will be
 461          * considered "OK" as long as no fatal errors occurred.
 462          *
 463          * @param $ops Array Set of operations to execute
 464          * @param $opts Array Batch operation options
 465          * @return Status
 466          * @since 1.20
 467          */
 468         final public function doQuickOperations( array $ops, array $opts = array() ) {
 469                 if ( empty( $opts['bypassReadOnly'] ) && $this->isReadOnly() ) {
 470                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 471                 }
 472                 foreach ( $ops as &$op ) {
 473                         $op['overwrite'] = true; // avoids RTTs in key/value stores
 474                 }
 475                 return $this->doQuickOperationsInternal( $ops );
 476         }
 477
 478         /**
 479          * @see FileBackend::doQuickOperations()
 480          * @since 1.20
 481          */
 482         abstract protected function doQuickOperationsInternal( array $ops );
 483
 484         /**
 485          * Same as doQuickOperations() except it takes a single operation.
 486          * If you are doing a batch of operations, then use that function instead.
 487          *
 488          * @see FileBackend::doQuickOperations()
 489          *
 490          * @param $op Array Operation
 491          * @return Status
 492          * @since 1.20
 493          */
 494         final public function doQuickOperation( array $op ) {
 495                 return $this->doQuickOperations( array( $op ) );
 496         }
 497
 498         /**
 499          * Performs a single quick create operation.
 500          * This sets $params['op'] to 'create' and passes it to doQuickOperation().
 501          *
 502          * @see FileBackend::doQuickOperation()
 503          *
 504          * @param $params Array Operation parameters
 505          * @return Status
 506          * @since 1.20
 507          */
 508         final public function quickCreate( array $params ) {
 509                 return $this->doQuickOperation( array( 'op' => 'create' ) + $params );
 510         }
 511
 512         /**
 513          * Performs a single quick store operation.
 514          * This sets $params['op'] to 'store' and passes it to doQuickOperation().
 515          *
 516          * @see FileBackend::doQuickOperation()
 517          *
 518          * @param $params Array Operation parameters
 519          * @return Status
 520          * @since 1.20
 521          */
 522         final public function quickStore( array $params ) {
 523                 return $this->doQuickOperation( array( 'op' => 'store' ) + $params );
 524         }
 525
 526         /**
 527          * Performs a single quick copy operation.
 528          * This sets $params['op'] to 'copy' and passes it to doQuickOperation().
 529          *
 530          * @see FileBackend::doQuickOperation()
 531          *
 532          * @param $params Array Operation parameters
 533          * @return Status
 534          * @since 1.20
 535          */
 536         final public function quickCopy( array $params ) {
 537                 return $this->doQuickOperation( array( 'op' => 'copy' ) + $params );
 538         }
 539
 540         /**
 541          * Performs a single quick move operation.
 542          * This sets $params['op'] to 'move' and passes it to doQuickOperation().
 543          *
 544          * @see FileBackend::doQuickOperation()
 545          *
 546          * @param $params Array Operation parameters
 547          * @return Status
 548          * @since 1.20
 549          */
 550         final public function quickMove( array $params ) {
 551                 return $this->doQuickOperation( array( 'op' => 'move' ) + $params );
 552         }
 553
 554         /**
 555          * Performs a single quick delete operation.
 556          * This sets $params['op'] to 'delete' and passes it to doQuickOperation().
 557          *
 558          * @see FileBackend::doQuickOperation()
 559          *
 560          * @param $params Array Operation parameters
 561          * @return Status
 562          * @since 1.20
 563          */
 564         final public function quickDelete( array $params ) {
 565                 return $this->doQuickOperation( array( 'op' => 'delete' ) + $params );
 566         }
 567
 568         /**
 569          * Concatenate a list of storage files into a single file system file.
 570          * The target path should refer to a file that is already locked or
 571          * otherwise safe from modification from other processes. Normally,
 572          * the file will be a new temp file, which should be adequate.
 573          *
 574          * @param $params Array Operation parameters
 575          * $params include:
 576          *   - srcs        : ordered source storage paths (e.g. chunk1, chunk2, ...)
 577          *   - dst         : file system path to 0-byte temp file
 578          *   - parallelize : try to do operations in parallel when possible
 579          * @return Status
 580          */
 581         abstract public function concatenate( array $params );
 582
 583         /**
 584          * Prepare a storage directory for usage.
 585          * This will create any required containers and parent directories.
 586          * Backends using key/value stores only need to create the container.
 587          *
 588          * The 'noAccess' and 'noListing' parameters works the same as in secure(),
 589          * except they are only applied *if* the directory/container had to be created.
 590          * These flags should always be set for directories that have private files.
 591          *
 592          * @param $params Array
 593          * $params include:
 594          *   - dir            : storage directory
 595          *   - noAccess       : try to deny file access (since 1.20)
 596          *   - noListing      : try to deny file listing (since 1.20)
 597          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 598          * @return Status
 599          */
 600         final public function prepare( array $params ) {
 601                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 602                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 603                 }
 604                 return $this->doPrepare( $params );
 605         }
 606
 607         /**
 608          * @see FileBackend::prepare()
 609          */
 610         abstract protected function doPrepare( array $params );
 611
 612         /**
 613          * Take measures to block web access to a storage directory and
 614          * the container it belongs to. FS backends might add .htaccess
 615          * files whereas key/value store backends might revoke container
 616          * access to the storage user representing end-users in web requests.
 617          * This is not guaranteed to actually do anything.
 618          *
 619          * @param $params Array
 620          * $params include:
 621          *   - dir            : storage directory
 622          *   - noAccess       : try to deny file access
 623          *   - noListing      : try to deny file listing
 624          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 625          * @return Status
 626          */
 627         final public function secure( array $params ) {
 628                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 629                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 630                 }
 631                 return $this->doSecure( $params );
 632         }
 633
 634         /**
 635          * @see FileBackend::secure()
 636          */
 637         abstract protected function doSecure( array $params );
 638
 639         /**
 640          * Remove measures to block web access to a storage directory and
 641          * the container it belongs to. FS backends might remove .htaccess
 642          * files whereas key/value store backends might grant container
 643          * access to the storage user representing end-users in web requests.
 644          * This essentially can undo the result of secure() calls.
 645          *
 646          * @param $params Array
 647          * $params include:
 648          *   - dir            : storage directory
 649          *   - access         : try to allow file access
 650          *   - listing        : try to allow file listing
 651          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 652          * @return Status
 653          * @since 1.20
 654          */
 655         final public function publish( array $params ) {
 656                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 657                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 658                 }
 659                 return $this->doPublish( $params );
 660         }
 661
 662         /**
 663          * @see FileBackend::publish()
 664          */
 665         abstract protected function doPublish( array $params );
 666
 667         /**
 668          * Delete a storage directory if it is empty.
 669          * Backends using key/value stores may do nothing unless the directory
 670          * is that of an empty container, in which case it will be deleted.
 671          *
 672          * @param $params Array
 673          * $params include:
 674          *   - dir            : storage directory
 675          *   - recursive      : recursively delete empty subdirectories first (since 1.20)
 676          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 677          * @return Status
 678          */
 679         final public function clean( array $params ) {
 680                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 681                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 682                 }
 683                 return $this->doClean( $params );
 684         }
 685
 686         /**
 687          * @see FileBackend::clean()
 688          */
 689         abstract protected function doClean( array $params );
 690
 691         /**
 692          * Check if a file exists at a storage path in the backend.
 693          * This returns false if only a directory exists at the path.
 694          *
 695          * @param $params Array
 696          * $params include:
 697          *   - src    : source storage path
 698          *   - latest : use the latest available data
 699          * @return bool|null Returns null on failure
 700          */
 701         abstract public function fileExists( array $params );
 702
 703         /**
 704          * Get the last-modified timestamp of the file at a storage path.
 705          *
 706          * @param $params Array
 707          * $params include:
 708          *   - src    : source storage path
 709          *   - latest : use the latest available data
 710          * @return string|bool TS_MW timestamp or false on failure
 711          */
 712         abstract public function getFileTimestamp( array $params );
 713
 714         /**
 715          * Get the contents of a file at a storage path in the backend.
 716          * This should be avoided for potentially large files.
 717          *
 718          * @param $params Array
 719          * $params include:
 720          *   - src    : source storage path
 721          *   - latest : use the latest available data
 722          * @return string|bool Returns false on failure
 723          */
 724         final public function getFileContents( array $params ) {
 725                 $contents = $this->getFileContentsMulti(
 726                         array( 'srcs' => array( $params['src'] ) ) + $params );
 727
 728                 return $contents[$params['src']];
 729         }
 730
 731         /**
 732          * Like getFileContents() except it takes an array of storage paths
 733          * and returns a map of storage paths to strings (or null on failure).
 734          * The map keys (paths) are in the same order as the provided list of paths.
 735          *
 736          * @see FileBackend::getFileContents()
 737          *
 738          * @param $params Array
 739          * $params include:
 740          *   - srcs        : list of source storage paths
 741          *   - latest      : use the latest available data
 742          *   - parallelize : try to do operations in parallel when possible
 743          * @return Array Map of (path name => string or false on failure)
 744          * @since 1.20
 745          */
 746         abstract public function getFileContentsMulti( array $params );
 747
 748         /**
 749          * Get the size (bytes) of a file at a storage path in the backend.
 750          *
 751          * @param $params Array
 752          * $params include:
 753          *   - src    : source storage path
 754          *   - latest : use the latest available data
 755          * @return integer|bool Returns false on failure
 756          */
 757         abstract public function getFileSize( array $params );
 758
 759         /**
 760          * Get quick information about a file at a storage path in the backend.
 761          * If the file does not exist, then this returns false.
 762          * Otherwise, the result is an associative array that includes:
 763          *   - mtime  : the last-modified timestamp (TS_MW)
 764          *   - size   : the file size (bytes)
 765          * Additional values may be included for internal use only.
 766          *
 767          * @param $params Array
 768          * $params include:
 769          *   - src    : source storage path
 770          *   - latest : use the latest available data
 771          * @return Array|bool|null Returns null on failure
 772          */
 773         abstract public function getFileStat( array $params );
 774
 775         /**
 776          * Get a SHA-1 hash of the file at a storage path in the backend.
 777          *
 778          * @param $params Array
 779          * $params include:
 780          *   - src    : source storage path
 781          *   - latest : use the latest available data
 782          * @return string|bool Hash string or false on failure
 783          */
 784         abstract public function getFileSha1Base36( array $params );
 785
 786         /**
 787          * Get the properties of the file at a storage path in the backend.
 788          * Returns FSFile::placeholderProps() on failure.
 789          *
 790          * @param $params Array
 791          * $params include:
 792          *   - src    : source storage path
 793          *   - latest : use the latest available data
 794          * @return Array
 795          */
 796         abstract public function getFileProps( array $params );
 797
 798         /**
 799          * Stream the file at a storage path in the backend.
 800          * If the file does not exists, an HTTP 404 error will be given.
 801          * Appropriate HTTP headers (Status, Content-Type, Content-Length)
 802          * will be sent if streaming began, while none will be sent otherwise.
 803          * Implementations should flush the output buffer before sending data.
 804          *
 805          * @param $params Array
 806          * $params include:
 807          *   - src     : source storage path
 808          *   - headers : list of additional HTTP headers to send on success
 809          *   - latest  : use the latest available data
 810          * @return Status
 811          */
 812         abstract public function streamFile( array $params );
 813
 814         /**
 815          * Returns a file system file, identical to the file at a storage path.
 816          * The file returned is either:
 817          *   - a) A local copy of the file at a storage path in the backend.
 818          *        The temporary copy will have the same extension as the source.
 819          *   - b) An original of the file at a storage path in the backend.
 820          * Temporary files may be purged when the file object falls out of scope.
 821          *
 822          * Write operations should *never* be done on this file as some backends
 823          * may do internal tracking or may be instances of FileBackendMultiWrite.
 824          * In that later case, there are copies of the file that must stay in sync.
 825          * Additionally, further calls to this function may return the same file.
 826          *
 827          * @param $params Array
 828          * $params include:
 829          *   - src    : source storage path
 830          *   - latest : use the latest available data
 831          * @return FSFile|null Returns null on failure
 832          */
 833         final public function getLocalReference( array $params ) {
 834                 $fsFiles = $this->getLocalReferenceMulti(
 835                         array( 'srcs' => array( $params['src'] ) ) + $params );
 836
 837                 return $fsFiles[$params['src']];
 838         }
 839
 840         /**
 841          * Like getLocalReference() except it takes an array of storage paths
 842          * and returns a map of storage paths to FSFile objects (or null on failure).
 843          * The map keys (paths) are in the same order as the provided list of paths.
 844          *
 845          * @see FileBackend::getLocalReference()
 846          *
 847          * @param $params Array
 848          * $params include:
 849          *   - srcs        : list of source storage paths
 850          *   - latest      : use the latest available data
 851          *   - parallelize : try to do operations in parallel when possible
 852          * @return Array Map of (path name => FSFile or null on failure)
 853          * @since 1.20
 854          */
 855         abstract public function getLocalReferenceMulti( array $params );
 856
 857         /**
 858          * Get a local copy on disk of the file at a storage path in the backend.
 859          * The temporary copy will have the same file extension as the source.
 860          * Temporary files may be purged when the file object falls out of scope.
 861          *
 862          * @param $params Array
 863          * $params include:
 864          *   - src    : source storage path
 865          *   - latest : use the latest available data
 866          * @return TempFSFile|null Returns null on failure
 867          */
 868         final public function getLocalCopy( array $params ) {
 869                 $tmpFiles = $this->getLocalCopyMulti(
 870                         array( 'srcs' => array( $params['src'] ) ) + $params );
 871
 872                 return $tmpFiles[$params['src']];
 873         }
 874
 875         /**
 876          * Like getLocalCopy() except it takes an array of storage paths and
 877          * returns a map of storage paths to TempFSFile objects (or null on failure).
 878          * The map keys (paths) are in the same order as the provided list of paths.
 879          *
 880          * @see FileBackend::getLocalCopy()
 881          *
 882          * @param $params Array
 883          * $params include:
 884          *   - srcs        : list of source storage paths
 885          *   - latest      : use the latest available data
 886          *   - parallelize : try to do operations in parallel when possible
 887          * @return Array Map of (path name => TempFSFile or null on failure)
 888          * @since 1.20
 889          */
 890         abstract public function getLocalCopyMulti( array $params );
 891
 892         /**
 893          * Check if a directory exists at a given storage path.
 894          * Backends using key/value stores will check if the path is a
 895          * virtual directory, meaning there are files under the given directory.
 896          *
 897          * Storage backends with eventual consistency might return stale data.
 898          *
 899          * @param $params array
 900          * $params include:
 901          *   - dir : storage directory
 902          * @return bool|null Returns null on failure
 903          * @since 1.20
 904          */
 905         abstract public function directoryExists( array $params );
 906
 907         /**
 908          * Get an iterator to list *all* directories under a storage directory.
 909          * If the directory is of the form "mwstore://backend/container",
 910          * then all directories in the container will be listed.
 911          * If the directory is of form "mwstore://backend/container/dir",
 912          * then all directories directly under that directory will be listed.
 913          * Results will be storage directories relative to the given directory.
 914          *
 915          * Storage backends with eventual consistency might return stale data.
 916          *
 917          * @param $params array
 918          * $params include:
 919          *   - dir     : storage directory
 920          *   - topOnly : only return direct child dirs of the directory
 921          * @return Traversable|Array|null Returns null on failure
 922          * @since 1.20
 923          */
 924         abstract public function getDirectoryList( array $params );
 925
 926         /**
 927          * Same as FileBackend::getDirectoryList() except only lists
 928          * directories that are immediately under the given directory.
 929          *
 930          * Storage backends with eventual consistency might return stale data.
 931          *
 932          * @param $params array
 933          * $params include:
 934          *   - dir : storage directory
 935          * @return Traversable|Array|null Returns null on failure
 936          * @since 1.20
 937          */
 938         final public function getTopDirectoryList( array $params ) {
 939                 return $this->getDirectoryList( array( 'topOnly' => true ) + $params );
 940         }
 941
 942         /**
 943          * Get an iterator to list *all* stored files under a storage directory.
 944          * If the directory is of the form "mwstore://backend/container",
 945          * then all files in the container will be listed.
 946          * If the directory is of form "mwstore://backend/container/dir",
 947          * then all files under that directory will be listed.
 948          * Results will be storage paths relative to the given directory.
 949          *
 950          * Storage backends with eventual consistency might return stale data.
 951          *
 952          * @param $params array
 953          * $params include:
 954          *   - dir     : storage directory
 955          *   - topOnly : only return direct child files of the directory (since 1.20)
 956          * @return Traversable|Array|null Returns null on failure
 957          */
 958         abstract public function getFileList( array $params );
 959
 960         /**
 961          * Same as FileBackend::getFileList() except only lists
 962          * files that are immediately under the given directory.
 963          *
 964          * Storage backends with eventual consistency might return stale data.
 965          *
 966          * @param $params array
 967          * $params include:
 968          *   - dir : storage directory
 969          * @return Traversable|Array|null Returns null on failure
 970          * @since 1.20
 971          */
 972         final public function getTopFileList( array $params ) {
 973                 return $this->getFileList( array( 'topOnly' => true ) + $params );
 974         }
 975
 976         /**
 977          * Preload persistent file stat and property cache into in-process cache.
 978          * This should be used when stat calls will be made on a known list of a many files.
 979          *
 980          * @param $paths Array Storage paths
 981          * @return void
 982          */
 983         public function preloadCache( array $paths ) {}
 984
 985         /**
 986          * Invalidate any in-process file stat and property cache.
 987          * If $paths is given, then only the cache for those files will be cleared.
 988          *
 989          * @param $paths Array Storage paths (optional)
 990          * @return void
 991          */
 992         public function clearCache( array $paths = null ) {}
 993
 994         /**
 995          * Lock the files at the given storage paths in the backend.
 996          * This will either lock all the files or none (on failure).
 997          *
 998          * Callers should consider using getScopedFileLocks() instead.
 999          *
1000          * @param $paths Array Storage paths
1001          * @param $type integer LockManager::LOCK_* constant
1002          * @return Status
1003          */
1004         final public function lockFiles( array $paths, $type ) {
1005                 return $this->lockManager->lock( $paths, $type );
1006         }
1007
1008         /**
1009          * Unlock the files at the given storage paths in the backend.
1010          *
1011          * @param $paths Array Storage paths
1012          * @param $type integer LockManager::LOCK_* constant
1013          * @return Status
1014          */
1015         final public function unlockFiles( array $paths, $type ) {
1016                 return $this->lockManager->unlock( $paths, $type );
1017         }
1018
1019         /**
1020          * Lock the files at the given storage paths in the backend.
1021          * This will either lock all the files or none (on failure).
1022          * On failure, the status object will be updated with errors.
1023          *
1024          * Once the return value goes out scope, the locks will be released and
1025          * the status updated. Unlock fatals will not change the status "OK" value.
1026          *
1027          * @param $paths Array Storage paths
1028          * @param $type integer LockManager::LOCK_* constant
1029          * @param $status Status Status to update on lock/unlock
1030          * @return ScopedLock|null Returns null on failure
1031          */
1032         final public function getScopedFileLocks( array $paths, $type, Status $status ) {
1033                 return ScopedLock::factory( $this->lockManager, $paths, $type, $status );
1034         }
1035
1036         /**
1037          * Get an array of scoped locks needed for a batch of file operations.
1038          *
1039          * Normally, FileBackend::doOperations() handles locking, unless
1040          * the 'nonLocking' param is passed in. This function is useful if you
1041          * want the files to be locked for a broader scope than just when the
1042          * files are changing. For example, if you need to update DB metadata,
1043          * you may want to keep the files locked until finished.
1044          *
1045          * @see FileBackend::doOperations()
1046          *
1047          * @param $ops Array List of file operations to FileBackend::doOperations()
1048          * @param $status Status Status to update on lock/unlock
1049          * @return Array List of ScopedFileLocks or null values
1050          * @since 1.20
1051          */
1052         abstract public function getScopedLocksForOps( array $ops, Status $status );
1053
1054         /**
1055          * Get the root storage path of this backend.
1056          * All container paths are "subdirectories" of this path.
1057          *
1058          * @return string Storage path
1059          * @since 1.20
1060          */
1061         final public function getRootStoragePath() {
1062                 return "mwstore://{$this->name}";
1063         }
1064
1065         /**
1066          * Get the file journal object for this backend
1067          *
1068          * @return FileJournal
1069          */
1070         final public function getJournal() {
1071                 return $this->fileJournal;
1072         }
1073
1074         /**
1075          * Check if a given path is a "mwstore://" path.
1076          * This does not do any further validation or any existence checks.
1077          *
1078          * @param $path string
1079          * @return bool
1080          */
1081         final public static function isStoragePath( $path ) {
1082                 return ( strpos( $path, 'mwstore://' ) === 0 );
1083         }
1084
1085         /**
1086          * Split a storage path into a backend name, a container name,
1087          * and a relative file path. The relative path may be the empty string.
1088          * This does not do any path normalization or traversal checks.
1089          *
1090          * @param $storagePath string
1091          * @return Array (backend, container, rel object) or (null, null, null)
1092          */
1093         final public static function splitStoragePath( $storagePath ) {
1094                 if ( self::isStoragePath( $storagePath ) ) {
1095                         // Remove the "mwstore://" prefix and split the path
1096                         $parts = explode( '/', substr( $storagePath, 10 ), 3 );
1097                         if ( count( $parts ) >= 2 && $parts[0] != '' && $parts[1] != '' ) {
1098                                 if ( count( $parts ) == 3 ) {
1099                                         return $parts; // e.g. "backend/container/path"
1100                                 } else {
1101                                         return array( $parts[0], $parts[1], '' ); // e.g. "backend/container"
1102                                 }
1103                         }
1104                 }
1105                 return array( null, null, null );
1106         }
1107
1108         /**
1109          * Normalize a storage path by cleaning up directory separators.
1110          * Returns null if the path is not of the format of a valid storage path.
1111          *
1112          * @param $storagePath string
1113          * @return string|null
1114          */
1115         final public static function normalizeStoragePath( $storagePath ) {
1116                 list( $backend, $container, $relPath ) = self::splitStoragePath( $storagePath );
1117                 if ( $relPath !== null ) { // must be for this backend
1118                         $relPath = self::normalizeContainerPath( $relPath );
1119                         if ( $relPath !== null ) {
1120                                 return ( $relPath != '' )
1121                                         ? "mwstore://{$backend}/{$container}/{$relPath}"
1122                                         : "mwstore://{$backend}/{$container}";
1123                         }
1124                 }
1125                 return null;
1126         }
1127
1128         /**
1129          * Get the parent storage directory of a storage path.
1130          * This returns a path like "mwstore://backend/container",
1131          * "mwstore://backend/container/...", or null if there is no parent.
1132          *
1133          * @param $storagePath string
1134          * @return string|null
1135          */
1136         final public static function parentStoragePath( $storagePath ) {
1137                 $storagePath = dirname( $storagePath );
1138                 list( $b, $cont, $rel ) = self::splitStoragePath( $storagePath );
1139                 return ( $rel === null ) ? null : $storagePath;
1140         }
1141
1142         /**
1143          * Get the final extension from a storage or FS path
1144          *
1145          * @param $path string
1146          * @return string
1147          */
1148         final public static function extensionFromPath( $path ) {
1149                 $i = strrpos( $path, '.' );
1150                 return strtolower( $i ? substr( $path, $i + 1 ) : '' );
1151         }
1152
1153         /**
1154          * Check if a relative path has no directory traversals
1155          *
1156          * @param $path string
1157          * @return bool
1158          * @since 1.20
1159          */
1160         final public static function isPathTraversalFree( $path ) {
1161                 return ( self::normalizeContainerPath( $path ) !== null );
1162         }
1163
1164         /**
1165          * Build a Content-Disposition header value per RFC 6266.
1166          *
1167          * @param $type string One of (attachment, inline)
1168          * @param $filename string Suggested file name (should not contain slashes)
1169          * @return string
1170          * @since 1.20
1171          */
1172         final public static function makeContentDisposition( $type, $filename = '' ) {
1173                 $parts = array();
1174
1175                 $type = strtolower( $type );
1176                 if ( !in_array( $type, array( 'inline', 'attachment' ) ) ) {
1177                         throw new MWException( "Invalid Content-Disposition type '$type'." );
1178                 }
1179                 $parts[] = $type;
1180
1181                 if ( strlen( $filename ) ) {
1182                         $parts[] = "filename*=UTF-8''" . rawurlencode( basename( $filename ) );
1183                 }
1184
1185                 return implode( ';', $parts );
1186         }
1187
1188         /**
1189          * Validate and normalize a relative storage path.
1190          * Null is returned if the path involves directory traversal.
1191          * Traversal is insecure for FS backends and broken for others.
1192          *
1193          * This uses the same traversal protection as Title::secureAndSplit().
1194          *
1195          * @param $path string Storage path relative to a container
1196          * @return string|null
1197          */
1198         final protected static function normalizeContainerPath( $path ) {
1199                 // Normalize directory separators
1200                 $path = strtr( $path, '\\', '/' );
1201                 // Collapse any consecutive directory separators
1202                 $path = preg_replace( '![/]{2,}!', '/', $path );
1203                 // Remove any leading directory separator
1204                 $path = ltrim( $path, '/' );
1205                 // Use the same traversal protection as Title::secureAndSplit()
1206                 if ( strpos( $path, '.' ) !== false ) {
1207                         if (
1208                                 $path === '.' ||
1209                                 $path === '..' ||
1210                                 strpos( $path, './' ) === 0 ||
1211                                 strpos( $path, '../' ) === 0 ||
1212                                 strpos( $path, '/./' ) !== false ||
1213                                 strpos( $path, '/../' ) !== false
1214                         ) {
1215                                 return null;
1216                         }
1217                 }
1218                 return $path;
1219         }
1220 }