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          *         'ignoreMissingSource' => <boolean>, # since 1.21
 210          *         'disposition'         => <Content-Disposition header value>
 211          *     )
 212          * @endcode
 213          *
 214          * d) Move a file within storage
 215          * @code
 216          *     array(
 217          *         'op'                  => 'move',
 218          *         'src'                 => <storage path>,
 219          *         'dst'                 => <storage path>,
 220          *         'overwrite'           => <boolean>,
 221          *         'overwriteSame'       => <boolean>,
 222          *         'ignoreMissingSource' => <boolean>, # since 1.21
 223          *         'disposition'         => <Content-Disposition header value>
 224          *     )
 225          * @endcode
 226          *
 227          * e) Delete a file within storage
 228          * @code
 229          *     array(
 230          *         'op'                  => 'delete',
 231          *         'src'                 => <storage path>,
 232          *         'ignoreMissingSource' => <boolean>
 233          *     )
 234          * @endcode
 235          *
 236          * f) Do nothing (no-op)
 237          * @code
 238          *     array(
 239          *         'op'                  => 'null',
 240          *     )
 241          * @endcode
 242          *
 243          * Boolean flags for operations (operation-specific):
 244          *   - ignoreMissingSource : The operation will simply succeed and do
 245          *                           nothing if the source file does not exist.
 246          *   - overwrite           : Any destination file will be overwritten.
 247          *   - overwriteSame       : An error will not be given if a file already
 248          *                           exists at the destination that has the same
 249          *                           contents as the new contents to be written there.
 250          *   - disposition         : When supplied, the backend will add a Content-Disposition
 251          *                           header when GETs/HEADs of the destination file are made.
 252          *                           Backends that don't support file metadata will ignore this.
 253          *                           See http://tools.ietf.org/html/rfc6266 (since 1.20).
 254          *
 255          * $opts is an associative of boolean flags, including:
 256          *   - force               : Operation precondition errors no longer trigger an abort.
 257          *                           Any remaining operations are still attempted. Unexpected
 258          *                           failures may still cause remaning operations to be aborted.
 259          *   - nonLocking          : No locks are acquired for the operations.
 260          *                           This can increase performance for non-critical writes.
 261          *                           This has no effect unless the 'force' flag is set.
 262          *   - allowStale          : Don't require the latest available data.
 263          *                           This can increase performance for non-critical writes.
 264          *                           This has no effect unless the 'force' flag is set.
 265          *   - nonJournaled        : Don't log this operation batch in the file journal.
 266          *                           This limits the ability of recovery scripts.
 267          *   - parallelize         : Try to do operations in parallel when possible.
 268          *   - bypassReadOnly      : Allow writes in read-only mode (since 1.20).
 269          *   - preserveCache       : Don't clear the process cache before checking files.
 270          *                           This should only be used if all entries in the process
 271          *                           cache were added after the files were already locked (since 1.20).
 272          *
 273          * @remarks Remarks on locking:
 274          * File system paths given to operations should refer to files that are
 275          * already locked or otherwise safe from modification from other processes.
 276          * Normally these files will be new temp files, which should be adequate.
 277          *
 278          * @par Return value:
 279          *
 280          * This returns a Status, which contains all warnings and fatals that occurred
 281          * during the operation. The 'failCount', 'successCount', and 'success' members
 282          * will reflect each operation attempted.
 283          *
 284          * The status will be "OK" unless:
 285          *   - a) unexpected operation errors occurred (network partitions, disk full...)
 286          *   - b) significant operation errors occurred and 'force' was not set
 287          *
 288          * @param $ops Array List of operations to execute in order
 289          * @param $opts Array Batch operation options
 290          * @return Status
 291          */
 292         final public function doOperations( array $ops, array $opts = array() ) {
 293                 if ( empty( $opts['bypassReadOnly'] ) && $this->isReadOnly() ) {
 294                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 295                 }
 296                 if ( empty( $opts['force'] ) ) { // sanity
 297                         unset( $opts['nonLocking'] );
 298                         unset( $opts['allowStale'] );
 299                 }
 300                 return $this->doOperationsInternal( $ops, $opts );
 301         }
 302
 303         /**
 304          * @see FileBackend::doOperations()
 305          */
 306         abstract protected function doOperationsInternal( array $ops, array $opts );
 307
 308         /**
 309          * Same as doOperations() except it takes a single operation.
 310          * If you are doing a batch of operations that should either
 311          * all succeed or all fail, then use that function instead.
 312          *
 313          * @see FileBackend::doOperations()
 314          *
 315          * @param $op Array Operation
 316          * @param $opts Array Operation options
 317          * @return Status
 318          */
 319         final public function doOperation( array $op, array $opts = array() ) {
 320                 return $this->doOperations( array( $op ), $opts );
 321         }
 322
 323         /**
 324          * Performs a single create operation.
 325          * This sets $params['op'] to 'create' and passes it to doOperation().
 326          *
 327          * @see FileBackend::doOperation()
 328          *
 329          * @param $params Array Operation parameters
 330          * @param $opts Array Operation options
 331          * @return Status
 332          */
 333         final public function create( array $params, array $opts = array() ) {
 334                 return $this->doOperation( array( 'op' => 'create' ) + $params, $opts );
 335         }
 336
 337         /**
 338          * Performs a single store operation.
 339          * This sets $params['op'] to 'store' and passes it to doOperation().
 340          *
 341          * @see FileBackend::doOperation()
 342          *
 343          * @param $params Array Operation parameters
 344          * @param $opts Array Operation options
 345          * @return Status
 346          */
 347         final public function store( array $params, array $opts = array() ) {
 348                 return $this->doOperation( array( 'op' => 'store' ) + $params, $opts );
 349         }
 350
 351         /**
 352          * Performs a single copy operation.
 353          * This sets $params['op'] to 'copy' and passes it to doOperation().
 354          *
 355          * @see FileBackend::doOperation()
 356          *
 357          * @param $params Array Operation parameters
 358          * @param $opts Array Operation options
 359          * @return Status
 360          */
 361         final public function copy( array $params, array $opts = array() ) {
 362                 return $this->doOperation( array( 'op' => 'copy' ) + $params, $opts );
 363         }
 364
 365         /**
 366          * Performs a single move operation.
 367          * This sets $params['op'] to 'move' and passes it to doOperation().
 368          *
 369          * @see FileBackend::doOperation()
 370          *
 371          * @param $params Array Operation parameters
 372          * @param $opts Array Operation options
 373          * @return Status
 374          */
 375         final public function move( array $params, array $opts = array() ) {
 376                 return $this->doOperation( array( 'op' => 'move' ) + $params, $opts );
 377         }
 378
 379         /**
 380          * Performs a single delete operation.
 381          * This sets $params['op'] to 'delete' and passes it to doOperation().
 382          *
 383          * @see FileBackend::doOperation()
 384          *
 385          * @param $params Array Operation parameters
 386          * @param $opts Array Operation options
 387          * @return Status
 388          */
 389         final public function delete( array $params, array $opts = array() ) {
 390                 return $this->doOperation( array( 'op' => 'delete' ) + $params, $opts );
 391         }
 392
 393         /**
 394          * Perform a set of independent file operations on some files.
 395          *
 396          * This does no locking, nor journaling, and possibly no stat calls.
 397          * Any destination files that already exist will be overwritten.
 398          * This should *only* be used on non-original files, like cache files.
 399          *
 400          * Supported operations and their parameters:
 401          *  - create
 402          *  - store
 403          *  - copy
 404          *  - move
 405          *  - delete
 406          *  - null
 407          *
 408          * a) Create a new file in storage with the contents of a string
 409          * @code
 410          *     array(
 411          *         'op'                  => 'create',
 412          *         'dst'                 => <storage path>,
 413          *         'content'             => <string of new file contents>,
 414          *         'disposition'         => <Content-Disposition header value>
 415          *     )
 416          * @endcode
 417          * b) Copy a file system file into storage
 418          * @code
 419          *     array(
 420          *         'op'                  => 'store',
 421          *         'src'                 => <file system path>,
 422          *         'dst'                 => <storage path>,
 423          *         'disposition'         => <Content-Disposition header value>
 424          *     )
 425          * @endcode
 426          * c) Copy a file within storage
 427          * @code
 428          *     array(
 429          *         'op'                  => 'copy',
 430          *         'src'                 => <storage path>,
 431          *         'dst'                 => <storage path>,
 432          *         'ignoreMissingSource' => <boolean>, # since 1.21
 433          *         'disposition'         => <Content-Disposition header value>
 434          *     )
 435          * @endcode
 436          * d) Move a file within storage
 437          * @code
 438          *     array(
 439          *         'op'                  => 'move',
 440          *         'src'                 => <storage path>,
 441          *         'dst'                 => <storage path>,
 442          *         'ignoreMissingSource' => <boolean>, # since 1.21
 443          *         'disposition'         => <Content-Disposition header value>
 444          *     )
 445          * @endcode
 446          * e) Delete a file within storage
 447          * @code
 448          *     array(
 449          *         'op'                  => 'delete',
 450          *         'src'                 => <storage path>,
 451          *         'ignoreMissingSource' => <boolean>
 452          *     )
 453          * @endcode
 454          * f) Do nothing (no-op)
 455          * @code
 456          *     array(
 457          *         'op'                  => 'null',
 458          *     )
 459          * @endcode
 460          *
 461          * @par Boolean flags for operations (operation-specific):
 462          *   - ignoreMissingSource : The operation will simply succeed and do
 463          *                           nothing if the source file does not exist.
 464          *   - disposition         : When supplied, the backend will add a Content-Disposition
 465          *                           header when GETs/HEADs of the destination file are made.
 466          *                           Backends that don't support file metadata will ignore this.
 467          *                           See http://tools.ietf.org/html/rfc6266 (since 1.20).
 468          *
 469          * $opts is an associative of boolean flags, including:
 470          *   - bypassReadOnly      : Allow writes in read-only mode (since 1.20)
 471          *
 472          * @par Return value:
 473          * This returns a Status, which contains all warnings and fatals that occurred
 474          * during the operation. The 'failCount', 'successCount', and 'success' members
 475          * will reflect each operation attempted for the given files. The status will be
 476          * considered "OK" as long as no fatal errors occurred.
 477          *
 478          * @param $ops Array Set of operations to execute
 479          * @param $opts Array Batch operation options
 480          * @return Status
 481          * @since 1.20
 482          */
 483         final public function doQuickOperations( array $ops, array $opts = array() ) {
 484                 if ( empty( $opts['bypassReadOnly'] ) && $this->isReadOnly() ) {
 485                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 486                 }
 487                 foreach ( $ops as &$op ) {
 488                         $op['overwrite'] = true; // avoids RTTs in key/value stores
 489                 }
 490                 return $this->doQuickOperationsInternal( $ops );
 491         }
 492
 493         /**
 494          * @see FileBackend::doQuickOperations()
 495          * @since 1.20
 496          */
 497         abstract protected function doQuickOperationsInternal( array $ops );
 498
 499         /**
 500          * Same as doQuickOperations() except it takes a single operation.
 501          * If you are doing a batch of operations, then use that function instead.
 502          *
 503          * @see FileBackend::doQuickOperations()
 504          *
 505          * @param $op Array Operation
 506          * @return Status
 507          * @since 1.20
 508          */
 509         final public function doQuickOperation( array $op ) {
 510                 return $this->doQuickOperations( array( $op ) );
 511         }
 512
 513         /**
 514          * Performs a single quick create operation.
 515          * This sets $params['op'] to 'create' and passes it to doQuickOperation().
 516          *
 517          * @see FileBackend::doQuickOperation()
 518          *
 519          * @param $params Array Operation parameters
 520          * @return Status
 521          * @since 1.20
 522          */
 523         final public function quickCreate( array $params ) {
 524                 return $this->doQuickOperation( array( 'op' => 'create' ) + $params );
 525         }
 526
 527         /**
 528          * Performs a single quick store operation.
 529          * This sets $params['op'] to 'store' and passes it to doQuickOperation().
 530          *
 531          * @see FileBackend::doQuickOperation()
 532          *
 533          * @param $params Array Operation parameters
 534          * @return Status
 535          * @since 1.20
 536          */
 537         final public function quickStore( array $params ) {
 538                 return $this->doQuickOperation( array( 'op' => 'store' ) + $params );
 539         }
 540
 541         /**
 542          * Performs a single quick copy operation.
 543          * This sets $params['op'] to 'copy' and passes it to doQuickOperation().
 544          *
 545          * @see FileBackend::doQuickOperation()
 546          *
 547          * @param $params Array Operation parameters
 548          * @return Status
 549          * @since 1.20
 550          */
 551         final public function quickCopy( array $params ) {
 552                 return $this->doQuickOperation( array( 'op' => 'copy' ) + $params );
 553         }
 554
 555         /**
 556          * Performs a single quick move operation.
 557          * This sets $params['op'] to 'move' and passes it to doQuickOperation().
 558          *
 559          * @see FileBackend::doQuickOperation()
 560          *
 561          * @param $params Array Operation parameters
 562          * @return Status
 563          * @since 1.20
 564          */
 565         final public function quickMove( array $params ) {
 566                 return $this->doQuickOperation( array( 'op' => 'move' ) + $params );
 567         }
 568
 569         /**
 570          * Performs a single quick delete operation.
 571          * This sets $params['op'] to 'delete' and passes it to doQuickOperation().
 572          *
 573          * @see FileBackend::doQuickOperation()
 574          *
 575          * @param $params Array Operation parameters
 576          * @return Status
 577          * @since 1.20
 578          */
 579         final public function quickDelete( array $params ) {
 580                 return $this->doQuickOperation( array( 'op' => 'delete' ) + $params );
 581         }
 582
 583         /**
 584          * Concatenate a list of storage files into a single file system file.
 585          * The target path should refer to a file that is already locked or
 586          * otherwise safe from modification from other processes. Normally,
 587          * the file will be a new temp file, which should be adequate.
 588          *
 589          * @param $params Array Operation parameters
 590          * $params include:
 591          *   - srcs        : ordered source storage paths (e.g. chunk1, chunk2, ...)
 592          *   - dst         : file system path to 0-byte temp file
 593          *   - parallelize : try to do operations in parallel when possible
 594          * @return Status
 595          */
 596         abstract public function concatenate( array $params );
 597
 598         /**
 599          * Prepare a storage directory for usage.
 600          * This will create any required containers and parent directories.
 601          * Backends using key/value stores only need to create the container.
 602          *
 603          * The 'noAccess' and 'noListing' parameters works the same as in secure(),
 604          * except they are only applied *if* the directory/container had to be created.
 605          * These flags should always be set for directories that have private files.
 606          *
 607          * @param $params Array
 608          * $params include:
 609          *   - dir            : storage directory
 610          *   - noAccess       : try to deny file access (since 1.20)
 611          *   - noListing      : try to deny file listing (since 1.20)
 612          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 613          * @return Status
 614          */
 615         final public function prepare( array $params ) {
 616                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 617                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 618                 }
 619                 return $this->doPrepare( $params );
 620         }
 621
 622         /**
 623          * @see FileBackend::prepare()
 624          */
 625         abstract protected function doPrepare( array $params );
 626
 627         /**
 628          * Take measures to block web access to a storage directory and
 629          * the container it belongs to. FS backends might add .htaccess
 630          * files whereas key/value store backends might revoke container
 631          * access to the storage user representing end-users in web requests.
 632          * This is not guaranteed to actually do anything.
 633          *
 634          * @param $params Array
 635          * $params include:
 636          *   - dir            : storage directory
 637          *   - noAccess       : try to deny file access
 638          *   - noListing      : try to deny file listing
 639          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 640          * @return Status
 641          */
 642         final public function secure( array $params ) {
 643                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 644                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 645                 }
 646                 return $this->doSecure( $params );
 647         }
 648
 649         /**
 650          * @see FileBackend::secure()
 651          */
 652         abstract protected function doSecure( array $params );
 653
 654         /**
 655          * Remove measures to block web access to a storage directory and
 656          * the container it belongs to. FS backends might remove .htaccess
 657          * files whereas key/value store backends might grant container
 658          * access to the storage user representing end-users in web requests.
 659          * This essentially can undo the result of secure() calls.
 660          *
 661          * @param $params Array
 662          * $params include:
 663          *   - dir            : storage directory
 664          *   - access         : try to allow file access
 665          *   - listing        : try to allow file listing
 666          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 667          * @return Status
 668          * @since 1.20
 669          */
 670         final public function publish( array $params ) {
 671                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 672                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 673                 }
 674                 return $this->doPublish( $params );
 675         }
 676
 677         /**
 678          * @see FileBackend::publish()
 679          */
 680         abstract protected function doPublish( array $params );
 681
 682         /**
 683          * Delete a storage directory if it is empty.
 684          * Backends using key/value stores may do nothing unless the directory
 685          * is that of an empty container, in which case it will be deleted.
 686          *
 687          * @param $params Array
 688          * $params include:
 689          *   - dir            : storage directory
 690          *   - recursive      : recursively delete empty subdirectories first (since 1.20)
 691          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 692          * @return Status
 693          */
 694         final public function clean( array $params ) {
 695                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 696                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 697                 }
 698                 return $this->doClean( $params );
 699         }
 700
 701         /**
 702          * @see FileBackend::clean()
 703          */
 704         abstract protected function doClean( array $params );
 705
 706         /**
 707          * Check if a file exists at a storage path in the backend.
 708          * This returns false if only a directory exists at the path.
 709          *
 710          * @param $params Array
 711          * $params include:
 712          *   - src    : source storage path
 713          *   - latest : use the latest available data
 714          * @return bool|null Returns null on failure
 715          */
 716         abstract public function fileExists( array $params );
 717
 718         /**
 719          * Get the last-modified timestamp of the file at a storage path.
 720          *
 721          * @param $params Array
 722          * $params include:
 723          *   - src    : source storage path
 724          *   - latest : use the latest available data
 725          * @return string|bool TS_MW timestamp or false on failure
 726          */
 727         abstract public function getFileTimestamp( array $params );
 728
 729         /**
 730          * Get the contents of a file at a storage path in the backend.
 731          * This should be avoided for potentially large files.
 732          *
 733          * @param $params Array
 734          * $params include:
 735          *   - src    : source storage path
 736          *   - latest : use the latest available data
 737          * @return string|bool Returns false on failure
 738          */
 739         final public function getFileContents( array $params ) {
 740                 $contents = $this->getFileContentsMulti(
 741                         array( 'srcs' => array( $params['src'] ) ) + $params );
 742
 743                 return $contents[$params['src']];
 744         }
 745
 746         /**
 747          * Like getFileContents() except it takes an array of storage paths
 748          * and returns a map of storage paths to strings (or null on failure).
 749          * The map keys (paths) are in the same order as the provided list of paths.
 750          *
 751          * @see FileBackend::getFileContents()
 752          *
 753          * @param $params Array
 754          * $params include:
 755          *   - srcs        : list of source storage paths
 756          *   - latest      : use the latest available data
 757          *   - parallelize : try to do operations in parallel when possible
 758          * @return Array Map of (path name => string or false on failure)
 759          * @since 1.20
 760          */
 761         abstract public function getFileContentsMulti( array $params );
 762
 763         /**
 764          * Get the size (bytes) of a 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 integer|bool Returns false on failure
 771          */
 772         abstract public function getFileSize( array $params );
 773
 774         /**
 775          * Get quick information about a file at a storage path in the backend.
 776          * If the file does not exist, then this returns false.
 777          * Otherwise, the result is an associative array that includes:
 778          *   - mtime  : the last-modified timestamp (TS_MW)
 779          *   - size   : the file size (bytes)
 780          * Additional values may be included for internal use only.
 781          *
 782          * @param $params Array
 783          * $params include:
 784          *   - src    : source storage path
 785          *   - latest : use the latest available data
 786          * @return Array|bool|null Returns null on failure
 787          */
 788         abstract public function getFileStat( array $params );
 789
 790         /**
 791          * Get a SHA-1 hash of the file at a storage path in the backend.
 792          *
 793          * @param $params Array
 794          * $params include:
 795          *   - src    : source storage path
 796          *   - latest : use the latest available data
 797          * @return string|bool Hash string or false on failure
 798          */
 799         abstract public function getFileSha1Base36( array $params );
 800
 801         /**
 802          * Get the properties of the file at a storage path in the backend.
 803          * Returns FSFile::placeholderProps() on failure.
 804          *
 805          * @param $params Array
 806          * $params include:
 807          *   - src    : source storage path
 808          *   - latest : use the latest available data
 809          * @return Array
 810          */
 811         abstract public function getFileProps( array $params );
 812
 813         /**
 814          * Stream the file at a storage path in the backend.
 815          * If the file does not exists, an HTTP 404 error will be given.
 816          * Appropriate HTTP headers (Status, Content-Type, Content-Length)
 817          * will be sent if streaming began, while none will be sent otherwise.
 818          * Implementations should flush the output buffer before sending data.
 819          *
 820          * @param $params Array
 821          * $params include:
 822          *   - src     : source storage path
 823          *   - headers : list of additional HTTP headers to send on success
 824          *   - latest  : use the latest available data
 825          * @return Status
 826          */
 827         abstract public function streamFile( array $params );
 828
 829         /**
 830          * Returns a file system file, identical to the file at a storage path.
 831          * The file returned is either:
 832          *   - a) A local copy of the file at a storage path in the backend.
 833          *        The temporary copy will have the same extension as the source.
 834          *   - b) An original of the file at a storage path in the backend.
 835          * Temporary files may be purged when the file object falls out of scope.
 836          *
 837          * Write operations should *never* be done on this file as some backends
 838          * may do internal tracking or may be instances of FileBackendMultiWrite.
 839          * In that later case, there are copies of the file that must stay in sync.
 840          * Additionally, further calls to this function may return the same file.
 841          *
 842          * @param $params Array
 843          * $params include:
 844          *   - src    : source storage path
 845          *   - latest : use the latest available data
 846          * @return FSFile|null Returns null on failure
 847          */
 848         final public function getLocalReference( array $params ) {
 849                 $fsFiles = $this->getLocalReferenceMulti(
 850                         array( 'srcs' => array( $params['src'] ) ) + $params );
 851
 852                 return $fsFiles[$params['src']];
 853         }
 854
 855         /**
 856          * Like getLocalReference() except it takes an array of storage paths
 857          * and returns a map of storage paths to FSFile objects (or null on failure).
 858          * The map keys (paths) are in the same order as the provided list of paths.
 859          *
 860          * @see FileBackend::getLocalReference()
 861          *
 862          * @param $params Array
 863          * $params include:
 864          *   - srcs        : list of source storage paths
 865          *   - latest      : use the latest available data
 866          *   - parallelize : try to do operations in parallel when possible
 867          * @return Array Map of (path name => FSFile or null on failure)
 868          * @since 1.20
 869          */
 870         abstract public function getLocalReferenceMulti( array $params );
 871
 872         /**
 873          * Get a local copy on disk of the file at a storage path in the backend.
 874          * The temporary copy will have the same file extension as the source.
 875          * Temporary files may be purged when the file object falls out of scope.
 876          *
 877          * @param $params Array
 878          * $params include:
 879          *   - src    : source storage path
 880          *   - latest : use the latest available data
 881          * @return TempFSFile|null Returns null on failure
 882          */
 883         final public function getLocalCopy( array $params ) {
 884                 $tmpFiles = $this->getLocalCopyMulti(
 885                         array( 'srcs' => array( $params['src'] ) ) + $params );
 886
 887                 return $tmpFiles[$params['src']];
 888         }
 889
 890         /**
 891          * Like getLocalCopy() except it takes an array of storage paths and
 892          * returns a map of storage paths to TempFSFile objects (or null on failure).
 893          * The map keys (paths) are in the same order as the provided list of paths.
 894          *
 895          * @see FileBackend::getLocalCopy()
 896          *
 897          * @param $params Array
 898          * $params include:
 899          *   - srcs        : list of source storage paths
 900          *   - latest      : use the latest available data
 901          *   - parallelize : try to do operations in parallel when possible
 902          * @return Array Map of (path name => TempFSFile or null on failure)
 903          * @since 1.20
 904          */
 905         abstract public function getLocalCopyMulti( array $params );
 906
 907         /**
 908          * Return an HTTP URL to a given file that requires no authentication to use.
 909          * The URL may be pre-authenticated (via some token in the URL) and temporary.
 910          * This will return null if the backend cannot make an HTTP URL for the file.
 911          *
 912          * This is useful for key/value stores when using scripts that seek around
 913          * large files and those scripts (and the backend) support HTTP Range headers.
 914          * Otherwise, one would need to use getLocalReference(), which involves loading
 915          * the entire file on to local disk.
 916          *
 917          * @param $params Array
 918          * $params include:
 919          *   - src : source storage path
 920          * @return string|null
 921          * @since 1.21
 922          */
 923         abstract public function getFileHttpUrl( array $params );
 924
 925         /**
 926          * Check if a directory exists at a given storage path.
 927          * Backends using key/value stores will check if the path is a
 928          * virtual directory, meaning there are files 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 bool|null Returns null on failure
 936          * @since 1.20
 937          */
 938         abstract public function directoryExists( array $params );
 939
 940         /**
 941          * Get an iterator to list *all* directories under a storage directory.
 942          * If the directory is of the form "mwstore://backend/container",
 943          * then all directories in the container will be listed.
 944          * If the directory is of form "mwstore://backend/container/dir",
 945          * then all directories directly under that directory will be listed.
 946          * Results will be storage directories relative to the given directory.
 947          *
 948          * Storage backends with eventual consistency might return stale data.
 949          *
 950          * @param $params array
 951          * $params include:
 952          *   - dir     : storage directory
 953          *   - topOnly : only return direct child dirs of the directory
 954          * @return Traversable|Array|null Returns null on failure
 955          * @since 1.20
 956          */
 957         abstract public function getDirectoryList( array $params );
 958
 959         /**
 960          * Same as FileBackend::getDirectoryList() except only lists
 961          * directories that are immediately under the given directory.
 962          *
 963          * Storage backends with eventual consistency might return stale data.
 964          *
 965          * @param $params array
 966          * $params include:
 967          *   - dir : storage directory
 968          * @return Traversable|Array|null Returns null on failure
 969          * @since 1.20
 970          */
 971         final public function getTopDirectoryList( array $params ) {
 972                 return $this->getDirectoryList( array( 'topOnly' => true ) + $params );
 973         }
 974
 975         /**
 976          * Get an iterator to list *all* stored files under a storage directory.
 977          * If the directory is of the form "mwstore://backend/container",
 978          * then all files in the container will be listed.
 979          * If the directory is of form "mwstore://backend/container/dir",
 980          * then all files under that directory will be listed.
 981          * Results will be storage paths relative to the given directory.
 982          *
 983          * Storage backends with eventual consistency might return stale data.
 984          *
 985          * @param $params array
 986          * $params include:
 987          *   - dir     : storage directory
 988          *   - topOnly : only return direct child files of the directory (since 1.20)
 989          * @return Traversable|Array|null Returns null on failure
 990          */
 991         abstract public function getFileList( array $params );
 992
 993         /**
 994          * Same as FileBackend::getFileList() except only lists
 995          * files that are immediately under the given directory.
 996          *
 997          * Storage backends with eventual consistency might return stale data.
 998          *
 999          * @param $params array
1000          * $params include:
1001          *   - dir : storage directory
1002          * @return Traversable|Array|null Returns null on failure
1003          * @since 1.20
1004          */
1005         final public function getTopFileList( array $params ) {
1006                 return $this->getFileList( array( 'topOnly' => true ) + $params );
1007         }
1008
1009         /**
1010          * Preload persistent file stat and property cache into in-process cache.
1011          * This should be used when stat calls will be made on a known list of a many files.
1012          *
1013          * @param $paths Array Storage paths
1014          * @return void
1015          */
1016         public function preloadCache( array $paths ) {}
1017
1018         /**
1019          * Invalidate any in-process file stat and property cache.
1020          * If $paths is given, then only the cache for those files will be cleared.
1021          *
1022          * @param $paths Array Storage paths (optional)
1023          * @return void
1024          */
1025         public function clearCache( array $paths = null ) {}
1026
1027         /**
1028          * Lock the files at the given storage paths in the backend.
1029          * This will either lock all the files or none (on failure).
1030          *
1031          * Callers should consider using getScopedFileLocks() instead.
1032          *
1033          * @param $paths Array Storage paths
1034          * @param $type integer LockManager::LOCK_* constant
1035          * @return Status
1036          */
1037         final public function lockFiles( array $paths, $type ) {
1038                 return $this->lockManager->lock( $paths, $type );
1039         }
1040
1041         /**
1042          * Unlock the files at the given storage paths in the backend.
1043          *
1044          * @param $paths Array Storage paths
1045          * @param $type integer LockManager::LOCK_* constant
1046          * @return Status
1047          */
1048         final public function unlockFiles( array $paths, $type ) {
1049                 return $this->lockManager->unlock( $paths, $type );
1050         }
1051
1052         /**
1053          * Lock the files at the given storage paths in the backend.
1054          * This will either lock all the files or none (on failure).
1055          * On failure, the status object will be updated with errors.
1056          *
1057          * Once the return value goes out scope, the locks will be released and
1058          * the status updated. Unlock fatals will not change the status "OK" value.
1059          *
1060          * @param $paths Array Storage paths
1061          * @param $type integer LockManager::LOCK_* constant
1062          * @param $status Status Status to update on lock/unlock
1063          * @return ScopedLock|null Returns null on failure
1064          */
1065         final public function getScopedFileLocks( array $paths, $type, Status $status ) {
1066                 return ScopedLock::factory( $this->lockManager, $paths, $type, $status );
1067         }
1068
1069         /**
1070          * Get an array of scoped locks needed for a batch of file operations.
1071          *
1072          * Normally, FileBackend::doOperations() handles locking, unless
1073          * the 'nonLocking' param is passed in. This function is useful if you
1074          * want the files to be locked for a broader scope than just when the
1075          * files are changing. For example, if you need to update DB metadata,
1076          * you may want to keep the files locked until finished.
1077          *
1078          * @see FileBackend::doOperations()
1079          *
1080          * @param $ops Array List of file operations to FileBackend::doOperations()
1081          * @param $status Status Status to update on lock/unlock
1082          * @return Array List of ScopedFileLocks or null values
1083          * @since 1.20
1084          */
1085         abstract public function getScopedLocksForOps( array $ops, Status $status );
1086
1087         /**
1088          * Get the root storage path of this backend.
1089          * All container paths are "subdirectories" of this path.
1090          *
1091          * @return string Storage path
1092          * @since 1.20
1093          */
1094         final public function getRootStoragePath() {
1095                 return "mwstore://{$this->name}";
1096         }
1097
1098         /**
1099          * Get the storage path for the given container for this backend
1100          *
1101          * @param $container string Container name
1102          * @return string Storage path
1103          * @since 1.21
1104          */
1105         final public function getContainerStoragePath( $container ) {
1106                 return $this->getRootStoragePath() . "/{$container}";
1107         }
1108
1109         /**
1110          * Get the file journal object for this backend
1111          *
1112          * @return FileJournal
1113          */
1114         final public function getJournal() {
1115                 return $this->fileJournal;
1116         }
1117
1118         /**
1119          * Check if a given path is a "mwstore://" path.
1120          * This does not do any further validation or any existence checks.
1121          *
1122          * @param $path string
1123          * @return bool
1124          */
1125         final public static function isStoragePath( $path ) {
1126                 return ( strpos( $path, 'mwstore://' ) === 0 );
1127         }
1128
1129         /**
1130          * Split a storage path into a backend name, a container name,
1131          * and a relative file path. The relative path may be the empty string.
1132          * This does not do any path normalization or traversal checks.
1133          *
1134          * @param $storagePath string
1135          * @return Array (backend, container, rel object) or (null, null, null)
1136          */
1137         final public static function splitStoragePath( $storagePath ) {
1138                 if ( self::isStoragePath( $storagePath ) ) {
1139                         // Remove the "mwstore://" prefix and split the path
1140                         $parts = explode( '/', substr( $storagePath, 10 ), 3 );
1141                         if ( count( $parts ) >= 2 && $parts[0] != '' && $parts[1] != '' ) {
1142                                 if ( count( $parts ) == 3 ) {
1143                                         return $parts; // e.g. "backend/container/path"
1144                                 } else {
1145                                         return array( $parts[0], $parts[1], '' ); // e.g. "backend/container"
1146                                 }
1147                         }
1148                 }
1149                 return array( null, null, null );
1150         }
1151
1152         /**
1153          * Normalize a storage path by cleaning up directory separators.
1154          * Returns null if the path is not of the format of a valid storage path.
1155          *
1156          * @param $storagePath string
1157          * @return string|null
1158          */
1159         final public static function normalizeStoragePath( $storagePath ) {
1160                 list( $backend, $container, $relPath ) = self::splitStoragePath( $storagePath );
1161                 if ( $relPath !== null ) { // must be for this backend
1162                         $relPath = self::normalizeContainerPath( $relPath );
1163                         if ( $relPath !== null ) {
1164                                 return ( $relPath != '' )
1165                                         ? "mwstore://{$backend}/{$container}/{$relPath}"
1166                                         : "mwstore://{$backend}/{$container}";
1167                         }
1168                 }
1169                 return null;
1170         }
1171
1172         /**
1173          * Get the parent storage directory of a storage path.
1174          * This returns a path like "mwstore://backend/container",
1175          * "mwstore://backend/container/...", or null if there is no parent.
1176          *
1177          * @param $storagePath string
1178          * @return string|null
1179          */
1180         final public static function parentStoragePath( $storagePath ) {
1181                 $storagePath = dirname( $storagePath );
1182                 list( $b, $cont, $rel ) = self::splitStoragePath( $storagePath );
1183                 return ( $rel === null ) ? null : $storagePath;
1184         }
1185
1186         /**
1187          * Get the final extension from a storage or FS path
1188          *
1189          * @param $path string
1190          * @return string
1191          */
1192         final public static function extensionFromPath( $path ) {
1193                 $i = strrpos( $path, '.' );
1194                 return strtolower( $i ? substr( $path, $i + 1 ) : '' );
1195         }
1196
1197         /**
1198          * Check if a relative path has no directory traversals
1199          *
1200          * @param $path string
1201          * @return bool
1202          * @since 1.20
1203          */
1204         final public static function isPathTraversalFree( $path ) {
1205                 return ( self::normalizeContainerPath( $path ) !== null );
1206         }
1207
1208         /**
1209          * Build a Content-Disposition header value per RFC 6266.
1210          *
1211          * @param $type string One of (attachment, inline)
1212          * @param $filename string Suggested file name (should not contain slashes)
1213          * @throws MWException
1214          * @return string
1215          * @since 1.20
1216          */
1217         final public static function makeContentDisposition( $type, $filename = '' ) {
1218                 $parts = array();
1219
1220                 $type = strtolower( $type );
1221                 if ( !in_array( $type, array( 'inline', 'attachment' ) ) ) {
1222                         throw new MWException( "Invalid Content-Disposition type '$type'." );
1223                 }
1224                 $parts[] = $type;
1225
1226                 if ( strlen( $filename ) ) {
1227                         $parts[] = "filename*=UTF-8''" . rawurlencode( basename( $filename ) );
1228                 }
1229
1230                 return implode( ';', $parts );
1231         }
1232
1233         /**
1234          * Validate and normalize a relative storage path.
1235          * Null is returned if the path involves directory traversal.
1236          * Traversal is insecure for FS backends and broken for others.
1237          *
1238          * This uses the same traversal protection as Title::secureAndSplit().
1239          *
1240          * @param $path string Storage path relative to a container
1241          * @return string|null
1242          */
1243         final protected static function normalizeContainerPath( $path ) {
1244                 // Normalize directory separators
1245                 $path = strtr( $path, '\\', '/' );
1246                 // Collapse any consecutive directory separators
1247                 $path = preg_replace( '![/]{2,}!', '/', $path );
1248                 // Remove any leading directory separator
1249                 $path = ltrim( $path, '/' );
1250                 // Use the same traversal protection as Title::secureAndSplit()
1251                 if ( strpos( $path, '.' ) !== false ) {
1252                         if (
1253                                 $path === '.' ||
1254                                 $path === '..' ||
1255                                 strpos( $path, './' ) === 0 ||
1256                                 strpos( $path, '../' ) === 0 ||
1257                                 strpos( $path, '/./' ) !== false ||
1258                                 strpos( $path, '/../' ) !== false
1259                         ) {
1260                                 return null;
1261                         }
1262                 }
1263                 return $path;
1264         }
1265 }