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          *  - describe (since 1.21)
 176          *  - null
 177          *
 178          * a) Create a new file in storage with the contents of a string
 179          * @code
 180          *     array(
 181          *         'op'                  => 'create',
 182          *         'dst'                 => <storage path>,
 183          *         'content'             => <string of new file contents>,
 184          *         'overwrite'           => <boolean>,
 185          *         'overwriteSame'       => <boolean>,
 186          *         'disposition'         => <Content-Disposition header value>,
 187          *         'headers'             => <HTTP header name/value map> # since 1.21
 188          *     );
 189          * @endcode
 190          *
 191          * b) Copy a file system file into storage
 192          * @code
 193          *     array(
 194          *         'op'                  => 'store',
 195          *         'src'                 => <file system path>,
 196          *         'dst'                 => <storage path>,
 197          *         'overwrite'           => <boolean>,
 198          *         'overwriteSame'       => <boolean>,
 199          *         'disposition'         => <Content-Disposition header value>,
 200          *         'headers'             => <HTTP header name/value map> # since 1.21
 201          *     )
 202          * @endcode
 203          *
 204          * c) Copy a file within storage
 205          * @code
 206          *     array(
 207          *         'op'                  => 'copy',
 208          *         'src'                 => <storage path>,
 209          *         'dst'                 => <storage path>,
 210          *         'overwrite'           => <boolean>,
 211          *         'overwriteSame'       => <boolean>,
 212          *         'ignoreMissingSource' => <boolean>, # since 1.21
 213          *         'disposition'         => <Content-Disposition header value>
 214          *     )
 215          * @endcode
 216          *
 217          * d) Move a file within storage
 218          * @code
 219          *     array(
 220          *         'op'                  => 'move',
 221          *         'src'                 => <storage path>,
 222          *         'dst'                 => <storage path>,
 223          *         'overwrite'           => <boolean>,
 224          *         'overwriteSame'       => <boolean>,
 225          *         'ignoreMissingSource' => <boolean>, # since 1.21
 226          *         'disposition'         => <Content-Disposition header value>
 227          *     )
 228          * @endcode
 229          *
 230          * e) Delete a file within storage
 231          * @code
 232          *     array(
 233          *         'op'                  => 'delete',
 234          *         'src'                 => <storage path>,
 235          *         'ignoreMissingSource' => <boolean>
 236          *     )
 237          * @endcode
 238          *
 239          * f) Update metadata for a file within storage
 240          * @code
 241          *     array(
 242          *         'op'                  => 'describe',
 243          *         'src'                 => <storage path>,
 244          *         'disposition'         => <Content-Disposition header value>,
 245          *         'headers'             => <HTTP header name/value map>
 246          *     )
 247          * @endcode
 248          *
 249          * g) Do nothing (no-op)
 250          * @code
 251          *     array(
 252          *         'op'                  => 'null',
 253          *     )
 254          * @endcode
 255          *
 256          * Boolean flags for operations (operation-specific):
 257          *   - ignoreMissingSource : The operation will simply succeed and do
 258          *                           nothing if the source file does not exist.
 259          *   - overwrite           : Any destination file will be overwritten.
 260          *   - overwriteSame       : An error will not be given if a file already
 261          *                           exists at the destination that has the same
 262          *                           contents as the new contents to be written there.
 263          *   - disposition         : If supplied, the backend will return a Content-Disposition
 264          *                           header when GETs/HEADs of the destination file are made.
 265          *                           Backends that don't support metadata ignore this.
 266          *                           See http://tools.ietf.org/html/rfc6266. (since 1.20)
 267          *   - headers             : If supplied, the backend will return these headers when
 268          *                           GETs/HEADs of the destination file are made. Header values
 269          *                           should be smaller than 256 bytes, often options or numbers.
 270          *                           Existing headers will remain, but these will replace any
 271          *                           conflicting previous headers, and headers will be removed
 272          *                           if they are set to an empty string.
 273          *                           Backends that don't support metadata ignore this. (since 1.21)
 274          *
 275          * $opts is an associative of boolean flags, including:
 276          *   - force               : Operation precondition errors no longer trigger an abort.
 277          *                           Any remaining operations are still attempted. Unexpected
 278          *                           failures may still cause remaning operations to be aborted.
 279          *   - nonLocking          : No locks are acquired for the operations.
 280          *                           This can increase performance for non-critical writes.
 281          *                           This has no effect unless the 'force' flag is set.
 282          *   - allowStale          : Don't require the latest available data.
 283          *                           This can increase performance for non-critical writes.
 284          *                           This has no effect unless the 'force' flag is set.
 285          *   - nonJournaled        : Don't log this operation batch in the file journal.
 286          *                           This limits the ability of recovery scripts.
 287          *   - parallelize         : Try to do operations in parallel when possible.
 288          *   - bypassReadOnly      : Allow writes in read-only mode. (since 1.20)
 289          *   - preserveCache       : Don't clear the process cache before checking files.
 290          *                           This should only be used if all entries in the process
 291          *                           cache were added after the files were already locked. (since 1.20)
 292          *
 293          * @remarks Remarks on locking:
 294          * File system paths given to operations should refer to files that are
 295          * already locked or otherwise safe from modification from other processes.
 296          * Normally these files will be new temp files, which should be adequate.
 297          *
 298          * @par Return value:
 299          *
 300          * This returns a Status, which contains all warnings and fatals that occurred
 301          * during the operation. The 'failCount', 'successCount', and 'success' members
 302          * will reflect each operation attempted.
 303          *
 304          * The status will be "OK" unless:
 305          *   - a) unexpected operation errors occurred (network partitions, disk full...)
 306          *   - b) significant operation errors occurred and 'force' was not set
 307          *
 308          * @param $ops Array List of operations to execute in order
 309          * @param $opts Array Batch operation options
 310          * @return Status
 311          */
 312         final public function doOperations( array $ops, array $opts = array() ) {
 313                 if ( empty( $opts['bypassReadOnly'] ) && $this->isReadOnly() ) {
 314                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 315                 }
 316                 if ( empty( $opts['force'] ) ) { // sanity
 317                         unset( $opts['nonLocking'] );
 318                         unset( $opts['allowStale'] );
 319                 }
 320                 return $this->doOperationsInternal( $ops, $opts );
 321         }
 322
 323         /**
 324          * @see FileBackend::doOperations()
 325          */
 326         abstract protected function doOperationsInternal( array $ops, array $opts );
 327
 328         /**
 329          * Same as doOperations() except it takes a single operation.
 330          * If you are doing a batch of operations that should either
 331          * all succeed or all fail, then use that function instead.
 332          *
 333          * @see FileBackend::doOperations()
 334          *
 335          * @param $op Array Operation
 336          * @param $opts Array Operation options
 337          * @return Status
 338          */
 339         final public function doOperation( array $op, array $opts = array() ) {
 340                 return $this->doOperations( array( $op ), $opts );
 341         }
 342
 343         /**
 344          * Performs a single create operation.
 345          * This sets $params['op'] to 'create' and passes it to doOperation().
 346          *
 347          * @see FileBackend::doOperation()
 348          *
 349          * @param $params Array Operation parameters
 350          * @param $opts Array Operation options
 351          * @return Status
 352          */
 353         final public function create( array $params, array $opts = array() ) {
 354                 return $this->doOperation( array( 'op' => 'create' ) + $params, $opts );
 355         }
 356
 357         /**
 358          * Performs a single store operation.
 359          * This sets $params['op'] to 'store' and passes it to doOperation().
 360          *
 361          * @see FileBackend::doOperation()
 362          *
 363          * @param $params Array Operation parameters
 364          * @param $opts Array Operation options
 365          * @return Status
 366          */
 367         final public function store( array $params, array $opts = array() ) {
 368                 return $this->doOperation( array( 'op' => 'store' ) + $params, $opts );
 369         }
 370
 371         /**
 372          * Performs a single copy operation.
 373          * This sets $params['op'] to 'copy' and passes it to doOperation().
 374          *
 375          * @see FileBackend::doOperation()
 376          *
 377          * @param $params Array Operation parameters
 378          * @param $opts Array Operation options
 379          * @return Status
 380          */
 381         final public function copy( array $params, array $opts = array() ) {
 382                 return $this->doOperation( array( 'op' => 'copy' ) + $params, $opts );
 383         }
 384
 385         /**
 386          * Performs a single move operation.
 387          * This sets $params['op'] to 'move' and passes it to doOperation().
 388          *
 389          * @see FileBackend::doOperation()
 390          *
 391          * @param $params Array Operation parameters
 392          * @param $opts Array Operation options
 393          * @return Status
 394          */
 395         final public function move( array $params, array $opts = array() ) {
 396                 return $this->doOperation( array( 'op' => 'move' ) + $params, $opts );
 397         }
 398
 399         /**
 400          * Performs a single delete operation.
 401          * This sets $params['op'] to 'delete' and passes it to doOperation().
 402          *
 403          * @see FileBackend::doOperation()
 404          *
 405          * @param $params Array Operation parameters
 406          * @param $opts Array Operation options
 407          * @return Status
 408          */
 409         final public function delete( array $params, array $opts = array() ) {
 410                 return $this->doOperation( array( 'op' => 'delete' ) + $params, $opts );
 411         }
 412
 413         /**
 414          * Performs a single describe operation.
 415          * This sets $params['op'] to 'describe' and passes it to doOperation().
 416          *
 417          * @see FileBackend::doOperation()
 418          *
 419          * @param $params Array Operation parameters
 420          * @param $opts Array Operation options
 421          * @return Status
 422          * @since 1.21
 423          */
 424         final public function describe( array $params, array $opts = array() ) {
 425                 return $this->doOperation( array( 'op' => 'describe' ) + $params, $opts );
 426         }
 427
 428         /**
 429          * Perform a set of independent file operations on some files.
 430          *
 431          * This does no locking, nor journaling, and possibly no stat calls.
 432          * Any destination files that already exist will be overwritten.
 433          * This should *only* be used on non-original files, like cache files.
 434          *
 435          * Supported operations and their parameters:
 436          *  - create
 437          *  - store
 438          *  - copy
 439          *  - move
 440          *  - delete
 441          *  - describe (since 1.21)
 442          *  - null
 443          *
 444          * a) Create a new file in storage with the contents of a string
 445          * @code
 446          *     array(
 447          *         'op'                  => 'create',
 448          *         'dst'                 => <storage path>,
 449          *         'content'             => <string of new file contents>,
 450          *         'disposition'         => <Content-Disposition header value>,
 451          *         'headers'             => <HTTP header name/value map> # since 1.21
 452          *     )
 453          * @endcode
 454          *
 455          * b) Copy a file system file into storage
 456          * @code
 457          *     array(
 458          *         'op'                  => 'store',
 459          *         'src'                 => <file system path>,
 460          *         'dst'                 => <storage path>,
 461          *         'disposition'         => <Content-Disposition header value>,
 462          *         'headers'             => <HTTP header name/value map> # since 1.21
 463          *     )
 464          * @endcode
 465          *
 466          * c) Copy a file within storage
 467          * @code
 468          *     array(
 469          *         'op'                  => 'copy',
 470          *         'src'                 => <storage path>,
 471          *         'dst'                 => <storage path>,
 472          *         'ignoreMissingSource' => <boolean>, # since 1.21
 473          *         'disposition'         => <Content-Disposition header value>
 474          *     )
 475          * @endcode
 476          *
 477          * d) Move a file within storage
 478          * @code
 479          *     array(
 480          *         'op'                  => 'move',
 481          *         'src'                 => <storage path>,
 482          *         'dst'                 => <storage path>,
 483          *         'ignoreMissingSource' => <boolean>, # since 1.21
 484          *         'disposition'         => <Content-Disposition header value>
 485          *     )
 486          * @endcode
 487          *
 488          * e) Delete a file within storage
 489          * @code
 490          *     array(
 491          *         'op'                  => 'delete',
 492          *         'src'                 => <storage path>,
 493          *         'ignoreMissingSource' => <boolean>
 494          *     )
 495          * @endcode
 496          *
 497          * f) Update metadata for a file within storage
 498          * @code
 499          *     array(
 500          *         'op'                  => 'describe',
 501          *         'src'                 => <storage path>,
 502          *         'disposition'         => <Content-Disposition header value>,
 503          *         'headers'             => <HTTP header name/value map>
 504          *     )
 505          * @endcode
 506          *
 507          * g) Do nothing (no-op)
 508          * @code
 509          *     array(
 510          *         'op'                  => 'null',
 511          *     )
 512          * @endcode
 513          *
 514          * @par Boolean flags for operations (operation-specific):
 515          *   - ignoreMissingSource : The operation will simply succeed and do
 516          *                           nothing if the source file does not exist.
 517          *   - disposition         : When supplied, the backend will add a Content-Disposition
 518          *                           header when GETs/HEADs of the destination file are made.
 519          *                           Backends that don't support file metadata will ignore this.
 520          *                           See http://tools.ietf.org/html/rfc6266 (since 1.20).
 521          *   - headers             : If supplied, the backend will return these headers when
 522          *                           GETs/HEADs of the destination file are made. Header values
 523          *                           should be smaller than 256 bytes, often options or numbers.
 524          *                           Existing headers will remain, but these will replace any
 525          *                           conflicting previous headers, and headers will be removed
 526          *                           if they are set to an empty string.
 527          *                           Backends that don't support metadata ignore this. (since 1.21)
 528          *
 529          * $opts is an associative of boolean flags, including:
 530          *   - bypassReadOnly      : Allow writes in read-only mode (since 1.20)
 531          *
 532          * @par Return value:
 533          * This returns a Status, which contains all warnings and fatals that occurred
 534          * during the operation. The 'failCount', 'successCount', and 'success' members
 535          * will reflect each operation attempted for the given files. The status will be
 536          * considered "OK" as long as no fatal errors occurred.
 537          *
 538          * @param $ops Array Set of operations to execute
 539          * @param $opts Array Batch operation options
 540          * @return Status
 541          * @since 1.20
 542          */
 543         final public function doQuickOperations( array $ops, array $opts = array() ) {
 544                 if ( empty( $opts['bypassReadOnly'] ) && $this->isReadOnly() ) {
 545                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 546                 }
 547                 foreach ( $ops as &$op ) {
 548                         $op['overwrite'] = true; // avoids RTTs in key/value stores
 549                 }
 550                 return $this->doQuickOperationsInternal( $ops );
 551         }
 552
 553         /**
 554          * @see FileBackend::doQuickOperations()
 555          * @since 1.20
 556          */
 557         abstract protected function doQuickOperationsInternal( array $ops );
 558
 559         /**
 560          * Same as doQuickOperations() except it takes a single operation.
 561          * If you are doing a batch of operations, then use that function instead.
 562          *
 563          * @see FileBackend::doQuickOperations()
 564          *
 565          * @param $op Array Operation
 566          * @return Status
 567          * @since 1.20
 568          */
 569         final public function doQuickOperation( array $op ) {
 570                 return $this->doQuickOperations( array( $op ) );
 571         }
 572
 573         /**
 574          * Performs a single quick create operation.
 575          * This sets $params['op'] to 'create' and passes it to doQuickOperation().
 576          *
 577          * @see FileBackend::doQuickOperation()
 578          *
 579          * @param $params Array Operation parameters
 580          * @return Status
 581          * @since 1.20
 582          */
 583         final public function quickCreate( array $params ) {
 584                 return $this->doQuickOperation( array( 'op' => 'create' ) + $params );
 585         }
 586
 587         /**
 588          * Performs a single quick store operation.
 589          * This sets $params['op'] to 'store' and passes it to doQuickOperation().
 590          *
 591          * @see FileBackend::doQuickOperation()
 592          *
 593          * @param $params Array Operation parameters
 594          * @return Status
 595          * @since 1.20
 596          */
 597         final public function quickStore( array $params ) {
 598                 return $this->doQuickOperation( array( 'op' => 'store' ) + $params );
 599         }
 600
 601         /**
 602          * Performs a single quick copy operation.
 603          * This sets $params['op'] to 'copy' and passes it to doQuickOperation().
 604          *
 605          * @see FileBackend::doQuickOperation()
 606          *
 607          * @param $params Array Operation parameters
 608          * @return Status
 609          * @since 1.20
 610          */
 611         final public function quickCopy( array $params ) {
 612                 return $this->doQuickOperation( array( 'op' => 'copy' ) + $params );
 613         }
 614
 615         /**
 616          * Performs a single quick move operation.
 617          * This sets $params['op'] to 'move' and passes it to doQuickOperation().
 618          *
 619          * @see FileBackend::doQuickOperation()
 620          *
 621          * @param $params Array Operation parameters
 622          * @return Status
 623          * @since 1.20
 624          */
 625         final public function quickMove( array $params ) {
 626                 return $this->doQuickOperation( array( 'op' => 'move' ) + $params );
 627         }
 628
 629         /**
 630          * Performs a single quick delete operation.
 631          * This sets $params['op'] to 'delete' and passes it to doQuickOperation().
 632          *
 633          * @see FileBackend::doQuickOperation()
 634          *
 635          * @param $params Array Operation parameters
 636          * @return Status
 637          * @since 1.20
 638          */
 639         final public function quickDelete( array $params ) {
 640                 return $this->doQuickOperation( array( 'op' => 'delete' ) + $params );
 641         }
 642
 643         /**
 644          * Performs a single quick describe operation.
 645          * This sets $params['op'] to 'describe' and passes it to doQuickOperation().
 646          *
 647          * @see FileBackend::doQuickOperation()
 648          *
 649          * @param $params Array Operation parameters
 650          * @return Status
 651          * @since 1.21
 652          */
 653         final public function quickDescribe( array $params ) {
 654                 return $this->doQuickOperation( array( 'op' => 'describe' ) + $params );
 655         }
 656
 657         /**
 658          * Concatenate a list of storage files into a single file system file.
 659          * The target path should refer to a file that is already locked or
 660          * otherwise safe from modification from other processes. Normally,
 661          * the file will be a new temp file, which should be adequate.
 662          *
 663          * @param $params Array Operation parameters
 664          * $params include:
 665          *   - srcs        : ordered source storage paths (e.g. chunk1, chunk2, ...)
 666          *   - dst         : file system path to 0-byte temp file
 667          *   - parallelize : try to do operations in parallel when possible
 668          * @return Status
 669          */
 670         abstract public function concatenate( array $params );
 671
 672         /**
 673          * Prepare a storage directory for usage.
 674          * This will create any required containers and parent directories.
 675          * Backends using key/value stores only need to create the container.
 676          *
 677          * The 'noAccess' and 'noListing' parameters works the same as in secure(),
 678          * except they are only applied *if* the directory/container had to be created.
 679          * These flags should always be set for directories that have private files.
 680          *
 681          * @param $params Array
 682          * $params include:
 683          *   - dir            : storage directory
 684          *   - noAccess       : try to deny file access (since 1.20)
 685          *   - noListing      : try to deny file listing (since 1.20)
 686          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 687          * @return Status
 688          */
 689         final public function prepare( array $params ) {
 690                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 691                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 692                 }
 693                 return $this->doPrepare( $params );
 694         }
 695
 696         /**
 697          * @see FileBackend::prepare()
 698          */
 699         abstract protected function doPrepare( array $params );
 700
 701         /**
 702          * Take measures to block web access to a storage directory and
 703          * the container it belongs to. FS backends might add .htaccess
 704          * files whereas key/value store backends might revoke container
 705          * access to the storage user representing end-users in web requests.
 706          * This is not guaranteed to actually do anything.
 707          *
 708          * @param $params Array
 709          * $params include:
 710          *   - dir            : storage directory
 711          *   - noAccess       : try to deny file access
 712          *   - noListing      : try to deny file listing
 713          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 714          * @return Status
 715          */
 716         final public function secure( array $params ) {
 717                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 718                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 719                 }
 720                 return $this->doSecure( $params );
 721         }
 722
 723         /**
 724          * @see FileBackend::secure()
 725          */
 726         abstract protected function doSecure( array $params );
 727
 728         /**
 729          * Remove measures to block web access to a storage directory and
 730          * the container it belongs to. FS backends might remove .htaccess
 731          * files whereas key/value store backends might grant container
 732          * access to the storage user representing end-users in web requests.
 733          * This essentially can undo the result of secure() calls.
 734          *
 735          * @param $params Array
 736          * $params include:
 737          *   - dir            : storage directory
 738          *   - access         : try to allow file access
 739          *   - listing        : try to allow file listing
 740          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 741          * @return Status
 742          * @since 1.20
 743          */
 744         final public function publish( array $params ) {
 745                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 746                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 747                 }
 748                 return $this->doPublish( $params );
 749         }
 750
 751         /**
 752          * @see FileBackend::publish()
 753          */
 754         abstract protected function doPublish( array $params );
 755
 756         /**
 757          * Delete a storage directory if it is empty.
 758          * Backends using key/value stores may do nothing unless the directory
 759          * is that of an empty container, in which case it will be deleted.
 760          *
 761          * @param $params Array
 762          * $params include:
 763          *   - dir            : storage directory
 764          *   - recursive      : recursively delete empty subdirectories first (since 1.20)
 765          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 766          * @return Status
 767          */
 768         final public function clean( array $params ) {
 769                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 770                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 771                 }
 772                 return $this->doClean( $params );
 773         }
 774
 775         /**
 776          * @see FileBackend::clean()
 777          */
 778         abstract protected function doClean( array $params );
 779
 780         /**
 781          * Check if a file exists at a storage path in the backend.
 782          * This returns false if only a directory exists at the path.
 783          *
 784          * @param $params Array
 785          * $params include:
 786          *   - src    : source storage path
 787          *   - latest : use the latest available data
 788          * @return bool|null Returns null on failure
 789          */
 790         abstract public function fileExists( array $params );
 791
 792         /**
 793          * Get the last-modified timestamp of the file at a storage path.
 794          *
 795          * @param $params Array
 796          * $params include:
 797          *   - src    : source storage path
 798          *   - latest : use the latest available data
 799          * @return string|bool TS_MW timestamp or false on failure
 800          */
 801         abstract public function getFileTimestamp( array $params );
 802
 803         /**
 804          * Get the contents of a file at a storage path in the backend.
 805          * This should be avoided for potentially large files.
 806          *
 807          * @param $params Array
 808          * $params include:
 809          *   - src    : source storage path
 810          *   - latest : use the latest available data
 811          * @return string|bool Returns false on failure
 812          */
 813         final public function getFileContents( array $params ) {
 814                 $contents = $this->getFileContentsMulti(
 815                         array( 'srcs' => array( $params['src'] ) ) + $params );
 816
 817                 return $contents[$params['src']];
 818         }
 819
 820         /**
 821          * Like getFileContents() except it takes an array of storage paths
 822          * and returns a map of storage paths to strings (or null on failure).
 823          * The map keys (paths) are in the same order as the provided list of paths.
 824          *
 825          * @see FileBackend::getFileContents()
 826          *
 827          * @param $params Array
 828          * $params include:
 829          *   - srcs        : list of source storage paths
 830          *   - latest      : use the latest available data
 831          *   - parallelize : try to do operations in parallel when possible
 832          * @return Array Map of (path name => string or false on failure)
 833          * @since 1.20
 834          */
 835         abstract public function getFileContentsMulti( array $params );
 836
 837         /**
 838          * Get the size (bytes) of a file at a storage path in the backend.
 839          *
 840          * @param $params Array
 841          * $params include:
 842          *   - src    : source storage path
 843          *   - latest : use the latest available data
 844          * @return integer|bool Returns false on failure
 845          */
 846         abstract public function getFileSize( array $params );
 847
 848         /**
 849          * Get quick information about a file at a storage path in the backend.
 850          * If the file does not exist, then this returns false.
 851          * Otherwise, the result is an associative array that includes:
 852          *   - mtime  : the last-modified timestamp (TS_MW)
 853          *   - size   : the file size (bytes)
 854          * Additional values may be included for internal use only.
 855          *
 856          * @param $params Array
 857          * $params include:
 858          *   - src    : source storage path
 859          *   - latest : use the latest available data
 860          * @return Array|bool|null Returns null on failure
 861          */
 862         abstract public function getFileStat( array $params );
 863
 864         /**
 865          * Get a SHA-1 hash of the file at a storage path in the backend.
 866          *
 867          * @param $params Array
 868          * $params include:
 869          *   - src    : source storage path
 870          *   - latest : use the latest available data
 871          * @return string|bool Hash string or false on failure
 872          */
 873         abstract public function getFileSha1Base36( array $params );
 874
 875         /**
 876          * Get the properties of the file at a storage path in the backend.
 877          * This gives the result of FSFile::getProps() on a local copy of the file.
 878          *
 879          * @param $params Array
 880          * $params include:
 881          *   - src    : source storage path
 882          *   - latest : use the latest available data
 883          * @return Array Returns FSFile::placeholderProps() on failure
 884          */
 885         abstract public function getFileProps( array $params );
 886
 887         /**
 888          * Stream the file at a storage path in the backend.
 889          * If the file does not exists, an HTTP 404 error will be given.
 890          * Appropriate HTTP headers (Status, Content-Type, Content-Length)
 891          * will be sent if streaming began, while none will be sent otherwise.
 892          * Implementations should flush the output buffer before sending data.
 893          *
 894          * @param $params Array
 895          * $params include:
 896          *   - src     : source storage path
 897          *   - headers : list of additional HTTP headers to send on success
 898          *   - latest  : use the latest available data
 899          * @return Status
 900          */
 901         abstract public function streamFile( array $params );
 902
 903         /**
 904          * Returns a file system file, identical to the file at a storage path.
 905          * The file returned is either:
 906          *   - a) A local copy of the file at a storage path in the backend.
 907          *        The temporary copy will have the same extension as the source.
 908          *   - b) An original of the file at a storage path in the backend.
 909          * Temporary files may be purged when the file object falls out of scope.
 910          *
 911          * Write operations should *never* be done on this file as some backends
 912          * may do internal tracking or may be instances of FileBackendMultiWrite.
 913          * In that later case, there are copies of the file that must stay in sync.
 914          * Additionally, further calls to this function may return the same file.
 915          *
 916          * @param $params Array
 917          * $params include:
 918          *   - src    : source storage path
 919          *   - latest : use the latest available data
 920          * @return FSFile|null Returns null on failure
 921          */
 922         final public function getLocalReference( array $params ) {
 923                 $fsFiles = $this->getLocalReferenceMulti(
 924                         array( 'srcs' => array( $params['src'] ) ) + $params );
 925
 926                 return $fsFiles[$params['src']];
 927         }
 928
 929         /**
 930          * Like getLocalReference() except it takes an array of storage paths
 931          * and returns a map of storage paths to FSFile objects (or null on failure).
 932          * The map keys (paths) are in the same order as the provided list of paths.
 933          *
 934          * @see FileBackend::getLocalReference()
 935          *
 936          * @param $params Array
 937          * $params include:
 938          *   - srcs        : list of source storage paths
 939          *   - latest      : use the latest available data
 940          *   - parallelize : try to do operations in parallel when possible
 941          * @return Array Map of (path name => FSFile or null on failure)
 942          * @since 1.20
 943          */
 944         abstract public function getLocalReferenceMulti( array $params );
 945
 946         /**
 947          * Get a local copy on disk of the file at a storage path in the backend.
 948          * The temporary copy will have the same file extension as the source.
 949          * Temporary files may be purged when the file object falls out of scope.
 950          *
 951          * @param $params Array
 952          * $params include:
 953          *   - src    : source storage path
 954          *   - latest : use the latest available data
 955          * @return TempFSFile|null Returns null on failure
 956          */
 957         final public function getLocalCopy( array $params ) {
 958                 $tmpFiles = $this->getLocalCopyMulti(
 959                         array( 'srcs' => array( $params['src'] ) ) + $params );
 960
 961                 return $tmpFiles[$params['src']];
 962         }
 963
 964         /**
 965          * Like getLocalCopy() except it takes an array of storage paths and
 966          * returns a map of storage paths to TempFSFile objects (or null on failure).
 967          * The map keys (paths) are in the same order as the provided list of paths.
 968          *
 969          * @see FileBackend::getLocalCopy()
 970          *
 971          * @param $params Array
 972          * $params include:
 973          *   - srcs        : list of source storage paths
 974          *   - latest      : use the latest available data
 975          *   - parallelize : try to do operations in parallel when possible
 976          * @return Array Map of (path name => TempFSFile or null on failure)
 977          * @since 1.20
 978          */
 979         abstract public function getLocalCopyMulti( array $params );
 980
 981         /**
 982          * Return an HTTP URL to a given file that requires no authentication to use.
 983          * The URL may be pre-authenticated (via some token in the URL) and temporary.
 984          * This will return null if the backend cannot make an HTTP URL for the file.
 985          *
 986          * This is useful for key/value stores when using scripts that seek around
 987          * large files and those scripts (and the backend) support HTTP Range headers.
 988          * Otherwise, one would need to use getLocalReference(), which involves loading
 989          * the entire file on to local disk.
 990          *
 991          * @param $params Array
 992          * $params include:
 993          *   - src : source storage path
 994          * @return string|null
 995          * @since 1.21
 996          */
 997         abstract public function getFileHttpUrl( array $params );
 998
 999         /**
1000          * Check if a directory exists at a given storage path.
1001          * Backends using key/value stores will check if the path is a
1002          * virtual directory, meaning there are files under the given directory.
1003          *
1004          * Storage backends with eventual consistency might return stale data.
1005          *
1006          * @param $params array
1007          * $params include:
1008          *   - dir : storage directory
1009          * @return bool|null Returns null on failure
1010          * @since 1.20
1011          */
1012         abstract public function directoryExists( array $params );
1013
1014         /**
1015          * Get an iterator to list *all* directories under a storage directory.
1016          * If the directory is of the form "mwstore://backend/container",
1017          * then all directories in the container will be listed.
1018          * If the directory is of form "mwstore://backend/container/dir",
1019          * then all directories directly under that directory will be listed.
1020          * Results will be storage directories relative to the given directory.
1021          *
1022          * Storage backends with eventual consistency might return stale data.
1023          *
1024          * @param $params array
1025          * $params include:
1026          *   - dir     : storage directory
1027          *   - topOnly : only return direct child dirs of the directory
1028          * @return Traversable|Array|null Returns null on failure
1029          * @since 1.20
1030          */
1031         abstract public function getDirectoryList( array $params );
1032
1033         /**
1034          * Same as FileBackend::getDirectoryList() except only lists
1035          * directories that are immediately under the given directory.
1036          *
1037          * Storage backends with eventual consistency might return stale data.
1038          *
1039          * @param $params array
1040          * $params include:
1041          *   - dir : storage directory
1042          * @return Traversable|Array|null Returns null on failure
1043          * @since 1.20
1044          */
1045         final public function getTopDirectoryList( array $params ) {
1046                 return $this->getDirectoryList( array( 'topOnly' => true ) + $params );
1047         }
1048
1049         /**
1050          * Get an iterator to list *all* stored files under a storage directory.
1051          * If the directory is of the form "mwstore://backend/container",
1052          * then all files in the container will be listed.
1053          * If the directory is of form "mwstore://backend/container/dir",
1054          * then all files under that directory will be listed.
1055          * Results will be storage paths relative to the given directory.
1056          *
1057          * Storage backends with eventual consistency might return stale data.
1058          *
1059          * @param $params array
1060          * $params include:
1061          *   - dir     : storage directory
1062          *   - topOnly : only return direct child files of the directory (since 1.20)
1063          * @return Traversable|Array|null Returns null on failure
1064          */
1065         abstract public function getFileList( array $params );
1066
1067         /**
1068          * Same as FileBackend::getFileList() except only lists
1069          * files that are immediately under the given directory.
1070          *
1071          * Storage backends with eventual consistency might return stale data.
1072          *
1073          * @param $params array
1074          * $params include:
1075          *   - dir : storage directory
1076          * @return Traversable|Array|null Returns null on failure
1077          * @since 1.20
1078          */
1079         final public function getTopFileList( array $params ) {
1080                 return $this->getFileList( array( 'topOnly' => true ) + $params );
1081         }
1082
1083         /**
1084          * Preload persistent file stat and property cache into in-process cache.
1085          * This should be used when stat calls will be made on a known list of a many files.
1086          *
1087          * @param $paths Array Storage paths
1088          * @return void
1089          */
1090         public function preloadCache( array $paths ) {}
1091
1092         /**
1093          * Invalidate any in-process file stat and property cache.
1094          * If $paths is given, then only the cache for those files will be cleared.
1095          *
1096          * @param $paths Array Storage paths (optional)
1097          * @return void
1098          */
1099         public function clearCache( array $paths = null ) {}
1100
1101         /**
1102          * Lock the files at the given storage paths in the backend.
1103          * This will either lock all the files or none (on failure).
1104          *
1105          * Callers should consider using getScopedFileLocks() instead.
1106          *
1107          * @param $paths Array Storage paths
1108          * @param $type integer LockManager::LOCK_* constant
1109          * @return Status
1110          */
1111         final public function lockFiles( array $paths, $type ) {
1112                 return $this->lockManager->lock( $paths, $type );
1113         }
1114
1115         /**
1116          * Unlock the files at the given storage paths in the backend.
1117          *
1118          * @param $paths Array Storage paths
1119          * @param $type integer LockManager::LOCK_* constant
1120          * @return Status
1121          */
1122         final public function unlockFiles( array $paths, $type ) {
1123                 return $this->lockManager->unlock( $paths, $type );
1124         }
1125
1126         /**
1127          * Lock the files at the given storage paths in the backend.
1128          * This will either lock all the files or none (on failure).
1129          * On failure, the status object will be updated with errors.
1130          *
1131          * Once the return value goes out scope, the locks will be released and
1132          * the status updated. Unlock fatals will not change the status "OK" value.
1133          *
1134          * @param $paths Array Storage paths
1135          * @param $type integer LockManager::LOCK_* constant
1136          * @param $status Status Status to update on lock/unlock
1137          * @return ScopedLock|null Returns null on failure
1138          */
1139         final public function getScopedFileLocks( array $paths, $type, Status $status ) {
1140                 return ScopedLock::factory( $this->lockManager, $paths, $type, $status );
1141         }
1142
1143         /**
1144          * Get an array of scoped locks needed for a batch of file operations.
1145          *
1146          * Normally, FileBackend::doOperations() handles locking, unless
1147          * the 'nonLocking' param is passed in. This function is useful if you
1148          * want the files to be locked for a broader scope than just when the
1149          * files are changing. For example, if you need to update DB metadata,
1150          * you may want to keep the files locked until finished.
1151          *
1152          * @see FileBackend::doOperations()
1153          *
1154          * @param $ops Array List of file operations to FileBackend::doOperations()
1155          * @param $status Status Status to update on lock/unlock
1156          * @return Array List of ScopedFileLocks or null values
1157          * @since 1.20
1158          */
1159         abstract public function getScopedLocksForOps( array $ops, Status $status );
1160
1161         /**
1162          * Get the root storage path of this backend.
1163          * All container paths are "subdirectories" of this path.
1164          *
1165          * @return string Storage path
1166          * @since 1.20
1167          */
1168         final public function getRootStoragePath() {
1169                 return "mwstore://{$this->name}";
1170         }
1171
1172         /**
1173          * Get the storage path for the given container for this backend
1174          *
1175          * @param $container string Container name
1176          * @return string Storage path
1177          * @since 1.21
1178          */
1179         final public function getContainerStoragePath( $container ) {
1180                 return $this->getRootStoragePath() . "/{$container}";
1181         }
1182
1183         /**
1184          * Get the file journal object for this backend
1185          *
1186          * @return FileJournal
1187          */
1188         final public function getJournal() {
1189                 return $this->fileJournal;
1190         }
1191
1192         /**
1193          * Check if a given path is a "mwstore://" path.
1194          * This does not do any further validation or any existence checks.
1195          *
1196          * @param $path string
1197          * @return bool
1198          */
1199         final public static function isStoragePath( $path ) {
1200                 return ( strpos( $path, 'mwstore://' ) === 0 );
1201         }
1202
1203         /**
1204          * Split a storage path into a backend name, a container name,
1205          * and a relative file path. The relative path may be the empty string.
1206          * This does not do any path normalization or traversal checks.
1207          *
1208          * @param $storagePath string
1209          * @return Array (backend, container, rel object) or (null, null, null)
1210          */
1211         final public static function splitStoragePath( $storagePath ) {
1212                 if ( self::isStoragePath( $storagePath ) ) {
1213                         // Remove the "mwstore://" prefix and split the path
1214                         $parts = explode( '/', substr( $storagePath, 10 ), 3 );
1215                         if ( count( $parts ) >= 2 && $parts[0] != '' && $parts[1] != '' ) {
1216                                 if ( count( $parts ) == 3 ) {
1217                                         return $parts; // e.g. "backend/container/path"
1218                                 } else {
1219                                         return array( $parts[0], $parts[1], '' ); // e.g. "backend/container"
1220                                 }
1221                         }
1222                 }
1223                 return array( null, null, null );
1224         }
1225
1226         /**
1227          * Normalize a storage path by cleaning up directory separators.
1228          * Returns null if the path is not of the format of a valid storage path.
1229          *
1230          * @param $storagePath string
1231          * @return string|null
1232          */
1233         final public static function normalizeStoragePath( $storagePath ) {
1234                 list( $backend, $container, $relPath ) = self::splitStoragePath( $storagePath );
1235                 if ( $relPath !== null ) { // must be for this backend
1236                         $relPath = self::normalizeContainerPath( $relPath );
1237                         if ( $relPath !== null ) {
1238                                 return ( $relPath != '' )
1239                                         ? "mwstore://{$backend}/{$container}/{$relPath}"
1240                                         : "mwstore://{$backend}/{$container}";
1241                         }
1242                 }
1243                 return null;
1244         }
1245
1246         /**
1247          * Get the parent storage directory of a storage path.
1248          * This returns a path like "mwstore://backend/container",
1249          * "mwstore://backend/container/...", or null if there is no parent.
1250          *
1251          * @param $storagePath string
1252          * @return string|null
1253          */
1254         final public static function parentStoragePath( $storagePath ) {
1255                 $storagePath = dirname( $storagePath );
1256                 list( $b, $cont, $rel ) = self::splitStoragePath( $storagePath );
1257                 return ( $rel === null ) ? null : $storagePath;
1258         }
1259
1260         /**
1261          * Get the final extension from a storage or FS path
1262          *
1263          * @param $path string
1264          * @return string
1265          */
1266         final public static function extensionFromPath( $path ) {
1267                 $i = strrpos( $path, '.' );
1268                 return strtolower( $i ? substr( $path, $i + 1 ) : '' );
1269         }
1270
1271         /**
1272          * Check if a relative path has no directory traversals
1273          *
1274          * @param $path string
1275          * @return bool
1276          * @since 1.20
1277          */
1278         final public static function isPathTraversalFree( $path ) {
1279                 return ( self::normalizeContainerPath( $path ) !== null );
1280         }
1281
1282         /**
1283          * Build a Content-Disposition header value per RFC 6266.
1284          *
1285          * @param $type string One of (attachment, inline)
1286          * @param $filename string Suggested file name (should not contain slashes)
1287          * @throws MWException
1288          * @return string
1289          * @since 1.20
1290          */
1291         final public static function makeContentDisposition( $type, $filename = '' ) {
1292                 $parts = array();
1293
1294                 $type = strtolower( $type );
1295                 if ( !in_array( $type, array( 'inline', 'attachment' ) ) ) {
1296                         throw new MWException( "Invalid Content-Disposition type '$type'." );
1297                 }
1298                 $parts[] = $type;
1299
1300                 if ( strlen( $filename ) ) {
1301                         $parts[] = "filename*=UTF-8''" . rawurlencode( basename( $filename ) );
1302                 }
1303
1304                 return implode( ';', $parts );
1305         }
1306
1307         /**
1308          * Validate and normalize a relative storage path.
1309          * Null is returned if the path involves directory traversal.
1310          * Traversal is insecure for FS backends and broken for others.
1311          *
1312          * This uses the same traversal protection as Title::secureAndSplit().
1313          *
1314          * @param $path string Storage path relative to a container
1315          * @return string|null
1316          */
1317         final protected static function normalizeContainerPath( $path ) {
1318                 // Normalize directory separators
1319                 $path = strtr( $path, '\\', '/' );
1320                 // Collapse any consecutive directory separators
1321                 $path = preg_replace( '![/]{2,}!', '/', $path );
1322                 // Remove any leading directory separator
1323                 $path = ltrim( $path, '/' );
1324                 // Use the same traversal protection as Title::secureAndSplit()
1325                 if ( strpos( $path, '.' ) !== false ) {
1326                         if (
1327                                 $path === '.' ||
1328                                 $path === '..' ||
1329                                 strpos( $path, './' ) === 0 ||
1330                                 strpos( $path, '../' ) === 0 ||
1331                                 strpos( $path, '/./' ) !== false ||
1332                                 strpos( $path, '/../' ) !== false
1333                         ) {
1334                                 return null;
1335                         }
1336                 }
1337                 return $path;
1338         }
1339 }