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          * If a callback function is given, it will be called each time a segment is
 664          * appended and when the overall concatenate operation completes or fails.
 665          * The arguments passed in are:
 666          *   - 1) A Status object containing errors if any problems occurred.
 667          *   - 2) The index of the relevant segment (starting at 1) if a segment was appended
 668          *        (including the last one) or null in the case of overall success or failure.
 669          * When a good Status is returned with a null segment, then the operation completed.
 670          * Callbacks should generally avoid throwing exceptions.
 671          *
 672          * @param $params Array Operation parameters
 673          * $params include:
 674          *   - srcs        : ordered source storage paths (e.g. chunk1, chunk2, ...)
 675          *   - dst         : file system path to 0-byte temp file
 676          *   - parallelize : try to do operations in parallel when possible
 677          *   - callback    : closure called when chunks are appended and on success/failure
 678          * @return Status
 679          */
 680         abstract public function concatenate( array $params );
 681
 682         /**
 683          * Prepare a storage directory for usage.
 684          * This will create any required containers and parent directories.
 685          * Backends using key/value stores only need to create the container.
 686          *
 687          * The 'noAccess' and 'noListing' parameters works the same as in secure(),
 688          * except they are only applied *if* the directory/container had to be created.
 689          * These flags should always be set for directories that have private files.
 690          *
 691          * @param $params Array
 692          * $params include:
 693          *   - dir            : storage directory
 694          *   - noAccess       : try to deny file access (since 1.20)
 695          *   - noListing      : try to deny file listing (since 1.20)
 696          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 697          * @return Status
 698          */
 699         final public function prepare( array $params ) {
 700                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 701                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 702                 }
 703                 return $this->doPrepare( $params );
 704         }
 705
 706         /**
 707          * @see FileBackend::prepare()
 708          */
 709         abstract protected function doPrepare( array $params );
 710
 711         /**
 712          * Take measures to block web access to a storage directory and
 713          * the container it belongs to. FS backends might add .htaccess
 714          * files whereas key/value store backends might revoke container
 715          * access to the storage user representing end-users in web requests.
 716          * This is not guaranteed to actually do anything.
 717          *
 718          * @param $params Array
 719          * $params include:
 720          *   - dir            : storage directory
 721          *   - noAccess       : try to deny file access
 722          *   - noListing      : try to deny file listing
 723          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 724          * @return Status
 725          */
 726         final public function secure( array $params ) {
 727                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 728                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 729                 }
 730                 return $this->doSecure( $params );
 731         }
 732
 733         /**
 734          * @see FileBackend::secure()
 735          */
 736         abstract protected function doSecure( array $params );
 737
 738         /**
 739          * Remove measures to block web access to a storage directory and
 740          * the container it belongs to. FS backends might remove .htaccess
 741          * files whereas key/value store backends might grant container
 742          * access to the storage user representing end-users in web requests.
 743          * This essentially can undo the result of secure() calls.
 744          *
 745          * @param $params Array
 746          * $params include:
 747          *   - dir            : storage directory
 748          *   - access         : try to allow file access
 749          *   - listing        : try to allow file listing
 750          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 751          * @return Status
 752          * @since 1.20
 753          */
 754         final public function publish( array $params ) {
 755                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 756                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 757                 }
 758                 return $this->doPublish( $params );
 759         }
 760
 761         /**
 762          * @see FileBackend::publish()
 763          */
 764         abstract protected function doPublish( array $params );
 765
 766         /**
 767          * Delete a storage directory if it is empty.
 768          * Backends using key/value stores may do nothing unless the directory
 769          * is that of an empty container, in which case it will be deleted.
 770          *
 771          * @param $params Array
 772          * $params include:
 773          *   - dir            : storage directory
 774          *   - recursive      : recursively delete empty subdirectories first (since 1.20)
 775          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 776          * @return Status
 777          */
 778         final public function clean( array $params ) {
 779                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 780                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 781                 }
 782                 return $this->doClean( $params );
 783         }
 784
 785         /**
 786          * @see FileBackend::clean()
 787          */
 788         abstract protected function doClean( array $params );
 789
 790         /**
 791          * Check if a file exists at a storage path in the backend.
 792          * This returns false if only a directory exists at the path.
 793          *
 794          * @param $params Array
 795          * $params include:
 796          *   - src    : source storage path
 797          *   - latest : use the latest available data
 798          * @return bool|null Returns null on failure
 799          */
 800         abstract public function fileExists( array $params );
 801
 802         /**
 803          * Get the last-modified timestamp of the file at a storage path.
 804          *
 805          * @param $params Array
 806          * $params include:
 807          *   - src    : source storage path
 808          *   - latest : use the latest available data
 809          * @return string|bool TS_MW timestamp or false on failure
 810          */
 811         abstract public function getFileTimestamp( array $params );
 812
 813         /**
 814          * Get the contents of a file at a storage path in the backend.
 815          * This should be avoided for potentially large files.
 816          *
 817          * @param $params Array
 818          * $params include:
 819          *   - src    : source storage path
 820          *   - latest : use the latest available data
 821          * @return string|bool Returns false on failure
 822          */
 823         final public function getFileContents( array $params ) {
 824                 $contents = $this->getFileContentsMulti(
 825                         array( 'srcs' => array( $params['src'] ) ) + $params );
 826
 827                 return $contents[$params['src']];
 828         }
 829
 830         /**
 831          * Like getFileContents() except it takes an array of storage paths
 832          * and returns a map of storage paths to strings (or null on failure).
 833          * The map keys (paths) are in the same order as the provided list of paths.
 834          *
 835          * @see FileBackend::getFileContents()
 836          *
 837          * @param $params Array
 838          * $params include:
 839          *   - srcs        : list of source storage paths
 840          *   - latest      : use the latest available data
 841          *   - parallelize : try to do operations in parallel when possible
 842          * @return Array Map of (path name => string or false on failure)
 843          * @since 1.20
 844          */
 845         abstract public function getFileContentsMulti( array $params );
 846
 847         /**
 848          * Get the size (bytes) of a file at a storage path in the backend.
 849          *
 850          * @param $params Array
 851          * $params include:
 852          *   - src    : source storage path
 853          *   - latest : use the latest available data
 854          * @return integer|bool Returns false on failure
 855          */
 856         abstract public function getFileSize( array $params );
 857
 858         /**
 859          * Get quick information about a file at a storage path in the backend.
 860          * If the file does not exist, then this returns false.
 861          * Otherwise, the result is an associative array that includes:
 862          *   - mtime  : the last-modified timestamp (TS_MW)
 863          *   - size   : the file size (bytes)
 864          * Additional values may be included for internal use only.
 865          *
 866          * @param $params Array
 867          * $params include:
 868          *   - src    : source storage path
 869          *   - latest : use the latest available data
 870          * @return Array|bool|null Returns null on failure
 871          */
 872         abstract public function getFileStat( array $params );
 873
 874         /**
 875          * Get a SHA-1 hash of the file at a storage path in the backend.
 876          *
 877          * @param $params Array
 878          * $params include:
 879          *   - src    : source storage path
 880          *   - latest : use the latest available data
 881          * @return string|bool Hash string or false on failure
 882          */
 883         abstract public function getFileSha1Base36( array $params );
 884
 885         /**
 886          * Get the properties of the file at a storage path in the backend.
 887          * Returns FSFile::placeholderProps() on failure.
 888          *
 889          * @param $params Array
 890          * $params include:
 891          *   - src    : source storage path
 892          *   - latest : use the latest available data
 893          * @return Array
 894          */
 895         abstract public function getFileProps( array $params );
 896
 897         /**
 898          * Stream the file at a storage path in the backend.
 899          * If the file does not exists, an HTTP 404 error will be given.
 900          * Appropriate HTTP headers (Status, Content-Type, Content-Length)
 901          * will be sent if streaming began, while none will be sent otherwise.
 902          * Implementations should flush the output buffer before sending data.
 903          *
 904          * @param $params Array
 905          * $params include:
 906          *   - src     : source storage path
 907          *   - headers : list of additional HTTP headers to send on success
 908          *   - latest  : use the latest available data
 909          * @return Status
 910          */
 911         abstract public function streamFile( array $params );
 912
 913         /**
 914          * Returns a file system file, identical to the file at a storage path.
 915          * The file returned is either:
 916          *   - a) A local copy of the file at a storage path in the backend.
 917          *        The temporary copy will have the same extension as the source.
 918          *   - b) An original of the file at a storage path in the backend.
 919          * Temporary files may be purged when the file object falls out of scope.
 920          *
 921          * Write operations should *never* be done on this file as some backends
 922          * may do internal tracking or may be instances of FileBackendMultiWrite.
 923          * In that later case, there are copies of the file that must stay in sync.
 924          * Additionally, further calls to this function may return the same file.
 925          *
 926          * @param $params Array
 927          * $params include:
 928          *   - src    : source storage path
 929          *   - latest : use the latest available data
 930          * @return FSFile|null Returns null on failure
 931          */
 932         final public function getLocalReference( array $params ) {
 933                 $fsFiles = $this->getLocalReferenceMulti(
 934                         array( 'srcs' => array( $params['src'] ) ) + $params );
 935
 936                 return $fsFiles[$params['src']];
 937         }
 938
 939         /**
 940          * Like getLocalReference() except it takes an array of storage paths
 941          * and returns a map of storage paths to FSFile objects (or null on failure).
 942          * The map keys (paths) are in the same order as the provided list of paths.
 943          *
 944          * @see FileBackend::getLocalReference()
 945          *
 946          * @param $params Array
 947          * $params include:
 948          *   - srcs        : list of source storage paths
 949          *   - latest      : use the latest available data
 950          *   - parallelize : try to do operations in parallel when possible
 951          * @return Array Map of (path name => FSFile or null on failure)
 952          * @since 1.20
 953          */
 954         abstract public function getLocalReferenceMulti( array $params );
 955
 956         /**
 957          * Get a local copy on disk of the file at a storage path in the backend.
 958          * The temporary copy will have the same file extension as the source.
 959          * Temporary files may be purged when the file object falls out of scope.
 960          *
 961          * @param $params Array
 962          * $params include:
 963          *   - src    : source storage path
 964          *   - latest : use the latest available data
 965          * @return TempFSFile|null Returns null on failure
 966          */
 967         final public function getLocalCopy( array $params ) {
 968                 $tmpFiles = $this->getLocalCopyMulti(
 969                         array( 'srcs' => array( $params['src'] ) ) + $params );
 970
 971                 return $tmpFiles[$params['src']];
 972         }
 973
 974         /**
 975          * Like getLocalCopy() except it takes an array of storage paths and
 976          * returns a map of storage paths to TempFSFile objects (or null on failure).
 977          * The map keys (paths) are in the same order as the provided list of paths.
 978          *
 979          * @see FileBackend::getLocalCopy()
 980          *
 981          * @param $params Array
 982          * $params include:
 983          *   - srcs        : list of source storage paths
 984          *   - latest      : use the latest available data
 985          *   - parallelize : try to do operations in parallel when possible
 986          * @return Array Map of (path name => TempFSFile or null on failure)
 987          * @since 1.20
 988          */
 989         abstract public function getLocalCopyMulti( array $params );
 990
 991         /**
 992          * Return an HTTP URL to a given file that requires no authentication to use.
 993          * The URL may be pre-authenticated (via some token in the URL) and temporary.
 994          * This will return null if the backend cannot make an HTTP URL for the file.
 995          *
 996          * This is useful for key/value stores when using scripts that seek around
 997          * large files and those scripts (and the backend) support HTTP Range headers.
 998          * Otherwise, one would need to use getLocalReference(), which involves loading
 999          * the entire file on to local disk.
1000          *
1001          * @param $params Array
1002          * $params include:
1003          *   - src : source storage path
1004          * @return string|null
1005          * @since 1.21
1006          */
1007         abstract public function getFileHttpUrl( array $params );
1008
1009         /**
1010          * Check if a directory exists at a given storage path.
1011          * Backends using key/value stores will check if the path is a
1012          * virtual directory, meaning there are files under the given directory.
1013          *
1014          * Storage backends with eventual consistency might return stale data.
1015          *
1016          * @param $params array
1017          * $params include:
1018          *   - dir : storage directory
1019          * @return bool|null Returns null on failure
1020          * @since 1.20
1021          */
1022         abstract public function directoryExists( array $params );
1023
1024         /**
1025          * Get an iterator to list *all* directories under a storage directory.
1026          * If the directory is of the form "mwstore://backend/container",
1027          * then all directories in the container will be listed.
1028          * If the directory is of form "mwstore://backend/container/dir",
1029          * then all directories directly under that directory will be listed.
1030          * Results will be storage directories relative to the given directory.
1031          *
1032          * Storage backends with eventual consistency might return stale data.
1033          *
1034          * @param $params array
1035          * $params include:
1036          *   - dir     : storage directory
1037          *   - topOnly : only return direct child dirs of the directory
1038          * @return Traversable|Array|null Returns null on failure
1039          * @since 1.20
1040          */
1041         abstract public function getDirectoryList( array $params );
1042
1043         /**
1044          * Same as FileBackend::getDirectoryList() except only lists
1045          * directories that are immediately under the given directory.
1046          *
1047          * Storage backends with eventual consistency might return stale data.
1048          *
1049          * @param $params array
1050          * $params include:
1051          *   - dir : storage directory
1052          * @return Traversable|Array|null Returns null on failure
1053          * @since 1.20
1054          */
1055         final public function getTopDirectoryList( array $params ) {
1056                 return $this->getDirectoryList( array( 'topOnly' => true ) + $params );
1057         }
1058
1059         /**
1060          * Get an iterator to list *all* stored files under a storage directory.
1061          * If the directory is of the form "mwstore://backend/container",
1062          * then all files in the container will be listed.
1063          * If the directory is of form "mwstore://backend/container/dir",
1064          * then all files under that directory will be listed.
1065          * Results will be storage paths relative to the given directory.
1066          *
1067          * Storage backends with eventual consistency might return stale data.
1068          *
1069          * @param $params array
1070          * $params include:
1071          *   - dir     : storage directory
1072          *   - topOnly : only return direct child files of the directory (since 1.20)
1073          * @return Traversable|Array|null Returns null on failure
1074          */
1075         abstract public function getFileList( array $params );
1076
1077         /**
1078          * Same as FileBackend::getFileList() except only lists
1079          * files that are immediately under the given directory.
1080          *
1081          * Storage backends with eventual consistency might return stale data.
1082          *
1083          * @param $params array
1084          * $params include:
1085          *   - dir : storage directory
1086          * @return Traversable|Array|null Returns null on failure
1087          * @since 1.20
1088          */
1089         final public function getTopFileList( array $params ) {
1090                 return $this->getFileList( array( 'topOnly' => true ) + $params );
1091         }
1092
1093         /**
1094          * Preload persistent file stat and property cache into in-process cache.
1095          * This should be used when stat calls will be made on a known list of a many files.
1096          *
1097          * @param $paths Array Storage paths
1098          * @return void
1099          */
1100         public function preloadCache( array $paths ) {}
1101
1102         /**
1103          * Invalidate any in-process file stat and property cache.
1104          * If $paths is given, then only the cache for those files will be cleared.
1105          *
1106          * @param $paths Array Storage paths (optional)
1107          * @return void
1108          */
1109         public function clearCache( array $paths = null ) {}
1110
1111         /**
1112          * Lock the files at the given storage paths in the backend.
1113          * This will either lock all the files or none (on failure).
1114          *
1115          * Callers should consider using getScopedFileLocks() instead.
1116          *
1117          * @param $paths Array Storage paths
1118          * @param $type integer LockManager::LOCK_* constant
1119          * @return Status
1120          */
1121         final public function lockFiles( array $paths, $type ) {
1122                 return $this->lockManager->lock( $paths, $type );
1123         }
1124
1125         /**
1126          * Unlock the files at the given storage paths in the backend.
1127          *
1128          * @param $paths Array Storage paths
1129          * @param $type integer LockManager::LOCK_* constant
1130          * @return Status
1131          */
1132         final public function unlockFiles( array $paths, $type ) {
1133                 return $this->lockManager->unlock( $paths, $type );
1134         }
1135
1136         /**
1137          * Lock the files at the given storage paths in the backend.
1138          * This will either lock all the files or none (on failure).
1139          * On failure, the status object will be updated with errors.
1140          *
1141          * Once the return value goes out scope, the locks will be released and
1142          * the status updated. Unlock fatals will not change the status "OK" value.
1143          *
1144          * @param $paths Array Storage paths
1145          * @param $type integer LockManager::LOCK_* constant
1146          * @param $status Status Status to update on lock/unlock
1147          * @return ScopedLock|null Returns null on failure
1148          */
1149         final public function getScopedFileLocks( array $paths, $type, Status $status ) {
1150                 return ScopedLock::factory( $this->lockManager, $paths, $type, $status );
1151         }
1152
1153         /**
1154          * Get an array of scoped locks needed for a batch of file operations.
1155          *
1156          * Normally, FileBackend::doOperations() handles locking, unless
1157          * the 'nonLocking' param is passed in. This function is useful if you
1158          * want the files to be locked for a broader scope than just when the
1159          * files are changing. For example, if you need to update DB metadata,
1160          * you may want to keep the files locked until finished.
1161          *
1162          * @see FileBackend::doOperations()
1163          *
1164          * @param $ops Array List of file operations to FileBackend::doOperations()
1165          * @param $status Status Status to update on lock/unlock
1166          * @return Array List of ScopedFileLocks or null values
1167          * @since 1.20
1168          */
1169         abstract public function getScopedLocksForOps( array $ops, Status $status );
1170
1171         /**
1172          * Get the root storage path of this backend.
1173          * All container paths are "subdirectories" of this path.
1174          *
1175          * @return string Storage path
1176          * @since 1.20
1177          */
1178         final public function getRootStoragePath() {
1179                 return "mwstore://{$this->name}";
1180         }
1181
1182         /**
1183          * Get the storage path for the given container for this backend
1184          *
1185          * @param $container string Container name
1186          * @return string Storage path
1187          * @since 1.21
1188          */
1189         final public function getContainerStoragePath( $container ) {
1190                 return $this->getRootStoragePath() . "/{$container}";
1191         }
1192
1193         /**
1194          * Get the file journal object for this backend
1195          *
1196          * @return FileJournal
1197          */
1198         final public function getJournal() {
1199                 return $this->fileJournal;
1200         }
1201
1202         /**
1203          * Check if a given path is a "mwstore://" path.
1204          * This does not do any further validation or any existence checks.
1205          *
1206          * @param $path string
1207          * @return bool
1208          */
1209         final public static function isStoragePath( $path ) {
1210                 return ( strpos( $path, 'mwstore://' ) === 0 );
1211         }
1212
1213         /**
1214          * Split a storage path into a backend name, a container name,
1215          * and a relative file path. The relative path may be the empty string.
1216          * This does not do any path normalization or traversal checks.
1217          *
1218          * @param $storagePath string
1219          * @return Array (backend, container, rel object) or (null, null, null)
1220          */
1221         final public static function splitStoragePath( $storagePath ) {
1222                 if ( self::isStoragePath( $storagePath ) ) {
1223                         // Remove the "mwstore://" prefix and split the path
1224                         $parts = explode( '/', substr( $storagePath, 10 ), 3 );
1225                         if ( count( $parts ) >= 2 && $parts[0] != '' && $parts[1] != '' ) {
1226                                 if ( count( $parts ) == 3 ) {
1227                                         return $parts; // e.g. "backend/container/path"
1228                                 } else {
1229                                         return array( $parts[0], $parts[1], '' ); // e.g. "backend/container"
1230                                 }
1231                         }
1232                 }
1233                 return array( null, null, null );
1234         }
1235
1236         /**
1237          * Normalize a storage path by cleaning up directory separators.
1238          * Returns null if the path is not of the format of a valid storage path.
1239          *
1240          * @param $storagePath string
1241          * @return string|null
1242          */
1243         final public static function normalizeStoragePath( $storagePath ) {
1244                 list( $backend, $container, $relPath ) = self::splitStoragePath( $storagePath );
1245                 if ( $relPath !== null ) { // must be for this backend
1246                         $relPath = self::normalizeContainerPath( $relPath );
1247                         if ( $relPath !== null ) {
1248                                 return ( $relPath != '' )
1249                                         ? "mwstore://{$backend}/{$container}/{$relPath}"
1250                                         : "mwstore://{$backend}/{$container}";
1251                         }
1252                 }
1253                 return null;
1254         }
1255
1256         /**
1257          * Get the parent storage directory of a storage path.
1258          * This returns a path like "mwstore://backend/container",
1259          * "mwstore://backend/container/...", or null if there is no parent.
1260          *
1261          * @param $storagePath string
1262          * @return string|null
1263          */
1264         final public static function parentStoragePath( $storagePath ) {
1265                 $storagePath = dirname( $storagePath );
1266                 list( $b, $cont, $rel ) = self::splitStoragePath( $storagePath );
1267                 return ( $rel === null ) ? null : $storagePath;
1268         }
1269
1270         /**
1271          * Get the final extension from a storage or FS path
1272          *
1273          * @param $path string
1274          * @return string
1275          */
1276         final public static function extensionFromPath( $path ) {
1277                 $i = strrpos( $path, '.' );
1278                 return strtolower( $i ? substr( $path, $i + 1 ) : '' );
1279         }
1280
1281         /**
1282          * Check if a relative path has no directory traversals
1283          *
1284          * @param $path string
1285          * @return bool
1286          * @since 1.20
1287          */
1288         final public static function isPathTraversalFree( $path ) {
1289                 return ( self::normalizeContainerPath( $path ) !== null );
1290         }
1291
1292         /**
1293          * Build a Content-Disposition header value per RFC 6266.
1294          *
1295          * @param $type string One of (attachment, inline)
1296          * @param $filename string Suggested file name (should not contain slashes)
1297          * @throws MWException
1298          * @return string
1299          * @since 1.20
1300          */
1301         final public static function makeContentDisposition( $type, $filename = '' ) {
1302                 $parts = array();
1303
1304                 $type = strtolower( $type );
1305                 if ( !in_array( $type, array( 'inline', 'attachment' ) ) ) {
1306                         throw new MWException( "Invalid Content-Disposition type '$type'." );
1307                 }
1308                 $parts[] = $type;
1309
1310                 if ( strlen( $filename ) ) {
1311                         $parts[] = "filename*=UTF-8''" . rawurlencode( basename( $filename ) );
1312                 }
1313
1314                 return implode( ';', $parts );
1315         }
1316
1317         /**
1318          * Validate and normalize a relative storage path.
1319          * Null is returned if the path involves directory traversal.
1320          * Traversal is insecure for FS backends and broken for others.
1321          *
1322          * This uses the same traversal protection as Title::secureAndSplit().
1323          *
1324          * @param $path string Storage path relative to a container
1325          * @return string|null
1326          */
1327         final protected static function normalizeContainerPath( $path ) {
1328                 // Normalize directory separators
1329                 $path = strtr( $path, '\\', '/' );
1330                 // Collapse any consecutive directory separators
1331                 $path = preg_replace( '![/]{2,}!', '/', $path );
1332                 // Remove any leading directory separator
1333                 $path = ltrim( $path, '/' );
1334                 // Use the same traversal protection as Title::secureAndSplit()
1335                 if ( strpos( $path, '.' ) !== false ) {
1336                         if (
1337                                 $path === '.' ||
1338                                 $path === '..' ||
1339                                 strpos( $path, './' ) === 0 ||
1340                                 strpos( $path, '../' ) === 0 ||
1341                                 strpos( $path, '/./' ) !== false ||
1342                                 strpos( $path, '/../' ) !== false
1343                         ) {
1344                                 return null;
1345                         }
1346                 }
1347                 return $path;
1348         }
1349 }