includes/filerepo/backend/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.
  43  * Therefore, the relative paths are only virtual.
  44  *
  45  * Backend contents are stored under wiki-specific container names by default.
  46  * For legacy reasons, this has no effect for the FS backend class, and per-wiki
  47  * segregation must be done by setting the container paths appropriately.
  48  *
  49  * FS-based backends are somewhat more restrictive due to the existence of real
  50  * directory files; a regular file cannot have the same name as a directory. Other
  51  * backends with virtual directories may not have this limitation. Callers should
  52  * store files in such a way that no files and directories are under the same path.
  53  *
  54  * Methods should avoid throwing exceptions at all costs.
  55  * As a corollary, external dependencies should be kept to a minimum.
  56  *
  57  * @ingroup FileBackend
  58  * @since 1.19
  59  */
  60 abstract class FileBackend {
  61         protected $name; // string; unique backend name
  62         protected $wikiId; // string; unique wiki name
  63         protected $readOnly; // string; read-only explanation message
  64         protected $parallelize; // string; when to do operations in parallel
  65         protected $concurrency; // integer; how many operations can be done in parallel
  66
  67         /** @var LockManager */
  68         protected $lockManager;
  69         /** @var FileJournal */
  70         protected $fileJournal;
  71
  72         /**
  73          * Create a new backend instance from configuration.
  74          * This should only be called from within FileBackendGroup.
  75          *
  76          * $config includes:
  77          *   - name        : The unique name of this backend.
  78          *                   This should consist of alphanumberic, '-', and '_' characters.
  79          *                   This name should not be changed after use.
  80          *   - wikiId      : Prefix to container names that is unique to this wiki.
  81          *                   It should only consist of alphanumberic, '-', and '_' characters.
  82          *   - lockManager : Registered name of a file lock manager to use.
  83          *   - fileJournal : File journal configuration; see FileJournal::factory().
  84          *                   Journals simply log changes to files stored in the backend.
  85          *   - readOnly    : Write operations are disallowed if this is a non-empty string.
  86          *                   It should be an explanation for the backend being read-only.
  87          *   - parallelize : When to do file operations in parallel (when possible).
  88          *                   Allowed values are "implicit", "explicit" and "off".
  89          *   - concurrency : How many file operations can be done in parallel.
  90          *
  91          * @param $config Array
  92          * @throws MWException
  93          */
  94         public function __construct( array $config ) {
  95                 $this->name = $config['name'];
  96                 if ( !preg_match( '!^[a-zA-Z0-9-_]{1,255}$!', $this->name ) ) {
  97                         throw new MWException( "Backend name `{$this->name}` is invalid." );
  98                 }
  99                 $this->wikiId = isset( $config['wikiId'] )
 100                         ? $config['wikiId']
 101                         : wfWikiID(); // e.g. "my_wiki-en_"
 102                 $this->lockManager = ( $config['lockManager'] instanceof LockManager )
 103                         ? $config['lockManager']
 104                         : LockManagerGroup::singleton()->get( $config['lockManager'] );
 105                 $this->fileJournal = isset( $config['fileJournal'] )
 106                         ? FileJournal::factory( $config['fileJournal'], $this->name )
 107                         : FileJournal::factory( array( 'class' => 'NullFileJournal' ), $this->name );
 108                 $this->readOnly = isset( $config['readOnly'] )
 109                         ? (string)$config['readOnly']
 110                         : '';
 111                 $this->parallelize = isset( $config['parallelize'] )
 112                         ? (string)$config['parallelize']
 113                         : 'off';
 114                 $this->concurrency = isset( $config['concurrency'] )
 115                         ? (int)$config['concurrency']
 116                         : 50;
 117         }
 118
 119         /**
 120          * Get the unique backend name.
 121          * We may have multiple different backends of the same type.
 122          * For example, we can have two Swift backends using different proxies.
 123          *
 124          * @return string
 125          */
 126         final public function getName() {
 127                 return $this->name;
 128         }
 129
 130         /**
 131          * Check if this backend is read-only
 132          *
 133          * @return bool
 134          */
 135         final public function isReadOnly() {
 136                 return ( $this->readOnly != '' );
 137         }
 138
 139         /**
 140          * Get an explanatory message if this backend is read-only
 141          *
 142          * @return string|bool Returns false if the backend is not read-only
 143          */
 144         final public function getReadOnlyReason() {
 145                 return ( $this->readOnly != '' ) ? $this->readOnly : false;
 146         }
 147
 148         /**
 149          * This is the main entry point into the backend for write operations.
 150          * Callers supply an ordered list of operations to perform as a transaction.
 151          * Files will be locked, the stat cache cleared, and then the operations attempted.
 152          * If any serious errors occur, all attempted operations will be rolled back.
 153          *
 154          * $ops is an array of arrays. The outer array holds a list of operations.
 155          * Each inner array is a set of key value pairs that specify an operation.
 156          *
 157          * Supported operations and their parameters. The supported actions are:
 158          *  - create
 159          *  - store
 160          *  - copy
 161          *  - move
 162          *  - delete
 163          *  - null
 164          *
 165          * a) Create a new file in storage with the contents of a string
 166          * @code
 167          *     array(
 168          *         'op'                  => 'create',
 169          *         'dst'                 => <storage path>,
 170          *         'content'             => <string of new file contents>,
 171          *         'overwrite'           => <boolean>,
 172          *         'overwriteSame'       => <boolean>
 173          *     );
 174          * @endcode
 175          *
 176          * b) Copy a file system file into storage
 177          * @code
 178          *     array(
 179          *         'op'                  => 'store',
 180          *         'src'                 => <file system path>,
 181          *         'dst'                 => <storage path>,
 182          *         'overwrite'           => <boolean>,
 183          *         'overwriteSame'       => <boolean>
 184          *     )
 185          * @endcode
 186          *
 187          * c) Copy a file within storage
 188          * @code
 189          *     array(
 190          *         'op'                  => 'copy',
 191          *         'src'                 => <storage path>,
 192          *         'dst'                 => <storage path>,
 193          *         'overwrite'           => <boolean>,
 194          *         'overwriteSame'       => <boolean>
 195          *     )
 196          * @endcode
 197          *
 198          * d) Move a file within storage
 199          * @code
 200          *     array(
 201          *         'op'                  => 'move',
 202          *         'src'                 => <storage path>,
 203          *         'dst'                 => <storage path>,
 204          *         'overwrite'           => <boolean>,
 205          *         'overwriteSame'       => <boolean>
 206          *     )
 207          * @endcode
 208          *
 209          * e) Delete a file within storage
 210          * @code
 211          *     array(
 212          *         'op'                  => 'delete',
 213          *         'src'                 => <storage path>,
 214          *         'ignoreMissingSource' => <boolean>
 215          *     )
 216          * @endcode
 217          *
 218          * f) Do nothing (no-op)
 219          * @code
 220          *     array(
 221          *         'op'                  => 'null',
 222          *     )
 223          * @endcode
 224          *
 225          * Boolean flags for operations (operation-specific):
 226          *   - ignoreMissingSource : The operation will simply succeed and do
 227          *                           nothing if the source file does not exist.
 228          *   - overwrite           : Any destination file will be overwritten.
 229          *   - overwriteSame       : An error will not be given if a file already
 230          *                           exists at the destination that has the same
 231          *                           contents as the new contents to be written there.
 232          *
 233          * $opts is an associative of boolean flags, including:
 234          *   - force               : Operation precondition errors no longer trigger an abort.
 235          *                           Any remaining operations are still attempted. Unexpected
 236          *                           failures may still cause remaning operations to be aborted.
 237          *   - nonLocking          : No locks are acquired for the operations.
 238          *                           This can increase performance for non-critical writes.
 239          *                           This has no effect unless the 'force' flag is set.
 240          *   - allowStale          : Don't require the latest available data.
 241          *                           This can increase performance for non-critical writes.
 242          *                           This has no effect unless the 'force' flag is set.
 243          *   - nonJournaled        : Don't log this operation batch in the file journal.
 244          *                           This limits the ability of recovery scripts.
 245          *   - parallelize'        : Try to do operations in parallel when possible.
 246          *
 247          * @remarks Remarks on locking:
 248          * File system paths given to operations should refer to files that are
 249          * already locked or otherwise safe from modification from other processes.
 250          * Normally these files will be new temp files, which should be adequate.
 251          *
 252          * @par Return value:
 253          *
 254          * This returns a Status, which contains all warnings and fatals that occured
 255          * during the operation. The 'failCount', 'successCount', and 'success' members
 256          * will reflect each operation attempted.
 257          *
 258          * The status will be "OK" unless:
 259          *     a) unexpected operation errors occurred (network partitions, disk full...)
 260          *     b) significant operation errors occured and 'force' was not set
 261          *
 262          * @param $ops Array List of operations to execute in order
 263          * @param $opts Array Batch operation options
 264          * @return Status
 265          */
 266         final public function doOperations( array $ops, array $opts = array() ) {
 267                 if ( $this->isReadOnly() ) {
 268                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 269                 }
 270                 if ( empty( $opts['force'] ) ) { // sanity
 271                         unset( $opts['nonLocking'] );
 272                         unset( $opts['allowStale'] );
 273                 }
 274                 $opts['concurrency'] = 1; // off
 275                 if ( $this->parallelize === 'implicit' ) {
 276                         if ( !isset( $opts['parallelize'] ) || $opts['parallelize'] ) {
 277                                 $opts['concurrency'] = $this->concurrency;
 278                         }
 279                 } elseif ( $this->parallelize === 'explicit' ) {
 280                         if ( !empty( $opts['parallelize'] ) ) {
 281                                 $opts['concurrency'] = $this->concurrency;
 282                         }
 283                 }
 284                 return $this->doOperationsInternal( $ops, $opts );
 285         }
 286
 287         /**
 288          * @see FileBackend::doOperations()
 289          */
 290         abstract protected function doOperationsInternal( array $ops, array $opts );
 291
 292         /**
 293          * Same as doOperations() except it takes a single operation.
 294          * If you are doing a batch of operations that should either
 295          * all succeed or all fail, then use that function instead.
 296          *
 297          * @see FileBackend::doOperations()
 298          *
 299          * @param $op Array Operation
 300          * @param $opts Array Operation options
 301          * @return Status
 302          */
 303         final public function doOperation( array $op, array $opts = array() ) {
 304                 return $this->doOperations( array( $op ), $opts );
 305         }
 306
 307         /**
 308          * Performs a single create operation.
 309          * This sets $params['op'] to 'create' and passes it to doOperation().
 310          *
 311          * @see FileBackend::doOperation()
 312          *
 313          * @param $params Array Operation parameters
 314          * @param $opts Array Operation options
 315          * @return Status
 316          */
 317         final public function create( array $params, array $opts = array() ) {
 318                 return $this->doOperation( array( 'op' => 'create' ) + $params, $opts );
 319         }
 320
 321         /**
 322          * Performs a single store operation.
 323          * This sets $params['op'] to 'store' and passes it to doOperation().
 324          *
 325          * @see FileBackend::doOperation()
 326          *
 327          * @param $params Array Operation parameters
 328          * @param $opts Array Operation options
 329          * @return Status
 330          */
 331         final public function store( array $params, array $opts = array() ) {
 332                 return $this->doOperation( array( 'op' => 'store' ) + $params, $opts );
 333         }
 334
 335         /**
 336          * Performs a single copy operation.
 337          * This sets $params['op'] to 'copy' and passes it to doOperation().
 338          *
 339          * @see FileBackend::doOperation()
 340          *
 341          * @param $params Array Operation parameters
 342          * @param $opts Array Operation options
 343          * @return Status
 344          */
 345         final public function copy( array $params, array $opts = array() ) {
 346                 return $this->doOperation( array( 'op' => 'copy' ) + $params, $opts );
 347         }
 348
 349         /**
 350          * Performs a single move operation.
 351          * This sets $params['op'] to 'move' and passes it to doOperation().
 352          *
 353          * @see FileBackend::doOperation()
 354          *
 355          * @param $params Array Operation parameters
 356          * @param $opts Array Operation options
 357          * @return Status
 358          */
 359         final public function move( array $params, array $opts = array() ) {
 360                 return $this->doOperation( array( 'op' => 'move' ) + $params, $opts );
 361         }
 362
 363         /**
 364          * Performs a single delete operation.
 365          * This sets $params['op'] to 'delete' and passes it to doOperation().
 366          *
 367          * @see FileBackend::doOperation()
 368          *
 369          * @param $params Array Operation parameters
 370          * @param $opts Array Operation options
 371          * @return Status
 372          */
 373         final public function delete( array $params, array $opts = array() ) {
 374                 return $this->doOperation( array( 'op' => 'delete' ) + $params, $opts );
 375         }
 376
 377         /**
 378          * Perform a set of independent file operations on some files.
 379          *
 380          * This does no locking, nor journaling, and possibly no stat calls.
 381          * Any destination files that already exist will be overwritten.
 382          * This should *only* be used on non-original files, like cache files.
 383          *
 384          * Supported operations and their parameters:
 385          *  - create
 386          *  - store
 387          *  - copy
 388          *  - move
 389          *  - delete
 390          *  - null
 391          * a) Create a new file in storage with the contents of a string
 392          * @code
 393          *     array(
 394          *         'op'                  => 'create',
 395          *         'dst'                 => <storage path>,
 396          *         'content'             => <string of new file contents>
 397          *     )
 398          * @endcode
 399          * b) Copy a file system file into storage
 400          * @code
 401          *     array(
 402          *         'op'                  => 'store',
 403          *         'src'                 => <file system path>,
 404          *         'dst'                 => <storage path>
 405          *     )
 406          * @endcode
 407          * c) Copy a file within storage
 408          * @code
 409          *     array(
 410          *         'op'                  => 'copy',
 411          *         'src'                 => <storage path>,
 412          *         'dst'                 => <storage path>
 413          *     )
 414          * @endcode
 415          * d) Move a file within storage
 416          * @code
 417          *     array(
 418          *         'op'                  => 'move',
 419          *         'src'                 => <storage path>,
 420          *         'dst'                 => <storage path>
 421          *     )
 422          * @endcode
 423          * e) Delete a file within storage
 424          * @code
 425          *     array(
 426          *         'op'                  => 'delete',
 427          *         'src'                 => <storage path>,
 428          *         'ignoreMissingSource' => <boolean>
 429          *     )
 430          * @endcode
 431          * f) Do nothing (no-op)
 432          * @code
 433          *     array(
 434          *         'op'                  => 'null',
 435          *     )
 436          * @endcode
 437          *
 438          * @par Boolean flags for operations (operation-specific):
 439          *   - ignoreMissingSource : The operation will simply succeed and do
 440          *                           nothing if the source file does not exist.
 441          *
 442          * @par Return value:
 443          * This returns a Status, which contains all warnings and fatals that occured
 444          * during the operation. The 'failCount', 'successCount', and 'success' members
 445          * will reflect each operation attempted for the given files. The status will be
 446          * considered "OK" as long as no fatal errors occured.
 447          *
 448          * @param $ops Array Set of operations to execute
 449          * @return Status
 450          * @since 1.20
 451          */
 452         final public function doQuickOperations( array $ops ) {
 453                 if ( $this->isReadOnly() ) {
 454                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 455                 }
 456                 foreach ( $ops as &$op ) {
 457                         $op['overwrite'] = true; // avoids RTTs in key/value stores
 458                 }
 459                 return $this->doQuickOperationsInternal( $ops );
 460         }
 461
 462         /**
 463          * @see FileBackend::doQuickOperations()
 464          * @since 1.20
 465          */
 466         abstract protected function doQuickOperationsInternal( array $ops );
 467
 468         /**
 469          * Same as doQuickOperations() except it takes a single operation.
 470          * If you are doing a batch of operations, then use that function instead.
 471          *
 472          * @see FileBackend::doQuickOperations()
 473          *
 474          * @param $op Array Operation
 475          * @return Status
 476          * @since 1.20
 477          */
 478         final public function doQuickOperation( array $op ) {
 479                 return $this->doQuickOperations( array( $op ) );
 480         }
 481
 482         /**
 483          * Performs a single quick create operation.
 484          * This sets $params['op'] to 'create' and passes it to doQuickOperation().
 485          *
 486          * @see FileBackend::doQuickOperation()
 487          *
 488          * @param $params Array Operation parameters
 489          * @return Status
 490          * @since 1.20
 491          */
 492         final public function quickCreate( array $params ) {
 493                 return $this->doQuickOperation( array( 'op' => 'create' ) + $params );
 494         }
 495
 496         /**
 497          * Performs a single quick store operation.
 498          * This sets $params['op'] to 'store' and passes it to doQuickOperation().
 499          *
 500          * @see FileBackend::doQuickOperation()
 501          *
 502          * @param $params Array Operation parameters
 503          * @return Status
 504          * @since 1.20
 505          */
 506         final public function quickStore( array $params ) {
 507                 return $this->doQuickOperation( array( 'op' => 'store' ) + $params );
 508         }
 509
 510         /**
 511          * Performs a single quick copy operation.
 512          * This sets $params['op'] to 'copy' and passes it to doQuickOperation().
 513          *
 514          * @see FileBackend::doQuickOperation()
 515          *
 516          * @param $params Array Operation parameters
 517          * @return Status
 518          * @since 1.20
 519          */
 520         final public function quickCopy( array $params ) {
 521                 return $this->doQuickOperation( array( 'op' => 'copy' ) + $params );
 522         }
 523
 524         /**
 525          * Performs a single quick move operation.
 526          * This sets $params['op'] to 'move' and passes it to doQuickOperation().
 527          *
 528          * @see FileBackend::doQuickOperation()
 529          *
 530          * @param $params Array Operation parameters
 531          * @return Status
 532          * @since 1.20
 533          */
 534         final public function quickMove( array $params ) {
 535                 return $this->doQuickOperation( array( 'op' => 'move' ) + $params );
 536         }
 537
 538         /**
 539          * Performs a single quick delete operation.
 540          * This sets $params['op'] to 'delete' and passes it to doQuickOperation().
 541          *
 542          * @see FileBackend::doQuickOperation()
 543          *
 544          * @param $params Array Operation parameters
 545          * @return Status
 546          * @since 1.20
 547          */
 548         final public function quickDelete( array $params ) {
 549                 return $this->doQuickOperation( array( 'op' => 'delete' ) + $params );
 550         }
 551
 552         /**
 553          * Concatenate a list of storage files into a single file system file.
 554          * The target path should refer to a file that is already locked or
 555          * otherwise safe from modification from other processes. Normally,
 556          * the file will be a new temp file, which should be adequate.
 557          *
 558          * @param $params Array Operation parameters
 559          * $params include:
 560          *   - srcs : ordered source storage paths (e.g. chunk1, chunk2, ...)
 561          *   - dst  : file system path to 0-byte temp file
 562          * @return Status
 563          */
 564         abstract public function concatenate( array $params );
 565
 566         /**
 567          * Prepare a storage directory for usage.
 568          * This will create any required containers and parent directories.
 569          * Backends using key/value stores only need to create the container.
 570          *
 571          * The 'noAccess' and 'noListing' parameters works the same as in secure(),
 572          * except they are only applied *if* the directory/container had to be created.
 573          * These flags should always be set for directories that have private files.
 574          *
 575          * $params include:
 576          *   - dir       : storage directory
 577          *   - noAccess  : try to deny file access (@since 1.20)
 578          *   - noListing : try to deny file listing (@since 1.20)
 579          *
 580          * @param $params Array
 581          * @return Status
 582          */
 583         final public function prepare( array $params ) {
 584                 if ( $this->isReadOnly() ) {
 585                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 586                 }
 587                 return $this->doPrepare( $params );
 588         }
 589
 590         /**
 591          * @see FileBackend::prepare()
 592          */
 593         abstract protected function doPrepare( array $params );
 594
 595         /**
 596          * Take measures to block web access to a storage directory and
 597          * the container it belongs to. FS backends might add .htaccess
 598          * files whereas key/value store backends might revoke container
 599          * access to the storage user representing end-users in web requests.
 600          * This is not guaranteed to actually do anything.
 601          *
 602          * @param $params Array
 603          * $params include:
 604          *   - dir       : storage directory
 605          *   - noAccess  : try to deny file access
 606          *   - noListing : try to deny file listing
 607          * @return Status
 608          */
 609         final public function secure( array $params ) {
 610                 if ( $this->isReadOnly() ) {
 611                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 612                 }
 613                 return $this->doSecure( $params );
 614         }
 615
 616         /**
 617          * @see FileBackend::secure()
 618          */
 619         abstract protected function doSecure( array $params );
 620
 621         /**
 622          * Remove measures to block web access to a storage directory and
 623          * the container it belongs to. FS backends might remove .htaccess
 624          * files whereas key/value store backends might grant container
 625          * access to the storage user representing end-users in web requests.
 626          * This essentially can undo the result of secure() calls.
 627          *
 628          * $params include:
 629          *   - dir     : storage directory
 630          *   - access  : try to allow file access
 631          *   - listing : try to allow file listing
 632          *
 633          * @param $params Array
 634          * @return Status
 635          * @since 1.20
 636          */
 637         final public function publish( array $params ) {
 638                 if ( $this->isReadOnly() ) {
 639                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 640                 }
 641                 return $this->doPublish( $params );
 642         }
 643
 644         /**
 645          * @see FileBackend::publish()
 646          */
 647         abstract protected function doPublish( array $params );
 648
 649         /**
 650          * Delete a storage directory if it is empty.
 651          * Backends using key/value stores may do nothing unless the directory
 652          * is that of an empty container, in which case it should be deleted.
 653          *
 654          * @param $params Array
 655          * $params include:
 656          *   - dir       : storage directory
 657          *   - recursive : recursively delete empty subdirectories first (@since 1.20)
 658          * @return Status
 659          */
 660         final public function clean( array $params ) {
 661                 if ( $this->isReadOnly() ) {
 662                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 663                 }
 664                 return $this->doClean( $params );
 665         }
 666
 667         /**
 668          * @see FileBackend::clean()
 669          */
 670         abstract protected function doClean( array $params );
 671
 672         /**
 673          * Check if a file exists at a storage path in the backend.
 674          * This returns false if only a directory exists at the path.
 675          *
 676          * @param $params Array
 677          * $params include:
 678          *   - src    : source storage path
 679          *   - latest : use the latest available data
 680          * @return bool|null Returns null on failure
 681          */
 682         abstract public function fileExists( array $params );
 683
 684         /**
 685          * Get the last-modified timestamp of the file at a storage path.
 686          *
 687          * @param $params Array
 688          * $params include:
 689          *   - src    : source storage path
 690          *   - latest : use the latest available data
 691          * @return string|bool TS_MW timestamp or false on failure
 692          */
 693         abstract public function getFileTimestamp( array $params );
 694
 695         /**
 696          * Get the contents of a file at a storage path in the backend.
 697          * This should be avoided for potentially large files.
 698          *
 699          * @param $params Array
 700          * $params include:
 701          *   - src    : source storage path
 702          *   - latest : use the latest available data
 703          * @return string|bool Returns false on failure
 704          */
 705         abstract public function getFileContents( array $params );
 706
 707         /**
 708          * Get the size (bytes) of a file at a storage path in the backend.
 709          *
 710          * @param $params Array
 711          * $params include:
 712          *   - src    : source storage path
 713          *   - latest : use the latest available data
 714          * @return integer|bool Returns false on failure
 715          */
 716         abstract public function getFileSize( array $params );
 717
 718         /**
 719          * Get quick information about a file at a storage path in the backend.
 720          * If the file does not exist, then this returns false.
 721          * Otherwise, the result is an associative array that includes:
 722          *   - mtime  : the last-modified timestamp (TS_MW)
 723          *   - size   : the file size (bytes)
 724          * Additional values may be included for internal use only.
 725          *
 726          * @param $params Array
 727          * $params include:
 728          *   - src    : source storage path
 729          *   - latest : use the latest available data
 730          * @return Array|bool|null Returns null on failure
 731          */
 732         abstract public function getFileStat( array $params );
 733
 734         /**
 735          * Get a SHA-1 hash of the file at a storage path in the backend.
 736          *
 737          * @param $params Array
 738          * $params include:
 739          *   - src    : source storage path
 740          *   - latest : use the latest available data
 741          * @return string|bool Hash string or false on failure
 742          */
 743         abstract public function getFileSha1Base36( array $params );
 744
 745         /**
 746          * Get the properties of the file at a storage path in the backend.
 747          * Returns FSFile::placeholderProps() on failure.
 748          *
 749          * @param $params Array
 750          * $params include:
 751          *   - src    : source storage path
 752          *   - latest : use the latest available data
 753          * @return Array
 754          */
 755         abstract public function getFileProps( array $params );
 756
 757         /**
 758          * Stream the file at a storage path in the backend.
 759          * If the file does not exists, a 404 error will be given.
 760          * Appropriate HTTP headers (Status, Content-Type, Content-Length)
 761          * must be sent if streaming began, while none should be sent otherwise.
 762          * Implementations should flush the output buffer before sending data.
 763          *
 764          * @param $params Array
 765          * $params include:
 766          *   - src     : source storage path
 767          *   - headers : additional HTTP headers to send on success
 768          *   - latest  : use the latest available data
 769          * @return Status
 770          */
 771         abstract public function streamFile( array $params );
 772
 773         /**
 774          * Returns a file system file, identical to the file at a storage path.
 775          * The file returned is either:
 776          * a) A local copy of the file at a storage path in the backend.
 777          *    The temporary copy will have the same extension as the source.
 778          * b) An original of the file at a storage path in the backend.
 779          * Temporary files may be purged when the file object falls out of scope.
 780          *
 781          * Write operations should *never* be done on this file as some backends
 782          * may do internal tracking or may be instances of FileBackendMultiWrite.
 783          * In that later case, there are copies of the file that must stay in sync.
 784          * Additionally, further calls to this function may return the same file.
 785          *
 786          * @param $params Array
 787          * $params include:
 788          *   - src    : source storage path
 789          *   - latest : use the latest available data
 790          * @return FSFile|null Returns null on failure
 791          */
 792         abstract public function getLocalReference( array $params );
 793
 794         /**
 795          * Get a local copy on disk of the file at a storage path in the backend.
 796          * The temporary copy will have the same file extension as the source.
 797          * Temporary files may be purged when the file object falls out of scope.
 798          *
 799          * @param $params Array
 800          * $params include:
 801          *   - src    : source storage path
 802          *   - latest : use the latest available data
 803          * @return TempFSFile|null Returns null on failure
 804          */
 805         abstract public function getLocalCopy( array $params );
 806
 807         /**
 808          * Check if a directory exists at a given storage path.
 809          * Backends using key/value stores will check if the path is a
 810          * virtual directory, meaning there are files under the given directory.
 811          *
 812          * Storage backends with eventual consistency might return stale data.
 813          *
 814          * @param $params array
 815          * $params include:
 816          *   - dir : storage directory
 817          * @return bool|null Returns null on failure
 818          * @since 1.20
 819          */
 820         abstract public function directoryExists( array $params );
 821
 822         /**
 823          * Get an iterator to list *all* directories under a storage directory.
 824          * If the directory is of the form "mwstore://backend/container",
 825          * then all directories in the container should be listed.
 826          * If the directory is of form "mwstore://backend/container/dir",
 827          * then all directories directly under that directory should be listed.
 828          * Results should be storage directories relative to the given directory.
 829          *
 830          * Storage backends with eventual consistency might return stale data.
 831          *
 832          * @param $params array
 833          * $params include:
 834          *   - dir     : storage directory
 835          *   - topOnly : only return direct child dirs of the directory
 836          * @return Traversable|Array|null Returns null on failure
 837          * @since 1.20
 838          */
 839         abstract public function getDirectoryList( array $params );
 840
 841         /**
 842          * Same as FileBackend::getDirectoryList() except only lists
 843          * directories that are immediately under the given directory.
 844          *
 845          * Storage backends with eventual consistency might return stale data.
 846          *
 847          * @param $params array
 848          * $params include:
 849          *   - dir : storage directory
 850          * @return Traversable|Array|null Returns null on failure
 851          * @since 1.20
 852          */
 853         final public function getTopDirectoryList( array $params ) {
 854                 return $this->getDirectoryList( array( 'topOnly' => true ) + $params );
 855         }
 856
 857         /**
 858          * Get an iterator to list *all* stored files under a storage directory.
 859          * If the directory is of the form "mwstore://backend/container",
 860          * then all files in the container should be listed.
 861          * If the directory is of form "mwstore://backend/container/dir",
 862          * then all files under that directory should be listed.
 863          * Results should be storage paths relative to the given directory.
 864          *
 865          * Storage backends with eventual consistency might return stale data.
 866          *
 867          * @param $params array
 868          * $params include:
 869          *   - dir     : storage directory
 870          *   - topOnly : only return direct child files of the directory (@since 1.20)
 871          * @return Traversable|Array|null Returns null on failure
 872          */
 873         abstract public function getFileList( array $params );
 874
 875         /**
 876          * Same as FileBackend::getFileList() except only lists
 877          * files that are immediately under the given directory.
 878          *
 879          * Storage backends with eventual consistency might return stale data.
 880          *
 881          * @param $params array
 882          * $params include:
 883          *   - dir : storage directory
 884          * @return Traversable|Array|null Returns null on failure
 885          * @since 1.20
 886          */
 887         final public function getTopFileList( array $params ) {
 888                 return $this->getFileList( array( 'topOnly' => true ) + $params );
 889         }
 890
 891         /**
 892          * Invalidate any in-process file existence and property cache.
 893          * If $paths is given, then only the cache for those files will be cleared.
 894          *
 895          * @param $paths Array Storage paths (optional)
 896          * @return void
 897          */
 898         public function clearCache( array $paths = null ) {}
 899
 900         /**
 901          * Lock the files at the given storage paths in the backend.
 902          * This will either lock all the files or none (on failure).
 903          *
 904          * Callers should consider using getScopedFileLocks() instead.
 905          *
 906          * @param $paths Array Storage paths
 907          * @param $type integer LockManager::LOCK_* constant
 908          * @return Status
 909          */
 910         final public function lockFiles( array $paths, $type ) {
 911                 return $this->lockManager->lock( $paths, $type );
 912         }
 913
 914         /**
 915          * Unlock the files at the given storage paths in the backend.
 916          *
 917          * @param $paths Array Storage paths
 918          * @param $type integer LockManager::LOCK_* constant
 919          * @return Status
 920          */
 921         final public function unlockFiles( array $paths, $type ) {
 922                 return $this->lockManager->unlock( $paths, $type );
 923         }
 924
 925         /**
 926          * Lock the files at the given storage paths in the backend.
 927          * This will either lock all the files or none (on failure).
 928          * On failure, the status object will be updated with errors.
 929          *
 930          * Once the return value goes out scope, the locks will be released and
 931          * the status updated. Unlock fatals will not change the status "OK" value.
 932          *
 933          * @param $paths Array Storage paths
 934          * @param $type integer LockManager::LOCK_* constant
 935          * @param $status Status Status to update on lock/unlock
 936          * @return ScopedLock|null Returns null on failure
 937          */
 938         final public function getScopedFileLocks( array $paths, $type, Status $status ) {
 939                 return ScopedLock::factory( $this->lockManager, $paths, $type, $status );
 940         }
 941
 942         /**
 943          * Get an array of scoped locks needed for a batch of file operations.
 944          *
 945          * Normally, FileBackend::doOperations() handles locking, unless
 946          * the 'nonLocking' param is passed in. This function is useful if you
 947          * want the files to be locked for a broader scope than just when the
 948          * files are changing. For example, if you need to update DB metadata,
 949          * you may want to keep the files locked until finished.
 950          *
 951          * @see FileBackend::doOperations()
 952          *
 953          * @param $ops Array List of file operations to FileBackend::doOperations()
 954          * @param $status Status Status to update on lock/unlock
 955          * @return Array List of ScopedFileLocks or null values
 956          * @since 1.20
 957          */
 958         abstract public function getScopedLocksForOps( array $ops, Status $status );
 959
 960         /**
 961          * Get the root storage path of this backend.
 962          * All container paths are "subdirectories" of this path.
 963          *
 964          * @return string Storage path
 965          * @since 1.20
 966          */
 967         final public function getRootStoragePath() {
 968                 return "mwstore://{$this->name}";
 969         }
 970
 971         /**
 972          * Get the file journal object for this backend
 973          *
 974          * @return FileJournal
 975          */
 976         final public function getJournal() {
 977                 return $this->fileJournal;
 978         }
 979
 980         /**
 981          * Check if a given path is a "mwstore://" path.
 982          * This does not do any further validation or any existence checks.
 983          *
 984          * @param $path string
 985          * @return bool
 986          */
 987         final public static function isStoragePath( $path ) {
 988                 return ( strpos( $path, 'mwstore://' ) === 0 );
 989         }
 990
 991         /**
 992          * Split a storage path into a backend name, a container name,
 993          * and a relative file path. The relative path may be the empty string.
 994          * This does not do any path normalization or traversal checks.
 995          *
 996          * @param $storagePath string
 997          * @return Array (backend, container, rel object) or (null, null, null)
 998          */
 999         final public static function splitStoragePath( $storagePath ) {
1000                 if ( self::isStoragePath( $storagePath ) ) {
1001                         // Remove the "mwstore://" prefix and split the path
1002                         $parts = explode( '/', substr( $storagePath, 10 ), 3 );
1003                         if ( count( $parts ) >= 2 && $parts[0] != '' && $parts[1] != '' ) {
1004                                 if ( count( $parts ) == 3 ) {
1005                                         return $parts; // e.g. "backend/container/path"
1006                                 } else {
1007                                         return array( $parts[0], $parts[1], '' ); // e.g. "backend/container"
1008                                 }
1009                         }
1010                 }
1011                 return array( null, null, null );
1012         }
1013
1014         /**
1015          * Normalize a storage path by cleaning up directory separators.
1016          * Returns null if the path is not of the format of a valid storage path.
1017          *
1018          * @param $storagePath string
1019          * @return string|null
1020          */
1021         final public static function normalizeStoragePath( $storagePath ) {
1022                 list( $backend, $container, $relPath ) = self::splitStoragePath( $storagePath );
1023                 if ( $relPath !== null ) { // must be for this backend
1024                         $relPath = self::normalizeContainerPath( $relPath );
1025                         if ( $relPath !== null ) {
1026                                 return ( $relPath != '' )
1027                                         ? "mwstore://{$backend}/{$container}/{$relPath}"
1028                                         : "mwstore://{$backend}/{$container}";
1029                         }
1030                 }
1031                 return null;
1032         }
1033
1034         /**
1035          * Get the parent storage directory of a storage path.
1036          * This returns a path like "mwstore://backend/container",
1037          * "mwstore://backend/container/...", or null if there is no parent.
1038          *
1039          * @param $storagePath string
1040          * @return string|null
1041          */
1042         final public static function parentStoragePath( $storagePath ) {
1043                 $storagePath = dirname( $storagePath );
1044                 list( $b, $cont, $rel ) = self::splitStoragePath( $storagePath );
1045                 return ( $rel === null ) ? null : $storagePath;
1046         }
1047
1048         /**
1049          * Get the final extension from a storage or FS path
1050          *
1051          * @param $path string
1052          * @return string
1053          */
1054         final public static function extensionFromPath( $path ) {
1055                 $i = strrpos( $path, '.' );
1056                 return strtolower( $i ? substr( $path, $i + 1 ) : '' );
1057         }
1058
1059         /**
1060          * Check if a relative path has no directory traversals
1061          *
1062          * @param $path string
1063          * @return bool
1064          * @since 1.20
1065          */
1066         final public static function isPathTraversalFree( $path ) {
1067                 return ( self::normalizeContainerPath( $path ) !== null );
1068         }
1069
1070         /**
1071          * Validate and normalize a relative storage path.
1072          * Null is returned if the path involves directory traversal.
1073          * Traversal is insecure for FS backends and broken for others.
1074          *
1075          * This uses the same traversal protection as Title::secureAndSplit().
1076          *
1077          * @param $path string Storage path relative to a container
1078          * @return string|null
1079          */
1080         final protected static function normalizeContainerPath( $path ) {
1081                 // Normalize directory separators
1082                 $path = strtr( $path, '\\', '/' );
1083                 // Collapse any consecutive directory separators
1084                 $path = preg_replace( '![/]{2,}!', '/', $path );
1085                 // Remove any leading directory separator
1086                 $path = ltrim( $path, '/' );
1087                 // Use the same traversal protection as Title::secureAndSplit()
1088                 if ( strpos( $path, '.' ) !== false ) {
1089                         if (
1090                                 $path === '.' ||
1091                                 $path === '..' ||
1092                                 strpos( $path, './' ) === 0 ||
1093                                 strpos( $path, '../' ) === 0 ||
1094                                 strpos( $path, '/./' ) !== false ||
1095                                 strpos( $path, '/../' ) !== false
1096                         ) {
1097                                 return null;
1098                         }
1099                 }
1100                 return $path;
1101         }
1102 }