includes/libs/filebackend/FileBackend.php

   1 <?php
   2 /**
   3  * @defgroup FileBackend File backend
   4  *
   5  * File backend is used to interact with file storage systems,
   6  * such as the local file system, NFS, or cloud storage systems.
   7  */
   8
   9 /**
  10  * Base class for all file backends.
  11  *
  12  * This program is free software; you can redistribute it and/or modify
  13  * it under the terms of the GNU General Public License as published by
  14  * the Free Software Foundation; either version 2 of the License, or
  15  * (at your option) any later version.
  16  *
  17  * This program is distributed in the hope that it will be useful,
  18  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  19  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  20  * GNU General Public License for more details.
  21  *
  22  * You should have received a copy of the GNU General Public License along
  23  * with this program; if not, write to the Free Software Foundation, Inc.,
  24  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  25  * http://www.gnu.org/copyleft/gpl.html
  26  *
  27  * @file
  28  * @ingroup FileBackend
  29  */
  30 use MediaWiki\FileBackend\FSFile\TempFSFileFactory;
  31 use Psr\Log\LoggerAwareInterface;
  32 use Psr\Log\LoggerInterface;
  33 use Wikimedia\ScopedCallback;
  34 use Psr\Log\NullLogger;
  35
  36 /**
  37  * @brief Base class for all file backend classes (including multi-write backends).
  38  *
  39  * This class defines the methods as abstract that subclasses must implement.
  40  * Outside callers can assume that all backends will have these functions.
  41  *
  42  * All "storage paths" are of the format "mwstore://<backend>/<container>/<path>".
  43  * The "backend" portion is unique name for the application to refer to a backend, while
  44  * the "container" portion is a top-level directory of the backend. The "path" portion
  45  * is a relative path that uses UNIX file system (FS) notation, though any particular
  46  * backend may not actually be using a local filesystem. Therefore, the relative paths
  47  * are only virtual.
  48  *
  49  * Backend contents are stored under "domain"-specific container names by default.
  50  * A domain is simply a logical umbrella for entities, such as those belonging to a certain
  51  * application or portion of a website, for example. A domain can be local or global.
  52  * Global (qualified) backends are achieved by configuring the "domain ID" to a constant.
  53  * Global domains are simpler, but local domains can be used by choosing a domain ID based on
  54  * the current context, such as which language of a website is being used.
  55  *
  56  * For legacy reasons, the FSFileBackend class allows manually setting the paths of
  57  * containers to ones that do not respect the "domain ID".
  58  *
  59  * In key/value (object) stores, containers are the only hierarchy (the rest is emulated).
  60  * FS-based backends are somewhat more restrictive due to the existence of real
  61  * directory files; a regular file cannot have the same name as a directory. Other
  62  * backends with virtual directories may not have this limitation. Callers should
  63  * store files in such a way that no files and directories are under the same path.
  64  *
  65  * In general, this class allows for callers to access storage through the same
  66  * interface, without regard to the underlying storage system. However, calling code
  67  * must follow certain patterns and be aware of certain things to ensure compatibility:
  68  *   - a) Always call prepare() on the parent directory before trying to put a file there;
  69  *        key/value stores only need the container to exist first, but filesystems need
  70  *        all the parent directories to exist first (prepare() is aware of all this)
  71  *   - b) Always call clean() on a directory when it might become empty to avoid empty
  72  *        directory buildup on filesystems; key/value stores never have empty directories,
  73  *        so doing this helps preserve consistency in both cases
  74  *   - c) Likewise, do not rely on the existence of empty directories for anything;
  75  *        calling directoryExists() on a path that prepare() was previously called on
  76  *        will return false for key/value stores if there are no files under that path
  77  *   - d) Never alter the resulting FSFile returned from getLocalReference(), as it could
  78  *        either be a copy of the source file in /tmp or the original source file itself
  79  *   - e) Use a file layout that results in never attempting to store files over directories
  80  *        or directories over files; key/value stores allow this but filesystems do not
  81  *   - f) Use ASCII file names (e.g. base32, IDs, hashes) to avoid Unicode issues in Windows
  82  *   - g) Do not assume that move operations are atomic (difficult with key/value stores)
  83  *   - h) Do not assume that file stat or read operations always have immediate consistency;
  84  *        various methods have a "latest" flag that should always be used if up-to-date
  85  *        information is required (this trades performance for correctness as needed)
  86  *   - i) Do not assume that directory listings have immediate consistency
  87  *
  88  * Methods of subclasses should avoid throwing exceptions at all costs.
  89  * As a corollary, external dependencies should be kept to a minimum.
  90  *
  91  * @ingroup FileBackend
  92  * @since 1.19
  93  */
  94 abstract class FileBackend implements LoggerAwareInterface {
  95         /** @var string Unique backend name */
  96         protected $name;
  97
  98         /** @var string Unique domain name */
  99         protected $domainId;
 100
 101         /** @var string Read-only explanation message */
 102         protected $readOnly;
 103
 104         /** @var string When to do operations in parallel */
 105         protected $parallelize;
 106
 107         /** @var int How many operations can be done in parallel */
 108         protected $concurrency;
 109
 110         /** @var TempFSFileFactory */
 111         protected $tmpFileFactory;
 112
 113         /** @var LockManager */
 114         protected $lockManager;
 115         /** @var FileJournal */
 116         protected $fileJournal;
 117         /** @var LoggerInterface */
 118         protected $logger;
 119         /** @var callable|null */
 120         protected $profiler;
 121
 122         /** @var callable */
 123         protected $obResetFunc;
 124         /** @var callable */
 125         protected $streamMimeFunc;
 126         /** @var callable */
 127         protected $statusWrapper;
 128
 129         /** Bitfield flags for supported features */
 130         const ATTR_HEADERS = 1; // files can be tagged with standard HTTP headers
 131         const ATTR_METADATA = 2; // files can be stored with metadata key/values
 132         const ATTR_UNICODE_PATHS = 4; // files can have Unicode paths (not just ASCII)
 133
 134         /** @var null Idiom for "could not determine due to I/O errors" */
 135         const UNKNOWN = null;
 136
 137         /**
 138          * Create a new backend instance from configuration.
 139          * This should only be called from within FileBackendGroup.
 140          *
 141          * @param array $config Parameters include:
 142          *   - name : The unique name of this backend.
 143          *      This should consist of alphanumberic, '-', and '_' characters.
 144          *      This name should not be changed after use (e.g. with journaling).
 145          *      Note that the name is *not* used in actual container names.
 146          *   - domainId : Prefix to container names that is unique to this backend.
 147          *      It should only consist of alphanumberic, '-', and '_' characters.
 148          *      This ID is what avoids collisions if multiple logical backends
 149          *      use the same storage system, so this should be set carefully.
 150          *   - lockManager : LockManager object to use for any file locking.
 151          *      If not provided, then no file locking will be enforced.
 152          *   - fileJournal : FileJournal object to use for logging changes to files.
 153          *      If not provided, then change journaling will be disabled.
 154          *   - readOnly : Write operations are disallowed if this is a non-empty string.
 155          *      It should be an explanation for the backend being read-only.
 156          *   - parallelize : When to do file operations in parallel (when possible).
 157          *      Allowed values are "implicit", "explicit" and "off".
 158          *   - concurrency : How many file operations can be done in parallel.
 159          *   - tmpDirectory : Directory to use for temporary files.
 160          *   - tmpFileFactory : Optional TempFSFileFactory object. Only has an effect if tmpDirectory is
 161          *      not set. If both are unset or null, then the backend will try to discover a usable
 162          *      temporary directory.
 163          *   - obResetFunc : alternative callback to clear the output buffer
 164          *   - streamMimeFunc : alternative method to determine the content type from the path
 165          *   - logger : Optional PSR logger object.
 166          *   - profiler : Optional callback that takes a section name argument and returns
 167          *      a ScopedCallback instance that ends the profile section in its destructor.
 168          * @throws InvalidArgumentException
 169          */
 170         public function __construct( array $config ) {
 171                 $this->name = $config['name'];
 172                 $this->domainId = $config['domainId'] // e.g. "my_wiki-en_"
 173                         ?? $config['wikiId']; // b/c alias
 174                 if ( !preg_match( '!^[a-zA-Z0-9-_]{1,255}$!', $this->name ) ) {
 175                         throw new InvalidArgumentException( "Backend name '{$this->name}' is invalid." );
 176                 } elseif ( !is_string( $this->domainId ) ) {
 177                         throw new InvalidArgumentException(
 178                                 "Backend domain ID not provided for '{$this->name}'." );
 179                 }
 180                 $this->lockManager = $config['lockManager'] ?? new NullLockManager( [] );
 181                 $this->fileJournal = $config['fileJournal']
 182                         ?? FileJournal::factory( [ 'class' => NullFileJournal::class ], $this->name );
 183                 $this->readOnly = isset( $config['readOnly'] )
 184                         ? (string)$config['readOnly']
 185                         : '';
 186                 $this->parallelize = isset( $config['parallelize'] )
 187                         ? (string)$config['parallelize']
 188                         : 'off';
 189                 $this->concurrency = isset( $config['concurrency'] )
 190                         ? (int)$config['concurrency']
 191                         : 50;
 192                 $this->obResetFunc = $config['obResetFunc'] ?? [ $this, 'resetOutputBuffer' ];
 193                 $this->streamMimeFunc = $config['streamMimeFunc'] ?? null;
 194                 $this->statusWrapper = $config['statusWrapper'] ?? null;
 195
 196                 $this->profiler = $config['profiler'] ?? null;
 197                 if ( !is_callable( $this->profiler ) ) {
 198                         $this->profiler = null;
 199                 }
 200                 $this->logger = $config['logger'] ?? new NullLogger();
 201                 $this->statusWrapper = $config['statusWrapper'] ?? null;
 202                 // tmpDirectory gets precedence for backward compatibility
 203                 if ( isset( $config['tmpDirectory'] ) ) {
 204                         $this->tmpFileFactory = new TempFSFileFactory( $config['tmpDirectory'] );
 205                 } else {
 206                         $this->tmpFileFactory = $config['tmpFileFactory'] ?? new TempFSFileFactory();
 207                 }
 208         }
 209
 210         public function setLogger( LoggerInterface $logger ) {
 211                 $this->logger = $logger;
 212         }
 213
 214         /**
 215          * Get the unique backend name
 216          *
 217          * We may have multiple different backends of the same type.
 218          * For example, we can have two Swift backends using different proxies.
 219          *
 220          * @return string
 221          */
 222         final public function getName() {
 223                 return $this->name;
 224         }
 225
 226         /**
 227          * Get the domain identifier used for this backend (possibly empty).
 228          *
 229          * @return string
 230          * @since 1.28
 231          */
 232         final public function getDomainId() {
 233                 return $this->domainId;
 234         }
 235
 236         /**
 237          * Alias to getDomainId()
 238          *
 239          * @return string
 240          * @since 1.20
 241          * @deprecated Since 1.34 Use getDomainId()
 242          */
 243         final public function getWikiId() {
 244                 return $this->getDomainId();
 245         }
 246
 247         /**
 248          * Check if this backend is read-only
 249          *
 250          * @return bool
 251          */
 252         final public function isReadOnly() {
 253                 return ( $this->readOnly != '' );
 254         }
 255
 256         /**
 257          * Get an explanatory message if this backend is read-only
 258          *
 259          * @return string|bool Returns false if the backend is not read-only
 260          */
 261         final public function getReadOnlyReason() {
 262                 return ( $this->readOnly != '' ) ? $this->readOnly : false;
 263         }
 264
 265         /**
 266          * Get the a bitfield of extra features supported by the backend medium
 267          *
 268          * @return int Bitfield of FileBackend::ATTR_* flags
 269          * @since 1.23
 270          */
 271         public function getFeatures() {
 272                 return self::ATTR_UNICODE_PATHS;
 273         }
 274
 275         /**
 276          * Check if the backend medium supports a field of extra features
 277          *
 278          * @param int $bitfield Bitfield of FileBackend::ATTR_* flags
 279          * @return bool
 280          * @since 1.23
 281          */
 282         final public function hasFeatures( $bitfield ) {
 283                 return ( $this->getFeatures() & $bitfield ) === $bitfield;
 284         }
 285
 286         /**
 287          * This is the main entry point into the backend for write operations.
 288          * Callers supply an ordered list of operations to perform as a transaction.
 289          * Files will be locked, the stat cache cleared, and then the operations attempted.
 290          * If any serious errors occur, all attempted operations will be rolled back.
 291          *
 292          * $ops is an array of arrays. The outer array holds a list of operations.
 293          * Each inner array is a set of key value pairs that specify an operation.
 294          *
 295          * Supported operations and their parameters. The supported actions are:
 296          *  - create
 297          *  - store
 298          *  - copy
 299          *  - move
 300          *  - delete
 301          *  - describe (since 1.21)
 302          *  - null
 303          *
 304          * FSFile/TempFSFile object support was added in 1.27.
 305          *
 306          * a) Create a new file in storage with the contents of a string
 307          * @code
 308          *     [
 309          *         'op'                  => 'create',
 310          *         'dst'                 => <storage path>,
 311          *         'content'             => <string of new file contents>,
 312          *         'overwrite'           => <boolean>,
 313          *         'overwriteSame'       => <boolean>,
 314          *         'headers'             => <HTTP header name/value map> # since 1.21
 315          *     ]
 316          * @endcode
 317          *
 318          * b) Copy a file system file into storage
 319          * @code
 320          *     [
 321          *         'op'                  => 'store',
 322          *         'src'                 => <file system path, FSFile, or TempFSFile>,
 323          *         'dst'                 => <storage path>,
 324          *         'overwrite'           => <boolean>,
 325          *         'overwriteSame'       => <boolean>,
 326          *         'headers'             => <HTTP header name/value map> # since 1.21
 327          *     ]
 328          * @endcode
 329          *
 330          * c) Copy a file within storage
 331          * @code
 332          *     [
 333          *         'op'                  => 'copy',
 334          *         'src'                 => <storage path>,
 335          *         'dst'                 => <storage path>,
 336          *         'overwrite'           => <boolean>,
 337          *         'overwriteSame'       => <boolean>,
 338          *         'ignoreMissingSource' => <boolean>, # since 1.21
 339          *         'headers'             => <HTTP header name/value map> # since 1.21
 340          *     ]
 341          * @endcode
 342          *
 343          * d) Move a file within storage
 344          * @code
 345          *     [
 346          *         'op'                  => 'move',
 347          *         'src'                 => <storage path>,
 348          *         'dst'                 => <storage path>,
 349          *         'overwrite'           => <boolean>,
 350          *         'overwriteSame'       => <boolean>,
 351          *         'ignoreMissingSource' => <boolean>, # since 1.21
 352          *         'headers'             => <HTTP header name/value map> # since 1.21
 353          *     ]
 354          * @endcode
 355          *
 356          * e) Delete a file within storage
 357          * @code
 358          *     [
 359          *         'op'                  => 'delete',
 360          *         'src'                 => <storage path>,
 361          *         'ignoreMissingSource' => <boolean>
 362          *     ]
 363          * @endcode
 364          *
 365          * f) Update metadata for a file within storage
 366          * @code
 367          *     [
 368          *         'op'                  => 'describe',
 369          *         'src'                 => <storage path>,
 370          *         'headers'             => <HTTP header name/value map>
 371          *     ]
 372          * @endcode
 373          *
 374          * g) Do nothing (no-op)
 375          * @code
 376          *     [
 377          *         'op'                  => 'null',
 378          *     ]
 379          * @endcode
 380          *
 381          * Boolean flags for operations (operation-specific):
 382          *   - ignoreMissingSource : The operation will simply succeed and do
 383          *                           nothing if the source file does not exist.
 384          *   - overwrite           : Any destination file will be overwritten.
 385          *   - overwriteSame       : If a file already exists at the destination with the
 386          *                           same contents, then do nothing to the destination file
 387          *                           instead of giving an error. This does not compare headers.
 388          *                           This option is ignored if 'overwrite' is already provided.
 389          *   - headers             : If supplied, the result of merging these headers with any
 390          *                           existing source file headers (replacing conflicting ones)
 391          *                           will be set as the destination file headers. Headers are
 392          *                           deleted if their value is set to the empty string. When a
 393          *                           file has headers they are included in responses to GET and
 394          *                           HEAD requests to the backing store for that file.
 395          *                           Header values should be no larger than 255 bytes, except for
 396          *                           Content-Disposition. The system might ignore or truncate any
 397          *                           headers that are too long to store (exact limits will vary).
 398          *                           Backends that don't support metadata ignore this. (since 1.21)
 399          *
 400          * $opts is an associative of boolean flags, including:
 401          *   - force               : Operation precondition errors no longer trigger an abort.
 402          *                           Any remaining operations are still attempted. Unexpected
 403          *                           failures may still cause remaining operations to be aborted.
 404          *   - nonLocking          : No locks are acquired for the operations.
 405          *                           This can increase performance for non-critical writes.
 406          *                           This has no effect unless the 'force' flag is set.
 407          *   - nonJournaled        : Don't log this operation batch in the file journal.
 408          *                           This limits the ability of recovery scripts.
 409          *   - parallelize         : Try to do operations in parallel when possible.
 410          *   - bypassReadOnly      : Allow writes in read-only mode. (since 1.20)
 411          *   - preserveCache       : Don't clear the process cache before checking files.
 412          *                           This should only be used if all entries in the process
 413          *                           cache were added after the files were already locked. (since 1.20)
 414          *
 415          * @note Remarks on locking:
 416          * File system paths given to operations should refer to files that are
 417          * already locked or otherwise safe from modification from other processes.
 418          * Normally these files will be new temp files, which should be adequate.
 419          *
 420          * @par Return value:
 421          *
 422          * This returns a Status, which contains all warnings and fatals that occurred
 423          * during the operation. The 'failCount', 'successCount', and 'success' members
 424          * will reflect each operation attempted.
 425          *
 426          * The StatusValue will be "OK" unless:
 427          *   - a) unexpected operation errors occurred (network partitions, disk full...)
 428          *   - b) predicted operation errors occurred and 'force' was not set
 429          *
 430          * @param array $ops List of operations to execute in order
 431          * @codingStandardsIgnoreStart
 432          * @phan-param array{ignoreMissingSource?:bool,overwrite?:bool,overwriteSame?:bool,headers?:bool} $ops
 433          * @param array $opts Batch operation options
 434          * @phan-param array{force?:bool,nonLocking?:bool,nonJournaled?:bool,parallelize?:bool,bypassReadOnly?:bool,preserveCache?:bool} $opts
 435          * @codingStandardsIgnoreEnd
 436          * @return StatusValue
 437          */
 438         final public function doOperations( array $ops, array $opts = [] ) {
 439                 if ( empty( $opts['bypassReadOnly'] ) && $this->isReadOnly() ) {
 440                         return $this->newStatus( 'backend-fail-readonly', $this->name, $this->readOnly );
 441                 }
 442                 if ( $ops === [] ) {
 443                         return $this->newStatus(); // nothing to do
 444                 }
 445
 446                 $ops = $this->resolveFSFileObjects( $ops );
 447                 if ( empty( $opts['force'] ) ) { // sanity
 448                         unset( $opts['nonLocking'] );
 449                 }
 450
 451                 /** @noinspection PhpUnusedLocalVariableInspection */
 452                 $scope = ScopedCallback::newScopedIgnoreUserAbort(); // try to ignore client aborts
 453
 454                 return $this->doOperationsInternal( $ops, $opts );
 455         }
 456
 457         /**
 458          * @see FileBackend::doOperations()
 459          * @param array $ops
 460          * @param array $opts
 461          * @return StatusValue
 462          */
 463         abstract protected function doOperationsInternal( array $ops, array $opts );
 464
 465         /**
 466          * Same as doOperations() except it takes a single operation.
 467          * If you are doing a batch of operations that should either
 468          * all succeed or all fail, then use that function instead.
 469          *
 470          * @see FileBackend::doOperations()
 471          *
 472          * @param array $op Operation
 473          * @param array $opts Operation options
 474          * @return StatusValue
 475          */
 476         final public function doOperation( array $op, array $opts = [] ) {
 477                 return $this->doOperations( [ $op ], $opts );
 478         }
 479
 480         /**
 481          * Performs a single create operation.
 482          * This sets $params['op'] to 'create' and passes it to doOperation().
 483          *
 484          * @see FileBackend::doOperation()
 485          *
 486          * @param array $params Operation parameters
 487          * @param array $opts Operation options
 488          * @return StatusValue
 489          */
 490         final public function create( array $params, array $opts = [] ) {
 491                 return $this->doOperation( [ 'op' => 'create' ] + $params, $opts );
 492         }
 493
 494         /**
 495          * Performs a single store operation.
 496          * This sets $params['op'] to 'store' and passes it to doOperation().
 497          *
 498          * @see FileBackend::doOperation()
 499          *
 500          * @param array $params Operation parameters
 501          * @param array $opts Operation options
 502          * @return StatusValue
 503          */
 504         final public function store( array $params, array $opts = [] ) {
 505                 return $this->doOperation( [ 'op' => 'store' ] + $params, $opts );
 506         }
 507
 508         /**
 509          * Performs a single copy operation.
 510          * This sets $params['op'] to 'copy' and passes it to doOperation().
 511          *
 512          * @see FileBackend::doOperation()
 513          *
 514          * @param array $params Operation parameters
 515          * @param array $opts Operation options
 516          * @return StatusValue
 517          */
 518         final public function copy( array $params, array $opts = [] ) {
 519                 return $this->doOperation( [ 'op' => 'copy' ] + $params, $opts );
 520         }
 521
 522         /**
 523          * Performs a single move operation.
 524          * This sets $params['op'] to 'move' and passes it to doOperation().
 525          *
 526          * @see FileBackend::doOperation()
 527          *
 528          * @param array $params Operation parameters
 529          * @param array $opts Operation options
 530          * @return StatusValue
 531          */
 532         final public function move( array $params, array $opts = [] ) {
 533                 return $this->doOperation( [ 'op' => 'move' ] + $params, $opts );
 534         }
 535
 536         /**
 537          * Performs a single delete operation.
 538          * This sets $params['op'] to 'delete' and passes it to doOperation().
 539          *
 540          * @see FileBackend::doOperation()
 541          *
 542          * @param array $params Operation parameters
 543          * @param array $opts Operation options
 544          * @return StatusValue
 545          */
 546         final public function delete( array $params, array $opts = [] ) {
 547                 return $this->doOperation( [ 'op' => 'delete' ] + $params, $opts );
 548         }
 549
 550         /**
 551          * Performs a single describe operation.
 552          * This sets $params['op'] to 'describe' and passes it to doOperation().
 553          *
 554          * @see FileBackend::doOperation()
 555          *
 556          * @param array $params Operation parameters
 557          * @param array $opts Operation options
 558          * @return StatusValue
 559          * @since 1.21
 560          */
 561         final public function describe( array $params, array $opts = [] ) {
 562                 return $this->doOperation( [ 'op' => 'describe' ] + $params, $opts );
 563         }
 564
 565         /**
 566          * Perform a set of independent file operations on some files.
 567          *
 568          * This does no locking, nor journaling, and possibly no stat calls.
 569          * Any destination files that already exist will be overwritten.
 570          * This should *only* be used on non-original files, like cache files.
 571          *
 572          * Supported operations and their parameters:
 573          *  - create
 574          *  - store
 575          *  - copy
 576          *  - move
 577          *  - delete
 578          *  - describe (since 1.21)
 579          *  - null
 580          *
 581          * FSFile/TempFSFile object support was added in 1.27.
 582          *
 583          * a) Create a new file in storage with the contents of a string
 584          * @code
 585          *     [
 586          *         'op'                  => 'create',
 587          *         'dst'                 => <storage path>,
 588          *         'content'             => <string of new file contents>,
 589          *         'headers'             => <HTTP header name/value map> # since 1.21
 590          *     ]
 591          * @endcode
 592          *
 593          * b) Copy a file system file into storage
 594          * @code
 595          *     [
 596          *         'op'                  => 'store',
 597          *         'src'                 => <file system path, FSFile, or TempFSFile>,
 598          *         'dst'                 => <storage path>,
 599          *         'headers'             => <HTTP header name/value map> # since 1.21
 600          *     ]
 601          * @endcode
 602          *
 603          * c) Copy a file within storage
 604          * @code
 605          *     [
 606          *         'op'                  => 'copy',
 607          *         'src'                 => <storage path>,
 608          *         'dst'                 => <storage path>,
 609          *         'ignoreMissingSource' => <boolean>, # since 1.21
 610          *         'headers'             => <HTTP header name/value map> # since 1.21
 611          *     ]
 612          * @endcode
 613          *
 614          * d) Move a file within storage
 615          * @code
 616          *     [
 617          *         'op'                  => 'move',
 618          *         'src'                 => <storage path>,
 619          *         'dst'                 => <storage path>,
 620          *         'ignoreMissingSource' => <boolean>, # since 1.21
 621          *         'headers'             => <HTTP header name/value map> # since 1.21
 622          *     ]
 623          * @endcode
 624          *
 625          * e) Delete a file within storage
 626          * @code
 627          *     [
 628          *         'op'                  => 'delete',
 629          *         'src'                 => <storage path>,
 630          *         'ignoreMissingSource' => <boolean>
 631          *     ]
 632          * @endcode
 633          *
 634          * f) Update metadata for a file within storage
 635          * @code
 636          *     [
 637          *         'op'                  => 'describe',
 638          *         'src'                 => <storage path>,
 639          *         'headers'             => <HTTP header name/value map>
 640          *     ]
 641          * @endcode
 642          *
 643          * g) Do nothing (no-op)
 644          * @code
 645          *     [
 646          *         'op'                  => 'null',
 647          *     ]
 648          * @endcode
 649          *
 650          * @par Boolean flags for operations (operation-specific):
 651          *   - ignoreMissingSource : The operation will simply succeed and do
 652          *                           nothing if the source file does not exist.
 653          *   - headers             : If supplied with a header name/value map, the backend will
 654          *                           reply with these headers when GETs/HEADs of the destination
 655          *                           file are made. Header values should be smaller than 256 bytes.
 656          *                           Content-Disposition headers can be longer, though the system
 657          *                           might ignore or truncate ones that are too long to store.
 658          *                           Existing headers will remain, but these will replace any
 659          *                           conflicting previous headers, and headers will be removed
 660          *                           if they are set to an empty string.
 661          *                           Backends that don't support metadata ignore this. (since 1.21)
 662          *
 663          * $opts is an associative of boolean flags, including:
 664          *   - bypassReadOnly      : Allow writes in read-only mode (since 1.20)
 665          *
 666          * @par Return value:
 667          * This returns a Status, which contains all warnings and fatals that occurred
 668          * during the operation. The 'failCount', 'successCount', and 'success' members
 669          * will reflect each operation attempted for the given files. The StatusValue will be
 670          * considered "OK" as long as no fatal errors occurred.
 671          *
 672          * @param array $ops Set of operations to execute
 673          * @phan-param array{ignoreMissingSource?:bool,headers?:bool} $ops
 674          * @param array $opts Batch operation options
 675          * @phan-param array{bypassReadOnly?:bool} $opts
 676          * @return StatusValue
 677          * @since 1.20
 678          */
 679         final public function doQuickOperations( array $ops, array $opts = [] ) {
 680                 if ( empty( $opts['bypassReadOnly'] ) && $this->isReadOnly() ) {
 681                         return $this->newStatus( 'backend-fail-readonly', $this->name, $this->readOnly );
 682                 }
 683                 if ( $ops === [] ) {
 684                         return $this->newStatus(); // nothing to do
 685                 }
 686
 687                 $ops = $this->resolveFSFileObjects( $ops );
 688                 foreach ( $ops as &$op ) {
 689                         $op['overwrite'] = true; // avoids RTTs in key/value stores
 690                 }
 691
 692                 /** @noinspection PhpUnusedLocalVariableInspection */
 693                 $scope = ScopedCallback::newScopedIgnoreUserAbort(); // try to ignore client aborts
 694
 695                 return $this->doQuickOperationsInternal( $ops );
 696         }
 697
 698         /**
 699          * @see FileBackend::doQuickOperations()
 700          * @param array $ops
 701          * @return StatusValue
 702          * @since 1.20
 703          */
 704         abstract protected function doQuickOperationsInternal( array $ops );
 705
 706         /**
 707          * Same as doQuickOperations() except it takes a single operation.
 708          * If you are doing a batch of operations, then use that function instead.
 709          *
 710          * @see FileBackend::doQuickOperations()
 711          *
 712          * @param array $op Operation
 713          * @return StatusValue
 714          * @since 1.20
 715          */
 716         final public function doQuickOperation( array $op ) {
 717                 return $this->doQuickOperations( [ $op ] );
 718         }
 719
 720         /**
 721          * Performs a single quick create operation.
 722          * This sets $params['op'] to 'create' and passes it to doQuickOperation().
 723          *
 724          * @see FileBackend::doQuickOperation()
 725          *
 726          * @param array $params Operation parameters
 727          * @return StatusValue
 728          * @since 1.20
 729          */
 730         final public function quickCreate( array $params ) {
 731                 return $this->doQuickOperation( [ 'op' => 'create' ] + $params );
 732         }
 733
 734         /**
 735          * Performs a single quick store operation.
 736          * This sets $params['op'] to 'store' and passes it to doQuickOperation().
 737          *
 738          * @see FileBackend::doQuickOperation()
 739          *
 740          * @param array $params Operation parameters
 741          * @return StatusValue
 742          * @since 1.20
 743          */
 744         final public function quickStore( array $params ) {
 745                 return $this->doQuickOperation( [ 'op' => 'store' ] + $params );
 746         }
 747
 748         /**
 749          * Performs a single quick copy operation.
 750          * This sets $params['op'] to 'copy' and passes it to doQuickOperation().
 751          *
 752          * @see FileBackend::doQuickOperation()
 753          *
 754          * @param array $params Operation parameters
 755          * @return StatusValue
 756          * @since 1.20
 757          */
 758         final public function quickCopy( array $params ) {
 759                 return $this->doQuickOperation( [ 'op' => 'copy' ] + $params );
 760         }
 761
 762         /**
 763          * Performs a single quick move operation.
 764          * This sets $params['op'] to 'move' and passes it to doQuickOperation().
 765          *
 766          * @see FileBackend::doQuickOperation()
 767          *
 768          * @param array $params Operation parameters
 769          * @return StatusValue
 770          * @since 1.20
 771          */
 772         final public function quickMove( array $params ) {
 773                 return $this->doQuickOperation( [ 'op' => 'move' ] + $params );
 774         }
 775
 776         /**
 777          * Performs a single quick delete operation.
 778          * This sets $params['op'] to 'delete' and passes it to doQuickOperation().
 779          *
 780          * @see FileBackend::doQuickOperation()
 781          *
 782          * @param array $params Operation parameters
 783          * @return StatusValue
 784          * @since 1.20
 785          */
 786         final public function quickDelete( array $params ) {
 787                 return $this->doQuickOperation( [ 'op' => 'delete' ] + $params );
 788         }
 789
 790         /**
 791          * Performs a single quick describe operation.
 792          * This sets $params['op'] to 'describe' and passes it to doQuickOperation().
 793          *
 794          * @see FileBackend::doQuickOperation()
 795          *
 796          * @param array $params Operation parameters
 797          * @return StatusValue
 798          * @since 1.21
 799          */
 800         final public function quickDescribe( array $params ) {
 801                 return $this->doQuickOperation( [ 'op' => 'describe' ] + $params );
 802         }
 803
 804         /**
 805          * Concatenate a list of storage files into a single file system file.
 806          * The target path should refer to a file that is already locked or
 807          * otherwise safe from modification from other processes. Normally,
 808          * the file will be a new temp file, which should be adequate.
 809          *
 810          * @param array $params Operation parameters, include:
 811          *   - srcs        : ordered source storage paths (e.g. chunk1, chunk2, ...)
 812          *   - dst         : file system path to 0-byte temp file
 813          *   - parallelize : try to do operations in parallel when possible
 814          * @return StatusValue
 815          */
 816         abstract public function concatenate( array $params );
 817
 818         /**
 819          * Prepare a storage directory for usage.
 820          * This will create any required containers and parent directories.
 821          * Backends using key/value stores only need to create the container.
 822          *
 823          * The 'noAccess' and 'noListing' parameters works the same as in secure(),
 824          * except they are only applied *if* the directory/container had to be created.
 825          * These flags should always be set for directories that have private files.
 826          * However, setting them is not guaranteed to actually do anything.
 827          * Additional server configuration may be needed to achieve the desired effect.
 828          *
 829          * @param array $params Parameters include:
 830          *   - dir            : storage directory
 831          *   - noAccess       : try to deny file access (since 1.20)
 832          *   - noListing      : try to deny file listing (since 1.20)
 833          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 834          * @return StatusValue
 835          */
 836         final public function prepare( array $params ) {
 837                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 838                         return $this->newStatus( 'backend-fail-readonly', $this->name, $this->readOnly );
 839                 }
 840                 /** @noinspection PhpUnusedLocalVariableInspection */
 841                 $scope = ScopedCallback::newScopedIgnoreUserAbort(); // try to ignore client aborts
 842                 return $this->doPrepare( $params );
 843         }
 844
 845         /**
 846          * @see FileBackend::prepare()
 847          * @param array $params
 848          * @return StatusValue
 849          */
 850         abstract protected function doPrepare( array $params );
 851
 852         /**
 853          * Take measures to block web access to a storage directory and
 854          * the container it belongs to. FS backends might add .htaccess
 855          * files whereas key/value store backends might revoke container
 856          * access to the storage user representing end-users in web requests.
 857          *
 858          * This is not guaranteed to actually make files or listings publicly hidden.
 859          * Additional server configuration may be needed to achieve the desired effect.
 860          *
 861          * @param array $params Parameters include:
 862          *   - dir            : storage directory
 863          *   - noAccess       : try to deny file access
 864          *   - noListing      : try to deny file listing
 865          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 866          * @return StatusValue
 867          */
 868         final public function secure( array $params ) {
 869                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 870                         return $this->newStatus( 'backend-fail-readonly', $this->name, $this->readOnly );
 871                 }
 872                 /** @noinspection PhpUnusedLocalVariableInspection */
 873                 $scope = ScopedCallback::newScopedIgnoreUserAbort(); // try to ignore client aborts
 874                 return $this->doSecure( $params );
 875         }
 876
 877         /**
 878          * @see FileBackend::secure()
 879          * @param array $params
 880          * @return StatusValue
 881          */
 882         abstract protected function doSecure( array $params );
 883
 884         /**
 885          * Remove measures to block web access to a storage directory and
 886          * the container it belongs to. FS backends might remove .htaccess
 887          * files whereas key/value store backends might grant container
 888          * access to the storage user representing end-users in web requests.
 889          * This essentially can undo the result of secure() calls.
 890          *
 891          * This is not guaranteed to actually make files or listings publicly viewable.
 892          * Additional server configuration may be needed to achieve the desired effect.
 893          *
 894          * @param array $params Parameters include:
 895          *   - dir            : storage directory
 896          *   - access         : try to allow file access
 897          *   - listing        : try to allow file listing
 898          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 899          * @return StatusValue
 900          * @since 1.20
 901          */
 902         final public function publish( array $params ) {
 903                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 904                         return $this->newStatus( 'backend-fail-readonly', $this->name, $this->readOnly );
 905                 }
 906                 /** @noinspection PhpUnusedLocalVariableInspection */
 907                 $scope = ScopedCallback::newScopedIgnoreUserAbort(); // try to ignore client aborts
 908                 return $this->doPublish( $params );
 909         }
 910
 911         /**
 912          * @see FileBackend::publish()
 913          * @param array $params
 914          * @return StatusValue
 915          */
 916         abstract protected function doPublish( array $params );
 917
 918         /**
 919          * Delete a storage directory if it is empty.
 920          * Backends using key/value stores may do nothing unless the directory
 921          * is that of an empty container, in which case it will be deleted.
 922          *
 923          * @param array $params Parameters include:
 924          *   - dir            : storage directory
 925          *   - recursive      : recursively delete empty subdirectories first (since 1.20)
 926          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 927          * @return StatusValue
 928          */
 929         final public function clean( array $params ) {
 930                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 931                         return $this->newStatus( 'backend-fail-readonly', $this->name, $this->readOnly );
 932                 }
 933                 /** @noinspection PhpUnusedLocalVariableInspection */
 934                 $scope = ScopedCallback::newScopedIgnoreUserAbort(); // try to ignore client aborts
 935                 return $this->doClean( $params );
 936         }
 937
 938         /**
 939          * @see FileBackend::clean()
 940          * @param array $params
 941          * @return StatusValue
 942          */
 943         abstract protected function doClean( array $params );
 944
 945         /**
 946          * Check if a file exists at a storage path in the backend.
 947          * This returns false if only a directory exists at the path.
 948          *
 949          * @param array $params Parameters include:
 950          *   - src    : source storage path
 951          *   - latest : use the latest available data
 952          * @return bool|null Returns null on failure
 953          */
 954         abstract public function fileExists( array $params );
 955
 956         /**
 957          * Get the last-modified timestamp of the file at a storage path.
 958          *
 959          * @param array $params Parameters include:
 960          *   - src    : source storage path
 961          *   - latest : use the latest available data
 962          * @return string|bool TS_MW timestamp or false on failure
 963          */
 964         abstract public function getFileTimestamp( array $params );
 965
 966         /**
 967          * Get the contents of a file at a storage path in the backend.
 968          * This should be avoided for potentially large files.
 969          *
 970          * @param array $params Parameters include:
 971          *   - src    : source storage path
 972          *   - latest : use the latest available data
 973          * @return string|bool Returns false on failure
 974          */
 975         final public function getFileContents( array $params ) {
 976                 $contents = $this->getFileContentsMulti(
 977                         [ 'srcs' => [ $params['src'] ] ] + $params );
 978
 979                 return $contents[$params['src']];
 980         }
 981
 982         /**
 983          * Like getFileContents() except it takes an array of storage paths
 984          * and returns a map of storage paths to strings (or null on failure).
 985          * The map keys (paths) are in the same order as the provided list of paths.
 986          *
 987          * @see FileBackend::getFileContents()
 988          *
 989          * @param array $params Parameters include:
 990          *   - srcs        : list of source storage paths
 991          *   - latest      : use the latest available data
 992          *   - parallelize : try to do operations in parallel when possible
 993          * @return array Map of (path name => string or false on failure)
 994          * @since 1.20
 995          */
 996         abstract public function getFileContentsMulti( array $params );
 997
 998         /**
 999          * Get metadata about a file at a storage path in the backend.
1000          * If the file does not exist, then this returns false.
1001          * Otherwise, the result is an associative array that includes:
1002          *   - headers  : map of HTTP headers used for GET/HEAD requests (name => value)
1003          *   - metadata : map of file metadata (name => value)
1004          * Metadata keys and headers names will be returned in all lower-case.
1005          * Additional values may be included for internal use only.
1006          *
1007          * Use FileBackend::hasFeatures() to check how well this is supported.
1008          *
1009          * @param array $params
1010          * $params include:
1011          *   - src    : source storage path
1012          *   - latest : use the latest available data
1013          * @return array|bool Returns false on failure
1014          * @since 1.23
1015          */
1016         abstract public function getFileXAttributes( array $params );
1017
1018         /**
1019          * Get the size (bytes) of a file at a storage path in the backend.
1020          *
1021          * @param array $params Parameters include:
1022          *   - src    : source storage path
1023          *   - latest : use the latest available data
1024          * @return int|bool Returns false on failure
1025          */
1026         abstract public function getFileSize( array $params );
1027
1028         /**
1029          * Get quick information about a file at a storage path in the backend.
1030          * If the file does not exist, then this returns false.
1031          * Otherwise, the result is an associative array that includes:
1032          *   - mtime  : the last-modified timestamp (TS_MW)
1033          *   - size   : the file size (bytes)
1034          * Additional values may be included for internal use only.
1035          *
1036          * @param array $params Parameters include:
1037          *   - src    : source storage path
1038          *   - latest : use the latest available data
1039          * @return array|bool|null Returns null on failure
1040          */
1041         abstract public function getFileStat( array $params );
1042
1043         /**
1044          * Get a SHA-1 hash of the file at a storage path in the backend.
1045          *
1046          * @param array $params Parameters include:
1047          *   - src    : source storage path
1048          *   - latest : use the latest available data
1049          * @return string|bool Hash string or false on failure
1050          */
1051         abstract public function getFileSha1Base36( array $params );
1052
1053         /**
1054          * Get the properties of the file at a storage path in the backend.
1055          * This gives the result of FSFile::getProps() on a local copy of the file.
1056          *
1057          * @param array $params Parameters include:
1058          *   - src    : source storage path
1059          *   - latest : use the latest available data
1060          * @return array Returns FSFile::placeholderProps() on failure
1061          */
1062         abstract public function getFileProps( array $params );
1063
1064         /**
1065          * Stream the file at a storage path in the backend.
1066          *
1067          * If the file does not exists, an HTTP 404 error will be given.
1068          * Appropriate HTTP headers (Status, Content-Type, Content-Length)
1069          * will be sent if streaming began, while none will be sent otherwise.
1070          * Implementations should flush the output buffer before sending data.
1071          *
1072          * @param array $params Parameters include:
1073          *   - src      : source storage path
1074          *   - headers  : list of additional HTTP headers to send if the file exists
1075          *   - options  : HTTP request header map with lower case keys (since 1.28). Supports:
1076          *                range             : format is "bytes=(\d*-\d*)"
1077          *                if-modified-since : format is an HTTP date
1078          *   - headless : only include the body (and headers from "headers") (since 1.28)
1079          *   - latest   : use the latest available data
1080          *   - allowOB  : preserve any output buffers (since 1.28)
1081          * @return StatusValue
1082          */
1083         abstract public function streamFile( array $params );
1084
1085         /**
1086          * Returns a file system file, identical to the file at a storage path.
1087          * The file returned is either:
1088          *   - a) A local copy of the file at a storage path in the backend.
1089          *        The temporary copy will have the same extension as the source.
1090          *   - b) An original of the file at a storage path in the backend.
1091          * Temporary files may be purged when the file object falls out of scope.
1092          *
1093          * Write operations should *never* be done on this file as some backends
1094          * may do internal tracking or may be instances of FileBackendMultiWrite.
1095          * In that latter case, there are copies of the file that must stay in sync.
1096          * Additionally, further calls to this function may return the same file.
1097          *
1098          * @param array $params Parameters include:
1099          *   - src    : source storage path
1100          *   - latest : use the latest available data
1101          * @return FSFile|null Returns null on failure
1102          */
1103         final public function getLocalReference( array $params ) {
1104                 $fsFiles = $this->getLocalReferenceMulti(
1105                         [ 'srcs' => [ $params['src'] ] ] + $params );
1106
1107                 return $fsFiles[$params['src']];
1108         }
1109
1110         /**
1111          * Like getLocalReference() except it takes an array of storage paths
1112          * and returns a map of storage paths to FSFile objects (or null on failure).
1113          * The map keys (paths) are in the same order as the provided list of paths.
1114          *
1115          * @see FileBackend::getLocalReference()
1116          *
1117          * @param array $params Parameters include:
1118          *   - srcs        : list of source storage paths
1119          *   - latest      : use the latest available data
1120          *   - parallelize : try to do operations in parallel when possible
1121          * @return array Map of (path name => FSFile or null on failure)
1122          * @since 1.20
1123          */
1124         abstract public function getLocalReferenceMulti( array $params );
1125
1126         /**
1127          * Get a local copy on disk of the file at a storage path in the backend.
1128          * The temporary copy will have the same file extension as the source.
1129          * Temporary files may be purged when the file object falls out of scope.
1130          *
1131          * @param array $params Parameters include:
1132          *   - src    : source storage path
1133          *   - latest : use the latest available data
1134          * @return TempFSFile|null Returns null on failure
1135          */
1136         final public function getLocalCopy( array $params ) {
1137                 $tmpFiles = $this->getLocalCopyMulti(
1138                         [ 'srcs' => [ $params['src'] ] ] + $params );
1139
1140                 return $tmpFiles[$params['src']];
1141         }
1142
1143         /**
1144          * Like getLocalCopy() except it takes an array of storage paths and
1145          * returns a map of storage paths to TempFSFile objects (or null on failure).
1146          * The map keys (paths) are in the same order as the provided list of paths.
1147          *
1148          * @see FileBackend::getLocalCopy()
1149          *
1150          * @param array $params Parameters include:
1151          *   - srcs        : list of source storage paths
1152          *   - latest      : use the latest available data
1153          *   - parallelize : try to do operations in parallel when possible
1154          * @return array Map of (path name => TempFSFile or null on failure)
1155          * @since 1.20
1156          */
1157         abstract public function getLocalCopyMulti( array $params );
1158
1159         /**
1160          * Return an HTTP URL to a given file that requires no authentication to use.
1161          * The URL may be pre-authenticated (via some token in the URL) and temporary.
1162          * This will return null if the backend cannot make an HTTP URL for the file.
1163          *
1164          * This is useful for key/value stores when using scripts that seek around
1165          * large files and those scripts (and the backend) support HTTP Range headers.
1166          * Otherwise, one would need to use getLocalReference(), which involves loading
1167          * the entire file on to local disk.
1168          *
1169          * @param array $params Parameters include:
1170          *   - src : source storage path
1171          *   - ttl : lifetime (seconds) if pre-authenticated; default is 1 day
1172          * @return string|null
1173          * @since 1.21
1174          */
1175         abstract public function getFileHttpUrl( array $params );
1176
1177         /**
1178          * Check if a directory exists at a given storage path
1179          *
1180          * For backends using key/value stores, a directory is said to exist whenever
1181          * there exist any files with paths using the given directory path as a prefix
1182          * followed by a forward slash. For example, if there is a file called
1183          * "mwstore://backend/container/dir/path.svg" then directories are said to exist
1184          * at "mwstore://backend/container" and "mwstore://backend/container/dir". These
1185          * can be thought of as "virtual" directories.
1186          *
1187          * Backends that directly use a filesystem layer might enumerate empty directories.
1188          * The clean() method should always be used when files are deleted or moved if this
1189          * is a concern. This is a trade-off to avoid write amplication/contention on file
1190          * changes or read amplification when calling this method.
1191          *
1192          * Storage backends with eventual consistency might return stale data.
1193          *
1194          * @see FileBackend::clean()
1195          *
1196          * @param array $params Parameters include:
1197          *   - dir : storage directory
1198          * @return bool|null Whether a directory exists or null on failure
1199          * @since 1.20
1200          */
1201         abstract public function directoryExists( array $params );
1202
1203         /**
1204          * Get an iterator to list *all* directories under a storage directory
1205          *
1206          * If the directory is of the form "mwstore://backend/container",
1207          * then all directories in the container will be listed.
1208          * If the directory is of form "mwstore://backend/container/dir",
1209          * then all directories directly under that directory will be listed.
1210          * Results will be storage directories relative to the given directory.
1211          *
1212          * Storage backends with eventual consistency might return stale data.
1213          *
1214          * Failures during iteration can result in FileBackendError exceptions (since 1.22).
1215          *
1216          * @see FileBackend::directoryExists()
1217          *
1218          * @param array $params Parameters include:
1219          *   - dir     : storage directory
1220          *   - topOnly : only return direct child dirs of the directory
1221          * @return Traversable|array|null Directory list enumerator null on failure
1222          * @since 1.20
1223          */
1224         abstract public function getDirectoryList( array $params );
1225
1226         /**
1227          * Same as FileBackend::getDirectoryList() except only lists
1228          * directories that are immediately under the given directory.
1229          *
1230          * Storage backends with eventual consistency might return stale data.
1231          *
1232          * Failures during iteration can result in FileBackendError exceptions (since 1.22).
1233          *
1234          * @see FileBackend::directoryExists()
1235          *
1236          * @param array $params Parameters include:
1237          *   - dir : storage directory
1238          * @return Traversable|array|null Directory list enumerator or null on failure
1239          * @since 1.20
1240          */
1241         final public function getTopDirectoryList( array $params ) {
1242                 return $this->getDirectoryList( [ 'topOnly' => true ] + $params );
1243         }
1244
1245         /**
1246          * Get an iterator to list *all* stored files under a storage directory
1247          *
1248          * If the directory is of the form "mwstore://backend/container", then all
1249          * files in the container will be listed. If the directory is of form
1250          * "mwstore://backend/container/dir", then all files under that directory will
1251          * be listed. Results will be storage paths relative to the given directory.
1252          *
1253          * Storage backends with eventual consistency might return stale data.
1254          *
1255          * Failures during iteration can result in FileBackendError exceptions (since 1.22).
1256          *
1257          * @param array $params Parameters include:
1258          *   - dir        : storage directory
1259          *   - topOnly    : only return direct child files of the directory (since 1.20)
1260          *   - adviseStat : set to true if stat requests will be made on the files (since 1.22)
1261          * @return Traversable|array|null File list enumerator or null on failure
1262          */
1263         abstract public function getFileList( array $params );
1264
1265         /**
1266          * Same as FileBackend::getFileList() except only lists
1267          * files that are immediately under the given directory.
1268          *
1269          * Storage backends with eventual consistency might return stale data.
1270          *
1271          * Failures during iteration can result in FileBackendError exceptions (since 1.22).
1272          *
1273          * @param array $params Parameters include:
1274          *   - dir        : storage directory
1275          *   - adviseStat : set to true if stat requests will be made on the files (since 1.22)
1276          * @return Traversable|array|null File list enumerator or null on failure
1277          * @since 1.20
1278          */
1279         final public function getTopFileList( array $params ) {
1280                 return $this->getFileList( [ 'topOnly' => true ] + $params );
1281         }
1282
1283         /**
1284          * Preload persistent file stat cache and property cache into in-process cache.
1285          * This should be used when stat calls will be made on a known list of a many files.
1286          *
1287          * @see FileBackend::getFileStat()
1288          *
1289          * @param array $paths Storage paths
1290          */
1291         abstract public function preloadCache( array $paths );
1292
1293         /**
1294          * Invalidate any in-process file stat and property cache.
1295          * If $paths is given, then only the cache for those files will be cleared.
1296          *
1297          * @see FileBackend::getFileStat()
1298          *
1299          * @param array|null $paths Storage paths (optional)
1300          */
1301         abstract public function clearCache( array $paths = null );
1302
1303         /**
1304          * Preload file stat information (concurrently if possible) into in-process cache.
1305          *
1306          * This should be used when stat calls will be made on a known list of a many files.
1307          * This does not make use of the persistent file stat cache.
1308          *
1309          * @see FileBackend::getFileStat()
1310          *
1311          * @param array $params Parameters include:
1312          *   - srcs        : list of source storage paths
1313          *   - latest      : use the latest available data
1314          * @return bool Whether all requests proceeded without I/O errors (since 1.24)
1315          * @since 1.23
1316          */
1317         abstract public function preloadFileStat( array $params );
1318
1319         /**
1320          * Lock the files at the given storage paths in the backend.
1321          * This will either lock all the files or none (on failure).
1322          *
1323          * Callers should consider using getScopedFileLocks() instead.
1324          *
1325          * @param array $paths Storage paths
1326          * @param int $type LockManager::LOCK_* constant
1327          * @param int $timeout Timeout in seconds (0 means non-blocking) (since 1.24)
1328          * @return StatusValue
1329          */
1330         final public function lockFiles( array $paths, $type, $timeout = 0 ) {
1331                 $paths = array_map( 'FileBackend::normalizeStoragePath', $paths );
1332
1333                 return $this->wrapStatus( $this->lockManager->lock( $paths, $type, $timeout ) );
1334         }
1335
1336         /**
1337          * Unlock the files at the given storage paths in the backend.
1338          *
1339          * @param array $paths Storage paths
1340          * @param int $type LockManager::LOCK_* constant
1341          * @return StatusValue
1342          */
1343         final public function unlockFiles( array $paths, $type ) {
1344                 $paths = array_map( 'FileBackend::normalizeStoragePath', $paths );
1345
1346                 return $this->wrapStatus( $this->lockManager->unlock( $paths, $type ) );
1347         }
1348
1349         /**
1350          * Lock the files at the given storage paths in the backend.
1351          * This will either lock all the files or none (on failure).
1352          * On failure, the StatusValue object will be updated with errors.
1353          *
1354          * Once the return value goes out scope, the locks will be released and
1355          * the StatusValue updated. Unlock fatals will not change the StatusValue "OK" value.
1356          *
1357          * @see ScopedLock::factory()
1358          *
1359          * @param array $paths List of storage paths or map of lock types to path lists
1360          * @param int|string $type LockManager::LOCK_* constant or "mixed"
1361          * @param StatusValue $status StatusValue to update on lock/unlock
1362          * @param int $timeout Timeout in seconds (0 means non-blocking) (since 1.24)
1363          * @return ScopedLock|null Returns null on failure
1364          */
1365         final public function getScopedFileLocks(
1366                 array $paths, $type, StatusValue $status, $timeout = 0
1367         ) {
1368                 if ( $type === 'mixed' ) {
1369                         foreach ( $paths as &$typePaths ) {
1370                                 $typePaths = array_map( 'FileBackend::normalizeStoragePath', $typePaths );
1371                         }
1372                 } else {
1373                         $paths = array_map( 'FileBackend::normalizeStoragePath', $paths );
1374                 }
1375
1376                 return ScopedLock::factory( $this->lockManager, $paths, $type, $status, $timeout );
1377         }
1378
1379         /**
1380          * Get an array of scoped locks needed for a batch of file operations.
1381          *
1382          * Normally, FileBackend::doOperations() handles locking, unless
1383          * the 'nonLocking' param is passed in. This function is useful if you
1384          * want the files to be locked for a broader scope than just when the
1385          * files are changing. For example, if you need to update DB metadata,
1386          * you may want to keep the files locked until finished.
1387          *
1388          * @see FileBackend::doOperations()
1389          *
1390          * @param array $ops List of file operations to FileBackend::doOperations()
1391          * @param StatusValue $status StatusValue to update on lock/unlock
1392          * @return ScopedLock|null
1393          * @since 1.20
1394          */
1395         abstract public function getScopedLocksForOps( array $ops, StatusValue $status );
1396
1397         /**
1398          * Get the root storage path of this backend.
1399          * All container paths are "subdirectories" of this path.
1400          *
1401          * @return string Storage path
1402          * @since 1.20
1403          */
1404         final public function getRootStoragePath() {
1405                 return "mwstore://{$this->name}";
1406         }
1407
1408         /**
1409          * Get the storage path for the given container for this backend
1410          *
1411          * @param string $container Container name
1412          * @return string Storage path
1413          * @since 1.21
1414          */
1415         final public function getContainerStoragePath( $container ) {
1416                 return $this->getRootStoragePath() . "/{$container}";
1417         }
1418
1419         /**
1420          * Get the file journal object for this backend
1421          *
1422          * @return FileJournal
1423          */
1424         final public function getJournal() {
1425                 return $this->fileJournal;
1426         }
1427
1428         /**
1429          * Convert FSFile 'src' paths to string paths (with an 'srcRef' field set to the FSFile)
1430          *
1431          * The 'srcRef' field keeps any TempFSFile objects in scope for the backend to have it
1432          * around as long it needs (which may vary greatly depending on configuration)
1433          *
1434          * @param array $ops File operation batch for FileBaclend::doOperations()
1435          * @return array File operation batch
1436          */
1437         protected function resolveFSFileObjects( array $ops ) {
1438                 foreach ( $ops as &$op ) {
1439                         $src = $op['src'] ?? null;
1440                         if ( $src instanceof FSFile ) {
1441                                 $op['srcRef'] = $src;
1442                                 $op['src'] = $src->getPath();
1443                         }
1444                 }
1445                 unset( $op );
1446
1447                 return $ops;
1448         }
1449
1450         /**
1451          * Check if a given path is a "mwstore://" path.
1452          * This does not do any further validation or any existence checks.
1453          *
1454          * @param string $path
1455          * @return bool
1456          */
1457         final public static function isStoragePath( $path ) {
1458                 return ( strpos( $path, 'mwstore://' ) === 0 );
1459         }
1460
1461         /**
1462          * Split a storage path into a backend name, a container name,
1463          * and a relative file path. The relative path may be the empty string.
1464          * This does not do any path normalization or traversal checks.
1465          *
1466          * @param string $storagePath
1467          * @return array (backend, container, rel object) or (null, null, null)
1468          */
1469         final public static function splitStoragePath( $storagePath ) {
1470                 if ( self::isStoragePath( $storagePath ) ) {
1471                         // Remove the "mwstore://" prefix and split the path
1472                         $parts = explode( '/', substr( $storagePath, 10 ), 3 );
1473                         if ( count( $parts ) >= 2 && $parts[0] != '' && $parts[1] != '' ) {
1474                                 if ( count( $parts ) == 3 ) {
1475                                         return $parts; // e.g. "backend/container/path"
1476                                 } else {
1477                                         return [ $parts[0], $parts[1], '' ]; // e.g. "backend/container"
1478                                 }
1479                         }
1480                 }
1481
1482                 return [ null, null, null ];
1483         }
1484
1485         /**
1486          * Normalize a storage path by cleaning up directory separators.
1487          * Returns null if the path is not of the format of a valid storage path.
1488          *
1489          * @param string $storagePath
1490          * @return string|null
1491          */
1492         final public static function normalizeStoragePath( $storagePath ) {
1493                 list( $backend, $container, $relPath ) = self::splitStoragePath( $storagePath );
1494                 if ( $relPath !== null ) { // must be for this backend
1495                         $relPath = self::normalizeContainerPath( $relPath );
1496                         if ( $relPath !== null ) {
1497                                 return ( $relPath != '' )
1498                                         ? "mwstore://{$backend}/{$container}/{$relPath}"
1499                                         : "mwstore://{$backend}/{$container}";
1500                         }
1501                 }
1502
1503                 return null;
1504         }
1505
1506         /**
1507          * Get the parent storage directory of a storage path.
1508          * This returns a path like "mwstore://backend/container",
1509          * "mwstore://backend/container/...", or null if there is no parent.
1510          *
1511          * @param string $storagePath
1512          * @return string|null
1513          */
1514         final public static function parentStoragePath( $storagePath ) {
1515                 $storagePath = dirname( $storagePath );
1516                 list( , , $rel ) = self::splitStoragePath( $storagePath );
1517
1518                 return ( $rel === null ) ? null : $storagePath;
1519         }
1520
1521         /**
1522          * Get the final extension from a storage or FS path
1523          *
1524          * @param string $path
1525          * @param string $case One of (rawcase, uppercase, lowercase) (since 1.24)
1526          * @return string
1527          */
1528         final public static function extensionFromPath( $path, $case = 'lowercase' ) {
1529                 $i = strrpos( $path, '.' );
1530                 $ext = $i ? substr( $path, $i + 1 ) : '';
1531
1532                 if ( $case === 'lowercase' ) {
1533                         $ext = strtolower( $ext );
1534                 } elseif ( $case === 'uppercase' ) {
1535                         $ext = strtoupper( $ext );
1536                 }
1537
1538                 return $ext;
1539         }
1540
1541         /**
1542          * Check if a relative path has no directory traversals
1543          *
1544          * @param string $path
1545          * @return bool
1546          * @since 1.20
1547          */
1548         final public static function isPathTraversalFree( $path ) {
1549                 return ( self::normalizeContainerPath( $path ) !== null );
1550         }
1551
1552         /**
1553          * Build a Content-Disposition header value per RFC 6266.
1554          *
1555          * @param string $type One of (attachment, inline)
1556          * @param string $filename Suggested file name (should not contain slashes)
1557          * @throws InvalidArgumentException
1558          * @return string
1559          * @since 1.20
1560          */
1561         final public static function makeContentDisposition( $type, $filename = '' ) {
1562                 $parts = [];
1563
1564                 $type = strtolower( $type );
1565                 if ( !in_array( $type, [ 'inline', 'attachment' ] ) ) {
1566                         throw new InvalidArgumentException( "Invalid Content-Disposition type '$type'." );
1567                 }
1568                 $parts[] = $type;
1569
1570                 if ( strlen( $filename ) ) {
1571                         $parts[] = "filename*=UTF-8''" . rawurlencode( basename( $filename ) );
1572                 }
1573
1574                 return implode( ';', $parts );
1575         }
1576
1577         /**
1578          * Validate and normalize a relative storage path.
1579          * Null is returned if the path involves directory traversal.
1580          * Traversal is insecure for FS backends and broken for others.
1581          *
1582          * This uses the same traversal protection as Title::secureAndSplit().
1583          *
1584          * @param string $path Storage path relative to a container
1585          * @return string|null
1586          */
1587         final protected static function normalizeContainerPath( $path ) {
1588                 // Normalize directory separators
1589                 $path = strtr( $path, '\\', '/' );
1590                 // Collapse any consecutive directory separators
1591                 $path = preg_replace( '![/]{2,}!', '/', $path );
1592                 // Remove any leading directory separator
1593                 $path = ltrim( $path, '/' );
1594                 // Use the same traversal protection as Title::secureAndSplit()
1595                 if ( strpos( $path, '.' ) !== false ) {
1596                         if (
1597                                 $path === '.' ||
1598                                 $path === '..' ||
1599                                 strpos( $path, './' ) === 0 ||
1600                                 strpos( $path, '../' ) === 0 ||
1601                                 strpos( $path, '/./' ) !== false ||
1602                                 strpos( $path, '/../' ) !== false
1603                         ) {
1604                                 return null;
1605                         }
1606                 }
1607
1608                 return $path;
1609         }
1610
1611         /**
1612          * Yields the result of the status wrapper callback on either:
1613          *   - StatusValue::newGood() if this method is called without parameters
1614          *   - StatusValue::newFatal() with all parameters to this method if passed in
1615          *
1616          * @param string ...$args
1617          * @return StatusValue
1618          */
1619         final protected function newStatus( ...$args ) {
1620                 if ( count( $args ) ) {
1621                         $sv = StatusValue::newFatal( ...$args );
1622                 } else {
1623                         $sv = StatusValue::newGood();
1624                 }
1625
1626                 return $this->wrapStatus( $sv );
1627         }
1628
1629         /**
1630          * @param StatusValue $sv
1631          * @return StatusValue Modified status or StatusValue subclass
1632          */
1633         final protected function wrapStatus( StatusValue $sv ) {
1634                 return $this->statusWrapper ? call_user_func( $this->statusWrapper, $sv ) : $sv;
1635         }
1636
1637         /**
1638          * @param string $section
1639          * @return ScopedCallback|null
1640          */
1641         protected function scopedProfileSection( $section ) {
1642                 return $this->profiler ? ( $this->profiler )( $section ) : null;
1643         }
1644
1645         protected function resetOutputBuffer() {
1646                 while ( ob_get_status() ) {
1647                         if ( !ob_end_clean() ) {
1648                                 // Could not remove output buffer handler; abort now
1649                                 // to avoid getting in some kind of infinite loop.
1650                                 break;
1651                         }
1652                 }
1653         }
1654 }