includes/filebackend/FileBackend.php

   1 <?php
   2 /**
   3  * @defgroup FileBackend File backend
   4  * @ingroup  FileRepo
   5  *
   6  * File backend is used to interact with file storage systems,
   7  * such as the local file system, NFS, or cloud storage systems.
   8  */
   9
  10 /**
  11  * Base class for all file backends.
  12  *
  13  * This program is free software; you can redistribute it and/or modify
  14  * it under the terms of the GNU General Public License as published by
  15  * the Free Software Foundation; either version 2 of the License, or
  16  * (at your option) any later version.
  17  *
  18  * This program is distributed in the hope that it will be useful,
  19  * but WITHOUT ANY WARRANTY; without even the implied warranty of
  20  * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
  21  * GNU General Public License for more details.
  22  *
  23  * You should have received a copy of the GNU General Public License along
  24  * with this program; if not, write to the Free Software Foundation, Inc.,
  25  * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
  26  * http://www.gnu.org/copyleft/gpl.html
  27  *
  28  * @file
  29  * @ingroup FileBackend
  30  * @author Aaron Schulz
  31  */
  32
  33 /**
  34  * @brief Base class for all file backend classes (including multi-write backends).
  35  *
  36  * This class defines the methods as abstract that subclasses must implement.
  37  * Outside callers can assume that all backends will have these functions.
  38  *
  39  * All "storage paths" are of the format "mwstore://<backend>/<container>/<path>".
  40  * The "<path>" portion is a relative path that uses UNIX file system (FS)
  41  * notation, though any particular backend may not actually be using a local
  42  * filesystem. Therefore, the relative paths are only virtual.
  43  *
  44  * Backend contents are stored under wiki-specific container names by default.
  45  * For legacy reasons, this has no effect for the FS backend class, and per-wiki
  46  * segregation must be done by setting the container paths appropriately.
  47  *
  48  * FS-based backends are somewhat more restrictive due to the existence of real
  49  * directory files; a regular file cannot have the same name as a directory. Other
  50  * backends with virtual directories may not have this limitation. Callers should
  51  * store files in such a way that no files and directories are under the same path.
  52  *
  53  * Methods of subclasses should avoid throwing exceptions at all costs.
  54  * As a corollary, external dependencies should be kept to a minimum.
  55  *
  56  * @ingroup FileBackend
  57  * @since 1.19
  58  */
  59 abstract class FileBackend {
  60         protected $name; // string; unique backend name
  61         protected $wikiId; // string; unique wiki name
  62         protected $readOnly; // string; read-only explanation message
  63         protected $parallelize; // string; when to do operations in parallel
  64         protected $concurrency; // integer; how many operations can be done in parallel
  65
  66         /** @var LockManager */
  67         protected $lockManager;
  68         /** @var FileJournal */
  69         protected $fileJournal;
  70
  71         /**
  72          * Create a new backend instance from configuration.
  73          * This should only be called from within FileBackendGroup.
  74          *
  75          * $config includes:
  76          *   - name        : The unique name of this backend.
  77          *                   This should consist of alphanumberic, '-', and '_' characters.
  78          *                   This name should not be changed after use.
  79          *   - wikiId      : Prefix to container names that is unique to this wiki.
  80          *                   If not provided, this defaults to the current wiki ID.
  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( $this->wikiId )->get( $config['lockManager'] );
 105                 $this->fileJournal = isset( $config['fileJournal'] )
 106                         ? ( ( $config['fileJournal'] instanceof FileJournal )
 107                                 ? $config['fileJournal']
 108                                 : FileJournal::factory( $config['fileJournal'], $this->name ) )
 109                         : FileJournal::factory( array( 'class' => 'NullFileJournal' ), $this->name );
 110                 $this->readOnly = isset( $config['readOnly'] )
 111                         ? (string)$config['readOnly']
 112                         : '';
 113                 $this->parallelize = isset( $config['parallelize'] )
 114                         ? (string)$config['parallelize']
 115                         : 'off';
 116                 $this->concurrency = isset( $config['concurrency'] )
 117                         ? (int)$config['concurrency']
 118                         : 50;
 119         }
 120
 121         /**
 122          * Get the unique backend name.
 123          * We may have multiple different backends of the same type.
 124          * For example, we can have two Swift backends using different proxies.
 125          *
 126          * @return string
 127          */
 128         final public function getName() {
 129                 return $this->name;
 130         }
 131
 132         /**
 133          * Get the wiki identifier used for this backend (possibly empty).
 134          * Note that this might *not* be in the same format as wfWikiID().
 135          *
 136          * @return string
 137          * @since 1.20
 138          */
 139         final public function getWikiId() {
 140                 return $this->wikiId;
 141         }
 142
 143         /**
 144          * Check if this backend is read-only
 145          *
 146          * @return bool
 147          */
 148         final public function isReadOnly() {
 149                 return ( $this->readOnly != '' );
 150         }
 151
 152         /**
 153          * Get an explanatory message if this backend is read-only
 154          *
 155          * @return string|bool Returns false if the backend is not read-only
 156          */
 157         final public function getReadOnlyReason() {
 158                 return ( $this->readOnly != '' ) ? $this->readOnly : false;
 159         }
 160
 161         /**
 162          * This is the main entry point into the backend for write operations.
 163          * Callers supply an ordered list of operations to perform as a transaction.
 164          * Files will be locked, the stat cache cleared, and then the operations attempted.
 165          * If any serious errors occur, all attempted operations will be rolled back.
 166          *
 167          * $ops is an array of arrays. The outer array holds a list of operations.
 168          * Each inner array is a set of key value pairs that specify an operation.
 169          *
 170          * Supported operations and their parameters. The supported actions are:
 171          *  - create
 172          *  - store
 173          *  - copy
 174          *  - move
 175          *  - delete
 176          *  - describe (since 1.21)
 177          *  - null
 178          *
 179          * a) Create a new file in storage with the contents of a string
 180          * @code
 181          *     array(
 182          *         'op'                  => 'create',
 183          *         'dst'                 => <storage path>,
 184          *         'content'             => <string of new file contents>,
 185          *         'overwrite'           => <boolean>,
 186          *         'overwriteSame'       => <boolean>,
 187          *         'disposition'         => <Content-Disposition header value>,
 188          *         'headers'             => <HTTP header name/value map> # since 1.21
 189          *     );
 190          * @endcode
 191          *
 192          * b) Copy a file system file into storage
 193          * @code
 194          *     array(
 195          *         'op'                  => 'store',
 196          *         'src'                 => <file system path>,
 197          *         'dst'                 => <storage path>,
 198          *         'overwrite'           => <boolean>,
 199          *         'overwriteSame'       => <boolean>,
 200          *         'disposition'         => <Content-Disposition header value>,
 201          *         'headers'             => <HTTP header name/value map> # since 1.21
 202          *     )
 203          * @endcode
 204          *
 205          * c) Copy a file within storage
 206          * @code
 207          *     array(
 208          *         'op'                  => 'copy',
 209          *         'src'                 => <storage path>,
 210          *         'dst'                 => <storage path>,
 211          *         'overwrite'           => <boolean>,
 212          *         'overwriteSame'       => <boolean>,
 213          *         'ignoreMissingSource' => <boolean>, # since 1.21
 214          *         'disposition'         => <Content-Disposition header value>
 215          *     )
 216          * @endcode
 217          *
 218          * d) Move a file within storage
 219          * @code
 220          *     array(
 221          *         'op'                  => 'move',
 222          *         'src'                 => <storage path>,
 223          *         'dst'                 => <storage path>,
 224          *         'overwrite'           => <boolean>,
 225          *         'overwriteSame'       => <boolean>,
 226          *         'ignoreMissingSource' => <boolean>, # since 1.21
 227          *         'disposition'         => <Content-Disposition header value>
 228          *     )
 229          * @endcode
 230          *
 231          * e) Delete a file within storage
 232          * @code
 233          *     array(
 234          *         'op'                  => 'delete',
 235          *         'src'                 => <storage path>,
 236          *         'ignoreMissingSource' => <boolean>
 237          *     )
 238          * @endcode
 239          *
 240          * f) Update metadata for a file within storage
 241          * @code
 242          *     array(
 243          *         'op'                  => 'describe',
 244          *         'src'                 => <storage path>,
 245          *         'disposition'         => <Content-Disposition header value>,
 246          *         'headers'             => <HTTP header name/value map>
 247          *     )
 248          * @endcode
 249          *
 250          * g) Do nothing (no-op)
 251          * @code
 252          *     array(
 253          *         'op'                  => 'null',
 254          *     )
 255          * @endcode
 256          *
 257          * Boolean flags for operations (operation-specific):
 258          *   - ignoreMissingSource : The operation will simply succeed and do
 259          *                           nothing if the source file does not exist.
 260          *   - overwrite           : Any destination file will be overwritten.
 261          *   - overwriteSame       : An error will not be given if a file already
 262          *                           exists at the destination that has the same
 263          *                           contents as the new contents to be written there.
 264          *   - disposition         : If supplied, the backend will return a Content-Disposition
 265          *                           header when GETs/HEADs of the destination file are made.
 266          *                           Backends that don't support metadata ignore this.
 267          *                           See http://tools.ietf.org/html/rfc6266. (since 1.20)
 268          *   - headers             : If supplied, the backend will return these headers when
 269          *                           GETs/HEADs of the destination file are made. Header values
 270          *                           should be smaller than 256 bytes, often options or numbers.
 271          *                           Existing headers will remain, but these will replace any
 272          *                           conflicting previous headers, and headers will be removed
 273          *                           if they are set to an empty string.
 274          *                           Backends that don't support metadata ignore this. (since 1.21)
 275          *
 276          * $opts is an associative of boolean flags, including:
 277          *   - force               : Operation precondition errors no longer trigger an abort.
 278          *                           Any remaining operations are still attempted. Unexpected
 279          *                           failures may still cause remaning operations to be aborted.
 280          *   - nonLocking          : No locks are acquired for the operations.
 281          *                           This can increase performance for non-critical writes.
 282          *                           This has no effect unless the 'force' flag is set.
 283          *   - nonJournaled        : Don't log this operation batch in the file journal.
 284          *                           This limits the ability of recovery scripts.
 285          *   - parallelize         : Try to do operations in parallel when possible.
 286          *   - bypassReadOnly      : Allow writes in read-only mode. (since 1.20)
 287          *   - preserveCache       : Don't clear the process cache before checking files.
 288          *                           This should only be used if all entries in the process
 289          *                           cache were added after the files were already locked. (since 1.20)
 290          *
 291          * @remarks Remarks on locking:
 292          * File system paths given to operations should refer to files that are
 293          * already locked or otherwise safe from modification from other processes.
 294          * Normally these files will be new temp files, which should be adequate.
 295          *
 296          * @par Return value:
 297          *
 298          * This returns a Status, which contains all warnings and fatals that occurred
 299          * during the operation. The 'failCount', 'successCount', and 'success' members
 300          * will reflect each operation attempted.
 301          *
 302          * The status will be "OK" unless:
 303          *   - a) unexpected operation errors occurred (network partitions, disk full...)
 304          *   - b) significant operation errors occurred and 'force' was not set
 305          *
 306          * @param $ops Array List of operations to execute in order
 307          * @param $opts Array Batch operation options
 308          * @return Status
 309          */
 310         final public function doOperations( array $ops, array $opts = array() ) {
 311                 if ( empty( $opts['bypassReadOnly'] ) && $this->isReadOnly() ) {
 312                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 313                 }
 314                 if ( empty( $opts['force'] ) ) { // sanity
 315                         unset( $opts['nonLocking'] );
 316                 }
 317                 return $this->doOperationsInternal( $ops, $opts );
 318         }
 319
 320         /**
 321          * @see FileBackend::doOperations()
 322          */
 323         abstract protected function doOperationsInternal( array $ops, array $opts );
 324
 325         /**
 326          * Same as doOperations() except it takes a single operation.
 327          * If you are doing a batch of operations that should either
 328          * all succeed or all fail, then use that function instead.
 329          *
 330          * @see FileBackend::doOperations()
 331          *
 332          * @param $op Array Operation
 333          * @param $opts Array Operation options
 334          * @return Status
 335          */
 336         final public function doOperation( array $op, array $opts = array() ) {
 337                 return $this->doOperations( array( $op ), $opts );
 338         }
 339
 340         /**
 341          * Performs a single create operation.
 342          * This sets $params['op'] to 'create' and passes it to doOperation().
 343          *
 344          * @see FileBackend::doOperation()
 345          *
 346          * @param $params Array Operation parameters
 347          * @param $opts Array Operation options
 348          * @return Status
 349          */
 350         final public function create( array $params, array $opts = array() ) {
 351                 return $this->doOperation( array( 'op' => 'create' ) + $params, $opts );
 352         }
 353
 354         /**
 355          * Performs a single store operation.
 356          * This sets $params['op'] to 'store' and passes it to doOperation().
 357          *
 358          * @see FileBackend::doOperation()
 359          *
 360          * @param $params Array Operation parameters
 361          * @param $opts Array Operation options
 362          * @return Status
 363          */
 364         final public function store( array $params, array $opts = array() ) {
 365                 return $this->doOperation( array( 'op' => 'store' ) + $params, $opts );
 366         }
 367
 368         /**
 369          * Performs a single copy operation.
 370          * This sets $params['op'] to 'copy' and passes it to doOperation().
 371          *
 372          * @see FileBackend::doOperation()
 373          *
 374          * @param $params Array Operation parameters
 375          * @param $opts Array Operation options
 376          * @return Status
 377          */
 378         final public function copy( array $params, array $opts = array() ) {
 379                 return $this->doOperation( array( 'op' => 'copy' ) + $params, $opts );
 380         }
 381
 382         /**
 383          * Performs a single move operation.
 384          * This sets $params['op'] to 'move' and passes it to doOperation().
 385          *
 386          * @see FileBackend::doOperation()
 387          *
 388          * @param $params Array Operation parameters
 389          * @param $opts Array Operation options
 390          * @return Status
 391          */
 392         final public function move( array $params, array $opts = array() ) {
 393                 return $this->doOperation( array( 'op' => 'move' ) + $params, $opts );
 394         }
 395
 396         /**
 397          * Performs a single delete operation.
 398          * This sets $params['op'] to 'delete' and passes it to doOperation().
 399          *
 400          * @see FileBackend::doOperation()
 401          *
 402          * @param $params Array Operation parameters
 403          * @param $opts Array Operation options
 404          * @return Status
 405          */
 406         final public function delete( array $params, array $opts = array() ) {
 407                 return $this->doOperation( array( 'op' => 'delete' ) + $params, $opts );
 408         }
 409
 410         /**
 411          * Performs a single describe operation.
 412          * This sets $params['op'] to 'describe' and passes it to doOperation().
 413          *
 414          * @see FileBackend::doOperation()
 415          *
 416          * @param $params Array Operation parameters
 417          * @param $opts Array Operation options
 418          * @return Status
 419          * @since 1.21
 420          */
 421         final public function describe( array $params, array $opts = array() ) {
 422                 return $this->doOperation( array( 'op' => 'describe' ) + $params, $opts );
 423         }
 424
 425         /**
 426          * Perform a set of independent file operations on some files.
 427          *
 428          * This does no locking, nor journaling, and possibly no stat calls.
 429          * Any destination files that already exist will be overwritten.
 430          * This should *only* be used on non-original files, like cache files.
 431          *
 432          * Supported operations and their parameters:
 433          *  - create
 434          *  - store
 435          *  - copy
 436          *  - move
 437          *  - delete
 438          *  - describe (since 1.21)
 439          *  - null
 440          *
 441          * a) Create a new file in storage with the contents of a string
 442          * @code
 443          *     array(
 444          *         'op'                  => 'create',
 445          *         'dst'                 => <storage path>,
 446          *         'content'             => <string of new file contents>,
 447          *         'disposition'         => <Content-Disposition header value>,
 448          *         'headers'             => <HTTP header name/value map> # since 1.21
 449          *     )
 450          * @endcode
 451          *
 452          * b) Copy a file system file into storage
 453          * @code
 454          *     array(
 455          *         'op'                  => 'store',
 456          *         'src'                 => <file system path>,
 457          *         'dst'                 => <storage path>,
 458          *         'disposition'         => <Content-Disposition header value>,
 459          *         'headers'             => <HTTP header name/value map> # since 1.21
 460          *     )
 461          * @endcode
 462          *
 463          * c) Copy a file within storage
 464          * @code
 465          *     array(
 466          *         'op'                  => 'copy',
 467          *         'src'                 => <storage path>,
 468          *         'dst'                 => <storage path>,
 469          *         'ignoreMissingSource' => <boolean>, # since 1.21
 470          *         'disposition'         => <Content-Disposition header value>
 471          *     )
 472          * @endcode
 473          *
 474          * d) Move a file within storage
 475          * @code
 476          *     array(
 477          *         'op'                  => 'move',
 478          *         'src'                 => <storage path>,
 479          *         'dst'                 => <storage path>,
 480          *         'ignoreMissingSource' => <boolean>, # since 1.21
 481          *         'disposition'         => <Content-Disposition header value>
 482          *     )
 483          * @endcode
 484          *
 485          * e) Delete a file within storage
 486          * @code
 487          *     array(
 488          *         'op'                  => 'delete',
 489          *         'src'                 => <storage path>,
 490          *         'ignoreMissingSource' => <boolean>
 491          *     )
 492          * @endcode
 493          *
 494          * f) Update metadata for a file within storage
 495          * @code
 496          *     array(
 497          *         'op'                  => 'describe',
 498          *         'src'                 => <storage path>,
 499          *         'disposition'         => <Content-Disposition header value>,
 500          *         'headers'             => <HTTP header name/value map>
 501          *     )
 502          * @endcode
 503          *
 504          * g) Do nothing (no-op)
 505          * @code
 506          *     array(
 507          *         'op'                  => 'null',
 508          *     )
 509          * @endcode
 510          *
 511          * @par Boolean flags for operations (operation-specific):
 512          *   - ignoreMissingSource : The operation will simply succeed and do
 513          *                           nothing if the source file does not exist.
 514          *   - disposition         : When supplied, the backend will add a Content-Disposition
 515          *                           header when GETs/HEADs of the destination file are made.
 516          *                           Backends that don't support file metadata will ignore this.
 517          *                           See http://tools.ietf.org/html/rfc6266 (since 1.20).
 518          *   - headers             : If supplied with a header name/value map, the backend will
 519          *                           reply with these headers when GETs/HEADs of the destination
 520          *                           file are made. Header values should be smaller than 256 bytes.
 521          *                           Existing headers will remain, but these will replace any
 522          *                           conflicting previous headers, and headers will be removed
 523          *                           if they are set to an empty string.
 524          *                           Backends that don't support metadata ignore this. (since 1.21)
 525          *
 526          * $opts is an associative of boolean flags, including:
 527          *   - bypassReadOnly      : Allow writes in read-only mode (since 1.20)
 528          *
 529          * @par Return value:
 530          * This returns a Status, which contains all warnings and fatals that occurred
 531          * during the operation. The 'failCount', 'successCount', and 'success' members
 532          * will reflect each operation attempted for the given files. The status will be
 533          * considered "OK" as long as no fatal errors occurred.
 534          *
 535          * @param $ops Array Set of operations to execute
 536          * @param $opts Array Batch operation options
 537          * @return Status
 538          * @since 1.20
 539          */
 540         final public function doQuickOperations( array $ops, array $opts = array() ) {
 541                 if ( empty( $opts['bypassReadOnly'] ) && $this->isReadOnly() ) {
 542                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 543                 }
 544                 foreach ( $ops as &$op ) {
 545                         $op['overwrite'] = true; // avoids RTTs in key/value stores
 546                 }
 547                 return $this->doQuickOperationsInternal( $ops );
 548         }
 549
 550         /**
 551          * @see FileBackend::doQuickOperations()
 552          * @since 1.20
 553          */
 554         abstract protected function doQuickOperationsInternal( array $ops );
 555
 556         /**
 557          * Same as doQuickOperations() except it takes a single operation.
 558          * If you are doing a batch of operations, then use that function instead.
 559          *
 560          * @see FileBackend::doQuickOperations()
 561          *
 562          * @param $op Array Operation
 563          * @return Status
 564          * @since 1.20
 565          */
 566         final public function doQuickOperation( array $op ) {
 567                 return $this->doQuickOperations( array( $op ) );
 568         }
 569
 570         /**
 571          * Performs a single quick create operation.
 572          * This sets $params['op'] to 'create' and passes it to doQuickOperation().
 573          *
 574          * @see FileBackend::doQuickOperation()
 575          *
 576          * @param $params Array Operation parameters
 577          * @return Status
 578          * @since 1.20
 579          */
 580         final public function quickCreate( array $params ) {
 581                 return $this->doQuickOperation( array( 'op' => 'create' ) + $params );
 582         }
 583
 584         /**
 585          * Performs a single quick store operation.
 586          * This sets $params['op'] to 'store' and passes it to doQuickOperation().
 587          *
 588          * @see FileBackend::doQuickOperation()
 589          *
 590          * @param $params Array Operation parameters
 591          * @return Status
 592          * @since 1.20
 593          */
 594         final public function quickStore( array $params ) {
 595                 return $this->doQuickOperation( array( 'op' => 'store' ) + $params );
 596         }
 597
 598         /**
 599          * Performs a single quick copy operation.
 600          * This sets $params['op'] to 'copy' and passes it to doQuickOperation().
 601          *
 602          * @see FileBackend::doQuickOperation()
 603          *
 604          * @param $params Array Operation parameters
 605          * @return Status
 606          * @since 1.20
 607          */
 608         final public function quickCopy( array $params ) {
 609                 return $this->doQuickOperation( array( 'op' => 'copy' ) + $params );
 610         }
 611
 612         /**
 613          * Performs a single quick move operation.
 614          * This sets $params['op'] to 'move' and passes it to doQuickOperation().
 615          *
 616          * @see FileBackend::doQuickOperation()
 617          *
 618          * @param $params Array Operation parameters
 619          * @return Status
 620          * @since 1.20
 621          */
 622         final public function quickMove( array $params ) {
 623                 return $this->doQuickOperation( array( 'op' => 'move' ) + $params );
 624         }
 625
 626         /**
 627          * Performs a single quick delete operation.
 628          * This sets $params['op'] to 'delete' and passes it to doQuickOperation().
 629          *
 630          * @see FileBackend::doQuickOperation()
 631          *
 632          * @param $params Array Operation parameters
 633          * @return Status
 634          * @since 1.20
 635          */
 636         final public function quickDelete( array $params ) {
 637                 return $this->doQuickOperation( array( 'op' => 'delete' ) + $params );
 638         }
 639
 640         /**
 641          * Performs a single quick describe operation.
 642          * This sets $params['op'] to 'describe' and passes it to doQuickOperation().
 643          *
 644          * @see FileBackend::doQuickOperation()
 645          *
 646          * @param $params Array Operation parameters
 647          * @return Status
 648          * @since 1.21
 649          */
 650         final public function quickDescribe( array $params ) {
 651                 return $this->doQuickOperation( array( 'op' => 'describe' ) + $params );
 652         }
 653
 654         /**
 655          * Concatenate a list of storage files into a single file system file.
 656          * The target path should refer to a file that is already locked or
 657          * otherwise safe from modification from other processes. Normally,
 658          * the file will be a new temp file, which should be adequate.
 659          *
 660          * @param $params Array Operation parameters
 661          * $params include:
 662          *   - srcs        : ordered source storage paths (e.g. chunk1, chunk2, ...)
 663          *   - dst         : file system path to 0-byte temp file
 664          *   - parallelize : try to do operations in parallel when possible
 665          * @return Status
 666          */
 667         abstract public function concatenate( array $params );
 668
 669         /**
 670          * Prepare a storage directory for usage.
 671          * This will create any required containers and parent directories.
 672          * Backends using key/value stores only need to create the container.
 673          *
 674          * The 'noAccess' and 'noListing' parameters works the same as in secure(),
 675          * except they are only applied *if* the directory/container had to be created.
 676          * These flags should always be set for directories that have private files.
 677          *
 678          * @param $params Array
 679          * $params include:
 680          *   - dir            : storage directory
 681          *   - noAccess       : try to deny file access (since 1.20)
 682          *   - noListing      : try to deny file listing (since 1.20)
 683          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 684          * @return Status
 685          */
 686         final public function prepare( array $params ) {
 687                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 688                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 689                 }
 690                 return $this->doPrepare( $params );
 691         }
 692
 693         /**
 694          * @see FileBackend::prepare()
 695          */
 696         abstract protected function doPrepare( array $params );
 697
 698         /**
 699          * Take measures to block web access to a storage directory and
 700          * the container it belongs to. FS backends might add .htaccess
 701          * files whereas key/value store backends might revoke container
 702          * access to the storage user representing end-users in web requests.
 703          * This is not guaranteed to actually do anything.
 704          *
 705          * @param $params Array
 706          * $params include:
 707          *   - dir            : storage directory
 708          *   - noAccess       : try to deny file access
 709          *   - noListing      : try to deny file listing
 710          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 711          * @return Status
 712          */
 713         final public function secure( array $params ) {
 714                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 715                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 716                 }
 717                 return $this->doSecure( $params );
 718         }
 719
 720         /**
 721          * @see FileBackend::secure()
 722          */
 723         abstract protected function doSecure( array $params );
 724
 725         /**
 726          * Remove measures to block web access to a storage directory and
 727          * the container it belongs to. FS backends might remove .htaccess
 728          * files whereas key/value store backends might grant container
 729          * access to the storage user representing end-users in web requests.
 730          * This essentially can undo the result of secure() calls.
 731          *
 732          * @param $params Array
 733          * $params include:
 734          *   - dir            : storage directory
 735          *   - access         : try to allow file access
 736          *   - listing        : try to allow file listing
 737          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 738          * @return Status
 739          * @since 1.20
 740          */
 741         final public function publish( array $params ) {
 742                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 743                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 744                 }
 745                 return $this->doPublish( $params );
 746         }
 747
 748         /**
 749          * @see FileBackend::publish()
 750          */
 751         abstract protected function doPublish( array $params );
 752
 753         /**
 754          * Delete a storage directory if it is empty.
 755          * Backends using key/value stores may do nothing unless the directory
 756          * is that of an empty container, in which case it will be deleted.
 757          *
 758          * @param $params Array
 759          * $params include:
 760          *   - dir            : storage directory
 761          *   - recursive      : recursively delete empty subdirectories first (since 1.20)
 762          *   - bypassReadOnly : allow writes in read-only mode (since 1.20)
 763          * @return Status
 764          */
 765         final public function clean( array $params ) {
 766                 if ( empty( $params['bypassReadOnly'] ) && $this->isReadOnly() ) {
 767                         return Status::newFatal( 'backend-fail-readonly', $this->name, $this->readOnly );
 768                 }
 769                 return $this->doClean( $params );
 770         }
 771
 772         /**
 773          * @see FileBackend::clean()
 774          */
 775         abstract protected function doClean( array $params );
 776
 777         /**
 778          * Check if a file exists at a storage path in the backend.
 779          * This returns false if only a directory exists at the path.
 780          *
 781          * @param $params Array
 782          * $params include:
 783          *   - src    : source storage path
 784          *   - latest : use the latest available data
 785          * @return bool|null Returns null on failure
 786          */
 787         abstract public function fileExists( array $params );
 788
 789         /**
 790          * Get the last-modified timestamp of the file at a storage path.
 791          *
 792          * @param $params Array
 793          * $params include:
 794          *   - src    : source storage path
 795          *   - latest : use the latest available data
 796          * @return string|bool TS_MW timestamp or false on failure
 797          */
 798         abstract public function getFileTimestamp( array $params );
 799
 800         /**
 801          * Get the contents of a file at a storage path in the backend.
 802          * This should be avoided for potentially large files.
 803          *
 804          * @param $params Array
 805          * $params include:
 806          *   - src    : source storage path
 807          *   - latest : use the latest available data
 808          * @return string|bool Returns false on failure
 809          */
 810         final public function getFileContents( array $params ) {
 811                 $contents = $this->getFileContentsMulti(
 812                         array( 'srcs' => array( $params['src'] ) ) + $params );
 813
 814                 return $contents[$params['src']];
 815         }
 816
 817         /**
 818          * Like getFileContents() except it takes an array of storage paths
 819          * and returns a map of storage paths to strings (or null on failure).
 820          * The map keys (paths) are in the same order as the provided list of paths.
 821          *
 822          * @see FileBackend::getFileContents()
 823          *
 824          * @param $params Array
 825          * $params include:
 826          *   - srcs        : list of source storage paths
 827          *   - latest      : use the latest available data
 828          *   - parallelize : try to do operations in parallel when possible
 829          * @return Array Map of (path name => string or false on failure)
 830          * @since 1.20
 831          */
 832         abstract public function getFileContentsMulti( array $params );
 833
 834         /**
 835          * Get the size (bytes) of a file at a storage path in the backend.
 836          *
 837          * @param $params Array
 838          * $params include:
 839          *   - src    : source storage path
 840          *   - latest : use the latest available data
 841          * @return integer|bool Returns false on failure
 842          */
 843         abstract public function getFileSize( array $params );
 844
 845         /**
 846          * Get quick information about a file at a storage path in the backend.
 847          * If the file does not exist, then this returns false.
 848          * Otherwise, the result is an associative array that includes:
 849          *   - mtime  : the last-modified timestamp (TS_MW)
 850          *   - size   : the file size (bytes)
 851          * Additional values may be included for internal use only.
 852          *
 853          * @param $params Array
 854          * $params include:
 855          *   - src    : source storage path
 856          *   - latest : use the latest available data
 857          * @return Array|bool|null Returns null on failure
 858          */
 859         abstract public function getFileStat( array $params );
 860
 861         /**
 862          * Get a SHA-1 hash of the file at a storage path in the backend.
 863          *
 864          * @param $params Array
 865          * $params include:
 866          *   - src    : source storage path
 867          *   - latest : use the latest available data
 868          * @return string|bool Hash string or false on failure
 869          */
 870         abstract public function getFileSha1Base36( array $params );
 871
 872         /**
 873          * Get the properties of the file at a storage path in the backend.
 874          * This gives the result of FSFile::getProps() on a local copy of the file.
 875          *
 876          * @param $params Array
 877          * $params include:
 878          *   - src    : source storage path
 879          *   - latest : use the latest available data
 880          * @return Array Returns FSFile::placeholderProps() on failure
 881          */
 882         abstract public function getFileProps( array $params );
 883
 884         /**
 885          * Stream the file at a storage path in the backend.
 886          * If the file does not exists, an HTTP 404 error will be given.
 887          * Appropriate HTTP headers (Status, Content-Type, Content-Length)
 888          * will be sent if streaming began, while none will be sent otherwise.
 889          * Implementations should flush the output buffer before sending data.
 890          *
 891          * @param $params Array
 892          * $params include:
 893          *   - src     : source storage path
 894          *   - headers : list of additional HTTP headers to send on success
 895          *   - latest  : use the latest available data
 896          * @return Status
 897          */
 898         abstract public function streamFile( array $params );
 899
 900         /**
 901          * Returns a file system file, identical to the file at a storage path.
 902          * The file returned is either:
 903          *   - a) A local copy of the file at a storage path in the backend.
 904          *        The temporary copy will have the same extension as the source.
 905          *   - b) An original of the file at a storage path in the backend.
 906          * Temporary files may be purged when the file object falls out of scope.
 907          *
 908          * Write operations should *never* be done on this file as some backends
 909          * may do internal tracking or may be instances of FileBackendMultiWrite.
 910          * In that later case, there are copies of the file that must stay in sync.
 911          * Additionally, further calls to this function may return the same file.
 912          *
 913          * @param $params Array
 914          * $params include:
 915          *   - src    : source storage path
 916          *   - latest : use the latest available data
 917          * @return FSFile|null Returns null on failure
 918          */
 919         final public function getLocalReference( array $params ) {
 920                 $fsFiles = $this->getLocalReferenceMulti(
 921                         array( 'srcs' => array( $params['src'] ) ) + $params );
 922
 923                 return $fsFiles[$params['src']];
 924         }
 925
 926         /**
 927          * Like getLocalReference() except it takes an array of storage paths
 928          * and returns a map of storage paths to FSFile objects (or null on failure).
 929          * The map keys (paths) are in the same order as the provided list of paths.
 930          *
 931          * @see FileBackend::getLocalReference()
 932          *
 933          * @param $params Array
 934          * $params include:
 935          *   - srcs        : list of source storage paths
 936          *   - latest      : use the latest available data
 937          *   - parallelize : try to do operations in parallel when possible
 938          * @return Array Map of (path name => FSFile or null on failure)
 939          * @since 1.20
 940          */
 941         abstract public function getLocalReferenceMulti( array $params );
 942
 943         /**
 944          * Get a local copy on disk of the file at a storage path in the backend.
 945          * The temporary copy will have the same file extension as the source.
 946          * Temporary files may be purged when the file object falls out of scope.
 947          *
 948          * @param $params Array
 949          * $params include:
 950          *   - src    : source storage path
 951          *   - latest : use the latest available data
 952          * @return TempFSFile|null Returns null on failure
 953          */
 954         final public function getLocalCopy( array $params ) {
 955                 $tmpFiles = $this->getLocalCopyMulti(
 956                         array( 'srcs' => array( $params['src'] ) ) + $params );
 957
 958                 return $tmpFiles[$params['src']];
 959         }
 960
 961         /**
 962          * Like getLocalCopy() except it takes an array of storage paths and
 963          * returns a map of storage paths to TempFSFile objects (or null on failure).
 964          * The map keys (paths) are in the same order as the provided list of paths.
 965          *
 966          * @see FileBackend::getLocalCopy()
 967          *
 968          * @param $params Array
 969          * $params include:
 970          *   - srcs        : list of source storage paths
 971          *   - latest      : use the latest available data
 972          *   - parallelize : try to do operations in parallel when possible
 973          * @return Array Map of (path name => TempFSFile or null on failure)
 974          * @since 1.20
 975          */
 976         abstract public function getLocalCopyMulti( array $params );
 977
 978         /**
 979          * Return an HTTP URL to a given file that requires no authentication to use.
 980          * The URL may be pre-authenticated (via some token in the URL) and temporary.
 981          * This will return null if the backend cannot make an HTTP URL for the file.
 982          *
 983          * This is useful for key/value stores when using scripts that seek around
 984          * large files and those scripts (and the backend) support HTTP Range headers.
 985          * Otherwise, one would need to use getLocalReference(), which involves loading
 986          * the entire file on to local disk.
 987          *
 988          * @param $params Array
 989          * $params include:
 990          *   - src : source storage path
 991          * @return string|null
 992          * @since 1.21
 993          */
 994         abstract public function getFileHttpUrl( array $params );
 995
 996         /**
 997          * Check if a directory exists at a given storage path.
 998          * Backends using key/value stores will check if the path is a
 999          * virtual directory, meaning there are files under the given directory.
1000          *
1001          * Storage backends with eventual consistency might return stale data.
1002          *
1003          * @param $params array
1004          * $params include:
1005          *   - dir : storage directory
1006          * @return bool|null Returns null on failure
1007          * @since 1.20
1008          */
1009         abstract public function directoryExists( array $params );
1010
1011         /**
1012          * Get an iterator to list *all* directories under a storage directory.
1013          * If the directory is of the form "mwstore://backend/container",
1014          * then all directories in the container will be listed.
1015          * If the directory is of form "mwstore://backend/container/dir",
1016          * then all directories directly under that directory will be listed.
1017          * Results will be storage directories relative to the given directory.
1018          *
1019          * Storage backends with eventual consistency might return stale data.
1020          *
1021          * @param $params array
1022          * $params include:
1023          *   - dir     : storage directory
1024          *   - topOnly : only return direct child dirs of the directory
1025          * @return Traversable|Array|null Returns null on failure
1026          * @since 1.20
1027          */
1028         abstract public function getDirectoryList( array $params );
1029
1030         /**
1031          * Same as FileBackend::getDirectoryList() except only lists
1032          * directories that are immediately under the given directory.
1033          *
1034          * Storage backends with eventual consistency might return stale data.
1035          *
1036          * @param $params array
1037          * $params include:
1038          *   - dir : storage directory
1039          * @return Traversable|Array|null Returns null on failure
1040          * @since 1.20
1041          */
1042         final public function getTopDirectoryList( array $params ) {
1043                 return $this->getDirectoryList( array( 'topOnly' => true ) + $params );
1044         }
1045
1046         /**
1047          * Get an iterator to list *all* stored files under a storage directory.
1048          * If the directory is of the form "mwstore://backend/container",
1049          * then all files in the container will be listed.
1050          * If the directory is of form "mwstore://backend/container/dir",
1051          * then all files under that directory will be listed.
1052          * Results will be storage paths relative to the given directory.
1053          *
1054          * Storage backends with eventual consistency might return stale data.
1055          *
1056          * @param $params array
1057          * $params include:
1058          *   - dir     : storage directory
1059          *   - topOnly : only return direct child files of the directory (since 1.20)
1060          * @return Traversable|Array|null Returns null on failure
1061          */
1062         abstract public function getFileList( array $params );
1063
1064         /**
1065          * Same as FileBackend::getFileList() except only lists
1066          * files that are immediately under the given directory.
1067          *
1068          * Storage backends with eventual consistency might return stale data.
1069          *
1070          * @param $params array
1071          * $params include:
1072          *   - dir : storage directory
1073          * @return Traversable|Array|null Returns null on failure
1074          * @since 1.20
1075          */
1076         final public function getTopFileList( array $params ) {
1077                 return $this->getFileList( array( 'topOnly' => true ) + $params );
1078         }
1079
1080         /**
1081          * Preload persistent file stat and property cache into in-process cache.
1082          * This should be used when stat calls will be made on a known list of a many files.
1083          *
1084          * @param $paths Array Storage paths
1085          * @return void
1086          */
1087         public function preloadCache( array $paths ) {}
1088
1089         /**
1090          * Invalidate any in-process file stat and property cache.
1091          * If $paths is given, then only the cache for those files will be cleared.
1092          *
1093          * @param $paths Array Storage paths (optional)
1094          * @return void
1095          */
1096         public function clearCache( array $paths = null ) {}
1097
1098         /**
1099          * Lock the files at the given storage paths in the backend.
1100          * This will either lock all the files or none (on failure).
1101          *
1102          * Callers should consider using getScopedFileLocks() instead.
1103          *
1104          * @param $paths Array Storage paths
1105          * @param $type integer LockManager::LOCK_* constant
1106          * @return Status
1107          */
1108         final public function lockFiles( array $paths, $type ) {
1109                 return $this->lockManager->lock( $paths, $type );
1110         }
1111
1112         /**
1113          * Unlock the files at the given storage paths in the backend.
1114          *
1115          * @param $paths Array Storage paths
1116          * @param $type integer LockManager::LOCK_* constant
1117          * @return Status
1118          */
1119         final public function unlockFiles( array $paths, $type ) {
1120                 return $this->lockManager->unlock( $paths, $type );
1121         }
1122
1123         /**
1124          * Lock the files at the given storage paths in the backend.
1125          * This will either lock all the files or none (on failure).
1126          * On failure, the status object will be updated with errors.
1127          *
1128          * Once the return value goes out scope, the locks will be released and
1129          * the status updated. Unlock fatals will not change the status "OK" value.
1130          *
1131          * @param $paths Array Storage paths
1132          * @param $type integer LockManager::LOCK_* constant
1133          * @param $status Status Status to update on lock/unlock
1134          * @return ScopedLock|null Returns null on failure
1135          */
1136         final public function getScopedFileLocks( array $paths, $type, Status $status ) {
1137                 return ScopedLock::factory( $this->lockManager, $paths, $type, $status );
1138         }
1139
1140         /**
1141          * Get an array of scoped locks needed for a batch of file operations.
1142          *
1143          * Normally, FileBackend::doOperations() handles locking, unless
1144          * the 'nonLocking' param is passed in. This function is useful if you
1145          * want the files to be locked for a broader scope than just when the
1146          * files are changing. For example, if you need to update DB metadata,
1147          * you may want to keep the files locked until finished.
1148          *
1149          * @see FileBackend::doOperations()
1150          *
1151          * @param $ops Array List of file operations to FileBackend::doOperations()
1152          * @param $status Status Status to update on lock/unlock
1153          * @return Array List of ScopedFileLocks or null values
1154          * @since 1.20
1155          */
1156         abstract public function getScopedLocksForOps( array $ops, Status $status );
1157
1158         /**
1159          * Get the root storage path of this backend.
1160          * All container paths are "subdirectories" of this path.
1161          *
1162          * @return string Storage path
1163          * @since 1.20
1164          */
1165         final public function getRootStoragePath() {
1166                 return "mwstore://{$this->name}";
1167         }
1168
1169         /**
1170          * Get the storage path for the given container for this backend
1171          *
1172          * @param $container string Container name
1173          * @return string Storage path
1174          * @since 1.21
1175          */
1176         final public function getContainerStoragePath( $container ) {
1177                 return $this->getRootStoragePath() . "/{$container}";
1178         }
1179
1180         /**
1181          * Get the file journal object for this backend
1182          *
1183          * @return FileJournal
1184          */
1185         final public function getJournal() {
1186                 return $this->fileJournal;
1187         }
1188
1189         /**
1190          * Check if a given path is a "mwstore://" path.
1191          * This does not do any further validation or any existence checks.
1192          *
1193          * @param $path string
1194          * @return bool
1195          */
1196         final public static function isStoragePath( $path ) {
1197                 return ( strpos( $path, 'mwstore://' ) === 0 );
1198         }
1199
1200         /**
1201          * Split a storage path into a backend name, a container name,
1202          * and a relative file path. The relative path may be the empty string.
1203          * This does not do any path normalization or traversal checks.
1204          *
1205          * @param $storagePath string
1206          * @return Array (backend, container, rel object) or (null, null, null)
1207          */
1208         final public static function splitStoragePath( $storagePath ) {
1209                 if ( self::isStoragePath( $storagePath ) ) {
1210                         // Remove the "mwstore://" prefix and split the path
1211                         $parts = explode( '/', substr( $storagePath, 10 ), 3 );
1212                         if ( count( $parts ) >= 2 && $parts[0] != '' && $parts[1] != '' ) {
1213                                 if ( count( $parts ) == 3 ) {
1214                                         return $parts; // e.g. "backend/container/path"
1215                                 } else {
1216                                         return array( $parts[0], $parts[1], '' ); // e.g. "backend/container"
1217                                 }
1218                         }
1219                 }
1220                 return array( null, null, null );
1221         }
1222
1223         /**
1224          * Normalize a storage path by cleaning up directory separators.
1225          * Returns null if the path is not of the format of a valid storage path.
1226          *
1227          * @param $storagePath string
1228          * @return string|null
1229          */
1230         final public static function normalizeStoragePath( $storagePath ) {
1231                 list( $backend, $container, $relPath ) = self::splitStoragePath( $storagePath );
1232                 if ( $relPath !== null ) { // must be for this backend
1233                         $relPath = self::normalizeContainerPath( $relPath );
1234                         if ( $relPath !== null ) {
1235                                 return ( $relPath != '' )
1236                                         ? "mwstore://{$backend}/{$container}/{$relPath}"
1237                                         : "mwstore://{$backend}/{$container}";
1238                         }
1239                 }
1240                 return null;
1241         }
1242
1243         /**
1244          * Get the parent storage directory of a storage path.
1245          * This returns a path like "mwstore://backend/container",
1246          * "mwstore://backend/container/...", or null if there is no parent.
1247          *
1248          * @param $storagePath string
1249          * @return string|null
1250          */
1251         final public static function parentStoragePath( $storagePath ) {
1252                 $storagePath = dirname( $storagePath );
1253                 list( $b, $cont, $rel ) = self::splitStoragePath( $storagePath );
1254                 return ( $rel === null ) ? null : $storagePath;
1255         }
1256
1257         /**
1258          * Get the final extension from a storage or FS path
1259          *
1260          * @param $path string
1261          * @return string
1262          */
1263         final public static function extensionFromPath( $path ) {
1264                 $i = strrpos( $path, '.' );
1265                 return strtolower( $i ? substr( $path, $i + 1 ) : '' );
1266         }
1267
1268         /**
1269          * Check if a relative path has no directory traversals
1270          *
1271          * @param $path string
1272          * @return bool
1273          * @since 1.20
1274          */
1275         final public static function isPathTraversalFree( $path ) {
1276                 return ( self::normalizeContainerPath( $path ) !== null );
1277         }
1278
1279         /**
1280          * Build a Content-Disposition header value per RFC 6266.
1281          *
1282          * @param $type string One of (attachment, inline)
1283          * @param $filename string Suggested file name (should not contain slashes)
1284          * @throws MWException
1285          * @return string
1286          * @since 1.20
1287          */
1288         final public static function makeContentDisposition( $type, $filename = '' ) {
1289                 $parts = array();
1290
1291                 $type = strtolower( $type );
1292                 if ( !in_array( $type, array( 'inline', 'attachment' ) ) ) {
1293                         throw new MWException( "Invalid Content-Disposition type '$type'." );
1294                 }
1295                 $parts[] = $type;
1296
1297                 if ( strlen( $filename ) ) {
1298                         $parts[] = "filename*=UTF-8''" . rawurlencode( basename( $filename ) );
1299                 }
1300
1301                 return implode( ';', $parts );
1302         }
1303
1304         /**
1305          * Validate and normalize a relative storage path.
1306          * Null is returned if the path involves directory traversal.
1307          * Traversal is insecure for FS backends and broken for others.
1308          *
1309          * This uses the same traversal protection as Title::secureAndSplit().
1310          *
1311          * @param $path string Storage path relative to a container
1312          * @return string|null
1313          */
1314         final protected static function normalizeContainerPath( $path ) {
1315                 // Normalize directory separators
1316                 $path = strtr( $path, '\\', '/' );
1317                 // Collapse any consecutive directory separators
1318                 $path = preg_replace( '![/]{2,}!', '/', $path );
1319                 // Remove any leading directory separator
1320                 $path = ltrim( $path, '/' );
1321                 // Use the same traversal protection as Title::secureAndSplit()
1322                 if ( strpos( $path, '.' ) !== false ) {
1323                         if (
1324                                 $path === '.' ||
1325                                 $path === '..' ||
1326                                 strpos( $path, './' ) === 0 ||
1327                                 strpos( $path, '../' ) === 0 ||
1328                                 strpos( $path, '/./' ) !== false ||
1329                                 strpos( $path, '/../' ) !== false
1330                         ) {
1331                                 return null;
1332                         }
1333                 }
1334                 return $path;
1335         }
1336 }