3 * Service for storing and loading data blobs representing revision content.
5 * This program is free software; you can redistribute it and/or modify
6 * it under the terms of the GNU General Public License as published by
7 * the Free Software Foundation; either version 2 of the License, or
8 * (at your option) any later version.
10 * This program is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 * GNU General Public License for more details.
15 * You should have received a copy of the GNU General Public License along
16 * with this program; if not, write to the Free Software Foundation, Inc.,
17 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
18 * http://www.gnu.org/copyleft/gpl.html
20 * Attribution notice: when this file was created, much of its content was taken
21 * from the Revision.php file as present in release 1.30. Refer to the history
22 * of that file for original authorship.
27 namespace MediaWiki\Storage
;
29 use DBAccessObjectUtils
;
33 use InvalidArgumentException
;
37 use Wikimedia\Assert\Assert
;
38 use Wikimedia\Rdbms\Database
;
39 use Wikimedia\Rdbms\IDatabase
;
40 use Wikimedia\Rdbms\LoadBalancer
;
43 * Service for storing and loading Content objects.
47 * @note This was written to act as a drop-in replacement for the corresponding
48 * static methods in Revision.
50 class SqlBlobStore
implements IDBAccessObject
, BlobStore
{
52 // Note: the name has been taken unchanged from the Revision class.
53 const TEXT_CACHE_GROUP
= 'revisiontext:10';
58 private $dbLoadBalancer;
66 * @var bool|string Wiki ID
73 private $cacheExpiry = 604800; // 7 days
78 private $compressBlobs = false;
83 private $legacyEncoding = false;
88 private $legacyEncodingConversionLang = null;
93 private $useExternalStore = false;
96 * @param LoadBalancer $dbLoadBalancer A load balancer for acquiring database connections
97 * @param WANObjectCache $cache A cache manager for caching blobs
98 * @param bool|string $wikiId The ID of the target wiki database. Use false for the local wiki.
100 public function __construct(
101 LoadBalancer
$dbLoadBalancer,
102 WANObjectCache
$cache,
105 $this->dbLoadBalancer
= $dbLoadBalancer;
106 $this->cache
= $cache;
107 $this->wikiId
= $wikiId;
111 * @return int time for which blobs can be cached, in seconds
113 public function getCacheExpiry() {
114 return $this->cacheExpiry
;
118 * @param int $cacheExpiry time for which blobs can be cached, in seconds
120 public function setCacheExpiry( $cacheExpiry ) {
121 Assert
::parameterType( 'integer', $cacheExpiry, '$cacheExpiry' );
123 $this->cacheExpiry
= $cacheExpiry;
127 * @return bool whether blobs should be compressed for storage
129 public function getCompressBlobs() {
130 return $this->compressBlobs
;
134 * @param bool $compressBlobs whether blobs should be compressed for storage
136 public function setCompressBlobs( $compressBlobs ) {
137 $this->compressBlobs
= $compressBlobs;
141 * @return false|string The legacy encoding to assume for blobs that are not marked as utf8.
142 * False means handling of legacy encoding is disabled, and utf8 assumed.
144 public function getLegacyEncoding() {
145 return $this->legacyEncoding
;
149 * @return Language|null The locale to use when decoding from a legacy encoding, or null
150 * if handling of legacy encoding is disabled.
152 public function getLegacyEncodingConversionLang() {
153 return $this->legacyEncodingConversionLang
;
157 * @param string $legacyEncoding The legacy encoding to assume for blobs that are
158 * not marked as utf8.
159 * @param Language $language The locale to use when decoding from a legacy encoding.
161 public function setLegacyEncoding( $legacyEncoding, Language
$language ) {
162 Assert
::parameterType( 'string', $legacyEncoding, '$legacyEncoding' );
164 $this->legacyEncoding
= $legacyEncoding;
165 $this->legacyEncodingConversionLang
= $language;
169 * @return bool Whether to use the ExternalStore mechanism for storing blobs.
171 public function getUseExternalStore() {
172 return $this->useExternalStore
;
176 * @param bool $useExternalStore Whether to use the ExternalStore mechanism for storing blobs.
178 public function setUseExternalStore( $useExternalStore ) {
179 Assert
::parameterType( 'boolean', $useExternalStore, '$useExternalStore' );
181 $this->useExternalStore
= $useExternalStore;
185 * @return LoadBalancer
187 private function getDBLoadBalancer() {
188 return $this->dbLoadBalancer
;
192 * @param int $index A database index, like DB_MASTER or DB_REPLICA
196 private function getDBConnection( $index ) {
197 $lb = $this->getDBLoadBalancer();
198 return $lb->getConnection( $index, [], $this->wikiId
);
202 * Stores an arbitrary blob of data and returns an address that can be used with
203 * getBlob() to retrieve the same blob of data,
205 * @param string $data
206 * @param array $hints An array of hints.
208 * @throws BlobAccessException
209 * @return string an address that can be used with getBlob() to retrieve the data.
211 public function storeBlob( $data, $hints = [] ) {
213 $flags = $this->compressData( $data );
215 # Write to external storage if required
216 if ( $this->useExternalStore
) {
217 // Store and get the URL
218 $data = ExternalStore
::insertToDefault( $data );
220 throw new BlobAccessException( "Failed to store text to external storage" );
225 $flags .= 'external';
227 // TODO: we could also return an address for the external store directly here.
228 // That would mean bypassing the text table entirely when the external store is
229 // used. We'll need to assess expected fallout before doing that.
232 $dbw = $this->getDBConnection( DB_MASTER
);
234 $old_id = $dbw->nextSequenceValue( 'text_old_id_seq' );
240 'old_flags' => $flags,
245 $textId = $dbw->insertId();
247 return 'tt:' . $textId;
248 } catch ( MWException
$e ) {
249 throw new BlobAccessException( $e->getMessage(), 0, $e );
254 * Retrieve a blob, given an address.
255 * Currently hardcoded to the 'text' table storage engine.
257 * MCR migration note: this replaces Revision::loadText
259 * @param string $blobAddress
260 * @param int $queryFlags
262 * @throws BlobAccessException
265 public function getBlob( $blobAddress, $queryFlags = 0 ) {
266 Assert
::parameterType( 'string', $blobAddress, '$blobAddress' );
268 // No negative caching; negative hits on text rows may be due to corrupted replica DBs
269 $blob = $this->cache
->getWithSetCallback(
270 // TODO: change key, since this is not necessarily revision text!
271 $this->cache
->makeKey( 'revisiontext', 'textid', $blobAddress ),
272 $this->getCacheTTL(),
273 function ( $unused, &$ttl, &$setOpts ) use ( $blobAddress, $queryFlags ) {
274 list( $index ) = DBAccessObjectUtils
::getDBOptions( $queryFlags );
275 $setOpts +
= Database
::getCacheSetOptions( $this->getDBConnection( $index ) );
277 return $this->fetchBlob( $blobAddress, $queryFlags );
279 [ 'pcGroup' => self
::TEXT_CACHE_GROUP
, 'pcTTL' => IExpiringStore
::TTL_PROC_LONG
]
282 if ( $blob === false ) {
283 throw new BlobAccessException( 'Failed to load blob from address ' . $blobAddress );
290 * MCR migration note: this corresponds to Revision::fetchText
292 * @param string $blobAddress
293 * @param int $queryFlags
295 * @throw BlobAccessException
296 * @return string|false
298 private function fetchBlob( $blobAddress, $queryFlags ) {
299 list( $schema, $id, ) = self
::splitBlobAddress( $blobAddress );
301 //TODO: MCR: also support 'ex' schema with ExternalStore URLs, plus flags encoded in the URL!
302 //TODO: MCR: also support 'ar' schema for content blobs in old style archive rows!
303 if ( $schema === 'tt' ) {
304 $textId = intval( $id );
306 // XXX: change to better exceptions! That makes migration more difficult, though.
307 throw new BlobAccessException( "Unknown blob address schema: $schema" );
310 if ( !$textId ||
$id !== (string)$textId ) {
311 // XXX: change to better exceptions! That makes migration more difficult, though.
312 throw new BlobAccessException( "Bad blob address: $blobAddress" );
315 // Callers doing updates will pass in READ_LATEST as usual. Since the text/blob tables
316 // do not normally get rows changed around, set READ_LATEST_IMMUTABLE in those cases.
317 $queryFlags |
= DBAccessObjectUtils
::hasFlags( $queryFlags, self
::READ_LATEST
)
318 ? self
::READ_LATEST_IMMUTABLE
321 list( $index, $options, $fallbackIndex, $fallbackOptions ) =
322 DBAccessObjectUtils
::getDBOptions( $queryFlags );
324 // Text data is immutable; check replica DBs first.
325 $row = $this->getDBConnection( $index )->selectRow(
327 [ 'old_text', 'old_flags' ],
328 [ 'old_id' => $textId ],
333 // Fallback to DB_MASTER in some cases if the row was not found, using the appropriate
334 // options, such as FOR UPDATE to avoid missing rows due to REPEATABLE-READ.
335 if ( !$row && $fallbackIndex !== null ) {
336 $row = $this->getDBConnection( $fallbackIndex )->selectRow(
338 [ 'old_text', 'old_flags' ],
339 [ 'old_id' => $textId ],
346 wfWarn( __METHOD__
. ": No text row with ID $textId." );
350 $blob = $this->expandBlob( $row->old_text
, $row->old_flags
, $blobAddress );
352 if ( $blob === false ) {
353 wfWarn( __METHOD__
. ": Bad data in text row $textId." );
361 * Expand a raw data blob according to the flags given.
363 * MCR migration note: this replaces Revision::getRevisionText
365 * @note direct use is deprecated, use getBlob() or SlotRecord::getContent() instead.
366 * @todo make this private, there should be no need to use this method outside this class.
368 * @param string $raw The raw blob data, to be processed according to $flags.
369 * May be the blob itself, or the blob compressed, or just the address
370 * of the actual blob, depending on $flags.
371 * @param string|string[] $flags Blob flags, such as 'external' or 'gzip'.
372 * Note that not including 'utf-8' in $flags will cause the data to be decoded
373 * according to the legacy encoding specified via setLegacyEncoding.
374 * @param string|null $cacheKey May be used for caching if given
376 * @return false|string The expanded blob or false on failure
378 public function expandBlob( $raw, $flags, $cacheKey = null ) {
379 if ( is_string( $flags ) ) {
380 $flags = explode( ',', $flags );
383 // Use external methods for external objects, text in table is URL-only then
384 if ( in_array( 'external', $flags ) ) {
386 $parts = explode( '://', $url, 2 );
387 if ( count( $parts ) == 1 ||
$parts[1] == '' ) {
391 if ( $cacheKey && $this->wikiId
=== false ) {
392 // Make use of the wiki-local revision text cache.
393 // The cached value should be decompressed, so handle that and return here.
394 // NOTE: we rely on $this->cache being the right cache for $this->wikiId!
395 return $this->cache
->getWithSetCallback(
396 // TODO: change key, since this is not necessarily revision text!
397 $this->cache
->makeKey( 'revisiontext', 'textid', $cacheKey ),
398 $this->getCacheTTL(),
399 function () use ( $url, $flags ) {
400 // No negative caching per BlobStore::getBlob()
401 $blob = ExternalStore
::fetchFromURL( $url, [ 'wiki' => $this->wikiId
] );
403 return $this->decompressData( $blob, $flags );
405 [ 'pcGroup' => self
::TEXT_CACHE_GROUP
, 'pcTTL' => WANObjectCache
::TTL_PROC_LONG
]
408 $blob = ExternalStore
::fetchFromURL( $url, [ 'wiki' => $this->wikiId
] );
409 return $this->decompressData( $blob, $flags );
412 return $this->decompressData( $raw, $flags );
417 * If $wgCompressRevisions is enabled, we will compress data.
418 * The input string is modified in place.
419 * Return value is the flags field: contains 'gzip' if the
420 * data is compressed, and 'utf-8' if we're saving in UTF-8
423 * MCR migration note: this replaces Revision::compressRevisionText
425 * @note direct use is deprecated!
426 * @todo make this private, there should be no need to use this method outside this class.
428 * @param mixed &$blob Reference to a text
432 public function compressData( &$blob ) {
435 // Revisions not marked as UTF-8 will have legacy decoding applied by decompressData().
436 // XXX: if $this->legacyEncoding is not set, we could skip this. That would however be
437 // risky, since $this->legacyEncoding being set in the future would lead to data corruption.
438 $blobFlags[] = 'utf-8';
440 if ( $this->compressBlobs
) {
441 if ( function_exists( 'gzdeflate' ) ) {
442 $deflated = gzdeflate( $blob );
444 if ( $deflated === false ) {
445 wfLogWarning( __METHOD__
. ': gzdeflate() failed' );
448 $blobFlags[] = 'gzip';
451 wfDebug( __METHOD__
. " -- no zlib support, not compressing\n" );
454 return implode( ',', $blobFlags );
458 * Re-converts revision text according to its flags.
460 * MCR migration note: this replaces Revision::decompressRevisionText
462 * @note direct use is deprecated, use getBlob() or SlotRecord::getContent() instead.
463 * @todo make this private, there should be no need to use this method outside this class.
465 * @param mixed $blob Reference to a text
466 * @param array $blobFlags Compression flags, such as 'gzip'.
467 * Note that not including 'utf-8' in $blobFlags will cause the data to be decoded
468 * according to the legacy encoding specified via setLegacyEncoding.
470 * @return string|bool Decompressed text, or false on failure
472 public function decompressData( $blob, array $blobFlags ) {
473 if ( $blob === false ) {
474 // Text failed to be fetched; nothing to do
478 if ( in_array( 'error', $blobFlags ) ) {
479 // Error row, return false
483 if ( in_array( 'gzip', $blobFlags ) ) {
484 # Deal with optional compression of archived pages.
485 # This can be done periodically via maintenance/compressOld.php, and
486 # as pages are saved if $wgCompressRevisions is set.
487 $blob = gzinflate( $blob );
489 if ( $blob === false ) {
490 wfLogWarning( __METHOD__
. ': gzinflate() failed' );
495 if ( in_array( 'object', $blobFlags ) ) {
496 # Generic compressed storage
497 $obj = unserialize( $blob );
498 if ( !is_object( $obj ) ) {
502 $blob = $obj->getText();
505 // Needed to support old revisions left over from from the 1.4 / 1.5 migration.
506 if ( $blob !== false && $this->legacyEncoding
&& $this->legacyEncodingConversionLang
507 && !in_array( 'utf-8', $blobFlags ) && !in_array( 'utf8', $blobFlags )
509 # Old revisions kept around in a legacy encoding?
510 # Upconvert on demand.
511 # ("utf8" checked for compatibility with some broken
512 # conversion scripts 2008-12-30)
513 $blob = $this->legacyEncodingConversionLang
->iconv( $this->legacyEncoding
, 'UTF-8', $blob );
520 * Get the text cache TTL
522 * MCR migration note: this replaces Revision::getCacheTTL
526 private function getCacheTTL() {
527 if ( $this->cache
->getQoS( WANObjectCache
::ATTR_EMULATION
)
528 <= WANObjectCache
::QOS_EMULATION_SQL
530 // Do not cache RDBMs blobs in...the RDBMs store
531 $ttl = WANObjectCache
::TTL_UNCACHEABLE
;
533 $ttl = $this->cacheExpiry ?
: WANObjectCache
::TTL_UNCACHEABLE
;
540 * Returns an ID corresponding to the old_id field in the text table, corresponding
541 * to the given $address.
543 * Currently, $address must start with 'tt:' followed by a decimal integer representing
544 * the old_id; if $address does not start with 'tt:', null is returned. However,
545 * the implementation may change to insert rows into the text table on the fly.
547 * @note This method exists for use with the text table based storage schema.
548 * It should not be assumed that is will function with all future kinds of content addresses.
550 * @deprecated since 1.31, so not assume that all blob addresses refer to a row in the text
551 * table. This method should become private once the relevant refactoring in WikiPage is
554 * @param string $address
558 public function getTextIdFromAddress( $address ) {
559 list( $schema, $id, ) = self
::splitBlobAddress( $address );
561 if ( $schema !== 'tt' ) {
565 $textId = intval( $id );
567 if ( !$textId ||
$id !== (string)$textId ) {
568 throw new InvalidArgumentException( "Malformed text_id: $id" );
575 * Splits a blob address into three parts: the schema, the ID, and parameters/flags.
577 * @param string $address
579 * @throws InvalidArgumentException
580 * @return array [ $schema, $id, $parameters ], with $parameters being an assoc array.
582 private static function splitBlobAddress( $address ) {
583 if ( !preg_match( '/^(\w+):(\w+)(\?(.*))?$/', $address, $m ) ) {
584 throw new InvalidArgumentException( "Bad blob address: $address" );
587 $schema = strtolower( $m[1] );
589 $parameters = isset( $m[4] ) ?
wfCgiToArray( $m[4] ) : [];
591 return [ $schema, $id, $parameters ];