Merge "Properly mark UserGetImplicitGroups hook as deprecated"
[lhc/web/wiklou.git] / includes / cache / LocalisationCache.php
1 <?php
2 /**
3 * Cache of the contents of localisation files.
4 *
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.
9 *
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.
14 *
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
19 *
20 * @file
21 */
22
23 use Cdb\Exception as CdbException;
24 use Cdb\Reader as CdbReader;
25 use Cdb\Writer as CdbWriter;
26 /**
27 * Class for caching the contents of localisation files, Messages*.php
28 * and *.i18n.php.
29 *
30 * An instance of this class is available using Language::getLocalisationCache().
31 *
32 * The values retrieved from here are merged, containing items from extension
33 * files, core messages files and the language fallback sequence (e.g. zh-cn ->
34 * zh-hans -> en ). Some common errors are corrected, for example namespace
35 * names with spaces instead of underscores, but heavyweight processing, such
36 * as grammatical transformation, is done by the caller.
37 */
38 class LocalisationCache {
39 const VERSION = 2;
40
41 /** Configuration associative array */
42 private $conf;
43
44 /**
45 * True if recaching should only be done on an explicit call to recache().
46 * Setting this reduces the overhead of cache freshness checking, which
47 * requires doing a stat() for every extension i18n file.
48 */
49 private $manualRecache = false;
50
51 /**
52 * True to treat all files as expired until they are regenerated by this object.
53 */
54 private $forceRecache = false;
55
56 /**
57 * The cache data. 3-d array, where the first key is the language code,
58 * the second key is the item key e.g. 'messages', and the third key is
59 * an item specific subkey index. Some items are not arrays and so for those
60 * items, there are no subkeys.
61 */
62 protected $data = array();
63
64 /**
65 * The persistent store object. An instance of LCStore.
66 *
67 * @var LCStore
68 */
69 private $store;
70
71 /**
72 * A 2-d associative array, code/key, where presence indicates that the item
73 * is loaded. Value arbitrary.
74 *
75 * For split items, if set, this indicates that all of the subitems have been
76 * loaded.
77 */
78 private $loadedItems = array();
79
80 /**
81 * A 3-d associative array, code/key/subkey, where presence indicates that
82 * the subitem is loaded. Only used for the split items, i.e. messages.
83 */
84 private $loadedSubitems = array();
85
86 /**
87 * An array where presence of a key indicates that that language has been
88 * initialised. Initialisation includes checking for cache expiry and doing
89 * any necessary updates.
90 */
91 private $initialisedLangs = array();
92
93 /**
94 * An array mapping non-existent pseudo-languages to fallback languages. This
95 * is filled by initShallowFallback() when data is requested from a language
96 * that lacks a Messages*.php file.
97 */
98 private $shallowFallbacks = array();
99
100 /**
101 * An array where the keys are codes that have been recached by this instance.
102 */
103 private $recachedLangs = array();
104
105 /**
106 * All item keys
107 */
108 static public $allKeys = array(
109 'fallback', 'namespaceNames', 'bookstoreList',
110 'magicWords', 'messages', 'rtl', 'capitalizeAllNouns', 'digitTransformTable',
111 'separatorTransformTable', 'fallback8bitEncoding', 'linkPrefixExtension',
112 'linkTrail', 'linkPrefixCharset', 'namespaceAliases',
113 'dateFormats', 'datePreferences', 'datePreferenceMigrationMap',
114 'defaultDateFormat', 'extraUserToggles', 'specialPageAliases',
115 'imageFiles', 'preloadedMessages', 'namespaceGenderAliases',
116 'digitGroupingPattern', 'pluralRules', 'pluralRuleTypes', 'compiledPluralRules',
117 );
118
119 /**
120 * Keys for items which consist of associative arrays, which may be merged
121 * by a fallback sequence.
122 */
123 static public $mergeableMapKeys = array( 'messages', 'namespaceNames',
124 'dateFormats', 'imageFiles', 'preloadedMessages'
125 );
126
127 /**
128 * Keys for items which are a numbered array.
129 */
130 static public $mergeableListKeys = array( 'extraUserToggles' );
131
132 /**
133 * Keys for items which contain an array of arrays of equivalent aliases
134 * for each subitem. The aliases may be merged by a fallback sequence.
135 */
136 static public $mergeableAliasListKeys = array( 'specialPageAliases' );
137
138 /**
139 * Keys for items which contain an associative array, and may be merged if
140 * the primary value contains the special array key "inherit". That array
141 * key is removed after the first merge.
142 */
143 static public $optionalMergeKeys = array( 'bookstoreList' );
144
145 /**
146 * Keys for items that are formatted like $magicWords
147 */
148 static public $magicWordKeys = array( 'magicWords' );
149
150 /**
151 * Keys for items where the subitems are stored in the backend separately.
152 */
153 static public $splitKeys = array( 'messages' );
154
155 /**
156 * Keys which are loaded automatically by initLanguage()
157 */
158 static public $preloadedKeys = array( 'dateFormats', 'namespaceNames' );
159
160 /**
161 * Associative array of cached plural rules. The key is the language code,
162 * the value is an array of plural rules for that language.
163 */
164 private $pluralRules = null;
165
166 /**
167 * Associative array of cached plural rule types. The key is the language
168 * code, the value is an array of plural rule types for that language. For
169 * example, $pluralRuleTypes['ar'] = ['zero', 'one', 'two', 'few', 'many'].
170 * The index for each rule type matches the index for the rule in
171 * $pluralRules, thus allowing correlation between the two. The reason we
172 * don't just use the type names as the keys in $pluralRules is because
173 * Language::convertPlural applies the rules based on numeric order (or
174 * explicit numeric parameter), not based on the name of the rule type. For
175 * example, {{plural:count|wordform1|wordform2|wordform3}}, rather than
176 * {{plural:count|one=wordform1|two=wordform2|many=wordform3}}.
177 */
178 private $pluralRuleTypes = null;
179
180 private $mergeableKeys = null;
181
182 /**
183 * Constructor.
184 * For constructor parameters, see the documentation in DefaultSettings.php
185 * for $wgLocalisationCacheConf.
186 *
187 * @param array $conf
188 * @throws MWException
189 */
190 function __construct( $conf ) {
191 global $wgCacheDirectory;
192
193 $this->conf = $conf;
194 $storeConf = array();
195 if ( !empty( $conf['storeClass'] ) ) {
196 $storeClass = $conf['storeClass'];
197 } else {
198 switch ( $conf['store'] ) {
199 case 'files':
200 case 'file':
201 $storeClass = 'LCStoreCDB';
202 break;
203 case 'db':
204 $storeClass = 'LCStoreDB';
205 break;
206 case 'detect':
207 $storeClass = $wgCacheDirectory ? 'LCStoreCDB' : 'LCStoreDB';
208 break;
209 default:
210 throw new MWException(
211 'Please set $wgLocalisationCacheConf[\'store\'] to something sensible.' );
212 }
213 }
214
215 wfDebugLog( 'caches', get_class( $this ) . ": using store $storeClass" );
216 if ( !empty( $conf['storeDirectory'] ) ) {
217 $storeConf['directory'] = $conf['storeDirectory'];
218 }
219
220 $this->store = new $storeClass( $storeConf );
221 foreach ( array( 'manualRecache', 'forceRecache' ) as $var ) {
222 if ( isset( $conf[$var] ) ) {
223 $this->$var = $conf[$var];
224 }
225 }
226 }
227
228 /**
229 * Returns true if the given key is mergeable, that is, if it is an associative
230 * array which can be merged through a fallback sequence.
231 * @param string $key
232 * @return bool
233 */
234 public function isMergeableKey( $key ) {
235 if ( $this->mergeableKeys === null ) {
236 $this->mergeableKeys = array_flip( array_merge(
237 self::$mergeableMapKeys,
238 self::$mergeableListKeys,
239 self::$mergeableAliasListKeys,
240 self::$optionalMergeKeys,
241 self::$magicWordKeys
242 ) );
243 }
244
245 return isset( $this->mergeableKeys[$key] );
246 }
247
248 /**
249 * Get a cache item.
250 *
251 * Warning: this may be slow for split items (messages), since it will
252 * need to fetch all of the subitems from the cache individually.
253 * @param string $code
254 * @param string $key
255 * @return mixed
256 */
257 public function getItem( $code, $key ) {
258 if ( !isset( $this->loadedItems[$code][$key] ) ) {
259 wfProfileIn( __METHOD__ . '-load' );
260 $this->loadItem( $code, $key );
261 wfProfileOut( __METHOD__ . '-load' );
262 }
263
264 if ( $key === 'fallback' && isset( $this->shallowFallbacks[$code] ) ) {
265 return $this->shallowFallbacks[$code];
266 }
267
268 return $this->data[$code][$key];
269 }
270
271 /**
272 * Get a subitem, for instance a single message for a given language.
273 * @param string $code
274 * @param string $key
275 * @param string $subkey
276 * @return mixed|null
277 */
278 public function getSubitem( $code, $key, $subkey ) {
279 if ( !isset( $this->loadedSubitems[$code][$key][$subkey] ) &&
280 !isset( $this->loadedItems[$code][$key] )
281 ) {
282 wfProfileIn( __METHOD__ . '-load' );
283 $this->loadSubitem( $code, $key, $subkey );
284 wfProfileOut( __METHOD__ . '-load' );
285 }
286
287 if ( isset( $this->data[$code][$key][$subkey] ) ) {
288 return $this->data[$code][$key][$subkey];
289 } else {
290 return null;
291 }
292 }
293
294 /**
295 * Get the list of subitem keys for a given item.
296 *
297 * This is faster than array_keys($lc->getItem(...)) for the items listed in
298 * self::$splitKeys.
299 *
300 * Will return null if the item is not found, or false if the item is not an
301 * array.
302 * @param string $code
303 * @param string $key
304 * @return bool|null|string
305 */
306 public function getSubitemList( $code, $key ) {
307 if ( in_array( $key, self::$splitKeys ) ) {
308 return $this->getSubitem( $code, 'list', $key );
309 } else {
310 $item = $this->getItem( $code, $key );
311 if ( is_array( $item ) ) {
312 return array_keys( $item );
313 } else {
314 return false;
315 }
316 }
317 }
318
319 /**
320 * Load an item into the cache.
321 * @param string $code
322 * @param string $key
323 */
324 protected function loadItem( $code, $key ) {
325 if ( !isset( $this->initialisedLangs[$code] ) ) {
326 $this->initLanguage( $code );
327 }
328
329 // Check to see if initLanguage() loaded it for us
330 if ( isset( $this->loadedItems[$code][$key] ) ) {
331 return;
332 }
333
334 if ( isset( $this->shallowFallbacks[$code] ) ) {
335 $this->loadItem( $this->shallowFallbacks[$code], $key );
336
337 return;
338 }
339
340 if ( in_array( $key, self::$splitKeys ) ) {
341 $subkeyList = $this->getSubitem( $code, 'list', $key );
342 foreach ( $subkeyList as $subkey ) {
343 if ( isset( $this->data[$code][$key][$subkey] ) ) {
344 continue;
345 }
346 $this->data[$code][$key][$subkey] = $this->getSubitem( $code, $key, $subkey );
347 }
348 } else {
349 $this->data[$code][$key] = $this->store->get( $code, $key );
350 }
351
352 $this->loadedItems[$code][$key] = true;
353 }
354
355 /**
356 * Load a subitem into the cache
357 * @param string $code
358 * @param string $key
359 * @param string $subkey
360 */
361 protected function loadSubitem( $code, $key, $subkey ) {
362 if ( !in_array( $key, self::$splitKeys ) ) {
363 $this->loadItem( $code, $key );
364
365 return;
366 }
367
368 if ( !isset( $this->initialisedLangs[$code] ) ) {
369 $this->initLanguage( $code );
370 }
371
372 // Check to see if initLanguage() loaded it for us
373 if ( isset( $this->loadedItems[$code][$key] ) ||
374 isset( $this->loadedSubitems[$code][$key][$subkey] )
375 ) {
376 return;
377 }
378
379 if ( isset( $this->shallowFallbacks[$code] ) ) {
380 $this->loadSubitem( $this->shallowFallbacks[$code], $key, $subkey );
381
382 return;
383 }
384
385 $value = $this->store->get( $code, "$key:$subkey" );
386 $this->data[$code][$key][$subkey] = $value;
387 $this->loadedSubitems[$code][$key][$subkey] = true;
388 }
389
390 /**
391 * Returns true if the cache identified by $code is missing or expired.
392 *
393 * @param string $code
394 *
395 * @return bool
396 */
397 public function isExpired( $code ) {
398 if ( $this->forceRecache && !isset( $this->recachedLangs[$code] ) ) {
399 wfDebug( __METHOD__ . "($code): forced reload\n" );
400
401 return true;
402 }
403
404 $deps = $this->store->get( $code, 'deps' );
405 $keys = $this->store->get( $code, 'list' );
406 $preload = $this->store->get( $code, 'preload' );
407 // Different keys may expire separately for some stores
408 if ( $deps === null || $keys === null || $preload === null ) {
409 wfDebug( __METHOD__ . "($code): cache missing, need to make one\n" );
410
411 return true;
412 }
413
414 foreach ( $deps as $dep ) {
415 // Because we're unserializing stuff from cache, we
416 // could receive objects of classes that don't exist
417 // anymore (e.g. uninstalled extensions)
418 // When this happens, always expire the cache
419 if ( !$dep instanceof CacheDependency || $dep->isExpired() ) {
420 wfDebug( __METHOD__ . "($code): cache for $code expired due to " .
421 get_class( $dep ) . "\n" );
422
423 return true;
424 }
425 }
426
427 return false;
428 }
429
430 /**
431 * Initialise a language in this object. Rebuild the cache if necessary.
432 * @param string $code
433 * @throws MWException
434 */
435 protected function initLanguage( $code ) {
436 if ( isset( $this->initialisedLangs[$code] ) ) {
437 return;
438 }
439
440 $this->initialisedLangs[$code] = true;
441
442 # If the code is of the wrong form for a Messages*.php file, do a shallow fallback
443 if ( !Language::isValidBuiltInCode( $code ) ) {
444 $this->initShallowFallback( $code, 'en' );
445
446 return;
447 }
448
449 # Recache the data if necessary
450 if ( !$this->manualRecache && $this->isExpired( $code ) ) {
451 if ( Language::isSupportedLanguage( $code ) ) {
452 $this->recache( $code );
453 } elseif ( $code === 'en' ) {
454 throw new MWException( 'MessagesEn.php is missing.' );
455 } else {
456 $this->initShallowFallback( $code, 'en' );
457 }
458
459 return;
460 }
461
462 # Preload some stuff
463 $preload = $this->getItem( $code, 'preload' );
464 if ( $preload === null ) {
465 if ( $this->manualRecache ) {
466 // No Messages*.php file. Do shallow fallback to en.
467 if ( $code === 'en' ) {
468 throw new MWException( 'No localisation cache found for English. ' .
469 'Please run maintenance/rebuildLocalisationCache.php.' );
470 }
471 $this->initShallowFallback( $code, 'en' );
472
473 return;
474 } else {
475 throw new MWException( 'Invalid or missing localisation cache.' );
476 }
477 }
478 $this->data[$code] = $preload;
479 foreach ( $preload as $key => $item ) {
480 if ( in_array( $key, self::$splitKeys ) ) {
481 foreach ( $item as $subkey => $subitem ) {
482 $this->loadedSubitems[$code][$key][$subkey] = true;
483 }
484 } else {
485 $this->loadedItems[$code][$key] = true;
486 }
487 }
488 }
489
490 /**
491 * Create a fallback from one language to another, without creating a
492 * complete persistent cache.
493 * @param string $primaryCode
494 * @param string $fallbackCode
495 */
496 public function initShallowFallback( $primaryCode, $fallbackCode ) {
497 $this->data[$primaryCode] =& $this->data[$fallbackCode];
498 $this->loadedItems[$primaryCode] =& $this->loadedItems[$fallbackCode];
499 $this->loadedSubitems[$primaryCode] =& $this->loadedSubitems[$fallbackCode];
500 $this->shallowFallbacks[$primaryCode] = $fallbackCode;
501 }
502
503 /**
504 * Read a PHP file containing localisation data.
505 * @param string $_fileName
506 * @param string $_fileType
507 * @throws MWException
508 * @return array
509 */
510 protected function readPHPFile( $_fileName, $_fileType ) {
511 wfProfileIn( __METHOD__ );
512 // Disable APC caching
513 wfSuppressWarnings();
514 $_apcEnabled = ini_set( 'apc.cache_by_default', '0' );
515 wfRestoreWarnings();
516
517 include $_fileName;
518
519 wfSuppressWarnings();
520 ini_set( 'apc.cache_by_default', $_apcEnabled );
521 wfRestoreWarnings();
522
523 if ( $_fileType == 'core' || $_fileType == 'extension' ) {
524 $data = compact( self::$allKeys );
525 } elseif ( $_fileType == 'aliases' ) {
526 $data = compact( 'aliases' );
527 } else {
528 wfProfileOut( __METHOD__ );
529 throw new MWException( __METHOD__ . ": Invalid file type: $_fileType" );
530 }
531 wfProfileOut( __METHOD__ );
532
533 return $data;
534 }
535
536 /**
537 * Read a JSON file containing localisation messages.
538 * @param string $fileName Name of file to read
539 * @throws MWException If there is a syntax error in the JSON file
540 * @return array Array with a 'messages' key, or empty array if the file doesn't exist
541 */
542 public function readJSONFile( $fileName ) {
543 wfProfileIn( __METHOD__ );
544
545 if ( !is_readable( $fileName ) ) {
546 wfProfileOut( __METHOD__ );
547
548 return array();
549 }
550
551 $json = file_get_contents( $fileName );
552 if ( $json === false ) {
553 wfProfileOut( __METHOD__ );
554
555 return array();
556 }
557
558 $data = FormatJson::decode( $json, true );
559 if ( $data === null ) {
560 wfProfileOut( __METHOD__ );
561
562 throw new MWException( __METHOD__ . ": Invalid JSON file: $fileName" );
563 }
564
565 // Remove keys starting with '@', they're reserved for metadata and non-message data
566 foreach ( $data as $key => $unused ) {
567 if ( $key === '' || $key[0] === '@' ) {
568 unset( $data[$key] );
569 }
570 }
571
572 wfProfileOut( __METHOD__ );
573
574 // The JSON format only supports messages, none of the other variables, so wrap the data
575 return array( 'messages' => $data );
576 }
577
578 /**
579 * Get the compiled plural rules for a given language from the XML files.
580 * @since 1.20
581 * @param string $code
582 * @return array|null
583 */
584 public function getCompiledPluralRules( $code ) {
585 $rules = $this->getPluralRules( $code );
586 if ( $rules === null ) {
587 return null;
588 }
589 try {
590 $compiledRules = CLDRPluralRuleEvaluator::compile( $rules );
591 } catch ( CLDRPluralRuleError $e ) {
592 wfDebugLog( 'l10n', $e->getMessage() );
593
594 return array();
595 }
596
597 return $compiledRules;
598 }
599
600 /**
601 * Get the plural rules for a given language from the XML files.
602 * Cached.
603 * @since 1.20
604 * @param string $code
605 * @return array|null
606 */
607 public function getPluralRules( $code ) {
608 if ( $this->pluralRules === null ) {
609 $this->loadPluralFiles();
610 }
611 if ( !isset( $this->pluralRules[$code] ) ) {
612 return null;
613 } else {
614 return $this->pluralRules[$code];
615 }
616 }
617
618 /**
619 * Get the plural rule types for a given language from the XML files.
620 * Cached.
621 * @since 1.22
622 * @param string $code
623 * @return array|null
624 */
625 public function getPluralRuleTypes( $code ) {
626 if ( $this->pluralRuleTypes === null ) {
627 $this->loadPluralFiles();
628 }
629 if ( !isset( $this->pluralRuleTypes[$code] ) ) {
630 return null;
631 } else {
632 return $this->pluralRuleTypes[$code];
633 }
634 }
635
636 /**
637 * Load the plural XML files.
638 */
639 protected function loadPluralFiles() {
640 global $IP;
641 $cldrPlural = "$IP/languages/data/plurals.xml";
642 $mwPlural = "$IP/languages/data/plurals-mediawiki.xml";
643 // Load CLDR plural rules
644 $this->loadPluralFile( $cldrPlural );
645 if ( file_exists( $mwPlural ) ) {
646 // Override or extend
647 $this->loadPluralFile( $mwPlural );
648 }
649 }
650
651 /**
652 * Load a plural XML file with the given filename, compile the relevant
653 * rules, and save the compiled rules in a process-local cache.
654 *
655 * @param string $fileName
656 */
657 protected function loadPluralFile( $fileName ) {
658 $doc = new DOMDocument;
659 $doc->load( $fileName );
660 $rulesets = $doc->getElementsByTagName( "pluralRules" );
661 foreach ( $rulesets as $ruleset ) {
662 $codes = $ruleset->getAttribute( 'locales' );
663 $rules = array();
664 $ruleTypes = array();
665 $ruleElements = $ruleset->getElementsByTagName( "pluralRule" );
666 foreach ( $ruleElements as $elt ) {
667 $ruleType = $elt->getAttribute( 'count' );
668 if ( $ruleType === 'other' ) {
669 // Don't record "other" rules, which have an empty condition
670 continue;
671 }
672 $rules[] = $elt->nodeValue;
673 $ruleTypes[] = $ruleType;
674 }
675 foreach ( explode( ' ', $codes ) as $code ) {
676 $this->pluralRules[$code] = $rules;
677 $this->pluralRuleTypes[$code] = $ruleTypes;
678 }
679 }
680 }
681
682 /**
683 * Read the data from the source files for a given language, and register
684 * the relevant dependencies in the $deps array. If the localisation
685 * exists, the data array is returned, otherwise false is returned.
686 *
687 * @param string $code
688 * @param array $deps
689 * @return array
690 */
691 protected function readSourceFilesAndRegisterDeps( $code, &$deps ) {
692 global $IP;
693 wfProfileIn( __METHOD__ );
694
695 // This reads in the PHP i18n file with non-messages l10n data
696 $fileName = Language::getMessagesFileName( $code );
697 if ( !file_exists( $fileName ) ) {
698 $data = array();
699 } else {
700 $deps[] = new FileDependency( $fileName );
701 $data = $this->readPHPFile( $fileName, 'core' );
702 }
703
704 # Load CLDR plural rules for JavaScript
705 $data['pluralRules'] = $this->getPluralRules( $code );
706 # And for PHP
707 $data['compiledPluralRules'] = $this->getCompiledPluralRules( $code );
708 # Load plural rule types
709 $data['pluralRuleTypes'] = $this->getPluralRuleTypes( $code );
710
711 $deps['plurals'] = new FileDependency( "$IP/languages/data/plurals.xml" );
712 $deps['plurals-mw'] = new FileDependency( "$IP/languages/data/plurals-mediawiki.xml" );
713
714 wfProfileOut( __METHOD__ );
715
716 return $data;
717 }
718
719 /**
720 * Merge two localisation values, a primary and a fallback, overwriting the
721 * primary value in place.
722 * @param string $key
723 * @param mixed $value
724 * @param mixed $fallbackValue
725 */
726 protected function mergeItem( $key, &$value, $fallbackValue ) {
727 if ( !is_null( $value ) ) {
728 if ( !is_null( $fallbackValue ) ) {
729 if ( in_array( $key, self::$mergeableMapKeys ) ) {
730 $value = $value + $fallbackValue;
731 } elseif ( in_array( $key, self::$mergeableListKeys ) ) {
732 $value = array_unique( array_merge( $fallbackValue, $value ) );
733 } elseif ( in_array( $key, self::$mergeableAliasListKeys ) ) {
734 $value = array_merge_recursive( $value, $fallbackValue );
735 } elseif ( in_array( $key, self::$optionalMergeKeys ) ) {
736 if ( !empty( $value['inherit'] ) ) {
737 $value = array_merge( $fallbackValue, $value );
738 }
739
740 if ( isset( $value['inherit'] ) ) {
741 unset( $value['inherit'] );
742 }
743 } elseif ( in_array( $key, self::$magicWordKeys ) ) {
744 $this->mergeMagicWords( $value, $fallbackValue );
745 }
746 }
747 } else {
748 $value = $fallbackValue;
749 }
750 }
751
752 /**
753 * @param mixed $value
754 * @param mixed $fallbackValue
755 */
756 protected function mergeMagicWords( &$value, $fallbackValue ) {
757 foreach ( $fallbackValue as $magicName => $fallbackInfo ) {
758 if ( !isset( $value[$magicName] ) ) {
759 $value[$magicName] = $fallbackInfo;
760 } else {
761 $oldSynonyms = array_slice( $fallbackInfo, 1 );
762 $newSynonyms = array_slice( $value[$magicName], 1 );
763 $synonyms = array_values( array_unique( array_merge(
764 $newSynonyms, $oldSynonyms ) ) );
765 $value[$magicName] = array_merge( array( $fallbackInfo[0] ), $synonyms );
766 }
767 }
768 }
769
770 /**
771 * Given an array mapping language code to localisation value, such as is
772 * found in extension *.i18n.php files, iterate through a fallback sequence
773 * to merge the given data with an existing primary value.
774 *
775 * Returns true if any data from the extension array was used, false
776 * otherwise.
777 * @param array $codeSequence
778 * @param string $key
779 * @param mixed $value
780 * @param mixed $fallbackValue
781 * @return bool
782 */
783 protected function mergeExtensionItem( $codeSequence, $key, &$value, $fallbackValue ) {
784 $used = false;
785 foreach ( $codeSequence as $code ) {
786 if ( isset( $fallbackValue[$code] ) ) {
787 $this->mergeItem( $key, $value, $fallbackValue[$code] );
788 $used = true;
789 }
790 }
791
792 return $used;
793 }
794
795 /**
796 * Load localisation data for a given language for both core and extensions
797 * and save it to the persistent cache store and the process cache
798 * @param string $code
799 * @throws MWException
800 */
801 public function recache( $code ) {
802 global $wgExtensionMessagesFiles, $wgMessagesDirs;
803 wfProfileIn( __METHOD__ );
804
805 if ( !$code ) {
806 wfProfileOut( __METHOD__ );
807 throw new MWException( "Invalid language code requested" );
808 }
809 $this->recachedLangs[$code] = true;
810
811 # Initial values
812 $initialData = array_combine(
813 self::$allKeys,
814 array_fill( 0, count( self::$allKeys ), null ) );
815 $coreData = $initialData;
816 $deps = array();
817
818 # Load the primary localisation from the source file
819 $data = $this->readSourceFilesAndRegisterDeps( $code, $deps );
820 if ( $data === false ) {
821 wfDebug( __METHOD__ . ": no localisation file for $code, using fallback to en\n" );
822 $coreData['fallback'] = 'en';
823 } else {
824 wfDebug( __METHOD__ . ": got localisation for $code from source\n" );
825
826 # Merge primary localisation
827 foreach ( $data as $key => $value ) {
828 $this->mergeItem( $key, $coreData[$key], $value );
829 }
830 }
831
832 # Fill in the fallback if it's not there already
833 if ( is_null( $coreData['fallback'] ) ) {
834 $coreData['fallback'] = $code === 'en' ? false : 'en';
835 }
836 if ( $coreData['fallback'] === false ) {
837 $coreData['fallbackSequence'] = array();
838 } else {
839 $coreData['fallbackSequence'] = array_map( 'trim', explode( ',', $coreData['fallback'] ) );
840 $len = count( $coreData['fallbackSequence'] );
841
842 # Ensure that the sequence ends at en
843 if ( $coreData['fallbackSequence'][$len - 1] !== 'en' ) {
844 $coreData['fallbackSequence'][] = 'en';
845 }
846 }
847
848 $codeSequence = array_merge( array( $code ), $coreData['fallbackSequence'] );
849
850 wfProfileIn( __METHOD__ . '-fallbacks' );
851
852 # Load non-JSON localisation data for extensions
853 $extensionData = array_combine(
854 $codeSequence,
855 array_fill( 0, count( $codeSequence ), $initialData ) );
856 foreach ( $wgExtensionMessagesFiles as $extension => $fileName ) {
857 if ( isset( $wgMessagesDirs[$extension] ) ) {
858 # This extension has JSON message data; skip the PHP shim
859 continue;
860 }
861
862 $data = $this->readPHPFile( $fileName, 'extension' );
863 $used = false;
864
865 foreach ( $data as $key => $item ) {
866 foreach ( $codeSequence as $csCode ) {
867 if ( isset( $item[$csCode] ) ) {
868 $this->mergeItem( $key, $extensionData[$csCode][$key], $item[$csCode] );
869 $used = true;
870 }
871 }
872 }
873
874 if ( $used ) {
875 $deps[] = new FileDependency( $fileName );
876 }
877 }
878
879 # Load the localisation data for each fallback, then merge it into the full array
880 $allData = $initialData;
881 foreach ( $codeSequence as $csCode ) {
882 $csData = $initialData;
883
884 # Load core messages and the extension localisations.
885 foreach ( $wgMessagesDirs as $dirs ) {
886 foreach ( (array)$dirs as $dir ) {
887 $fileName = "$dir/$csCode.json";
888 $data = $this->readJSONFile( $fileName );
889
890 foreach ( $data as $key => $item ) {
891 $this->mergeItem( $key, $csData[$key], $item );
892 }
893
894 $deps[] = new FileDependency( $fileName );
895 }
896 }
897
898 # Merge non-JSON extension data
899 if ( isset( $extensionData[$csCode] ) ) {
900 foreach ( $extensionData[$csCode] as $key => $item ) {
901 $this->mergeItem( $key, $csData[$key], $item );
902 }
903 }
904
905 if ( $csCode === $code ) {
906 # Merge core data into extension data
907 foreach ( $coreData as $key => $item ) {
908 $this->mergeItem( $key, $csData[$key], $item );
909 }
910 } else {
911 # Load the secondary localisation from the source file to
912 # avoid infinite cycles on cyclic fallbacks
913 $fbData = $this->readSourceFilesAndRegisterDeps( $csCode, $deps );
914 if ( $fbData !== false ) {
915 # Only merge the keys that make sense to merge
916 foreach ( self::$allKeys as $key ) {
917 if ( !isset( $fbData[$key] ) ) {
918 continue;
919 }
920
921 if ( is_null( $coreData[$key] ) || $this->isMergeableKey( $key ) ) {
922 $this->mergeItem( $key, $csData[$key], $fbData[$key] );
923 }
924 }
925 }
926 }
927
928 # Allow extensions an opportunity to adjust the data for this
929 # fallback
930 wfRunHooks( 'LocalisationCacheRecacheFallback', array( $this, $csCode, &$csData ) );
931
932 # Merge the data for this fallback into the final array
933 if ( $csCode === $code ) {
934 $allData = $csData;
935 } else {
936 foreach ( self::$allKeys as $key ) {
937 if ( !isset( $csData[$key] ) ) {
938 continue;
939 }
940
941 if ( is_null( $allData[$key] ) || $this->isMergeableKey( $key ) ) {
942 $this->mergeItem( $key, $allData[$key], $csData[$key] );
943 }
944 }
945 }
946 }
947
948 wfProfileOut( __METHOD__ . '-fallbacks' );
949
950 # Add cache dependencies for any referenced globals
951 $deps['wgExtensionMessagesFiles'] = new GlobalDependency( 'wgExtensionMessagesFiles' );
952 $deps['wgMessagesDirs'] = new GlobalDependency( 'wgMessagesDirs' );
953 $deps['version'] = new ConstantDependency( 'LocalisationCache::VERSION' );
954
955 # Add dependencies to the cache entry
956 $allData['deps'] = $deps;
957
958 # Replace spaces with underscores in namespace names
959 $allData['namespaceNames'] = str_replace( ' ', '_', $allData['namespaceNames'] );
960
961 # And do the same for special page aliases. $page is an array.
962 foreach ( $allData['specialPageAliases'] as &$page ) {
963 $page = str_replace( ' ', '_', $page );
964 }
965 # Decouple the reference to prevent accidental damage
966 unset( $page );
967
968 # If there were no plural rules, return an empty array
969 if ( $allData['pluralRules'] === null ) {
970 $allData['pluralRules'] = array();
971 }
972 if ( $allData['compiledPluralRules'] === null ) {
973 $allData['compiledPluralRules'] = array();
974 }
975 # If there were no plural rule types, return an empty array
976 if ( $allData['pluralRuleTypes'] === null ) {
977 $allData['pluralRuleTypes'] = array();
978 }
979
980 # Set the list keys
981 $allData['list'] = array();
982 foreach ( self::$splitKeys as $key ) {
983 $allData['list'][$key] = array_keys( $allData[$key] );
984 }
985 # Run hooks
986 $purgeBlobs = true;
987 wfRunHooks( 'LocalisationCacheRecache', array( $this, $code, &$allData, &$purgeBlobs ) );
988
989 if ( is_null( $allData['namespaceNames'] ) ) {
990 wfProfileOut( __METHOD__ );
991 throw new MWException( __METHOD__ . ': Localisation data failed sanity check! ' .
992 'Check that your languages/messages/MessagesEn.php file is intact.' );
993 }
994
995 # Set the preload key
996 $allData['preload'] = $this->buildPreload( $allData );
997
998 # Save to the process cache and register the items loaded
999 $this->data[$code] = $allData;
1000 foreach ( $allData as $key => $item ) {
1001 $this->loadedItems[$code][$key] = true;
1002 }
1003
1004 # Save to the persistent cache
1005 wfProfileIn( __METHOD__ . '-write' );
1006 $this->store->startWrite( $code );
1007 foreach ( $allData as $key => $value ) {
1008 if ( in_array( $key, self::$splitKeys ) ) {
1009 foreach ( $value as $subkey => $subvalue ) {
1010 $this->store->set( "$key:$subkey", $subvalue );
1011 }
1012 } else {
1013 $this->store->set( $key, $value );
1014 }
1015 }
1016 $this->store->finishWrite();
1017 wfProfileOut( __METHOD__ . '-write' );
1018
1019 # Clear out the MessageBlobStore
1020 # HACK: If using a null (i.e. disabled) storage backend, we
1021 # can't write to the MessageBlobStore either
1022 if ( $purgeBlobs && !$this->store instanceof LCStoreNull ) {
1023 MessageBlobStore::getInstance()->clear();
1024 }
1025
1026 wfProfileOut( __METHOD__ );
1027 }
1028
1029 /**
1030 * Build the preload item from the given pre-cache data.
1031 *
1032 * The preload item will be loaded automatically, improving performance
1033 * for the commonly-requested items it contains.
1034 * @param array $data
1035 * @return array
1036 */
1037 protected function buildPreload( $data ) {
1038 $preload = array( 'messages' => array() );
1039 foreach ( self::$preloadedKeys as $key ) {
1040 $preload[$key] = $data[$key];
1041 }
1042
1043 foreach ( $data['preloadedMessages'] as $subkey ) {
1044 if ( isset( $data['messages'][$subkey] ) ) {
1045 $subitem = $data['messages'][$subkey];
1046 } else {
1047 $subitem = null;
1048 }
1049 $preload['messages'][$subkey] = $subitem;
1050 }
1051
1052 return $preload;
1053 }
1054
1055 /**
1056 * Unload the data for a given language from the object cache.
1057 * Reduces memory usage.
1058 * @param string $code
1059 */
1060 public function unload( $code ) {
1061 unset( $this->data[$code] );
1062 unset( $this->loadedItems[$code] );
1063 unset( $this->loadedSubitems[$code] );
1064 unset( $this->initialisedLangs[$code] );
1065 unset( $this->shallowFallbacks[$code] );
1066
1067 foreach ( $this->shallowFallbacks as $shallowCode => $fbCode ) {
1068 if ( $fbCode === $code ) {
1069 $this->unload( $shallowCode );
1070 }
1071 }
1072 }
1073
1074 /**
1075 * Unload all data
1076 */
1077 public function unloadAll() {
1078 foreach ( $this->initialisedLangs as $lang => $unused ) {
1079 $this->unload( $lang );
1080 }
1081 }
1082
1083 /**
1084 * Disable the storage backend
1085 */
1086 public function disableBackend() {
1087 $this->store = new LCStoreNull;
1088 $this->manualRecache = false;
1089 }
1090 }
1091
1092 /**
1093 * Interface for the persistence layer of LocalisationCache.
1094 *
1095 * The persistence layer is two-level hierarchical cache. The first level
1096 * is the language, the second level is the item or subitem.
1097 *
1098 * Since the data for a whole language is rebuilt in one operation, it needs
1099 * to have a fast and atomic method for deleting or replacing all of the
1100 * current data for a given language. The interface reflects this bulk update
1101 * operation. Callers writing to the cache must first call startWrite(), then
1102 * will call set() a couple of thousand times, then will call finishWrite()
1103 * to commit the operation. When finishWrite() is called, the cache is
1104 * expected to delete all data previously stored for that language.
1105 *
1106 * The values stored are PHP variables suitable for serialize(). Implementations
1107 * of LCStore are responsible for serializing and unserializing.
1108 */
1109 interface LCStore {
1110 /**
1111 * Get a value.
1112 * @param string $code Language code
1113 * @param string $key Cache key
1114 */
1115 function get( $code, $key );
1116
1117 /**
1118 * Start a write transaction.
1119 * @param string $code Language code
1120 */
1121 function startWrite( $code );
1122
1123 /**
1124 * Finish a write transaction.
1125 */
1126 function finishWrite();
1127
1128 /**
1129 * Set a key to a given value. startWrite() must be called before this
1130 * is called, and finishWrite() must be called afterwards.
1131 * @param string $key
1132 * @param mixed $value
1133 */
1134 function set( $key, $value );
1135 }
1136
1137 /**
1138 * LCStore implementation which uses the standard DB functions to store data.
1139 * This will work on any MediaWiki installation.
1140 */
1141 class LCStoreDB implements LCStore {
1142 private $currentLang;
1143 private $writesDone = false;
1144
1145 /** @var DatabaseBase */
1146 private $dbw;
1147 /** @var array */
1148 private $batch = array();
1149
1150 private $readOnly = false;
1151
1152 public function get( $code, $key ) {
1153 if ( $this->writesDone ) {
1154 $db = wfGetDB( DB_MASTER );
1155 } else {
1156 $db = wfGetDB( DB_SLAVE );
1157 }
1158 $row = $db->selectRow( 'l10n_cache', array( 'lc_value' ),
1159 array( 'lc_lang' => $code, 'lc_key' => $key ), __METHOD__ );
1160 if ( $row ) {
1161 return unserialize( $db->decodeBlob( $row->lc_value ) );
1162 } else {
1163 return null;
1164 }
1165 }
1166
1167 public function startWrite( $code ) {
1168 if ( $this->readOnly ) {
1169 return;
1170 } elseif ( !$code ) {
1171 throw new MWException( __METHOD__ . ": Invalid language \"$code\"" );
1172 }
1173
1174 $this->dbw = wfGetDB( DB_MASTER );
1175
1176 $this->currentLang = $code;
1177 $this->batch = array();
1178 }
1179
1180 public function finishWrite() {
1181 if ( $this->readOnly ) {
1182 return;
1183 } elseif ( is_null( $this->currentLang ) ) {
1184 throw new MWException( __CLASS__ . ': must call startWrite() before finishWrite()' );
1185 }
1186
1187 $this->dbw->begin( __METHOD__ );
1188 try {
1189 $this->dbw->delete( 'l10n_cache',
1190 array( 'lc_lang' => $this->currentLang ), __METHOD__ );
1191 foreach ( array_chunk( $this->batch, 500 ) as $rows ) {
1192 $this->dbw->insert( 'l10n_cache', $rows, __METHOD__ );
1193 }
1194 $this->writesDone = true;
1195 } catch ( DBQueryError $e ) {
1196 if ( $this->dbw->wasReadOnlyError() ) {
1197 $this->readOnly = true; // just avoid site down time
1198 } else {
1199 throw $e;
1200 }
1201 }
1202 $this->dbw->commit( __METHOD__ );
1203
1204 $this->currentLang = null;
1205 $this->batch = array();
1206 }
1207
1208 public function set( $key, $value ) {
1209 if ( $this->readOnly ) {
1210 return;
1211 } elseif ( is_null( $this->currentLang ) ) {
1212 throw new MWException( __CLASS__ . ': must call startWrite() before set()' );
1213 }
1214
1215 $this->batch[] = array(
1216 'lc_lang' => $this->currentLang,
1217 'lc_key' => $key,
1218 'lc_value' => $this->dbw->encodeBlob( serialize( $value ) ) );
1219 }
1220 }
1221
1222 /**
1223 * LCStore implementation which stores data as a collection of CDB files in the
1224 * directory given by $wgCacheDirectory. If $wgCacheDirectory is not set, this
1225 * will throw an exception.
1226 *
1227 * Profiling indicates that on Linux, this implementation outperforms MySQL if
1228 * the directory is on a local filesystem and there is ample kernel cache
1229 * space. The performance advantage is greater when the DBA extension is
1230 * available than it is with the PHP port.
1231 *
1232 * See Cdb.php and http://cr.yp.to/cdb.html
1233 */
1234 class LCStoreCDB implements LCStore {
1235 /** @var CdbReader[] */
1236 private $readers;
1237
1238 /** @var CdbWriter */
1239 private $writer;
1240
1241 /** @var string Current language code */
1242 private $currentLang;
1243
1244 /** @var bool|string Cache directory. False if not set */
1245 private $directory;
1246
1247 function __construct( $conf = array() ) {
1248 global $wgCacheDirectory;
1249
1250 if ( isset( $conf['directory'] ) ) {
1251 $this->directory = $conf['directory'];
1252 } else {
1253 $this->directory = $wgCacheDirectory;
1254 }
1255 }
1256
1257 public function get( $code, $key ) {
1258 if ( !isset( $this->readers[$code] ) ) {
1259 $fileName = $this->getFileName( $code );
1260
1261 $this->readers[$code] = false;
1262 if ( file_exists( $fileName ) ) {
1263 try {
1264 $this->readers[$code] = CdbReader::open( $fileName );
1265 } catch ( CdbException $e ) {
1266 wfDebug( __METHOD__ . ": unable to open cdb file for reading\n" );
1267 }
1268 }
1269 }
1270
1271 if ( !$this->readers[$code] ) {
1272 return null;
1273 } else {
1274 $value = false;
1275 try {
1276 $value = $this->readers[$code]->get( $key );
1277 } catch ( CdbException $e ) {
1278 wfDebug( __METHOD__ . ": CdbException caught, error message was "
1279 . $e->getMessage() . "\n" );
1280 }
1281 if ( $value === false ) {
1282 return null;
1283 }
1284
1285 return unserialize( $value );
1286 }
1287 }
1288
1289 public function startWrite( $code ) {
1290 if ( !file_exists( $this->directory ) ) {
1291 if ( !wfMkdirParents( $this->directory, null, __METHOD__ ) ) {
1292 throw new MWException( "Unable to create the localisation store " .
1293 "directory \"{$this->directory}\"" );
1294 }
1295 }
1296
1297 // Close reader to stop permission errors on write
1298 if ( !empty( $this->readers[$code] ) ) {
1299 $this->readers[$code]->close();
1300 }
1301
1302 try {
1303 $this->writer = CdbWriter::open( $this->getFileName( $code ) );
1304 } catch ( CdbException $e ) {
1305 throw new MWException( $e->getMessage() );
1306 }
1307 $this->currentLang = $code;
1308 }
1309
1310 public function finishWrite() {
1311 // Close the writer
1312 try {
1313 $this->writer->close();
1314 } catch ( CdbException $e ) {
1315 throw new MWException( $e->getMessage() );
1316 }
1317 $this->writer = null;
1318 unset( $this->readers[$this->currentLang] );
1319 $this->currentLang = null;
1320 }
1321
1322 public function set( $key, $value ) {
1323 if ( is_null( $this->writer ) ) {
1324 throw new MWException( __CLASS__ . ': must call startWrite() before calling set()' );
1325 }
1326 try {
1327 $this->writer->set( $key, serialize( $value ) );
1328 } catch ( CdbException $e ) {
1329 throw new MWException( $e->getMessage() );
1330 }
1331 }
1332
1333 protected function getFileName( $code ) {
1334 if ( strval( $code ) === '' || strpos( $code, '/' ) !== false ) {
1335 throw new MWException( __METHOD__ . ": Invalid language \"$code\"" );
1336 }
1337
1338 return "{$this->directory}/l10n_cache-$code.cdb";
1339 }
1340 }
1341
1342 /**
1343 * Null store backend, used to avoid DB errors during install
1344 */
1345 class LCStoreNull implements LCStore {
1346 public function get( $code, $key ) {
1347 return null;
1348 }
1349
1350 public function startWrite( $code ) {
1351 }
1352
1353 public function finishWrite() {
1354 }
1355
1356 public function set( $key, $value ) {
1357 }
1358 }
1359
1360 /**
1361 * A localisation cache optimised for loading large amounts of data for many
1362 * languages. Used by rebuildLocalisationCache.php.
1363 */
1364 class LocalisationCacheBulkLoad extends LocalisationCache {
1365 /**
1366 * A cache of the contents of data files.
1367 * Core files are serialized to avoid using ~1GB of RAM during a recache.
1368 */
1369 private $fileCache = array();
1370
1371 /**
1372 * Most recently used languages. Uses the linked-list aspect of PHP hashtables
1373 * to keep the most recently used language codes at the end of the array, and
1374 * the language codes that are ready to be deleted at the beginning.
1375 */
1376 private $mruLangs = array();
1377
1378 /**
1379 * Maximum number of languages that may be loaded into $this->data
1380 */
1381 private $maxLoadedLangs = 10;
1382
1383 /**
1384 * @param string $fileName
1385 * @param string $fileType
1386 * @return array|mixed
1387 */
1388 protected function readPHPFile( $fileName, $fileType ) {
1389 $serialize = $fileType === 'core';
1390 if ( !isset( $this->fileCache[$fileName][$fileType] ) ) {
1391 $data = parent::readPHPFile( $fileName, $fileType );
1392
1393 if ( $serialize ) {
1394 $encData = serialize( $data );
1395 } else {
1396 $encData = $data;
1397 }
1398
1399 $this->fileCache[$fileName][$fileType] = $encData;
1400
1401 return $data;
1402 } elseif ( $serialize ) {
1403 return unserialize( $this->fileCache[$fileName][$fileType] );
1404 } else {
1405 return $this->fileCache[$fileName][$fileType];
1406 }
1407 }
1408
1409 /**
1410 * @param string $code
1411 * @param string $key
1412 * @return mixed
1413 */
1414 public function getItem( $code, $key ) {
1415 unset( $this->mruLangs[$code] );
1416 $this->mruLangs[$code] = true;
1417
1418 return parent::getItem( $code, $key );
1419 }
1420
1421 /**
1422 * @param string $code
1423 * @param string $key
1424 * @param string $subkey
1425 * @return mixed
1426 */
1427 public function getSubitem( $code, $key, $subkey ) {
1428 unset( $this->mruLangs[$code] );
1429 $this->mruLangs[$code] = true;
1430
1431 return parent::getSubitem( $code, $key, $subkey );
1432 }
1433
1434 /**
1435 * @param string $code
1436 */
1437 public function recache( $code ) {
1438 parent::recache( $code );
1439 unset( $this->mruLangs[$code] );
1440 $this->mruLangs[$code] = true;
1441 $this->trimCache();
1442 }
1443
1444 /**
1445 * @param string $code
1446 */
1447 public function unload( $code ) {
1448 unset( $this->mruLangs[$code] );
1449 parent::unload( $code );
1450 }
1451
1452 /**
1453 * Unload cached languages until there are less than $this->maxLoadedLangs
1454 */
1455 protected function trimCache() {
1456 while ( count( $this->data ) > $this->maxLoadedLangs && count( $this->mruLangs ) ) {
1457 reset( $this->mruLangs );
1458 $code = key( $this->mruLangs );
1459 wfDebug( __METHOD__ . ": unloading $code\n" );
1460 $this->unload( $code );
1461 }
1462 }
1463 }