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