Merge "Promote LivePreview from its experimental state"
[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 /**
28 * Class for caching the contents of localisation files, Messages*.php
29 * and *.i18n.php.
30 *
31 * An instance of this class is available using Language::getLocalisationCache().
32 *
33 * The values retrieved from here are merged, containing items from extension
34 * files, core messages files and the language fallback sequence (e.g. zh-cn ->
35 * zh-hans -> en ). Some common errors are corrected, for example namespace
36 * names with spaces instead of underscores, but heavyweight processing, such
37 * as grammatical transformation, is done by the caller.
38 */
39 class LocalisationCache {
40 const VERSION = 3;
41
42 /** Configuration associative array */
43 private $conf;
44
45 /**
46 * True if recaching should only be done on an explicit call to recache().
47 * Setting this reduces the overhead of cache freshness checking, which
48 * requires doing a stat() for every extension i18n file.
49 */
50 private $manualRecache = false;
51
52 /**
53 * True to treat all files as expired until they are regenerated by this object.
54 */
55 private $forceRecache = false;
56
57 /**
58 * The cache data. 3-d array, where the first key is the language code,
59 * the second key is the item key e.g. 'messages', and the third key is
60 * an item specific subkey index. Some items are not arrays and so for those
61 * items, there are no subkeys.
62 */
63 protected $data = array();
64
65 /**
66 * The persistent store object. An instance of LCStore.
67 *
68 * @var LCStore
69 */
70 private $store;
71
72 /**
73 * A 2-d associative array, code/key, where presence indicates that the item
74 * is loaded. Value arbitrary.
75 *
76 * For split items, if set, this indicates that all of the subitems have been
77 * loaded.
78 */
79 private $loadedItems = array();
80
81 /**
82 * A 3-d associative array, code/key/subkey, where presence indicates that
83 * the subitem is loaded. Only used for the split items, i.e. messages.
84 */
85 private $loadedSubitems = array();
86
87 /**
88 * An array where presence of a key indicates that that language has been
89 * initialised. Initialisation includes checking for cache expiry and doing
90 * any necessary updates.
91 */
92 private $initialisedLangs = array();
93
94 /**
95 * An array mapping non-existent pseudo-languages to fallback languages. This
96 * is filled by initShallowFallback() when data is requested from a language
97 * that lacks a Messages*.php file.
98 */
99 private $shallowFallbacks = array();
100
101 /**
102 * An array where the keys are codes that have been recached by this instance.
103 */
104 private $recachedLangs = array();
105
106 /**
107 * All item keys
108 */
109 static public $allKeys = array(
110 'fallback', 'namespaceNames', 'bookstoreList',
111 'magicWords', 'messages', 'rtl', 'capitalizeAllNouns', 'digitTransformTable',
112 'separatorTransformTable', 'fallback8bitEncoding', 'linkPrefixExtension',
113 'linkTrail', 'linkPrefixCharset', 'namespaceAliases',
114 'dateFormats', 'datePreferences', 'datePreferenceMigrationMap',
115 'defaultDateFormat', 'extraUserToggles', 'specialPageAliases',
116 'imageFiles', 'preloadedMessages', 'namespaceGenderAliases',
117 'digitGroupingPattern', 'pluralRules', 'pluralRuleTypes', 'compiledPluralRules',
118 );
119
120 /**
121 * Keys for items which consist of associative arrays, which may be merged
122 * by a fallback sequence.
123 */
124 static public $mergeableMapKeys = array( 'messages', 'namespaceNames',
125 'dateFormats', 'imageFiles', 'preloadedMessages'
126 );
127
128 /**
129 * Keys for items which are a numbered array.
130 */
131 static public $mergeableListKeys = array( 'extraUserToggles' );
132
133 /**
134 * Keys for items which contain an array of arrays of equivalent aliases
135 * for each subitem. The aliases may be merged by a fallback sequence.
136 */
137 static public $mergeableAliasListKeys = array( 'specialPageAliases' );
138
139 /**
140 * Keys for items which contain an associative array, and may be merged if
141 * the primary value contains the special array key "inherit". That array
142 * key is removed after the first merge.
143 */
144 static public $optionalMergeKeys = array( 'bookstoreList' );
145
146 /**
147 * Keys for items that are formatted like $magicWords
148 */
149 static public $magicWordKeys = array( 'magicWords' );
150
151 /**
152 * Keys for items where the subitems are stored in the backend separately.
153 */
154 static public $splitKeys = array( 'messages' );
155
156 /**
157 * Keys which are loaded automatically by initLanguage()
158 */
159 static public $preloadedKeys = array( 'dateFormats', 'namespaceNames' );
160
161 /**
162 * Associative array of cached plural rules. The key is the language code,
163 * the value is an array of plural rules for that language.
164 */
165 private $pluralRules = null;
166
167 /**
168 * Associative array of cached plural rule types. The key is the language
169 * code, the value is an array of plural rule types for that language. For
170 * example, $pluralRuleTypes['ar'] = ['zero', 'one', 'two', 'few', 'many'].
171 * The index for each rule type matches the index for the rule in
172 * $pluralRules, thus allowing correlation between the two. The reason we
173 * don't just use the type names as the keys in $pluralRules is because
174 * Language::convertPlural applies the rules based on numeric order (or
175 * explicit numeric parameter), not based on the name of the rule type. For
176 * example, {{plural:count|wordform1|wordform2|wordform3}}, rather than
177 * {{plural:count|one=wordform1|two=wordform2|many=wordform3}}.
178 */
179 private $pluralRuleTypes = null;
180
181 private $mergeableKeys = null;
182
183 /**
184 * Constructor.
185 * For constructor parameters, see the documentation in DefaultSettings.php
186 * for $wgLocalisationCacheConf.
187 *
188 * @param array $conf
189 * @throws MWException
190 */
191 function __construct( $conf ) {
192 global $wgCacheDirectory;
193
194 $this->conf = $conf;
195 $storeConf = array();
196 if ( !empty( $conf['storeClass'] ) ) {
197 $storeClass = $conf['storeClass'];
198 } else {
199 switch ( $conf['store'] ) {
200 case 'files':
201 case 'file':
202 $storeClass = 'LCStoreCDB';
203 break;
204 case 'db':
205 $storeClass = 'LCStoreDB';
206 break;
207 case 'detect':
208 $storeClass = $wgCacheDirectory ? 'LCStoreCDB' : 'LCStoreDB';
209 break;
210 default:
211 throw new MWException(
212 'Please set $wgLocalisationCacheConf[\'store\'] to something sensible.' );
213 }
214 }
215
216 wfDebugLog( 'caches', get_class( $this ) . ": using store $storeClass" );
217 if ( !empty( $conf['storeDirectory'] ) ) {
218 $storeConf['directory'] = $conf['storeDirectory'];
219 }
220
221 $this->store = new $storeClass( $storeConf );
222 foreach ( array( 'manualRecache', 'forceRecache' ) as $var ) {
223 if ( isset( $conf[$var] ) ) {
224 $this->$var = $conf[$var];
225 }
226 }
227 }
228
229 /**
230 * Returns true if the given key is mergeable, that is, if it is an associative
231 * array which can be merged through a fallback sequence.
232 * @param string $key
233 * @return bool
234 */
235 public function isMergeableKey( $key ) {
236 if ( $this->mergeableKeys === null ) {
237 $this->mergeableKeys = array_flip( array_merge(
238 self::$mergeableMapKeys,
239 self::$mergeableListKeys,
240 self::$mergeableAliasListKeys,
241 self::$optionalMergeKeys,
242 self::$magicWordKeys
243 ) );
244 }
245
246 return isset( $this->mergeableKeys[$key] );
247 }
248
249 /**
250 * Get a cache item.
251 *
252 * Warning: this may be slow for split items (messages), since it will
253 * need to fetch all of the subitems from the cache individually.
254 * @param string $code
255 * @param string $key
256 * @return mixed
257 */
258 public function getItem( $code, $key ) {
259 if ( !isset( $this->loadedItems[$code][$key] ) ) {
260 wfProfileIn( __METHOD__ . '-load' );
261 $this->loadItem( $code, $key );
262 wfProfileOut( __METHOD__ . '-load' );
263 }
264
265 if ( $key === 'fallback' && isset( $this->shallowFallbacks[$code] ) ) {
266 return $this->shallowFallbacks[$code];
267 }
268
269 return $this->data[$code][$key];
270 }
271
272 /**
273 * Get a subitem, for instance a single message for a given language.
274 * @param string $code
275 * @param string $key
276 * @param string $subkey
277 * @return mixed|null
278 */
279 public function getSubitem( $code, $key, $subkey ) {
280 if ( !isset( $this->loadedSubitems[$code][$key][$subkey] ) &&
281 !isset( $this->loadedItems[$code][$key] )
282 ) {
283 wfProfileIn( __METHOD__ . '-load' );
284 $this->loadSubitem( $code, $key, $subkey );
285 wfProfileOut( __METHOD__ . '-load' );
286 }
287
288 if ( isset( $this->data[$code][$key][$subkey] ) ) {
289 return $this->data[$code][$key][$subkey];
290 } else {
291 return null;
292 }
293 }
294
295 /**
296 * Get the list of subitem keys for a given item.
297 *
298 * This is faster than array_keys($lc->getItem(...)) for the items listed in
299 * self::$splitKeys.
300 *
301 * Will return null if the item is not found, or false if the item is not an
302 * array.
303 * @param string $code
304 * @param string $key
305 * @return bool|null|string
306 */
307 public function getSubitemList( $code, $key ) {
308 if ( in_array( $key, self::$splitKeys ) ) {
309 return $this->getSubitem( $code, 'list', $key );
310 } else {
311 $item = $this->getItem( $code, $key );
312 if ( is_array( $item ) ) {
313 return array_keys( $item );
314 } else {
315 return false;
316 }
317 }
318 }
319
320 /**
321 * Load an item into the cache.
322 * @param string $code
323 * @param string $key
324 */
325 protected function loadItem( $code, $key ) {
326 if ( !isset( $this->initialisedLangs[$code] ) ) {
327 $this->initLanguage( $code );
328 }
329
330 // Check to see if initLanguage() loaded it for us
331 if ( isset( $this->loadedItems[$code][$key] ) ) {
332 return;
333 }
334
335 if ( isset( $this->shallowFallbacks[$code] ) ) {
336 $this->loadItem( $this->shallowFallbacks[$code], $key );
337
338 return;
339 }
340
341 if ( in_array( $key, self::$splitKeys ) ) {
342 $subkeyList = $this->getSubitem( $code, 'list', $key );
343 foreach ( $subkeyList as $subkey ) {
344 if ( isset( $this->data[$code][$key][$subkey] ) ) {
345 continue;
346 }
347 $this->data[$code][$key][$subkey] = $this->getSubitem( $code, $key, $subkey );
348 }
349 } else {
350 $this->data[$code][$key] = $this->store->get( $code, $key );
351 }
352
353 $this->loadedItems[$code][$key] = true;
354 }
355
356 /**
357 * Load a subitem into the cache
358 * @param string $code
359 * @param string $key
360 * @param string $subkey
361 */
362 protected function loadSubitem( $code, $key, $subkey ) {
363 if ( !in_array( $key, self::$splitKeys ) ) {
364 $this->loadItem( $code, $key );
365
366 return;
367 }
368
369 if ( !isset( $this->initialisedLangs[$code] ) ) {
370 $this->initLanguage( $code );
371 }
372
373 // Check to see if initLanguage() loaded it for us
374 if ( isset( $this->loadedItems[$code][$key] ) ||
375 isset( $this->loadedSubitems[$code][$key][$subkey] )
376 ) {
377 return;
378 }
379
380 if ( isset( $this->shallowFallbacks[$code] ) ) {
381 $this->loadSubitem( $this->shallowFallbacks[$code], $key, $subkey );
382
383 return;
384 }
385
386 $value = $this->store->get( $code, "$key:$subkey" );
387 $this->data[$code][$key][$subkey] = $value;
388 $this->loadedSubitems[$code][$key][$subkey] = true;
389 }
390
391 /**
392 * Returns true if the cache identified by $code is missing or expired.
393 *
394 * @param string $code
395 *
396 * @return bool
397 */
398 public function isExpired( $code ) {
399 if ( $this->forceRecache && !isset( $this->recachedLangs[$code] ) ) {
400 wfDebug( __METHOD__ . "($code): forced reload\n" );
401
402 return true;
403 }
404
405 $deps = $this->store->get( $code, 'deps' );
406 $keys = $this->store->get( $code, 'list' );
407 $preload = $this->store->get( $code, 'preload' );
408 // Different keys may expire separately for some stores
409 if ( $deps === null || $keys === null || $preload === null ) {
410 wfDebug( __METHOD__ . "($code): cache missing, need to make one\n" );
411
412 return true;
413 }
414
415 foreach ( $deps as $dep ) {
416 // Because we're unserializing stuff from cache, we
417 // could receive objects of classes that don't exist
418 // anymore (e.g. uninstalled extensions)
419 // When this happens, always expire the cache
420 if ( !$dep instanceof CacheDependency || $dep->isExpired() ) {
421 wfDebug( __METHOD__ . "($code): cache for $code expired due to " .
422 get_class( $dep ) . "\n" );
423
424 return true;
425 }
426 }
427
428 return false;
429 }
430
431 /**
432 * Initialise a language in this object. Rebuild the cache if necessary.
433 * @param string $code
434 * @throws MWException
435 */
436 protected function initLanguage( $code ) {
437 if ( isset( $this->initialisedLangs[$code] ) ) {
438 return;
439 }
440
441 $this->initialisedLangs[$code] = true;
442
443 # If the code is of the wrong form for a Messages*.php file, do a shallow fallback
444 if ( !Language::isValidBuiltInCode( $code ) ) {
445 $this->initShallowFallback( $code, 'en' );
446
447 return;
448 }
449
450 # Recache the data if necessary
451 if ( !$this->manualRecache && $this->isExpired( $code ) ) {
452 if ( Language::isSupportedLanguage( $code ) ) {
453 $this->recache( $code );
454 } elseif ( $code === 'en' ) {
455 throw new MWException( 'MessagesEn.php is missing.' );
456 } else {
457 $this->initShallowFallback( $code, 'en' );
458 }
459
460 return;
461 }
462
463 # Preload some stuff
464 $preload = $this->getItem( $code, 'preload' );
465 if ( $preload === null ) {
466 if ( $this->manualRecache ) {
467 // No Messages*.php file. Do shallow fallback to en.
468 if ( $code === 'en' ) {
469 throw new MWException( 'No localisation cache found for English. ' .
470 'Please run maintenance/rebuildLocalisationCache.php.' );
471 }
472 $this->initShallowFallback( $code, 'en' );
473
474 return;
475 } else {
476 throw new MWException( 'Invalid or missing localisation cache.' );
477 }
478 }
479 $this->data[$code] = $preload;
480 foreach ( $preload as $key => $item ) {
481 if ( in_array( $key, self::$splitKeys ) ) {
482 foreach ( $item as $subkey => $subitem ) {
483 $this->loadedSubitems[$code][$key][$subkey] = true;
484 }
485 } else {
486 $this->loadedItems[$code][$key] = true;
487 }
488 }
489 }
490
491 /**
492 * Create a fallback from one language to another, without creating a
493 * complete persistent cache.
494 * @param string $primaryCode
495 * @param string $fallbackCode
496 */
497 public function initShallowFallback( $primaryCode, $fallbackCode ) {
498 $this->data[$primaryCode] =& $this->data[$fallbackCode];
499 $this->loadedItems[$primaryCode] =& $this->loadedItems[$fallbackCode];
500 $this->loadedSubitems[$primaryCode] =& $this->loadedSubitems[$fallbackCode];
501 $this->shallowFallbacks[$primaryCode] = $fallbackCode;
502 }
503
504 /**
505 * Read a PHP file containing localisation data.
506 * @param string $_fileName
507 * @param string $_fileType
508 * @throws MWException
509 * @return array
510 */
511 protected function readPHPFile( $_fileName, $_fileType ) {
512 wfProfileIn( __METHOD__ );
513 // Disable APC caching
514 wfSuppressWarnings();
515 $_apcEnabled = ini_set( 'apc.cache_by_default', '0' );
516 wfRestoreWarnings();
517
518 include $_fileName;
519
520 wfSuppressWarnings();
521 ini_set( 'apc.cache_by_default', $_apcEnabled );
522 wfRestoreWarnings();
523
524 if ( $_fileType == 'core' || $_fileType == 'extension' ) {
525 $data = compact( self::$allKeys );
526 } elseif ( $_fileType == 'aliases' ) {
527 $data = compact( 'aliases' );
528 } else {
529 wfProfileOut( __METHOD__ );
530 throw new MWException( __METHOD__ . ": Invalid file type: $_fileType" );
531 }
532 wfProfileOut( __METHOD__ );
533
534 return $data;
535 }
536
537 /**
538 * Read a JSON file containing localisation messages.
539 * @param string $fileName Name of file to read
540 * @throws MWException If there is a syntax error in the JSON file
541 * @return array Array with a 'messages' key, or empty array if the file doesn't exist
542 */
543 public function readJSONFile( $fileName ) {
544 wfProfileIn( __METHOD__ );
545
546 if ( !is_readable( $fileName ) ) {
547 wfProfileOut( __METHOD__ );
548
549 return array();
550 }
551
552 $json = file_get_contents( $fileName );
553 if ( $json === false ) {
554 wfProfileOut( __METHOD__ );
555
556 return array();
557 }
558
559 $data = FormatJson::decode( $json, true );
560 if ( $data === null ) {
561 wfProfileOut( __METHOD__ );
562
563 throw new MWException( __METHOD__ . ": Invalid JSON file: $fileName" );
564 }
565
566 // Remove keys starting with '@', they're reserved for metadata and non-message data
567 foreach ( $data as $key => $unused ) {
568 if ( $key === '' || $key[0] === '@' ) {
569 unset( $data[$key] );
570 }
571 }
572
573 wfProfileOut( __METHOD__ );
574
575 // The JSON format only supports messages, none of the other variables, so wrap the data
576 return array( 'messages' => $data );
577 }
578
579 /**
580 * Get the compiled plural rules for a given language from the XML files.
581 * @since 1.20
582 * @param string $code
583 * @return array|null
584 */
585 public function getCompiledPluralRules( $code ) {
586 $rules = $this->getPluralRules( $code );
587 if ( $rules === null ) {
588 return null;
589 }
590 try {
591 $compiledRules = CLDRPluralRuleEvaluator::compile( $rules );
592 } catch ( CLDRPluralRuleError $e ) {
593 wfDebugLog( 'l10n', $e->getMessage() );
594
595 return array();
596 }
597
598 return $compiledRules;
599 }
600
601 /**
602 * Get the plural rules for a given language from the XML files.
603 * Cached.
604 * @since 1.20
605 * @param string $code
606 * @return array|null
607 */
608 public function getPluralRules( $code ) {
609 if ( $this->pluralRules === null ) {
610 $this->loadPluralFiles();
611 }
612 if ( !isset( $this->pluralRules[$code] ) ) {
613 return null;
614 } else {
615 return $this->pluralRules[$code];
616 }
617 }
618
619 /**
620 * Get the plural rule types for a given language from the XML files.
621 * Cached.
622 * @since 1.22
623 * @param string $code
624 * @return array|null
625 */
626 public function getPluralRuleTypes( $code ) {
627 if ( $this->pluralRuleTypes === null ) {
628 $this->loadPluralFiles();
629 }
630 if ( !isset( $this->pluralRuleTypes[$code] ) ) {
631 return null;
632 } else {
633 return $this->pluralRuleTypes[$code];
634 }
635 }
636
637 /**
638 * Load the plural XML files.
639 */
640 protected function loadPluralFiles() {
641 global $IP;
642 $cldrPlural = "$IP/languages/data/plurals.xml";
643 $mwPlural = "$IP/languages/data/plurals-mediawiki.xml";
644 // Load CLDR plural rules
645 $this->loadPluralFile( $cldrPlural );
646 if ( file_exists( $mwPlural ) ) {
647 // Override or extend
648 $this->loadPluralFile( $mwPlural );
649 }
650 }
651
652 /**
653 * Load a plural XML file with the given filename, compile the relevant
654 * rules, and save the compiled rules in a process-local cache.
655 *
656 * @param string $fileName
657 */
658 protected function loadPluralFile( $fileName ) {
659 $doc = new DOMDocument;
660 $doc->load( $fileName );
661 $rulesets = $doc->getElementsByTagName( "pluralRules" );
662 foreach ( $rulesets as $ruleset ) {
663 $codes = $ruleset->getAttribute( 'locales' );
664 $rules = array();
665 $ruleTypes = array();
666 $ruleElements = $ruleset->getElementsByTagName( "pluralRule" );
667 foreach ( $ruleElements as $elt ) {
668 $ruleType = $elt->getAttribute( 'count' );
669 if ( $ruleType === 'other' ) {
670 // Don't record "other" rules, which have an empty condition
671 continue;
672 }
673 $rules[] = $elt->nodeValue;
674 $ruleTypes[] = $ruleType;
675 }
676 foreach ( explode( ' ', $codes ) as $code ) {
677 $this->pluralRules[$code] = $rules;
678 $this->pluralRuleTypes[$code] = $ruleTypes;
679 }
680 }
681 }
682
683 /**
684 * Read the data from the source files for a given language, and register
685 * the relevant dependencies in the $deps array. If the localisation
686 * exists, the data array is returned, otherwise false is returned.
687 *
688 * @param string $code
689 * @param array $deps
690 * @return array
691 */
692 protected function readSourceFilesAndRegisterDeps( $code, &$deps ) {
693 global $IP;
694 wfProfileIn( __METHOD__ );
695
696 // This reads in the PHP i18n file with non-messages l10n data
697 $fileName = Language::getMessagesFileName( $code );
698 if ( !file_exists( $fileName ) ) {
699 $data = array();
700 } else {
701 $deps[] = new FileDependency( $fileName );
702 $data = $this->readPHPFile( $fileName, 'core' );
703 }
704
705 # Load CLDR plural rules for JavaScript
706 $data['pluralRules'] = $this->getPluralRules( $code );
707 # And for PHP
708 $data['compiledPluralRules'] = $this->getCompiledPluralRules( $code );
709 # Load plural rule types
710 $data['pluralRuleTypes'] = $this->getPluralRuleTypes( $code );
711
712 $deps['plurals'] = new FileDependency( "$IP/languages/data/plurals.xml" );
713 $deps['plurals-mw'] = new FileDependency( "$IP/languages/data/plurals-mediawiki.xml" );
714
715 wfProfileOut( __METHOD__ );
716
717 return $data;
718 }
719
720 /**
721 * Merge two localisation values, a primary and a fallback, overwriting the
722 * primary value in place.
723 * @param string $key
724 * @param mixed $value
725 * @param mixed $fallbackValue
726 */
727 protected function mergeItem( $key, &$value, $fallbackValue ) {
728 if ( !is_null( $value ) ) {
729 if ( !is_null( $fallbackValue ) ) {
730 if ( in_array( $key, self::$mergeableMapKeys ) ) {
731 $value = $value + $fallbackValue;
732 } elseif ( in_array( $key, self::$mergeableListKeys ) ) {
733 $value = array_unique( array_merge( $fallbackValue, $value ) );
734 } elseif ( in_array( $key, self::$mergeableAliasListKeys ) ) {
735 $value = array_merge_recursive( $value, $fallbackValue );
736 } elseif ( in_array( $key, self::$optionalMergeKeys ) ) {
737 if ( !empty( $value['inherit'] ) ) {
738 $value = array_merge( $fallbackValue, $value );
739 }
740
741 if ( isset( $value['inherit'] ) ) {
742 unset( $value['inherit'] );
743 }
744 } elseif ( in_array( $key, self::$magicWordKeys ) ) {
745 $this->mergeMagicWords( $value, $fallbackValue );
746 }
747 }
748 } else {
749 $value = $fallbackValue;
750 }
751 }
752
753 /**
754 * @param mixed $value
755 * @param mixed $fallbackValue
756 */
757 protected function mergeMagicWords( &$value, $fallbackValue ) {
758 foreach ( $fallbackValue as $magicName => $fallbackInfo ) {
759 if ( !isset( $value[$magicName] ) ) {
760 $value[$magicName] = $fallbackInfo;
761 } else {
762 $oldSynonyms = array_slice( $fallbackInfo, 1 );
763 $newSynonyms = array_slice( $value[$magicName], 1 );
764 $synonyms = array_values( array_unique( array_merge(
765 $newSynonyms, $oldSynonyms ) ) );
766 $value[$magicName] = array_merge( array( $fallbackInfo[0] ), $synonyms );
767 }
768 }
769 }
770
771 /**
772 * Given an array mapping language code to localisation value, such as is
773 * found in extension *.i18n.php files, iterate through a fallback sequence
774 * to merge the given data with an existing primary value.
775 *
776 * Returns true if any data from the extension array was used, false
777 * otherwise.
778 * @param array $codeSequence
779 * @param string $key
780 * @param mixed $value
781 * @param mixed $fallbackValue
782 * @return bool
783 */
784 protected function mergeExtensionItem( $codeSequence, $key, &$value, $fallbackValue ) {
785 $used = false;
786 foreach ( $codeSequence as $code ) {
787 if ( isset( $fallbackValue[$code] ) ) {
788 $this->mergeItem( $key, $value, $fallbackValue[$code] );
789 $used = true;
790 }
791 }
792
793 return $used;
794 }
795
796 /**
797 * Gets the combined list of messages dirs from
798 * core and extensions
799 *
800 * @since 1.25
801 * @return array
802 */
803 public function getMessagesDirs() {
804 global $wgMessagesDirs, $IP;
805 return array(
806 'core' => "$IP/languages/i18n",
807 'api' => "$IP/includes/api/i18n",
808 'oojs-ui' => "$IP/resources/lib/oojs-ui/i18n",
809 ) + $wgMessagesDirs;
810 }
811
812 /**
813 * Load localisation data for a given language for both core and extensions
814 * and save it to the persistent cache store and the process cache
815 * @param string $code
816 * @throws MWException
817 */
818 public function recache( $code ) {
819 global $wgExtensionMessagesFiles;
820 wfProfileIn( __METHOD__ );
821
822 if ( !$code ) {
823 wfProfileOut( __METHOD__ );
824 throw new MWException( "Invalid language code requested" );
825 }
826 $this->recachedLangs[$code] = true;
827
828 # Initial values
829 $initialData = array_combine(
830 self::$allKeys,
831 array_fill( 0, count( self::$allKeys ), null ) );
832 $coreData = $initialData;
833 $deps = array();
834
835 # Load the primary localisation from the source file
836 $data = $this->readSourceFilesAndRegisterDeps( $code, $deps );
837 if ( $data === false ) {
838 wfDebug( __METHOD__ . ": no localisation file for $code, using fallback to en\n" );
839 $coreData['fallback'] = 'en';
840 } else {
841 wfDebug( __METHOD__ . ": got localisation for $code from source\n" );
842
843 # Merge primary localisation
844 foreach ( $data as $key => $value ) {
845 $this->mergeItem( $key, $coreData[$key], $value );
846 }
847 }
848
849 # Fill in the fallback if it's not there already
850 if ( is_null( $coreData['fallback'] ) ) {
851 $coreData['fallback'] = $code === 'en' ? false : 'en';
852 }
853 if ( $coreData['fallback'] === false ) {
854 $coreData['fallbackSequence'] = array();
855 } else {
856 $coreData['fallbackSequence'] = array_map( 'trim', explode( ',', $coreData['fallback'] ) );
857 $len = count( $coreData['fallbackSequence'] );
858
859 # Ensure that the sequence ends at en
860 if ( $coreData['fallbackSequence'][$len - 1] !== 'en' ) {
861 $coreData['fallbackSequence'][] = 'en';
862 }
863 }
864
865 $codeSequence = array_merge( array( $code ), $coreData['fallbackSequence'] );
866 $messageDirs = $this->getMessagesDirs();
867
868 wfProfileIn( __METHOD__ . '-fallbacks' );
869
870 # Load non-JSON localisation data for extensions
871 $extensionData = array_combine(
872 $codeSequence,
873 array_fill( 0, count( $codeSequence ), $initialData ) );
874 foreach ( $wgExtensionMessagesFiles as $extension => $fileName ) {
875 if ( isset( $messageDirs[$extension] ) ) {
876 # This extension has JSON message data; skip the PHP shim
877 continue;
878 }
879
880 $data = $this->readPHPFile( $fileName, 'extension' );
881 $used = false;
882
883 foreach ( $data as $key => $item ) {
884 foreach ( $codeSequence as $csCode ) {
885 if ( isset( $item[$csCode] ) ) {
886 $this->mergeItem( $key, $extensionData[$csCode][$key], $item[$csCode] );
887 $used = true;
888 }
889 }
890 }
891
892 if ( $used ) {
893 $deps[] = new FileDependency( $fileName );
894 }
895 }
896
897 # Load the localisation data for each fallback, then merge it into the full array
898 $allData = $initialData;
899 foreach ( $codeSequence as $csCode ) {
900 $csData = $initialData;
901
902 # Load core messages and the extension localisations.
903 foreach ( $messageDirs as $dirs ) {
904 foreach ( (array)$dirs as $dir ) {
905 $fileName = "$dir/$csCode.json";
906 $data = $this->readJSONFile( $fileName );
907
908 foreach ( $data as $key => $item ) {
909 $this->mergeItem( $key, $csData[$key], $item );
910 }
911
912 $deps[] = new FileDependency( $fileName );
913 }
914 }
915
916 # Merge non-JSON extension data
917 if ( isset( $extensionData[$csCode] ) ) {
918 foreach ( $extensionData[$csCode] as $key => $item ) {
919 $this->mergeItem( $key, $csData[$key], $item );
920 }
921 }
922
923 if ( $csCode === $code ) {
924 # Merge core data into extension data
925 foreach ( $coreData as $key => $item ) {
926 $this->mergeItem( $key, $csData[$key], $item );
927 }
928 } else {
929 # Load the secondary localisation from the source file to
930 # avoid infinite cycles on cyclic fallbacks
931 $fbData = $this->readSourceFilesAndRegisterDeps( $csCode, $deps );
932 if ( $fbData !== false ) {
933 # Only merge the keys that make sense to merge
934 foreach ( self::$allKeys as $key ) {
935 if ( !isset( $fbData[$key] ) ) {
936 continue;
937 }
938
939 if ( is_null( $coreData[$key] ) || $this->isMergeableKey( $key ) ) {
940 $this->mergeItem( $key, $csData[$key], $fbData[$key] );
941 }
942 }
943 }
944 }
945
946 # Allow extensions an opportunity to adjust the data for this
947 # fallback
948 wfRunHooks( 'LocalisationCacheRecacheFallback', array( $this, $csCode, &$csData ) );
949
950 # Merge the data for this fallback into the final array
951 if ( $csCode === $code ) {
952 $allData = $csData;
953 } else {
954 foreach ( self::$allKeys as $key ) {
955 if ( !isset( $csData[$key] ) ) {
956 continue;
957 }
958
959 if ( is_null( $allData[$key] ) || $this->isMergeableKey( $key ) ) {
960 $this->mergeItem( $key, $allData[$key], $csData[$key] );
961 }
962 }
963 }
964 }
965
966 wfProfileOut( __METHOD__ . '-fallbacks' );
967
968 # Add cache dependencies for any referenced globals
969 $deps['wgExtensionMessagesFiles'] = new GlobalDependency( 'wgExtensionMessagesFiles' );
970 // $wgMessagesDirs is used in LocalisationCache::getMessagesDirs()
971 $deps['wgMessagesDirs'] = new GlobalDependency( 'wgMessagesDirs' );
972 $deps['version'] = new ConstantDependency( 'LocalisationCache::VERSION' );
973
974 # Add dependencies to the cache entry
975 $allData['deps'] = $deps;
976
977 # Replace spaces with underscores in namespace names
978 $allData['namespaceNames'] = str_replace( ' ', '_', $allData['namespaceNames'] );
979
980 # And do the same for special page aliases. $page is an array.
981 foreach ( $allData['specialPageAliases'] as &$page ) {
982 $page = str_replace( ' ', '_', $page );
983 }
984 # Decouple the reference to prevent accidental damage
985 unset( $page );
986
987 # If there were no plural rules, return an empty array
988 if ( $allData['pluralRules'] === null ) {
989 $allData['pluralRules'] = array();
990 }
991 if ( $allData['compiledPluralRules'] === null ) {
992 $allData['compiledPluralRules'] = array();
993 }
994 # If there were no plural rule types, return an empty array
995 if ( $allData['pluralRuleTypes'] === null ) {
996 $allData['pluralRuleTypes'] = array();
997 }
998
999 # Set the list keys
1000 $allData['list'] = array();
1001 foreach ( self::$splitKeys as $key ) {
1002 $allData['list'][$key] = array_keys( $allData[$key] );
1003 }
1004 # Run hooks
1005 $purgeBlobs = true;
1006 wfRunHooks( 'LocalisationCacheRecache', array( $this, $code, &$allData, &$purgeBlobs ) );
1007
1008 if ( is_null( $allData['namespaceNames'] ) ) {
1009 wfProfileOut( __METHOD__ );
1010 throw new MWException( __METHOD__ . ': Localisation data failed sanity check! ' .
1011 'Check that your languages/messages/MessagesEn.php file is intact.' );
1012 }
1013
1014 # Set the preload key
1015 $allData['preload'] = $this->buildPreload( $allData );
1016
1017 # Save to the process cache and register the items loaded
1018 $this->data[$code] = $allData;
1019 foreach ( $allData as $key => $item ) {
1020 $this->loadedItems[$code][$key] = true;
1021 }
1022
1023 # Save to the persistent cache
1024 wfProfileIn( __METHOD__ . '-write' );
1025 $this->store->startWrite( $code );
1026 foreach ( $allData as $key => $value ) {
1027 if ( in_array( $key, self::$splitKeys ) ) {
1028 foreach ( $value as $subkey => $subvalue ) {
1029 $this->store->set( "$key:$subkey", $subvalue );
1030 }
1031 } else {
1032 $this->store->set( $key, $value );
1033 }
1034 }
1035 $this->store->finishWrite();
1036 wfProfileOut( __METHOD__ . '-write' );
1037
1038 # Clear out the MessageBlobStore
1039 # HACK: If using a null (i.e. disabled) storage backend, we
1040 # can't write to the MessageBlobStore either
1041 if ( $purgeBlobs && !$this->store instanceof LCStoreNull ) {
1042 MessageBlobStore::getInstance()->clear();
1043 }
1044
1045 wfProfileOut( __METHOD__ );
1046 }
1047
1048 /**
1049 * Build the preload item from the given pre-cache data.
1050 *
1051 * The preload item will be loaded automatically, improving performance
1052 * for the commonly-requested items it contains.
1053 * @param array $data
1054 * @return array
1055 */
1056 protected function buildPreload( $data ) {
1057 $preload = array( 'messages' => array() );
1058 foreach ( self::$preloadedKeys as $key ) {
1059 $preload[$key] = $data[$key];
1060 }
1061
1062 foreach ( $data['preloadedMessages'] as $subkey ) {
1063 if ( isset( $data['messages'][$subkey] ) ) {
1064 $subitem = $data['messages'][$subkey];
1065 } else {
1066 $subitem = null;
1067 }
1068 $preload['messages'][$subkey] = $subitem;
1069 }
1070
1071 return $preload;
1072 }
1073
1074 /**
1075 * Unload the data for a given language from the object cache.
1076 * Reduces memory usage.
1077 * @param string $code
1078 */
1079 public function unload( $code ) {
1080 unset( $this->data[$code] );
1081 unset( $this->loadedItems[$code] );
1082 unset( $this->loadedSubitems[$code] );
1083 unset( $this->initialisedLangs[$code] );
1084 unset( $this->shallowFallbacks[$code] );
1085
1086 foreach ( $this->shallowFallbacks as $shallowCode => $fbCode ) {
1087 if ( $fbCode === $code ) {
1088 $this->unload( $shallowCode );
1089 }
1090 }
1091 }
1092
1093 /**
1094 * Unload all data
1095 */
1096 public function unloadAll() {
1097 foreach ( $this->initialisedLangs as $lang => $unused ) {
1098 $this->unload( $lang );
1099 }
1100 }
1101
1102 /**
1103 * Disable the storage backend
1104 */
1105 public function disableBackend() {
1106 $this->store = new LCStoreNull;
1107 $this->manualRecache = false;
1108 }
1109 }
1110
1111 /**
1112 * Interface for the persistence layer of LocalisationCache.
1113 *
1114 * The persistence layer is two-level hierarchical cache. The first level
1115 * is the language, the second level is the item or subitem.
1116 *
1117 * Since the data for a whole language is rebuilt in one operation, it needs
1118 * to have a fast and atomic method for deleting or replacing all of the
1119 * current data for a given language. The interface reflects this bulk update
1120 * operation. Callers writing to the cache must first call startWrite(), then
1121 * will call set() a couple of thousand times, then will call finishWrite()
1122 * to commit the operation. When finishWrite() is called, the cache is
1123 * expected to delete all data previously stored for that language.
1124 *
1125 * The values stored are PHP variables suitable for serialize(). Implementations
1126 * of LCStore are responsible for serializing and unserializing.
1127 */
1128 interface LCStore {
1129 /**
1130 * Get a value.
1131 * @param string $code Language code
1132 * @param string $key Cache key
1133 */
1134 function get( $code, $key );
1135
1136 /**
1137 * Start a write transaction.
1138 * @param string $code Language code
1139 */
1140 function startWrite( $code );
1141
1142 /**
1143 * Finish a write transaction.
1144 */
1145 function finishWrite();
1146
1147 /**
1148 * Set a key to a given value. startWrite() must be called before this
1149 * is called, and finishWrite() must be called afterwards.
1150 * @param string $key
1151 * @param mixed $value
1152 */
1153 function set( $key, $value );
1154 }
1155
1156 /**
1157 * LCStore implementation which uses the standard DB functions to store data.
1158 * This will work on any MediaWiki installation.
1159 */
1160 class LCStoreDB implements LCStore {
1161 private $currentLang;
1162 private $writesDone = false;
1163
1164 /** @var DatabaseBase */
1165 private $dbw;
1166 /** @var array */
1167 private $batch = array();
1168
1169 private $readOnly = false;
1170
1171 public function get( $code, $key ) {
1172 if ( $this->writesDone ) {
1173 $db = wfGetDB( DB_MASTER );
1174 } else {
1175 $db = wfGetDB( DB_SLAVE );
1176 }
1177 $row = $db->selectRow( 'l10n_cache', array( 'lc_value' ),
1178 array( 'lc_lang' => $code, 'lc_key' => $key ), __METHOD__ );
1179 if ( $row ) {
1180 return unserialize( $db->decodeBlob( $row->lc_value ) );
1181 } else {
1182 return null;
1183 }
1184 }
1185
1186 public function startWrite( $code ) {
1187 if ( $this->readOnly ) {
1188 return;
1189 } elseif ( !$code ) {
1190 throw new MWException( __METHOD__ . ": Invalid language \"$code\"" );
1191 }
1192
1193 $this->dbw = wfGetDB( DB_MASTER );
1194
1195 $this->currentLang = $code;
1196 $this->batch = array();
1197 }
1198
1199 public function finishWrite() {
1200 if ( $this->readOnly ) {
1201 return;
1202 } elseif ( is_null( $this->currentLang ) ) {
1203 throw new MWException( __CLASS__ . ': must call startWrite() before finishWrite()' );
1204 }
1205
1206 $this->dbw->begin( __METHOD__ );
1207 try {
1208 $this->dbw->delete( 'l10n_cache',
1209 array( 'lc_lang' => $this->currentLang ), __METHOD__ );
1210 foreach ( array_chunk( $this->batch, 500 ) as $rows ) {
1211 $this->dbw->insert( 'l10n_cache', $rows, __METHOD__ );
1212 }
1213 $this->writesDone = true;
1214 } catch ( DBQueryError $e ) {
1215 if ( $this->dbw->wasReadOnlyError() ) {
1216 $this->readOnly = true; // just avoid site down time
1217 } else {
1218 throw $e;
1219 }
1220 }
1221 $this->dbw->commit( __METHOD__ );
1222
1223 $this->currentLang = null;
1224 $this->batch = array();
1225 }
1226
1227 public function set( $key, $value ) {
1228 if ( $this->readOnly ) {
1229 return;
1230 } elseif ( is_null( $this->currentLang ) ) {
1231 throw new MWException( __CLASS__ . ': must call startWrite() before set()' );
1232 }
1233
1234 $this->batch[] = array(
1235 'lc_lang' => $this->currentLang,
1236 'lc_key' => $key,
1237 'lc_value' => $this->dbw->encodeBlob( serialize( $value ) ) );
1238 }
1239 }
1240
1241 /**
1242 * LCStore implementation which stores data as a collection of CDB files in the
1243 * directory given by $wgCacheDirectory. If $wgCacheDirectory is not set, this
1244 * will throw an exception.
1245 *
1246 * Profiling indicates that on Linux, this implementation outperforms MySQL if
1247 * the directory is on a local filesystem and there is ample kernel cache
1248 * space. The performance advantage is greater when the DBA extension is
1249 * available than it is with the PHP port.
1250 *
1251 * See Cdb.php and http://cr.yp.to/cdb.html
1252 */
1253 class LCStoreCDB implements LCStore {
1254 /** @var CdbReader[] */
1255 private $readers;
1256
1257 /** @var CdbWriter */
1258 private $writer;
1259
1260 /** @var string Current language code */
1261 private $currentLang;
1262
1263 /** @var bool|string Cache directory. False if not set */
1264 private $directory;
1265
1266 function __construct( $conf = array() ) {
1267 global $wgCacheDirectory;
1268
1269 if ( isset( $conf['directory'] ) ) {
1270 $this->directory = $conf['directory'];
1271 } else {
1272 $this->directory = $wgCacheDirectory;
1273 }
1274 }
1275
1276 public function get( $code, $key ) {
1277 if ( !isset( $this->readers[$code] ) ) {
1278 $fileName = $this->getFileName( $code );
1279
1280 $this->readers[$code] = false;
1281 if ( file_exists( $fileName ) ) {
1282 try {
1283 $this->readers[$code] = CdbReader::open( $fileName );
1284 } catch ( CdbException $e ) {
1285 wfDebug( __METHOD__ . ": unable to open cdb file for reading\n" );
1286 }
1287 }
1288 }
1289
1290 if ( !$this->readers[$code] ) {
1291 return null;
1292 } else {
1293 $value = false;
1294 try {
1295 $value = $this->readers[$code]->get( $key );
1296 } catch ( CdbException $e ) {
1297 wfDebug( __METHOD__ . ": CdbException caught, error message was "
1298 . $e->getMessage() . "\n" );
1299 }
1300 if ( $value === false ) {
1301 return null;
1302 }
1303
1304 return unserialize( $value );
1305 }
1306 }
1307
1308 public function startWrite( $code ) {
1309 if ( !file_exists( $this->directory ) ) {
1310 if ( !wfMkdirParents( $this->directory, null, __METHOD__ ) ) {
1311 throw new MWException( "Unable to create the localisation store " .
1312 "directory \"{$this->directory}\"" );
1313 }
1314 }
1315
1316 // Close reader to stop permission errors on write
1317 if ( !empty( $this->readers[$code] ) ) {
1318 $this->readers[$code]->close();
1319 }
1320
1321 try {
1322 $this->writer = CdbWriter::open( $this->getFileName( $code ) );
1323 } catch ( CdbException $e ) {
1324 throw new MWException( $e->getMessage() );
1325 }
1326 $this->currentLang = $code;
1327 }
1328
1329 public function finishWrite() {
1330 // Close the writer
1331 try {
1332 $this->writer->close();
1333 } catch ( CdbException $e ) {
1334 throw new MWException( $e->getMessage() );
1335 }
1336 $this->writer = null;
1337 unset( $this->readers[$this->currentLang] );
1338 $this->currentLang = null;
1339 }
1340
1341 public function set( $key, $value ) {
1342 if ( is_null( $this->writer ) ) {
1343 throw new MWException( __CLASS__ . ': must call startWrite() before calling set()' );
1344 }
1345 try {
1346 $this->writer->set( $key, serialize( $value ) );
1347 } catch ( CdbException $e ) {
1348 throw new MWException( $e->getMessage() );
1349 }
1350 }
1351
1352 protected function getFileName( $code ) {
1353 if ( strval( $code ) === '' || strpos( $code, '/' ) !== false ) {
1354 throw new MWException( __METHOD__ . ": Invalid language \"$code\"" );
1355 }
1356
1357 return "{$this->directory}/l10n_cache-$code.cdb";
1358 }
1359 }
1360
1361 /**
1362 * Null store backend, used to avoid DB errors during install
1363 */
1364 class LCStoreNull implements LCStore {
1365 public function get( $code, $key ) {
1366 return null;
1367 }
1368
1369 public function startWrite( $code ) {
1370 }
1371
1372 public function finishWrite() {
1373 }
1374
1375 public function set( $key, $value ) {
1376 }
1377 }
1378
1379 /**
1380 * A localisation cache optimised for loading large amounts of data for many
1381 * languages. Used by rebuildLocalisationCache.php.
1382 */
1383 class LocalisationCacheBulkLoad extends LocalisationCache {
1384 /**
1385 * A cache of the contents of data files.
1386 * Core files are serialized to avoid using ~1GB of RAM during a recache.
1387 */
1388 private $fileCache = array();
1389
1390 /**
1391 * Most recently used languages. Uses the linked-list aspect of PHP hashtables
1392 * to keep the most recently used language codes at the end of the array, and
1393 * the language codes that are ready to be deleted at the beginning.
1394 */
1395 private $mruLangs = array();
1396
1397 /**
1398 * Maximum number of languages that may be loaded into $this->data
1399 */
1400 private $maxLoadedLangs = 10;
1401
1402 /**
1403 * @param string $fileName
1404 * @param string $fileType
1405 * @return array|mixed
1406 */
1407 protected function readPHPFile( $fileName, $fileType ) {
1408 $serialize = $fileType === 'core';
1409 if ( !isset( $this->fileCache[$fileName][$fileType] ) ) {
1410 $data = parent::readPHPFile( $fileName, $fileType );
1411
1412 if ( $serialize ) {
1413 $encData = serialize( $data );
1414 } else {
1415 $encData = $data;
1416 }
1417
1418 $this->fileCache[$fileName][$fileType] = $encData;
1419
1420 return $data;
1421 } elseif ( $serialize ) {
1422 return unserialize( $this->fileCache[$fileName][$fileType] );
1423 } else {
1424 return $this->fileCache[$fileName][$fileType];
1425 }
1426 }
1427
1428 /**
1429 * @param string $code
1430 * @param string $key
1431 * @return mixed
1432 */
1433 public function getItem( $code, $key ) {
1434 unset( $this->mruLangs[$code] );
1435 $this->mruLangs[$code] = true;
1436
1437 return parent::getItem( $code, $key );
1438 }
1439
1440 /**
1441 * @param string $code
1442 * @param string $key
1443 * @param string $subkey
1444 * @return mixed
1445 */
1446 public function getSubitem( $code, $key, $subkey ) {
1447 unset( $this->mruLangs[$code] );
1448 $this->mruLangs[$code] = true;
1449
1450 return parent::getSubitem( $code, $key, $subkey );
1451 }
1452
1453 /**
1454 * @param string $code
1455 */
1456 public function recache( $code ) {
1457 parent::recache( $code );
1458 unset( $this->mruLangs[$code] );
1459 $this->mruLangs[$code] = true;
1460 $this->trimCache();
1461 }
1462
1463 /**
1464 * @param string $code
1465 */
1466 public function unload( $code ) {
1467 unset( $this->mruLangs[$code] );
1468 parent::unload( $code );
1469 }
1470
1471 /**
1472 * Unload cached languages until there are less than $this->maxLoadedLangs
1473 */
1474 protected function trimCache() {
1475 while ( count( $this->data ) > $this->maxLoadedLangs && count( $this->mruLangs ) ) {
1476 reset( $this->mruLangs );
1477 $code = key( $this->mruLangs );
1478 wfDebug( __METHOD__ . ": unloading $code\n" );
1479 $this->unload( $code );
1480 }
1481 }
1482 }