Add special pages aliases for Luri (lrc) from translatewiki
[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 $this->loadItem( $code, $key );
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 string $code
273 * @param string $key
274 * @param string $subkey
275 * @return mixed|null
276 */
277 public function getSubitem( $code, $key, $subkey ) {
278 if ( !isset( $this->loadedSubitems[$code][$key][$subkey] ) &&
279 !isset( $this->loadedItems[$code][$key] )
280 ) {
281 $this->loadSubitem( $code, $key, $subkey );
282 }
283
284 if ( isset( $this->data[$code][$key][$subkey] ) ) {
285 return $this->data[$code][$key][$subkey];
286 } else {
287 return null;
288 }
289 }
290
291 /**
292 * Get the list of subitem keys for a given item.
293 *
294 * This is faster than array_keys($lc->getItem(...)) for the items listed in
295 * self::$splitKeys.
296 *
297 * Will return null if the item is not found, or false if the item is not an
298 * array.
299 * @param string $code
300 * @param string $key
301 * @return bool|null|string
302 */
303 public function getSubitemList( $code, $key ) {
304 if ( in_array( $key, self::$splitKeys ) ) {
305 return $this->getSubitem( $code, 'list', $key );
306 } else {
307 $item = $this->getItem( $code, $key );
308 if ( is_array( $item ) ) {
309 return array_keys( $item );
310 } else {
311 return false;
312 }
313 }
314 }
315
316 /**
317 * Load an item into the cache.
318 * @param string $code
319 * @param string $key
320 */
321 protected function loadItem( $code, $key ) {
322 if ( !isset( $this->initialisedLangs[$code] ) ) {
323 $this->initLanguage( $code );
324 }
325
326 // Check to see if initLanguage() loaded it for us
327 if ( isset( $this->loadedItems[$code][$key] ) ) {
328 return;
329 }
330
331 if ( isset( $this->shallowFallbacks[$code] ) ) {
332 $this->loadItem( $this->shallowFallbacks[$code], $key );
333
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 string $code
355 * @param string $key
356 * @param string $subkey
357 */
358 protected function loadSubitem( $code, $key, $subkey ) {
359 if ( !in_array( $key, self::$splitKeys ) ) {
360 $this->loadItem( $code, $key );
361
362 return;
363 }
364
365 if ( !isset( $this->initialisedLangs[$code] ) ) {
366 $this->initLanguage( $code );
367 }
368
369 // Check to see if initLanguage() loaded it for us
370 if ( isset( $this->loadedItems[$code][$key] ) ||
371 isset( $this->loadedSubitems[$code][$key][$subkey] )
372 ) {
373 return;
374 }
375
376 if ( isset( $this->shallowFallbacks[$code] ) ) {
377 $this->loadSubitem( $this->shallowFallbacks[$code], $key, $subkey );
378
379 return;
380 }
381
382 $value = $this->store->get( $code, "$key:$subkey" );
383 $this->data[$code][$key][$subkey] = $value;
384 $this->loadedSubitems[$code][$key][$subkey] = true;
385 }
386
387 /**
388 * Returns true if the cache identified by $code is missing or expired.
389 *
390 * @param string $code
391 *
392 * @return bool
393 */
394 public function isExpired( $code ) {
395 if ( $this->forceRecache && !isset( $this->recachedLangs[$code] ) ) {
396 wfDebug( __METHOD__ . "($code): forced reload\n" );
397
398 return true;
399 }
400
401 $deps = $this->store->get( $code, 'deps' );
402 $keys = $this->store->get( $code, 'list' );
403 $preload = $this->store->get( $code, 'preload' );
404 // Different keys may expire separately for some stores
405 if ( $deps === null || $keys === null || $preload === null ) {
406 wfDebug( __METHOD__ . "($code): cache missing, need to make one\n" );
407
408 return true;
409 }
410
411 foreach ( $deps as $dep ) {
412 // Because we're unserializing stuff from cache, we
413 // could receive objects of classes that don't exist
414 // anymore (e.g. uninstalled extensions)
415 // When this happens, always expire the cache
416 if ( !$dep instanceof CacheDependency || $dep->isExpired() ) {
417 wfDebug( __METHOD__ . "($code): cache for $code expired due to " .
418 get_class( $dep ) . "\n" );
419
420 return true;
421 }
422 }
423
424 return false;
425 }
426
427 /**
428 * Initialise a language in this object. Rebuild the cache if necessary.
429 * @param string $code
430 * @throws MWException
431 */
432 protected function initLanguage( $code ) {
433 if ( isset( $this->initialisedLangs[$code] ) ) {
434 return;
435 }
436
437 $this->initialisedLangs[$code] = true;
438
439 # If the code is of the wrong form for a Messages*.php file, do a shallow fallback
440 if ( !Language::isValidBuiltInCode( $code ) ) {
441 $this->initShallowFallback( $code, 'en' );
442
443 return;
444 }
445
446 # Recache the data if necessary
447 if ( !$this->manualRecache && $this->isExpired( $code ) ) {
448 if ( Language::isSupportedLanguage( $code ) ) {
449 $this->recache( $code );
450 } elseif ( $code === 'en' ) {
451 throw new MWException( 'MessagesEn.php is missing.' );
452 } else {
453 $this->initShallowFallback( $code, 'en' );
454 }
455
456 return;
457 }
458
459 # Preload some stuff
460 $preload = $this->getItem( $code, 'preload' );
461 if ( $preload === null ) {
462 if ( $this->manualRecache ) {
463 // No Messages*.php file. Do shallow fallback to en.
464 if ( $code === 'en' ) {
465 throw new MWException( 'No localisation cache found for English. ' .
466 'Please run maintenance/rebuildLocalisationCache.php.' );
467 }
468 $this->initShallowFallback( $code, 'en' );
469
470 return;
471 } else {
472 throw new MWException( 'Invalid or missing localisation cache.' );
473 }
474 }
475 $this->data[$code] = $preload;
476 foreach ( $preload as $key => $item ) {
477 if ( in_array( $key, self::$splitKeys ) ) {
478 foreach ( $item as $subkey => $subitem ) {
479 $this->loadedSubitems[$code][$key][$subkey] = true;
480 }
481 } else {
482 $this->loadedItems[$code][$key] = true;
483 }
484 }
485 }
486
487 /**
488 * Create a fallback from one language to another, without creating a
489 * complete persistent cache.
490 * @param string $primaryCode
491 * @param string $fallbackCode
492 */
493 public function initShallowFallback( $primaryCode, $fallbackCode ) {
494 $this->data[$primaryCode] =& $this->data[$fallbackCode];
495 $this->loadedItems[$primaryCode] =& $this->loadedItems[$fallbackCode];
496 $this->loadedSubitems[$primaryCode] =& $this->loadedSubitems[$fallbackCode];
497 $this->shallowFallbacks[$primaryCode] = $fallbackCode;
498 }
499
500 /**
501 * Read a PHP file containing localisation data.
502 * @param string $_fileName
503 * @param string $_fileType
504 * @throws MWException
505 * @return array
506 */
507 protected function readPHPFile( $_fileName, $_fileType ) {
508 // Disable APC caching
509 wfSuppressWarnings();
510 $_apcEnabled = ini_set( 'apc.cache_by_default', '0' );
511 wfRestoreWarnings();
512
513 include $_fileName;
514
515 wfSuppressWarnings();
516 ini_set( 'apc.cache_by_default', $_apcEnabled );
517 wfRestoreWarnings();
518
519 if ( $_fileType == 'core' || $_fileType == 'extension' ) {
520 $data = compact( self::$allKeys );
521 } elseif ( $_fileType == 'aliases' ) {
522 $data = compact( 'aliases' );
523 } else {
524 throw new MWException( __METHOD__ . ": Invalid file type: $_fileType" );
525 }
526
527 return $data;
528 }
529
530 /**
531 * Read a JSON file containing localisation messages.
532 * @param string $fileName Name of file to read
533 * @throws MWException If there is a syntax error in the JSON file
534 * @return array Array with a 'messages' key, or empty array if the file doesn't exist
535 */
536 public function readJSONFile( $fileName ) {
537
538 if ( !is_readable( $fileName ) ) {
539
540 return array();
541 }
542
543 $json = file_get_contents( $fileName );
544 if ( $json === false ) {
545
546 return array();
547 }
548
549 $data = FormatJson::decode( $json, true );
550 if ( $data === null ) {
551
552 throw new MWException( __METHOD__ . ": Invalid JSON file: $fileName" );
553 }
554
555 // Remove keys starting with '@', they're reserved for metadata and non-message data
556 foreach ( $data as $key => $unused ) {
557 if ( $key === '' || $key[0] === '@' ) {
558 unset( $data[$key] );
559 }
560 }
561
562 // The JSON format only supports messages, none of the other variables, so wrap the data
563 return array( 'messages' => $data );
564 }
565
566 /**
567 * Get the compiled plural rules for a given language from the XML files.
568 * @since 1.20
569 * @param string $code
570 * @return array|null
571 */
572 public function getCompiledPluralRules( $code ) {
573 $rules = $this->getPluralRules( $code );
574 if ( $rules === null ) {
575 return null;
576 }
577 try {
578 $compiledRules = CLDRPluralRuleEvaluator::compile( $rules );
579 } catch ( CLDRPluralRuleError $e ) {
580 wfDebugLog( 'l10n', $e->getMessage() );
581
582 return array();
583 }
584
585 return $compiledRules;
586 }
587
588 /**
589 * Get the plural rules for a given language from the XML files.
590 * Cached.
591 * @since 1.20
592 * @param string $code
593 * @return array|null
594 */
595 public function getPluralRules( $code ) {
596 if ( $this->pluralRules === null ) {
597 $this->loadPluralFiles();
598 }
599 if ( !isset( $this->pluralRules[$code] ) ) {
600 return null;
601 } else {
602 return $this->pluralRules[$code];
603 }
604 }
605
606 /**
607 * Get the plural rule types for a given language from the XML files.
608 * Cached.
609 * @since 1.22
610 * @param string $code
611 * @return array|null
612 */
613 public function getPluralRuleTypes( $code ) {
614 if ( $this->pluralRuleTypes === null ) {
615 $this->loadPluralFiles();
616 }
617 if ( !isset( $this->pluralRuleTypes[$code] ) ) {
618 return null;
619 } else {
620 return $this->pluralRuleTypes[$code];
621 }
622 }
623
624 /**
625 * Load the plural XML files.
626 */
627 protected function loadPluralFiles() {
628 global $IP;
629 $cldrPlural = "$IP/languages/data/plurals.xml";
630 $mwPlural = "$IP/languages/data/plurals-mediawiki.xml";
631 // Load CLDR plural rules
632 $this->loadPluralFile( $cldrPlural );
633 if ( file_exists( $mwPlural ) ) {
634 // Override or extend
635 $this->loadPluralFile( $mwPlural );
636 }
637 }
638
639 /**
640 * Load a plural XML file with the given filename, compile the relevant
641 * rules, and save the compiled rules in a process-local cache.
642 *
643 * @param string $fileName
644 * @throws MWException
645 */
646 protected function loadPluralFile( $fileName ) {
647 // Use file_get_contents instead of DOMDocument::load (T58439)
648 $xml = file_get_contents( $fileName );
649 if ( !$xml ) {
650 throw new MWException( "Unable to read plurals file $fileName" );
651 }
652 $doc = new DOMDocument;
653 $doc->loadXML( $xml );
654 $rulesets = $doc->getElementsByTagName( "pluralRules" );
655 foreach ( $rulesets as $ruleset ) {
656 $codes = $ruleset->getAttribute( 'locales' );
657 $rules = array();
658 $ruleTypes = array();
659 $ruleElements = $ruleset->getElementsByTagName( "pluralRule" );
660 foreach ( $ruleElements as $elt ) {
661 $ruleType = $elt->getAttribute( 'count' );
662 if ( $ruleType === 'other' ) {
663 // Don't record "other" rules, which have an empty condition
664 continue;
665 }
666 $rules[] = $elt->nodeValue;
667 $ruleTypes[] = $ruleType;
668 }
669 foreach ( explode( ' ', $codes ) as $code ) {
670 $this->pluralRules[$code] = $rules;
671 $this->pluralRuleTypes[$code] = $ruleTypes;
672 }
673 }
674 }
675
676 /**
677 * Read the data from the source files for a given language, and register
678 * the relevant dependencies in the $deps array. If the localisation
679 * exists, the data array is returned, otherwise false is returned.
680 *
681 * @param string $code
682 * @param array $deps
683 * @return array
684 */
685 protected function readSourceFilesAndRegisterDeps( $code, &$deps ) {
686 global $IP;
687
688 // This reads in the PHP i18n file with non-messages l10n data
689 $fileName = Language::getMessagesFileName( $code );
690 if ( !file_exists( $fileName ) ) {
691 $data = array();
692 } else {
693 $deps[] = new FileDependency( $fileName );
694 $data = $this->readPHPFile( $fileName, 'core' );
695 }
696
697 # Load CLDR plural rules for JavaScript
698 $data['pluralRules'] = $this->getPluralRules( $code );
699 # And for PHP
700 $data['compiledPluralRules'] = $this->getCompiledPluralRules( $code );
701 # Load plural rule types
702 $data['pluralRuleTypes'] = $this->getPluralRuleTypes( $code );
703
704 $deps['plurals'] = new FileDependency( "$IP/languages/data/plurals.xml" );
705 $deps['plurals-mw'] = new FileDependency( "$IP/languages/data/plurals-mediawiki.xml" );
706
707 return $data;
708 }
709
710 /**
711 * Merge two localisation values, a primary and a fallback, overwriting the
712 * primary value in place.
713 * @param string $key
714 * @param mixed $value
715 * @param mixed $fallbackValue
716 */
717 protected function mergeItem( $key, &$value, $fallbackValue ) {
718 if ( !is_null( $value ) ) {
719 if ( !is_null( $fallbackValue ) ) {
720 if ( in_array( $key, self::$mergeableMapKeys ) ) {
721 $value = $value + $fallbackValue;
722 } elseif ( in_array( $key, self::$mergeableListKeys ) ) {
723 $value = array_unique( array_merge( $fallbackValue, $value ) );
724 } elseif ( in_array( $key, self::$mergeableAliasListKeys ) ) {
725 $value = array_merge_recursive( $value, $fallbackValue );
726 } elseif ( in_array( $key, self::$optionalMergeKeys ) ) {
727 if ( !empty( $value['inherit'] ) ) {
728 $value = array_merge( $fallbackValue, $value );
729 }
730
731 if ( isset( $value['inherit'] ) ) {
732 unset( $value['inherit'] );
733 }
734 } elseif ( in_array( $key, self::$magicWordKeys ) ) {
735 $this->mergeMagicWords( $value, $fallbackValue );
736 }
737 }
738 } else {
739 $value = $fallbackValue;
740 }
741 }
742
743 /**
744 * @param mixed $value
745 * @param mixed $fallbackValue
746 */
747 protected function mergeMagicWords( &$value, $fallbackValue ) {
748 foreach ( $fallbackValue as $magicName => $fallbackInfo ) {
749 if ( !isset( $value[$magicName] ) ) {
750 $value[$magicName] = $fallbackInfo;
751 } else {
752 $oldSynonyms = array_slice( $fallbackInfo, 1 );
753 $newSynonyms = array_slice( $value[$magicName], 1 );
754 $synonyms = array_values( array_unique( array_merge(
755 $newSynonyms, $oldSynonyms ) ) );
756 $value[$magicName] = array_merge( array( $fallbackInfo[0] ), $synonyms );
757 }
758 }
759 }
760
761 /**
762 * Given an array mapping language code to localisation value, such as is
763 * found in extension *.i18n.php files, iterate through a fallback sequence
764 * to merge the given data with an existing primary value.
765 *
766 * Returns true if any data from the extension array was used, false
767 * otherwise.
768 * @param array $codeSequence
769 * @param string $key
770 * @param mixed $value
771 * @param mixed $fallbackValue
772 * @return bool
773 */
774 protected function mergeExtensionItem( $codeSequence, $key, &$value, $fallbackValue ) {
775 $used = false;
776 foreach ( $codeSequence as $code ) {
777 if ( isset( $fallbackValue[$code] ) ) {
778 $this->mergeItem( $key, $value, $fallbackValue[$code] );
779 $used = true;
780 }
781 }
782
783 return $used;
784 }
785
786 /**
787 * Gets the combined list of messages dirs from
788 * core and extensions
789 *
790 * @since 1.25
791 * @return array
792 */
793 public function getMessagesDirs() {
794 global $wgMessagesDirs, $IP;
795 return array(
796 'core' => "$IP/languages/i18n",
797 'api' => "$IP/includes/api/i18n",
798 'oojs-ui' => "$IP/resources/lib/oojs-ui/i18n",
799 ) + $wgMessagesDirs;
800 }
801
802 /**
803 * Load localisation data for a given language for both core and extensions
804 * and save it to the persistent cache store and the process cache
805 * @param string $code
806 * @throws MWException
807 */
808 public function recache( $code ) {
809 global $wgExtensionMessagesFiles;
810
811 if ( !$code ) {
812 throw new MWException( "Invalid language code requested" );
813 }
814 $this->recachedLangs[$code] = true;
815
816 # Initial values
817 $initialData = array_combine(
818 self::$allKeys,
819 array_fill( 0, count( self::$allKeys ), null ) );
820 $coreData = $initialData;
821 $deps = array();
822
823 # Load the primary localisation from the source file
824 $data = $this->readSourceFilesAndRegisterDeps( $code, $deps );
825 if ( $data === false ) {
826 wfDebug( __METHOD__ . ": no localisation file for $code, using fallback to en\n" );
827 $coreData['fallback'] = 'en';
828 } else {
829 wfDebug( __METHOD__ . ": got localisation for $code from source\n" );
830
831 # Merge primary localisation
832 foreach ( $data as $key => $value ) {
833 $this->mergeItem( $key, $coreData[$key], $value );
834 }
835 }
836
837 # Fill in the fallback if it's not there already
838 if ( is_null( $coreData['fallback'] ) ) {
839 $coreData['fallback'] = $code === 'en' ? false : 'en';
840 }
841 if ( $coreData['fallback'] === false ) {
842 $coreData['fallbackSequence'] = array();
843 } else {
844 $coreData['fallbackSequence'] = array_map( 'trim', explode( ',', $coreData['fallback'] ) );
845 $len = count( $coreData['fallbackSequence'] );
846
847 # Ensure that the sequence ends at en
848 if ( $coreData['fallbackSequence'][$len - 1] !== 'en' ) {
849 $coreData['fallbackSequence'][] = 'en';
850 }
851 }
852
853 $codeSequence = array_merge( array( $code ), $coreData['fallbackSequence'] );
854 $messageDirs = $this->getMessagesDirs();
855
856 # Load non-JSON localisation data for extensions
857 $extensionData = array_combine(
858 $codeSequence,
859 array_fill( 0, count( $codeSequence ), $initialData ) );
860 foreach ( $wgExtensionMessagesFiles as $extension => $fileName ) {
861 if ( isset( $messageDirs[$extension] ) ) {
862 # This extension has JSON message data; skip the PHP shim
863 continue;
864 }
865
866 $data = $this->readPHPFile( $fileName, 'extension' );
867 $used = false;
868
869 foreach ( $data as $key => $item ) {
870 foreach ( $codeSequence as $csCode ) {
871 if ( isset( $item[$csCode] ) ) {
872 $this->mergeItem( $key, $extensionData[$csCode][$key], $item[$csCode] );
873 $used = true;
874 }
875 }
876 }
877
878 if ( $used ) {
879 $deps[] = new FileDependency( $fileName );
880 }
881 }
882
883 # Load the localisation data for each fallback, then merge it into the full array
884 $allData = $initialData;
885 foreach ( $codeSequence as $csCode ) {
886 $csData = $initialData;
887
888 # Load core messages and the extension localisations.
889 foreach ( $messageDirs as $dirs ) {
890 foreach ( (array)$dirs as $dir ) {
891 $fileName = "$dir/$csCode.json";
892 $data = $this->readJSONFile( $fileName );
893
894 foreach ( $data as $key => $item ) {
895 $this->mergeItem( $key, $csData[$key], $item );
896 }
897
898 $deps[] = new FileDependency( $fileName );
899 }
900 }
901
902 # Merge non-JSON extension data
903 if ( isset( $extensionData[$csCode] ) ) {
904 foreach ( $extensionData[$csCode] as $key => $item ) {
905 $this->mergeItem( $key, $csData[$key], $item );
906 }
907 }
908
909 if ( $csCode === $code ) {
910 # Merge core data into extension data
911 foreach ( $coreData as $key => $item ) {
912 $this->mergeItem( $key, $csData[$key], $item );
913 }
914 } else {
915 # Load the secondary localisation from the source file to
916 # avoid infinite cycles on cyclic fallbacks
917 $fbData = $this->readSourceFilesAndRegisterDeps( $csCode, $deps );
918 if ( $fbData !== false ) {
919 # Only merge the keys that make sense to merge
920 foreach ( self::$allKeys as $key ) {
921 if ( !isset( $fbData[$key] ) ) {
922 continue;
923 }
924
925 if ( is_null( $coreData[$key] ) || $this->isMergeableKey( $key ) ) {
926 $this->mergeItem( $key, $csData[$key], $fbData[$key] );
927 }
928 }
929 }
930 }
931
932 # Allow extensions an opportunity to adjust the data for this
933 # fallback
934 Hooks::run( 'LocalisationCacheRecacheFallback', array( $this, $csCode, &$csData ) );
935
936 # Merge the data for this fallback into the final array
937 if ( $csCode === $code ) {
938 $allData = $csData;
939 } else {
940 foreach ( self::$allKeys as $key ) {
941 if ( !isset( $csData[$key] ) ) {
942 continue;
943 }
944
945 if ( is_null( $allData[$key] ) || $this->isMergeableKey( $key ) ) {
946 $this->mergeItem( $key, $allData[$key], $csData[$key] );
947 }
948 }
949 }
950 }
951
952 # Add cache dependencies for any referenced globals
953 $deps['wgExtensionMessagesFiles'] = new GlobalDependency( 'wgExtensionMessagesFiles' );
954 // $wgMessagesDirs is used in LocalisationCache::getMessagesDirs()
955 $deps['wgMessagesDirs'] = new GlobalDependency( 'wgMessagesDirs' );
956 $deps['version'] = new ConstantDependency( 'LocalisationCache::VERSION' );
957
958 # Add dependencies to the cache entry
959 $allData['deps'] = $deps;
960
961 # Replace spaces with underscores in namespace names
962 $allData['namespaceNames'] = str_replace( ' ', '_', $allData['namespaceNames'] );
963
964 # And do the same for special page aliases. $page is an array.
965 foreach ( $allData['specialPageAliases'] as &$page ) {
966 $page = str_replace( ' ', '_', $page );
967 }
968 # Decouple the reference to prevent accidental damage
969 unset( $page );
970
971 # If there were no plural rules, return an empty array
972 if ( $allData['pluralRules'] === null ) {
973 $allData['pluralRules'] = array();
974 }
975 if ( $allData['compiledPluralRules'] === null ) {
976 $allData['compiledPluralRules'] = array();
977 }
978 # If there were no plural rule types, return an empty array
979 if ( $allData['pluralRuleTypes'] === null ) {
980 $allData['pluralRuleTypes'] = array();
981 }
982
983 # Set the list keys
984 $allData['list'] = array();
985 foreach ( self::$splitKeys as $key ) {
986 $allData['list'][$key] = array_keys( $allData[$key] );
987 }
988 # Run hooks
989 $purgeBlobs = true;
990 Hooks::run( 'LocalisationCacheRecache', array( $this, $code, &$allData, &$purgeBlobs ) );
991
992 if ( is_null( $allData['namespaceNames'] ) ) {
993 throw new MWException( __METHOD__ . ': Localisation data failed sanity check! ' .
994 'Check that your languages/messages/MessagesEn.php file is intact.' );
995 }
996
997 # Set the preload key
998 $allData['preload'] = $this->buildPreload( $allData );
999
1000 # Save to the process cache and register the items loaded
1001 $this->data[$code] = $allData;
1002 foreach ( $allData as $key => $item ) {
1003 $this->loadedItems[$code][$key] = true;
1004 }
1005
1006 # Save to the persistent cache
1007 $this->store->startWrite( $code );
1008 foreach ( $allData as $key => $value ) {
1009 if ( in_array( $key, self::$splitKeys ) ) {
1010 foreach ( $value as $subkey => $subvalue ) {
1011 $this->store->set( "$key:$subkey", $subvalue );
1012 }
1013 } else {
1014 $this->store->set( $key, $value );
1015 }
1016 }
1017 $this->store->finishWrite();
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 }
1027
1028 /**
1029 * Build the preload item from the given pre-cache data.
1030 *
1031 * The preload item will be loaded automatically, improving performance
1032 * for the commonly-requested items it contains.
1033 * @param array $data
1034 * @return array
1035 */
1036 protected function buildPreload( $data ) {
1037 $preload = array( 'messages' => array() );
1038 foreach ( self::$preloadedKeys as $key ) {
1039 $preload[$key] = $data[$key];
1040 }
1041
1042 foreach ( $data['preloadedMessages'] as $subkey ) {
1043 if ( isset( $data['messages'][$subkey] ) ) {
1044 $subitem = $data['messages'][$subkey];
1045 } else {
1046 $subitem = null;
1047 }
1048 $preload['messages'][$subkey] = $subitem;
1049 }
1050
1051 return $preload;
1052 }
1053
1054 /**
1055 * Unload the data for a given language from the object cache.
1056 * Reduces memory usage.
1057 * @param string $code
1058 */
1059 public function unload( $code ) {
1060 unset( $this->data[$code] );
1061 unset( $this->loadedItems[$code] );
1062 unset( $this->loadedSubitems[$code] );
1063 unset( $this->initialisedLangs[$code] );
1064 unset( $this->shallowFallbacks[$code] );
1065
1066 foreach ( $this->shallowFallbacks as $shallowCode => $fbCode ) {
1067 if ( $fbCode === $code ) {
1068 $this->unload( $shallowCode );
1069 }
1070 }
1071 }
1072
1073 /**
1074 * Unload all data
1075 */
1076 public function unloadAll() {
1077 foreach ( $this->initialisedLangs as $lang => $unused ) {
1078 $this->unload( $lang );
1079 }
1080 }
1081
1082 /**
1083 * Disable the storage backend
1084 */
1085 public function disableBackend() {
1086 $this->store = new LCStoreNull;
1087 $this->manualRecache = false;
1088 }
1089 }
1090
1091 /**
1092 * Interface for the persistence layer of LocalisationCache.
1093 *
1094 * The persistence layer is two-level hierarchical cache. The first level
1095 * is the language, the second level is the item or subitem.
1096 *
1097 * Since the data for a whole language is rebuilt in one operation, it needs
1098 * to have a fast and atomic method for deleting or replacing all of the
1099 * current data for a given language. The interface reflects this bulk update
1100 * operation. Callers writing to the cache must first call startWrite(), then
1101 * will call set() a couple of thousand times, then will call finishWrite()
1102 * to commit the operation. When finishWrite() is called, the cache is
1103 * expected to delete all data previously stored for that language.
1104 *
1105 * The values stored are PHP variables suitable for serialize(). Implementations
1106 * of LCStore are responsible for serializing and unserializing.
1107 */
1108 interface LCStore {
1109 /**
1110 * Get a value.
1111 * @param string $code Language code
1112 * @param string $key Cache key
1113 */
1114 function get( $code, $key );
1115
1116 /**
1117 * Start a write transaction.
1118 * @param string $code Language code
1119 */
1120 function startWrite( $code );
1121
1122 /**
1123 * Finish a write transaction.
1124 */
1125 function finishWrite();
1126
1127 /**
1128 * Set a key to a given value. startWrite() must be called before this
1129 * is called, and finishWrite() must be called afterwards.
1130 * @param string $key
1131 * @param mixed $value
1132 */
1133 function set( $key, $value );
1134 }
1135
1136 /**
1137 * LCStore implementation which uses the standard DB functions to store data.
1138 * This will work on any MediaWiki installation.
1139 */
1140 class LCStoreDB implements LCStore {
1141 private $currentLang;
1142 private $writesDone = false;
1143
1144 /** @var DatabaseBase */
1145 private $dbw;
1146 /** @var array */
1147 private $batch = array();
1148
1149 private $readOnly = false;
1150
1151 public function get( $code, $key ) {
1152 if ( $this->writesDone ) {
1153 $db = wfGetDB( DB_MASTER );
1154 } else {
1155 $db = wfGetDB( DB_SLAVE );
1156 }
1157 $row = $db->selectRow( 'l10n_cache', array( 'lc_value' ),
1158 array( 'lc_lang' => $code, 'lc_key' => $key ), __METHOD__ );
1159 if ( $row ) {
1160 return unserialize( $db->decodeBlob( $row->lc_value ) );
1161 } else {
1162 return null;
1163 }
1164 }
1165
1166 public function startWrite( $code ) {
1167 if ( $this->readOnly ) {
1168 return;
1169 } elseif ( !$code ) {
1170 throw new MWException( __METHOD__ . ": Invalid language \"$code\"" );
1171 }
1172
1173 $this->dbw = wfGetDB( DB_MASTER );
1174
1175 $this->currentLang = $code;
1176 $this->batch = array();
1177 }
1178
1179 public function finishWrite() {
1180 if ( $this->readOnly ) {
1181 return;
1182 } elseif ( is_null( $this->currentLang ) ) {
1183 throw new MWException( __CLASS__ . ': must call startWrite() before finishWrite()' );
1184 }
1185
1186 $this->dbw->begin( __METHOD__ );
1187 try {
1188 $this->dbw->delete( 'l10n_cache',
1189 array( 'lc_lang' => $this->currentLang ), __METHOD__ );
1190 foreach ( array_chunk( $this->batch, 500 ) as $rows ) {
1191 $this->dbw->insert( 'l10n_cache', $rows, __METHOD__ );
1192 }
1193 $this->writesDone = true;
1194 } catch ( DBQueryError $e ) {
1195 if ( $this->dbw->wasReadOnlyError() ) {
1196 $this->readOnly = true; // just avoid site down time
1197 } else {
1198 throw $e;
1199 }
1200 }
1201 $this->dbw->commit( __METHOD__ );
1202
1203 $this->currentLang = null;
1204 $this->batch = array();
1205 }
1206
1207 public function set( $key, $value ) {
1208 if ( $this->readOnly ) {
1209 return;
1210 } elseif ( is_null( $this->currentLang ) ) {
1211 throw new MWException( __CLASS__ . ': must call startWrite() before set()' );
1212 }
1213
1214 $this->batch[] = array(
1215 'lc_lang' => $this->currentLang,
1216 'lc_key' => $key,
1217 'lc_value' => $this->dbw->encodeBlob( serialize( $value ) ) );
1218 }
1219 }
1220
1221 /**
1222 * LCStore implementation which stores data as a collection of CDB files in the
1223 * directory given by $wgCacheDirectory. If $wgCacheDirectory is not set, this
1224 * will throw an exception.
1225 *
1226 * Profiling indicates that on Linux, this implementation outperforms MySQL if
1227 * the directory is on a local filesystem and there is ample kernel cache
1228 * space. The performance advantage is greater when the DBA extension is
1229 * available than it is with the PHP port.
1230 *
1231 * See Cdb.php and http://cr.yp.to/cdb.html
1232 */
1233 class LCStoreCDB implements LCStore {
1234 /** @var CdbReader[] */
1235 private $readers;
1236
1237 /** @var CdbWriter */
1238 private $writer;
1239
1240 /** @var string Current language code */
1241 private $currentLang;
1242
1243 /** @var bool|string Cache directory. False if not set */
1244 private $directory;
1245
1246 function __construct( $conf = array() ) {
1247 global $wgCacheDirectory;
1248
1249 if ( isset( $conf['directory'] ) ) {
1250 $this->directory = $conf['directory'];
1251 } else {
1252 $this->directory = $wgCacheDirectory;
1253 }
1254 }
1255
1256 public function get( $code, $key ) {
1257 if ( !isset( $this->readers[$code] ) ) {
1258 $fileName = $this->getFileName( $code );
1259
1260 $this->readers[$code] = false;
1261 if ( file_exists( $fileName ) ) {
1262 try {
1263 $this->readers[$code] = CdbReader::open( $fileName );
1264 } catch ( CdbException $e ) {
1265 wfDebug( __METHOD__ . ": unable to open cdb file for reading\n" );
1266 }
1267 }
1268 }
1269
1270 if ( !$this->readers[$code] ) {
1271 return null;
1272 } else {
1273 $value = false;
1274 try {
1275 $value = $this->readers[$code]->get( $key );
1276 } catch ( CdbException $e ) {
1277 wfDebug( __METHOD__ . ": CdbException caught, error message was "
1278 . $e->getMessage() . "\n" );
1279 }
1280 if ( $value === false ) {
1281 return null;
1282 }
1283
1284 return unserialize( $value );
1285 }
1286 }
1287
1288 public function startWrite( $code ) {
1289 if ( !file_exists( $this->directory ) ) {
1290 if ( !wfMkdirParents( $this->directory, null, __METHOD__ ) ) {
1291 throw new MWException( "Unable to create the localisation store " .
1292 "directory \"{$this->directory}\"" );
1293 }
1294 }
1295
1296 // Close reader to stop permission errors on write
1297 if ( !empty( $this->readers[$code] ) ) {
1298 $this->readers[$code]->close();
1299 }
1300
1301 try {
1302 $this->writer = CdbWriter::open( $this->getFileName( $code ) );
1303 } catch ( CdbException $e ) {
1304 throw new MWException( $e->getMessage() );
1305 }
1306 $this->currentLang = $code;
1307 }
1308
1309 public function finishWrite() {
1310 // Close the writer
1311 try {
1312 $this->writer->close();
1313 } catch ( CdbException $e ) {
1314 throw new MWException( $e->getMessage() );
1315 }
1316 $this->writer = null;
1317 unset( $this->readers[$this->currentLang] );
1318 $this->currentLang = null;
1319 }
1320
1321 public function set( $key, $value ) {
1322 if ( is_null( $this->writer ) ) {
1323 throw new MWException( __CLASS__ . ': must call startWrite() before calling set()' );
1324 }
1325 try {
1326 $this->writer->set( $key, serialize( $value ) );
1327 } catch ( CdbException $e ) {
1328 throw new MWException( $e->getMessage() );
1329 }
1330 }
1331
1332 protected function getFileName( $code ) {
1333 if ( strval( $code ) === '' || strpos( $code, '/' ) !== false ) {
1334 throw new MWException( __METHOD__ . ": Invalid language \"$code\"" );
1335 }
1336
1337 return "{$this->directory}/l10n_cache-$code.cdb";
1338 }
1339 }
1340
1341 /**
1342 * Null store backend, used to avoid DB errors during install
1343 */
1344 class LCStoreNull implements LCStore {
1345 public function get( $code, $key ) {
1346 return null;
1347 }
1348
1349 public function startWrite( $code ) {
1350 }
1351
1352 public function finishWrite() {
1353 }
1354
1355 public function set( $key, $value ) {
1356 }
1357 }
1358
1359 /**
1360 * A localisation cache optimised for loading large amounts of data for many
1361 * languages. Used by rebuildLocalisationCache.php.
1362 */
1363 class LocalisationCacheBulkLoad extends LocalisationCache {
1364 /**
1365 * A cache of the contents of data files.
1366 * Core files are serialized to avoid using ~1GB of RAM during a recache.
1367 */
1368 private $fileCache = array();
1369
1370 /**
1371 * Most recently used languages. Uses the linked-list aspect of PHP hashtables
1372 * to keep the most recently used language codes at the end of the array, and
1373 * the language codes that are ready to be deleted at the beginning.
1374 */
1375 private $mruLangs = array();
1376
1377 /**
1378 * Maximum number of languages that may be loaded into $this->data
1379 */
1380 private $maxLoadedLangs = 10;
1381
1382 /**
1383 * @param string $fileName
1384 * @param string $fileType
1385 * @return array|mixed
1386 */
1387 protected function readPHPFile( $fileName, $fileType ) {
1388 $serialize = $fileType === 'core';
1389 if ( !isset( $this->fileCache[$fileName][$fileType] ) ) {
1390 $data = parent::readPHPFile( $fileName, $fileType );
1391
1392 if ( $serialize ) {
1393 $encData = serialize( $data );
1394 } else {
1395 $encData = $data;
1396 }
1397
1398 $this->fileCache[$fileName][$fileType] = $encData;
1399
1400 return $data;
1401 } elseif ( $serialize ) {
1402 return unserialize( $this->fileCache[$fileName][$fileType] );
1403 } else {
1404 return $this->fileCache[$fileName][$fileType];
1405 }
1406 }
1407
1408 /**
1409 * @param string $code
1410 * @param string $key
1411 * @return mixed
1412 */
1413 public function getItem( $code, $key ) {
1414 unset( $this->mruLangs[$code] );
1415 $this->mruLangs[$code] = true;
1416
1417 return parent::getItem( $code, $key );
1418 }
1419
1420 /**
1421 * @param string $code
1422 * @param string $key
1423 * @param string $subkey
1424 * @return mixed
1425 */
1426 public function getSubitem( $code, $key, $subkey ) {
1427 unset( $this->mruLangs[$code] );
1428 $this->mruLangs[$code] = true;
1429
1430 return parent::getSubitem( $code, $key, $subkey );
1431 }
1432
1433 /**
1434 * @param string $code
1435 */
1436 public function recache( $code ) {
1437 parent::recache( $code );
1438 unset( $this->mruLangs[$code] );
1439 $this->mruLangs[$code] = true;
1440 $this->trimCache();
1441 }
1442
1443 /**
1444 * @param string $code
1445 */
1446 public function unload( $code ) {
1447 unset( $this->mruLangs[$code] );
1448 parent::unload( $code );
1449 }
1450
1451 /**
1452 * Unload cached languages until there are less than $this->maxLoadedLangs
1453 */
1454 protected function trimCache() {
1455 while ( count( $this->data ) > $this->maxLoadedLangs && count( $this->mruLangs ) ) {
1456 reset( $this->mruLangs );
1457 $code = key( $this->mruLangs );
1458 wfDebug( __METHOD__ . ": unloading $code\n" );
1459 $this->unload( $code );
1460 }
1461 }
1462 }