Merge "Darken the text color in successbox"
[lhc/web/wiklou.git] / includes / cache / LocalisationCache.php
1 <?php
2 /**
3 * Cache of the contents of localisation files.
4 *
5 * This program is free software; you can redistribute it and/or modify
6 * it under the terms of the GNU General Public License as published by
7 * the Free Software Foundation; either version 2 of the License, or
8 * (at your option) any later version.
9 *
10 * This program is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 * GNU General Public License for more details.
14 *
15 * You should have received a copy of the GNU General Public License along
16 * with this program; if not, write to the Free Software Foundation, Inc.,
17 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
18 * http://www.gnu.org/copyleft/gpl.html
19 *
20 * @file
21 */
22
23 use Cdb\Exception as CdbException;
24 use Cdb\Reader as CdbReader;
25 use Cdb\Writer as CdbWriter;
26 /**
27 * Class for caching the contents of localisation files, Messages*.php
28 * and *.i18n.php.
29 *
30 * An instance of this class is available using Language::getLocalisationCache().
31 *
32 * The values retrieved from here are merged, containing items from extension
33 * files, core messages files and the language fallback sequence (e.g. zh-cn ->
34 * zh-hans -> en ). Some common errors are corrected, for example namespace
35 * names with spaces instead of underscores, but heavyweight processing, such
36 * as grammatical transformation, is done by the caller.
37 */
38 class LocalisationCache {
39 const VERSION = 3;
40
41 /** Configuration associative array */
42 private $conf;
43
44 /**
45 * True if recaching should only be done on an explicit call to recache().
46 * Setting this reduces the overhead of cache freshness checking, which
47 * requires doing a stat() for every extension i18n file.
48 */
49 private $manualRecache = false;
50
51 /**
52 * True to treat all files as expired until they are regenerated by this object.
53 */
54 private $forceRecache = false;
55
56 /**
57 * The cache data. 3-d array, where the first key is the language code,
58 * the second key is the item key e.g. 'messages', and the third key is
59 * an item specific subkey index. Some items are not arrays and so for those
60 * items, there are no subkeys.
61 */
62 protected $data = array();
63
64 /**
65 * The persistent store object. An instance of LCStore.
66 *
67 * @var LCStore
68 */
69 private $store;
70
71 /**
72 * A 2-d associative array, code/key, where presence indicates that the item
73 * is loaded. Value arbitrary.
74 *
75 * For split items, if set, this indicates that all of the subitems have been
76 * loaded.
77 */
78 private $loadedItems = array();
79
80 /**
81 * A 3-d associative array, code/key/subkey, where presence indicates that
82 * the subitem is loaded. Only used for the split items, i.e. messages.
83 */
84 private $loadedSubitems = array();
85
86 /**
87 * An array where presence of a key indicates that that language has been
88 * initialised. Initialisation includes checking for cache expiry and doing
89 * any necessary updates.
90 */
91 private $initialisedLangs = array();
92
93 /**
94 * An array mapping non-existent pseudo-languages to fallback languages. This
95 * is filled by initShallowFallback() when data is requested from a language
96 * that lacks a Messages*.php file.
97 */
98 private $shallowFallbacks = array();
99
100 /**
101 * An array where the keys are codes that have been recached by this instance.
102 */
103 private $recachedLangs = array();
104
105 /**
106 * All item keys
107 */
108 static public $allKeys = array(
109 'fallback', 'namespaceNames', 'bookstoreList',
110 'magicWords', 'messages', 'rtl', 'capitalizeAllNouns', 'digitTransformTable',
111 'separatorTransformTable', 'fallback8bitEncoding', 'linkPrefixExtension',
112 'linkTrail', 'linkPrefixCharset', 'namespaceAliases',
113 'dateFormats', 'datePreferences', 'datePreferenceMigrationMap',
114 'defaultDateFormat', 'extraUserToggles', 'specialPageAliases',
115 'imageFiles', 'preloadedMessages', 'namespaceGenderAliases',
116 'digitGroupingPattern', 'pluralRules', 'pluralRuleTypes', 'compiledPluralRules',
117 );
118
119 /**
120 * Keys for items which consist of associative arrays, which may be merged
121 * by a fallback sequence.
122 */
123 static public $mergeableMapKeys = array( 'messages', 'namespaceNames',
124 'dateFormats', 'imageFiles', 'preloadedMessages'
125 );
126
127 /**
128 * Keys for items which are a numbered array.
129 */
130 static public $mergeableListKeys = array( 'extraUserToggles' );
131
132 /**
133 * Keys for items which contain an array of arrays of equivalent aliases
134 * for each subitem. The aliases may be merged by a fallback sequence.
135 */
136 static public $mergeableAliasListKeys = array( 'specialPageAliases' );
137
138 /**
139 * Keys for items which contain an associative array, and may be merged if
140 * the primary value contains the special array key "inherit". That array
141 * key is removed after the first merge.
142 */
143 static public $optionalMergeKeys = array( 'bookstoreList' );
144
145 /**
146 * Keys for items that are formatted like $magicWords
147 */
148 static public $magicWordKeys = array( 'magicWords' );
149
150 /**
151 * Keys for items where the subitems are stored in the backend separately.
152 */
153 static public $splitKeys = array( 'messages' );
154
155 /**
156 * Keys which are loaded automatically by initLanguage()
157 */
158 static public $preloadedKeys = array( 'dateFormats', 'namespaceNames' );
159
160 /**
161 * Associative array of cached plural rules. The key is the language code,
162 * the value is an array of plural rules for that language.
163 */
164 private $pluralRules = null;
165
166 /**
167 * Associative array of cached plural rule types. The key is the language
168 * code, the value is an array of plural rule types for that language. For
169 * example, $pluralRuleTypes['ar'] = ['zero', 'one', 'two', 'few', 'many'].
170 * The index for each rule type matches the index for the rule in
171 * $pluralRules, thus allowing correlation between the two. The reason we
172 * don't just use the type names as the keys in $pluralRules is because
173 * Language::convertPlural applies the rules based on numeric order (or
174 * explicit numeric parameter), not based on the name of the rule type. For
175 * example, {{plural:count|wordform1|wordform2|wordform3}}, rather than
176 * {{plural:count|one=wordform1|two=wordform2|many=wordform3}}.
177 */
178 private $pluralRuleTypes = null;
179
180 private $mergeableKeys = null;
181
182 /**
183 * Constructor.
184 * For constructor parameters, see the documentation in DefaultSettings.php
185 * for $wgLocalisationCacheConf.
186 *
187 * @param array $conf
188 * @throws MWException
189 */
190 function __construct( $conf ) {
191 global $wgCacheDirectory;
192
193 $this->conf = $conf;
194 $storeConf = array();
195 if ( !empty( $conf['storeClass'] ) ) {
196 $storeClass = $conf['storeClass'];
197 } else {
198 switch ( $conf['store'] ) {
199 case 'files':
200 case 'file':
201 $storeClass = 'LCStoreCDB';
202 break;
203 case 'db':
204 $storeClass = 'LCStoreDB';
205 break;
206 case 'detect':
207 $storeClass = $wgCacheDirectory ? 'LCStoreCDB' : 'LCStoreDB';
208 break;
209 default:
210 throw new MWException(
211 'Please set $wgLocalisationCacheConf[\'store\'] to something sensible.' );
212 }
213 }
214
215 wfDebugLog( 'caches', get_class( $this ) . ": using store $storeClass" );
216 if ( !empty( $conf['storeDirectory'] ) ) {
217 $storeConf['directory'] = $conf['storeDirectory'];
218 }
219
220 $this->store = new $storeClass( $storeConf );
221 foreach ( array( 'manualRecache', 'forceRecache' ) as $var ) {
222 if ( isset( $conf[$var] ) ) {
223 $this->$var = $conf[$var];
224 }
225 }
226 }
227
228 /**
229 * Returns true if the given key is mergeable, that is, if it is an associative
230 * array which can be merged through a fallback sequence.
231 * @param string $key
232 * @return bool
233 */
234 public function isMergeableKey( $key ) {
235 if ( $this->mergeableKeys === null ) {
236 $this->mergeableKeys = array_flip( array_merge(
237 self::$mergeableMapKeys,
238 self::$mergeableListKeys,
239 self::$mergeableAliasListKeys,
240 self::$optionalMergeKeys,
241 self::$magicWordKeys
242 ) );
243 }
244
245 return isset( $this->mergeableKeys[$key] );
246 }
247
248 /**
249 * Get a cache item.
250 *
251 * Warning: this may be slow for split items (messages), since it will
252 * need to fetch all of the subitems from the cache individually.
253 * @param string $code
254 * @param string $key
255 * @return mixed
256 */
257 public function getItem( $code, $key ) {
258 if ( !isset( $this->loadedItems[$code][$key] ) ) {
259 wfProfileIn( __METHOD__ . '-load' );
260 $this->loadItem( $code, $key );
261 wfProfileOut( __METHOD__ . '-load' );
262 }
263
264 if ( $key === 'fallback' && isset( $this->shallowFallbacks[$code] ) ) {
265 return $this->shallowFallbacks[$code];
266 }
267
268 return $this->data[$code][$key];
269 }
270
271 /**
272 * Get a subitem, for instance a single message for a given language.
273 * @param string $code
274 * @param string $key
275 * @param string $subkey
276 * @return mixed|null
277 */
278 public function getSubitem( $code, $key, $subkey ) {
279 if ( !isset( $this->loadedSubitems[$code][$key][$subkey] ) &&
280 !isset( $this->loadedItems[$code][$key] )
281 ) {
282 wfProfileIn( __METHOD__ . '-load' );
283 $this->loadSubitem( $code, $key, $subkey );
284 wfProfileOut( __METHOD__ . '-load' );
285 }
286
287 if ( isset( $this->data[$code][$key][$subkey] ) ) {
288 return $this->data[$code][$key][$subkey];
289 } else {
290 return null;
291 }
292 }
293
294 /**
295 * Get the list of subitem keys for a given item.
296 *
297 * This is faster than array_keys($lc->getItem(...)) for the items listed in
298 * self::$splitKeys.
299 *
300 * Will return null if the item is not found, or false if the item is not an
301 * array.
302 * @param string $code
303 * @param string $key
304 * @return bool|null|string
305 */
306 public function getSubitemList( $code, $key ) {
307 if ( in_array( $key, self::$splitKeys ) ) {
308 return $this->getSubitem( $code, 'list', $key );
309 } else {
310 $item = $this->getItem( $code, $key );
311 if ( is_array( $item ) ) {
312 return array_keys( $item );
313 } else {
314 return false;
315 }
316 }
317 }
318
319 /**
320 * Load an item into the cache.
321 * @param string $code
322 * @param string $key
323 */
324 protected function loadItem( $code, $key ) {
325 if ( !isset( $this->initialisedLangs[$code] ) ) {
326 $this->initLanguage( $code );
327 }
328
329 // Check to see if initLanguage() loaded it for us
330 if ( isset( $this->loadedItems[$code][$key] ) ) {
331 return;
332 }
333
334 if ( isset( $this->shallowFallbacks[$code] ) ) {
335 $this->loadItem( $this->shallowFallbacks[$code], $key );
336
337 return;
338 }
339
340 if ( in_array( $key, self::$splitKeys ) ) {
341 $subkeyList = $this->getSubitem( $code, 'list', $key );
342 foreach ( $subkeyList as $subkey ) {
343 if ( isset( $this->data[$code][$key][$subkey] ) ) {
344 continue;
345 }
346 $this->data[$code][$key][$subkey] = $this->getSubitem( $code, $key, $subkey );
347 }
348 } else {
349 $this->data[$code][$key] = $this->store->get( $code, $key );
350 }
351
352 $this->loadedItems[$code][$key] = true;
353 }
354
355 /**
356 * Load a subitem into the cache
357 * @param string $code
358 * @param string $key
359 * @param string $subkey
360 */
361 protected function loadSubitem( $code, $key, $subkey ) {
362 if ( !in_array( $key, self::$splitKeys ) ) {
363 $this->loadItem( $code, $key );
364
365 return;
366 }
367
368 if ( !isset( $this->initialisedLangs[$code] ) ) {
369 $this->initLanguage( $code );
370 }
371
372 // Check to see if initLanguage() loaded it for us
373 if ( isset( $this->loadedItems[$code][$key] ) ||
374 isset( $this->loadedSubitems[$code][$key][$subkey] )
375 ) {
376 return;
377 }
378
379 if ( isset( $this->shallowFallbacks[$code] ) ) {
380 $this->loadSubitem( $this->shallowFallbacks[$code], $key, $subkey );
381
382 return;
383 }
384
385 $value = $this->store->get( $code, "$key:$subkey" );
386 $this->data[$code][$key][$subkey] = $value;
387 $this->loadedSubitems[$code][$key][$subkey] = true;
388 }
389
390 /**
391 * Returns true if the cache identified by $code is missing or expired.
392 *
393 * @param string $code
394 *
395 * @return bool
396 */
397 public function isExpired( $code ) {
398 if ( $this->forceRecache && !isset( $this->recachedLangs[$code] ) ) {
399 wfDebug( __METHOD__ . "($code): forced reload\n" );
400
401 return true;
402 }
403
404 $deps = $this->store->get( $code, 'deps' );
405 $keys = $this->store->get( $code, 'list' );
406 $preload = $this->store->get( $code, 'preload' );
407 // Different keys may expire separately for some stores
408 if ( $deps === null || $keys === null || $preload === null ) {
409 wfDebug( __METHOD__ . "($code): cache missing, need to make one\n" );
410
411 return true;
412 }
413
414 foreach ( $deps as $dep ) {
415 // Because we're unserializing stuff from cache, we
416 // could receive objects of classes that don't exist
417 // anymore (e.g. uninstalled extensions)
418 // When this happens, always expire the cache
419 if ( !$dep instanceof CacheDependency || $dep->isExpired() ) {
420 wfDebug( __METHOD__ . "($code): cache for $code expired due to " .
421 get_class( $dep ) . "\n" );
422
423 return true;
424 }
425 }
426
427 return false;
428 }
429
430 /**
431 * Initialise a language in this object. Rebuild the cache if necessary.
432 * @param string $code
433 * @throws MWException
434 */
435 protected function initLanguage( $code ) {
436 if ( isset( $this->initialisedLangs[$code] ) ) {
437 return;
438 }
439
440 $this->initialisedLangs[$code] = true;
441
442 # If the code is of the wrong form for a Messages*.php file, do a shallow fallback
443 if ( !Language::isValidBuiltInCode( $code ) ) {
444 $this->initShallowFallback( $code, 'en' );
445
446 return;
447 }
448
449 # Recache the data if necessary
450 if ( !$this->manualRecache && $this->isExpired( $code ) ) {
451 if ( Language::isSupportedLanguage( $code ) ) {
452 $this->recache( $code );
453 } elseif ( $code === 'en' ) {
454 throw new MWException( 'MessagesEn.php is missing.' );
455 } else {
456 $this->initShallowFallback( $code, 'en' );
457 }
458
459 return;
460 }
461
462 # Preload some stuff
463 $preload = $this->getItem( $code, 'preload' );
464 if ( $preload === null ) {
465 if ( $this->manualRecache ) {
466 // No Messages*.php file. Do shallow fallback to en.
467 if ( $code === 'en' ) {
468 throw new MWException( 'No localisation cache found for English. ' .
469 'Please run maintenance/rebuildLocalisationCache.php.' );
470 }
471 $this->initShallowFallback( $code, 'en' );
472
473 return;
474 } else {
475 throw new MWException( 'Invalid or missing localisation cache.' );
476 }
477 }
478 $this->data[$code] = $preload;
479 foreach ( $preload as $key => $item ) {
480 if ( in_array( $key, self::$splitKeys ) ) {
481 foreach ( $item as $subkey => $subitem ) {
482 $this->loadedSubitems[$code][$key][$subkey] = true;
483 }
484 } else {
485 $this->loadedItems[$code][$key] = true;
486 }
487 }
488 }
489
490 /**
491 * Create a fallback from one language to another, without creating a
492 * complete persistent cache.
493 * @param string $primaryCode
494 * @param string $fallbackCode
495 */
496 public function initShallowFallback( $primaryCode, $fallbackCode ) {
497 $this->data[$primaryCode] =& $this->data[$fallbackCode];
498 $this->loadedItems[$primaryCode] =& $this->loadedItems[$fallbackCode];
499 $this->loadedSubitems[$primaryCode] =& $this->loadedSubitems[$fallbackCode];
500 $this->shallowFallbacks[$primaryCode] = $fallbackCode;
501 }
502
503 /**
504 * Read a PHP file containing localisation data.
505 * @param string $_fileName
506 * @param string $_fileType
507 * @throws MWException
508 * @return array
509 */
510 protected function readPHPFile( $_fileName, $_fileType ) {
511 wfProfileIn( __METHOD__ );
512 // Disable APC caching
513 wfSuppressWarnings();
514 $_apcEnabled = ini_set( 'apc.cache_by_default', '0' );
515 wfRestoreWarnings();
516
517 include $_fileName;
518
519 wfSuppressWarnings();
520 ini_set( 'apc.cache_by_default', $_apcEnabled );
521 wfRestoreWarnings();
522
523 if ( $_fileType == 'core' || $_fileType == 'extension' ) {
524 $data = compact( self::$allKeys );
525 } elseif ( $_fileType == 'aliases' ) {
526 $data = compact( 'aliases' );
527 } else {
528 wfProfileOut( __METHOD__ );
529 throw new MWException( __METHOD__ . ": Invalid file type: $_fileType" );
530 }
531 wfProfileOut( __METHOD__ );
532
533 return $data;
534 }
535
536 /**
537 * Read a JSON file containing localisation messages.
538 * @param string $fileName Name of file to read
539 * @throws MWException If there is a syntax error in the JSON file
540 * @return array Array with a 'messages' key, or empty array if the file doesn't exist
541 */
542 public function readJSONFile( $fileName ) {
543 wfProfileIn( __METHOD__ );
544
545 if ( !is_readable( $fileName ) ) {
546 wfProfileOut( __METHOD__ );
547
548 return array();
549 }
550
551 $json = file_get_contents( $fileName );
552 if ( $json === false ) {
553 wfProfileOut( __METHOD__ );
554
555 return array();
556 }
557
558 $data = FormatJson::decode( $json, true );
559 if ( $data === null ) {
560 wfProfileOut( __METHOD__ );
561
562 throw new MWException( __METHOD__ . ": Invalid JSON file: $fileName" );
563 }
564
565 // Remove keys starting with '@', they're reserved for metadata and non-message data
566 foreach ( $data as $key => $unused ) {
567 if ( $key === '' || $key[0] === '@' ) {
568 unset( $data[$key] );
569 }
570 }
571
572 wfProfileOut( __METHOD__ );
573
574 // The JSON format only supports messages, none of the other variables, so wrap the data
575 return array( 'messages' => $data );
576 }
577
578 /**
579 * Get the compiled plural rules for a given language from the XML files.
580 * @since 1.20
581 * @param string $code
582 * @return array|null
583 */
584 public function getCompiledPluralRules( $code ) {
585 $rules = $this->getPluralRules( $code );
586 if ( $rules === null ) {
587 return null;
588 }
589 try {
590 $compiledRules = CLDRPluralRuleEvaluator::compile( $rules );
591 } catch ( CLDRPluralRuleError $e ) {
592 wfDebugLog( 'l10n', $e->getMessage() );
593
594 return array();
595 }
596
597 return $compiledRules;
598 }
599
600 /**
601 * Get the plural rules for a given language from the XML files.
602 * Cached.
603 * @since 1.20
604 * @param string $code
605 * @return array|null
606 */
607 public function getPluralRules( $code ) {
608 if ( $this->pluralRules === null ) {
609 $this->loadPluralFiles();
610 }
611 if ( !isset( $this->pluralRules[$code] ) ) {
612 return null;
613 } else {
614 return $this->pluralRules[$code];
615 }
616 }
617
618 /**
619 * Get the plural rule types for a given language from the XML files.
620 * Cached.
621 * @since 1.22
622 * @param string $code
623 * @return array|null
624 */
625 public function getPluralRuleTypes( $code ) {
626 if ( $this->pluralRuleTypes === null ) {
627 $this->loadPluralFiles();
628 }
629 if ( !isset( $this->pluralRuleTypes[$code] ) ) {
630 return null;
631 } else {
632 return $this->pluralRuleTypes[$code];
633 }
634 }
635
636 /**
637 * Load the plural XML files.
638 */
639 protected function loadPluralFiles() {
640 global $IP;
641 $cldrPlural = "$IP/languages/data/plurals.xml";
642 $mwPlural = "$IP/languages/data/plurals-mediawiki.xml";
643 // Load CLDR plural rules
644 $this->loadPluralFile( $cldrPlural );
645 if ( file_exists( $mwPlural ) ) {
646 // Override or extend
647 $this->loadPluralFile( $mwPlural );
648 }
649 }
650
651 /**
652 * Load a plural XML file with the given filename, compile the relevant
653 * rules, and save the compiled rules in a process-local cache.
654 *
655 * @param string $fileName
656 */
657 protected function loadPluralFile( $fileName ) {
658 $doc = new DOMDocument;
659 $doc->load( $fileName );
660 $rulesets = $doc->getElementsByTagName( "pluralRules" );
661 foreach ( $rulesets as $ruleset ) {
662 $codes = $ruleset->getAttribute( 'locales' );
663 $rules = array();
664 $ruleTypes = array();
665 $ruleElements = $ruleset->getElementsByTagName( "pluralRule" );
666 foreach ( $ruleElements as $elt ) {
667 $ruleType = $elt->getAttribute( 'count' );
668 if ( $ruleType === 'other' ) {
669 // Don't record "other" rules, which have an empty condition
670 continue;
671 }
672 $rules[] = $elt->nodeValue;
673 $ruleTypes[] = $ruleType;
674 }
675 foreach ( explode( ' ', $codes ) as $code ) {
676 $this->pluralRules[$code] = $rules;
677 $this->pluralRuleTypes[$code] = $ruleTypes;
678 }
679 }
680 }
681
682 /**
683 * Read the data from the source files for a given language, and register
684 * the relevant dependencies in the $deps array. If the localisation
685 * exists, the data array is returned, otherwise false is returned.
686 *
687 * @param string $code
688 * @param array $deps
689 * @return array
690 */
691 protected function readSourceFilesAndRegisterDeps( $code, &$deps ) {
692 global $IP;
693 wfProfileIn( __METHOD__ );
694
695 // This reads in the PHP i18n file with non-messages l10n data
696 $fileName = Language::getMessagesFileName( $code );
697 if ( !file_exists( $fileName ) ) {
698 $data = array();
699 } else {
700 $deps[] = new FileDependency( $fileName );
701 $data = $this->readPHPFile( $fileName, 'core' );
702 }
703
704 # Load CLDR plural rules for JavaScript
705 $data['pluralRules'] = $this->getPluralRules( $code );
706 # And for PHP
707 $data['compiledPluralRules'] = $this->getCompiledPluralRules( $code );
708 # Load plural rule types
709 $data['pluralRuleTypes'] = $this->getPluralRuleTypes( $code );
710
711 $deps['plurals'] = new FileDependency( "$IP/languages/data/plurals.xml" );
712 $deps['plurals-mw'] = new FileDependency( "$IP/languages/data/plurals-mediawiki.xml" );
713
714 wfProfileOut( __METHOD__ );
715
716 return $data;
717 }
718
719 /**
720 * Merge two localisation values, a primary and a fallback, overwriting the
721 * primary value in place.
722 * @param string $key
723 * @param mixed $value
724 * @param mixed $fallbackValue
725 */
726 protected function mergeItem( $key, &$value, $fallbackValue ) {
727 if ( !is_null( $value ) ) {
728 if ( !is_null( $fallbackValue ) ) {
729 if ( in_array( $key, self::$mergeableMapKeys ) ) {
730 $value = $value + $fallbackValue;
731 } elseif ( in_array( $key, self::$mergeableListKeys ) ) {
732 $value = array_unique( array_merge( $fallbackValue, $value ) );
733 } elseif ( in_array( $key, self::$mergeableAliasListKeys ) ) {
734 $value = array_merge_recursive( $value, $fallbackValue );
735 } elseif ( in_array( $key, self::$optionalMergeKeys ) ) {
736 if ( !empty( $value['inherit'] ) ) {
737 $value = array_merge( $fallbackValue, $value );
738 }
739
740 if ( isset( $value['inherit'] ) ) {
741 unset( $value['inherit'] );
742 }
743 } elseif ( in_array( $key, self::$magicWordKeys ) ) {
744 $this->mergeMagicWords( $value, $fallbackValue );
745 }
746 }
747 } else {
748 $value = $fallbackValue;
749 }
750 }
751
752 /**
753 * @param mixed $value
754 * @param mixed $fallbackValue
755 */
756 protected function mergeMagicWords( &$value, $fallbackValue ) {
757 foreach ( $fallbackValue as $magicName => $fallbackInfo ) {
758 if ( !isset( $value[$magicName] ) ) {
759 $value[$magicName] = $fallbackInfo;
760 } else {
761 $oldSynonyms = array_slice( $fallbackInfo, 1 );
762 $newSynonyms = array_slice( $value[$magicName], 1 );
763 $synonyms = array_values( array_unique( array_merge(
764 $newSynonyms, $oldSynonyms ) ) );
765 $value[$magicName] = array_merge( array( $fallbackInfo[0] ), $synonyms );
766 }
767 }
768 }
769
770 /**
771 * Given an array mapping language code to localisation value, such as is
772 * found in extension *.i18n.php files, iterate through a fallback sequence
773 * to merge the given data with an existing primary value.
774 *
775 * Returns true if any data from the extension array was used, false
776 * otherwise.
777 * @param array $codeSequence
778 * @param string $key
779 * @param mixed $value
780 * @param mixed $fallbackValue
781 * @return bool
782 */
783 protected function mergeExtensionItem( $codeSequence, $key, &$value, $fallbackValue ) {
784 $used = false;
785 foreach ( $codeSequence as $code ) {
786 if ( isset( $fallbackValue[$code] ) ) {
787 $this->mergeItem( $key, $value, $fallbackValue[$code] );
788 $used = true;
789 }
790 }
791
792 return $used;
793 }
794
795 /**
796 * Gets the combined list of messages dirs from
797 * core and extensions
798 *
799 * @since 1.25
800 * @return array
801 */
802 public function getMessagesDirs() {
803 global $wgMessagesDirs, $IP;
804 return array(
805 'core' => "$IP/languages/i18n",
806 'api' => "$IP/includes/api/i18n",
807 'oojs-ui' => "$IP/resources/lib/oojs-ui/i18n",
808 ) + $wgMessagesDirs;
809 }
810
811 /**
812 * Load localisation data for a given language for both core and extensions
813 * and save it to the persistent cache store and the process cache
814 * @param string $code
815 * @throws MWException
816 */
817 public function recache( $code ) {
818 global $wgExtensionMessagesFiles;
819 wfProfileIn( __METHOD__ );
820
821 if ( !$code ) {
822 wfProfileOut( __METHOD__ );
823 throw new MWException( "Invalid language code requested" );
824 }
825 $this->recachedLangs[$code] = true;
826
827 # Initial values
828 $initialData = array_combine(
829 self::$allKeys,
830 array_fill( 0, count( self::$allKeys ), null ) );
831 $coreData = $initialData;
832 $deps = array();
833
834 # Load the primary localisation from the source file
835 $data = $this->readSourceFilesAndRegisterDeps( $code, $deps );
836 if ( $data === false ) {
837 wfDebug( __METHOD__ . ": no localisation file for $code, using fallback to en\n" );
838 $coreData['fallback'] = 'en';
839 } else {
840 wfDebug( __METHOD__ . ": got localisation for $code from source\n" );
841
842 # Merge primary localisation
843 foreach ( $data as $key => $value ) {
844 $this->mergeItem( $key, $coreData[$key], $value );
845 }
846 }
847
848 # Fill in the fallback if it's not there already
849 if ( is_null( $coreData['fallback'] ) ) {
850 $coreData['fallback'] = $code === 'en' ? false : 'en';
851 }
852 if ( $coreData['fallback'] === false ) {
853 $coreData['fallbackSequence'] = array();
854 } else {
855 $coreData['fallbackSequence'] = array_map( 'trim', explode( ',', $coreData['fallback'] ) );
856 $len = count( $coreData['fallbackSequence'] );
857
858 # Ensure that the sequence ends at en
859 if ( $coreData['fallbackSequence'][$len - 1] !== 'en' ) {
860 $coreData['fallbackSequence'][] = 'en';
861 }
862 }
863
864 $codeSequence = array_merge( array( $code ), $coreData['fallbackSequence'] );
865 $messageDirs = $this->getMessagesDirs();
866
867 wfProfileIn( __METHOD__ . '-fallbacks' );
868
869 # Load non-JSON localisation data for extensions
870 $extensionData = array_combine(
871 $codeSequence,
872 array_fill( 0, count( $codeSequence ), $initialData ) );
873 foreach ( $wgExtensionMessagesFiles as $extension => $fileName ) {
874 if ( isset( $messageDirs[$extension] ) ) {
875 # This extension has JSON message data; skip the PHP shim
876 continue;
877 }
878
879 $data = $this->readPHPFile( $fileName, 'extension' );
880 $used = false;
881
882 foreach ( $data as $key => $item ) {
883 foreach ( $codeSequence as $csCode ) {
884 if ( isset( $item[$csCode] ) ) {
885 $this->mergeItem( $key, $extensionData[$csCode][$key], $item[$csCode] );
886 $used = true;
887 }
888 }
889 }
890
891 if ( $used ) {
892 $deps[] = new FileDependency( $fileName );
893 }
894 }
895
896 # Load the localisation data for each fallback, then merge it into the full array
897 $allData = $initialData;
898 foreach ( $codeSequence as $csCode ) {
899 $csData = $initialData;
900
901 # Load core messages and the extension localisations.
902 foreach ( $messageDirs as $dirs ) {
903 foreach ( (array)$dirs as $dir ) {
904 $fileName = "$dir/$csCode.json";
905 $data = $this->readJSONFile( $fileName );
906
907 foreach ( $data as $key => $item ) {
908 $this->mergeItem( $key, $csData[$key], $item );
909 }
910
911 $deps[] = new FileDependency( $fileName );
912 }
913 }
914
915 # Merge non-JSON extension data
916 if ( isset( $extensionData[$csCode] ) ) {
917 foreach ( $extensionData[$csCode] as $key => $item ) {
918 $this->mergeItem( $key, $csData[$key], $item );
919 }
920 }
921
922 if ( $csCode === $code ) {
923 # Merge core data into extension data
924 foreach ( $coreData as $key => $item ) {
925 $this->mergeItem( $key, $csData[$key], $item );
926 }
927 } else {
928 # Load the secondary localisation from the source file to
929 # avoid infinite cycles on cyclic fallbacks
930 $fbData = $this->readSourceFilesAndRegisterDeps( $csCode, $deps );
931 if ( $fbData !== false ) {
932 # Only merge the keys that make sense to merge
933 foreach ( self::$allKeys as $key ) {
934 if ( !isset( $fbData[$key] ) ) {
935 continue;
936 }
937
938 if ( is_null( $coreData[$key] ) || $this->isMergeableKey( $key ) ) {
939 $this->mergeItem( $key, $csData[$key], $fbData[$key] );
940 }
941 }
942 }
943 }
944
945 # Allow extensions an opportunity to adjust the data for this
946 # fallback
947 wfRunHooks( 'LocalisationCacheRecacheFallback', array( $this, $csCode, &$csData ) );
948
949 # Merge the data for this fallback into the final array
950 if ( $csCode === $code ) {
951 $allData = $csData;
952 } else {
953 foreach ( self::$allKeys as $key ) {
954 if ( !isset( $csData[$key] ) ) {
955 continue;
956 }
957
958 if ( is_null( $allData[$key] ) || $this->isMergeableKey( $key ) ) {
959 $this->mergeItem( $key, $allData[$key], $csData[$key] );
960 }
961 }
962 }
963 }
964
965 wfProfileOut( __METHOD__ . '-fallbacks' );
966
967 # Add cache dependencies for any referenced globals
968 $deps['wgExtensionMessagesFiles'] = new GlobalDependency( 'wgExtensionMessagesFiles' );
969 // $wgMessagesDirs is used in LocalisationCache::getMessagesDirs()
970 $deps['wgMessagesDirs'] = new GlobalDependency( 'wgMessagesDirs' );
971 $deps['version'] = new ConstantDependency( 'LocalisationCache::VERSION' );
972
973 # Add dependencies to the cache entry
974 $allData['deps'] = $deps;
975
976 # Replace spaces with underscores in namespace names
977 $allData['namespaceNames'] = str_replace( ' ', '_', $allData['namespaceNames'] );
978
979 # And do the same for special page aliases. $page is an array.
980 foreach ( $allData['specialPageAliases'] as &$page ) {
981 $page = str_replace( ' ', '_', $page );
982 }
983 # Decouple the reference to prevent accidental damage
984 unset( $page );
985
986 # If there were no plural rules, return an empty array
987 if ( $allData['pluralRules'] === null ) {
988 $allData['pluralRules'] = array();
989 }
990 if ( $allData['compiledPluralRules'] === null ) {
991 $allData['compiledPluralRules'] = array();
992 }
993 # If there were no plural rule types, return an empty array
994 if ( $allData['pluralRuleTypes'] === null ) {
995 $allData['pluralRuleTypes'] = array();
996 }
997
998 # Set the list keys
999 $allData['list'] = array();
1000 foreach ( self::$splitKeys as $key ) {
1001 $allData['list'][$key] = array_keys( $allData[$key] );
1002 }
1003 # Run hooks
1004 $purgeBlobs = true;
1005 wfRunHooks( 'LocalisationCacheRecache', array( $this, $code, &$allData, &$purgeBlobs ) );
1006
1007 if ( is_null( $allData['namespaceNames'] ) ) {
1008 wfProfileOut( __METHOD__ );
1009 throw new MWException( __METHOD__ . ': Localisation data failed sanity check! ' .
1010 'Check that your languages/messages/MessagesEn.php file is intact.' );
1011 }
1012
1013 # Set the preload key
1014 $allData['preload'] = $this->buildPreload( $allData );
1015
1016 # Save to the process cache and register the items loaded
1017 $this->data[$code] = $allData;
1018 foreach ( $allData as $key => $item ) {
1019 $this->loadedItems[$code][$key] = true;
1020 }
1021
1022 # Save to the persistent cache
1023 wfProfileIn( __METHOD__ . '-write' );
1024 $this->store->startWrite( $code );
1025 foreach ( $allData as $key => $value ) {
1026 if ( in_array( $key, self::$splitKeys ) ) {
1027 foreach ( $value as $subkey => $subvalue ) {
1028 $this->store->set( "$key:$subkey", $subvalue );
1029 }
1030 } else {
1031 $this->store->set( $key, $value );
1032 }
1033 }
1034 $this->store->finishWrite();
1035 wfProfileOut( __METHOD__ . '-write' );
1036
1037 # Clear out the MessageBlobStore
1038 # HACK: If using a null (i.e. disabled) storage backend, we
1039 # can't write to the MessageBlobStore either
1040 if ( $purgeBlobs && !$this->store instanceof LCStoreNull ) {
1041 MessageBlobStore::getInstance()->clear();
1042 }
1043
1044 wfProfileOut( __METHOD__ );
1045 }
1046
1047 /**
1048 * Build the preload item from the given pre-cache data.
1049 *
1050 * The preload item will be loaded automatically, improving performance
1051 * for the commonly-requested items it contains.
1052 * @param array $data
1053 * @return array
1054 */
1055 protected function buildPreload( $data ) {
1056 $preload = array( 'messages' => array() );
1057 foreach ( self::$preloadedKeys as $key ) {
1058 $preload[$key] = $data[$key];
1059 }
1060
1061 foreach ( $data['preloadedMessages'] as $subkey ) {
1062 if ( isset( $data['messages'][$subkey] ) ) {
1063 $subitem = $data['messages'][$subkey];
1064 } else {
1065 $subitem = null;
1066 }
1067 $preload['messages'][$subkey] = $subitem;
1068 }
1069
1070 return $preload;
1071 }
1072
1073 /**
1074 * Unload the data for a given language from the object cache.
1075 * Reduces memory usage.
1076 * @param string $code
1077 */
1078 public function unload( $code ) {
1079 unset( $this->data[$code] );
1080 unset( $this->loadedItems[$code] );
1081 unset( $this->loadedSubitems[$code] );
1082 unset( $this->initialisedLangs[$code] );
1083 unset( $this->shallowFallbacks[$code] );
1084
1085 foreach ( $this->shallowFallbacks as $shallowCode => $fbCode ) {
1086 if ( $fbCode === $code ) {
1087 $this->unload( $shallowCode );
1088 }
1089 }
1090 }
1091
1092 /**
1093 * Unload all data
1094 */
1095 public function unloadAll() {
1096 foreach ( $this->initialisedLangs as $lang => $unused ) {
1097 $this->unload( $lang );
1098 }
1099 }
1100
1101 /**
1102 * Disable the storage backend
1103 */
1104 public function disableBackend() {
1105 $this->store = new LCStoreNull;
1106 $this->manualRecache = false;
1107 }
1108 }
1109
1110 /**
1111 * Interface for the persistence layer of LocalisationCache.
1112 *
1113 * The persistence layer is two-level hierarchical cache. The first level
1114 * is the language, the second level is the item or subitem.
1115 *
1116 * Since the data for a whole language is rebuilt in one operation, it needs
1117 * to have a fast and atomic method for deleting or replacing all of the
1118 * current data for a given language. The interface reflects this bulk update
1119 * operation. Callers writing to the cache must first call startWrite(), then
1120 * will call set() a couple of thousand times, then will call finishWrite()
1121 * to commit the operation. When finishWrite() is called, the cache is
1122 * expected to delete all data previously stored for that language.
1123 *
1124 * The values stored are PHP variables suitable for serialize(). Implementations
1125 * of LCStore are responsible for serializing and unserializing.
1126 */
1127 interface LCStore {
1128 /**
1129 * Get a value.
1130 * @param string $code Language code
1131 * @param string $key Cache key
1132 */
1133 function get( $code, $key );
1134
1135 /**
1136 * Start a write transaction.
1137 * @param string $code Language code
1138 */
1139 function startWrite( $code );
1140
1141 /**
1142 * Finish a write transaction.
1143 */
1144 function finishWrite();
1145
1146 /**
1147 * Set a key to a given value. startWrite() must be called before this
1148 * is called, and finishWrite() must be called afterwards.
1149 * @param string $key
1150 * @param mixed $value
1151 */
1152 function set( $key, $value );
1153 }
1154
1155 /**
1156 * LCStore implementation which uses the standard DB functions to store data.
1157 * This will work on any MediaWiki installation.
1158 */
1159 class LCStoreDB implements LCStore {
1160 private $currentLang;
1161 private $writesDone = false;
1162
1163 /** @var DatabaseBase */
1164 private $dbw;
1165 /** @var array */
1166 private $batch = array();
1167
1168 private $readOnly = false;
1169
1170 public function get( $code, $key ) {
1171 if ( $this->writesDone ) {
1172 $db = wfGetDB( DB_MASTER );
1173 } else {
1174 $db = wfGetDB( DB_SLAVE );
1175 }
1176 $row = $db->selectRow( 'l10n_cache', array( 'lc_value' ),
1177 array( 'lc_lang' => $code, 'lc_key' => $key ), __METHOD__ );
1178 if ( $row ) {
1179 return unserialize( $db->decodeBlob( $row->lc_value ) );
1180 } else {
1181 return null;
1182 }
1183 }
1184
1185 public function startWrite( $code ) {
1186 if ( $this->readOnly ) {
1187 return;
1188 } elseif ( !$code ) {
1189 throw new MWException( __METHOD__ . ": Invalid language \"$code\"" );
1190 }
1191
1192 $this->dbw = wfGetDB( DB_MASTER );
1193
1194 $this->currentLang = $code;
1195 $this->batch = array();
1196 }
1197
1198 public function finishWrite() {
1199 if ( $this->readOnly ) {
1200 return;
1201 } elseif ( is_null( $this->currentLang ) ) {
1202 throw new MWException( __CLASS__ . ': must call startWrite() before finishWrite()' );
1203 }
1204
1205 $this->dbw->begin( __METHOD__ );
1206 try {
1207 $this->dbw->delete( 'l10n_cache',
1208 array( 'lc_lang' => $this->currentLang ), __METHOD__ );
1209 foreach ( array_chunk( $this->batch, 500 ) as $rows ) {
1210 $this->dbw->insert( 'l10n_cache', $rows, __METHOD__ );
1211 }
1212 $this->writesDone = true;
1213 } catch ( DBQueryError $e ) {
1214 if ( $this->dbw->wasReadOnlyError() ) {
1215 $this->readOnly = true; // just avoid site down time
1216 } else {
1217 throw $e;
1218 }
1219 }
1220 $this->dbw->commit( __METHOD__ );
1221
1222 $this->currentLang = null;
1223 $this->batch = array();
1224 }
1225
1226 public function set( $key, $value ) {
1227 if ( $this->readOnly ) {
1228 return;
1229 } elseif ( is_null( $this->currentLang ) ) {
1230 throw new MWException( __CLASS__ . ': must call startWrite() before set()' );
1231 }
1232
1233 $this->batch[] = array(
1234 'lc_lang' => $this->currentLang,
1235 'lc_key' => $key,
1236 'lc_value' => $this->dbw->encodeBlob( serialize( $value ) ) );
1237 }
1238 }
1239
1240 /**
1241 * LCStore implementation which stores data as a collection of CDB files in the
1242 * directory given by $wgCacheDirectory. If $wgCacheDirectory is not set, this
1243 * will throw an exception.
1244 *
1245 * Profiling indicates that on Linux, this implementation outperforms MySQL if
1246 * the directory is on a local filesystem and there is ample kernel cache
1247 * space. The performance advantage is greater when the DBA extension is
1248 * available than it is with the PHP port.
1249 *
1250 * See Cdb.php and http://cr.yp.to/cdb.html
1251 */
1252 class LCStoreCDB implements LCStore {
1253 /** @var CdbReader[] */
1254 private $readers;
1255
1256 /** @var CdbWriter */
1257 private $writer;
1258
1259 /** @var string Current language code */
1260 private $currentLang;
1261
1262 /** @var bool|string Cache directory. False if not set */
1263 private $directory;
1264
1265 function __construct( $conf = array() ) {
1266 global $wgCacheDirectory;
1267
1268 if ( isset( $conf['directory'] ) ) {
1269 $this->directory = $conf['directory'];
1270 } else {
1271 $this->directory = $wgCacheDirectory;
1272 }
1273 }
1274
1275 public function get( $code, $key ) {
1276 if ( !isset( $this->readers[$code] ) ) {
1277 $fileName = $this->getFileName( $code );
1278
1279 $this->readers[$code] = false;
1280 if ( file_exists( $fileName ) ) {
1281 try {
1282 $this->readers[$code] = CdbReader::open( $fileName );
1283 } catch ( CdbException $e ) {
1284 wfDebug( __METHOD__ . ": unable to open cdb file for reading\n" );
1285 }
1286 }
1287 }
1288
1289 if ( !$this->readers[$code] ) {
1290 return null;
1291 } else {
1292 $value = false;
1293 try {
1294 $value = $this->readers[$code]->get( $key );
1295 } catch ( CdbException $e ) {
1296 wfDebug( __METHOD__ . ": CdbException caught, error message was "
1297 . $e->getMessage() . "\n" );
1298 }
1299 if ( $value === false ) {
1300 return null;
1301 }
1302
1303 return unserialize( $value );
1304 }
1305 }
1306
1307 public function startWrite( $code ) {
1308 if ( !file_exists( $this->directory ) ) {
1309 if ( !wfMkdirParents( $this->directory, null, __METHOD__ ) ) {
1310 throw new MWException( "Unable to create the localisation store " .
1311 "directory \"{$this->directory}\"" );
1312 }
1313 }
1314
1315 // Close reader to stop permission errors on write
1316 if ( !empty( $this->readers[$code] ) ) {
1317 $this->readers[$code]->close();
1318 }
1319
1320 try {
1321 $this->writer = CdbWriter::open( $this->getFileName( $code ) );
1322 } catch ( CdbException $e ) {
1323 throw new MWException( $e->getMessage() );
1324 }
1325 $this->currentLang = $code;
1326 }
1327
1328 public function finishWrite() {
1329 // Close the writer
1330 try {
1331 $this->writer->close();
1332 } catch ( CdbException $e ) {
1333 throw new MWException( $e->getMessage() );
1334 }
1335 $this->writer = null;
1336 unset( $this->readers[$this->currentLang] );
1337 $this->currentLang = null;
1338 }
1339
1340 public function set( $key, $value ) {
1341 if ( is_null( $this->writer ) ) {
1342 throw new MWException( __CLASS__ . ': must call startWrite() before calling set()' );
1343 }
1344 try {
1345 $this->writer->set( $key, serialize( $value ) );
1346 } catch ( CdbException $e ) {
1347 throw new MWException( $e->getMessage() );
1348 }
1349 }
1350
1351 protected function getFileName( $code ) {
1352 if ( strval( $code ) === '' || strpos( $code, '/' ) !== false ) {
1353 throw new MWException( __METHOD__ . ": Invalid language \"$code\"" );
1354 }
1355
1356 return "{$this->directory}/l10n_cache-$code.cdb";
1357 }
1358 }
1359
1360 /**
1361 * Null store backend, used to avoid DB errors during install
1362 */
1363 class LCStoreNull implements LCStore {
1364 public function get( $code, $key ) {
1365 return null;
1366 }
1367
1368 public function startWrite( $code ) {
1369 }
1370
1371 public function finishWrite() {
1372 }
1373
1374 public function set( $key, $value ) {
1375 }
1376 }
1377
1378 /**
1379 * A localisation cache optimised for loading large amounts of data for many
1380 * languages. Used by rebuildLocalisationCache.php.
1381 */
1382 class LocalisationCacheBulkLoad extends LocalisationCache {
1383 /**
1384 * A cache of the contents of data files.
1385 * Core files are serialized to avoid using ~1GB of RAM during a recache.
1386 */
1387 private $fileCache = array();
1388
1389 /**
1390 * Most recently used languages. Uses the linked-list aspect of PHP hashtables
1391 * to keep the most recently used language codes at the end of the array, and
1392 * the language codes that are ready to be deleted at the beginning.
1393 */
1394 private $mruLangs = array();
1395
1396 /**
1397 * Maximum number of languages that may be loaded into $this->data
1398 */
1399 private $maxLoadedLangs = 10;
1400
1401 /**
1402 * @param string $fileName
1403 * @param string $fileType
1404 * @return array|mixed
1405 */
1406 protected function readPHPFile( $fileName, $fileType ) {
1407 $serialize = $fileType === 'core';
1408 if ( !isset( $this->fileCache[$fileName][$fileType] ) ) {
1409 $data = parent::readPHPFile( $fileName, $fileType );
1410
1411 if ( $serialize ) {
1412 $encData = serialize( $data );
1413 } else {
1414 $encData = $data;
1415 }
1416
1417 $this->fileCache[$fileName][$fileType] = $encData;
1418
1419 return $data;
1420 } elseif ( $serialize ) {
1421 return unserialize( $this->fileCache[$fileName][$fileType] );
1422 } else {
1423 return $this->fileCache[$fileName][$fileType];
1424 }
1425 }
1426
1427 /**
1428 * @param string $code
1429 * @param string $key
1430 * @return mixed
1431 */
1432 public function getItem( $code, $key ) {
1433 unset( $this->mruLangs[$code] );
1434 $this->mruLangs[$code] = true;
1435
1436 return parent::getItem( $code, $key );
1437 }
1438
1439 /**
1440 * @param string $code
1441 * @param string $key
1442 * @param string $subkey
1443 * @return mixed
1444 */
1445 public function getSubitem( $code, $key, $subkey ) {
1446 unset( $this->mruLangs[$code] );
1447 $this->mruLangs[$code] = true;
1448
1449 return parent::getSubitem( $code, $key, $subkey );
1450 }
1451
1452 /**
1453 * @param string $code
1454 */
1455 public function recache( $code ) {
1456 parent::recache( $code );
1457 unset( $this->mruLangs[$code] );
1458 $this->mruLangs[$code] = true;
1459 $this->trimCache();
1460 }
1461
1462 /**
1463 * @param string $code
1464 */
1465 public function unload( $code ) {
1466 unset( $this->mruLangs[$code] );
1467 parent::unload( $code );
1468 }
1469
1470 /**
1471 * Unload cached languages until there are less than $this->maxLoadedLangs
1472 */
1473 protected function trimCache() {
1474 while ( count( $this->data ) > $this->maxLoadedLangs && count( $this->mruLangs ) ) {
1475 reset( $this->mruLangs );
1476 $code = key( $this->mruLangs );
1477 wfDebug( __METHOD__ . ": unloading $code\n" );
1478 $this->unload( $code );
1479 }
1480 }
1481 }