EditPage::newSectionSummary should return a value in all code paths
[lhc/web/wiklou.git] / includes / cache / LocalisationCache.php
1 <?php
2 /**
3 * Cache of the contents of localisation files.
4 *
5 * This program is free software; you can redistribute it and/or modify
6 * it under the terms of the GNU General Public License as published by
7 * the Free Software Foundation; either version 2 of the License, or
8 * (at your option) any later version.
9 *
10 * This program is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 * GNU General Public License for more details.
14 *
15 * You should have received a copy of the GNU General Public License along
16 * with this program; if not, write to the Free Software Foundation, Inc.,
17 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
18 * http://www.gnu.org/copyleft/gpl.html
19 *
20 * @file
21 */
22
23 define( 'MW_LC_VERSION', 2 );
24
25 /**
26 * Class for caching the contents of localisation files, Messages*.php
27 * and *.i18n.php.
28 *
29 * An instance of this class is available using Language::getLocalisationCache().
30 *
31 * The values retrieved from here are merged, containing items from extension
32 * files, core messages files and the language fallback sequence (e.g. zh-cn ->
33 * zh-hans -> en ). Some common errors are corrected, for example namespace
34 * names with spaces instead of underscores, but heavyweight processing, such
35 * as grammatical transformation, is done by the caller.
36 */
37 class LocalisationCache {
38 /** Configuration associative array */
39 private $conf;
40
41 /**
42 * True if recaching should only be done on an explicit call to recache().
43 * Setting this reduces the overhead of cache freshness checking, which
44 * requires doing a stat() for every extension i18n file.
45 */
46 private $manualRecache = false;
47
48 /**
49 * True to treat all files as expired until they are regenerated by this object.
50 */
51 private $forceRecache = false;
52
53 /**
54 * The cache data. 3-d array, where the first key is the language code,
55 * the second key is the item key e.g. 'messages', and the third key is
56 * an item specific subkey index. Some items are not arrays and so for those
57 * items, there are no subkeys.
58 */
59 protected $data = array();
60
61 /**
62 * The persistent store object. An instance of LCStore.
63 *
64 * @var LCStore
65 */
66 private $store;
67
68 /**
69 * A 2-d associative array, code/key, where presence indicates that the item
70 * is loaded. Value arbitrary.
71 *
72 * For split items, if set, this indicates that all of the subitems have been
73 * loaded.
74 */
75 private $loadedItems = array();
76
77 /**
78 * A 3-d associative array, code/key/subkey, where presence indicates that
79 * the subitem is loaded. Only used for the split items, i.e. messages.
80 */
81 private $loadedSubitems = array();
82
83 /**
84 * An array where presence of a key indicates that that language has been
85 * initialised. Initialisation includes checking for cache expiry and doing
86 * any necessary updates.
87 */
88 private $initialisedLangs = array();
89
90 /**
91 * An array mapping non-existent pseudo-languages to fallback languages. This
92 * is filled by initShallowFallback() when data is requested from a language
93 * that lacks a Messages*.php file.
94 */
95 private $shallowFallbacks = array();
96
97 /**
98 * An array where the keys are codes that have been recached by this instance.
99 */
100 private $recachedLangs = array();
101
102 /**
103 * All item keys
104 */
105 static public $allKeys = array(
106 'fallback', 'namespaceNames', 'bookstoreList',
107 'magicWords', 'messages', 'rtl', 'capitalizeAllNouns', 'digitTransformTable',
108 'separatorTransformTable', 'fallback8bitEncoding', 'linkPrefixExtension',
109 'linkTrail', 'linkPrefixCharset', 'namespaceAliases',
110 'dateFormats', 'datePreferences', 'datePreferenceMigrationMap',
111 'defaultDateFormat', 'extraUserToggles', 'specialPageAliases',
112 'imageFiles', 'preloadedMessages', 'namespaceGenderAliases',
113 'digitGroupingPattern', 'pluralRules', 'pluralRuleTypes', 'compiledPluralRules',
114 );
115
116 /**
117 * Keys for items which consist of associative arrays, which may be merged
118 * by a fallback sequence.
119 */
120 static public $mergeableMapKeys = array( 'messages', 'namespaceNames',
121 'dateFormats', 'imageFiles', 'preloadedMessages'
122 );
123
124 /**
125 * Keys for items which are a numbered array.
126 */
127 static public $mergeableListKeys = array( 'extraUserToggles' );
128
129 /**
130 * Keys for items which contain an array of arrays of equivalent aliases
131 * for each subitem. The aliases may be merged by a fallback sequence.
132 */
133 static public $mergeableAliasListKeys = array( 'specialPageAliases' );
134
135 /**
136 * Keys for items which contain an associative array, and may be merged if
137 * the primary value contains the special array key "inherit". That array
138 * key is removed after the first merge.
139 */
140 static public $optionalMergeKeys = array( 'bookstoreList' );
141
142 /**
143 * Keys for items that are formatted like $magicWords
144 */
145 static public $magicWordKeys = array( 'magicWords' );
146
147 /**
148 * Keys for items where the subitems are stored in the backend separately.
149 */
150 static public $splitKeys = array( 'messages' );
151
152 /**
153 * Keys which are loaded automatically by initLanguage()
154 */
155 static public $preloadedKeys = array( 'dateFormats', 'namespaceNames' );
156
157 /**
158 * Associative array of cached plural rules. The key is the language code,
159 * the value is an array of plural rules for that language.
160 */
161 private $pluralRules = null;
162
163 /**
164 * Associative array of cached plural rule types. The key is the language
165 * code, the value is an array of plural rule types for that language. For
166 * example, $pluralRuleTypes['ar'] = ['zero', 'one', 'two', 'few', 'many'].
167 * The index for each rule type matches the index for the rule in
168 * $pluralRules, thus allowing correlation between the two. The reason we
169 * don't just use the type names as the keys in $pluralRules is because
170 * Language::convertPlural applies the rules based on numeric order (or
171 * explicit numeric parameter), not based on the name of the rule type. For
172 * example, {{plural:count|wordform1|wordform2|wordform3}}, rather than
173 * {{plural:count|one=wordform1|two=wordform2|many=wordform3}}.
174 */
175 private $pluralRuleTypes = null;
176
177 private $mergeableKeys = null;
178
179 /**
180 * Constructor.
181 * For constructor parameters, see the documentation in DefaultSettings.php
182 * for $wgLocalisationCacheConf.
183 *
184 * @param array $conf
185 * @throws MWException
186 */
187 function __construct( $conf ) {
188 global $wgCacheDirectory;
189
190 $this->conf = $conf;
191 $storeConf = array();
192 if ( !empty( $conf['storeClass'] ) ) {
193 $storeClass = $conf['storeClass'];
194 } else {
195 switch ( $conf['store'] ) {
196 case 'files':
197 case 'file':
198 $storeClass = 'LCStoreCDB';
199 break;
200 case 'db':
201 $storeClass = 'LCStoreDB';
202 break;
203 case 'accel':
204 $storeClass = 'LCStoreAccel';
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, at least in LCStoreAccel
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 */
690 protected function readSourceFilesAndRegisterDeps( $code, &$deps ) {
691 global $IP;
692 wfProfileIn( __METHOD__ );
693
694 // This reads in the PHP i18n file with non-messages l10n data
695 $fileName = Language::getMessagesFileName( $code );
696 if ( !file_exists( $fileName ) ) {
697 $data = array();
698 } else {
699 $deps[] = new FileDependency( $fileName );
700 $data = $this->readPHPFile( $fileName, 'core' );
701 }
702
703 # Load CLDR plural rules for JavaScript
704 $data['pluralRules'] = $this->getPluralRules( $code );
705 # And for PHP
706 $data['compiledPluralRules'] = $this->getCompiledPluralRules( $code );
707 # Load plural rule types
708 $data['pluralRuleTypes'] = $this->getPluralRuleTypes( $code );
709
710 $deps['plurals'] = new FileDependency( "$IP/languages/data/plurals.xml" );
711 $deps['plurals-mw'] = new FileDependency( "$IP/languages/data/plurals-mediawiki.xml" );
712
713 wfProfileOut( __METHOD__ );
714
715 return $data;
716 }
717
718 /**
719 * Merge two localisation values, a primary and a fallback, overwriting the
720 * primary value in place.
721 * @param string $key
722 * @param mixed $value
723 * @param mixed $fallbackValue
724 */
725 protected function mergeItem( $key, &$value, $fallbackValue ) {
726 if ( !is_null( $value ) ) {
727 if ( !is_null( $fallbackValue ) ) {
728 if ( in_array( $key, self::$mergeableMapKeys ) ) {
729 $value = $value + $fallbackValue;
730 } elseif ( in_array( $key, self::$mergeableListKeys ) ) {
731 $value = array_unique( array_merge( $fallbackValue, $value ) );
732 } elseif ( in_array( $key, self::$mergeableAliasListKeys ) ) {
733 $value = array_merge_recursive( $value, $fallbackValue );
734 } elseif ( in_array( $key, self::$optionalMergeKeys ) ) {
735 if ( !empty( $value['inherit'] ) ) {
736 $value = array_merge( $fallbackValue, $value );
737 }
738
739 if ( isset( $value['inherit'] ) ) {
740 unset( $value['inherit'] );
741 }
742 } elseif ( in_array( $key, self::$magicWordKeys ) ) {
743 $this->mergeMagicWords( $value, $fallbackValue );
744 }
745 }
746 } else {
747 $value = $fallbackValue;
748 }
749 }
750
751 /**
752 * @param mixed $value
753 * @param mixed $fallbackValue
754 */
755 protected function mergeMagicWords( &$value, $fallbackValue ) {
756 foreach ( $fallbackValue as $magicName => $fallbackInfo ) {
757 if ( !isset( $value[$magicName] ) ) {
758 $value[$magicName] = $fallbackInfo;
759 } else {
760 $oldSynonyms = array_slice( $fallbackInfo, 1 );
761 $newSynonyms = array_slice( $value[$magicName], 1 );
762 $synonyms = array_values( array_unique( array_merge(
763 $newSynonyms, $oldSynonyms ) ) );
764 $value[$magicName] = array_merge( array( $fallbackInfo[0] ), $synonyms );
765 }
766 }
767 }
768
769 /**
770 * Given an array mapping language code to localisation value, such as is
771 * found in extension *.i18n.php files, iterate through a fallback sequence
772 * to merge the given data with an existing primary value.
773 *
774 * Returns true if any data from the extension array was used, false
775 * otherwise.
776 * @param string $codeSequence
777 * @param string $key
778 * @param mixed $value
779 * @param mixed $fallbackValue
780 * @return bool
781 */
782 protected function mergeExtensionItem( $codeSequence, $key, &$value, $fallbackValue ) {
783 $used = false;
784 foreach ( $codeSequence as $code ) {
785 if ( isset( $fallbackValue[$code] ) ) {
786 $this->mergeItem( $key, $value, $fallbackValue[$code] );
787 $used = true;
788 }
789 }
790
791 return $used;
792 }
793
794 /**
795 * Load localisation data for a given language for both core and extensions
796 * and save it to the persistent cache store and the process cache
797 * @param string $code
798 * @throws MWException
799 */
800 public function recache( $code ) {
801 global $wgExtensionMessagesFiles, $wgMessagesDirs;
802 wfProfileIn( __METHOD__ );
803
804 if ( !$code ) {
805 wfProfileOut( __METHOD__ );
806 throw new MWException( "Invalid language code requested" );
807 }
808 $this->recachedLangs[$code] = true;
809
810 # Initial values
811 $initialData = array_combine(
812 self::$allKeys,
813 array_fill( 0, count( self::$allKeys ), null ) );
814 $coreData = $initialData;
815 $deps = array();
816
817 # Load the primary localisation from the source file
818 $data = $this->readSourceFilesAndRegisterDeps( $code, $deps );
819 if ( $data === false ) {
820 wfDebug( __METHOD__ . ": no localisation file for $code, using fallback to en\n" );
821 $coreData['fallback'] = 'en';
822 } else {
823 wfDebug( __METHOD__ . ": got localisation for $code from source\n" );
824
825 # Merge primary localisation
826 foreach ( $data as $key => $value ) {
827 $this->mergeItem( $key, $coreData[$key], $value );
828 }
829 }
830
831 # Fill in the fallback if it's not there already
832 if ( is_null( $coreData['fallback'] ) ) {
833 $coreData['fallback'] = $code === 'en' ? false : 'en';
834 }
835 if ( $coreData['fallback'] === false ) {
836 $coreData['fallbackSequence'] = array();
837 } else {
838 $coreData['fallbackSequence'] = array_map( 'trim', explode( ',', $coreData['fallback'] ) );
839 $len = count( $coreData['fallbackSequence'] );
840
841 # Ensure that the sequence ends at en
842 if ( $coreData['fallbackSequence'][$len - 1] !== 'en' ) {
843 $coreData['fallbackSequence'][] = 'en';
844 }
845
846 # Load the fallback localisation item by item and merge it
847 foreach ( $coreData['fallbackSequence'] as $fbCode ) {
848 # Load the secondary localisation from the source file to
849 # avoid infinite cycles on cyclic fallbacks
850 $fbData = $this->readSourceFilesAndRegisterDeps( $fbCode, $deps );
851 if ( $fbData === false ) {
852 continue;
853 }
854
855 foreach ( self::$allKeys as $key ) {
856 if ( !isset( $fbData[$key] ) ) {
857 continue;
858 }
859
860 if ( is_null( $coreData[$key] ) || $this->isMergeableKey( $key ) ) {
861 $this->mergeItem( $key, $coreData[$key], $fbData[$key] );
862 }
863 }
864 }
865 }
866
867 $codeSequence = array_merge( array( $code ), $coreData['fallbackSequence'] );
868
869 # Load core messages and the extension localisations.
870 wfProfileIn( __METHOD__ . '-extensions' );
871 $allData = $initialData;
872 foreach ( $wgMessagesDirs as $dirs ) {
873 foreach ( (array)$dirs as $dir ) {
874 foreach ( $codeSequence as $csCode ) {
875 $fileName = "$dir/$csCode.json";
876 $data = $this->readJSONFile( $fileName );
877
878 foreach ( $data as $key => $item ) {
879 $this->mergeItem( $key, $allData[$key], $item );
880 }
881
882 $deps[] = new FileDependency( $fileName );
883 }
884 }
885 }
886
887 foreach ( $wgExtensionMessagesFiles as $extension => $fileName ) {
888 if ( isset( $wgMessagesDirs[$extension] ) ) {
889 # Already loaded the JSON files for this extension; skip the PHP shim
890 continue;
891 }
892
893 $data = $this->readPHPFile( $fileName, 'extension' );
894 $used = false;
895
896 foreach ( $data as $key => $item ) {
897 if ( $this->mergeExtensionItem( $codeSequence, $key, $allData[$key], $item ) ) {
898 $used = true;
899 }
900 }
901
902 if ( $used ) {
903 $deps[] = new FileDependency( $fileName );
904 }
905 }
906
907 # Merge core data into extension data
908 foreach ( $coreData as $key => $item ) {
909 $this->mergeItem( $key, $allData[$key], $item );
910 }
911 wfProfileOut( __METHOD__ . '-extensions' );
912
913 # Add cache dependencies for any referenced globals
914 $deps['wgExtensionMessagesFiles'] = new GlobalDependency( 'wgExtensionMessagesFiles' );
915 $deps['wgMessagesDirs'] = new GlobalDependency( 'wgMessagesDirs' );
916 $deps['version'] = new ConstantDependency( 'MW_LC_VERSION' );
917
918 # Add dependencies to the cache entry
919 $allData['deps'] = $deps;
920
921 # Replace spaces with underscores in namespace names
922 $allData['namespaceNames'] = str_replace( ' ', '_', $allData['namespaceNames'] );
923
924 # And do the same for special page aliases. $page is an array.
925 foreach ( $allData['specialPageAliases'] as &$page ) {
926 $page = str_replace( ' ', '_', $page );
927 }
928 # Decouple the reference to prevent accidental damage
929 unset( $page );
930
931 # If there were no plural rules, return an empty array
932 if ( $allData['pluralRules'] === null ) {
933 $allData['pluralRules'] = array();
934 }
935 if ( $allData['compiledPluralRules'] === null ) {
936 $allData['compiledPluralRules'] = array();
937 }
938 # If there were no plural rule types, return an empty array
939 if ( $allData['pluralRuleTypes'] === null ) {
940 $allData['pluralRuleTypes'] = array();
941 }
942
943 # Set the list keys
944 $allData['list'] = array();
945 foreach ( self::$splitKeys as $key ) {
946 $allData['list'][$key] = array_keys( $allData[$key] );
947 }
948 # Run hooks
949 $purgeBlobs = true;
950 wfRunHooks( 'LocalisationCacheRecache', array( $this, $code, &$allData, &$purgeBlobs ) );
951
952 if ( is_null( $allData['namespaceNames'] ) ) {
953 wfProfileOut( __METHOD__ );
954 throw new MWException( __METHOD__ . ': Localisation data failed sanity check! ' .
955 'Check that your languages/messages/MessagesEn.php file is intact.' );
956 }
957
958 # Set the preload key
959 $allData['preload'] = $this->buildPreload( $allData );
960
961 # Save to the process cache and register the items loaded
962 $this->data[$code] = $allData;
963 foreach ( $allData as $key => $item ) {
964 $this->loadedItems[$code][$key] = true;
965 }
966
967 # Save to the persistent cache
968 wfProfileIn( __METHOD__ . '-write' );
969 $this->store->startWrite( $code );
970 foreach ( $allData as $key => $value ) {
971 if ( in_array( $key, self::$splitKeys ) ) {
972 foreach ( $value as $subkey => $subvalue ) {
973 $this->store->set( "$key:$subkey", $subvalue );
974 }
975 } else {
976 $this->store->set( $key, $value );
977 }
978 }
979 $this->store->finishWrite();
980 wfProfileOut( __METHOD__ . '-write' );
981
982 # Clear out the MessageBlobStore
983 # HACK: If using a null (i.e. disabled) storage backend, we
984 # can't write to the MessageBlobStore either
985 if ( $purgeBlobs && !$this->store instanceof LCStoreNull ) {
986 MessageBlobStore::clear();
987 }
988
989 wfProfileOut( __METHOD__ );
990 }
991
992 /**
993 * Build the preload item from the given pre-cache data.
994 *
995 * The preload item will be loaded automatically, improving performance
996 * for the commonly-requested items it contains.
997 * @param array $data
998 * @return array
999 */
1000 protected function buildPreload( $data ) {
1001 $preload = array( 'messages' => array() );
1002 foreach ( self::$preloadedKeys as $key ) {
1003 $preload[$key] = $data[$key];
1004 }
1005
1006 foreach ( $data['preloadedMessages'] as $subkey ) {
1007 if ( isset( $data['messages'][$subkey] ) ) {
1008 $subitem = $data['messages'][$subkey];
1009 } else {
1010 $subitem = null;
1011 }
1012 $preload['messages'][$subkey] = $subitem;
1013 }
1014
1015 return $preload;
1016 }
1017
1018 /**
1019 * Unload the data for a given language from the object cache.
1020 * Reduces memory usage.
1021 * @param string $code
1022 */
1023 public function unload( $code ) {
1024 unset( $this->data[$code] );
1025 unset( $this->loadedItems[$code] );
1026 unset( $this->loadedSubitems[$code] );
1027 unset( $this->initialisedLangs[$code] );
1028 unset( $this->shallowFallbacks[$code] );
1029
1030 foreach ( $this->shallowFallbacks as $shallowCode => $fbCode ) {
1031 if ( $fbCode === $code ) {
1032 $this->unload( $shallowCode );
1033 }
1034 }
1035 }
1036
1037 /**
1038 * Unload all data
1039 */
1040 public function unloadAll() {
1041 foreach ( $this->initialisedLangs as $lang => $unused ) {
1042 $this->unload( $lang );
1043 }
1044 }
1045
1046 /**
1047 * Disable the storage backend
1048 */
1049 public function disableBackend() {
1050 $this->store = new LCStoreNull;
1051 $this->manualRecache = false;
1052 }
1053 }
1054
1055 /**
1056 * Interface for the persistence layer of LocalisationCache.
1057 *
1058 * The persistence layer is two-level hierarchical cache. The first level
1059 * is the language, the second level is the item or subitem.
1060 *
1061 * Since the data for a whole language is rebuilt in one operation, it needs
1062 * to have a fast and atomic method for deleting or replacing all of the
1063 * current data for a given language. The interface reflects this bulk update
1064 * operation. Callers writing to the cache must first call startWrite(), then
1065 * will call set() a couple of thousand times, then will call finishWrite()
1066 * to commit the operation. When finishWrite() is called, the cache is
1067 * expected to delete all data previously stored for that language.
1068 *
1069 * The values stored are PHP variables suitable for serialize(). Implementations
1070 * of LCStore are responsible for serializing and unserializing.
1071 */
1072 interface LCStore {
1073 /**
1074 * Get a value.
1075 * @param string $code Language code
1076 * @param string $key Cache key
1077 */
1078 function get( $code, $key );
1079
1080 /**
1081 * Start a write transaction.
1082 * @param string $code Language code
1083 */
1084 function startWrite( $code );
1085
1086 /**
1087 * Finish a write transaction.
1088 */
1089 function finishWrite();
1090
1091 /**
1092 * Set a key to a given value. startWrite() must be called before this
1093 * is called, and finishWrite() must be called afterwards.
1094 * @param string $key
1095 * @param mixed $value
1096 */
1097 function set( $key, $value );
1098 }
1099
1100 /**
1101 * LCStore implementation which uses PHP accelerator to store data.
1102 * This will work if one of XCache, WinCache or APC cacher is configured.
1103 * (See ObjectCache.php)
1104 */
1105 class LCStoreAccel implements LCStore {
1106 private $currentLang;
1107 private $keys;
1108
1109 public function __construct() {
1110 $this->cache = wfGetCache( CACHE_ACCEL );
1111 }
1112
1113 public function get( $code, $key ) {
1114 $k = wfMemcKey( 'l10n', $code, 'k', $key );
1115 $r = $this->cache->get( $k );
1116
1117 return $r === false ? null : $r;
1118 }
1119
1120 public function startWrite( $code ) {
1121 $k = wfMemcKey( 'l10n', $code, 'l' );
1122 $keys = $this->cache->get( $k );
1123 if ( $keys ) {
1124 foreach ( $keys as $k ) {
1125 $this->cache->delete( $k );
1126 }
1127 }
1128 $this->currentLang = $code;
1129 $this->keys = array();
1130 }
1131
1132 public function finishWrite() {
1133 if ( $this->currentLang ) {
1134 $k = wfMemcKey( 'l10n', $this->currentLang, 'l' );
1135 $this->cache->set( $k, array_keys( $this->keys ) );
1136 }
1137 $this->currentLang = null;
1138 $this->keys = array();
1139 }
1140
1141 public function set( $key, $value ) {
1142 if ( $this->currentLang ) {
1143 $k = wfMemcKey( 'l10n', $this->currentLang, 'k', $key );
1144 $this->keys[$k] = true;
1145 $this->cache->set( $k, $value );
1146 }
1147 }
1148 }
1149
1150 /**
1151 * LCStore implementation which uses the standard DB functions to store data.
1152 * This will work on any MediaWiki installation.
1153 */
1154 class LCStoreDB implements LCStore {
1155 private $currentLang;
1156 private $writesDone = false;
1157
1158 /** @var DatabaseBase */
1159 private $dbw;
1160 /** @var array */
1161 private $batch = array();
1162
1163 private $readOnly = false;
1164
1165 public function get( $code, $key ) {
1166 if ( $this->writesDone && $this->dbw ) {
1167 $db = $this->dbw;
1168 } else {
1169 $db = wfGetDB( DB_SLAVE );
1170 }
1171 $row = $db->selectRow( 'l10n_cache', array( 'lc_value' ),
1172 array( 'lc_lang' => $code, 'lc_key' => $key ), __METHOD__ );
1173 if ( $row ) {
1174 return unserialize( $db->decodeBlob( $row->lc_value ) );
1175 } else {
1176 return null;
1177 }
1178 }
1179
1180 public function startWrite( $code ) {
1181 if ( $this->readOnly ) {
1182 return;
1183 } elseif ( !$code ) {
1184 throw new MWException( __METHOD__ . ": Invalid language \"$code\"" );
1185 }
1186
1187 // We must keep a separate connection to MySQL in order to avoid breaking
1188 // main transactions. However, SQLite deadlocks when using two connections.
1189 // @todo get this trick to work on PostgreSQL too
1190 if ( wfGetDB( DB_MASTER )->getType() == 'mysql' ) {
1191 $lb = wfGetLBFactory()->newMainLB();
1192 $this->dbw = $lb->getConnection( DB_MASTER );
1193 $this->dbw->clearFlag( DBO_TRX ); // auto-commit mode
1194 } else {
1195 $this->dbw = wfGetDB( DB_MASTER );
1196 }
1197
1198 $this->currentLang = $code;
1199 $this->batch = array();
1200 }
1201
1202 public function finishWrite() {
1203 if ( $this->readOnly ) {
1204 return;
1205 } elseif ( is_null( $this->currentLang ) ) {
1206 throw new MWException( __CLASS__ . ': must call startWrite() before finishWrite()' );
1207 }
1208
1209 $this->dbw->begin( __METHOD__ );
1210 try {
1211 $this->dbw->delete( 'l10n_cache',
1212 array( 'lc_lang' => $this->currentLang ), __METHOD__ );
1213 foreach ( array_chunk( $this->batch, 500 ) as $rows ) {
1214 $this->dbw->insert( 'l10n_cache', $rows, __METHOD__ );
1215 }
1216 $this->writesDone = true;
1217 } catch ( DBQueryError $e ) {
1218 if ( $this->dbw->wasReadOnlyError() ) {
1219 $this->readOnly = true; // just avoid site down time
1220 } else {
1221 throw $e;
1222 }
1223 }
1224 $this->dbw->commit( __METHOD__ );
1225
1226 $this->currentLang = null;
1227 $this->batch = array();
1228 }
1229
1230 public function set( $key, $value ) {
1231 if ( $this->readOnly ) {
1232 return;
1233 } elseif ( is_null( $this->currentLang ) ) {
1234 throw new MWException( __CLASS__ . ': must call startWrite() before set()' );
1235 }
1236
1237 $this->batch[] = array(
1238 'lc_lang' => $this->currentLang,
1239 'lc_key' => $key,
1240 'lc_value' => $this->dbw->encodeBlob( serialize( $value ) ) );
1241 }
1242 }
1243
1244 /**
1245 * LCStore implementation which stores data as a collection of CDB files in the
1246 * directory given by $wgCacheDirectory. If $wgCacheDirectory is not set, this
1247 * will throw an exception.
1248 *
1249 * Profiling indicates that on Linux, this implementation outperforms MySQL if
1250 * the directory is on a local filesystem and there is ample kernel cache
1251 * space. The performance advantage is greater when the DBA extension is
1252 * available than it is with the PHP port.
1253 *
1254 * See Cdb.php and http://cr.yp.to/cdb.html
1255 */
1256 class LCStoreCDB implements LCStore {
1257 /** @var CdbReader[] */
1258 private $readers;
1259
1260 /** @var CdbWriter */
1261 private $writer;
1262
1263 /** @var string Current language code */
1264 private $currentLang;
1265
1266 /** @var bool|string Cache directory. False if not set */
1267 private $directory;
1268
1269 function __construct( $conf = array() ) {
1270 global $wgCacheDirectory;
1271
1272 if ( isset( $conf['directory'] ) ) {
1273 $this->directory = $conf['directory'];
1274 } else {
1275 $this->directory = $wgCacheDirectory;
1276 }
1277 }
1278
1279 public function get( $code, $key ) {
1280 if ( !isset( $this->readers[$code] ) ) {
1281 $fileName = $this->getFileName( $code );
1282
1283 $this->readers[$code] = false;
1284 if ( file_exists( $fileName ) ) {
1285 try {
1286 $this->readers[$code] = CdbReader::open( $fileName );
1287 } catch ( CdbException $e ) {
1288 wfDebug( __METHOD__ . ": unable to open cdb file for reading\n" );
1289 }
1290 }
1291 }
1292
1293 if ( !$this->readers[$code] ) {
1294 return null;
1295 } else {
1296 $value = false;
1297 try {
1298 $value = $this->readers[$code]->get( $key );
1299 } catch ( CdbException $e ) {
1300 wfDebug( __METHOD__ . ": CdbException caught, error message was "
1301 . $e->getMessage() . "\n" );
1302 }
1303 if ( $value === false ) {
1304 return null;
1305 }
1306
1307 return unserialize( $value );
1308 }
1309 }
1310
1311 public function startWrite( $code ) {
1312 if ( !file_exists( $this->directory ) ) {
1313 if ( !wfMkdirParents( $this->directory, null, __METHOD__ ) ) {
1314 throw new MWException( "Unable to create the localisation store " .
1315 "directory \"{$this->directory}\"" );
1316 }
1317 }
1318
1319 // Close reader to stop permission errors on write
1320 if ( !empty( $this->readers[$code] ) ) {
1321 $this->readers[$code]->close();
1322 }
1323
1324 try {
1325 $this->writer = CdbWriter::open( $this->getFileName( $code ) );
1326 } catch ( CdbException $e ) {
1327 throw new MWException( $e->getMessage() );
1328 }
1329 $this->currentLang = $code;
1330 }
1331
1332 public function finishWrite() {
1333 // Close the writer
1334 try {
1335 $this->writer->close();
1336 } catch ( CdbException $e ) {
1337 throw new MWException( $e->getMessage() );
1338 }
1339 $this->writer = null;
1340 unset( $this->readers[$this->currentLang] );
1341 $this->currentLang = null;
1342 }
1343
1344 public function set( $key, $value ) {
1345 if ( is_null( $this->writer ) ) {
1346 throw new MWException( __CLASS__ . ': must call startWrite() before calling set()' );
1347 }
1348 try {
1349 $this->writer->set( $key, serialize( $value ) );
1350 } catch ( CdbException $e ) {
1351 throw new MWException( $e->getMessage() );
1352 }
1353 }
1354
1355 protected function getFileName( $code ) {
1356 if ( strval( $code ) === '' || strpos( $code, '/' ) !== false ) {
1357 throw new MWException( __METHOD__ . ": Invalid language \"$code\"" );
1358 }
1359
1360 return "{$this->directory}/l10n_cache-$code.cdb";
1361 }
1362 }
1363
1364 /**
1365 * Null store backend, used to avoid DB errors during install
1366 */
1367 class LCStoreNull implements LCStore {
1368 public function get( $code, $key ) {
1369 return null;
1370 }
1371
1372 public function startWrite( $code ) {
1373 }
1374
1375 public function finishWrite() {
1376 }
1377
1378 public function set( $key, $value ) {
1379 }
1380 }
1381
1382 /**
1383 * A localisation cache optimised for loading large amounts of data for many
1384 * languages. Used by rebuildLocalisationCache.php.
1385 */
1386 class LocalisationCacheBulkLoad extends LocalisationCache {
1387 /**
1388 * A cache of the contents of data files.
1389 * Core files are serialized to avoid using ~1GB of RAM during a recache.
1390 */
1391 private $fileCache = array();
1392
1393 /**
1394 * Most recently used languages. Uses the linked-list aspect of PHP hashtables
1395 * to keep the most recently used language codes at the end of the array, and
1396 * the language codes that are ready to be deleted at the beginning.
1397 */
1398 private $mruLangs = array();
1399
1400 /**
1401 * Maximum number of languages that may be loaded into $this->data
1402 */
1403 private $maxLoadedLangs = 10;
1404
1405 /**
1406 * @param string $fileName
1407 * @param string $fileType
1408 * @return array|mixed
1409 */
1410 protected function readPHPFile( $fileName, $fileType ) {
1411 $serialize = $fileType === 'core';
1412 if ( !isset( $this->fileCache[$fileName][$fileType] ) ) {
1413 $data = parent::readPHPFile( $fileName, $fileType );
1414
1415 if ( $serialize ) {
1416 $encData = serialize( $data );
1417 } else {
1418 $encData = $data;
1419 }
1420
1421 $this->fileCache[$fileName][$fileType] = $encData;
1422
1423 return $data;
1424 } elseif ( $serialize ) {
1425 return unserialize( $this->fileCache[$fileName][$fileType] );
1426 } else {
1427 return $this->fileCache[$fileName][$fileType];
1428 }
1429 }
1430
1431 /**
1432 * @param string $code
1433 * @param string $key
1434 * @return mixed
1435 */
1436 public function getItem( $code, $key ) {
1437 unset( $this->mruLangs[$code] );
1438 $this->mruLangs[$code] = true;
1439
1440 return parent::getItem( $code, $key );
1441 }
1442
1443 /**
1444 * @param string $code
1445 * @param string $key
1446 * @param string $subkey
1447 * @return mixed
1448 */
1449 public function getSubitem( $code, $key, $subkey ) {
1450 unset( $this->mruLangs[$code] );
1451 $this->mruLangs[$code] = true;
1452
1453 return parent::getSubitem( $code, $key, $subkey );
1454 }
1455
1456 /**
1457 * @param string $code
1458 */
1459 public function recache( $code ) {
1460 parent::recache( $code );
1461 unset( $this->mruLangs[$code] );
1462 $this->mruLangs[$code] = true;
1463 $this->trimCache();
1464 }
1465
1466 /**
1467 * @param string $code
1468 */
1469 public function unload( $code ) {
1470 unset( $this->mruLangs[$code] );
1471 parent::unload( $code );
1472 }
1473
1474 /**
1475 * Unload cached languages until there are less than $this->maxLoadedLangs
1476 */
1477 protected function trimCache() {
1478 while ( count( $this->data ) > $this->maxLoadedLangs && count( $this->mruLangs ) ) {
1479 reset( $this->mruLangs );
1480 $code = key( $this->mruLangs );
1481 wfDebug( __METHOD__ . ": unloading $code\n" );
1482 $this->unload( $code );
1483 }
1484 }
1485 }