Merge "Added ProfilerSectionOnly class"
[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
563 // The JSON format only supports messages, none of the other variables, so wrap the data
564 return array( 'messages' => $data );
565 }
566
567 /**
568 * Get the compiled plural rules for a given language from the XML files.
569 * @since 1.20
570 * @param string $code
571 * @return array|null
572 */
573 public function getCompiledPluralRules( $code ) {
574 $rules = $this->getPluralRules( $code );
575 if ( $rules === null ) {
576 return null;
577 }
578 try {
579 $compiledRules = CLDRPluralRuleEvaluator::compile( $rules );
580 } catch ( CLDRPluralRuleError $e ) {
581 wfDebugLog( 'l10n', $e->getMessage() );
582
583 return array();
584 }
585
586 return $compiledRules;
587 }
588
589 /**
590 * Get the plural rules for a given language from the XML files.
591 * Cached.
592 * @since 1.20
593 * @param string $code
594 * @return array|null
595 */
596 public function getPluralRules( $code ) {
597 if ( $this->pluralRules === null ) {
598 $this->loadPluralFiles();
599 }
600 if ( !isset( $this->pluralRules[$code] ) ) {
601 return null;
602 } else {
603 return $this->pluralRules[$code];
604 }
605 }
606
607 /**
608 * Get the plural rule types for a given language from the XML files.
609 * Cached.
610 * @since 1.22
611 * @param string $code
612 * @return array|null
613 */
614 public function getPluralRuleTypes( $code ) {
615 if ( $this->pluralRuleTypes === null ) {
616 $this->loadPluralFiles();
617 }
618 if ( !isset( $this->pluralRuleTypes[$code] ) ) {
619 return null;
620 } else {
621 return $this->pluralRuleTypes[$code];
622 }
623 }
624
625 /**
626 * Load the plural XML files.
627 */
628 protected function loadPluralFiles() {
629 global $IP;
630 $cldrPlural = "$IP/languages/data/plurals.xml";
631 $mwPlural = "$IP/languages/data/plurals-mediawiki.xml";
632 // Load CLDR plural rules
633 $this->loadPluralFile( $cldrPlural );
634 if ( file_exists( $mwPlural ) ) {
635 // Override or extend
636 $this->loadPluralFile( $mwPlural );
637 }
638 }
639
640 /**
641 * Load a plural XML file with the given filename, compile the relevant
642 * rules, and save the compiled rules in a process-local cache.
643 *
644 * @param string $fileName
645 * @throws MWException
646 */
647 protected function loadPluralFile( $fileName ) {
648 // Use file_get_contents instead of DOMDocument::load (T58439)
649 $xml = file_get_contents( $fileName );
650 if ( !$xml ) {
651 throw new MWException( "Unable to read plurals file $fileName" );
652 }
653 $doc = new DOMDocument;
654 $doc->loadXML( $xml );
655 $rulesets = $doc->getElementsByTagName( "pluralRules" );
656 foreach ( $rulesets as $ruleset ) {
657 $codes = $ruleset->getAttribute( 'locales' );
658 $rules = array();
659 $ruleTypes = array();
660 $ruleElements = $ruleset->getElementsByTagName( "pluralRule" );
661 foreach ( $ruleElements as $elt ) {
662 $ruleType = $elt->getAttribute( 'count' );
663 if ( $ruleType === 'other' ) {
664 // Don't record "other" rules, which have an empty condition
665 continue;
666 }
667 $rules[] = $elt->nodeValue;
668 $ruleTypes[] = $ruleType;
669 }
670 foreach ( explode( ' ', $codes ) as $code ) {
671 $this->pluralRules[$code] = $rules;
672 $this->pluralRuleTypes[$code] = $ruleTypes;
673 }
674 }
675 }
676
677 /**
678 * Read the data from the source files for a given language, and register
679 * the relevant dependencies in the $deps array. If the localisation
680 * exists, the data array is returned, otherwise false is returned.
681 *
682 * @param string $code
683 * @param array $deps
684 * @return array
685 */
686 protected function readSourceFilesAndRegisterDeps( $code, &$deps ) {
687 global $IP;
688
689 // This reads in the PHP i18n file with non-messages l10n data
690 $fileName = Language::getMessagesFileName( $code );
691 if ( !file_exists( $fileName ) ) {
692 $data = array();
693 } else {
694 $deps[] = new FileDependency( $fileName );
695 $data = $this->readPHPFile( $fileName, 'core' );
696 }
697
698 # Load CLDR plural rules for JavaScript
699 $data['pluralRules'] = $this->getPluralRules( $code );
700 # And for PHP
701 $data['compiledPluralRules'] = $this->getCompiledPluralRules( $code );
702 # Load plural rule types
703 $data['pluralRuleTypes'] = $this->getPluralRuleTypes( $code );
704
705 $deps['plurals'] = new FileDependency( "$IP/languages/data/plurals.xml" );
706 $deps['plurals-mw'] = new FileDependency( "$IP/languages/data/plurals-mediawiki.xml" );
707
708
709 return $data;
710 }
711
712 /**
713 * Merge two localisation values, a primary and a fallback, overwriting the
714 * primary value in place.
715 * @param string $key
716 * @param mixed $value
717 * @param mixed $fallbackValue
718 */
719 protected function mergeItem( $key, &$value, $fallbackValue ) {
720 if ( !is_null( $value ) ) {
721 if ( !is_null( $fallbackValue ) ) {
722 if ( in_array( $key, self::$mergeableMapKeys ) ) {
723 $value = $value + $fallbackValue;
724 } elseif ( in_array( $key, self::$mergeableListKeys ) ) {
725 $value = array_unique( array_merge( $fallbackValue, $value ) );
726 } elseif ( in_array( $key, self::$mergeableAliasListKeys ) ) {
727 $value = array_merge_recursive( $value, $fallbackValue );
728 } elseif ( in_array( $key, self::$optionalMergeKeys ) ) {
729 if ( !empty( $value['inherit'] ) ) {
730 $value = array_merge( $fallbackValue, $value );
731 }
732
733 if ( isset( $value['inherit'] ) ) {
734 unset( $value['inherit'] );
735 }
736 } elseif ( in_array( $key, self::$magicWordKeys ) ) {
737 $this->mergeMagicWords( $value, $fallbackValue );
738 }
739 }
740 } else {
741 $value = $fallbackValue;
742 }
743 }
744
745 /**
746 * @param mixed $value
747 * @param mixed $fallbackValue
748 */
749 protected function mergeMagicWords( &$value, $fallbackValue ) {
750 foreach ( $fallbackValue as $magicName => $fallbackInfo ) {
751 if ( !isset( $value[$magicName] ) ) {
752 $value[$magicName] = $fallbackInfo;
753 } else {
754 $oldSynonyms = array_slice( $fallbackInfo, 1 );
755 $newSynonyms = array_slice( $value[$magicName], 1 );
756 $synonyms = array_values( array_unique( array_merge(
757 $newSynonyms, $oldSynonyms ) ) );
758 $value[$magicName] = array_merge( array( $fallbackInfo[0] ), $synonyms );
759 }
760 }
761 }
762
763 /**
764 * Given an array mapping language code to localisation value, such as is
765 * found in extension *.i18n.php files, iterate through a fallback sequence
766 * to merge the given data with an existing primary value.
767 *
768 * Returns true if any data from the extension array was used, false
769 * otherwise.
770 * @param array $codeSequence
771 * @param string $key
772 * @param mixed $value
773 * @param mixed $fallbackValue
774 * @return bool
775 */
776 protected function mergeExtensionItem( $codeSequence, $key, &$value, $fallbackValue ) {
777 $used = false;
778 foreach ( $codeSequence as $code ) {
779 if ( isset( $fallbackValue[$code] ) ) {
780 $this->mergeItem( $key, $value, $fallbackValue[$code] );
781 $used = true;
782 }
783 }
784
785 return $used;
786 }
787
788 /**
789 * Gets the combined list of messages dirs from
790 * core and extensions
791 *
792 * @since 1.25
793 * @return array
794 */
795 public function getMessagesDirs() {
796 global $wgMessagesDirs, $IP;
797 return array(
798 'core' => "$IP/languages/i18n",
799 'api' => "$IP/includes/api/i18n",
800 'oojs-ui' => "$IP/resources/lib/oojs-ui/i18n",
801 ) + $wgMessagesDirs;
802 }
803
804 /**
805 * Load localisation data for a given language for both core and extensions
806 * and save it to the persistent cache store and the process cache
807 * @param string $code
808 * @throws MWException
809 */
810 public function recache( $code ) {
811 global $wgExtensionMessagesFiles;
812
813 if ( !$code ) {
814 throw new MWException( "Invalid language code requested" );
815 }
816 $this->recachedLangs[$code] = true;
817
818 # Initial values
819 $initialData = array_combine(
820 self::$allKeys,
821 array_fill( 0, count( self::$allKeys ), null ) );
822 $coreData = $initialData;
823 $deps = array();
824
825 # Load the primary localisation from the source file
826 $data = $this->readSourceFilesAndRegisterDeps( $code, $deps );
827 if ( $data === false ) {
828 wfDebug( __METHOD__ . ": no localisation file for $code, using fallback to en\n" );
829 $coreData['fallback'] = 'en';
830 } else {
831 wfDebug( __METHOD__ . ": got localisation for $code from source\n" );
832
833 # Merge primary localisation
834 foreach ( $data as $key => $value ) {
835 $this->mergeItem( $key, $coreData[$key], $value );
836 }
837 }
838
839 # Fill in the fallback if it's not there already
840 if ( is_null( $coreData['fallback'] ) ) {
841 $coreData['fallback'] = $code === 'en' ? false : 'en';
842 }
843 if ( $coreData['fallback'] === false ) {
844 $coreData['fallbackSequence'] = array();
845 } else {
846 $coreData['fallbackSequence'] = array_map( 'trim', explode( ',', $coreData['fallback'] ) );
847 $len = count( $coreData['fallbackSequence'] );
848
849 # Ensure that the sequence ends at en
850 if ( $coreData['fallbackSequence'][$len - 1] !== 'en' ) {
851 $coreData['fallbackSequence'][] = 'en';
852 }
853 }
854
855 $codeSequence = array_merge( array( $code ), $coreData['fallbackSequence'] );
856 $messageDirs = $this->getMessagesDirs();
857
858
859 # Load non-JSON localisation data for extensions
860 $extensionData = array_combine(
861 $codeSequence,
862 array_fill( 0, count( $codeSequence ), $initialData ) );
863 foreach ( $wgExtensionMessagesFiles as $extension => $fileName ) {
864 if ( isset( $messageDirs[$extension] ) ) {
865 # This extension has JSON message data; skip the PHP shim
866 continue;
867 }
868
869 $data = $this->readPHPFile( $fileName, 'extension' );
870 $used = false;
871
872 foreach ( $data as $key => $item ) {
873 foreach ( $codeSequence as $csCode ) {
874 if ( isset( $item[$csCode] ) ) {
875 $this->mergeItem( $key, $extensionData[$csCode][$key], $item[$csCode] );
876 $used = true;
877 }
878 }
879 }
880
881 if ( $used ) {
882 $deps[] = new FileDependency( $fileName );
883 }
884 }
885
886 # Load the localisation data for each fallback, then merge it into the full array
887 $allData = $initialData;
888 foreach ( $codeSequence as $csCode ) {
889 $csData = $initialData;
890
891 # Load core messages and the extension localisations.
892 foreach ( $messageDirs as $dirs ) {
893 foreach ( (array)$dirs as $dir ) {
894 $fileName = "$dir/$csCode.json";
895 $data = $this->readJSONFile( $fileName );
896
897 foreach ( $data as $key => $item ) {
898 $this->mergeItem( $key, $csData[$key], $item );
899 }
900
901 $deps[] = new FileDependency( $fileName );
902 }
903 }
904
905 # Merge non-JSON extension data
906 if ( isset( $extensionData[$csCode] ) ) {
907 foreach ( $extensionData[$csCode] as $key => $item ) {
908 $this->mergeItem( $key, $csData[$key], $item );
909 }
910 }
911
912 if ( $csCode === $code ) {
913 # Merge core data into extension data
914 foreach ( $coreData as $key => $item ) {
915 $this->mergeItem( $key, $csData[$key], $item );
916 }
917 } else {
918 # Load the secondary localisation from the source file to
919 # avoid infinite cycles on cyclic fallbacks
920 $fbData = $this->readSourceFilesAndRegisterDeps( $csCode, $deps );
921 if ( $fbData !== false ) {
922 # Only merge the keys that make sense to merge
923 foreach ( self::$allKeys as $key ) {
924 if ( !isset( $fbData[$key] ) ) {
925 continue;
926 }
927
928 if ( is_null( $coreData[$key] ) || $this->isMergeableKey( $key ) ) {
929 $this->mergeItem( $key, $csData[$key], $fbData[$key] );
930 }
931 }
932 }
933 }
934
935 # Allow extensions an opportunity to adjust the data for this
936 # fallback
937 Hooks::run( 'LocalisationCacheRecacheFallback', array( $this, $csCode, &$csData ) );
938
939 # Merge the data for this fallback into the final array
940 if ( $csCode === $code ) {
941 $allData = $csData;
942 } else {
943 foreach ( self::$allKeys as $key ) {
944 if ( !isset( $csData[$key] ) ) {
945 continue;
946 }
947
948 if ( is_null( $allData[$key] ) || $this->isMergeableKey( $key ) ) {
949 $this->mergeItem( $key, $allData[$key], $csData[$key] );
950 }
951 }
952 }
953 }
954
955
956 # Add cache dependencies for any referenced globals
957 $deps['wgExtensionMessagesFiles'] = new GlobalDependency( 'wgExtensionMessagesFiles' );
958 // $wgMessagesDirs is used in LocalisationCache::getMessagesDirs()
959 $deps['wgMessagesDirs'] = new GlobalDependency( 'wgMessagesDirs' );
960 $deps['version'] = new ConstantDependency( 'LocalisationCache::VERSION' );
961
962 # Add dependencies to the cache entry
963 $allData['deps'] = $deps;
964
965 # Replace spaces with underscores in namespace names
966 $allData['namespaceNames'] = str_replace( ' ', '_', $allData['namespaceNames'] );
967
968 # And do the same for special page aliases. $page is an array.
969 foreach ( $allData['specialPageAliases'] as &$page ) {
970 $page = str_replace( ' ', '_', $page );
971 }
972 # Decouple the reference to prevent accidental damage
973 unset( $page );
974
975 # If there were no plural rules, return an empty array
976 if ( $allData['pluralRules'] === null ) {
977 $allData['pluralRules'] = array();
978 }
979 if ( $allData['compiledPluralRules'] === null ) {
980 $allData['compiledPluralRules'] = array();
981 }
982 # If there were no plural rule types, return an empty array
983 if ( $allData['pluralRuleTypes'] === null ) {
984 $allData['pluralRuleTypes'] = array();
985 }
986
987 # Set the list keys
988 $allData['list'] = array();
989 foreach ( self::$splitKeys as $key ) {
990 $allData['list'][$key] = array_keys( $allData[$key] );
991 }
992 # Run hooks
993 $purgeBlobs = true;
994 Hooks::run( 'LocalisationCacheRecache', array( $this, $code, &$allData, &$purgeBlobs ) );
995
996 if ( is_null( $allData['namespaceNames'] ) ) {
997 throw new MWException( __METHOD__ . ': Localisation data failed sanity check! ' .
998 'Check that your languages/messages/MessagesEn.php file is intact.' );
999 }
1000
1001 # Set the preload key
1002 $allData['preload'] = $this->buildPreload( $allData );
1003
1004 # Save to the process cache and register the items loaded
1005 $this->data[$code] = $allData;
1006 foreach ( $allData as $key => $item ) {
1007 $this->loadedItems[$code][$key] = true;
1008 }
1009
1010 # Save to the persistent cache
1011 $this->store->startWrite( $code );
1012 foreach ( $allData as $key => $value ) {
1013 if ( in_array( $key, self::$splitKeys ) ) {
1014 foreach ( $value as $subkey => $subvalue ) {
1015 $this->store->set( "$key:$subkey", $subvalue );
1016 }
1017 } else {
1018 $this->store->set( $key, $value );
1019 }
1020 }
1021 $this->store->finishWrite();
1022
1023 # Clear out the MessageBlobStore
1024 # HACK: If using a null (i.e. disabled) storage backend, we
1025 # can't write to the MessageBlobStore either
1026 if ( $purgeBlobs && !$this->store instanceof LCStoreNull ) {
1027 MessageBlobStore::getInstance()->clear();
1028 }
1029
1030 }
1031
1032 /**
1033 * Build the preload item from the given pre-cache data.
1034 *
1035 * The preload item will be loaded automatically, improving performance
1036 * for the commonly-requested items it contains.
1037 * @param array $data
1038 * @return array
1039 */
1040 protected function buildPreload( $data ) {
1041 $preload = array( 'messages' => array() );
1042 foreach ( self::$preloadedKeys as $key ) {
1043 $preload[$key] = $data[$key];
1044 }
1045
1046 foreach ( $data['preloadedMessages'] as $subkey ) {
1047 if ( isset( $data['messages'][$subkey] ) ) {
1048 $subitem = $data['messages'][$subkey];
1049 } else {
1050 $subitem = null;
1051 }
1052 $preload['messages'][$subkey] = $subitem;
1053 }
1054
1055 return $preload;
1056 }
1057
1058 /**
1059 * Unload the data for a given language from the object cache.
1060 * Reduces memory usage.
1061 * @param string $code
1062 */
1063 public function unload( $code ) {
1064 unset( $this->data[$code] );
1065 unset( $this->loadedItems[$code] );
1066 unset( $this->loadedSubitems[$code] );
1067 unset( $this->initialisedLangs[$code] );
1068 unset( $this->shallowFallbacks[$code] );
1069
1070 foreach ( $this->shallowFallbacks as $shallowCode => $fbCode ) {
1071 if ( $fbCode === $code ) {
1072 $this->unload( $shallowCode );
1073 }
1074 }
1075 }
1076
1077 /**
1078 * Unload all data
1079 */
1080 public function unloadAll() {
1081 foreach ( $this->initialisedLangs as $lang => $unused ) {
1082 $this->unload( $lang );
1083 }
1084 }
1085
1086 /**
1087 * Disable the storage backend
1088 */
1089 public function disableBackend() {
1090 $this->store = new LCStoreNull;
1091 $this->manualRecache = false;
1092 }
1093 }
1094
1095 /**
1096 * Interface for the persistence layer of LocalisationCache.
1097 *
1098 * The persistence layer is two-level hierarchical cache. The first level
1099 * is the language, the second level is the item or subitem.
1100 *
1101 * Since the data for a whole language is rebuilt in one operation, it needs
1102 * to have a fast and atomic method for deleting or replacing all of the
1103 * current data for a given language. The interface reflects this bulk update
1104 * operation. Callers writing to the cache must first call startWrite(), then
1105 * will call set() a couple of thousand times, then will call finishWrite()
1106 * to commit the operation. When finishWrite() is called, the cache is
1107 * expected to delete all data previously stored for that language.
1108 *
1109 * The values stored are PHP variables suitable for serialize(). Implementations
1110 * of LCStore are responsible for serializing and unserializing.
1111 */
1112 interface LCStore {
1113 /**
1114 * Get a value.
1115 * @param string $code Language code
1116 * @param string $key Cache key
1117 */
1118 function get( $code, $key );
1119
1120 /**
1121 * Start a write transaction.
1122 * @param string $code Language code
1123 */
1124 function startWrite( $code );
1125
1126 /**
1127 * Finish a write transaction.
1128 */
1129 function finishWrite();
1130
1131 /**
1132 * Set a key to a given value. startWrite() must be called before this
1133 * is called, and finishWrite() must be called afterwards.
1134 * @param string $key
1135 * @param mixed $value
1136 */
1137 function set( $key, $value );
1138 }
1139
1140 /**
1141 * LCStore implementation which uses the standard DB functions to store data.
1142 * This will work on any MediaWiki installation.
1143 */
1144 class LCStoreDB implements LCStore {
1145 private $currentLang;
1146 private $writesDone = false;
1147
1148 /** @var DatabaseBase */
1149 private $dbw;
1150 /** @var array */
1151 private $batch = array();
1152
1153 private $readOnly = false;
1154
1155 public function get( $code, $key ) {
1156 if ( $this->writesDone ) {
1157 $db = wfGetDB( DB_MASTER );
1158 } else {
1159 $db = wfGetDB( DB_SLAVE );
1160 }
1161 $row = $db->selectRow( 'l10n_cache', array( 'lc_value' ),
1162 array( 'lc_lang' => $code, 'lc_key' => $key ), __METHOD__ );
1163 if ( $row ) {
1164 return unserialize( $db->decodeBlob( $row->lc_value ) );
1165 } else {
1166 return null;
1167 }
1168 }
1169
1170 public function startWrite( $code ) {
1171 if ( $this->readOnly ) {
1172 return;
1173 } elseif ( !$code ) {
1174 throw new MWException( __METHOD__ . ": Invalid language \"$code\"" );
1175 }
1176
1177 $this->dbw = wfGetDB( DB_MASTER );
1178
1179 $this->currentLang = $code;
1180 $this->batch = array();
1181 }
1182
1183 public function finishWrite() {
1184 if ( $this->readOnly ) {
1185 return;
1186 } elseif ( is_null( $this->currentLang ) ) {
1187 throw new MWException( __CLASS__ . ': must call startWrite() before finishWrite()' );
1188 }
1189
1190 $this->dbw->begin( __METHOD__ );
1191 try {
1192 $this->dbw->delete( 'l10n_cache',
1193 array( 'lc_lang' => $this->currentLang ), __METHOD__ );
1194 foreach ( array_chunk( $this->batch, 500 ) as $rows ) {
1195 $this->dbw->insert( 'l10n_cache', $rows, __METHOD__ );
1196 }
1197 $this->writesDone = true;
1198 } catch ( DBQueryError $e ) {
1199 if ( $this->dbw->wasReadOnlyError() ) {
1200 $this->readOnly = true; // just avoid site down time
1201 } else {
1202 throw $e;
1203 }
1204 }
1205 $this->dbw->commit( __METHOD__ );
1206
1207 $this->currentLang = null;
1208 $this->batch = array();
1209 }
1210
1211 public function set( $key, $value ) {
1212 if ( $this->readOnly ) {
1213 return;
1214 } elseif ( is_null( $this->currentLang ) ) {
1215 throw new MWException( __CLASS__ . ': must call startWrite() before set()' );
1216 }
1217
1218 $this->batch[] = array(
1219 'lc_lang' => $this->currentLang,
1220 'lc_key' => $key,
1221 'lc_value' => $this->dbw->encodeBlob( serialize( $value ) ) );
1222 }
1223 }
1224
1225 /**
1226 * LCStore implementation which stores data as a collection of CDB files in the
1227 * directory given by $wgCacheDirectory. If $wgCacheDirectory is not set, this
1228 * will throw an exception.
1229 *
1230 * Profiling indicates that on Linux, this implementation outperforms MySQL if
1231 * the directory is on a local filesystem and there is ample kernel cache
1232 * space. The performance advantage is greater when the DBA extension is
1233 * available than it is with the PHP port.
1234 *
1235 * See Cdb.php and http://cr.yp.to/cdb.html
1236 */
1237 class LCStoreCDB implements LCStore {
1238 /** @var CdbReader[] */
1239 private $readers;
1240
1241 /** @var CdbWriter */
1242 private $writer;
1243
1244 /** @var string Current language code */
1245 private $currentLang;
1246
1247 /** @var bool|string Cache directory. False if not set */
1248 private $directory;
1249
1250 function __construct( $conf = array() ) {
1251 global $wgCacheDirectory;
1252
1253 if ( isset( $conf['directory'] ) ) {
1254 $this->directory = $conf['directory'];
1255 } else {
1256 $this->directory = $wgCacheDirectory;
1257 }
1258 }
1259
1260 public function get( $code, $key ) {
1261 if ( !isset( $this->readers[$code] ) ) {
1262 $fileName = $this->getFileName( $code );
1263
1264 $this->readers[$code] = false;
1265 if ( file_exists( $fileName ) ) {
1266 try {
1267 $this->readers[$code] = CdbReader::open( $fileName );
1268 } catch ( CdbException $e ) {
1269 wfDebug( __METHOD__ . ": unable to open cdb file for reading\n" );
1270 }
1271 }
1272 }
1273
1274 if ( !$this->readers[$code] ) {
1275 return null;
1276 } else {
1277 $value = false;
1278 try {
1279 $value = $this->readers[$code]->get( $key );
1280 } catch ( CdbException $e ) {
1281 wfDebug( __METHOD__ . ": CdbException caught, error message was "
1282 . $e->getMessage() . "\n" );
1283 }
1284 if ( $value === false ) {
1285 return null;
1286 }
1287
1288 return unserialize( $value );
1289 }
1290 }
1291
1292 public function startWrite( $code ) {
1293 if ( !file_exists( $this->directory ) ) {
1294 if ( !wfMkdirParents( $this->directory, null, __METHOD__ ) ) {
1295 throw new MWException( "Unable to create the localisation store " .
1296 "directory \"{$this->directory}\"" );
1297 }
1298 }
1299
1300 // Close reader to stop permission errors on write
1301 if ( !empty( $this->readers[$code] ) ) {
1302 $this->readers[$code]->close();
1303 }
1304
1305 try {
1306 $this->writer = CdbWriter::open( $this->getFileName( $code ) );
1307 } catch ( CdbException $e ) {
1308 throw new MWException( $e->getMessage() );
1309 }
1310 $this->currentLang = $code;
1311 }
1312
1313 public function finishWrite() {
1314 // Close the writer
1315 try {
1316 $this->writer->close();
1317 } catch ( CdbException $e ) {
1318 throw new MWException( $e->getMessage() );
1319 }
1320 $this->writer = null;
1321 unset( $this->readers[$this->currentLang] );
1322 $this->currentLang = null;
1323 }
1324
1325 public function set( $key, $value ) {
1326 if ( is_null( $this->writer ) ) {
1327 throw new MWException( __CLASS__ . ': must call startWrite() before calling set()' );
1328 }
1329 try {
1330 $this->writer->set( $key, serialize( $value ) );
1331 } catch ( CdbException $e ) {
1332 throw new MWException( $e->getMessage() );
1333 }
1334 }
1335
1336 protected function getFileName( $code ) {
1337 if ( strval( $code ) === '' || strpos( $code, '/' ) !== false ) {
1338 throw new MWException( __METHOD__ . ": Invalid language \"$code\"" );
1339 }
1340
1341 return "{$this->directory}/l10n_cache-$code.cdb";
1342 }
1343 }
1344
1345 /**
1346 * Null store backend, used to avoid DB errors during install
1347 */
1348 class LCStoreNull implements LCStore {
1349 public function get( $code, $key ) {
1350 return null;
1351 }
1352
1353 public function startWrite( $code ) {
1354 }
1355
1356 public function finishWrite() {
1357 }
1358
1359 public function set( $key, $value ) {
1360 }
1361 }
1362
1363 /**
1364 * A localisation cache optimised for loading large amounts of data for many
1365 * languages. Used by rebuildLocalisationCache.php.
1366 */
1367 class LocalisationCacheBulkLoad extends LocalisationCache {
1368 /**
1369 * A cache of the contents of data files.
1370 * Core files are serialized to avoid using ~1GB of RAM during a recache.
1371 */
1372 private $fileCache = array();
1373
1374 /**
1375 * Most recently used languages. Uses the linked-list aspect of PHP hashtables
1376 * to keep the most recently used language codes at the end of the array, and
1377 * the language codes that are ready to be deleted at the beginning.
1378 */
1379 private $mruLangs = array();
1380
1381 /**
1382 * Maximum number of languages that may be loaded into $this->data
1383 */
1384 private $maxLoadedLangs = 10;
1385
1386 /**
1387 * @param string $fileName
1388 * @param string $fileType
1389 * @return array|mixed
1390 */
1391 protected function readPHPFile( $fileName, $fileType ) {
1392 $serialize = $fileType === 'core';
1393 if ( !isset( $this->fileCache[$fileName][$fileType] ) ) {
1394 $data = parent::readPHPFile( $fileName, $fileType );
1395
1396 if ( $serialize ) {
1397 $encData = serialize( $data );
1398 } else {
1399 $encData = $data;
1400 }
1401
1402 $this->fileCache[$fileName][$fileType] = $encData;
1403
1404 return $data;
1405 } elseif ( $serialize ) {
1406 return unserialize( $this->fileCache[$fileName][$fileType] );
1407 } else {
1408 return $this->fileCache[$fileName][$fileType];
1409 }
1410 }
1411
1412 /**
1413 * @param string $code
1414 * @param string $key
1415 * @return mixed
1416 */
1417 public function getItem( $code, $key ) {
1418 unset( $this->mruLangs[$code] );
1419 $this->mruLangs[$code] = true;
1420
1421 return parent::getItem( $code, $key );
1422 }
1423
1424 /**
1425 * @param string $code
1426 * @param string $key
1427 * @param string $subkey
1428 * @return mixed
1429 */
1430 public function getSubitem( $code, $key, $subkey ) {
1431 unset( $this->mruLangs[$code] );
1432 $this->mruLangs[$code] = true;
1433
1434 return parent::getSubitem( $code, $key, $subkey );
1435 }
1436
1437 /**
1438 * @param string $code
1439 */
1440 public function recache( $code ) {
1441 parent::recache( $code );
1442 unset( $this->mruLangs[$code] );
1443 $this->mruLangs[$code] = true;
1444 $this->trimCache();
1445 }
1446
1447 /**
1448 * @param string $code
1449 */
1450 public function unload( $code ) {
1451 unset( $this->mruLangs[$code] );
1452 parent::unload( $code );
1453 }
1454
1455 /**
1456 * Unload cached languages until there are less than $this->maxLoadedLangs
1457 */
1458 protected function trimCache() {
1459 while ( count( $this->data ) > $this->maxLoadedLangs && count( $this->mruLangs ) ) {
1460 reset( $this->mruLangs );
1461 $code = key( $this->mruLangs );
1462 wfDebug( __METHOD__ . ": unloading $code\n" );
1463 $this->unload( $code );
1464 }
1465 }
1466 }