Remove obvious function-level profiling
[lhc/web/wiklou.git] / includes / cache / LocalisationCache.php
1 <?php
2 /**
3 * Cache of the contents of localisation files.
4 *
5 * This program is free software; you can redistribute it and/or modify
6 * it under the terms of the GNU General Public License as published by
7 * the Free Software Foundation; either version 2 of the License, or
8 * (at your option) any later version.
9 *
10 * This program is distributed in the hope that it will be useful,
11 * but WITHOUT ANY WARRANTY; without even the implied warranty of
12 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
13 * GNU General Public License for more details.
14 *
15 * You should have received a copy of the GNU General Public License along
16 * with this program; if not, write to the Free Software Foundation, Inc.,
17 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
18 * http://www.gnu.org/copyleft/gpl.html
19 *
20 * @file
21 */
22
23 use Cdb\Exception as CdbException;
24 use Cdb\Reader as CdbReader;
25 use Cdb\Writer as CdbWriter;
26
27 /**
28 * Class for caching the contents of localisation files, Messages*.php
29 * and *.i18n.php.
30 *
31 * An instance of this class is available using Language::getLocalisationCache().
32 *
33 * The values retrieved from here are merged, containing items from extension
34 * files, core messages files and the language fallback sequence (e.g. zh-cn ->
35 * zh-hans -> en ). Some common errors are corrected, for example namespace
36 * names with spaces instead of underscores, but heavyweight processing, such
37 * as grammatical transformation, is done by the caller.
38 */
39 class LocalisationCache {
40 const VERSION = 3;
41
42 /** Configuration associative array */
43 private $conf;
44
45 /**
46 * True if recaching should only be done on an explicit call to recache().
47 * Setting this reduces the overhead of cache freshness checking, which
48 * requires doing a stat() for every extension i18n file.
49 */
50 private $manualRecache = false;
51
52 /**
53 * True to treat all files as expired until they are regenerated by this object.
54 */
55 private $forceRecache = false;
56
57 /**
58 * The cache data. 3-d array, where the first key is the language code,
59 * the second key is the item key e.g. 'messages', and the third key is
60 * an item specific subkey index. Some items are not arrays and so for those
61 * items, there are no subkeys.
62 */
63 protected $data = array();
64
65 /**
66 * The persistent store object. An instance of LCStore.
67 *
68 * @var LCStore
69 */
70 private $store;
71
72 /**
73 * A 2-d associative array, code/key, where presence indicates that the item
74 * is loaded. Value arbitrary.
75 *
76 * For split items, if set, this indicates that all of the subitems have been
77 * loaded.
78 */
79 private $loadedItems = array();
80
81 /**
82 * A 3-d associative array, code/key/subkey, where presence indicates that
83 * the subitem is loaded. Only used for the split items, i.e. messages.
84 */
85 private $loadedSubitems = array();
86
87 /**
88 * An array where presence of a key indicates that that language has been
89 * initialised. Initialisation includes checking for cache expiry and doing
90 * any necessary updates.
91 */
92 private $initialisedLangs = array();
93
94 /**
95 * An array mapping non-existent pseudo-languages to fallback languages. This
96 * is filled by initShallowFallback() when data is requested from a language
97 * that lacks a Messages*.php file.
98 */
99 private $shallowFallbacks = array();
100
101 /**
102 * An array where the keys are codes that have been recached by this instance.
103 */
104 private $recachedLangs = array();
105
106 /**
107 * All item keys
108 */
109 static public $allKeys = array(
110 'fallback', 'namespaceNames', 'bookstoreList',
111 'magicWords', 'messages', 'rtl', 'capitalizeAllNouns', 'digitTransformTable',
112 'separatorTransformTable', 'fallback8bitEncoding', 'linkPrefixExtension',
113 'linkTrail', 'linkPrefixCharset', 'namespaceAliases',
114 'dateFormats', 'datePreferences', 'datePreferenceMigrationMap',
115 'defaultDateFormat', 'extraUserToggles', 'specialPageAliases',
116 'imageFiles', 'preloadedMessages', 'namespaceGenderAliases',
117 'digitGroupingPattern', 'pluralRules', 'pluralRuleTypes', 'compiledPluralRules',
118 );
119
120 /**
121 * Keys for items which consist of associative arrays, which may be merged
122 * by a fallback sequence.
123 */
124 static public $mergeableMapKeys = array( 'messages', 'namespaceNames',
125 'dateFormats', 'imageFiles', 'preloadedMessages'
126 );
127
128 /**
129 * Keys for items which are a numbered array.
130 */
131 static public $mergeableListKeys = array( 'extraUserToggles' );
132
133 /**
134 * Keys for items which contain an array of arrays of equivalent aliases
135 * for each subitem. The aliases may be merged by a fallback sequence.
136 */
137 static public $mergeableAliasListKeys = array( 'specialPageAliases' );
138
139 /**
140 * Keys for items which contain an associative array, and may be merged if
141 * the primary value contains the special array key "inherit". That array
142 * key is removed after the first merge.
143 */
144 static public $optionalMergeKeys = array( 'bookstoreList' );
145
146 /**
147 * Keys for items that are formatted like $magicWords
148 */
149 static public $magicWordKeys = array( 'magicWords' );
150
151 /**
152 * Keys for items where the subitems are stored in the backend separately.
153 */
154 static public $splitKeys = array( 'messages' );
155
156 /**
157 * Keys which are loaded automatically by initLanguage()
158 */
159 static public $preloadedKeys = array( 'dateFormats', 'namespaceNames' );
160
161 /**
162 * Associative array of cached plural rules. The key is the language code,
163 * the value is an array of plural rules for that language.
164 */
165 private $pluralRules = null;
166
167 /**
168 * Associative array of cached plural rule types. The key is the language
169 * code, the value is an array of plural rule types for that language. For
170 * example, $pluralRuleTypes['ar'] = ['zero', 'one', 'two', 'few', 'many'].
171 * The index for each rule type matches the index for the rule in
172 * $pluralRules, thus allowing correlation between the two. The reason we
173 * don't just use the type names as the keys in $pluralRules is because
174 * Language::convertPlural applies the rules based on numeric order (or
175 * explicit numeric parameter), not based on the name of the rule type. For
176 * example, {{plural:count|wordform1|wordform2|wordform3}}, rather than
177 * {{plural:count|one=wordform1|two=wordform2|many=wordform3}}.
178 */
179 private $pluralRuleTypes = null;
180
181 private $mergeableKeys = null;
182
183 /**
184 * Constructor.
185 * For constructor parameters, see the documentation in DefaultSettings.php
186 * for $wgLocalisationCacheConf.
187 *
188 * @param array $conf
189 * @throws MWException
190 */
191 function __construct( $conf ) {
192 global $wgCacheDirectory;
193
194 $this->conf = $conf;
195 $storeConf = array();
196 if ( !empty( $conf['storeClass'] ) ) {
197 $storeClass = $conf['storeClass'];
198 } else {
199 switch ( $conf['store'] ) {
200 case 'files':
201 case 'file':
202 $storeClass = 'LCStoreCDB';
203 break;
204 case 'db':
205 $storeClass = 'LCStoreDB';
206 break;
207 case 'detect':
208 $storeClass = $wgCacheDirectory ? 'LCStoreCDB' : 'LCStoreDB';
209 break;
210 default:
211 throw new MWException(
212 'Please set $wgLocalisationCacheConf[\'store\'] to something sensible.' );
213 }
214 }
215
216 wfDebugLog( 'caches', get_class( $this ) . ": using store $storeClass" );
217 if ( !empty( $conf['storeDirectory'] ) ) {
218 $storeConf['directory'] = $conf['storeDirectory'];
219 }
220
221 $this->store = new $storeClass( $storeConf );
222 foreach ( array( 'manualRecache', 'forceRecache' ) as $var ) {
223 if ( isset( $conf[$var] ) ) {
224 $this->$var = $conf[$var];
225 }
226 }
227 }
228
229 /**
230 * Returns true if the given key is mergeable, that is, if it is an associative
231 * array which can be merged through a fallback sequence.
232 * @param string $key
233 * @return bool
234 */
235 public function isMergeableKey( $key ) {
236 if ( $this->mergeableKeys === null ) {
237 $this->mergeableKeys = array_flip( array_merge(
238 self::$mergeableMapKeys,
239 self::$mergeableListKeys,
240 self::$mergeableAliasListKeys,
241 self::$optionalMergeKeys,
242 self::$magicWordKeys
243 ) );
244 }
245
246 return isset( $this->mergeableKeys[$key] );
247 }
248
249 /**
250 * Get a cache item.
251 *
252 * Warning: this may be slow for split items (messages), since it will
253 * need to fetch all of the subitems from the cache individually.
254 * @param string $code
255 * @param string $key
256 * @return mixed
257 */
258 public function getItem( $code, $key ) {
259 if ( !isset( $this->loadedItems[$code][$key] ) ) {
260 wfProfileIn( __METHOD__ . '-load' );
261 $this->loadItem( $code, $key );
262 wfProfileOut( __METHOD__ . '-load' );
263 }
264
265 if ( $key === 'fallback' && isset( $this->shallowFallbacks[$code] ) ) {
266 return $this->shallowFallbacks[$code];
267 }
268
269 return $this->data[$code][$key];
270 }
271
272 /**
273 * Get a subitem, for instance a single message for a given language.
274 * @param string $code
275 * @param string $key
276 * @param string $subkey
277 * @return mixed|null
278 */
279 public function getSubitem( $code, $key, $subkey ) {
280 if ( !isset( $this->loadedSubitems[$code][$key][$subkey] ) &&
281 !isset( $this->loadedItems[$code][$key] )
282 ) {
283 wfProfileIn( __METHOD__ . '-load' );
284 $this->loadSubitem( $code, $key, $subkey );
285 wfProfileOut( __METHOD__ . '-load' );
286 }
287
288 if ( isset( $this->data[$code][$key][$subkey] ) ) {
289 return $this->data[$code][$key][$subkey];
290 } else {
291 return null;
292 }
293 }
294
295 /**
296 * Get the list of subitem keys for a given item.
297 *
298 * This is faster than array_keys($lc->getItem(...)) for the items listed in
299 * self::$splitKeys.
300 *
301 * Will return null if the item is not found, or false if the item is not an
302 * array.
303 * @param string $code
304 * @param string $key
305 * @return bool|null|string
306 */
307 public function getSubitemList( $code, $key ) {
308 if ( in_array( $key, self::$splitKeys ) ) {
309 return $this->getSubitem( $code, 'list', $key );
310 } else {
311 $item = $this->getItem( $code, $key );
312 if ( is_array( $item ) ) {
313 return array_keys( $item );
314 } else {
315 return false;
316 }
317 }
318 }
319
320 /**
321 * Load an item into the cache.
322 * @param string $code
323 * @param string $key
324 */
325 protected function loadItem( $code, $key ) {
326 if ( !isset( $this->initialisedLangs[$code] ) ) {
327 $this->initLanguage( $code );
328 }
329
330 // Check to see if initLanguage() loaded it for us
331 if ( isset( $this->loadedItems[$code][$key] ) ) {
332 return;
333 }
334
335 if ( isset( $this->shallowFallbacks[$code] ) ) {
336 $this->loadItem( $this->shallowFallbacks[$code], $key );
337
338 return;
339 }
340
341 if ( in_array( $key, self::$splitKeys ) ) {
342 $subkeyList = $this->getSubitem( $code, 'list', $key );
343 foreach ( $subkeyList as $subkey ) {
344 if ( isset( $this->data[$code][$key][$subkey] ) ) {
345 continue;
346 }
347 $this->data[$code][$key][$subkey] = $this->getSubitem( $code, $key, $subkey );
348 }
349 } else {
350 $this->data[$code][$key] = $this->store->get( $code, $key );
351 }
352
353 $this->loadedItems[$code][$key] = true;
354 }
355
356 /**
357 * Load a subitem into the cache
358 * @param string $code
359 * @param string $key
360 * @param string $subkey
361 */
362 protected function loadSubitem( $code, $key, $subkey ) {
363 if ( !in_array( $key, self::$splitKeys ) ) {
364 $this->loadItem( $code, $key );
365
366 return;
367 }
368
369 if ( !isset( $this->initialisedLangs[$code] ) ) {
370 $this->initLanguage( $code );
371 }
372
373 // Check to see if initLanguage() loaded it for us
374 if ( isset( $this->loadedItems[$code][$key] ) ||
375 isset( $this->loadedSubitems[$code][$key][$subkey] )
376 ) {
377 return;
378 }
379
380 if ( isset( $this->shallowFallbacks[$code] ) ) {
381 $this->loadSubitem( $this->shallowFallbacks[$code], $key, $subkey );
382
383 return;
384 }
385
386 $value = $this->store->get( $code, "$key:$subkey" );
387 $this->data[$code][$key][$subkey] = $value;
388 $this->loadedSubitems[$code][$key][$subkey] = true;
389 }
390
391 /**
392 * Returns true if the cache identified by $code is missing or expired.
393 *
394 * @param string $code
395 *
396 * @return bool
397 */
398 public function isExpired( $code ) {
399 if ( $this->forceRecache && !isset( $this->recachedLangs[$code] ) ) {
400 wfDebug( __METHOD__ . "($code): forced reload\n" );
401
402 return true;
403 }
404
405 $deps = $this->store->get( $code, 'deps' );
406 $keys = $this->store->get( $code, 'list' );
407 $preload = $this->store->get( $code, 'preload' );
408 // Different keys may expire separately for some stores
409 if ( $deps === null || $keys === null || $preload === null ) {
410 wfDebug( __METHOD__ . "($code): cache missing, need to make one\n" );
411
412 return true;
413 }
414
415 foreach ( $deps as $dep ) {
416 // Because we're unserializing stuff from cache, we
417 // could receive objects of classes that don't exist
418 // anymore (e.g. uninstalled extensions)
419 // When this happens, always expire the cache
420 if ( !$dep instanceof CacheDependency || $dep->isExpired() ) {
421 wfDebug( __METHOD__ . "($code): cache for $code expired due to " .
422 get_class( $dep ) . "\n" );
423
424 return true;
425 }
426 }
427
428 return false;
429 }
430
431 /**
432 * Initialise a language in this object. Rebuild the cache if necessary.
433 * @param string $code
434 * @throws MWException
435 */
436 protected function initLanguage( $code ) {
437 if ( isset( $this->initialisedLangs[$code] ) ) {
438 return;
439 }
440
441 $this->initialisedLangs[$code] = true;
442
443 # If the code is of the wrong form for a Messages*.php file, do a shallow fallback
444 if ( !Language::isValidBuiltInCode( $code ) ) {
445 $this->initShallowFallback( $code, 'en' );
446
447 return;
448 }
449
450 # Recache the data if necessary
451 if ( !$this->manualRecache && $this->isExpired( $code ) ) {
452 if ( Language::isSupportedLanguage( $code ) ) {
453 $this->recache( $code );
454 } elseif ( $code === 'en' ) {
455 throw new MWException( 'MessagesEn.php is missing.' );
456 } else {
457 $this->initShallowFallback( $code, 'en' );
458 }
459
460 return;
461 }
462
463 # Preload some stuff
464 $preload = $this->getItem( $code, 'preload' );
465 if ( $preload === null ) {
466 if ( $this->manualRecache ) {
467 // No Messages*.php file. Do shallow fallback to en.
468 if ( $code === 'en' ) {
469 throw new MWException( 'No localisation cache found for English. ' .
470 'Please run maintenance/rebuildLocalisationCache.php.' );
471 }
472 $this->initShallowFallback( $code, 'en' );
473
474 return;
475 } else {
476 throw new MWException( 'Invalid or missing localisation cache.' );
477 }
478 }
479 $this->data[$code] = $preload;
480 foreach ( $preload as $key => $item ) {
481 if ( in_array( $key, self::$splitKeys ) ) {
482 foreach ( $item as $subkey => $subitem ) {
483 $this->loadedSubitems[$code][$key][$subkey] = true;
484 }
485 } else {
486 $this->loadedItems[$code][$key] = true;
487 }
488 }
489 }
490
491 /**
492 * Create a fallback from one language to another, without creating a
493 * complete persistent cache.
494 * @param string $primaryCode
495 * @param string $fallbackCode
496 */
497 public function initShallowFallback( $primaryCode, $fallbackCode ) {
498 $this->data[$primaryCode] =& $this->data[$fallbackCode];
499 $this->loadedItems[$primaryCode] =& $this->loadedItems[$fallbackCode];
500 $this->loadedSubitems[$primaryCode] =& $this->loadedSubitems[$fallbackCode];
501 $this->shallowFallbacks[$primaryCode] = $fallbackCode;
502 }
503
504 /**
505 * Read a PHP file containing localisation data.
506 * @param string $_fileName
507 * @param string $_fileType
508 * @throws MWException
509 * @return array
510 */
511 protected function readPHPFile( $_fileName, $_fileType ) {
512 // 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 throw new MWException( __METHOD__ . ": Invalid file type: $_fileType" );
529 }
530
531 return $data;
532 }
533
534 /**
535 * Read a JSON file containing localisation messages.
536 * @param string $fileName Name of file to read
537 * @throws MWException If there is a syntax error in the JSON file
538 * @return array Array with a 'messages' key, or empty array if the file doesn't exist
539 */
540 public function readJSONFile( $fileName ) {
541
542 if ( !is_readable( $fileName ) ) {
543
544 return array();
545 }
546
547 $json = file_get_contents( $fileName );
548 if ( $json === false ) {
549
550 return array();
551 }
552
553 $data = FormatJson::decode( $json, true );
554 if ( $data === null ) {
555
556 throw new MWException( __METHOD__ . ": Invalid JSON file: $fileName" );
557 }
558
559 // Remove keys starting with '@', they're reserved for metadata and non-message data
560 foreach ( $data as $key => $unused ) {
561 if ( $key === '' || $key[0] === '@' ) {
562 unset( $data[$key] );
563 }
564 }
565
566
567 // The JSON format only supports messages, none of the other variables, so wrap the data
568 return array( 'messages' => $data );
569 }
570
571 /**
572 * Get the compiled plural rules for a given language from the XML files.
573 * @since 1.20
574 * @param string $code
575 * @return array|null
576 */
577 public function getCompiledPluralRules( $code ) {
578 $rules = $this->getPluralRules( $code );
579 if ( $rules === null ) {
580 return null;
581 }
582 try {
583 $compiledRules = CLDRPluralRuleEvaluator::compile( $rules );
584 } catch ( CLDRPluralRuleError $e ) {
585 wfDebugLog( 'l10n', $e->getMessage() );
586
587 return array();
588 }
589
590 return $compiledRules;
591 }
592
593 /**
594 * Get the plural rules for a given language from the XML files.
595 * Cached.
596 * @since 1.20
597 * @param string $code
598 * @return array|null
599 */
600 public function getPluralRules( $code ) {
601 if ( $this->pluralRules === null ) {
602 $this->loadPluralFiles();
603 }
604 if ( !isset( $this->pluralRules[$code] ) ) {
605 return null;
606 } else {
607 return $this->pluralRules[$code];
608 }
609 }
610
611 /**
612 * Get the plural rule types for a given language from the XML files.
613 * Cached.
614 * @since 1.22
615 * @param string $code
616 * @return array|null
617 */
618 public function getPluralRuleTypes( $code ) {
619 if ( $this->pluralRuleTypes === null ) {
620 $this->loadPluralFiles();
621 }
622 if ( !isset( $this->pluralRuleTypes[$code] ) ) {
623 return null;
624 } else {
625 return $this->pluralRuleTypes[$code];
626 }
627 }
628
629 /**
630 * Load the plural XML files.
631 */
632 protected function loadPluralFiles() {
633 global $IP;
634 $cldrPlural = "$IP/languages/data/plurals.xml";
635 $mwPlural = "$IP/languages/data/plurals-mediawiki.xml";
636 // Load CLDR plural rules
637 $this->loadPluralFile( $cldrPlural );
638 if ( file_exists( $mwPlural ) ) {
639 // Override or extend
640 $this->loadPluralFile( $mwPlural );
641 }
642 }
643
644 /**
645 * Load a plural XML file with the given filename, compile the relevant
646 * rules, and save the compiled rules in a process-local cache.
647 *
648 * @param string $fileName
649 * @throws MWException
650 */
651 protected function loadPluralFile( $fileName ) {
652 // Use file_get_contents instead of DOMDocument::load (T58439)
653 $xml = file_get_contents( $fileName );
654 if ( !$xml ) {
655 throw new MWException( "Unable to read plurals file $fileName" );
656 }
657 $doc = new DOMDocument;
658 $doc->loadXML( $xml );
659 $rulesets = $doc->getElementsByTagName( "pluralRules" );
660 foreach ( $rulesets as $ruleset ) {
661 $codes = $ruleset->getAttribute( 'locales' );
662 $rules = array();
663 $ruleTypes = array();
664 $ruleElements = $ruleset->getElementsByTagName( "pluralRule" );
665 foreach ( $ruleElements as $elt ) {
666 $ruleType = $elt->getAttribute( 'count' );
667 if ( $ruleType === 'other' ) {
668 // Don't record "other" rules, which have an empty condition
669 continue;
670 }
671 $rules[] = $elt->nodeValue;
672 $ruleTypes[] = $ruleType;
673 }
674 foreach ( explode( ' ', $codes ) as $code ) {
675 $this->pluralRules[$code] = $rules;
676 $this->pluralRuleTypes[$code] = $ruleTypes;
677 }
678 }
679 }
680
681 /**
682 * Read the data from the source files for a given language, and register
683 * the relevant dependencies in the $deps array. If the localisation
684 * exists, the data array is returned, otherwise false is returned.
685 *
686 * @param string $code
687 * @param array $deps
688 * @return array
689 */
690 protected function readSourceFilesAndRegisterDeps( $code, &$deps ) {
691 global $IP;
692
693 // This reads in the PHP i18n file with non-messages l10n data
694 $fileName = Language::getMessagesFileName( $code );
695 if ( !file_exists( $fileName ) ) {
696 $data = array();
697 } else {
698 $deps[] = new FileDependency( $fileName );
699 $data = $this->readPHPFile( $fileName, 'core' );
700 }
701
702 # Load CLDR plural rules for JavaScript
703 $data['pluralRules'] = $this->getPluralRules( $code );
704 # And for PHP
705 $data['compiledPluralRules'] = $this->getCompiledPluralRules( $code );
706 # Load plural rule types
707 $data['pluralRuleTypes'] = $this->getPluralRuleTypes( $code );
708
709 $deps['plurals'] = new FileDependency( "$IP/languages/data/plurals.xml" );
710 $deps['plurals-mw'] = new FileDependency( "$IP/languages/data/plurals-mediawiki.xml" );
711
712
713 return $data;
714 }
715
716 /**
717 * Merge two localisation values, a primary and a fallback, overwriting the
718 * primary value in place.
719 * @param string $key
720 * @param mixed $value
721 * @param mixed $fallbackValue
722 */
723 protected function mergeItem( $key, &$value, $fallbackValue ) {
724 if ( !is_null( $value ) ) {
725 if ( !is_null( $fallbackValue ) ) {
726 if ( in_array( $key, self::$mergeableMapKeys ) ) {
727 $value = $value + $fallbackValue;
728 } elseif ( in_array( $key, self::$mergeableListKeys ) ) {
729 $value = array_unique( array_merge( $fallbackValue, $value ) );
730 } elseif ( in_array( $key, self::$mergeableAliasListKeys ) ) {
731 $value = array_merge_recursive( $value, $fallbackValue );
732 } elseif ( in_array( $key, self::$optionalMergeKeys ) ) {
733 if ( !empty( $value['inherit'] ) ) {
734 $value = array_merge( $fallbackValue, $value );
735 }
736
737 if ( isset( $value['inherit'] ) ) {
738 unset( $value['inherit'] );
739 }
740 } elseif ( in_array( $key, self::$magicWordKeys ) ) {
741 $this->mergeMagicWords( $value, $fallbackValue );
742 }
743 }
744 } else {
745 $value = $fallbackValue;
746 }
747 }
748
749 /**
750 * @param mixed $value
751 * @param mixed $fallbackValue
752 */
753 protected function mergeMagicWords( &$value, $fallbackValue ) {
754 foreach ( $fallbackValue as $magicName => $fallbackInfo ) {
755 if ( !isset( $value[$magicName] ) ) {
756 $value[$magicName] = $fallbackInfo;
757 } else {
758 $oldSynonyms = array_slice( $fallbackInfo, 1 );
759 $newSynonyms = array_slice( $value[$magicName], 1 );
760 $synonyms = array_values( array_unique( array_merge(
761 $newSynonyms, $oldSynonyms ) ) );
762 $value[$magicName] = array_merge( array( $fallbackInfo[0] ), $synonyms );
763 }
764 }
765 }
766
767 /**
768 * Given an array mapping language code to localisation value, such as is
769 * found in extension *.i18n.php files, iterate through a fallback sequence
770 * to merge the given data with an existing primary value.
771 *
772 * Returns true if any data from the extension array was used, false
773 * otherwise.
774 * @param array $codeSequence
775 * @param string $key
776 * @param mixed $value
777 * @param mixed $fallbackValue
778 * @return bool
779 */
780 protected function mergeExtensionItem( $codeSequence, $key, &$value, $fallbackValue ) {
781 $used = false;
782 foreach ( $codeSequence as $code ) {
783 if ( isset( $fallbackValue[$code] ) ) {
784 $this->mergeItem( $key, $value, $fallbackValue[$code] );
785 $used = true;
786 }
787 }
788
789 return $used;
790 }
791
792 /**
793 * Gets the combined list of messages dirs from
794 * core and extensions
795 *
796 * @since 1.25
797 * @return array
798 */
799 public function getMessagesDirs() {
800 global $wgMessagesDirs, $IP;
801 return array(
802 'core' => "$IP/languages/i18n",
803 'api' => "$IP/includes/api/i18n",
804 'oojs-ui' => "$IP/resources/lib/oojs-ui/i18n",
805 ) + $wgMessagesDirs;
806 }
807
808 /**
809 * Load localisation data for a given language for both core and extensions
810 * and save it to the persistent cache store and the process cache
811 * @param string $code
812 * @throws MWException
813 */
814 public function recache( $code ) {
815 global $wgExtensionMessagesFiles;
816
817 if ( !$code ) {
818 throw new MWException( "Invalid language code requested" );
819 }
820 $this->recachedLangs[$code] = true;
821
822 # Initial values
823 $initialData = array_combine(
824 self::$allKeys,
825 array_fill( 0, count( self::$allKeys ), null ) );
826 $coreData = $initialData;
827 $deps = array();
828
829 # Load the primary localisation from the source file
830 $data = $this->readSourceFilesAndRegisterDeps( $code, $deps );
831 if ( $data === false ) {
832 wfDebug( __METHOD__ . ": no localisation file for $code, using fallback to en\n" );
833 $coreData['fallback'] = 'en';
834 } else {
835 wfDebug( __METHOD__ . ": got localisation for $code from source\n" );
836
837 # Merge primary localisation
838 foreach ( $data as $key => $value ) {
839 $this->mergeItem( $key, $coreData[$key], $value );
840 }
841 }
842
843 # Fill in the fallback if it's not there already
844 if ( is_null( $coreData['fallback'] ) ) {
845 $coreData['fallback'] = $code === 'en' ? false : 'en';
846 }
847 if ( $coreData['fallback'] === false ) {
848 $coreData['fallbackSequence'] = array();
849 } else {
850 $coreData['fallbackSequence'] = array_map( 'trim', explode( ',', $coreData['fallback'] ) );
851 $len = count( $coreData['fallbackSequence'] );
852
853 # Ensure that the sequence ends at en
854 if ( $coreData['fallbackSequence'][$len - 1] !== 'en' ) {
855 $coreData['fallbackSequence'][] = 'en';
856 }
857 }
858
859 $codeSequence = array_merge( array( $code ), $coreData['fallbackSequence'] );
860 $messageDirs = $this->getMessagesDirs();
861
862 wfProfileIn( __METHOD__ . '-fallbacks' );
863
864 # Load non-JSON localisation data for extensions
865 $extensionData = array_combine(
866 $codeSequence,
867 array_fill( 0, count( $codeSequence ), $initialData ) );
868 foreach ( $wgExtensionMessagesFiles as $extension => $fileName ) {
869 if ( isset( $messageDirs[$extension] ) ) {
870 # This extension has JSON message data; skip the PHP shim
871 continue;
872 }
873
874 $data = $this->readPHPFile( $fileName, 'extension' );
875 $used = false;
876
877 foreach ( $data as $key => $item ) {
878 foreach ( $codeSequence as $csCode ) {
879 if ( isset( $item[$csCode] ) ) {
880 $this->mergeItem( $key, $extensionData[$csCode][$key], $item[$csCode] );
881 $used = true;
882 }
883 }
884 }
885
886 if ( $used ) {
887 $deps[] = new FileDependency( $fileName );
888 }
889 }
890
891 # Load the localisation data for each fallback, then merge it into the full array
892 $allData = $initialData;
893 foreach ( $codeSequence as $csCode ) {
894 $csData = $initialData;
895
896 # Load core messages and the extension localisations.
897 foreach ( $messageDirs as $dirs ) {
898 foreach ( (array)$dirs as $dir ) {
899 $fileName = "$dir/$csCode.json";
900 $data = $this->readJSONFile( $fileName );
901
902 foreach ( $data as $key => $item ) {
903 $this->mergeItem( $key, $csData[$key], $item );
904 }
905
906 $deps[] = new FileDependency( $fileName );
907 }
908 }
909
910 # Merge non-JSON extension data
911 if ( isset( $extensionData[$csCode] ) ) {
912 foreach ( $extensionData[$csCode] as $key => $item ) {
913 $this->mergeItem( $key, $csData[$key], $item );
914 }
915 }
916
917 if ( $csCode === $code ) {
918 # Merge core data into extension data
919 foreach ( $coreData as $key => $item ) {
920 $this->mergeItem( $key, $csData[$key], $item );
921 }
922 } else {
923 # Load the secondary localisation from the source file to
924 # avoid infinite cycles on cyclic fallbacks
925 $fbData = $this->readSourceFilesAndRegisterDeps( $csCode, $deps );
926 if ( $fbData !== false ) {
927 # Only merge the keys that make sense to merge
928 foreach ( self::$allKeys as $key ) {
929 if ( !isset( $fbData[$key] ) ) {
930 continue;
931 }
932
933 if ( is_null( $coreData[$key] ) || $this->isMergeableKey( $key ) ) {
934 $this->mergeItem( $key, $csData[$key], $fbData[$key] );
935 }
936 }
937 }
938 }
939
940 # Allow extensions an opportunity to adjust the data for this
941 # fallback
942 Hooks::run( 'LocalisationCacheRecacheFallback', array( $this, $csCode, &$csData ) );
943
944 # Merge the data for this fallback into the final array
945 if ( $csCode === $code ) {
946 $allData = $csData;
947 } else {
948 foreach ( self::$allKeys as $key ) {
949 if ( !isset( $csData[$key] ) ) {
950 continue;
951 }
952
953 if ( is_null( $allData[$key] ) || $this->isMergeableKey( $key ) ) {
954 $this->mergeItem( $key, $allData[$key], $csData[$key] );
955 }
956 }
957 }
958 }
959
960 wfProfileOut( __METHOD__ . '-fallbacks' );
961
962 # Add cache dependencies for any referenced globals
963 $deps['wgExtensionMessagesFiles'] = new GlobalDependency( 'wgExtensionMessagesFiles' );
964 // $wgMessagesDirs is used in LocalisationCache::getMessagesDirs()
965 $deps['wgMessagesDirs'] = new GlobalDependency( 'wgMessagesDirs' );
966 $deps['version'] = new ConstantDependency( 'LocalisationCache::VERSION' );
967
968 # Add dependencies to the cache entry
969 $allData['deps'] = $deps;
970
971 # Replace spaces with underscores in namespace names
972 $allData['namespaceNames'] = str_replace( ' ', '_', $allData['namespaceNames'] );
973
974 # And do the same for special page aliases. $page is an array.
975 foreach ( $allData['specialPageAliases'] as &$page ) {
976 $page = str_replace( ' ', '_', $page );
977 }
978 # Decouple the reference to prevent accidental damage
979 unset( $page );
980
981 # If there were no plural rules, return an empty array
982 if ( $allData['pluralRules'] === null ) {
983 $allData['pluralRules'] = array();
984 }
985 if ( $allData['compiledPluralRules'] === null ) {
986 $allData['compiledPluralRules'] = array();
987 }
988 # If there were no plural rule types, return an empty array
989 if ( $allData['pluralRuleTypes'] === null ) {
990 $allData['pluralRuleTypes'] = array();
991 }
992
993 # Set the list keys
994 $allData['list'] = array();
995 foreach ( self::$splitKeys as $key ) {
996 $allData['list'][$key] = array_keys( $allData[$key] );
997 }
998 # Run hooks
999 $purgeBlobs = true;
1000 Hooks::run( 'LocalisationCacheRecache', array( $this, $code, &$allData, &$purgeBlobs ) );
1001
1002 if ( is_null( $allData['namespaceNames'] ) ) {
1003 throw new MWException( __METHOD__ . ': Localisation data failed sanity check! ' .
1004 'Check that your languages/messages/MessagesEn.php file is intact.' );
1005 }
1006
1007 # Set the preload key
1008 $allData['preload'] = $this->buildPreload( $allData );
1009
1010 # Save to the process cache and register the items loaded
1011 $this->data[$code] = $allData;
1012 foreach ( $allData as $key => $item ) {
1013 $this->loadedItems[$code][$key] = true;
1014 }
1015
1016 # Save to the persistent cache
1017 wfProfileIn( __METHOD__ . '-write' );
1018 $this->store->startWrite( $code );
1019 foreach ( $allData as $key => $value ) {
1020 if ( in_array( $key, self::$splitKeys ) ) {
1021 foreach ( $value as $subkey => $subvalue ) {
1022 $this->store->set( "$key:$subkey", $subvalue );
1023 }
1024 } else {
1025 $this->store->set( $key, $value );
1026 }
1027 }
1028 $this->store->finishWrite();
1029 wfProfileOut( __METHOD__ . '-write' );
1030
1031 # Clear out the MessageBlobStore
1032 # HACK: If using a null (i.e. disabled) storage backend, we
1033 # can't write to the MessageBlobStore either
1034 if ( $purgeBlobs && !$this->store instanceof LCStoreNull ) {
1035 MessageBlobStore::getInstance()->clear();
1036 }
1037
1038 }
1039
1040 /**
1041 * Build the preload item from the given pre-cache data.
1042 *
1043 * The preload item will be loaded automatically, improving performance
1044 * for the commonly-requested items it contains.
1045 * @param array $data
1046 * @return array
1047 */
1048 protected function buildPreload( $data ) {
1049 $preload = array( 'messages' => array() );
1050 foreach ( self::$preloadedKeys as $key ) {
1051 $preload[$key] = $data[$key];
1052 }
1053
1054 foreach ( $data['preloadedMessages'] as $subkey ) {
1055 if ( isset( $data['messages'][$subkey] ) ) {
1056 $subitem = $data['messages'][$subkey];
1057 } else {
1058 $subitem = null;
1059 }
1060 $preload['messages'][$subkey] = $subitem;
1061 }
1062
1063 return $preload;
1064 }
1065
1066 /**
1067 * Unload the data for a given language from the object cache.
1068 * Reduces memory usage.
1069 * @param string $code
1070 */
1071 public function unload( $code ) {
1072 unset( $this->data[$code] );
1073 unset( $this->loadedItems[$code] );
1074 unset( $this->loadedSubitems[$code] );
1075 unset( $this->initialisedLangs[$code] );
1076 unset( $this->shallowFallbacks[$code] );
1077
1078 foreach ( $this->shallowFallbacks as $shallowCode => $fbCode ) {
1079 if ( $fbCode === $code ) {
1080 $this->unload( $shallowCode );
1081 }
1082 }
1083 }
1084
1085 /**
1086 * Unload all data
1087 */
1088 public function unloadAll() {
1089 foreach ( $this->initialisedLangs as $lang => $unused ) {
1090 $this->unload( $lang );
1091 }
1092 }
1093
1094 /**
1095 * Disable the storage backend
1096 */
1097 public function disableBackend() {
1098 $this->store = new LCStoreNull;
1099 $this->manualRecache = false;
1100 }
1101 }
1102
1103 /**
1104 * Interface for the persistence layer of LocalisationCache.
1105 *
1106 * The persistence layer is two-level hierarchical cache. The first level
1107 * is the language, the second level is the item or subitem.
1108 *
1109 * Since the data for a whole language is rebuilt in one operation, it needs
1110 * to have a fast and atomic method for deleting or replacing all of the
1111 * current data for a given language. The interface reflects this bulk update
1112 * operation. Callers writing to the cache must first call startWrite(), then
1113 * will call set() a couple of thousand times, then will call finishWrite()
1114 * to commit the operation. When finishWrite() is called, the cache is
1115 * expected to delete all data previously stored for that language.
1116 *
1117 * The values stored are PHP variables suitable for serialize(). Implementations
1118 * of LCStore are responsible for serializing and unserializing.
1119 */
1120 interface LCStore {
1121 /**
1122 * Get a value.
1123 * @param string $code Language code
1124 * @param string $key Cache key
1125 */
1126 function get( $code, $key );
1127
1128 /**
1129 * Start a write transaction.
1130 * @param string $code Language code
1131 */
1132 function startWrite( $code );
1133
1134 /**
1135 * Finish a write transaction.
1136 */
1137 function finishWrite();
1138
1139 /**
1140 * Set a key to a given value. startWrite() must be called before this
1141 * is called, and finishWrite() must be called afterwards.
1142 * @param string $key
1143 * @param mixed $value
1144 */
1145 function set( $key, $value );
1146 }
1147
1148 /**
1149 * LCStore implementation which uses the standard DB functions to store data.
1150 * This will work on any MediaWiki installation.
1151 */
1152 class LCStoreDB implements LCStore {
1153 private $currentLang;
1154 private $writesDone = false;
1155
1156 /** @var DatabaseBase */
1157 private $dbw;
1158 /** @var array */
1159 private $batch = array();
1160
1161 private $readOnly = false;
1162
1163 public function get( $code, $key ) {
1164 if ( $this->writesDone ) {
1165 $db = wfGetDB( DB_MASTER );
1166 } else {
1167 $db = wfGetDB( DB_SLAVE );
1168 }
1169 $row = $db->selectRow( 'l10n_cache', array( 'lc_value' ),
1170 array( 'lc_lang' => $code, 'lc_key' => $key ), __METHOD__ );
1171 if ( $row ) {
1172 return unserialize( $db->decodeBlob( $row->lc_value ) );
1173 } else {
1174 return null;
1175 }
1176 }
1177
1178 public function startWrite( $code ) {
1179 if ( $this->readOnly ) {
1180 return;
1181 } elseif ( !$code ) {
1182 throw new MWException( __METHOD__ . ": Invalid language \"$code\"" );
1183 }
1184
1185 $this->dbw = wfGetDB( DB_MASTER );
1186
1187 $this->currentLang = $code;
1188 $this->batch = array();
1189 }
1190
1191 public function finishWrite() {
1192 if ( $this->readOnly ) {
1193 return;
1194 } elseif ( is_null( $this->currentLang ) ) {
1195 throw new MWException( __CLASS__ . ': must call startWrite() before finishWrite()' );
1196 }
1197
1198 $this->dbw->begin( __METHOD__ );
1199 try {
1200 $this->dbw->delete( 'l10n_cache',
1201 array( 'lc_lang' => $this->currentLang ), __METHOD__ );
1202 foreach ( array_chunk( $this->batch, 500 ) as $rows ) {
1203 $this->dbw->insert( 'l10n_cache', $rows, __METHOD__ );
1204 }
1205 $this->writesDone = true;
1206 } catch ( DBQueryError $e ) {
1207 if ( $this->dbw->wasReadOnlyError() ) {
1208 $this->readOnly = true; // just avoid site down time
1209 } else {
1210 throw $e;
1211 }
1212 }
1213 $this->dbw->commit( __METHOD__ );
1214
1215 $this->currentLang = null;
1216 $this->batch = array();
1217 }
1218
1219 public function set( $key, $value ) {
1220 if ( $this->readOnly ) {
1221 return;
1222 } elseif ( is_null( $this->currentLang ) ) {
1223 throw new MWException( __CLASS__ . ': must call startWrite() before set()' );
1224 }
1225
1226 $this->batch[] = array(
1227 'lc_lang' => $this->currentLang,
1228 'lc_key' => $key,
1229 'lc_value' => $this->dbw->encodeBlob( serialize( $value ) ) );
1230 }
1231 }
1232
1233 /**
1234 * LCStore implementation which stores data as a collection of CDB files in the
1235 * directory given by $wgCacheDirectory. If $wgCacheDirectory is not set, this
1236 * will throw an exception.
1237 *
1238 * Profiling indicates that on Linux, this implementation outperforms MySQL if
1239 * the directory is on a local filesystem and there is ample kernel cache
1240 * space. The performance advantage is greater when the DBA extension is
1241 * available than it is with the PHP port.
1242 *
1243 * See Cdb.php and http://cr.yp.to/cdb.html
1244 */
1245 class LCStoreCDB implements LCStore {
1246 /** @var CdbReader[] */
1247 private $readers;
1248
1249 /** @var CdbWriter */
1250 private $writer;
1251
1252 /** @var string Current language code */
1253 private $currentLang;
1254
1255 /** @var bool|string Cache directory. False if not set */
1256 private $directory;
1257
1258 function __construct( $conf = array() ) {
1259 global $wgCacheDirectory;
1260
1261 if ( isset( $conf['directory'] ) ) {
1262 $this->directory = $conf['directory'];
1263 } else {
1264 $this->directory = $wgCacheDirectory;
1265 }
1266 }
1267
1268 public function get( $code, $key ) {
1269 if ( !isset( $this->readers[$code] ) ) {
1270 $fileName = $this->getFileName( $code );
1271
1272 $this->readers[$code] = false;
1273 if ( file_exists( $fileName ) ) {
1274 try {
1275 $this->readers[$code] = CdbReader::open( $fileName );
1276 } catch ( CdbException $e ) {
1277 wfDebug( __METHOD__ . ": unable to open cdb file for reading\n" );
1278 }
1279 }
1280 }
1281
1282 if ( !$this->readers[$code] ) {
1283 return null;
1284 } else {
1285 $value = false;
1286 try {
1287 $value = $this->readers[$code]->get( $key );
1288 } catch ( CdbException $e ) {
1289 wfDebug( __METHOD__ . ": CdbException caught, error message was "
1290 . $e->getMessage() . "\n" );
1291 }
1292 if ( $value === false ) {
1293 return null;
1294 }
1295
1296 return unserialize( $value );
1297 }
1298 }
1299
1300 public function startWrite( $code ) {
1301 if ( !file_exists( $this->directory ) ) {
1302 if ( !wfMkdirParents( $this->directory, null, __METHOD__ ) ) {
1303 throw new MWException( "Unable to create the localisation store " .
1304 "directory \"{$this->directory}\"" );
1305 }
1306 }
1307
1308 // Close reader to stop permission errors on write
1309 if ( !empty( $this->readers[$code] ) ) {
1310 $this->readers[$code]->close();
1311 }
1312
1313 try {
1314 $this->writer = CdbWriter::open( $this->getFileName( $code ) );
1315 } catch ( CdbException $e ) {
1316 throw new MWException( $e->getMessage() );
1317 }
1318 $this->currentLang = $code;
1319 }
1320
1321 public function finishWrite() {
1322 // Close the writer
1323 try {
1324 $this->writer->close();
1325 } catch ( CdbException $e ) {
1326 throw new MWException( $e->getMessage() );
1327 }
1328 $this->writer = null;
1329 unset( $this->readers[$this->currentLang] );
1330 $this->currentLang = null;
1331 }
1332
1333 public function set( $key, $value ) {
1334 if ( is_null( $this->writer ) ) {
1335 throw new MWException( __CLASS__ . ': must call startWrite() before calling set()' );
1336 }
1337 try {
1338 $this->writer->set( $key, serialize( $value ) );
1339 } catch ( CdbException $e ) {
1340 throw new MWException( $e->getMessage() );
1341 }
1342 }
1343
1344 protected function getFileName( $code ) {
1345 if ( strval( $code ) === '' || strpos( $code, '/' ) !== false ) {
1346 throw new MWException( __METHOD__ . ": Invalid language \"$code\"" );
1347 }
1348
1349 return "{$this->directory}/l10n_cache-$code.cdb";
1350 }
1351 }
1352
1353 /**
1354 * Null store backend, used to avoid DB errors during install
1355 */
1356 class LCStoreNull implements LCStore {
1357 public function get( $code, $key ) {
1358 return null;
1359 }
1360
1361 public function startWrite( $code ) {
1362 }
1363
1364 public function finishWrite() {
1365 }
1366
1367 public function set( $key, $value ) {
1368 }
1369 }
1370
1371 /**
1372 * A localisation cache optimised for loading large amounts of data for many
1373 * languages. Used by rebuildLocalisationCache.php.
1374 */
1375 class LocalisationCacheBulkLoad extends LocalisationCache {
1376 /**
1377 * A cache of the contents of data files.
1378 * Core files are serialized to avoid using ~1GB of RAM during a recache.
1379 */
1380 private $fileCache = array();
1381
1382 /**
1383 * Most recently used languages. Uses the linked-list aspect of PHP hashtables
1384 * to keep the most recently used language codes at the end of the array, and
1385 * the language codes that are ready to be deleted at the beginning.
1386 */
1387 private $mruLangs = array();
1388
1389 /**
1390 * Maximum number of languages that may be loaded into $this->data
1391 */
1392 private $maxLoadedLangs = 10;
1393
1394 /**
1395 * @param string $fileName
1396 * @param string $fileType
1397 * @return array|mixed
1398 */
1399 protected function readPHPFile( $fileName, $fileType ) {
1400 $serialize = $fileType === 'core';
1401 if ( !isset( $this->fileCache[$fileName][$fileType] ) ) {
1402 $data = parent::readPHPFile( $fileName, $fileType );
1403
1404 if ( $serialize ) {
1405 $encData = serialize( $data );
1406 } else {
1407 $encData = $data;
1408 }
1409
1410 $this->fileCache[$fileName][$fileType] = $encData;
1411
1412 return $data;
1413 } elseif ( $serialize ) {
1414 return unserialize( $this->fileCache[$fileName][$fileType] );
1415 } else {
1416 return $this->fileCache[$fileName][$fileType];
1417 }
1418 }
1419
1420 /**
1421 * @param string $code
1422 * @param string $key
1423 * @return mixed
1424 */
1425 public function getItem( $code, $key ) {
1426 unset( $this->mruLangs[$code] );
1427 $this->mruLangs[$code] = true;
1428
1429 return parent::getItem( $code, $key );
1430 }
1431
1432 /**
1433 * @param string $code
1434 * @param string $key
1435 * @param string $subkey
1436 * @return mixed
1437 */
1438 public function getSubitem( $code, $key, $subkey ) {
1439 unset( $this->mruLangs[$code] );
1440 $this->mruLangs[$code] = true;
1441
1442 return parent::getSubitem( $code, $key, $subkey );
1443 }
1444
1445 /**
1446 * @param string $code
1447 */
1448 public function recache( $code ) {
1449 parent::recache( $code );
1450 unset( $this->mruLangs[$code] );
1451 $this->mruLangs[$code] = true;
1452 $this->trimCache();
1453 }
1454
1455 /**
1456 * @param string $code
1457 */
1458 public function unload( $code ) {
1459 unset( $this->mruLangs[$code] );
1460 parent::unload( $code );
1461 }
1462
1463 /**
1464 * Unload cached languages until there are less than $this->maxLoadedLangs
1465 */
1466 protected function trimCache() {
1467 while ( count( $this->data ) > $this->maxLoadedLangs && count( $this->mruLangs ) ) {
1468 reset( $this->mruLangs );
1469 $code = key( $this->mruLangs );
1470 wfDebug( __METHOD__ . ": unloading $code\n" );
1471 $this->unload( $code );
1472 }
1473 }
1474 }