3 * This program is free software; you can redistribute it and/or modify
4 * it under the terms of the GNU General Public License as published by
5 * the Free Software Foundation; either version 2 of the License, or
6 * (at your option) any later version.
8 * This program is distributed in the hope that it will be useful,
9 * but WITHOUT ANY WARRANTY; without even the implied warranty of
10 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
11 * GNU General Public License for more details.
13 * You should have received a copy of the GNU General Public License along
14 * with this program; if not, write to the Free Software Foundation, Inc.,
15 * 51 Franklin Street, Fifth Floor, Boston, MA 02110-1301, USA.
16 * http://www.gnu.org/copyleft/gpl.html
21 use MediaWiki\MediaWikiServices
;
24 * Base class for language conversion.
27 * @author Zhengzhu Feng <zhengzhu@gmail.com>
28 * @author fdcn <fdcn64@gmail.com>
29 * @author shinjiman <shinjiman@gmail.com>
30 * @author PhiLiP <philip.npc@gmail.com>
32 class LanguageConverter
{
34 * languages supporting variants
38 static public $languagesWithVariants = [
50 public $mMainLanguageCode;
52 public $mVariantFallbacks;
53 public $mVariantNames;
54 public $mTablesLoaded = false;
56 // 'bidirectional' 'unidirectional' 'disable' for each variant
60 * @var string Memcached key name
66 public $mDescCodeSep = ':', $mDescVarSep = ';';
67 public $mUcfirst = false;
68 public $mConvRuleTitle = false;
71 public $mHeaderVariant;
72 public $mMaxDepth = 10;
73 public $mVarSeparatorPattern;
75 const CACHE_VERSION_KEY
= 'VERSION 7';
80 * @param Language $langobj
81 * @param string $maincode The main language code of this language
82 * @param array $variants The supported variants of this language
83 * @param array $variantfallbacks The fallback language of each variant
84 * @param array $flags Defining the custom strings that maps to the flags
85 * @param array $manualLevel Limit for supported variants
87 public function __construct( $langobj, $maincode, $variants = [],
88 $variantfallbacks = [], $flags = [],
90 global $wgDisabledVariants;
91 $this->mLangObj
= $langobj;
92 $this->mMainLanguageCode
= $maincode;
93 $this->mVariants
= array_diff( $variants, $wgDisabledVariants );
94 $this->mVariantFallbacks
= $variantfallbacks;
95 $this->mVariantNames
= Language
::fetchLanguageNames();
96 $this->mCacheKey
= wfMemcKey( 'conversiontables', $maincode );
98 // 'S' show converted text
99 // '+' add rules for alltext
100 // 'E' the gave flags is error
101 // these flags above are reserved for program
102 'A' => 'A', // add rule for convert code (all text convert)
103 'T' => 'T', // title convert
104 'R' => 'R', // raw content
105 'D' => 'D', // convert description (subclass implement)
106 '-' => '-', // remove convert (not implement)
107 'H' => 'H', // add rule for convert code (but no display in placed code)
108 'N' => 'N' // current variant name
110 $this->mFlags
= array_merge( $defaultflags, $flags );
111 foreach ( $this->mVariants
as $v ) {
112 if ( array_key_exists( $v, $manualLevel ) ) {
113 $this->mManualLevel
[$v] = $manualLevel[$v];
115 $this->mManualLevel
[$v] = 'bidirectional';
117 $this->mFlags
[$v] = $v;
122 * Get all valid variants.
123 * Call this instead of using $this->mVariants directly.
125 * @return array Contains all valid variants
127 public function getVariants() {
128 return $this->mVariants
;
132 * In case some variant is not defined in the markup, we need
133 * to have some fallback. For example, in zh, normally people
134 * will define zh-hans and zh-hant, but less so for zh-sg or zh-hk.
135 * when zh-sg is preferred but not defined, we will pick zh-hans
136 * in this case. Right now this is only used by zh.
138 * @param string $variant The language code of the variant
139 * @return string|array The code of the fallback language or the
140 * main code if there is no fallback
142 public function getVariantFallbacks( $variant ) {
143 if ( isset( $this->mVariantFallbacks
[$variant] ) ) {
144 return $this->mVariantFallbacks
[$variant];
146 return $this->mMainLanguageCode
;
150 * Get the title produced by the conversion rule.
151 * @return string The converted title text
153 public function getConvRuleTitle() {
154 return $this->mConvRuleTitle
;
158 * Get preferred language variant.
159 * @return string The preferred language code
161 public function getPreferredVariant() {
162 global $wgDefaultLanguageVariant, $wgUser;
164 $req = $this->getURLVariant();
166 if ( $wgUser->isSafeToLoad() && $wgUser->isLoggedIn() && !$req ) {
167 $req = $this->getUserVariant();
169 $req = $this->getHeaderVariant();
172 if ( $wgDefaultLanguageVariant && !$req ) {
173 $req = $this->validateVariant( $wgDefaultLanguageVariant );
176 // This function, unlike the other get*Variant functions, is
177 // not memoized (i.e. there return value is not cached) since
178 // new information might appear during processing after this
180 if ( $this->validateVariant( $req ) ) {
183 return $this->mMainLanguageCode
;
187 * Get default variant.
188 * This function would not be affected by user's settings
189 * @return string The default variant code
191 public function getDefaultVariant() {
192 global $wgDefaultLanguageVariant;
194 $req = $this->getURLVariant();
197 $req = $this->getHeaderVariant();
200 if ( $wgDefaultLanguageVariant && !$req ) {
201 $req = $this->validateVariant( $wgDefaultLanguageVariant );
207 return $this->mMainLanguageCode
;
211 * Validate the variant
212 * @param string $variant The variant to validate
213 * @return mixed Returns the variant if it is valid, null otherwise
215 public function validateVariant( $variant = null ) {
216 if ( $variant !== null && in_array( $variant, $this->mVariants
) ) {
223 * Get the variant specified in the URL
225 * @return mixed Variant if one found, false otherwise.
227 public function getURLVariant() {
230 if ( $this->mURLVariant
) {
231 return $this->mURLVariant
;
234 // see if the preference is set in the request
235 $ret = $wgRequest->getText( 'variant' );
238 $ret = $wgRequest->getVal( 'uselang' );
241 $this->mURLVariant
= $this->validateVariant( $ret );
242 return $this->mURLVariant
;
246 * Determine if the user has a variant set.
248 * @return mixed Variant if one found, false otherwise.
250 protected function getUserVariant() {
251 global $wgUser, $wgContLang;
253 // memoizing this function wreaks havoc on parserTest.php
255 if ( $this->mUserVariant ) {
256 return $this->mUserVariant;
260 // Get language variant preference from logged in users
261 // Don't call this on stub objects because that causes infinite
262 // recursion during initialisation
263 if ( !$wgUser->isSafeToLoad() ) {
266 if ( $wgUser->isLoggedIn() ) {
267 if ( $this->mMainLanguageCode
== $wgContLang->getCode() ) {
268 $ret = $wgUser->getOption( 'variant' );
270 $ret = $wgUser->getOption( 'variant-' . $this->mMainLanguageCode
);
273 // figure out user lang without constructing wgLang to avoid
274 // infinite recursion
275 $ret = $wgUser->getOption( 'language' );
278 $this->mUserVariant
= $this->validateVariant( $ret );
279 return $this->mUserVariant
;
283 * Determine the language variant from the Accept-Language header.
285 * @return mixed Variant if one found, false otherwise.
287 protected function getHeaderVariant() {
290 if ( $this->mHeaderVariant
) {
291 return $this->mHeaderVariant
;
294 // see if some supported language variant is set in the
296 $languages = array_keys( $wgRequest->getAcceptLang() );
297 if ( empty( $languages ) ) {
301 $fallbackLanguages = [];
302 foreach ( $languages as $language ) {
303 $this->mHeaderVariant
= $this->validateVariant( $language );
304 if ( $this->mHeaderVariant
) {
308 // To see if there are fallbacks of current language.
309 // We record these fallback variants, and process
311 $fallbacks = $this->getVariantFallbacks( $language );
312 if ( is_string( $fallbacks ) && $fallbacks !== $this->mMainLanguageCode
) {
313 $fallbackLanguages[] = $fallbacks;
314 } elseif ( is_array( $fallbacks ) ) {
316 array_merge( $fallbackLanguages, $fallbacks );
320 if ( !$this->mHeaderVariant
) {
321 // process fallback languages now
322 $fallback_languages = array_unique( $fallbackLanguages );
323 foreach ( $fallback_languages as $language ) {
324 $this->mHeaderVariant
= $this->validateVariant( $language );
325 if ( $this->mHeaderVariant
) {
331 return $this->mHeaderVariant
;
335 * Dictionary-based conversion.
336 * This function would not parse the conversion rules.
337 * If you want to parse rules, try to use convert() or
340 * @param string $text The text to be converted
341 * @param bool|string $toVariant The target language code
342 * @return string The converted text
344 public function autoConvert( $text, $toVariant = false ) {
349 $toVariant = $this->getPreferredVariant();
355 if ( $this->guessVariant( $text, $toVariant ) ) {
359 /* we convert everything except:
360 1. HTML markups (anything between < and >)
362 3. placeholders created by the parser
364 $marker = '|' . Parser
::MARKER_PREFIX
. '[\-a-zA-Z0-9]+';
366 // this one is needed when the text is inside an HTML markup
367 $htmlfix = '|<[^>]+$|^[^<>]*>';
369 // disable convert to variants between <code> tags
370 $codefix = '<code>.+?<\/code>|';
371 // disable conversion of <script> tags
372 $scriptfix = '<script.*?>.*?<\/script>|';
373 // disable conversion of <pre> tags
374 $prefix = '<pre.*?>.*?<\/pre>|';
376 $reg = '/' . $codefix . $scriptfix . $prefix .
377 '<[^>]+>|&[a-zA-Z#][a-z0-9]+;' . $marker . $htmlfix . '/s';
382 // Guard against delimiter nulls in the input
383 $text = str_replace( "\000", '', $text );
385 $markupMatches = null;
386 $elementMatches = null;
387 while ( $startPos < strlen( $text ) ) {
388 if ( preg_match( $reg, $text, $markupMatches, PREG_OFFSET_CAPTURE
, $startPos ) ) {
389 $elementPos = $markupMatches[0][1];
390 $element = $markupMatches[0][0];
392 $elementPos = strlen( $text );
396 // Queue the part before the markup for translation in a batch
397 $sourceBlob .= substr( $text, $startPos, $elementPos - $startPos ) . "\000";
399 // Advance to the next position
400 $startPos = $elementPos +
strlen( $element );
402 // Translate any alt or title attributes inside the matched element
404 && preg_match( '/^(<[^>\s]*)\s([^>]*)(.*)$/', $element, $elementMatches )
406 $attrs = Sanitizer
::decodeTagAttributes( $elementMatches[2] );
408 foreach ( [ 'title', 'alt' ] as $attrName ) {
409 if ( !isset( $attrs[$attrName] ) ) {
412 $attr = $attrs[$attrName];
413 // Don't convert URLs
414 if ( !strpos( $attr, '://' ) ) {
415 $attr = $this->recursiveConvertTopLevel( $attr, $toVariant );
418 if ( $attr !== $attrs[$attrName] ) {
419 $attrs[$attrName] = $attr;
424 $element = $elementMatches[1] . Html
::expandAttributes( $attrs ) .
428 $literalBlob .= $element . "\000";
431 // Do the main translation batch
432 $translatedBlob = $this->translate( $sourceBlob, $toVariant );
434 // Put the output back together
435 $translatedIter = StringUtils
::explode( "\000", $translatedBlob );
436 $literalIter = StringUtils
::explode( "\000", $literalBlob );
438 while ( $translatedIter->valid() && $literalIter->valid() ) {
439 $output .= $translatedIter->current();
440 $output .= $literalIter->current();
441 $translatedIter->next();
442 $literalIter->next();
449 * Translate a string to a variant.
450 * Doesn't parse rules or do any of that other stuff, for that use
451 * convert() or convertTo().
453 * @param string $text Text to convert
454 * @param string $variant Variant language code
455 * @return string Translated text
457 public function translate( $text, $variant ) {
458 // If $text is empty or only includes spaces, do nothing
459 // Otherwise translate it
460 if ( trim( $text ) ) {
462 $text = $this->mTables
[$variant]->replace( $text );
468 * Call translate() to convert text to all valid variants.
470 * @param string $text The text to be converted
471 * @return array Variant => converted text
473 public function autoConvertToAllVariants( $text ) {
477 foreach ( $this->mVariants
as $variant ) {
478 $ret[$variant] = $this->translate( $text, $variant );
485 * Apply manual conversion rules.
487 * @param ConverterRule $convRule
489 protected function applyManualConv( $convRule ) {
490 // Use syntax -{T|zh-cn:TitleCN; zh-tw:TitleTw}- to custom
492 // Bug 24072: $mConvRuleTitle was overwritten by other manual
493 // rule(s) not for title, this breaks the title conversion.
494 $newConvRuleTitle = $convRule->getTitle();
495 if ( $newConvRuleTitle ) {
496 // So I add an empty check for getTitle()
497 $this->mConvRuleTitle
= $newConvRuleTitle;
500 // merge/remove manual conversion rules to/from global table
501 $convTable = $convRule->getConvTable();
502 $action = $convRule->getRulesAction();
503 foreach ( $convTable as $variant => $pair ) {
504 if ( !$this->validateVariant( $variant ) ) {
508 if ( $action == 'add' ) {
509 // More efficient than array_merge(), about 2.5 times.
510 foreach ( $pair as $from => $to ) {
511 $this->mTables
[$variant]->setPair( $from, $to );
513 } elseif ( $action == 'remove' ) {
514 $this->mTables
[$variant]->removeArray( $pair );
520 * Auto convert a Title object to a readable string in the
523 * @param Title $title A object of Title
524 * @return string Converted title text
526 public function convertTitle( $title ) {
527 $variant = $this->getPreferredVariant();
528 $index = $title->getNamespace();
529 if ( $index !== NS_MAIN
) {
530 $text = $this->convertNamespace( $index, $variant ) . ':';
534 $text .= $this->translate( $title->getText(), $variant );
539 * Get the namespace display name in the preferred variant.
541 * @param int $index Namespace id
542 * @param string|null $variant Variant code or null for preferred variant
543 * @return string Namespace name for display
545 public function convertNamespace( $index, $variant = null ) {
546 if ( $index === NS_MAIN
) {
550 if ( $variant === null ) {
551 $variant = $this->getPreferredVariant();
554 $cache = MediaWikiServices
::getInstance()->getLocalServerObjectCache();
555 $key = $cache->makeKey( 'languageconverter', 'namespace-text', $index, $variant );
556 $nsVariantText = $cache->get( $key );
557 if ( $nsVariantText !== false ) {
558 return $nsVariantText;
561 // First check if a message gives a converted name in the target variant.
562 $nsConvMsg = wfMessage( 'conversion-ns' . $index )->inLanguage( $variant );
563 if ( $nsConvMsg->exists() ) {
564 $nsVariantText = $nsConvMsg->plain();
567 // Then check if a message gives a converted name in content language
568 // which needs extra translation to the target variant.
569 if ( $nsVariantText === false ) {
570 $nsConvMsg = wfMessage( 'conversion-ns' . $index )->inContentLanguage();
571 if ( $nsConvMsg->exists() ) {
572 $nsVariantText = $this->translate( $nsConvMsg->plain(), $variant );
576 if ( $nsVariantText === false ) {
577 // No message exists, retrieve it from the target variant's namespace names.
578 $langObj = $this->mLangObj
->factory( $variant );
579 $nsVariantText = $langObj->getFormattedNsText( $index );
582 $cache->set( $key, $nsVariantText, 60 );
584 return $nsVariantText;
588 * Convert text to different variants of a language. The automatic
589 * conversion is done in autoConvert(). Here we parse the text
590 * marked with -{}-, which specifies special conversions of the
591 * text that can not be accomplished in autoConvert().
593 * Syntax of the markup:
594 * -{code1:text1;code2:text2;...}- or
595 * -{flags|code1:text1;code2:text2;...}- or
596 * -{text}- in which case no conversion should take place for text
598 * @param string $text Text to be converted
599 * @return string Converted text
601 public function convert( $text ) {
602 $variant = $this->getPreferredVariant();
603 return $this->convertTo( $text, $variant );
607 * Same as convert() except a extra parameter to custom variant.
609 * @param string $text Text to be converted
610 * @param string $variant The target variant code
611 * @return string Converted text
613 public function convertTo( $text, $variant ) {
614 global $wgDisableLangConversion;
615 if ( $wgDisableLangConversion ) {
618 // Reset converter state for a new converter run.
619 $this->mConvRuleTitle
= false;
620 return $this->recursiveConvertTopLevel( $text, $variant );
624 * Recursively convert text on the outside. Allow to use nested
625 * markups to custom rules.
627 * @param string $text Text to be converted
628 * @param string $variant The target variant code
629 * @param int $depth Depth of recursion
630 * @return string Converted text
632 protected function recursiveConvertTopLevel( $text, $variant, $depth = 0 ) {
635 $length = strlen( $text );
636 $shouldConvert = !$this->guessVariant( $text, $variant );
638 while ( $startPos < $length ) {
639 $pos = strpos( $text, '-{', $startPos );
641 if ( $pos === false ) {
642 // No more markup, append final segment
643 $fragment = substr( $text, $startPos );
644 $out .= $shouldConvert ?
$this->autoConvert( $fragment, $variant ) : $fragment;
649 // Append initial segment
650 $fragment = substr( $text, $startPos, $pos - $startPos );
651 $out .= $shouldConvert ?
$this->autoConvert( $fragment, $variant ) : $fragment;
656 // Do recursive conversion
657 $out .= $this->recursiveConvertRule( $text, $variant, $startPos, $depth +
1 );
664 * Recursively convert text on the inside.
666 * @param string $text Text to be converted
667 * @param string $variant The target variant code
668 * @param int $startPos
669 * @param int $depth Depth of recursion
671 * @throws MWException
672 * @return string Converted text
674 protected function recursiveConvertRule( $text, $variant, &$startPos, $depth = 0 ) {
675 // Quick sanity check (no function calls)
676 if ( $text[$startPos] !== '-' ||
$text[$startPos +
1] !== '{' ) {
677 throw new MWException( __METHOD__
. ': invalid input string' );
682 $warningDone = false;
683 $length = strlen( $text );
685 while ( $startPos < $length ) {
687 preg_match( '/-\{|\}-/', $text, $m, PREG_OFFSET_CAPTURE
, $startPos );
697 // Append initial segment
698 $inner .= substr( $text, $startPos, $pos - $startPos );
706 if ( $depth >= $this->mMaxDepth
) {
708 if ( !$warningDone ) {
709 $inner .= '<span class="error">' .
710 wfMessage( 'language-converter-depth-warning' )
711 ->numParams( $this->mMaxDepth
)->inContentLanguage()->text() .
718 // Recursively parse another rule
719 $inner .= $this->recursiveConvertRule( $text, $variant, $startPos, $depth +
1 );
724 $rule = new ConverterRule( $inner, $this );
725 $rule->parse( $variant );
726 $this->applyManualConv( $rule );
727 return $rule->getDisplay();
729 throw new MWException( __METHOD__
. ': invalid regex match' );
734 if ( $startPos < $length ) {
735 $inner .= substr( $text, $startPos );
738 return '-{' . $this->autoConvert( $inner, $variant );
742 * If a language supports multiple variants, it is possible that
743 * non-existing link in one variant actually exists in another variant.
744 * This function tries to find it. See e.g. LanguageZh.php
745 * The input parameters may be modified upon return
747 * @param string &$link The name of the link
748 * @param Title &$nt The title object of the link
749 * @param bool $ignoreOtherCond To disable other conditions when
750 * we need to transclude a template or update a category's link
752 public function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) {
753 # If the article has already existed, there is no need to
754 # check it again, otherwise it may cause a fault.
755 if ( is_object( $nt ) && $nt->exists() ) {
759 global $wgDisableLangConversion, $wgDisableTitleConversion, $wgRequest;
760 $isredir = $wgRequest->getText( 'redirect', 'yes' );
761 $action = $wgRequest->getText( 'action' );
762 if ( $action == 'edit' && $wgRequest->getBool( 'redlink' ) ) {
765 $linkconvert = $wgRequest->getText( 'linkconvert', 'yes' );
766 $disableLinkConversion = $wgDisableLangConversion
767 ||
$wgDisableTitleConversion;
768 $linkBatch = new LinkBatch();
772 if ( $disableLinkConversion ||
773 ( !$ignoreOtherCond &&
776 ||
$action == 'submit'
777 ||
$linkconvert == 'no' ) ) ) {
781 if ( is_object( $nt ) ) {
782 $ns = $nt->getNamespace();
785 $variants = $this->autoConvertToAllVariants( $link );
786 if ( !$variants ) { // give up
792 foreach ( $variants as $v ) {
794 $varnt = Title
::newFromText( $v, $ns );
795 if ( !is_null( $varnt ) ) {
796 $linkBatch->addObj( $varnt );
802 // fetch all variants in single query
803 $linkBatch->execute();
805 foreach ( $titles as $varnt ) {
806 if ( $varnt->getArticleID() > 0 ) {
808 $link = $varnt->getText();
815 * Returns language specific hash options.
819 public function getExtraHashOptions() {
820 $variant = $this->getPreferredVariant();
822 return '!' . $variant;
826 * Guess if a text is written in a variant. This should be implemented in subclasses.
828 * @param string $text The text to be checked
829 * @param string $variant Language code of the variant to be checked for
830 * @return bool True if $text appears to be written in $variant, false if not
832 * @author Nikola Smolenski <smolensk@eunet.rs>
835 public function guessVariant( $text, $variant ) {
840 * Load default conversion tables.
841 * This method must be implemented in derived class.
844 * @throws MWException
846 function loadDefaultTables() {
847 $name = get_class( $this );
849 throw new MWException( "Must implement loadDefaultTables() method in class $name" );
853 * Load conversion tables either from the cache or the disk.
855 * @param bool $fromCache Load from memcached? Defaults to true.
857 function loadTables( $fromCache = true ) {
858 global $wgLanguageConverterCacheType;
860 if ( $this->mTablesLoaded
) {
864 $this->mTablesLoaded
= true;
865 $this->mTables
= false;
866 $cache = ObjectCache
::getInstance( $wgLanguageConverterCacheType );
868 wfProfileIn( __METHOD__
. '-cache' );
869 $this->mTables
= $cache->get( $this->mCacheKey
);
870 wfProfileOut( __METHOD__
. '-cache' );
872 if ( !$this->mTables ||
!array_key_exists( self
::CACHE_VERSION_KEY
, $this->mTables
) ) {
873 wfProfileIn( __METHOD__
. '-recache' );
874 // not in cache, or we need a fresh reload.
875 // We will first load the default tables
876 // then update them using things in MediaWiki:Conversiontable/*
877 $this->loadDefaultTables();
878 foreach ( $this->mVariants
as $var ) {
879 $cached = $this->parseCachedTable( $var );
880 $this->mTables
[$var]->mergeArray( $cached );
883 $this->postLoadTables();
884 $this->mTables
[self
::CACHE_VERSION_KEY
] = true;
886 $cache->set( $this->mCacheKey
, $this->mTables
, 43200 );
887 wfProfileOut( __METHOD__
. '-recache' );
892 * Hook for post processing after conversion tables are loaded.
894 function postLoadTables() {
898 * Reload the conversion tables.
902 function reloadTables() {
903 if ( $this->mTables
) {
904 unset( $this->mTables
);
907 $this->mTablesLoaded
= false;
908 $this->loadTables( false );
912 * Parse the conversion table stored in the cache.
914 * The tables should be in blocks of the following form:
921 * To make the tables more manageable, subpages are allowed
922 * and will be parsed recursively if $recursive == true.
924 * @param string $code Language code
925 * @param string $subpage Subpage name
926 * @param bool $recursive Parse subpages recursively? Defaults to true.
930 function parseCachedTable( $code, $subpage = '', $recursive = true ) {
933 $key = 'Conversiontable/' . $code;
935 $key .= '/' . $subpage;
937 if ( array_key_exists( $key, $parsed ) ) {
941 $parsed[$key] = true;
943 if ( $subpage === '' ) {
944 $txt = MessageCache
::singleton()->getMsgFromNamespace( $key, $code );
947 $title = Title
::makeTitleSafe( NS_MEDIAWIKI
, $key );
948 if ( $title && $title->exists() ) {
949 $revision = Revision
::newFromTitle( $title );
951 if ( $revision->getContentModel() == CONTENT_MODEL_WIKITEXT
) {
952 $txt = $revision->getContent( Revision
::RAW
)->getNativeData();
955 // @todo in the future, use a specialized content model, perhaps based on json!
960 # Nothing to parse if there's no text
961 if ( $txt === false ||
$txt === null ||
$txt === '' ) {
965 // get all subpage links of the form
966 // [[MediaWiki:Conversiontable/zh-xx/...|...]]
967 $linkhead = $this->mLangObj
->getNsText( NS_MEDIAWIKI
) .
969 $subs = StringUtils
::explode( '[[', $txt );
971 foreach ( $subs as $sub ) {
972 $link = explode( ']]', $sub, 2 );
973 if ( count( $link ) != 2 ) {
976 $b = explode( '|', $link[0], 2 );
977 $b = explode( '/', trim( $b[0] ), 3 );
978 if ( count( $b ) == 3 ) {
984 if ( $b[0] == $linkhead && $b[1] == $code ) {
985 $sublinks[] = $sublink;
989 // parse the mappings in this page
990 $blocks = StringUtils
::explode( '-{', $txt );
993 foreach ( $blocks as $block ) {
995 // Skip the part before the first -{
999 $mappings = explode( '}-', $block, 2 )[0];
1000 $stripped = str_replace( [ "'", '"', '*', '#' ], '', $mappings );
1001 $table = StringUtils
::explode( ';', $stripped );
1002 foreach ( $table as $t ) {
1003 $m = explode( '=>', $t, 3 );
1004 if ( count( $m ) != 2 ) {
1007 // trim any trailling comments starting with '//'
1008 $tt = explode( '//', $m[1], 2 );
1009 $ret[trim( $m[0] )] = trim( $tt[0] );
1013 // recursively parse the subpages
1015 foreach ( $sublinks as $link ) {
1016 $s = $this->parseCachedTable( $code, $link, $recursive );
1021 if ( $this->mUcfirst
) {
1022 foreach ( $ret as $k => $v ) {
1023 $ret[$this->mLangObj
->ucfirst( $k )] = $this->mLangObj
->ucfirst( $v );
1030 * Enclose a string with the "no conversion" tag. This is used by
1031 * various functions in the Parser.
1033 * @param string $text Text to be tagged for no conversion
1034 * @param bool $noParse Unused
1035 * @return string The tagged text
1037 public function markNoConversion( $text, $noParse = false ) {
1038 # don't mark if already marked
1039 if ( strpos( $text, '-{' ) ||
strpos( $text, '}-' ) ) {
1043 $ret = "-{R|$text}-";
1048 * Convert the sorting key for category links. This should make different
1049 * keys that are variants of each other map to the same key.
1051 * @param string $key
1055 function convertCategoryKey( $key ) {
1060 * Refresh the cache of conversion tables when
1061 * MediaWiki:Conversiontable* is updated.
1063 * @param Title $titleobj The Title of the page being updated
1065 public function updateConversionTable( Title
$titleobj ) {
1066 if ( $titleobj->getNamespace() == NS_MEDIAWIKI
) {
1067 $title = $titleobj->getDBkey();
1068 $t = explode( '/', $title, 3 );
1070 if ( $c > 1 && $t[0] == 'Conversiontable' ) {
1071 if ( $this->validateVariant( $t[1] ) ) {
1072 $this->reloadTables();
1079 * Get the cached separator pattern for ConverterRule::parseRules()
1082 function getVarSeparatorPattern() {
1083 if ( is_null( $this->mVarSeparatorPattern
) ) {
1084 // varsep_pattern for preg_split:
1085 // text should be splited by ";" only if a valid variant
1086 // name exist after the markup, for example:
1087 // -{zh-hans:<span style="font-size:120%;">xxx</span>;zh-hant:\
1088 // <span style="font-size:120%;">yyy</span>;}-
1089 // we should split it as:
1091 // [0] => 'zh-hans:<span style="font-size:120%;">xxx</span>'
1092 // [1] => 'zh-hant:<span style="font-size:120%;">yyy</span>'
1096 foreach ( $this->mVariants
as $variant ) {
1097 // zh-hans:xxx;zh-hant:yyy
1098 $pat .= $variant . '\s*:|';
1099 // xxx=>zh-hans:yyy; xxx=>zh-hant:zzz
1100 $pat .= '[^;]*?=>\s*' . $variant . '\s*:|';
1103 $this->mVarSeparatorPattern
= $pat;
1105 return $this->mVarSeparatorPattern
;