X-Git-Url: https://git.heureux-cyclage.org/?a=blobdiff_plain;f=languages%2FLanguageConverter.php;h=5e4bce8d158509936472ed921e527d3c3d9faf2f;hb=5ccc792805398296ef91184cf729c231685132bb;hp=6e650c8aa82f604dbb26e793b8c53162381c532b;hpb=3f142983b335a8e3834426d42bacd35d7f61a6dc;p=lhc%2Fweb%2Fwiklou.git diff --git a/languages/LanguageConverter.php b/languages/LanguageConverter.php index 6e650c8aa8..5e4bce8d15 100644 --- a/languages/LanguageConverter.php +++ b/languages/LanguageConverter.php @@ -1,268 +1,345 @@ * @maintainers fdcn , shinjiman , PhiLiP */ class LanguageConverter { - var $mPreferredVariant=''; var $mMainLanguageCode; var $mVariants, $mVariantFallbacks, $mVariantNames; var $mTablesLoaded = false; var $mTables; - var $mNamespaceTables; - var $mTitleDisplay=''; - var $mDoTitleConvert=true, $mDoContentConvert=true; - var $mManualLevel; // 'bidirectional' 'unidirectional' 'disable' for each variants - var $mTitleFromFlag = false; + // 'bidirectional' 'unidirectional' 'disable' for each variant + var $mManualLevel; + + /** + * @var String: memcached key name + */ var $mCacheKey; + var $mLangObj; - var $mMarkup; var $mFlags; - var $mDescCodeSep = ':',$mDescVarSep = ';'; + var $mDescCodeSep = ':', $mDescVarSep = ';'; var $mUcfirst = false; + var $mConvRuleTitle = false; + var $mURLVariant; + var $mUserVariant; + var $mHeaderVariant; + var $mMaxDepth = 10; + var $mVarSeparatorPattern; const CACHE_VERSION_KEY = 'VERSION 6'; /** * Constructor * - * @param string $maincode the main language code of this language - * @param array $variants the supported variants of this language - * @param array $variantfallback the fallback language of each variant - * @param array $markup array defining the markup used for manual conversion - * @param array $flags array defining the custom strings that maps to the flags - * @param array $manualLevel limit for supported variants - * @public + * @param $langobj Language: the Language Object + * @param $maincode String: the main language code of this language + * @param $variants Array: the supported variants of this language + * @param $variantfallbacks Array: the fallback language of each variant + * @param $flags Array: defining the custom strings that maps to the flags + * @param $manualLevel Array: limit for supported variants */ - function __construct($langobj, $maincode, - $variants=array(), - $variantfallbacks=array(), - $markup=array(), - $flags = array(), + public function __construct( $langobj, $maincode, $variants = array(), + $variantfallbacks = array(), $flags = array(), $manualLevel = array() ) { + global $wgDisabledVariants; $this->mLangObj = $langobj; $this->mMainLanguageCode = $maincode; - $this->mVariants = $variants; + $this->mVariants = array_diff( $variants, $wgDisabledVariants ); $this->mVariantFallbacks = $variantfallbacks; - global $wgLanguageNames; - $this->mVariantNames = $wgLanguageNames; + $this->mVariantNames = Language::getLanguageNames(); $this->mCacheKey = wfMemcKey( 'conversiontables', $maincode ); - $m = array( - 'begin'=>'-{', - 'flagsep'=>'|', - 'unidsep'=>'=>', //for unidirectional conversion - 'codesep'=>':', - 'varsep'=>';', - 'end'=>'}-' - ); - $this->mMarkup = array_merge($m, $markup); - $f = array( + $defaultflags = array( // 'S' show converted text // '+' add rules for alltext // 'E' the gave flags is error // these flags above are reserved for program - 'A'=>'A', // add rule for convert code (all text convert) - 'T'=>'T', // title convert - 'R'=>'R', // raw content - 'D'=>'D', // convert description (subclass implement) - '-'=>'-', // remove convert (not implement) - 'H'=>'H', // add rule for convert code (but no display in placed code ) - 'N'=>'N' // current variant name + 'A' => 'A', // add rule for convert code (all text convert) + 'T' => 'T', // title convert + 'R' => 'R', // raw content + 'D' => 'D', // convert description (subclass implement) + '-' => '-', // remove convert (not implement) + 'H' => 'H', // add rule for convert code + // (but no display in placed code) + 'N' => 'N' // current variant name ); - $this->mFlags = array_merge($f, $flags); - foreach( $this->mVariants as $v) { - $this->mManualLevel[$v]=array_key_exists($v,$manualLevel) - ?$manualLevel[$v] - :'bidirectional'; - $this->mNamespaceTables[$v] = array(); + $this->mFlags = array_merge( $defaultflags, $flags ); + foreach ( $this->mVariants as $v ) { + if ( array_key_exists( $v, $manualLevel ) ) { + $this->mManualLevel[$v] = $manualLevel[$v]; + } else { + $this->mManualLevel[$v] = 'bidirectional'; + } $this->mFlags[$v] = $v; } } /** - * @public + * Get all valid variants. + * Call this instead of using $this->mVariants directly. + * + * @return Array: contains all valid variants */ - function getVariants() { + public function getVariants() { return $this->mVariants; } /** - * in case some variant is not defined in the markup, we need - * to have some fallback. for example, in zh, normally people + * In case some variant is not defined in the markup, we need + * to have some fallback. For example, in zh, normally people * will define zh-hans and zh-hant, but less so for zh-sg or zh-hk. * when zh-sg is preferred but not defined, we will pick zh-hans - * in this case. right now this is only used by zh. + * in this case. Right now this is only used by zh. * - * @param string $v the language code of the variant - * @return string array the code of the fallback language or false if there is no fallback - * @public + * @param $variant String: the language code of the variant + * @return String|array: The code of the fallback language or the + * main code if there is no fallback */ - function getVariantFallbacks($v) { - if( isset( $this->mVariantFallbacks[$v] ) ) { - return $this->mVariantFallbacks[$v]; + public function getVariantFallbacks( $variant ) { + if ( isset( $this->mVariantFallbacks[$variant] ) ) { + return $this->mVariantFallbacks[$variant]; } return $this->mMainLanguageCode; } /** - * get preferred language variants. - * @param boolean $fromUser Get it from $wgUser's preferences - * @return string the preferred language code - * @public + * Get the title produced by the conversion rule. + * @return String: The converted title text */ - function getPreferredVariant( $fromUser = true ) { - global $wgUser, $wgRequest, $wgVariantArticlePath, $wgDefaultLanguageVariant; + public function getConvRuleTitle() { + return $this->mConvRuleTitle; + } - if($this->mPreferredVariant) - return $this->mPreferredVariant; + /** + * Get preferred language variant. + * @return String: the preferred language code + */ + public function getPreferredVariant() { + global $wgDefaultLanguageVariant, $wgUser; - // figure out user lang without constructing wgLang to avoid infinite recursion - if( $fromUser ) - $defaultUserLang = $wgUser->getOption( 'language' ); - else - $defaultUserLang = $this->mMainLanguageCode; - $userLang = $wgRequest->getVal( 'uselang', $defaultUserLang ); - // see if interface language is same as content, if not, prevent conversion - if( ! in_array( $userLang, $this->mVariants ) ){ - $this->mPreferredVariant = $this->mMainLanguageCode; // no conversion - return $this->mPreferredVariant; + $req = $this->getURLVariant(); + + if ( $wgUser->isLoggedIn() && !$req ) { + $req = $this->getUserVariant(); + } elseif ( !$req ) { + $req = $this->getHeaderVariant(); } - // see if the preference is set in the request - $req = $wgRequest->getText( 'variant' ); - if( in_array( $req, $this->mVariants ) ) { - $this->mPreferredVariant = $req; - return $req; + if ( $wgDefaultLanguageVariant && !$req ) { + $req = $this->validateVariant( $wgDefaultLanguageVariant ); } - // check the syntax /code/ArticleTitle - if($wgVariantArticlePath!=false && isset($_SERVER['SCRIPT_NAME'])){ - // Note: SCRIPT_NAME probably won't hold the correct value if PHP is run as CGI - // (it will hold path to php.cgi binary), and might not exist on some very old PHP installations - $scriptBase = basename( $_SERVER['SCRIPT_NAME'] ); - if(in_array($scriptBase,$this->mVariants)){ - $this->mPreferredVariant = $scriptBase; - return $this->mPreferredVariant; - } + // This function, unlike the other get*Variant functions, is + // not memoized (i.e. there return value is not cached) since + // new information might appear during processing after this + // is first called. + if ( $this->validateVariant( $req ) ) { + return $req; } + return $this->mMainLanguageCode; + } - // get language variant preference from logged in users - // Don't call this on stub objects because that causes infinite - // recursion during initialisation - if( $fromUser && $wgUser->isLoggedIn() ) { - $this->mPreferredVariant = $wgUser->getOption('variant'); - return $this->mPreferredVariant; - } - - // see if default variant is globaly set - if($wgDefaultLanguageVariant != false && in_array( $wgDefaultLanguageVariant, $this->mVariants )){ - $this->mPreferredVariant = $wgDefaultLanguageVariant; - return $this->mPreferredVariant; - } - - if( !$this->mPreferredVariant ) { - // see if some supported language variant is set in the - // http header, but we don't set the mPreferredVariant - // variable in case this is called before the user's - // preference is loaded - if( array_key_exists( 'HTTP_ACCEPT_LANGUAGE', $_SERVER ) ) { - $acceptLanguage = strtolower( $_SERVER['HTTP_ACCEPT_LANGUAGE'] ); - - // explode by comma - $result = explode(',', $acceptLanguage); - - $languages = array(); - - foreach( $result as $elem ) { - // if $elem likes 'zh-cn;q=0.9' - if(($posi = strpos( $elem, ';' )) !== false ) { - // get the real language code likes 'zh-cn' - $languages[] = substr( $elem, 0, $posi ); - } - else { - $languages[] = $elem; - } - } + /** + * Get default variant. + * This function would not be affected by user's settings or headers + * @return String: the default variant code + */ + public function getDefaultVariant() { + global $wgDefaultLanguageVariant; - foreach( $languages as $language ) { - // strip whitespace - $language = trim( $language ); - if( in_array( $language, $this->mVariants ) ) { - $this->mPreferredVariant = $language; - return $this->mPreferredVariant; - } - } - } + $req = $this->getURLVariant(); + + if ( $wgDefaultLanguageVariant && !$req ) { + $req = $this->validateVariant( $wgDefaultLanguageVariant ); } - $this->mPreferredVariant = $this->mMainLanguageCode; - return $this->mPreferredVariant; + if ( $req ) { + return $req; + } + return $this->mMainLanguageCode; + } + + /** + * Validate the variant + * @param $variant String: the variant to validate + * @return Mixed: returns the variant if it is valid, null otherwise + */ + public function validateVariant( $variant = null ) { + if ( $variant !== null && in_array( $variant, $this->mVariants ) ) { + return $variant; + } + return null; } - + /** - * caption convert, base on preg_replace_callback + * Get the variant specified in the URL * - * to convert text in "title" or "alt", like 'textmURLVariant ) { + return $this->mURLVariant; + } + + // see if the preference is set in the request + $ret = $wgRequest->getText( 'variant' ); + + if ( !$ret ) { + $ret = $wgRequest->getVal( 'uselang' ); + } + + return $this->mURLVariant = $this->validateVariant( $ret ); + } + + /** + * Determine if the user has a variant set. * - * @return string like ' alt="yyyy"' or ' title="yyyy"' - * @private + * @return Mixed: variant if one found, false otherwise. */ - function captionConvert( $matches ) { - $toVariant = $this->getPreferredVariant(); - $title = $matches[1]; - $text = $matches[2]; - // we convert captions except URL - if( !strpos( $text, '://' ) ) - $text = $this->translate($text, $toVariant); - return " $title=\"$text\""; + protected function getUserVariant() { + global $wgUser; + + // memoizing this function wreaks havoc on parserTest.php + /* + if ( $this->mUserVariant ) { + return $this->mUserVariant; + } + */ + + // Get language variant preference from logged in users + // Don't call this on stub objects because that causes infinite + // recursion during initialisation + if ( $wgUser->isLoggedIn() ) { + $ret = $wgUser->getOption( 'variant' ); + } else { + // figure out user lang without constructing wgLang to avoid + // infinite recursion + $ret = $wgUser->getOption( 'language' ); + } + + return $this->mUserVariant = $this->validateVariant( $ret ); } /** - * dictionary-based conversion + * Determine the language variant from the Accept-Language header. * - * @param string $text the text to be converted - * @param string $toVariant the target language code - * @return string the converted text - * @private + * @return Mixed: variant if one found, false otherwise. */ - function autoConvert($text, $toVariant=false) { - $fname="LanguageConverter::autoConvert"; + protected function getHeaderVariant() { + global $wgRequest; - wfProfileIn( $fname ); + if ( $this->mHeaderVariant ) { + return $this->mHeaderVariant; + } - if(!$this->mTablesLoaded) - $this->loadTables(); + // see if some supported language variant is set in the + // HTTP header. + $languages = array_keys( $wgRequest->getAcceptLang() ); + if ( empty( $languages ) ) { + return null; + } + + $fallbackLanguages = array(); + foreach ( $languages as $language ) { + $this->mHeaderVariant = $this->validateVariant( $language ); + if ( $this->mHeaderVariant ) { + break; + } + + // To see if there are fallbacks of current language. + // We record these fallback variants, and process + // them later. + $fallbacks = $this->getVariantFallbacks( $language ); + if ( is_string( $fallbacks ) ) { + $fallbackLanguages[] = $fallbacks; + } elseif ( is_array( $fallbacks ) ) { + $fallbackLanguages = + array_merge( $fallbackLanguages, $fallbacks ); + } + } + + if ( !$this->mHeaderVariant ) { + // process fallback languages now + $fallback_languages = array_unique( $fallbackLanguages ); + foreach ( $fallback_languages as $language ) { + $this->mHeaderVariant = $this->validateVariant( $language ); + if ( $this->mHeaderVariant ) { + break; + } + } + } + + return $this->mHeaderVariant; + } + + /** + * Dictionary-based conversion. + * This function would not parse the conversion rules. + * If you want to parse rules, try to use convert() or + * convertTo(). + * + * @param $text String the text to be converted + * @param $toVariant bool|string the target language code + * @return String the converted text + */ + public function autoConvert( $text, $toVariant = false ) { + wfProfileIn( __METHOD__ ); - if(!$toVariant) + $this->loadTables(); + + if ( !$toVariant ) { $toVariant = $this->getPreferredVariant(); - if(!in_array($toVariant, $this->mVariants)) + if ( !$toVariant ) { + wfProfileOut( __METHOD__ ); + return $text; + } + } + + if( $this->guessVariant( $text, $toVariant ) ) { + wfProfileOut( __METHOD__ ); return $text; + } /* we convert everything except: - 1. html markups (anything between < and >) - 2. html entities - 3. place holders created by the parser + 1. HTML markups (anything between < and >) + 2. HTML entities + 3. placeholders created by the parser */ global $wgParser; - if (isset($wgParser) && $wgParser->UniqPrefix()!=''){ + if ( isset( $wgParser ) && $wgParser->UniqPrefix() != '' ) { $marker = '|' . $wgParser->UniqPrefix() . '[\-a-zA-Z0-9]+'; - } else - $marker = ""; + } else { + $marker = ''; + } - // this one is needed when the text is inside an html markup + // this one is needed when the text is inside an HTML markup $htmlfix = '|<[^>]+$|^[^<>]*>'; // disable convert to variants between tags @@ -272,314 +349,408 @@ class LanguageConverter { // disable conversion of
 ... 
$prefix = '.*?<\/pre>|'; - $reg = '/'.$codefix . $scriptfix . $prefix . '<[^>]+>|&[a-zA-Z#][a-z0-9]+;' . $marker . $htmlfix . '/s'; + $reg = '/' . $codefix . $scriptfix . $prefix . + '<[^>]+>|&[a-zA-Z#][a-z0-9]+;' . $marker . $htmlfix . '/s'; + $startPos = 0; + $sourceBlob = ''; + $literalBlob = ''; + + // Guard against delimiter nulls in the input + $text = str_replace( "\000", '', $text ); + + $markupMatches = null; + $elementMatches = null; + while ( $startPos < strlen( $text ) ) { + if ( preg_match( $reg, $text, $markupMatches, PREG_OFFSET_CAPTURE, $startPos ) ) { + $elementPos = $markupMatches[0][1]; + $element = $markupMatches[0][0]; + } else { + $elementPos = strlen( $text ); + $element = ''; + } - $matches = preg_split($reg, $text, -1, PREG_SPLIT_OFFSET_CAPTURE); + // Queue the part before the markup for translation in a batch + $sourceBlob .= substr( $text, $startPos, $elementPos - $startPos ) . "\000"; + + // Advance to the next position + $startPos = $elementPos + strlen( $element ); + + // Translate any alt or title attributes inside the matched element + if ( $element !== '' && preg_match( '/^(<[^>\s]*)\s([^>]*)(.*)$/', $element, + $elementMatches ) ) + { + $attrs = Sanitizer::decodeTagAttributes( $elementMatches[2] ); + $changed = false; + foreach ( array( 'title', 'alt' ) as $attrName ) { + if ( !isset( $attrs[$attrName] ) ) { + continue; + } + $attr = $attrs[$attrName]; + // Don't convert URLs + if ( !strpos( $attr, '://' ) ) { + $attr = $this->translate( $attr, $toVariant ); + } - $m = array_shift($matches); + // Remove HTML tags to avoid disrupting the layout + $attr = preg_replace( '/<[^>]+>/', '', $attr ); + if ( $attr !== $attrs[$attrName] ) { + $attrs[$attrName] = $attr; + $changed = true; + } + } + if ( $changed ) { + $element = $elementMatches[1] . Html::expandAttributes( $attrs ) . + $elementMatches[3]; + } + } + $literalBlob .= $element . "\000"; + } - $ret = $this->translate($m[0], $toVariant); - $mstart = $m[1]+strlen($m[0]); - - // enable convertsion of 'xxxxtranslate($m[0], $toVariant); - $mstart = $m[1] + strlen($m[0]); + // Do the main translation batch + $translatedBlob = $this->translate( $sourceBlob, $toVariant ); + + // Put the output back together + $translatedIter = StringUtils::explode( "\000", $translatedBlob ); + $literalIter = StringUtils::explode( "\000", $literalBlob ); + $output = ''; + while ( $translatedIter->valid() && $literalIter->valid() ) { + $output .= $translatedIter->current(); + $output .= $literalIter->current(); + $translatedIter->next(); + $literalIter->next(); } - wfProfileOut( $fname ); - return $ret; + + wfProfileOut( __METHOD__ ); + return $output; } /** - * Translate a string to a variant - * Doesn't process markup or do any of that other stuff, for that use convert() + * Translate a string to a variant. + * Doesn't parse rules or do any of that other stuff, for that use + * convert() or convertTo(). * - * @param string $text Text to convert - * @param string $variant Variant language code - * @return string Translated text - * @private + * @param $text String: text to convert + * @param $variant String: variant language code + * @return String: translated text */ - function translate( $text, $variant ) { + public function translate( $text, $variant ) { wfProfileIn( __METHOD__ ); - if( !$this->mTablesLoaded ) + // If $text is empty or only includes spaces, do nothing + // Otherwise translate it + if ( trim( $text ) ) { $this->loadTables(); - $text = $this->mTables[$variant]->replace( $text ); + $text = $this->mTables[$variant]->replace( $text ); + } wfProfileOut( __METHOD__ ); return $text; } /** - * convert text to all supported variants + * Call translate() to convert text to all valid variants. * - * @param string $text the text to be converted - * @return array of string - * @public + * @param $text String: the text to be converted + * @return Array: variant => converted text */ - function autoConvertToAllVariants($text) { - $fname="LanguageConverter::autoConvertToAllVariants"; - wfProfileIn( $fname ); - if( !$this->mTablesLoaded ) - $this->loadTables(); + public function autoConvertToAllVariants( $text ) { + wfProfileIn( __METHOD__ ); + $this->loadTables(); $ret = array(); - foreach($this->mVariants as $variant) { - $ret[$variant] = $this->translate($text, $variant); + foreach ( $this->mVariants as $variant ) { + $ret[$variant] = $this->translate( $text, $variant ); } - wfProfileOut( $fname ); + wfProfileOut( __METHOD__ ); return $ret; } /** - * convert link text to all supported variants + * Convert link text to all valid variants. + * In the first, this function only convert text outside the + * "-{" "}-" markups. Since the "{" and "}" are not allowed in + * titles, the text will get all converted always. + * So I removed this feature and deprecated the function. * - * @param string $text the text to be converted - * @return array of string - * @public + * @param $text String: the text to be converted + * @return Array: variant => converted text + * @deprecated since 1.17 Use autoConvertToAllVariants() instead */ - function convertLinkToAllVariants($text) { - if( !$this->mTablesLoaded ) - $this->loadTables(); - - $ret = array(); - $tarray = explode($this->mMarkup['begin'], $text); - $tfirst = array_shift($tarray); - - foreach($this->mVariants as $variant) - $ret[$variant] = $this->translate($tfirst,$variant); - - foreach($tarray as $txt) { - $marked = explode($this->mMarkup['end'], $txt, 2); - - foreach($this->mVariants as $variant){ - $ret[$variant] .= $this->mMarkup['begin'].$marked[0].$this->mMarkup['end']; - if(array_key_exists(1, $marked)) - $ret[$variant] .= $this->translate($marked[1],$variant); - } - - } - - return $ret; + public function convertLinkToAllVariants( $text ) { + return $this->autoConvertToAllVariants( $text ); } - + /** - * prepare manual conversion table - * @private + * Apply manual conversion rules. + * + * @param $convRule ConverterRule Object of ConverterRule */ - function applyManualConv( $convRule ){ - // use syntax -{T|zh:TitleZh;zh-tw:TitleTw}- for custom conversion in title - $title = $convRule->getTitle(); - if( $title ){ - $this->mTitleFromFlag = true; - $this->mTitleDisplay = $title; + protected function applyManualConv( $convRule ) { + // Use syntax -{T|zh-cn:TitleCN; zh-tw:TitleTw}- to custom + // title conversion. + // Bug 24072: $mConvRuleTitle was overwritten by other manual + // rule(s) not for title, this breaks the title conversion. + $newConvRuleTitle = $convRule->getTitle(); + if ( $newConvRuleTitle ) { + // So I add an empty check for getTitle() + $this->mConvRuleTitle = $newConvRuleTitle; } - //apply manual conversion table to global table + // merge/remove manual conversion rules to/from global table $convTable = $convRule->getConvTable(); $action = $convRule->getRulesAction(); - foreach( $convTable as $variant => $pair ) { - if( !in_array( $variant, $this->mVariants ) )continue; - if( $action=="add" ) { - foreach( $pair as $from => $to ) { + foreach ( $convTable as $variant => $pair ) { + if ( !$this->validateVariant( $variant ) ) { + continue; + } + + if ( $action == 'add' ) { + foreach ( $pair as $from => $to ) { // to ensure that $from and $to not be left blank // so $this->translate() could always return a string - if ( $from || $to ) + if ( $from || $to ) { // more efficient than array_merge(), about 2.5 times. $this->mTables[$variant]->setPair( $from, $to ); + } } - } - elseif ( $action == "remove" ) { + } elseif ( $action == 'remove' ) { $this->mTables[$variant]->removeArray( $pair ); } } } /** - * Convert text using a parser object for context - * @public + * Auto convert a Title object to a readable string in the + * preferred variant. + * + * @param $title Title a object of Title + * @return String: converted title text */ - function parserConvert( $text, &$parser ) { - global $wgDisableLangConversion; - /* don't do anything if this is the conversion table */ - if ( $parser->getTitle()->getNamespace() == NS_MEDIAWIKI && - strpos($parser->mTitle->getText(), "Conversiontable") !== false ) - { - return $text; + public function convertTitle( $title ) { + $variant = $this->getPreferredVariant(); + $index = $title->getNamespace(); + if ( $index === NS_MAIN ) { + $text = ''; + } else { + // first let's check if a message has given us a converted name + $nsConvMsg = wfMessage( 'conversion-ns' . $index )->inContentLanguage(); + if ( $nsConvMsg->exists() ) { + $text = $nsConvMsg->plain(); + } else { + // the message does not exist, try retrieve it from the current + // variant's namespace names. + $langObj = $this->mLangObj->factory( $variant ); + $text = $langObj->getFormattedNsText( $index ); + } + $text .= ':'; } - - if ( $wgDisableLangConversion ) - return $text; - - $text = $this->convert( $text ); - - if ( $this->mTitleFromFlag ) - $parser->mOutput->setTitleText( $this->mTitleDisplay ); + $text .= $title->getText(); + $text = $this->translate( $text, $variant ); return $text; } - + /** - * convert namespace - * @param string $title the title included namespace - * @return array of string - * @private + * Convert text to different variants of a language. The automatic + * conversion is done in autoConvert(). Here we parse the text + * marked with -{}-, which specifies special conversions of the + * text that can not be accomplished in autoConvert(). + * + * Syntax of the markup: + * -{code1:text1;code2:text2;...}- or + * -{flags|code1:text1;code2:text2;...}- or + * -{text}- in which case no conversion should take place for text + * + * @param $text String: text to be converted + * @return String: converted text */ - function convertNamespace( $title, $variant ) { - $splittitle = explode( ':', $title ); - if (count($splittitle) < 2) - return $title; - if ( isset( $this->mNamespaceTables[$variant][$splittitle[0]] ) ) - $splittitle[0] = $this->mNamespaceTables[$variant][$splittitle[0]]; - $ret = implode(':', $splittitle ); - return $ret; + public function convert( $text ) { + $variant = $this->getPreferredVariant(); + return $this->convertTo( $text, $variant ); } /** - * convert title - * @private + * Same as convert() except a extra parameter to custom variant. + * + * @param $text String: text to be converted + * @param $variant String: the target variant code + * @return String: converted text */ - function convertTitle( $text, $variant ){ - global $wgDisableTitleConversion, $wgUser; - - // check for global param and __NOTC__ tag - if( $wgDisableTitleConversion || !$this->mDoTitleConvert || $wgUser->getOption('noconvertlink') == 1 ) { - $this->mTitleDisplay = $text; + public function convertTo( $text, $variant ) { + global $wgDisableLangConversion; + if ( $wgDisableLangConversion || $this->guessVariant( $text, $variant ) ) { return $text; } - - // use the title from the T flag if any - if( $this->mTitleFromFlag ){ - $this->mTitleFromFlag = false; - return $this->mTitleDisplay; - } - - global $wgRequest; - $isredir = $wgRequest->getText( 'redirect', 'yes' ); - $action = $wgRequest->getText( 'action' ); - $linkconvert = $wgRequest->getText( 'linkconvert', 'yes' ); - if ( $isredir == 'no' || $action == 'edit' || $action == 'submit' || $linkconvert == 'no' ) { - return $text; - } else { - $text = $this->convertNamespace( $text, $variant ); - $this->mTitleDisplay = $this->convert( $text ); - return $this->mTitleDisplay; - } + return $this->recursiveConvertTopLevel( $text, $variant ); } /** - * convert text to different variants of a language. the automatic - * conversion is done in autoConvert(). here we parse the text - * marked with -{}-, which specifies special conversions of the - * text that can not be accomplished in autoConvert() + * Recursively convert text on the outside. Allow to use nested + * markups to custom rules. * - * syntax of the markup: - * -{code1:text1;code2:text2;...}- or - * -{flags|code1:text1;code2:text2;...}- or - * -{text}- in which case no conversion should take place for text - * - * @param string $text text to be converted - * @param bool $isTitle whether this conversion is for the article title - * @return string converted text - * @public + * @param $text String: text to be converted + * @param $variant String: the target variant code + * @param $depth Integer: depth of recursion + * @return String: converted text */ - function convert( $text, $isTitle = false ) { - - $mw =& MagicWord::get( 'notitleconvert' ); - if( $mw->matchAndRemove( $text ) ) - $this->mDoTitleConvert = false; - $mw =& MagicWord::get( 'nocontentconvert' ); - if( $mw->matchAndRemove( $text ) ) { - $this->mDoContentConvert = false; - } - - // no conversion if redirecting - $mw =& MagicWord::get( 'redirect' ); - if( $mw->matchStart( $text ) ) - return $text; + protected function recursiveConvertTopLevel( $text, $variant, $depth = 0 ) { + $startPos = 0; + $out = ''; + $length = strlen( $text ); + while ( $startPos < $length ) { + $pos = strpos( $text, '-{', $startPos ); + + if ( $pos === false ) { + // No more markup, append final segment + $out .= $this->autoConvert( substr( $text, $startPos ), $variant ); + return $out; + } - $plang = $this->getPreferredVariant(); + // Markup found + // Append initial segment + $out .= $this->autoConvert( substr( $text, $startPos, $pos - $startPos ), $variant ); - // for title convertion - if ( $isTitle ) return $this->convertTitle( $text, $plang ); + // Advance position + $startPos = $pos; - $tarray = StringUtils::explode( $this->mMarkup['end'], $text ); - $text = ''; + // Do recursive conversion + $out .= $this->recursiveConvertRule( $text, $variant, $startPos, $depth + 1 ); + } - foreach ( $tarray as $txt ) { + return $out; + } - $marked = explode( $this->mMarkup['begin'], $txt, 2 ); + /** + * Recursively convert text on the inside. + * + * @param $text String: text to be converted + * @param $variant String: the target variant code + * @param $startPos int + * @param $depth Integer: depth of recursion + * + * @return String: converted text + */ + protected function recursiveConvertRule( $text, $variant, &$startPos, $depth = 0 ) { + // Quick sanity check (no function calls) + if ( $text[$startPos] !== '-' || $text[$startPos + 1] !== '{' ) { + throw new MWException( __METHOD__ . ': invalid input string' ); + } - if( $this->mDoContentConvert ) - // Bug 19620: should convert a string immediately after a new rule added. - $text .= $this->autoConvert( $marked[0], $plang ); + $startPos += 2; + $inner = ''; + $warningDone = false; + $length = strlen( $text ); - if ( array_key_exists( 1, $marked ) ) { - $crule = new ConverterRule($marked[1], $this); - $crule->parse( $plang ); - $text .= $crule->getDisplay(); - $this->applyManualConv( $crule ); + while ( $startPos < $length ) { + $m = false; + preg_match( '/-\{|\}-/', $text, $m, PREG_OFFSET_CAPTURE, $startPos ); + if ( !$m ) { + // Unclosed rule + break; } - else - $text .= $this->mMarkup['end']; + $token = $m[0][0]; + $pos = $m[0][1]; + + // Markup found + // Append initial segment + $inner .= substr( $text, $startPos, $pos - $startPos ); + + // Advance position + $startPos = $pos; + + switch ( $token ) { + case '-{': + // Check max depth + if ( $depth >= $this->mMaxDepth ) { + $inner .= '-{'; + if ( !$warningDone ) { + $inner .= '' . + wfMsgForContent( 'language-converter-depth-warning', + $this->mMaxDepth ) . + ''; + $warningDone = true; + } + $startPos += 2; + continue; + } + // Recursively parse another rule + $inner .= $this->recursiveConvertRule( $text, $variant, $startPos, $depth + 1 ); + break; + case '}-': + // Apply the rule + $startPos += 2; + $rule = new ConverterRule( $inner, $this ); + $rule->parse( $variant ); + $this->applyManualConv( $rule ); + return $rule->getDisplay(); + default: + throw new MWException( __METHOD__ . ': invalid regex match' ); + } } - // Remove the last delimiter (wasn't real) - $text = substr( $text, 0, -strlen( $this->mMarkup['end'] ) ); - - return $text; + // Unclosed rule + if ( $startPos < $length ) { + $inner .= substr( $text, $startPos ); + } + $startPos = $length; + return '-{' . $this->autoConvert( $inner, $variant ); } /** - * if a language supports multiple variants, it is - * possible that non-existing link in one variant - * actually exists in another variant. this function - * tries to find it. See e.g. LanguageZh.php + * If a language supports multiple variants, it is possible that + * non-existing link in one variant actually exists in another variant. + * This function tries to find it. See e.g. LanguageZh.php * - * @param string $link the name of the link - * @param mixed $nt the title object of the link - * @param boolean $ignoreOtherCond: to disable other conditions when - * we need to transclude a template or update a category's link - * @return null the input parameters may be modified upon return - * @public + * @param $link String: the name of the link + * @param $nt Mixed: the title object of the link + * @param $ignoreOtherCond Boolean: to disable other conditions when + * we need to transclude a template or update a category's link + * @return Null, the input parameters may be modified upon return */ - function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) { + public function findVariantLink( &$link, &$nt, $ignoreOtherCond = false ) { # If the article has already existed, there is no need to # check it again, otherwise it may cause a fault. - if ( $nt->exists() ) + if ( is_object( $nt ) && $nt->exists() ) { return; + } - global $wgDisableLangConversion, $wgDisableTitleConversion, $wgRequest, $wgUser; + global $wgDisableLangConversion, $wgDisableTitleConversion, $wgRequest, + $wgUser; $isredir = $wgRequest->getText( 'redirect', 'yes' ); $action = $wgRequest->getText( 'action' ); $linkconvert = $wgRequest->getText( 'linkconvert', 'yes' ); - $disableLinkConversion = $wgDisableLangConversion || $wgDisableTitleConversion; + $disableLinkConversion = $wgDisableLangConversion + || $wgDisableTitleConversion; $linkBatch = new LinkBatch(); - $ns=NS_MAIN; + $ns = NS_MAIN; - if ( $disableLinkConversion || ( !$ignoreOtherCond && ( $isredir == 'no' || $action == 'edit' - || $action == 'submit' || $linkconvert == 'no' || $wgUser->getOption('noconvertlink') == 1 ) ) ) + if ( $disableLinkConversion || + ( !$ignoreOtherCond && + ( $isredir == 'no' + || $action == 'edit' + || $action == 'submit' + || $linkconvert == 'no' + || $wgUser->getOption( 'noconvertlink' ) == 1 ) ) ) { return; + } - if ( is_object( $nt ) ) + if ( is_object( $nt ) ) { $ns = $nt->getNamespace(); + } - $variants = $this->autoConvertToAllVariants($link); - if($variants == false) //give up + $variants = $this->autoConvertToAllVariants( $link ); + if ( !$variants ) { // give up return; + } $titles = array(); - foreach( $variants as $v ) { - if($v != $link){ + foreach ( $variants as $v ) { + if ( $v != $link ) { $varnt = Title::newFromText( $v, $ns ); - if(!is_null($varnt)){ - $linkBatch->addObj($varnt); - $titles[]=$varnt; + if ( !is_null( $varnt ) ) { + $linkBatch->addObj( $varnt ); + $titles[] = $varnt; } } } @@ -587,8 +758,8 @@ class LanguageConverter { // fetch all variants in single query $linkBatch->execute(); - foreach( $titles as $varnt ) { - if( $varnt->getArticleID() > 0 ) { + foreach ( $titles as $varnt ) { + if ( $varnt->getArticleID() > 0 ) { $nt = $varnt; $link = $varnt->getText(); break; @@ -596,214 +767,208 @@ class LanguageConverter { } } - /** - * returns language specific hash options + /** + * Returns language specific hash options. * - * @public + * @return string */ - function getExtraHashOptions() { + public function getExtraHashOptions() { $variant = $this->getPreferredVariant(); - return '!' . $variant ; - } - - /** - * get title text as defined in the body of the article text - * - * @public - */ - function getParsedTitle() { - return $this->mTitleDisplay; + return '!' . $variant; } /** - * a write lock to the cache + * Guess if a text is written in a variant. This should be implemented in subclasses. * - * @private - */ - function lockCache() { - global $wgMemc; - $success = false; - for($i=0; $i<30; $i++) { - if($success = $wgMemc->add($this->mCacheKey . "lock", 1, 10)) - break; - sleep(1); - } - return $success; - } - - /** - * unlock cache + * @param string $text the text to be checked + * @param string $variant language code of the variant to be checked for + * @return bool true if $text appears to be written in $variant, false if not * - * @private + * @author Nikola Smolenski + * @since 1.19 */ - function unlockCache() { - global $wgMemc; - $wgMemc->delete($this->mCacheKey . "lock"); + public function guessVariant($text, $variant) { + return false; } - /** - * Load default conversion tables - * This method must be implemented in derived class + * Load default conversion tables. + * This method must be implemented in derived class. * * @private */ function loadDefaultTables() { - $name = get_class($this); - wfDie("Must implement loadDefaultTables() method in class $name"); + $name = get_class( $this ); + throw new MWException( "Must implement loadDefaultTables() method in class $name" ); } /** - * load conversion tables either from the cache or the disk + * Load conversion tables either from the cache or the disk. * @private + * @param $fromCache Boolean: load from memcached? Defaults to true. */ - function loadTables($fromcache=true) { - global $wgMemc; - if( $this->mTablesLoaded ) + function loadTables( $fromCache = true ) { + if ( $this->mTablesLoaded ) { return; + } + global $wgMemc; wfProfileIn( __METHOD__ ); $this->mTablesLoaded = true; $this->mTables = false; - if($fromcache) { - wfProfileIn( __METHOD__.'-cache' ); + if ( $fromCache ) { + wfProfileIn( __METHOD__ . '-cache' ); $this->mTables = $wgMemc->get( $this->mCacheKey ); - wfProfileOut( __METHOD__.'-cache' ); + wfProfileOut( __METHOD__ . '-cache' ); } - if ( !$this->mTables || !isset( $this->mTables[self::CACHE_VERSION_KEY] ) ) { - wfProfileIn( __METHOD__.'-recache' ); + if ( !$this->mTables + || !array_key_exists( self::CACHE_VERSION_KEY, $this->mTables ) ) { + wfProfileIn( __METHOD__ . '-recache' ); // not in cache, or we need a fresh reload. - // we will first load the default tables - // then update them using things in MediaWiki:Zhconversiontable/* + // We will first load the default tables + // then update them using things in MediaWiki:Conversiontable/* $this->loadDefaultTables(); - foreach($this->mVariants as $var) { - $cached = $this->parseCachedTable($var); - $this->mTables[$var]->mergeArray($cached); + foreach ( $this->mVariants as $var ) { + $cached = $this->parseCachedTable( $var ); + $this->mTables[$var]->mergeArray( $cached ); } $this->postLoadTables(); $this->mTables[self::CACHE_VERSION_KEY] = true; - if($this->lockCache()) { - $wgMemc->set($this->mCacheKey, $this->mTables, 43200); - $this->unlockCache(); - } - wfProfileOut( __METHOD__.'-recache' ); + $wgMemc->set( $this->mCacheKey, $this->mTables, 43200 ); + wfProfileOut( __METHOD__ . '-recache' ); } wfProfileOut( __METHOD__ ); } - /** - * Hook for post processig after conversion tables are loaded - * + /** + * Hook for post processing after conversion tables are loaded. */ - function postLoadTables() {} + function postLoadTables() { } - /** - * Reload the conversion tables + /** + * Reload the conversion tables. * * @private */ function reloadTables() { - if($this->mTables) - unset($this->mTables); + if ( $this->mTables ) { + unset( $this->mTables ); + } $this->mTablesLoaded = false; - $this->loadTables(false); + $this->loadTables( false ); } - /** - * parse the conversion table stored in the cache + * Parse the conversion table stored in the cache. * - * the tables should be in blocks of the following form: + * The tables should be in blocks of the following form: * -{ * word => word ; * word => word ; * ... * }- * - * to make the tables more manageable, subpages are allowed - * and will be parsed recursively if $recursive=true + * To make the tables more manageable, subpages are allowed + * and will be parsed recursively if $recursive == true. * + * @param $code String: language code + * @param $subpage String: subpage name + * @param $recursive Boolean: parse subpages recursively? Defaults to true. + * + * @return array */ - function parseCachedTable($code, $subpage='', $recursive=true) { - global $wgMessageCache; + function parseCachedTable( $code, $subpage = '', $recursive = true ) { static $parsed = array(); - if(!is_object($wgMessageCache)) - return array(); - - $key = 'Conversiontable/'.$code; - if($subpage) + $key = 'Conversiontable/' . $code; + if ( $subpage ) { $key .= '/' . $subpage; - - if(array_key_exists($key, $parsed)) + } + if ( array_key_exists( $key, $parsed ) ) { return array(); + } - if ( strpos( $code, '/' ) === false ) { - $txt = $wgMessageCache->get( 'Conversiontable', true, $code ); + $parsed[$key] = true; + + if ( $subpage === '' ) { + $txt = MessageCache::singleton()->get( 'conversiontable', true, $code ); } else { - $title = Title::makeTitleSafe( NS_MEDIAWIKI, "Conversiontable/$code" ); + $txt = false; + $title = Title::makeTitleSafe( NS_MEDIAWIKI, $key ); if ( $title && $title->exists() ) { - $article = new Article( $title ); - $txt = $article->getContents(); - } else { - $txt = ''; + $revision = Revision::newFromTitle( $title ); + if ( $revision ) { + $txt = $revision->getRawText(); + } } } + # Nothing to parse if there's no text + if ( $txt === false || $txt === null || $txt === '' ) { + return array(); + } + // get all subpage links of the form - // [[MediaWiki:conversiontable/zh-xx/...|...]] - $linkhead = $this->mLangObj->getNsText(NS_MEDIAWIKI) . ':Conversiontable'; - $subs = explode('[[', $txt); + // [[MediaWiki:Conversiontable/zh-xx/...|...]] + $linkhead = $this->mLangObj->getNsText( NS_MEDIAWIKI ) . + ':Conversiontable'; + $subs = StringUtils::explode( '[[', $txt ); $sublinks = array(); - foreach( $subs as $sub ) { - $link = explode(']]', $sub, 2); - if(count($link) != 2) + foreach ( $subs as $sub ) { + $link = explode( ']]', $sub, 2 ); + if ( count( $link ) != 2 ) { continue; - $b = explode('|', $link[0]); - $b = explode('/', trim($b[0]), 3); - if(count($b)==3) + } + $b = explode( '|', $link[0], 2 ); + $b = explode( '/', trim( $b[0] ), 3 ); + if ( count( $b ) == 3 ) { $sublink = $b[2]; - else + } else { $sublink = ''; + } - if($b[0] == $linkhead && $b[1] == $code) { + if ( $b[0] == $linkhead && $b[1] == $code ) { $sublinks[] = $sublink; } } - // parse the mappings in this page - $blocks = explode($this->mMarkup['begin'], $txt); - array_shift($blocks); + $blocks = StringUtils::explode( '-{', $txt ); $ret = array(); - foreach($blocks as $block) { - $mappings = explode($this->mMarkup['end'], $block, 2); - $stripped = str_replace(array("'", '"', '*','#'), '', $mappings[0]); - $table = explode( ';', $stripped ); - foreach( $table as $t ) { - $m = explode( '=>', $t ); - if( count( $m ) != 2) + $first = true; + foreach ( $blocks as $block ) { + if ( $first ) { + // Skip the part before the first -{ + $first = false; + continue; + } + $mappings = explode( '}-', $block, 2 ); + $stripped = str_replace( array( "'", '"', '*', '#' ), '', + $mappings[0] ); + $table = StringUtils::explode( ';', $stripped ); + foreach ( $table as $t ) { + $m = explode( '=>', $t, 3 ); + if ( count( $m ) != 2 ) { continue; + } // trim any trailling comments starting with '//' - $tt = explode('//', $m[1], 2); - $ret[trim($m[0])] = trim($tt[0]); + $tt = explode( '//', $m[1], 2 ); + $ret[trim( $m[0] )] = trim( $tt[0] ); } } - $parsed[$key] = true; - // recursively parse the subpages - if($recursive) { - foreach($sublinks as $link) { - $s = $this->parseCachedTable($code, $link, $recursive); - $ret = array_merge($ret, $s); + if ( $recursive ) { + foreach ( $sublinks as $link ) { + $s = $this->parseCachedTable( $code, $link, $recursive ); + $ret = array_merge( $ret, $s ); } } - if ($this->mUcfirst) { - foreach ($ret as $k => $v) { - $ret[Language::ucfirst($k)] = Language::ucfirst($v); + if ( $this->mUcfirst ) { + foreach ( $ret as $k => $v ) { + $ret[$this->mLangObj->ucfirst( $k )] = $this->mLangObj->ucfirst( $v ); } } return $ret; @@ -811,42 +976,59 @@ class LanguageConverter { /** * Enclose a string with the "no conversion" tag. This is used by - * various functions in the Parser + * various functions in the Parser. * - * @param string $text text to be tagged for no conversion - * @return string the tagged text - * @public + * @param $text String: text to be tagged for no conversion + * @param $noParse Boolean: unused + * @return String: the tagged text */ - function markNoConversion($text, $noParse=false) { + public function markNoConversion( $text, $noParse = false ) { # don't mark if already marked - if(strpos($text, $this->mMarkup['begin']) || - strpos($text, $this->mMarkup['end'])) + if ( strpos( $text, '-{' ) || strpos( $text, '}-' ) ) { return $text; + } - $ret = $this->mMarkup['begin'] .'R|'. $text . $this->mMarkup['end']; + $ret = "-{R|$text}-"; return $ret; } /** - * convert the sorting key for category links. this should make different - * keys that are variants of each other map to the same key + * Convert the sorting key for category links. This should make different + * keys that are variants of each other map to the same key. + * + * @param $key string + * + * @return string */ function convertCategoryKey( $key ) { return $key; } + /** - * hook to refresh the cache of conversion tables when - * MediaWiki:conversiontable* is updated + * Hook to refresh the cache of conversion tables when + * MediaWiki:Conversiontable* is updated. * @private + * + * @param $article Article object + * @param $user Object: User object for the current user + * @param $text String: article text (?) + * @param $summary String: edit summary of the edit + * @param $isMinor Boolean: was the edit marked as minor? + * @param $isWatch Boolean: did the user watch this page or not? + * @param $section Unused + * @param $flags Bitfield + * @param $revision Object: new Revision object or null + * @return Boolean: true */ - function OnArticleSaveComplete($article, $user, $text, $summary, $isminor, $iswatch, $section, $flags, $revision) { + function OnArticleSaveComplete( $article, $user, $text, $summary, $isMinor, + $isWatch, $section, $flags, $revision ) { $titleobj = $article->getTitle(); - if($titleobj->getNamespace() == NS_MEDIAWIKI) { + if ( $titleobj->getNamespace() == NS_MEDIAWIKI ) { $title = $titleobj->getDBkey(); - $t = explode('/', $title, 3); - $c = count($t); - if( $c > 1 && $t[0] == 'Conversiontable' ) { - if(in_array($t[1], $this->mVariants)) { + $t = explode( '/', $title, 3 ); + $c = count( $t ); + if ( $c > 1 && $t[0] == 'Conversiontable' ) { + if ( $this->validateVariant( $t[1] ) ) { $this->reloadTables(); } } @@ -854,33 +1036,66 @@ class LanguageConverter { return true; } - /** - * Armour rendered math against conversion - * Wrap math into rawoutput -{R| math }- syntax - * @public + /** + * Armour rendered math against conversion. + * Escape special chars in parsed math text. (in most cases are img elements) + * + * @param $text String: text to armour against conversion + * @return String: armoured text where { and } have been converted to + * { and } */ - function armourMath($text){ - // we need to convert '-{' and '}-' to '-{' and '}-' - // to avoid a unwanted '}-' appeared after the math-image. - $text = strtr( $text, array('-{' => '-{', '}-' => '}-') ); - $ret = $this->mMarkup['begin'] . 'R|' . $text . $this->mMarkup['end']; - return $ret; + public function armourMath( $text ) { + // convert '-{' and '}-' to '-{' and '}-' to prevent + // any unwanted markup appearing in the math image tag. + $text = strtr( $text, array( '-{' => '-{', '}-' => '}-' ) ); + return $text; + } + + /** + * Get the cached separator pattern for ConverterRule::parseRules() + */ + function getVarSeparatorPattern() { + if ( is_null( $this->mVarSeparatorPattern ) ) { + // varsep_pattern for preg_split: + // text should be splited by ";" only if a valid variant + // name exist after the markup, for example: + // -{zh-hans:xxx;zh-hant:\ + // yyy;}- + // we should split it as: + // array( + // [0] => 'zh-hans:xxx' + // [1] => 'zh-hant:yyy' + // [2] => '' + // ) + $pat = '/;\s*(?='; + foreach ( $this->mVariants as $variant ) { + // zh-hans:xxx;zh-hant:yyy + $pat .= $variant . '\s*:|'; + // xxx=>zh-hans:yyy; xxx=>zh-hant:zzz + $pat .= '[^;]*?=>\s*' . $variant . '\s*:|'; + } + $pat .= '\s*$)/'; + $this->mVarSeparatorPattern = $pat; + } + return $this->mVarSeparatorPattern; } } /** - * parser for rules of language conversion , parse rules in -{ }- tag + * Parser for rules of language conversion , parse rules in -{ }- tag. * @ingroup Language * @author fdcn , PhiLiP */ class ConverterRule { var $mText; // original text in -{text}- - var $mConverter; // LanguageConverter object - var $mManualCodeError='code error!'; - var $mRuleDisplay = '',$mRuleTitle=false; + var $mConverter; // LanguageConverter object + var $mManualCodeError = 'code error!'; + var $mRuleDisplay = ''; + var $mRuleTitle = false; var $mRules = '';// string : the text of the rules var $mRulesAction = 'none'; var $mFlags = array(); + var $mVariantFlags = array(); var $mConvTable = array(); var $mBidtable = array();// array of the translation in each variant var $mUnidtable = array();// array of the translation in each variant @@ -888,158 +1103,135 @@ class ConverterRule { /** * Constructor * - * @param string $text the text between -{ and }- - * @param object $converter a LanguageConverter object - * @access public + * @param $text String: the text between -{ and }- + * @param $converter LanguageConverter object */ - function __construct($text,$converter){ + public function __construct( $text, $converter ) { $this->mText = $text; - $this->mConverter=$converter; - foreach($converter->mVariants as $v){ - $this->mConvTable[$v]=array(); - } + $this->mConverter = $converter; } /** - * check if variants array in convert array + * Check if variants array in convert array. * - * @param string $variant Variant language code - * @return string Translated text - * @public + * @param $variants Array or string: variant language code + * @return String: translated text */ - function getTextInBidtable($variants){ - if(is_string($variants)){ $variants=array($variants); } - if(!is_array($variants)) return false; - foreach ($variants as $variant){ - if(array_key_exists($variant, $this->mBidtable)){ + public function getTextInBidtable( $variants ) { + $variants = (array)$variants; + if ( !$variants ) { + return false; + } + foreach ( $variants as $variant ) { + if ( isset( $this->mBidtable[$variant] ) ) { return $this->mBidtable[$variant]; } } return false; } - + /** * Parse flags with syntax -{FLAG| ... }- * @private */ - function parseFlags(){ + function parseFlags() { $text = $this->mText; - if(strlen($text) < 2 ) { - $this->mFlags = array( 'R' ); - $this->mRules = $text; - return; - } - $flags = array(); - $markup = $this->mConverter->mMarkup; - $validFlags = $this->mConverter->mFlags; - $variants = $this->mConverter->mVariants; - - $tt = explode($markup['flagsep'], $text, 2); - if(count($tt) == 2) { - $f = explode($markup['varsep'], $tt[0]); - foreach($f as $ff) { - $ff = trim($ff); - if(array_key_exists($ff, $validFlags) && - !in_array($validFlags[$ff], $flags)) - $flags[] = $validFlags[$ff]; + $variantFlags = array(); + + $sepPos = strpos( $text, '|' ); + if ( $sepPos !== false ) { + $validFlags = $this->mConverter->mFlags; + $f = StringUtils::explode( ';', substr( $text, 0, $sepPos ) ); + foreach ( $f as $ff ) { + $ff = trim( $ff ); + if ( isset( $validFlags[$ff] ) ) { + $flags[$validFlags[$ff]] = true; + } } - $rules = $tt[1]; - } else { - $rules = $text; - } - - //check flags - if( in_array('R',$flags) ){ - $flags = array('R');// remove other flags - } elseif ( in_array('N',$flags) ){ - $flags = array('N');// remove other flags - } elseif ( in_array('-',$flags) ){ - $flags = array('-');// remove other flags - } elseif (count($flags)==1 && $flags[0]=='T'){ - $flags[]='H'; - } elseif ( in_array('H',$flags) ){ + $text = strval( substr( $text, $sepPos + 1 ) ); + } + + if ( !$flags ) { + $flags['S'] = true; + } elseif ( isset( $flags['R'] ) ) { + $flags = array( 'R' => true );// remove other flags + } elseif ( isset( $flags['N'] ) ) { + $flags = array( 'N' => true );// remove other flags + } elseif ( isset( $flags['-'] ) ) { + $flags = array( '-' => true );// remove other flags + } elseif ( count( $flags ) == 1 && isset( $flags['T'] ) ) { + $flags['H'] = true; + } elseif ( isset( $flags['H'] ) ) { // replace A flag, and remove other flags except T - $temp=array('+','H'); - if(in_array('T',$flags)) $temp[] = 'T'; - if(in_array('D',$flags)) $temp[] = 'D'; + $temp = array( '+' => true, 'H' => true ); + if ( isset( $flags['T'] ) ) { + $temp['T'] = true; + } + if ( isset( $flags['D'] ) ) { + $temp['D'] = true; + } $flags = $temp; } else { - if ( in_array('A',$flags) ) { - $flags[]='+'; - $flags[]='S'; + if ( isset( $flags['A'] ) ) { + $flags['+'] = true; + $flags['S'] = true; + } + if ( isset( $flags['D'] ) ) { + unset( $flags['S'] ); } - if ( in_array('D',$flags) ) - $flags=array_diff($flags,array('S')); - $flags_temp = array(); - foreach ($variants as $variant) { - // try to find flags like "zh-hans", "zh-hant" - // allow syntaxes like "-{zh-hans;zh-hant|XXXX}-" - if ( in_array($variant, $flags) ) - $flags_temp[] = $variant; + // try to find flags like "zh-hans", "zh-hant" + // allow syntaxes like "-{zh-hans;zh-hant|XXXX}-" + $variantFlags = array_intersect( array_keys( $flags ), $this->mConverter->mVariants ); + if ( $variantFlags ) { + $variantFlags = array_flip( $variantFlags ); + $flags = array(); } - if ( count($flags_temp) !== 0 ) - $flags = $flags_temp; } - if ( count($flags) == 0 ) - $flags = array('S'); - $this->mRules=$rules; - $this->mFlags=$flags; + $this->mVariantFlags = $variantFlags; + $this->mRules = $text; + $this->mFlags = $flags; } - + /** - * generate conversion table + * Generate conversion table. * @private */ function parseRules() { $rules = $this->mRules; - $flags = $this->mFlags; $bidtable = array(); $unidtable = array(); - $markup = $this->mConverter->mMarkup; $variants = $this->mConverter->mVariants; + $varsep_pattern = $this->mConverter->getVarSeparatorPattern(); - // varsep_pattern for preg_split: - // text should be splited by ";" only if a valid variant - // name exist after the markup, for example: - // -{zh-hans:xxx;zh-hant:yyy;}- - // we should split it as: - // array( - // [0] => 'zh-hans:xxx' - // [1] => 'zh-hant:yyy' - // [2] => '' - // ) - $varsep_pattern = '/' . $markup['varsep'] . '\s*' . '(?='; - foreach( $variants as $variant ) { - $varsep_pattern .= $variant . '\s*' . $markup['codesep'] . '|'; // zh-hans:xxx;zh-hant:yyy - $varsep_pattern .= '[^;]*?' . $markup['unidsep'] . '\s*' . $variant - . '\s*' . $markup['codesep'] . '|'; // xxx=>zh-hans:yyy; xxx=>zh-hant:zzz - } - $varsep_pattern .= '\s*$)/'; - - $choice = preg_split($varsep_pattern, $rules); - - foreach( $choice as $c ) { - $v = explode($markup['codesep'], $c, 2); - if( count($v) != 2 ) - continue;// syntax error, skip - $to = trim($v[1]); - $v = trim($v[0]); - $u = explode($markup['unidsep'], $v); + $choice = preg_split( $varsep_pattern, $rules ); + + foreach ( $choice as $c ) { + $v = explode( ':', $c, 2 ); + if ( count( $v ) != 2 ) { + // syntax error, skip + continue; + } + $to = trim( $v[1] ); + $v = trim( $v[0] ); + $u = explode( '=>', $v, 2 ); // if $to is empty, strtr() could return a wrong result - if( count($u) == 1 && $to && in_array( $v, $variants ) ) { + if ( count( $u ) == 1 && $to && in_array( $v, $variants ) ) { $bidtable[$v] = $to; - } else if(count($u) == 2){ - $from = trim($u[0]); - $v = trim($u[1]); - if( array_key_exists( $v, $unidtable ) && !is_array( $unidtable[$v] ) - && $to && in_array( $v, $variants ) ) - $unidtable[$v] = array( $from=>$to ); - elseif ( $to && in_array( $v, $variants ) ) + } elseif ( count( $u ) == 2 ) { + $from = trim( $u[0] ); + $v = trim( $u[1] ); + if ( array_key_exists( $v, $unidtable ) + && !is_array( $unidtable[$v] ) + && $to + && in_array( $v, $variants ) ) { + $unidtable[$v] = array( $from => $to ); + } elseif ( $to && in_array( $v, $variants ) ) { $unidtable[$v][$from] = $to; + } } // syntax error, pass - if ( !array_key_exists( $v, $this->mConverter->mVariantNames ) ){ + if ( !isset( $this->mConverter->mVariantNames[$v] ) ) { $bidtable = array(); $unidtable = array(); break; @@ -1051,238 +1243,287 @@ class ConverterRule { /** * @private + * + * @return string */ - function getRulesDesc(){ + function getRulesDesc() { $codesep = $this->mConverter->mDescCodeSep; $varsep = $this->mConverter->mDescVarSep; - $text=''; - foreach($this->mBidtable as $k => $v) - $text .= $this->mConverter->mVariantNames[$k]."$codesep$v$varsep"; - foreach($this->mUnidtable as $k => $a) - foreach($a as $from=>$to) - $text.=$from.'⇒'.$this->mConverter->mVariantNames[$k]."$codesep$to$varsep"; + $text = ''; + foreach ( $this->mBidtable as $k => $v ) { + $text .= $this->mConverter->mVariantNames[$k] . "$codesep$v$varsep"; + } + foreach ( $this->mUnidtable as $k => $a ) { + foreach ( $a as $from => $to ) { + $text .= $from . '⇒' . $this->mConverter->mVariantNames[$k] . + "$codesep$to$varsep"; + } + } return $text; } /** - * Parse rules conversion + * Parse rules conversion. * @private + * + * @param $variant + * + * @return string */ - function getRuleConvertedStr($variant,$doConvert){ + function getRuleConvertedStr( $variant ) { $bidtable = $this->mBidtable; $unidtable = $this->mUnidtable; - if( count($bidtable) + count($unidtable) == 0 ){ + if ( count( $bidtable ) + count( $unidtable ) == 0 ) { return $this->mRules; - } elseif ($doConvert){// the text converted + } else { // display current variant in bidirectional array - $disp = $this->getTextInBidtable($variant); + $disp = $this->getTextInBidtable( $variant ); // or display current variant in fallbacks - if(!$disp) + if ( !$disp ) { $disp = $this->getTextInBidtable( - $this->mConverter->getVariantFallbacks($variant)); + $this->mConverter->getVariantFallbacks( $variant ) ); + } // or display current variant in unidirectional array - if(!$disp && array_key_exists($variant,$unidtable)){ - $disp = array_values($unidtable[$variant]); + if ( !$disp && array_key_exists( $variant, $unidtable ) ) { + $disp = array_values( $unidtable[$variant] ); $disp = $disp[0]; } // or display frist text under disable manual convert - if(!$disp && $this->mConverter->mManualLevel[$variant]=='disable') { - if(count($bidtable)>0){ - $disp = array_values($bidtable); + if ( !$disp + && $this->mConverter->mManualLevel[$variant] == 'disable' ) { + if ( count( $bidtable ) > 0 ) { + $disp = array_values( $bidtable ); $disp = $disp[0]; } else { - $disp = array_values($unidtable); - $disp = array_values($disp[0]); + $disp = array_values( $unidtable ); + $disp = array_values( $disp[0] ); $disp = $disp[0]; } } return $disp; - } else {// no convert - return $this->mRules; } } /** - * generate conversion table for all text + * Generate conversion table for all text. * @private */ - function generateConvTable(){ - $flags = $this->mFlags; + function generateConvTable() { + // Special case optimisation + if ( !$this->mBidtable && !$this->mUnidtable ) { + $this->mConvTable = array(); + return; + } + $bidtable = $this->mBidtable; $unidtable = $this->mUnidtable; $manLevel = $this->mConverter->mManualLevel; - $vmarked=array(); - foreach($this->mConverter->mVariants as $v) { + $vmarked = array(); + foreach ( $this->mConverter->mVariants as $v ) { /* for bidirectional array fill in the missing variants, if any, with fallbacks */ - if(!array_key_exists($v, $bidtable)) { - $variantFallbacks = $this->mConverter->getVariantFallbacks($v); - $vf = $this->getTextInBidtable($variantFallbacks); - if($vf) $bidtable[$v] = $vf; + if ( !isset( $bidtable[$v] ) ) { + $variantFallbacks = + $this->mConverter->getVariantFallbacks( $v ); + $vf = $this->getTextInBidtable( $variantFallbacks ); + if ( $vf ) { + $bidtable[$v] = $vf; + } } - if(array_key_exists($v,$bidtable)){ - foreach($vmarked as $vo){ - // use syntax: -{A|zh:WordZh;zh-tw:WordTw}- - // or -{H|zh:WordZh;zh-tw:WordTw}- or -{-|zh:WordZh;zh-tw:WordTw}- + if ( isset( $bidtable[$v] ) ) { + foreach ( $vmarked as $vo ) { + // use syntax: -{A|zh:WordZh;zh-tw:WordTw}- + // or -{H|zh:WordZh;zh-tw:WordTw}- + // or -{-|zh:WordZh;zh-tw:WordTw}- // to introduce a custom mapping between - // words WordZh and WordTw in the whole text - if($manLevel[$v]=='bidirectional'){ - $this->mConvTable[$v][$bidtable[$vo]]=$bidtable[$v]; + // words WordZh and WordTw in the whole text + if ( $manLevel[$v] == 'bidirectional' ) { + $this->mConvTable[$v][$bidtable[$vo]] = $bidtable[$v]; } - if($manLevel[$vo]=='bidirectional'){ - $this->mConvTable[$vo][$bidtable[$v]]=$bidtable[$vo]; + if ( $manLevel[$vo] == 'bidirectional' ) { + $this->mConvTable[$vo][$bidtable[$v]] = $bidtable[$vo]; } } - $vmarked[]=$v; + $vmarked[] = $v; } - /*for unidirectional array - fill to convert tables */ - $allow_unid = $manLevel[$v]=='bidirectional' - || $manLevel[$v]=='unidirectional'; - if($allow_unid && array_key_exists($v,$unidtable)){ - $ct=$this->mConvTable[$v]; - $this->mConvTable[$v] = array_merge($ct,$unidtable[$v]); + /* for unidirectional array fill to convert tables */ + if ( ( $manLevel[$v] == 'bidirectional' || $manLevel[$v] == 'unidirectional' ) + && isset( $unidtable[$v] ) ) + { + if ( isset( $this->mConvTable[$v] ) ) { + $this->mConvTable[$v] = array_merge( $this->mConvTable[$v], $unidtable[$v] ); + } else { + $this->mConvTable[$v] = $unidtable[$v]; + } } } } /** - * Parse rules and flags - * @public + * Parse rules and flags. + * @param $variant String: variant language code */ - function parse($variant){ - if(!$variant) + public function parse( $variant = null ) { + if ( !$variant ) { $variant = $this->mConverter->getPreferredVariant(); + } - $variants = $this->mConverter->mVariants; $this->parseFlags(); $flags = $this->mFlags; // convert to specified variant // syntax: -{zh-hans;zh-hant[;...]|}- - if( count( array_diff( $flags, $variants ) ) == 0 and count( $flags ) != 0 ) { - if ( in_array( $variant, $flags ) ) // check if current variant in flags + if ( $this->mVariantFlags ) { + // check if current variant in flags + if ( isset( $this->mVariantFlags[$variant] ) ) { // then convert to current language - $this->mRules = $this->mConverter->autoConvert( $this->mRules, $variant ); - else { // if current variant no in flags, + $this->mRules = $this->mConverter->autoConvert( $this->mRules, + $variant ); + } else { // if current variant no in flags, // then we check its fallback variants. - $variantFallbacks = $this->mConverter->getVariantFallbacks($variant); - foreach ( $variantFallbacks as $variantFallback ) { - // if current variant's fallback exist in flags - if ( in_array( $variantFallback, $flags ) ) { - // then convert to fallback language - $this->mRules = $this->mConverter->autoConvert( $this->mRules, $variantFallback ); - break; + $variantFallbacks = + $this->mConverter->getVariantFallbacks( $variant ); + if( is_array( $variantFallbacks ) ) { + foreach ( $variantFallbacks as $variantFallback ) { + // if current variant's fallback exist in flags + if ( isset( $this->mVariantFlags[$variantFallback] ) ) { + // then convert to fallback language + $this->mRules = + $this->mConverter->autoConvert( $this->mRules, + $variantFallback ); + break; + } } } } - $this->mFlags = $flags = array('R'); + $this->mFlags = $flags = array( 'R' => true ); } - if( !in_array( 'R', $flags ) || !in_array( 'N', $flags ) ) { - // decode => HTML entities modified by Sanitizer::removeHTMLtags - $this->mRules = str_replace('=>','=>',$this->mRules); - + if ( !isset( $flags['R'] ) && !isset( $flags['N'] ) ) { + // decode => HTML entities modified by Sanitizer::removeHTMLtags + $this->mRules = str_replace( '=>', '=>', $this->mRules ); $this->parseRules(); } $rules = $this->mRules; - if( count( $this->mBidtable ) == 0 && count( $this->mUnidtable ) == 0 ){ - if(in_array('+',$flags) || in_array('-',$flags)) + if ( !$this->mBidtable && !$this->mUnidtable ) { + if ( isset( $flags['+'] ) || isset( $flags['-'] ) ) { // fill all variants if text in -{A/H/-|text} without rules - foreach($this->mConverter->mVariants as $v) + foreach ( $this->mConverter->mVariants as $v ) { $this->mBidtable[$v] = $rules; - elseif (!in_array('N',$flags) && !in_array('T',$flags) ) - $this->mFlags = $flags = array('R'); - } - - if( in_array('R',$flags) ) { - // if we don't do content convert, still strip the -{}- tags - $this->mRuleDisplay = $rules; - } elseif ( in_array('N',$flags) ){ - // proces N flag: output current variant name - $this->mRuleDisplay = $this->mConverter->mVariantNames[trim($rules)]; - } elseif ( in_array('D',$flags) ){ - // proces D flag: output rules description - $this->mRuleDisplay = $this->getRulesDesc(); - } elseif ( in_array('H',$flags) || in_array('-',$flags) ) { - // proces H,- flag or T only: output nothing - $this->mRuleDisplay = ''; - } elseif ( in_array('S',$flags) ){ - $this->mRuleDisplay = $this->getRuleConvertedStr($variant, - $this->mConverter->mDoContentConvert); - } else { - $this->mRuleDisplay= $this->mManualCodeError; + } + } elseif ( !isset( $flags['N'] ) && !isset( $flags['T'] ) ) { + $this->mFlags = $flags = array( 'R' => true ); + } } - // proces T flag - if ( in_array('T',$flags) ) { - $this->mRuleTitle = $this->getRuleConvertedStr($variant, - $this->mConverter->mDoTitleConvert); + + $this->mRuleDisplay = false; + foreach ( $flags as $flag => $unused ) { + switch ( $flag ) { + case 'R': + // if we don't do content convert, still strip the -{}- tags + $this->mRuleDisplay = $rules; + break; + case 'N': + // process N flag: output current variant name + $ruleVar = trim( $rules ); + if ( isset( $this->mConverter->mVariantNames[$ruleVar] ) ) { + $this->mRuleDisplay = $this->mConverter->mVariantNames[$ruleVar]; + } else { + $this->mRuleDisplay = ''; + } + break; + case 'D': + // process D flag: output rules description + $this->mRuleDisplay = $this->getRulesDesc(); + break; + case 'H': + // process H,- flag or T only: output nothing + $this->mRuleDisplay = ''; + break; + case '-': + $this->mRulesAction = 'remove'; + $this->mRuleDisplay = ''; + break; + case '+': + $this->mRulesAction = 'add'; + $this->mRuleDisplay = ''; + break; + case 'S': + $this->mRuleDisplay = $this->getRuleConvertedStr( $variant ); + break; + case 'T': + $this->mRuleTitle = $this->getRuleConvertedStr( $variant ); + $this->mRuleDisplay = ''; + break; + default: + // ignore unknown flags (but see error case below) + } + } + if ( $this->mRuleDisplay === false ) { + $this->mRuleDisplay = $this->mManualCodeError; } - if (in_array('-', $flags)) - $this->mRulesAction='remove'; - if (in_array('+', $flags)) - $this->mRulesAction='add'; - $this->generateConvTable(); } - + /** - * @public + * @todo FIXME: code this function :) */ - function hasRules(){ + public function hasRules() { // TODO: } /** - * get display text on markup -{...}- - * @public + * Get display text on markup -{...}- + * @return string */ - function getDisplay(){ + public function getDisplay() { return $this->mRuleDisplay; } + /** - * get converted title - * @public + * Get converted title. + * @return string */ - function getTitle(){ + public function getTitle() { return $this->mRuleTitle; } /** - * return how deal with conversion rules - * @public + * Return how deal with conversion rules. + * @return string */ - function getRulesAction(){ + public function getRulesAction() { return $this->mRulesAction; } /** - * get conversion table ( bidirectional and unidirectional conversion table ) - * @public + * Get conversion table. (bidirectional and unidirectional + * conversion table) + * @return array */ - function getConvTable(){ + public function getConvTable() { return $this->mConvTable; } /** - * get conversion rules string - * @public + * Get conversion rules string. + * @return string */ - function getRules(){ + public function getRules() { return $this->mRules; } /** - * get conversion flags - * @public + * Get conversion flags. + * @return array */ - function getFlags(){ + public function getFlags() { return $this->mFlags; } }