Mark Article::doEditContent() as deprecated
[lhc/web/wiklou.git] / languages / Language.php
index 3e28759..5bce76b 100644 (file)
@@ -45,7 +45,9 @@ class Language {
        public $dateFormatStrings = [];
        public $mExtendedSpecialPageAliases;
 
-       protected $namespaceNames, $mNamespaceIds, $namespaceAliases;
+       /** @var array|null */
+       protected $namespaceNames;
+       protected $mNamespaceIds, $namespaceAliases;
 
        /**
         * ReplacementArray object caches
@@ -137,6 +139,12 @@ class Language {
         */
        static private $fallbackLanguageCache = [];
 
+       /**
+        * Cache for grammar rules data
+        * @var MapCacheLRU|null
+        */
+       static private $grammarTransformations;
+
        /**
         * Cache for language names
         * @var HashBagOStuff|null
@@ -291,7 +299,7 @@ class Language {
                # Since these are limited, this is safe even later changes to the registry --
                # the only oddity is that it might change the type of the tag, and thus
                # the results from the capturing groups.
-               # http://www.iana.org/assignments/language-subtag-registry
+               # https://www.iana.org/assignments/language-subtag-registry
 
                $grandfathered = "en{$s}GB{$s}oed"
                        . "|i{$s}(?:ami|bnn|default|enochian|hak|klingon|lux|mingo|navajo|pwn|tao|tay|tsu)"
@@ -332,7 +340,7 @@ class Language {
                        // People think language codes are html safe, so enforce it.
                        // Ideally we should only allow a-zA-Z0-9-
                        // but, .+ and other chars are often used for {{int:}} hacks
-                       // see bugs 37564, 37587, 36938
+                       // see bugs T39564, T39587, T38938
                        $cache[$code] =
                                // Protect against path traversal
                                strcspn( $code, ":/\\\000&<>'\"" ) === strlen( $code )
@@ -457,10 +465,11 @@ class Language {
                if ( is_null( $this->namespaceNames ) ) {
                        global $wgMetaNamespace, $wgMetaNamespaceTalk, $wgExtraNamespaces;
 
-                       $this->namespaceNames = self::$dataCache->getItem( $this->mCode, 'namespaceNames' );
                        $validNamespaces = MWNamespace::getCanonicalNamespaces();
 
-                       $this->namespaceNames = $wgExtraNamespaces + $this->namespaceNames + $validNamespaces;
+                       $this->namespaceNames = $wgExtraNamespaces +
+                               self::$dataCache->getItem( $this->mCode, 'namespaceNames' );
+                       $this->namespaceNames += $validNamespaces;
 
                        $this->namespaceNames[NS_PROJECT] = $wgMetaNamespace;
                        if ( $wgMetaNamespaceTalk ) {
@@ -1623,7 +1632,7 @@ class Language {
         *
         * Based on a PHP-Nuke block by Sharjeel which is released under GNU/GPL license
         *
-        * @see http://phpnuke.org/modules.php?name=News&file=article&sid=8234&mode=thread&order=0&thold=0
+        * @see https://phpnuke.org/modules.php?name=News&file=article&sid=8234&mode=thread&order=0&thold=0
         *
         * @param string $ts
         *
@@ -1852,9 +1861,9 @@ class Language {
         * Algorithm to convert Gregorian dates to Thai solar dates,
         * Minguo dates or Minguo dates.
         *
-        * Link: http://en.wikipedia.org/wiki/Thai_solar_calendar
-        *       http://en.wikipedia.org/wiki/Minguo_calendar
-        *       http://en.wikipedia.org/wiki/Japanese_era_name
+        * Link: https://en.wikipedia.org/wiki/Thai_solar_calendar
+        *       https://en.wikipedia.org/wiki/Minguo_calendar
+        *       https://en.wikipedia.org/wiki/Japanese_era_name
         *
         * @param string $ts 14-character timestamp
         * @param string $cName Calender name
@@ -2091,17 +2100,15 @@ class Language {
                $data = explode( '|', $tz, 3 );
 
                if ( $data[0] == 'ZoneInfo' ) {
-                       MediaWiki\suppressWarnings();
-                       $userTZ = timezone_open( $data[2] );
-                       MediaWiki\restoreWarnings();
-                       if ( $userTZ !== false ) {
-                               $date = date_create( $ts, timezone_open( 'UTC' ) );
-                               date_timezone_set( $date, $userTZ );
-                               $date = date_format( $date, 'YmdHis' );
-                               return $date;
+                       try {
+                               $userTZ = new DateTimeZone( $data[2] );
+                               $date = new DateTime( $ts, new DateTimeZone( 'UTC' ) );
+                               $date->setTimezone( $userTZ );
+                               return $date->format( 'YmdHis' );
+                       } catch ( Exception $e ) {
+                               // Unrecognized timezone, default to 'Offset' with the stored offset.
+                               $data[0] = 'Offset';
                        }
-                       # Unrecognized timezone, default to 'Offset' with the stored offset.
-                       $data[0] = 'Offset';
                }
 
                if ( $data[0] == 'System' || $tz == '' ) {
@@ -2571,7 +2578,7 @@ class Language {
 
        /**
         * @param string $key
-        * @return array|null
+        * @return string|null
         */
        public function getMessage( $key ) {
                return self::$dataCache->getSubitem( $this->mCode, 'messages', $key );
@@ -2593,7 +2600,7 @@ class Language {
        public function iconv( $in, $out, $string ) {
                # Even with //IGNORE iconv can whine about illegal characters in
                # *input* string. We just ignore those too.
-               # REF: http://bugs.php.net/bug.php?id=37166
+               # REF: https://bugs.php.net/bug.php?id=37166
                # REF: https://phabricator.wikimedia.org/T18885
                MediaWiki\suppressWarnings();
                $text = iconv( $in, $out . '//IGNORE', $string );
@@ -3728,8 +3735,46 @@ class Language {
                        return $wgGrammarForms[$this->getCode()][$case][$word];
                }
 
+               $grammarTransformations = $this->getGrammarTransformations();
+
+               if ( isset( $grammarTransformations[$case] ) ) {
+                       $forms = $grammarTransformations[$case];
+
+                       // Some names of grammar rules are aliases for other rules.
+                       // In such cases the value is a string rather than object,
+                       // so load the actual rules.
+                       if ( is_string( $forms ) ) {
+                               $forms = $grammarTransformations[$forms];
+                       }
+
+                       foreach ( array_values( $forms ) as $rule ) {
+                               $form = $rule[0];
+
+                               if ( $form === '@metadata' ) {
+                                       continue;
+                               }
+
+                               $replacement = $rule[1];
+
+                               $regex = '/' . addcslashes( $form, '/' ) . '/u';
+                               $patternMatches = preg_match( $regex, $word );
+
+                               if ( $patternMatches === false ) {
+                                       wfLogWarning(
+                                               'An error occurred while processing grammar. ' .
+                                               "Word: '$word'. Regex: /$form/."
+                                       );
+                               } elseif ( $patternMatches === 1 ) {
+                                       $word = preg_replace( $regex, $replacement, $word );
+
+                                       break;
+                               }
+                       }
+               }
+
                return $word;
        }
+
        /**
         * Get the grammar forms for the content language
         * @return array Array of grammar forms
@@ -3745,6 +3790,46 @@ class Language {
 
                return [];
        }
+
+       /**
+        * Get the grammar transformations data for the language.
+        * Used like grammar forms, with {{GRAMMAR}} and cases,
+        * but uses pairs of regexes and replacements instead of code.
+        *
+        * @return array[] Array of grammar transformations.
+        * @throws MWException
+        * @since 1.28
+        */
+       public function getGrammarTransformations() {
+               $languageCode = $this->getCode();
+
+               if ( self::$grammarTransformations === null ) {
+                       self::$grammarTransformations = new MapCacheLRU( 10 );
+               }
+
+               if ( self::$grammarTransformations->has( $languageCode ) ) {
+                       return self::$grammarTransformations->get( $languageCode );
+               }
+
+               $data = [];
+
+               $grammarDataFile = __DIR__ . "/data/grammarTransformations/$languageCode.json";
+               if ( is_readable( $grammarDataFile ) ) {
+                       $data = FormatJson::decode(
+                               file_get_contents( $grammarDataFile ),
+                               true
+                       );
+
+                       if ( $data === null ) {
+                               throw new MWException( "Invalid grammar data for \"$languageCode\"." );
+                       }
+
+                       self::$grammarTransformations->set( $languageCode, $data );
+               }
+
+               return $data;
+       }
+
        /**
         * Provides an alternative text depending on specified gender.
         * Usage {{gender:username|masculine|feminine|unknown}}.
@@ -4461,14 +4546,15 @@ class Language {
        }
 
        /**
-        * @todo Document
+        * Formats a time given in seconds into a string representation of that time.
+        *
         * @param int|float $seconds
-        * @param array $format Optional
-        *   If $format['avoid'] === 'avoidseconds': don't mention seconds if $seconds >= 1 hour.
-        *   If $format['avoid'] === 'avoidminutes': don't mention seconds/minutes if $seconds > 48 hours.
+        * @param array $format An optional argument that formats the returned string in different ways:
+        *   If $format['avoid'] === 'avoidseconds': don't show seconds if $seconds >= 1 hour,
+        *   If $format['avoid'] === 'avoidminutes': don't show seconds/minutes if $seconds > 48 hours,
         *   If $format['noabbrevs'] is true: use 'seconds' and friends instead of 'seconds-abbrev'
         *     and friends.
-        *   For backwards compatibility, $format may also be one of the strings 'avoidseconds'
+        * @note For backwards compatibility, $format may also be one of the strings 'avoidseconds'
         *     or 'avoidminutes'.
         * @return string
         */