Merge "TimeAdjustTest: Use dataProvider"

[lhc/web/wiklou.git] / includes / Collation.php

diff --git a/includes/Collation.php b/includes/Collation.php
index a77e017..3f0fa7b 100644 (file)
--- a/includes/Collation.php
+++ b/includes/Collation.php
@@ -149,6 +149,8 @@ class IdentityCollation extends Collation {
 }
 
 class IcuCollation extends Collation {
+       const FIRST_LETTER_VERSION = 1;
+
        var $primaryCollator, $mainCollator, $locale;
        var $firstLetterData;
 
@@ -181,7 +183,10 @@ class IcuCollation extends Collation {
 
        /**
         * Additional characters (or character groups) to be considered separate
-        * letters for given languages, compared to the data stored in the
+        * letters for given languages, or to be removed from the list of such
+        * letters (denoted by keys starting with '-').
+        *
+        * These are additions to (or subtractions from) the data stored in the
         * first-letters-root.ser file (which among others includes full basic latin,
         * cyrillic and greek alphabets).
         *
@@ -201,9 +206,18 @@ class IcuCollation extends Collation {
                // Verified by native speakers
                'be' => array( "Ё" ),
                'be-tarask' => array( "Ё" ),
+               'en' => array(),
                'fi' => array( "Å", "Ä", "Ö" ),
+               '-fi' => array( "Ǥ", "Ŋ", "Ŧ", "Ʒ" ), // sorted like G, N, T, Z - bug 46330
+               'hu' => array( "Cs", "Dz", "Dzs", "Gy", "Ly", "Ny", "Ö", "Sz", "Ty", "Ü", "Zs" ),
+               'it' => array(),
                'pl' => array( "Ą", "Ć", "Ę", "Ł", "Ń", "Ó", "Ś", "Ź", "Ż" ),
+               'pt' => array(),
                'ru' => array(),
+               'sv' => array( "Å", "Ä", "Ö" ),
+               '-sv' => array( "Þ" ), // sorted as "th" in Swedish, causing unexpected output - bug 45446
+               'uk' => array( "Ґ", "Ь" ),
+               'vi' => array( "Ă", "Â", "Đ", "Ê", "Ô", "Ơ", "Ư" ),
                // Not verified, but likely correct
                'af' => array(),
                'ast' => array( "Ch", "Ll", "Ñ" ),
@@ -219,7 +233,6 @@ class IcuCollation extends Collation {
                'de' => array(),
                'dsb' => array( "Č", "Ć", "Dź", "Ě", "Ch", "Ł", "Ń", "Ŕ", "Š", "Ś", "Ž", "Ź" ),
                'el' => array(),
-               'en' => array(),
                'eo' => array( "Ĉ", "Ĝ", "Ĥ", "Ĵ", "Ŝ", "Ŭ" ),
                'es' => array( "Ñ" ),
                'et' => array( "Š", "Ž", "Õ", "Ä", "Ö", "Ü" ),
@@ -233,9 +246,7 @@ class IcuCollation extends Collation {
                'gl' => array( "Ch", "Ll", "Ñ" ),
                'hr' => array( "Č", "Ć", "Dž", "Đ", "Lj", "Nj", "Š", "Ž" ),
                'hsb' => array( "Č", "Dź", "Ě", "Ch", "Ł", "Ń", "Ř", "Š", "Ć", "Ž" ),
-               'hu' => array( "Cs", "Dz", "Dzs", "Gy", "Ly", "Ny", "Ö", "Sz", "Ty", "Ü", "Zs" ),
                'is' => array( "Á", "Ð", "É", "Í", "Ó", "Ú", "Ý", "Þ", "Æ", "Ö", "Å" ),
-               'it' => array(),
                'kk' => array( "Ү", "І" ),
                'kl' => array( "Æ", "Ø", "Å" ),
                'ku' => array( "Ç", "Ê", "Î", "Ş", "Û" ),
@@ -250,7 +261,6 @@ class IcuCollation extends Collation {
                'nl' => array(),
                'no' => array( "Æ", "Ø", "Å" ),
                'oc' => array(),
-               'pt' => array(),
                'rm' => array(),
                'ro' => array( "Ă", "Â", "Î", "Ş", "Ţ" ),
                'rup' => array( "Ă", "Â", "Î", "Ľ", "Ń", "Ş", "Ţ" ),
@@ -260,14 +270,11 @@ class IcuCollation extends Collation {
                'smn' => array( "Á", "Č", "Đ", "Ŋ", "Š", "Ŧ", "Ž", "Æ", "Ø", "Å", "Ä", "Ö" ),
                'sq' => array( "Ç", "Dh", "Ë", "Gj", "Ll", "Nj", "Rr", "Sh", "Th", "Xh", "Zh" ),
                'sr' => array(),
-               'sv' => array( "Å", "Ä", "Ö" ),
                'tk' => array( "Ç", "Ä", "Ž", "Ň", "Ö", "Ş", "Ü", "Ý" ),
                'tl' => array( "Ñ", "Ng" ),
                'tr' => array( "Ç", "Ğ", "İ", "Ö", "Ş", "Ü" ),
                'tt' => array( "Ә", "Ө", "Ү", "Җ", "Ң", "Һ" ),
-               'uk' => array( "Ґ", "Ь" ),
                'uz' => array( "Ch", "G'", "Ng", "O'", "Sh" ),
-               'vi' => array( "Ă", "Â", "Đ", "Ê", "Ô", "Ơ", "Ư" ),
        );
 
        const RECORD_LENGTH = 14;
@@ -343,7 +350,9 @@ class IcuCollation extends Collation {
                $cacheKey = wfMemcKey( 'first-letters', $this->locale );
                $cacheEntry = $cache->get( $cacheKey );
 
-               if ( $cacheEntry ) {
+               if ( $cacheEntry && isset( $cacheEntry['version'] )
+                       && $cacheEntry['version'] == self::FIRST_LETTER_VERSION ) 
+               {
                        $this->firstLetterData = $cacheEntry;
                        return $this->firstLetterData;
                }
@@ -352,7 +361,12 @@ class IcuCollation extends Collation {
 
                if ( isset ( self::$tailoringFirstLetters[$this->locale] ) ) {
                        $letters = wfGetPrecompiledData( "first-letters-root.ser" );
+                       // Append additional characters
                        $letters = array_merge( $letters, self::$tailoringFirstLetters[$this->locale] );
+                       // Remove unnecessary ones, if any
+                       if ( isset( self::$tailoringFirstLetters[ '-' . $this->locale ] ) ) {
+                               $letters = array_diff( $letters, self::$tailoringFirstLetters[ '-' . $this->locale ] );
+                       }
                } else {
                        $letters = wfGetPrecompiledData( "first-letters-{$this->locale}.ser" );
                        if ( $letters === false ) {
@@ -385,7 +399,8 @@ class IcuCollation extends Collation {
                ksort( $letterMap, SORT_STRING );
                $data = array(
                        'chars' => array_values( $letterMap ),
-                       'keys' => array_keys( $letterMap )
+                       'keys' => array_keys( $letterMap ),
+                       'version' => self::FIRST_LETTER_VERSION,
                );
 
                // Reduce memory usage before caching
@@ -422,13 +437,13 @@ class IcuCollation extends Collation {
         * Do a binary search, and return the index of the largest item that sorts
         * less than or equal to the target value.
         *
-        * @param $valueCallback array A function to call to get the value with
+        * @param array $valueCallback A function to call to get the value with
         *     a given array index.
-        * @param $valueCount int The number of items accessible via $valueCallback,
+        * @param int $valueCount The number of items accessible via $valueCallback,
         *     indexed from 0 to $valueCount - 1
-        * @param $comparisonCallback array A callback to compare two values, returning
+        * @param array $comparisonCallback A callback to compare two values, returning
         *     -1, 0 or 1 in the style of strcmp().
-        * @param $target string The target value to find.
+        * @param string $target The target value to find.
         *
         * @return int|bool The item index of the lower bound, or false if the target value
         *     sorts before all items.