Name: laC46010 Date: 10/19/97



Methods NumberFormat.getCurrencyInstance creates wrong format for ja_JP
(Japanese-Japan) localization.
This format returns the pattern "¥#,##0" instead of "\u5186#,##0".
                                 ^----\u00a5

As result JDK1.2 fails to pass the following JCK1.2 test case:

api/java_util/localeData/coreLocale.html#ja_JP

The contradiction comes from the source where
"yen" character is used in "currency pattern" but
"\u5186" is used as a "local currency symbol".
In all others locales identical characters are used
in these places.

See src/share/classes/java/text/resources/LocaleElements_ja.java

        "¥#,##0", /*currency pattern*/
...
        ",", /*group (thousands) separator*/
...
        "0", /*native 0 digit*/
        "#", /*pattern digit*/
        "-", /*minus sign*/
        "E", /*exponential*/
        "\u5186", /*local currency symbol*/
...
        ".", /*monetary decimal separator*/

 
An example below produces the following output:
--------------output----------------
failed: return: ¥#,##0 instead \u5186#,##0

--------------JA_JP_fails.java--------------------
import java.util.Locale;
import java.text.NumberFormat;
import java.text.DecimalFormat;

class JA_JP_fails {
    public static void main(String argv[]) {
Locale ja = new Locale("ja", "JP");
NumberFormat nf = NumberFormat.getCurrencyInstance(ja);
DecimalFormat df = (DecimalFormat)nf;

if ( df.toLocalizedPattern().equals("\u5186#,##0")) {
System.out.println("OK");
} else {
System.out.println("failed: return: "+
df.toLocalizedPattern() + " instead " +
"\\u5186#,##0");
}
    }
}

---------------------------------------------------------

======================================================================