One of our developers here is experimenting with Java SDK "1.3.0-C" on
WindowsNT has encountered what appears to be an incomplete Java
localization issue. Can you pass the following along to your development
folks to see if they concur that this is a bug?
Here is my observation on the Double.parseDouble(String) method.

I am using the latest jdk1.3 final release. It apparently shows that Sun's
implementation of Double does not accept a String which is in the France locale.

The source is attached.


richard.baxter@east 2000-06-30
recent customer feed back is:

 From the java2doc that came with Java SDK "1.3.0-C" on WindowsNT ,
Double.parseDouble makes use of the 'valueOf' method to perform its
function. It seems unlikely that, 'valueOf' is going to take into account
localization issues and know to replace a US locale separator '.' with a
French local separator of ','.

Further, when we look at Java Source code for Double.parseDouble(), it uses
'FloatingDecimal.readJavaFormatString(s).doubleValue();' The implementation
for readJavaFormatString is hard-coded to look for a decimal value as a
separator (see line 1033).

Finally, when we do use the workaround, localization fails. Here is sample
code and the result....
import java.util.*;
import java.text.*;
public class Test {
public static void main(String[] args)
{
String doubleStr = "12,34";
// Get the default locale
Locale l = Locale.getDefault();
try {
double d = Double.valueOf(doubleStr).doubleValue();
} catch (NumberFormatException e) {
e.printStackTrace();
System.out.println("The number parsed is: " + doubleStr);
System.out.println("Default locale is: " + l);
}
}
}
Here is the result:
U:\>java Test
java.lang.NumberFormatException: 12,34
at java.lang.FloatingDecimal.readJavaFormatString(Unknown Source)
at java.lang.Double.valueOf(Unknown Source)
at Test.main(Test.java:11)
The number parsed is: 12,34
Default locale is: fr_FR