gcj 3.0/unicode

[David Brownell <david-b at pacbell dot net>]

I was comparing the output of an api test suite running on
jdk 1.3 versus gcj 3.0, and came across a few errors.
These are demonstrable with the appended program.

I have no idea whether these are the only cases where
there's a difference; this particular suite doesn't test all
relevant Unicode characters.  

- Dave


public class bar
{
    public static void main (String argv [])
    {
        // GCJ/GIJ 3.0 gives different answers from JDK 1.3
        // for at least these values

        int starts [] = {
            0x0e2f,     // jdk 1.3: false
            0x0eaf      // jdk 1.3: false
        };
        int parts [] = {
            0x3006,     // jdk 1.3: false
            0x309b,     // jdk 1.3: true
            0x309c      // jdk 1.3: true
        };

        for (int i = 0; i < starts.length; i++)
            System.out.println ("Start: " + starts [i] + " ==> "
                    + Character.isUnicodeIdentifierStart ((char) starts [i]));
        for (int i = 0; i < parts.length; i++)
            System.out.println ("Part: " + parts [i] + " ==> "
                    + Character.isUnicodeIdentifierPart ((char) parts [i]));
    }
}

/*
GCJ 3.0 build using RedHat 7.0, i686-pc-linux-gnu
../gcc-3.0/configure --enable-threads=posix --enable-languages=c++,java --enable-checking
 */