What's the difference of character encoding between 1.4.0and1.4.2 in Linux

Question

As i find, the character encoding about chinese in jdk1.4.2 no langer the same of jdk1.4.0.
In jdk1.4.0, the character encoding used the "file.encoding" system property, we often set the
property with "gb2312".
But in jdk1.4.2, i find that the default character encoding no longer used the "file.encoding" system property.
Who knows the reason?
Test Program:
public class B{
public static void main(String args[]) throws Exception{
byte [] bytes = new byte[]{(byte)0xD6,(byte)0xD0,(byte)0xCE,(byte)0xC4};
String s1 = new String(bytes);
String s2 = new String(bytes,System.getProperty("file.encoding"));
System.out.println("s1="+s1+" , s2="+s2);
System.out.println("s1.length=" + s1.length() + " , s2.length="+s2.length());
run four times and the result list:
[root@app15 component]# /usr/local/j2sdk1.4.0/bin/java -Dfile.encoding=ISO-8859-1 -cp . B
s1=中文 , s2=中文
s1.length=4 , s2.length=4
[root@app15 component]# /usr/local/j2sdk1.4.0/bin/java -Dfile.encoding=gb2312 -cp . B
s1=中文 , s2=中文
s1.length=2 , s2.length=2
[root@app15 component]# /usr/local/j2sdk1.4.2/bin/java -Dfile.encoding=ISO-8859-1 -cp . B
s1=中文 , s2=中文
s1.length=4 , s2.length=4
[root@app15 component]# /usr/local/j2sdk1.4.2/bin/java -Dfile.encoding=gb2312 -cp . B
s1=中文 , s2=??
s1.length=4 , s2.length=2
[root@app15 component]#

Answer 1

I don't know for sure, but:
-- The API documentation for String says that "new String(byte[])" uses "the platform's default charset".
-- The API documentation for Charset says "The default charset is determined during virtual-machine startup and typically depends upon the locale and charset being used by the underlying operating system."
You'll notice that it doesn't say anything about using the file.encoding system value, so presumably (based on your experiments) it doesn't. I did a search for "java default charset" and didn't find anything specific, but this site says "As of Java 1.4.1, the default Charset varies from platform to platform" and suggests you explicitly hard-code your charset. I would agree with that.