Accented characters showing up as ? in JRE1.3 but ok in 1.2

Question

I'm implementing a database web interface product that utilizes JSPs (on SunOS 5.7).
The problem is in the search form. When using accented characters (French language), the JSP calls on URLEncode, but all accented characters show up as '?'.
However, when editing a record, using accented characters is not a problem (i.e., the accented characters are properly stored in the fields).
Back on the server, I ran a small program to output accented characters and also to call java.net.URLEncoder to convert the characters.
The default JDK is J2SE (1.3.1). Compiliing and running the program results in question marks.
Using JDK 1.2, the accented characters show up fine.
It would appear that URLEncoder is not at fault, but instead, JDK 1.3.1 doesn't seem to handle the accented characters.
I figure there must be a setting somewhere, but I'm not sure where.
Here's the program I used (written in Win98, using standard Win-based character set and Unicode format \u00xx; in Unix, "more" displays the Win accented characters fine but "vi" displays them as \xxx; compiles and displays perfectly when using JDK 1.2):
import java.net.*;
class mine {
public static void main(String args[]) {
System.out.println("�����������") ;
System.out.println(URLEncoder.encode("�����������")) ;
System.out.println("\u00e0\u00e2\u00e4");
System.out.println("\351");
System.out.println("\351\347\356\364\373\340\350\342\344\374\357") ;
The output with JDK 1.2 is:
�����������
%E9%E7%EE%F4%FB%E0%E8%E2%E4%FC%EF
���
The output with JDK 1.3.1 is:
%3F%3F%3F%3F%3F%3F%3F%3F%3F%3F%3F

Answer 1

Between jdk1.2 and jdk 1.3 the default encoding of the vm changed.
You can get it by executing:
System.out.println("Default Encoding:" + System.getProperty("file.encoding"));
or
System.out.println("Default Encoding:" + (new java.io.InputStreamReader(System.in)).getEncoding());
The default encoding is used during the conversion of bytes to strings and vice verca.
Assume your default encoding is ISO8859_1. Then calling new String(byte[]) is equivalent to calling
new String(byte[], "ISO8859_1")
Now if you are converting a character from one encoding scheme to another and there is no mapping
for this character in the target scheme. Then the character will be replaced by a default character
which is (quite often) the question mark.
You can set the default encoding for a vm by passing it as a command line parameter
java -Dfile.encoding=ISO8859_1
java -Dfile.encoding=Cp1252