Name: tb29552 Date: 06/28/99


/*

If this code sample is compiled using:
    jvc(Microsoft compiler) or
    bcj(JBuilder compiler) or
    javac from jdk1.0.2
    javac from jdk1.1.x
Then the output of running the .class file (with any runtime
version) will be:

    s='949500000000000'

However, if this code is compiled using javac (including the
Early Access javac-ea compiler) from 1.2.x, the output will be
something like:

    s='9495[C@fe655feb0'

It appears that the .class from the 1.1.x compiler is including
the contents of the char [] in the string concatenation, whereas
the 1.2.x compiler is using the object identifier instead.

I don't see this change in behavior documented on the
compatability page at:
   http://java.sun.com/products/jdk/1.2/compatibility.html
but perhaps I'm missing something.

Related reading:
  "String Concatenation Operator +"
    http://java.sun.com/docs/books/jls/html/15.doc.html#39990

  "public String toString()"
    http://java.sun.com/docs/books/jls/html/javalang.doc1.html#13783

  "An Array of Characters is Not a String"
    http://java.sun.com/docs/books/jls/html/10.doc.html#25726

*/
class stringTest {
    public static void main(String[] args) {
        char[] buf = new char[11];
        for (int i = 0; i < buf.length; i++) {
            buf[i] = '0';
        }
        String s = "9.495E11";
        s = s.substring(0, 1) + s.substring(2, 5) + buf + "0";
        System.out.println("s='" + s + "'");
        // System.out.println(buf);
    }
}

(Review ID: 84861)
======================================================================