Name: tb29552			Date: 09/15/97


java.io.StreamTokenizer can't read valid floating point
numbers such as:

    -6.77e-005

Strings of this type can be read are valid in
Java.   Double.valueOf("-6.77e-005") give the
value I expect when it is printed out.  StreamTokenizer
doesn't handle this.

The following java program illustrates this problem:

-------

import java.io.*;
 
public class test2 {
 
        public static void main(String argv[]) {
                try {
                Reader r = new BufferedReader(new FileReader("data"));
                StreamTokenizer st = new StreamTokenizer(r);
 
                st.parseNumbers();
 
                st.nextToken();
                while (st.ttype != StreamTokenizer.TT_EOF) {
                        if (st.ttype == StreamTokenizer.TT_NUMBER)
                                System.out.println(st.nval);
                        else if (st.ttype == StreamTokenizer.TT_WORD) {
                                System.out.println("error!  got word!");
                                System.out.println("saw "+st.sval+" on line "+st
.lineno());
                        } else
                                System.out.println("error!");
                        st.nextToken();
                }
                } catch (Exception e) {
                        System.out.println(e);
                }
 
        }
}

-------
The following data file can be used with the program
above:
-------cut here - only include the numbers not this line
1
2.2
3.33
0.44
-6.77e-005
-------cut here - only include the numbers not this line


company - National Center for Supercomputing Applications , email - ###@###.###
======================================================================

Name: skT45625			Date: 03/08/2000


java version "1.2.2"
Classic VM (build JDK-1.2.2-W, native threads, symcjit)


StreamTokenizer currently extracts numbers expressed in scientific notation as a
number plus a word.  For example, the string "1.23E4" is parsed as number 1.23
and word "E4". This request is to add a method (sample name
StreamTokenizer.parseFloats( boolean flag)), that would allow the user to
specify the behavior of StreamTokenizer (similar to the slashStartComments() and
slashSlashComments() methods).  When flag = true, 1.23E4 would be extracted as a
single number, when flag = false 1.23E4 would be extracted as a number plus a
word, as is current.  To maintain compatibility with current releases this
setting should default to false.
(Review ID: 102224)
======================================================================