Path: utzoo!utgpu!jarvis.csri.toronto.edu!clyde.concordia.ca!uunet!bu.edu!snorkelwacker!think!zaphod.mps.ohio-state.edu!swrinde!emory!mephisto!ncar!boulder!rutgers!netnews.upenn.edu!chin
From: chin@sg1.chem.upenn.edu (Chin Wu)
Newsgroups: gnu.g++.bug
Subject: Is it a bug
Message-ID: <CHIN.90Feb17160902@sg1.chem.upenn.edu>
Date: 17 Feb 90 21:09:02 GMT
Sender: news@netnews.upenn.edu
Distribution: gnu
Organization: University of Pennsylvania, Philadelphia, PA
Lines: 27


    I just installed G++ on our IRIS 4D using gcc-1.37, g++-1.36.4 and
libg++-1.36.3. Since I have to do a few modifications during my
installion. I don't know if this is a bug or because of some errors I
made when I install the compiler. Following is the program and its
output. It behaves erraneous when deal with "float", but seems all
right using "double". Although I can work around using the "form()",
I, however, like to see the second way to work.

PROGRAM:
#include <stream.h>
main()
{
    float f = 0.2;
    double d = 0.4;
    cout << form("%f\n%f\n", d, f);
    cout << d << "\n" << f << "\n";
}

OUTPUT:
0.000000
0.200000
0.4
1.3411e-08  <-- ERROR


chin@sg1.chem.upenn.edu