From: matz@... (Yukihiro Matsumoto)
Date: 2000-10-04T13:22:14+09:00
Subject: [ruby-talk:5274] Re: Compile Problem with 1.6.1

Hi,

In message "[ruby-talk:5267] Re: Compile Problem with 1.6.1"
    on 00/10/04, Scott Billings <aerogemsX@netins.net> writes:

|You're in luck:
|
|[root@localhost qt-gui]# ruby -e 'p(3.330669074e-16 < 0.001)'
|false

Not the worst, but I still have no idea about what make this false.

If this program gives something like "3.330669074e-16, 0.001: 1",
it's the compiler's fault.

  int
  main() {
    double a = 3.330669074e-16;
    double b = 0.001
    printf("%lf < %lf: %d\n", a, b, a < b);
    return 0;
  }

If not, try:

  # gdb ruby
  (gdb) b rb_flo_lt
  (gdb) r -e 'p(3.330669074e-16 < 0.001)'
  (gdb) p *(struct RFloat*)a
  (gdb) p *(struct RFloat*)b

Thank you.

							matz.