How to query FLT_MAX from Ruby code?

ruby-talk
1

I hope I’m not setting myself up for a Doh! moment…

Is there some way to determine (from Ruby code) the minimum and maximum
limits for Ruby Floats (i.e. along the lines of the C constants FLT_MIN,
FLT_MAX, DBL_MIN and DBL_MAX from the <limits.h> header file)?

2

I hope I’m not setting myself up for a Doh! moment…

Is there some way to determine (from Ruby code) the minimum and maximum
limits for Ruby Floats (i.e. along the lines of the C constants FLT_MIN,
FLT_MAX, DBL_MIN and DBL_MAX from the <limits.h> header file)?

Beats me. But if there’s not, maybe there should be.
Time for an RCR? Something like Float::FLT_MAX etc.?

Hal

···

----- Original Message -----
From: “Lyle Johnson” lyle@users.sourceforge.net
Newsgroups: comp.lang.ruby
To: “ruby-talk ML” ruby-talk@ruby-lang.org
Sent: Friday, December 20, 2002 7:17 PM
Subject: How to query FLT_MAX from Ruby code?

3

Hal E. Fulton wrote:

Lyle Johnson wrote:

Is there some way to determine (from Ruby code) the minimum and maximum
limits for Ruby Floats (i.e. along the lines of the C constants FLT_MIN,
FLT_MAX, DBL_MIN and DBL_MAX from the <limits.h> header file)?

Beats me. But if there’s not, maybe there should be.
Time for an RCR? Something like Float::FLT_MAX etc.?

Dave sat with me a few years ago and helped me enter an RCR for this. I haven’t
checked it’s status in a year or so. For the time being we apply our own patch
whenever we install ruby. (Included after signature.)

···

–
Bil Kleb
NASA Langley Research Center
Hampton, Virginia, USA

*** numeric.c 2002/03/12 05:29:13 1.1
— numeric.c 2002/03/18 12:55:24

*** 13,18 ****
— 13,21 ----
#include “ruby.h”
#include <math.h>
#include <stdio.h>

  • #ifdef HAVE_FLOAT_H
  • #include <float.h>
  • #endif
    #if defined(FreeBSD) && FreeBSD < 4
    #include <floatingpoint.h>
    #endif

*** 1646,1651 ****
— 1649,1668 ----
rb_define_singleton_method(rb_cFloat, “induced_from”, rb_flo_induced_from, 1);
rb_include_module(rb_cFloat, rb_mPrecision);

  • #ifdef HAVE_FLOAT_H
  • rb_define_const(rb_cFloat, "ROUNDS", INT2FIX(FLT_ROUNDS));

  • rb_define_const(rb_cFloat, "RADIX", INT2FIX(FLT_RADIX));

  • rb_define_const(rb_cFloat, "MANT_DIG", INT2FIX(DBL_MANT_DIG));

  • rb_define_const(rb_cFloat, "DIG", INT2FIX(DBL_DIG));

  • rb_define_const(rb_cFloat, "MIN_EXP", INT2FIX(DBL_MIN_EXP));

  • rb_define_const(rb_cFloat, "MAX_EXP", INT2FIX(DBL_MAX_EXP));

  • rb_define_const(rb_cFloat, "MIN_10_EXP", INT2FIX(DBL_MIN_10_EXP));

  • rb_define_const(rb_cFloat, "MAX_10_EXP", INT2FIX(DBL_MAX_10_EXP));

  • rb_define_const(rb_cFloat, "MIN", rb_float_new(DBL_MIN));

  • rb_define_const(rb_cFloat, "MAX", rb_float_new(DBL_MAX));

  • rb_define_const(rb_cFloat, "EPSILON", rb_float_new(DBL_EPSILON));
  • #endif
  •  rb_define_method(rb_cFloat, "to_s", flo_to_s, 0);
 rb_define_method(rb_cFloat, "coerce", flo_coerce, 1);
 rb_define_method(rb_cFloat, "-@", flo_uminus, 0);

*** configure.in 2002/03/18 12:52:21 1.1
— configure.in 2002/03/18 12:52:26

*** 256,262 ****
AC_HEADER_DIRENT
AC_HEADER_STDC
AC_HEADER_SYS_WAIT
! AC_CHECK_HEADERS(stdlib.h string.h unistd.h limits.h sys/file.h sys/ioctl.h
fcntl.h sys/fcntl.h sys/select.h sys/time.h sys/times.h sys/param.h
syscall.h pwd.h a.out.h utime.h memory.h direct.h sys/resource.h)

— 256,262 ----
AC_HEADER_DIRENT
AC_HEADER_STDC
AC_HEADER_SYS_WAIT
! AC_CHECK_HEADERS(stdlib.h string.h unistd.h limits.h float.h sys/file.h sys/ioctl.h
fcntl.h sys/fcntl.h sys/select.h sys/time.h sys/times.h sys/param.h
syscall.h pwd.h a.out.h utime.h memory.h direct.h sys/resource.h)

4

Bil Kleb wrote:

Dave sat with me a few years ago and helped me enter an RCR for this.

OK, found it here:

http://www.rubygarden.org/article.php?sid=103

Date is September 21 (not sure of which year).

I haven’t checked its status in a year or so.

Looks like it’s stuck in RCR limbo, neither accepted nor rejected. I
just added my vote :wink: