GNU bug report logs - #25785
scm_c_make_polar broken on APPLE

Previous Next

Package: guile;

Reported by: Matt Wette <matt.wette <at> gmail.com>

Date: Sat, 18 Feb 2017 17:39:01 UTC

Severity: normal

Done: Andy Wingo <wingo <at> pobox.com>

Bug is archived. No further changes may be made.

Full log


View this message in rfc822 format

From: help-debbugs <at> gnu.org (GNU bug Tracking System)
To: Matt Wette <matt.wette <at> gmail.com>
Subject: bug#25785: closed (Re: bug#25785: scm_c_make_polar broken on APPLE)
Date: Tue, 21 Feb 2017 21:11:02 +0000
[Message part 1 (text/plain, inline)]
Your bug report

#25785: scm_c_make_polar broken on APPLE

which was filed against the guile package, has been closed.

The explanation is attached below, along with your original report.
If you require more details, please reply to 25785 <at> debbugs.gnu.org.

-- 
25785: http://debbugs.gnu.org/cgi/bugreport.cgi?bug=25785
GNU Bug Tracking System
Contact help-debbugs <at> gnu.org with problems
[Message part 2 (message/rfc822, inline)]
From: Andy Wingo <wingo <at> pobox.com>
To: 25785-done <at> debbugs.gnu.org
Subject: Re: bug#25785: scm_c_make_polar broken on APPLE
Date: Tue, 21 Feb 2017 22:10:40 +0100
Thanks!

[Message part 3 (message/rfc822, inline)]
From: Matt Wette <matt.wette <at> gmail.com>
To: bug-guile <at> gnu.org
Subject: scm_c_make_polar broken on APPLE
Date: Sat, 18 Feb 2017 09:37:54 -0800
[Message part 4 (text/plain, inline)]
numbers.c:scm_c_make_polar in guile-2.1.7 breaks on APPLE (macOS 10.12).

Reason: APPLE does not have sincos(), and gcc will optimize sin/cos to use cexp(), and cexp() does not preserve sign of zero.

Patch includes (and attached) for guile-2.1.7.  This uses macOS __sincos().

Example:
In scm_c_make_polar, “gcc -O2”  turns sin(), cos() into cexp(), since cexp(i*x) = cos(x) + i*sin(x):

gcc -O0 =>
LM4339:
	movq	-32(%rbp), %rax
	movd	%rax, %xmm0
	call	_sin
	movd	%xmm0, %rax
	movq	%rax, -8(%rbp)
LM4340:
	movq	-32(%rbp), %rax
	movd	%rax, %xmm0
	call	_cos
	movd	%xmm0, %rax
	movq	%rax, -16(%rbp)

gcc -O2 =>
	pxor	%xmm0, %xmm0
LVL2703:
	call	_cexp

I wrote a little C program to show that cexp() does not preserve the zero-signed-ness:

 cos,sin: +1.000000 -0.000000
__sincos: +1.000000 -0.000000
    cexp: +1.000000 +0.000000

The scm_c_make_polar will use sincos() if available, but macOS does not have sincos(), it has __sincos().

#include <stdio.h>
#include <stdint.h>
#include <math.h>
#include <complex.h>

extern double z, p, n;

int main() {
  double complex c;
  double d, sine, cosine;

  d = z*n;
  printf(" cos,sin: %+f %+f\n", cos(d), sin(d));
  __sincos(d, &sine, &cosine);
  printf("__sincos: %+f %+f\n", cosine, sine);
  c = cexp(CMPLX(0.0, d));
  printf("    cexp: %+f %+f\n", creal(c), cimag(c));
}

double z = 0.0, p = +1.0, n = -1.0;


--- configure.ac.orig	2017-02-18 08:35:06.000000000 -0800
+++ configure.ac	2017-02-18 08:35:26.000000000 -0800
@@ -1152,8 +1152,9 @@
 #   asinh, acosh, atanh, trunc - C99 standard, generally not available on
 #                                older systems
 #   sincos - GLIBC extension
+#   __sincos - APPLE extension
 #
-AC_CHECK_FUNCS(asinh acosh atanh copysign finite sincos trunc)
+AC_CHECK_FUNCS(asinh acosh atanh copysign finite sincos __sincos trunc)
 
 # C99 specifies isinf and isnan as macros.
 # HP-UX provides only macros, no functions.
--- libguile/numbers.c.patch.labs	2017-02-18 08:31:21.000000000 -0800
+++ libguile/numbers.c	2017-02-18 08:34:18.000000000 -0800
@@ -9109,6 +9109,8 @@
      details.  */
 #if (defined HAVE_SINCOS) && (defined __GLIBC__) && (defined _GNU_SOURCE)
   sincos (ang, &s, &c);
+#elif (defined HAVE___SINCOS)
+  __sincos (ang, &s, &c);
 #else
   s = sin (ang);
   c = cos (ang);


[Message part 5 (text/html, inline)]
[sincos.patch (application/octet-stream, attachment)]
[Message part 7 (text/html, inline)]

This bug report was last modified 8 years and 89 days ago.

Previous Next


GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.