[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [PATCH] strtod: make it more-accurate typically, and don't require l
From: |
Bruno Haible |
Subject: |
Re: [PATCH] strtod: make it more-accurate typically, and don't require libm |
Date: |
Tue, 13 Jul 2010 10:10:14 +0200 |
User-agent: |
KMail/1.9.9 |
Hi Paul,
> but won't the "return 0.0" cause some compilers to
> issue bogus warnings about X and EXPONENT not being used?
> That's why I had the "x + exponent" in there.
We have a macro for this purpose in gnulib (actually, even two): _GL_UNUSED
and _UNUSED_PARAMETER_. Alternatively, a cast to void could also be used.
I committed this:
2010-07-13 Bruno Haible <address@hidden>
strtod: Add safety check.
* lib/strtod.c (ldexp): Abort if this dummy replacement gets called.
--- lib/strtod.c.orig Tue Jul 13 10:05:26 2010
+++ lib/strtod.c Tue Jul 13 10:04:54 2010
@@ -45,7 +45,12 @@
#if !HAVE_LDEXP_IN_LIBC
#define ldexp dummy_ldexp
- static double ldexp (double x, int exponent) { return x + exponent; }
+ /* A dummy definition that will never be invoked. */
+ static double ldexp (double x _GL_UNUSED, int exponent _GL_UNUSED)
+ {
+ abort ();
+ return 0.0;
+ }
#endif
/* Return X * BASE**EXPONENT. Return an extreme value and set errno
- [PATCH] strtod: make it more-accurate typically, and don't require libm, Paul Eggert, 2010/07/12
- Re: [PATCH] strtod: make it more-accurate typically, and don't require libm, Bruno Haible, 2010/07/12
- Re: [PATCH] strtod: make it more-accurate typically, and don't require libm, Bruno Haible, 2010/07/12
- Re: [PATCH] strtod: make it more-accurate typically, and don't require libm, Paul Eggert, 2010/07/12
- Re: [PATCH] strtod: make it more-accurate typically, and don't require libm,
Bruno Haible <=