Bug 64173
| Summary: | C99 requires some <math.h> macros to be constant expressions | ||
|---|---|---|---|
| Product: | Base System | Reporter: | Stefan Farfeleder <stefan> |
| Component: | standards | Assignee: | Stefan Farfeleder <stefanf> |
| Status: | Closed FIXED | ||
| Severity: | Affects Only Me | CC: | stefan |
| Priority: | Normal | ||
| Version: | 5.2-CURRENT | ||
| Hardware: | Any | ||
| OS: | Any | ||
Responsible Changed From-To: freebsd-standards->stefanf Over to me. I don't think there's a strictly conforming way to produce constant expressions
with the values infinity and NaN but since GCC has builtins exactly for that
reason, we might as well use them:
Index: src/lib/msun/src/math.h
===================================================================
RCS file: /usr/home/ncvs/src/lib/msun/src/math.h,v
retrieving revision 1.31
diff -u -r1.31 math.h
--- src/lib/msun/src/math.h 25 Apr 2004 02:35:42 -0000 1.31
+++ src/lib/msun/src/math.h 7 May 2004 10:01:06 -0000
@@ -32,7 +32,11 @@
float __uf;
} __nan;
+#ifdef __GNUC__
+#define HUGE_VAL __builtin_huge_val()
+#else
#define HUGE_VAL (__infinity.__ud)
+#endif
#if __ISO_C_VISIBLE >= 1999
#define FP_ILOGB0 (-0x7fffffff - 1) /* INT_MIN */
@@ -40,7 +44,11 @@
#define HUGE_VALF (float)HUGE_VAL
#define HUGE_VALL (long double)HUGE_VAL
#define INFINITY HUGE_VALF
+#ifdef __GNUC__
+#define NAN __builtin_nanf("")
+#else
#define NAN (__nan.__uf)
+#endif
#define MATH_ERRNO 1
#define MATH_ERREXCEPT 2
State Changed From-To: open->closed Fixed for GCC in math.h's revision 1.39. |
Quoting from ISO/IEC 9899:1999: # 7.12 Mathematics <math.h> # 3 The macro # HUGE_VAL # expands to a positive double constant expression, not necessarily representable as a # float. The macros # HUGE_VALF # HUGE_VALL # are respectively float and long double analogs of HUGE_VAL.191) # # 4 The macro # INFINITY # expands to a constant expression of type float representing positive or unsigned # infinity, if available; else to a positive constant of type float that overflows at # translation time.192) # # 5 The macro # NAN # is defined if and only if the implementation supports quiet NaNs for the float type. It # expands to a constant expression of type float representing a quiet NaN. These macros expand to the variables __infinity and __nan currently (defined in src/lib/${ARCH}/gen/infinity.c), thus violating the requirement of being constant expressions. Fix: None yet. How-To-Repeat: This strictly conformant C99 program fails to compile: $ cat test.c #include <math.h> double d = HUGE_VAL; int main(void) {} $ c99 test.c test.c:2: error: initializer element is not constant