2 @section @file{float.h}
4 POSIX specification:@* @url{http://www.opengroup.org/onlinepubs/9699919799/basedefs/float.h.html}
8 Portability problems fixed by Gnulib:
11 The conversion from @code{int} to @code{long double} in incorrect on some
13 glibc 2.7 on Linux/SPARC64.
15 The values of @code{LDBL_*} macros are incorrect on some platforms:
16 On OpenBSD 4.0, MirBSD 10, and BeOS, they are the same as the values of the
17 @code{DBL_*} macros, although @samp{long double} is a larger type than
19 On FreeBSD/x86 6.4, they represent the incorrect 53-bit precision assumptions
20 in the compiler, not the real 64-bit precision at runtime.
21 On Linux/PowerPC with GCC 4.4, on AIX 7.1 with GCC 4.2, and on IRIX 6.5,
22 they don't reflect the ``double double'' representation of @code{long double}
26 Portability problems not fixed by Gnulib:
29 The macro @code{FLT_ROUNDS} is a constant expression and does not represent
30 the current rounding mode on some platforms:
31 glibc 2.11, HP-UX 11, mingw.