__NO_LONG_DOUBLE_MATH?

[Mark Mitchell]

I'm mystified as to why many platforms defined __NO_LONG_DOUBLE_MATH, 
which prevents the long double prototypes for math functions (e.g., 
acosl) from being declard in <math.h>.

I understand that the motivation was to avoid declaring these functions 
on platforms where double and long double have the same precision. 
However, I'm not sure why that should matter.  GCC makes double and long 
double separate types on all platforms, and, AFAIK, there are no 
platforms where double exists, but long double does not.

We define the long double functions in the library anyhow (as aliases 
for the double functions), but we make it harder for users to use them 
by not declaring the prototypes.

I'd be interested in the rationale for the current state of affairs.

Thanks,

-- 
Mark Mitchell
CodeSourcery, LLC
mark@codesourcery.com
(916) 791-8304