Re: cmath and pow

[Benjamin Kosnik <bkoz at cygnus dot com>]

No. Pre-process. There's a bunch of autoconf hackery in bits/std_cmath.h 
to select between powf, pow, and a using a potential __builtin_pow or 
whatever, depending on the host platform. The

The problem probably is probably that bits/std_cmath.h is wrong, and the 
wrong function is being called, or the right function is being called 
with the wrong alignment. 

So, do a 

g++ -E yourtestfile.cc -o yourtestfile.ii

And either tar/gzip it and post to this list or send to Gaby and or I. 
Another alternative is to post the generated bits/c++config.h, and it's 
possible to deduce the rest from the defines in bits/std_cmath.h

Thanks,
Benjamin


On Thu, 30 Sep 1999, Chris Prince wrote:

> > Then I'd like to see pow() prototype (in <math.h>)
> 
>   Sorry about that, the prototype is actually in __math.h and looks like
> this...
> __MATH_INLINE double pow (double __x, double __y);
> __MATH_INLINE double
> pow (double __x, double __y)
> {