Re: isnan/isinf problems still exist

[Johannes Schindelin]

Hi,

On Sun, 18 Sep 2005, Graham Percival wrote:

> 
> On 16-Sep-05, at 8:50 AM, Johannes Schindelin wrote:
> 
> > Concur. What I found irritating is that after #include'ing <cmath> in
> > flower/include/real.hh, isinf() is not #define'd, even when it is still
> > defined when checking within <cmath>! I have OSX 10.2.8.
> > 
> > My workaround for now is to append this after "#include <cmath>":
> > 
> > #ifndef isinf
> > #define isinf(x) __isinfd(x)
> > #define isnan(x) __isnand(x)
> > #endif
> 
> Thanks, that solves that problem.  Han-Wen, can I commit this fix to CVS?

I would be a bit more careful with that: isinf does not need to be a 
macro on every platform, and it probably is defined to be __isinfd(x) only 
on Darwin/MaxOSX. Please wrap this into a "#ifdef __apple__" (I hope that 
is correct, please check, as I don´t have access to my machine right now) 
if you want to commit it.

Ciao,
Dscho