Thread-topic: Cmath and math.h Changes in Xcode 2.4
User-agent: Microsoft-Entourage/11.2.5.060620
On 8/28/06 10:59 PM, "Mike Jackson" <email@hidden> wrote:
> Was there some change to Xcode (and the entire tool chain) in Xcode 2.4 that
> would affect the isnan(x) function?
>
> I have now come across two open source packages that get hung up on
> "isnan(x) is undeclared in this context" type errors.
>
> The projects are OpenSceneGraph (www.openscenegraph.net) and ITK
> (www.itk.org).
>
> In ITK's case I had to change a #define to look like the following:
>
> #if defined(__APPLE__)
> #if (__GNUC__ >= 4)
> #include <cmath>
> #define isnan(x) std::isnan(x)
> #else
> #include <math.h>
> #define isnan(x) __isnand((double)x)
> #endif
> #endif
>
> Where as it used to be just:
> #if defined(__APPLE__)
> #include <math.h>
> #define isnan(x) __isnand((double)x)
> #endif
>
> Just curious....
I came up against the same problem and Tommy Nordgren (whom I could not get
a reply thank-you-email to) found that:
Preprocessing and grepping for isnand gives the following line:
static __inline__ int __inline_isnand( double __x ) { return __x !=
__x; }
So it looks as the correct name is : __inline_isnand
This is only for Intel builds, it is still __isnand on PowePC.
Lyndsey
--
Lyndsey Ferguson
Software Engineer
email@hidden
VectorWorks 12
Go from life without 12 to life without limits.
Visit http://www.vectorworks.net for more information.
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Xcode-users mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
http://lists.apple.com/mailman/options/xcode-users/email@hidden
This email sent to email@hidden
