Brian Havard wrote:
>
> >
> >yeah, I was afraid of that. Where is the macro defined?
>
> I belive I did say math.h above :)
>
Oops :) Sorry :)
>
> >We could
> >first check for implementations that use macros and then true
> >functions. Despite the nasty error (which I'm unsure why we're
> >seeing... such messages should be going to /dev/null) does
> >the build process completely die?
>
> It still builds fine, it just doesn't define HAVE_ISNAN even though we do
> have it. Just how important an impact that has, I have no idea. AFAIK
> nothing in Apache uses ap_snprintf to print a float.
>
For now, let's just override it in the correct *.h file (ala what
OtherBill just did for Win32). A false positive is worse, in
this case, than a false negative.
--
===========================================================================
Jim Jagielski [|] [EMAIL PROTECTED] [|] http://www.jaguNET.com/
"Hell is hot, that's never been disputed by anybody."
