Hi Werner,
You are probably right. So long as the C++ bindings and examples work
for you...
Andrew
On Mon, Mar 16, 2009 at 09:05:14PM +0100, Werner Smekal wrote:
> Hi Andrew,
>
> for Visual C++ _isnan is defined in float.h and not in math.h and/or
> cmath. So this test is not valid for Visua
Hi Andrew,
for Visual C++ _isnan is defined in float.h and not in math.h and/or
cmath. So this test is not valid for Visual C++ anyways, since _isnan
isn't defined in either cmath or math.h. So it's maybe really the best
to not run this test for Visual C++.
Regards,
Werner
Andrew Ross wrote:
Werner,
Some history here. There were some unix platforms (AIX perhaps?) which at
the time had C++ implementations that were behind the times, so although
isnan was in math.h it was not in cmath. This test was to catch those.
For modernish compilers I would suspect math.h and cmath to be in sync
Hi,
we test in TestBrokenIsnanCXX.cmake if isnan is in either cmath or
math.h - problem is Visual C++ doesn't pass this test, since the
functions is called _isnan and not isnan. We test this already in
plplot.cmake where we look for isnan or _isnan. The problem now is these
are contradictory c
