Hi,
I stumbled upon a problem compiling FSL on ARMEL machines. Apparently
ARMEL does not have long double arithmetics. Therefore it doesn't have
roundl().
The only place that function is used is in src/basisfield/splinefield.cpp
My question is now whether I introduce a #ifdef __ARMEL__ solution, or
whether I can revert to MISCMATHS::round() or even std::round(). I'm
asking because the function looks like this:
unsigned int splinefield::fake_old_ksp(double nvxs,
unsigned int nksp,
double ovxs) const
{
return(static_cast<unsigned int>(roundl((ovxs/nvxs)*double(nksp))));
}
To me it looks like there is no chance that a long double will ever be
fed into roundl(), so it should be safe to revert to round().
Which path should I take?
Thanks in advance,
Michael
--
GPG key: 1024D/3144BE0F Michael Hanke
http://mih.voxindeserto.de
|