Re: cmath vs. math.h vs isfinite

[Robert Elz <kre%munnari.OZ.AU@localhost>]

Then after actually looking at the standards, I see that
isfinite() (and its friends) are required to be defined as
macros, nothing I can see actually demands functions with
those names (both the C and POSIX standards say the same).

So, the problem is in <cmath> which shouldn't be #undef'ing
them if the names are to be allowed to be used from something
which includes <cmath>

On the other hand, I don't see how <cmath> can define its
versions of isfinite (one for each possible floating point type)
unless it #undef's the macros -- as otherwise it is just a mess.

What all that suggests is a question I wondered about in the
first place, but didn't ask in the previous message:   What is a
C++ program doing calling C library functions, when perfectly sane
alternatives are available from C++ ??

If that is what the problem is, then it seems like it might be a
python problem - it simply shouldn't be including <math.h> in a C++
program, and it should be using the C++ namespace stuff to get access
to the functions it wants to call.

kre