Trapping Floating Point Errors

[countryone77]

We would like to trap floating point errors (e.g., division by zero) by
setting flags on the command line.  We tried using the following, but
they do not trap a floating point division by zero when using either
version 3.4.3 or 4.0.2:

        -fsignaling-nans -ftrapping-math

Is it possible to turn on floating point error trapping via command
line options?  If so, how is it done?

Thanks,
Bev in TX