Porting stuff from the net (under ape) which wants to use IEEE floats. i.e. ieee754_float32_t, powf(), fabsf() and log10f().
I could map these to plan9 floats and call the double version of the transient functions, but is there a better solution? -Steve
