-ffpe-trap=precision,denormal in gfortran

[GerritGroot]

Hi,

Does anyone know what the option

-ffpe-trap=precision

does in gfortran

If I read 4 real values given in a text file like "1. 2. 3. 4."
and compile with the option -ffpe-trap=precision, the program breaks down, supposedly because it's losing precision, but ehhh..., losing precision when reading values in is a bit weird, isn't it?

Does someone know what causes this strange behaviour?

Thanks,

Gerrit

P.S.: I also would like to know what the option, "denormal" does in ffpe-trap, I couldn't find any useful info on that one either in my gcc pdf.