overflow makes my script crash

[Dragan Vidovic]

Hi

I just installed the native port of Octave 3.0 in Windows, and I am
trying my old scripts that I used to run in Linux. There I use the
erfc function on a matrix which has some too large values so they
produce overflow (underflow? Doesn't matter). As a result in Linux no
error would be reported and the return matrix would contain NaN values
in places where the overflow happened. However, in Windows an error in
DERFC is reported ("X SO BIG ERFC UNDERFLOWS") and the script crashes.
I tried to use the "try-catch" block, but then I don't get the good
values in the return matrix, i.e. those for which the overflow did not
happen. I understand that this behavior may be safer, but my scripts
don't work any more and I'd like to avoid rewriting. My question is if
I can get the Linux behavior in Windows.

Thanks in advance

Dragan