> James Giles writes: > > Well, actually -1 might very well be a valid unit number. Nothing in > > the standard requires unit numbers to be positive. > > 3rd para (not counting the constraint as a para) after R903 in section > 9.3 of f95. > > Actually, it says zero or positive. Must have been written by a C programmer. :-) %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%