(j3.2006) READ unit ambiguity

[Keith Bierman]

>> 
> 
> C812 (R819) The do-variable shall be a variable of type integer.
> 
> Is that constraint really necessary?
> 
> 
Allowing floating point indices opens up large cans of portability worms. If memory serves that was the point of the constraint in days of yore (portability in terms of a consistent number of iterations or in some possible pathological cases infinite (or virtually so) if underflow and trapping to software interacted pessimally. 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://j3-fortran.org/pipermail/j3/attachments/20120305/018f3337/attachment.html>