Loop-and-a-half (Re: Curious assignment behaviour)

[Paul Rubin]

Greg Ewing <greg at cosc.canterbury.ac.nz> writes:
> > (The real code stored the
> > number of the last item processed, not the number of items processed
> 
> In that case, if there are 0 items, there is no "last
> item processed", so that's going to be a special case
> anyway.

The last item processed in that case is the last item processed by
a previous run of the program.