Thought I'd also have a go at translating ch-1.pl into C..

ch-1.c produces identical (non-debugging) output to my Perl original.

ch-1.c uses the command-line argument processing module args.[ch].

The C version (ch-1.c) is approx twice as fast than even my fastest
Perl version (ch-1g.pl).

As a further optimization in ch-1a.c: let's pre-compute all the powers
of digits: this runs about 30% faster than ch-1.c did.