Drinkable Chicken

More efforts to make s1 code shorter...

I added/changed the following:

• slice function (this will come in handy for string manipulation, although (slice s a b) is still longer than s[a:b]
• nf variable (indicates the number of fields in a line)
• $ now accepts any expression, not just a number literal (so we can say $nf or $(- x 1) or whatever)
• -f command line option to quickly set the field separator
• created an alias s1 for 'csi -ss /path/to/s1.scm' (OK, this is not a change to the tool proper, but it helps... nobody wants to type csi -ss with the full path, all the time)

None of this is particularly original (as it's all borrowed from Awk and Python), but it does help to make code much shorter. I can now write a one-liner like this:

find "${1-.}" -type d | s1 -f/ '(print "   |" (make-string nf #\-) $nf)'

...which produces almost the same tree as described here. Almost, because it uses one "-" for each directory rather than two. We can fix that by making the example somewhat longer (and less readable to non-Schemers, although admittedly the Awk example isn't very readable to non-Awkers either):

find "${1-.}" -type d | s1 -f/ '(print "   |" (make-string (* (- nf 1) 2) #\-) $nf)'

Next up: regular expressions...