Hi Kashyap,
> While working on the json library I had to write a string to float
I recommend to use upper case for locally bound symbols
(de isNum (N)
(or (= "." N) (num? (format N))) )
The 'num?' can be omitted, because 'format' returns NIL if it does not succeed.
(or (= "." N) (format N)) )
> function. In a list of characters when I encounter a digit, I'd like to go
> over the rest of the list and consume all the digits (and possibly a
> period) that form a number. I'd appreciate some feedback from a Lispiness
> and PicoLispiness point of view :)
Just a few notes:
(de read-num (N Lst)
(let
(C 0 # count of digits found
P F # period found?
'F' (or 'f' in the original) is probably wrong here. It is not bound and
contains an unpredictable value. You meant
P NIL # period found?
Ds (pack N
(make
(while (isNum (car Lst))
(ifn P (setq P (= "." (car Lst))))
#-> (and (= "." (car Lst)) (on P))
(setq C (inc C))
#-> (inc 'C)
(link (pop Lst))
(setq Lst (cdr Lst)) ) ) )
#-> (link (++ Lst)) # Same as (link (pop 'Lst))
(pop Lst) does not pop the list, but pops from the CAR of that list which
happens to be a transient symbol with its value being itself. So it luckily
worked ;)
result (if P Ds (pack Ds ".0")) ) # if a period was not present then
add a ".0"
(list C result)) )
There are many ways to write this function. I don't know which would be the best
one, perhaps I would go with this:
(de read-num (N Lst)
(let L
(make
(while (isNum (car Lst))
(link (++ Lst)) ) )
(list
(length L)
(pack N L (unless (member "." L) ".0")) ) ) )
> On a related note - Is chopping a reasonable way to break an input file
> into characters while parsing?
Yes, but you don't need to 'chop' explicitly. You can call (char) to get a
single character, (line) to get a (chopped) list of characters, or (till) to get
a list of all characters in the file (usually not the best solution).
You may look at e.g. @lib/xm.l to see a non-trivial (yet still simple) parser.
> I mean, would each character now occupy one
> cell?
Yes.
☺/ A!ex
--
UNSUBSCRIBE: mailto:picolisp@software-lab.de?subject=Unsubscribe
