I have no reason to use something like that in real code, so this is just pure curiosity. Nevertheless, I would like to understand why this piece of code does what it does.
In newLisp, the following code
(println (f "A" 1) (f "B") (f "C"))
will print AnilD. I don't quite understand where nil comes from. I would expect either ABD or, less intuitive but still understandable, ABC. It gets even more confusing. When I do
(f "A")
the interpreter tells me:
And, indeed, in this case I have to restart the shell to make it work again.
In Common Lisp, Clojure and Ruby the following pieces of code seem to be equivalent:
(format t "~a~a~a" (f "A" 1) (f "B") (f "C"))
(println (str (f "A" 1) (f "B") (f "C")))
p f("A", 1) + f("B") + f("C")
All of them print ABD.
It does literally what you tell it to do. newLISP is a fully self reflective language which can introspect or modify itself at run-time. In your example the function
When functions modify themselves, care must be taken to not replace code parts executing at the same moment.
See about self-modifying code in newLISP on this page: http://www.newlisp.org/index.cgi?Closures the paragraphs
Thanks for the answer! So the old function definition is collected right after we redefine the function? That makes sense for me. Then, as far as I understand, modifying the function without actually redefining it should be safe (is it?):
(println (f "A" 1) (f "B") (f "C"))
It prints ADD, as expected.
Is there a way to self-modify a function so that the changes only take place
Some functions in newLISP delay memory cleanup, but not applicable in this case.
Thanks.