Hi
I have a general question of understanding
after having read
//http://www.newlisp.org/MemoryManagement.html
It seems that parameters are always passed by value to the functions. This means that a parameter value is copied, i.e. duplicated, when calling a function. This seems to be the case even for lists.
So if I have a function with a list as an argument which calls itself recursively let's say 50 times, do I have then 50 times a copy of the same list in the memory? Or am I missing something?
Thank you for the answer in advance
--HJH
You can pass a list or any other data object by reference using contexts as described in the manual chapter 16 "Programming with context objects" in the sub chapter: "Passing objects by reference":
;; pass an object by reference
(set 'mydb:lst (sequence 1 10000))
(define (change-db obj idx value)
(nth-set idx obj:lst value))
(change-db mydb 1234 "abcdefg")
(nth 1234 mydb:lst) => "abcdefg"
and of course you could pass the database by symbol without the usage of contexts:
(define (change-db lst idx value)
(nth-set idx (eval lst) value))
(change-db 'mydb 1234 "abcdefg") ; quoted 'mydb
(nth 1234 mydb) => "abcdefg"
both methods work at same speed, but the first one may be easier to read/understand.
Lutz
Thank you for the answer. In fact for me the second version with passing the large data structure (the data base list) as a symbol is easier to understand.
--HJH
Quote from: "Lutz"
...
[code]
(define (change lst idx value)
(nth-set idx (eval lst) value))
(change-db 'mydb 1234 "abcdefg") ; quoted 'mydb
...
Lutz
Shouldn't that be: (change 'mydb 1234 "abcdefg") ?
.
Yes, I corrected the original post to all 'change-db'
Lutz
This inconsistent function name also occurs in the "Design elements and patterns" doc:
Chapter 6. Creating accessing and modifying lists
section: Passing lists by reference
Quote
(define (change-data obj i value)
(nth-set i obj:data value))
(set 'DB:data '(a b c d e f))
(change DB 3 'x) => d
DB:data => (a b c x e f)
Thanks for the correction. A new version is posted here: http://newlisp.org/DesignPatterns.html
Lutz
Thank you for this links to the design patterns page. It is very useful. I was not aware of it.
A suggestion for the future:
A section of a side by side comparison of some of the more prominent patterns with COMMON LISP would be useful. I am currently learning COMMON LISP (CLISP and ACL) as well so I am willing to contribute if I am pointed to the right spots where to start working out the differences.
A nice thing would be as well something like a "bilingual dictionary" which gives the translations (or only rough equivalents if there is a difference in meaning) between the predicates and function of newLisp and COMMON LISP. Even an incomplete list will be helpful. I am currently only using 30 functions / predicates of newLisp and can do useful things with it.
My aim is to translate the script I just did to COMMON LISP because I would like to get a feeling to see how different it is and to give it to people who probably are not willing to experiment with newLisp.
I may post this 30 word lists of functions / predicates, if you like.
(it is about file IO, list manipulation and string manipulation. With these functions you can already accomplish a lot in general data processing or more precisely - scripting).
--HJH
BTW Has somebody a newLisp function which takes the name of a newLisp source file as input and produces a statistics of the built-in predicates used?
(the result beeing a list of key-value paris; the key is newLisp predicate/function, value is the number of times the thing is used)
Try this one:
;; count primes in a newLISP source file
;;
;; returns a list of built in function symbols in
;; a newLISP source file sorted by frequency
;;
;; example:
;;
;; (count-primes "somefile.lsp")
(define (count-primes file-name)
(letn ( (tkns (parse (read-file file-name)))
(utkns (unique tkns))
(primes (filter (fn (s) (primitive? (eval (sym s)))) utkns) ))
(sort (map list (count primes tkns) (map sym primes)) > )
)
)
Lutz