Re: Making alistp matching plistp requirements

Heime <[email protected]> writes:

> On Monday, April 13th, 2026 at 10:55 AM, tpeplt <[email protected]> wrote:
>
>> Heime <[email protected]> writes:
>> 
>> > On Monday, April 13th, 2026 at 9:59 AM, tpeplt <[email protected]> wrote:
>> >
>> >>
>> >> Here is a simple definition that does not check the contents
>> >> of the items in the alist:
>> >>
>> >>   (defun alistp (obj-list)
>> >>     "Determine whether OBJ-LIST is an association list.
>> >>      This function definition says that an empty list is not
>> >>      an association list."
>> >>     (when (and obj-list (proper-list-p obj-list))
>> >>       (cl-every #'consp obj-list)))
>> >
>> > Emacs Lisp plistp allows nil as both keys and values.
>> > Have been thinking of keeping with same idea to accept
>> > nil.
>> >
>> 
>> That is a simple change to the definition, above:
>> 
>>   (defun alistp (obj-list)
>>     "Determine whether OBJ-LIST is an association list."
>>     (when (proper-list-p obj-list)
>>       (cl-every #'consp obj-list)))
>> 
>> Be aware that you may need code to distinguish between a
>> value set to nil in an alist and an empty alist:
>> 
>> (alist-get 'a '((a . nil)))
>> ;;=> nil
>> (alist-get 'a '())
>> ;;=> nil
>
> Have come up with this
>
> (defun lana-alistp (object)
>   (declare (pure t) (side-effect-free error-free))
>   (cond ((atom object)  (eq object nil))
>         (t  (and (consp (car object))
>                  (lana-alistp (cdr object))))))
>
> How do you think this compares with your version?
>  

So long as your alist is not too long, this should work.
Emacs has a nested-call limit that you should be aware of,
see ‘max-lisp-eval-depth’ in (info "(elisp) Eval").

Here is what a trace of your function looks like:

(lana-alistp '((a . b) (c . d) (1 . 2) (f . g)))
=>
1 -> (lana-alistp ((a . b) (c . d) (1 . 2) (f . g)))
| 2 -> (lana-alistp ((c . d) (1 . 2) (f . g)))
| | 3 -> (lana-alistp ((1 . 2) (f . g)))
| | | 4 -> (lana-alistp ((f . g)))
| | | | 5 -> (lana-alistp nil)
| | | | 5 <- lana-alistp: t
| | | 4 <- lana-alistp: t
| | 3 <- lana-alistp: t
| 2 <- lana-alistp: t
1 <- lana-alistp: t

Here is an example of a long alist (2000 pairs):

(defvar long-list
  (loop
   for i upto 2000
   collect '(a . b)))

When (lana-alistp long-list) is evaluated, Emacs reports:

Entering debugger...
cl-prin1: (excessive-lisp-nesting 1622)
cl-prin1: (error "Lisp nesting exceeds ‘max-lisp-eval-depth’") [1599 times]

The definition of ‘alistp’ using ‘cl-every’ will return t
when given ‘long-list’ as its argument.

-- 
The lyf so short, the craft so long to lerne.
- Geoffrey Chaucer, The Parliament of Birds.