Hi, swipl 8.4.1 : (ins)?- A=call(A), call(A). ERROR: Stack limit (1.0Gb) exceeded ERROR: Stack sizes: local: 0.7Gb, global: 0.2Gb, trail: 0Kb ERROR: Stack depth: 10,735,430, last-call: 0%, Choice points: 3 ERROR: Probable infinite recursion (cycle): ERROR: [10,735,430] system:call(<compo…

What do you want that would actually be useful? :slightly_smiling_face: An infinite loop is not useful, in itself, e.g. 10 goto 10 in BASIC.

[image] brebs: What do you want that would actually be useful? I could write “hello world” in a loop: (ins)?- A = ( write('hello '), call(B)), B = ( writeln( 'world'), call(A)), call(A). hello world hello world hello world ... And I could decide how the loop begins: (ins)?- A = ( write('he…

But why are you using call/1 for this? It works as expected (???) if it is a normal predicate, like: a :- a. and then: ?- a. % loops forever You can have it with mutual recursion, too: a :- b. b :- a. or as in your example: hello :- write(hello), world. world :- write(world), hello. I hope s…

[image] Boris: a :- a. [image] Boris: a :- b. b :- a. Both don’t interrupt, so swipl uses tail recursion this time. But my usecase is the call of the cyclic term. Regards.

[image] Frank_Schwidom: But my usecase is the call of the cyclic term. My comment might come across as argumentative, but to me this looks like a mechanic, not a use case. I mean that “calling a cyclic term” is a mechanic but to what end? Surely you must have some use case for it and it might…

[image] Boris: My comment might come across as argumentative, but to me this looks like a mechanic, not a use case. I mean that “calling a cyclic term” is a mechanic but to what end? Surely you must have some use case for it and it might be that there is a different mechanic to achieve the same?…

OK, I really don’t understand. Maybe I should stay out of it.

Are you trying to make a meta-interpreter, using a cyclic term that’s derived from the clauses? That is, the cyclic term is the fix-point of the recursive term that represents the clause? If so, it might be useful to show how you create that cyclic term. (I tried something like this a long long tim…

As is, call(X) does not do last call optimization. It could of course. There are some scenarios where this is probably fairly easy to do. The same code is also used by many other meta-calling primitives where last call optimization should not be done. Also, if the argument to call/1 is a co…

A=call(A), call(A). % can I get tail recursion?

Help!

brebs April 23, 2022, 10:04am 7

Seems relevant: Compiler support for cyclic terms? - #3 by jan

Topic		Replies	Views
Goals in freeze are not tail call optimized General	8	162	November 5, 2025
Declarative predicates by limiting resources Help!	28	2040	August 31, 2019
How to do tail recursion without growing the stack? Help!	2	1171	June 11, 2021
Compiler support for cyclic terms? Discussion	3	642	March 1, 2022
Change in performance measures with non-tail recursion Nice to know	9	157	July 16, 2024

A=call(A), call(A). % can I get tail recursion?

Related topics