Re: recursion

On 2024-07-16 02:41, Gerry Jackson wrote:

On 15/07/2024 20:37, Ruvim wrote:

>

>
RECURSIVE also allows you to tick the word in its own definition (not
possible with RECURSE), a feature that I actually have used;

>
I think, there should be a standard method to get the xt of the current definition (regardless whether it is a named definition, or nameless definition).
>

 It can be done by using DEFER as a forward definition
e.g.
defer foo
:noname ... ['] foo defer@ ... ; is foo

Or this way:
   variable _foo
   :noname ... _foo @ ... ; _foo !
Of course, in some particular cases this can be done in a standard way.
But what if we are creating the definition programmatically? Things become more cumbersome without intrinsic necessity.
And this is not an API —  this is not a method of getting the xt of the current definition in the general case.
In the general case, you don't know how compilation of the current definition was started, and you cannot require a special way of starting compilation of the definition.

 using DEFER@ gives the xt of the code, omittimg it gives the xt of the name.

In this use case the named definition (created by "defer") is only auxiliary, its xt is not needed.
[...]

As DEFER can be used as a forward definition it can also be used for mutual recursion
 

This is possible but, awkward. Why else do some systems provide the words like "recursive" and "forward"?
--
Ruvim