Re: "Back & Forth" - Local variables

On 12/03/2025 9:29 pm, albert@spenarnc.xs4all.nl wrote:

In article <858080e0b3faa16a4b9ba24e3b34e12a555494b4@i2pn2.org>,
dxf  <dxforth@gmail.com> wrote:

On 12/03/2025 8:14 am, Hans Bezemer wrote:

...
Since the LOCAL definition has already moved the return address to

     the data stack, ;: works.. In an earlier version I moved the return
     address back to the return stack and executed YIELD. Which (of course)
     worked. So they're not that far apart IMHO.

>
But of equal value?  As you say YIELD can be made to do ;:
>

 
This discussion is typical of Forth. A specification of a word
is essential,  then I can implement it.
I quarrel about the meaning of ;: , and you casually mention
"do ;:" as if everybody knows what this is supposed to mean.
 
This is the specification of CO .
Note how the stack effect is ( -- ) .
For most Forthers this is a apparently a sufficient specification.
 
"
   CO
 
  STACKEFFECT:
 
  DESCRIPTION: []
 
  Return to the caller, suspending interpretation of the current
  definition, such that when the caller exits, this definition is
  resumed. The return stack must not be engaged, such as between >R and
  R> , or DO and LOOP .

It seems a digression to formally specify something you previously
commented on and whose implementation worked on all popular forths.
But to borrow yours, the definition would go something like:

   ;:  ( nest-sys -- )
 
   Return to the caller specified by nest-sys, suspending [etc etc]