David N. Williams wrote:
>
> Sorry for the weird subject title -- couldn't think what to call
> this.
>
> A problem I've been having with gforth distills to the following:
>
> --------------
> s" [UNDEFINED]" pad c! pad char+ pad c@ move
> pad find nip 0=
> [IF]
> : [UNDEFINED] ( "name" -- flag )
> bl word find nip 0= ; immediate
> [THEN]
>
> [UNDEFINED] (f: [IF] : (f: postpone ( ; immediate [THEN]
>
> [UNDEFINED] (f: [IF] : (f: postpone ( ; immediate [THEN]
>
> [UNDEFINED] f>s [IF] : f>s (f: r -- s: n ) f>d drop ; [THEN]
>
> [UNDEFINED] f>s [IF] cr .( f>s is undefined in fileb.fs ) cr
> [THEN]
> -------------
>
> In my situation, the last three [UNDEFINED]'s were included from
> another file, but that's irrelevant.
>
> When I run this, F>S does not get defined.
>
> Can anybody see what I'm doing wrong?
The [IF]-processor is too smart, but not smart enough. In the "0
[IF]" case it sees a '(' and ignores everything until the next ')'
(including the [THEN]).
A workaround is to define (f: like this:
[UNDEFINED] (f: [IF] : (f: postpone ( ; immediate \ )
[THEN]
Since this (-parsing is against the ANS Forth specification, this is a
bug in Gforth.
- anton
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
