[DEFINED] name

description:

precompiler-like test if a word is defined. The result
is tested with => [IF]

discussion:

[DEFINED] is part of the => wordset/FPH
also in FPH is [UNDEFINED]
[DEFINED] and [UNDEFINED] are also in the OTA standard.
There are quite a few words that do something similar
but are named differently.

> Guido Draheim wrote:
>   gforth:  [IFDEF] and [IFUNDEF] in base system,
>            does the nesting with a dynamic parser vector ;
>             has neither [DEFINED] nor [NOT], but
>             has immediate DEFINED ( "name" -- 0|1 )
>   old pfe:  none in base system, strictly ansi
>   new pfe:  [DEFINED] and [NOT] in base system,
>             has also an independent C-like syntax that can not
>             mix with [IF] [ELSE] (#IFDEF .. #ELSE .. #ENDIF)
>   win32for: DEFINED and HAVE in base system, actually
>             DEFINED ( "name" -- str 0 | cfa flag ) non-immediate
>             HAVE ( "name" -- 0|flag ) non-immediate
>   bigforth: [IFDEF] and [IFUNDEF] in base system,
>             implemented just like gforth does ;
>             neither [DEFINED] nor [NOT] ;
>             has a non-immediate DEFINED?

possible implementations:

: [DEFINED] ( IN: word -- CS: flag )
    BL WORD FIND NIP \ incompatible with OTA
; IMMEDIATE
: [DEFINED] ( IN: word -- CS: flag )
    BL WORD FIND NIP 0<> \ as in OTA
; IMMEDIATE
: [UNDEFINED] ( IN: word -- CS: flag )
    BL WORD FIND NIP 0=
; IMMEDIATE

( The difference between the two definitions of [DEFINED] is
in the value returned for IMMEDIATE words, 1 or -1.