While updating kForth-Win32 from its current version, 1.x, to version
2.x, to make it functionally similar to kForth-32 for Linux, I was
confounded for a short while by a bug:

create e 4 allot

e
' e execute
' e >body

The last three statements should return the same address on a standard
Forth system, and they weren't. Turned out that, in moving from ver 1.x
to 2.x, I had added another implementation of >BODY and the Jump Table
was referring to the wrong version (the older version of >BODY hadn't
been removed for some reason).

Anyway, this sort of check is not part of any of the CORE tests, and
maybe should be considered for a basic test of a standard system.

Cheers,
Krishna

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)