I've just been trying out the revised Hayes tester from the Forth 200X
web site (so it ought to be the latest version) and found a problem
with compiling it with a system with no floating point.

After the preliminaries the constants HAS-FLOATING and HAS-FLOATING-
STACK both end up as FALSE.

This leads to the system trying to compile the following

DECIMAL
: COMPUTE-CELLS-PER-FP ( -- U )
DEPTH 0E DEPTH 1- >R FDROP R> SWAP - ;
HEX

when it fails to recognise the 0E in the third line. I presume this is
meant to be recognised as floating point zero. But how is a system
that has no floating point supposed to recognise that 0E is a floating
point number? Even if it did then the FDROP wouldn't be recognised.

Presumably this could be corrected by surrounding this bit of code
with something like

HAS-FLOATING [IF] ( DECIMAL etc )
[ELSE] 0
[THEN]

On 2 Apr, 14:29, Gerry jackson9000.fsnet.co.uk> wrote: