So, one trick with WIDTH is to create hexadecimal-literal words $xx and $xxxx with WIDTH set to 1. They'd be IMMEDIATE, and parse their text after the first character as a hexadecimal literal. I forget where I came across this trick, and I don't believe I ever used it.

I've never tried this and Fleet Forth does not have WIDTH , but it does set the high bit of the count and the high bit of the last character in the name ( to keep (FIND) fast ) so the following would accomplish the same thing in Fleet Forth:

Since this is for a Commodore 64, just read each double forward slash as a backslash.

I've been playing around with Fig Forth and one thing mystifies me. I can't see how it initializes the X register for the argument stack. Does it just use page zero as a circular buffer and starts with whatever X value has originally? Or does it somehow stealth initializes it?

_________________
http://WilsonMinesCo.com/ lots of 6502 resources
The "second front page" is http://wilsonminesco.com/links.html .
What's an additional VIA among friends, anyhow?

Thanks. I was looking for an ldx instructions, but I see how it's done now.

Here is a slight speed improvement for CMOVE

From this

To this

_________________
http://WilsonMinesCo.com/ lots of 6502 resources
The "second front page" is http://wilsonminesco.com/links.html .
What's an additional VIA among friends, anyhow?

I had broken the code, but nobody complained. Shows the fix with indented lines.

ProForth, which is based on FigForth but is fully compatible with the '79 standard, and I have also added some '83 standard words, for some further enhanced compatibility.

One of the changes I have made that doesn't seem to have broken any compatibility with FigForth, is to have -FIND return with the CFA instead of the PFA.

Steps to do that are:
' (find) 36 dup c@ ( it is decimal 36 ($24)( make sure it says 5)( do a memory DUMP if not showing the 5 and adjust the 36)
' (find) 36 dup + 3 swap c! ; ( change it to 3)

once you do that you can now see that (find) has changed to the CFA and requires 36 2+
' (find) 38 dup c@

next change these definitions from
: cfa pfa 2 - ;
: lfa pfa 4 - ;
: nfa pfa 5 - -1 traverse ;
: pfa traverse 5 + ;

to
: cfa ; ( make this into a NULL word)
: lfa 4 - ;
: nfa 3 - -1 traverse
: pfa 1 traverse 5 + ;
or
: pfa dup @ 1f and + 5 + ;

INTERPRET has two occurrences of CFA that are no longer needed, but since the CFA word above was made into a NULL word, they have no effect.

Make sure you are saving this on a backup of the Forth System.

At this point, the only word that is broken after all the changes is (;CODE). (;CODE) has PFA CFA, but since CFA is now a NUL word, the CFA word cannot be used. A couple of options: one is to make 2- a word definition higher up in memory and manually link to it or since (;CODE) is so short, just make a new definition for it:

: (;CODE) ( --- ) R> LATEST PFA 2 - ! ;

At this point, nothing is broken and we are now closer to being able to use new words that are supported by newer standards which were created to aid in Forth compatibility across platforms.

words that can now be used are
: >body 2+ ;