Thank you Tristan,
That would explain alot.
When I saw that itype was undefined on gforth & ESP32Forth for initial
quick testing, I commented it out
without thinking (me bad - I should have known better by now). I should
have defined it properly for 32 bit.

Thanks again
John S
PS: I couldn't believe the price on CH32V307 dev boards so I ordered a few
https://www.aliexpress.com/item/1005009417544259.html?spm=a2g0o.order_list.order_list_main.4.448c18021QAp1G



On Sun, Jan 4, 2026 at 4:12 PM <[email protected]> wrote:

> Hi John,
>
> > Do the testing words t{...}t work properly on your amForth on CH32V307
> > ?
>
> Yes. If I #include common/lib/forth2012/tester/tester-amforth.frt and
> give
> it some simple tests.
>
> |S|    3|t{ 0 0 = -> true }t
> |S|    4|t{ 1 1 = -> true }t
> |S|    5|t{ -1 -1 = -> true }t
> |S|    6|t{ 1 0 = -> false }t
> |S|    7|t{ -1 0 = -> false }t
> |S|    8|t{ 0 1 = -> false }t
> |S|    9|t{ 0 -1 = -> false }t
> |C|   10|\ and now one that is a deliberate error
> |S|   11|t{ $a 0 = -> true }t
> |O|   11|INCORRECT RESULT: t{ $a 0 = -> true }t
>
> The responses are correct. Will it properly pass all the tests as
> given below? No.
>
> [1] https://github.com/gerryjackson/forth2012-test-suite
>
> Some it will never pass because of the memory organisation, some because
> the words are not applicable and some because they are (currently) just
> missing.
>
> > In my testing I have found that these words don't work properly in
> > a 32 bit Forth be it RISCV (ESP32Forth) or even I386 (gForth - Linux ).
> > These words weren't giving an error even if the test was obvious to
> > give an error message (false positive).
>
> I don't have any of these 32 bit Forths to hand but IIRC
> tester-amforth.frt
> is a modified version of tester.fr (see [1]), so that it could run on
> AmForth. From tester-amforth.frt, the word below
>
> : ERROR     \ ( C-ADDR U -- ) DISPLAY AN ERROR MESSAGE FOLLOWED BY
>          \ THE LINE THAT HAD THE ERROR.
>     itype source type cr         \ DISPLAY LINE CORRESPONDING TO ERROR
>     EMPTY-STACK              \ THROW AWAY EVERY THING ELSE
> ;
>
> uses itype which is an AmForth word, one I would not expect to be in
> ESP32Forth. I would be surprised if it would work correctly on an
> ESP32Forth. I also would not expect tester-amforth.frt to work at all on
> the unmodified AmForth RISC-V svn repo source because of its
> experimental
> status[2]. It will, however, work perfectly on AmForth avr8. As an
> aside,
> AmForth's avr8 heritage, born of the capabilities of ATmega mcus,
> ripples through the codebase. Something that has caught me out a few
> times in my AmForth RISC-V endeavours.
>
>
> Best wishes,
> Tristan
>
> [2] https://sourceforge.net/p/amforth/mailman/message/36511255/
>
>
> On 2026-01-04 17:40, John Sarabacha wrote:
> > Hi Tristan,
> > Do the testing words t{...}t work properly on your amForth on CH32V307
> > ?
> > These were found in tester-amforth.frt used for Avr8. In my testing I
> > have
> > found that these words don't work properly in a 32 bit Forth be it
> > RISCV
> > (ESP32Forth) or even I386 (gForth - Linux ). These words weren't giving
> > an
> > error even if the test was obvious to give an error message (false
> > positive).
> >
> > Thanks,
> > John S
> >
> > _______________________________________________
> > Amforth-devel mailing list for http://amforth.sf.net/
> > [email protected]
> > https://lists.sourceforge.net/lists/listinfo/amforth-devel
>
>
> _______________________________________________
> Amforth-devel mailing list for http://amforth.sf.net/
> [email protected]
> https://lists.sourceforge.net/lists/listinfo/amforth-devel
>

_______________________________________________
Amforth-devel mailing list for http://amforth.sf.net/
[email protected]
https://lists.sourceforge.net/lists/listinfo/amforth-devel

Reply via email to