flashforth-devel

[Flashforth-devel] Strange behavior DO ... LOOP

[Oleg V. <mer...@gm...>]

Hello, Mikael Nordman.

I recently started more closely studying the FORTH language with which I
met Even in my childhood when the first 8-bit computers appeared. Now
when Microcontrollers
have become more powerful than the first 8-bit computers FORTH Starting to
gain momentum and I found your compiler which I really liked.
I read various books on the FORTH language and try examples and tasks from
books Execute on your version of FlashFORTH. There are many examples and
tasks where The base statement of a cycle in language FORTH is used is DO
... LOOP, But FlashFORTH does not have such an operator, but I found the
"avr / forth / doloop.txt" file which adds the possibility of using the DO
... LOOP operator in FlashFORTH.
"doloop.txt" is perfectly compiled without errors, but when trying to use In
their definitions of op. DO's words, for example

: do_test cr 10 0 do i. cr loop;

when you try to execute do_test hangs tightly. And I tried it On different
boards arduino UNO, arduino MEGA and just on Atmega128, the result the
same. But I found the old version of FlashFORTH on 06.07.2015 and so It all
works fine! Therefore, I have a question for you as a developer Is there
something I'm doing wrong or is it some kind of bug in new versions of
FlashFORTH?

I also have this question, as far as I understand FlashFORTH can address A
maximum of 64Kb of Flash memory because the size of cell = 16 bit and the
operation Read and Write will only work within the low 64KB Flash, ie There
is no sense in using microcontrollers with Flash memory > 64kb such as
Atmega128,
Atmega2560 or it is possible to access Flash memory > 64kb for example for
storing text strings of the menu or other similar information?

P.S. Sorry for my bad English.

All the best,
Oleg.

Re: [Flashforth-devel] Strange behavior DO ... LOOP

[Mikael N. <mik...@fl...>]

Hi Oleg,
Just uploaded a corrected doloop.txt.

I had moved the do loop index to use the P register but did not update 
the doloop code.
That also means the the P register and the doloop cannot be used/nested 
together.

BR Mikael


-- 
--
Mikael

Re: [Flashforth-devel] Strange behavior DO ... LOOP

[Mikael N. <mik...@fl...>]

Hi,
you can use x@ x! to fetch and store cells on 64, 128 and 256 KB 
devices.
The adresses are unmapped double numbers.
There is also xdump.

Here are examples from atm2560

pfl  ok<$,ram> 3200
33200. hex 40 xdump
033200: 5c f5 86 66 6f 72 67 65 74 ff 0f 94 10 f8 0f 94 \..forget.......
033210: 7a f4 0f 94 23 f8 0f 94 d0 fe 0f 94 c5 f4 0f 94 z...#...........
033220: da f6 0f 94 ba f4 02 97 9a 93 8a 93 0f 94 d0 fe ................
033230: 0f 94 da f6 9a 93 8a 93 0f 94 45 f0 0f 94 5a f0 ..........E...Z.
  ok<$,ram>
see forget
320a 940f f810 call  bl
320e 940f f47a call  word
3212 940f f823 call  latest
3216 940f fed0 call  @
321a 940f f4c5 call  (f)
321e 940f f6da call  ?abort?
3222 940f f4ba call  c>n
3226 9702
3228 939a      st    -y r25
322a 938a      st    -y r24
322c 940f fed0 call  @
3230 940f f6da call  ?abort?
3234 939a      st    -y r25
3236 938a      st    -y r24
3238 940f f045 call  flash
323c 940f f05a call  dp
3240 940f fe93 call  !
3244 940f fed0 call  @
3248 940f f823 call  latest
324c 940f fe93 call  !
3250 940d f053 jmp   ram
  ok<$,ram>

33200. x@  ok<$,ram> f55c

1234 12400. x!  ok<$,ram>
12400. x@  ok<$,ram> 1234

On 2017-04-11 21:44, Oleg Voitiuk wrote:
> Hello, Mikael.
> 
> Many thanks, now everything works well!
> 
> I still want to know how to deal with flash memory which is above
> 64kb, is it possible to use it now or will it be possible to use it in
> the near future?
> 
> Cheers,
> Oleg.