amforth-devel

[Amforth] Release 4.4 tagged and published

[Matthias T. <mt...@we...>]

Hi,

I've just tagged and published a new release named 4.4.

Noteworthy changes are

- The Game of Life with terminal output (requires controllers
  with at least many KB RAM, otherwise make the world small
  enough). A queens puzzle solver as well (more kind of a
  benchmark however).

- The reference card got a more compact layout, many bugs
  from the documentation headers in the source file are gone.

- new interrupt handling. The ideas are relativly old but
  now it works well. Many thanks to Wojciech and Al, I beg
  your pardon that it took such a long time to implement. But
  sometimes some active tests such as Erichs will speed up
  things ;=)

- The recognizer thing is now documented (to some degree);
  I sent an article to the German Forth FIG about it, that
  should be translated into plain English ASAP. Leons floating
  point library (a few patches are needed to work with it, just
  use the float.frt that is included) can now be included at
  runtime and the text interpreter can deal afterwards with
  the numbers nativly. Great stuff.

- internally there are even more source files and the
  amforth.asm file got smaller than ever. COLD is now
  WARM and the new COLD is what was in amforth.asm...

- not really my favorite: If you set WANT_IGNORECASE to
  1 and rebuild amforth you get a case insensive dictionary
  lookup. It took a few lines of code in icompare.asm, The
  discussion about it took more time than the implementation. Brr

Have fun, esp with Life

Matthias

Re: [Amforth] Release 4.4 tagged and published

[Hannu V. <vu...@ms...>]

Hi!

Amforth is releasing quite fast. It was like yesterday when 4.2 came out :)
Big thank you everyone who has contributed.

Would it be good idea to change the 8k requirement from web page to something like Amforth takes about 8k Flash but doesn't fit out of box to 8k MCUs?

> - The reference card got a more compact layout, many bugs
> from the documentation headers in the source file are gone.

 rx rx? tx tx? +usart and also @e and !e are described several times. e words have mention about XMEGA but tx and rx have same description. This has probably something to do with the fact we have UART USART and USART0. 

Also numeric I/O part u words have maybe little bit too compact description. ud. ud.r u. u.r are double cell output. I'm still learning to do things with Amforth and that kind of documentation doesn't help much. It is probably huge task to do but someone should check out the descriptions of words. I would do that however I'm not familiar enough to do that.

Web page wordlist and refcard.pdf differs little bit but that isn't big problem. Bigger problem is that some words give error 404.

One nice feature would also be generation of refcard on compilation. Interactively added words can't be there but let's assume that one creates Amforth system and writes some frt files which defines words, they could be put to Application part of card. The more work requiring thing would be list not included words in the project. That would have save so much my time if I had some list which had included and excluded words. I have already written few words to do some tasks and usually start working with all words included and when system is working I'll take words out one by one until I have just working system :)

But I'll give 4.4 a try when I have more time. Now I have some thing called life to live :)

Best regards,
Hannu Vuolasaho
 		 	   		  

Re: [Amforth] Release 4.4 tagged and published

[Matthias T. <mt...@we...>]

Hi Hannu


> Would it be good idea to change the 8k requirement from web page to
> something like Amforth takes about 8k Flash but doesn't fit out of
> box to 8k MCUs?

Well, maybe. Microcontrollers in general and amforth in
particular are no easy toys anyway... The German Forth
FIG prepares some workshops with pre-configured hardware.
That is an easier start.

> 
>> - The reference card got a more compact layout, many bugs from the
>> documentation headers in the source file are gone.
> 
> rx rx? tx tx? +usart and also @e and !e are described several times.
> e words have mention about XMEGA but tx and rx have same description.
> This has probably something to do with the fact we have UART USART
> and USART0.

The reason for that many words with the same name is simple:
they serve the same purpose in different environments. One
rx is a receive on an atmega using the interrupt driver, the
other one is the same but for an atxmega etc. But every rx
does the same. Anyways I've updated the words to give them better
(longer) names and added some more text to the PNO words (pictured
numeric output). The new pdf is on the project home page.

> Web page wordlist and refcard.pdf differs little bit but that isn't
> big problem. Bigger problem is that some words give error 404.

That sometimes happens, I did not found out, who is guilty for that.
Currently it seems to work again.

> One nice feature would also be generation of refcard on compilation.

That would change the refcard to an application specific one. Nice
idea...

> Interactively added words can't be there but let's assume that one
> creates Amforth system and writes some frt files which defines words,
> they could be put to Application part of card.

The refcard generator is already overcomplicated, dealing with forth
source code would make it to explode...

> The more work
> requiring thing would be list not included words in the project. That
> would have save so much my time if I had some list which had included
> and excluded words.

good idea as well, for an application specific build report.

Matthias

Re: [Amforth] Release 4.4 tagged and published

[pito <pi...@vo...>]

Matthias, great news!

> afterwards with the numbers nativly. Great stuff.

Last fall I did it like:

\ floating point input with >float, "& 1.23456e12"
: & bl word count >float not abort" NaN" 
state @ if postpone fliteral then ; immediate

& -3.44556e-16 & 0.23233E13 f* fs.

Does it mean we can input the float now like this?:

-3.44556e-16 0.23233E13 f* fs.

Is the Leon's float patched in the 4.4 already? (as I can remember
there was an issue with printing the floats with 4.2, it worked with
4.0)

> It took a few lines of code in icompare.asm, The
> discussion about it took more time than the implementation. Brr
Who needs that actually?

Pito.


-- 
IHNED.cz je nový, přehlednější a rychlejší. Přesvědčte se na:
www.ihned.cz

Re: [Amforth] Release 4.4 tagged and published

[Matthias T. <mt...@we...>]

Pito,

> Does it mean we can input the float now like this?:
> 
> -3.44556e-16 0.23233E13 f* fs.

yes, that is the goal of it. You can even compile
floats in colon definitions and they will work.


> Is the Leon's float patched in the 4.4 already? (as I can remember
> there was an issue with printing the floats with 4.2, it worked with
> 4.0)

yeah, there were a few regressions there. Now it works again. see my
other mail.

Matthias

Re: [Amforth] Release 4.4 tagged and published

[<an...@ki...>]

Excellent, thanks guys.

Hopefully I will get some time to give it a spin over the nexy couple of
weeks.

Think I will stick with case sensitive.

Cheers

Andy Kirby

On 24/05/11 18:54, Matthias Trute wrote:
> Hi,
> 
> I've just tagged and published a new release named 4.4.
> 
> Noteworthy changes are
> 
> - The Game of Life with terminal output (requires controllers
>   with at least many KB RAM, otherwise make the world small
>   enough). A queens puzzle solver as well (more kind of a
>   benchmark however).
> 
> - The reference card got a more compact layout, many bugs
>   from the documentation headers in the source file are gone.
> 
> - new interrupt handling. The ideas are relativly old but
>   now it works well. Many thanks to Wojciech and Al, I beg
>   your pardon that it took such a long time to implement. But
>   sometimes some active tests such as Erichs will speed up
>   things ;=)
> 
> - The recognizer thing is now documented (to some degree);
>   I sent an article to the German Forth FIG about it, that
>   should be translated into plain English ASAP. Leons floating
>   point library (a few patches are needed to work with it, just
>   use the float.frt that is included) can now be included at
>   runtime and the text interpreter can deal afterwards with
>   the numbers nativly. Great stuff.
> 
> - internally there are even more source files and the
>   amforth.asm file got smaller than ever. COLD is now
>   WARM and the new COLD is what was in amforth.asm...
> 
> - not really my favorite: If you set WANT_IGNORECASE to
>   1 and rebuild amforth you get a case insensive dictionary
>   lookup. It took a few lines of code in icompare.asm, The
>   discussion about it took more time than the implementation. Brr
> 
> Have fun, esp with Life
> 
> Matthias
> 
> ------------------------------------------------------------------------------
> vRanger cuts backup time in half-while increasing security.
> With the market-leading solution for virtual backup and recovery, 
> you get blazing-fast, flexible, and affordable data protection.
> Download your free trial now. 
> http://p.sf.net/sfu/quest-d2dcopy1
> _______________________________________________
> Amforth-devel mailing list
> Amf...@li...
> https://lists.sourceforge.net/lists/listinfo/amforth-devel
> 

Re: [Amforth] Release 4.4 tagged and published

[Leon N. M. <leo...@gm...>]

On Tuesday, May 24, 2011 07:54:08 PM Matthias Trute wrote:
> - The recognizer thing is now documented (to some degree);
>   I sent an article to the German Forth FIG about it, that
>   should be translated into plain English ASAP. Leons floating
>   point library (a few patches are needed to work with it, just
>   use the float.frt that is included) can now be included at
>   runtime and the text interpreter can deal afterwards with
>   the numbers nativly. Great stuff.

Nice! I haven't played around with the floating point code (or amforth in general) in a while, but I'll try to make some time with it this weekend. It looks like some of the bugs are old ones 
from my code (e.g. >float failing for numbers with too many digits), so I'll see what I can do.

-Leon