amforth-devel — Development discussion and user support

Re: [Amforth-devel] XMEGA

[Bernard M. <bme...@gm...>]

Ok, so the existing assembler should work fine. On first glance I do see
some differences, the main one is that the interrupt vectors are 4 bytes
apart not two bytes apart as in the other devices.
Also, the names in ATxmega128A1.inc a quite a bit different so names in the
.asm files will have to change accordingly ..

On Tue, Apr 29, 2008 at 6:20 PM, Matthias Trute <mt...@we...> wrote:

>
> > Quick question: Does amforth support the XMEGA with just the addition of
> > the
> > .asm in the /devices folder and the appropriate .inc file from ATMEL
> > Studio?
>
> I hope so, but do not know it. The assembler code works with
> very little modifications (mostly regarding the serial interface).
>
> > Or is a bit harder than that?
>
> maybe, but Atmel mentioned that existing atmega code works for atxmega
> without modifications.
>
> bye
> Matthias
>
>
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by the 2008 JavaOne(SM) Conference
> Don't miss this year's exciting event. There's still time to save $100.
> Use priority code J8TL2D2.
>
> http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
> _______________________________________________
> Amforth-devel mailing list
> Amf...@li...
> https://lists.sourceforge.net/lists/listinfo/amforth-devel
>

Re: [Amforth-devel] avrdude error downloading image

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

Bernard,

debugging hardware remotly is difficult. Did you check google, there
was a short discussion around this with some possible reasons that cause
a serial line to not work, the most stupid one is probably keeping
the programmer connected (that may keep the reset line).

> I have a 16mhz crystal connected to an external oscillator which feeds the
> chip via XTAL1. Using the link you mentioned I get LFUSE = 0xC0 HFUSE=
> 0xD9
> nad EFUSE= 0xff, is this correct?

Cannot check, but the calculator works fine, usually.

> Not sure what to debug next.

just join on #forth-ev this evening

Bye
Matthias

Re: [Amforth-devel] avrdude error downloading image

[Bernard M. <bme...@gm...>]

Thanks Mathias,

I have a 16mhz crystal connected to an external oscillator which feeds the
chip via XTAL1. Using the link you mentioned I get LFUSE = 0xC0 HFUSE= 0xD9
nad EFUSE= 0xff, is this correct?

I then added these values to the sample makefile and did a "make write-fuse"
and then reset the device. I also have frequency=16000000 in the bf.asm
file.

I still have no activity out the serial port. I am using miniterm and a
loopback on the cable works fine at 9600 baud (what I have in bf.asm).

Not sure what to debug next.

Thanks again,
Bernard

On Tue, Apr 29, 2008 at 8:06 PM, Matthias Trute <mt...@we...> wrote:

>
> > 1. the line ".equ cpu_frequency = 8000000" in the bf.asm file, is that
> the
> > crystal frequency of the crystal frequency devided by two?
>
> That is the frequency, the controller runs with in Hz. which
> clock source is used and the possible clock modification depend
> on the fuse settings. http://palmavr.sf.net/cgi-bin/fc.cgi
> may be useful here.
>
> Keep in mind, that the usart interface has a 2x setting that doubles
> the baudrate, but this is unused in the default sources.
>
> btw: The template directory is a much better starting point since the
> relevant files are better documented.
>
> > 2. Which serial port of the mega128 device will I expect the forth to
> use?
>
> If you use the default sources, connect with usart0.
>
> Bye
> Matthias
>
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by the 2008 JavaOne(SM) Conference
> Don't miss this year's exciting event. There's still time to save $100.
> Use priority code J8TL2D2.
>
> http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
> _______________________________________________
> Amforth-devel mailing list
> Amf...@li...
> https://lists.sourceforge.net/lists/listinfo/amforth-devel
>

Re: [Amforth-devel] avrdude error downloading image

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

> 1. the line ".equ cpu_frequency = 8000000" in the bf.asm file, is that the
> crystal frequency of the crystal frequency devided by two?

That is the frequency, the controller runs with in Hz. which
clock source is used and the possible clock modification depend
on the fuse settings. http://palmavr.sf.net/cgi-bin/fc.cgi
may be useful here.

Keep in mind, that the usart interface has a 2x setting that doubles
the baudrate, but this is unused in the default sources.

btw: The template directory is a much better starting point since the
relevant files are better documented.

> 2. Which serial port of the mega128 device will I expect the forth to use?

If you use the default sources, connect with usart0.

Bye
Matthias

Re: [Amforth-devel] avrdude error downloading image

[Bernard M. <bme...@gm...>]

Hi Mathias,

Thanks but I found that setting the device flag to "-c stk200" solved the
issue and downloading and verification finished fine.

I am now trying to get a prompt out of the Forth via the serial port. I have
a couple of questions:

1. the line ".equ cpu_frequency = 8000000" in the bf.asm file, is that the
crystal frequency of the crystal frequency devided by two?

2. Which serial port of the mega128 device will I expect the forth to use?

Thanks,
Bernard

On Tue, Apr 29, 2008 at 6:18 PM, Matthias Trute <mt...@we...> wrote:

>
> > Hi All,
> >
> > I have an STK500/STK501 and am trying to load a forth image to the
> > atmega128
> > that is on the STK501.
> > I have modified the bf.asm file etc for this micro and everything
> > assembles
> > ok on a Linux host.
> >
> > The problem is communication to the parallel port doesn't seem to work,
> > anyone have any suggestions?
>
> Looks like a permission problem on the parport device. You may try
> running avrdude as root (sudo recommended). Otherwise you may need
> to contact the avrdude folks.
>
> Bye
> Matthias
>
>
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by the 2008 JavaOne(SM) Conference
> Don't miss this year's exciting event. There's still time to save $100.
> Use priority code J8TL2D2.
>
> http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
> _______________________________________________
> Amforth-devel mailing list
> Amf...@li...
> https://lists.sourceforge.net/lists/listinfo/amforth-devel
>

Re: [Amforth-devel] Deferred number?

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

> Matthias,
>
> Thanks for the personal reply.
>
>> emit/key and friends are deferred since they should easily
>> re-targettable from the serial to say can or I2C. For number
>> there was so such scenario. Why do you want to behave number
>> differently? A solution could be to CATCH conversion errors,
>> since number THROWs on them
>
>... or to allow entering double precision numbers.

That may be added into the main-stream code base sometimes. Feel
free to send a patch ;=) But note, that the number conversion in
amforth is only good enough and not fully standard compatible...

> Of course I am only playing, but when I used Forth as a sort of
> super-monitor on an embedded system controlling a lathe I did this to
> allow me to enter hex as $abcd whilst keeping BASE set to 10.

Ahh. There are ongoing standardization tasks running with the
forth2000x folks on this topic, and I remender a (short) discussion
in the usenet (clf) this year around this topic too. We'll see, what will
happen.

bye
Matthias

Re: [Amforth-devel] XMEGA

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

> Quick question: Does amforth support the XMEGA with just the addition of
> the
> .asm in the /devices folder and the appropriate .inc file from ATMEL
> Studio?

I hope so, but do not know it. The assembler code works with
very little modifications (mostly regarding the serial interface).

> Or is a bit harder than that?

maybe, but Atmel mentioned that existing atmega code works for atxmega
without modifications.

bye
Matthias

Re: [Amforth-devel] avrdude error downloading image

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

> Hi All,
>
> I have an STK500/STK501 and am trying to load a forth image to the
> atmega128
> that is on the STK501.
> I have modified the bf.asm file etc for this micro and everything
> assembles
> ok on a Linux host.
>
> The problem is communication to the parallel port doesn't seem to work,
> anyone have any suggestions?

Looks like a permission problem on the parport device. You may try
running avrdude as root (sudo recommended). Otherwise you may need
to contact the avrdude folks.

Bye
Matthias

[Amforth-devel] avrdude error downloading image

[Bernard M. <bme...@gm...>]

Hi All,

I have an STK500/STK501 and am trying to load a forth image to the atmega128
that is on the STK501.
I have modified the bf.asm file etc for this micro and everything assembles
ok on a Linux host.

The problem is communication to the parallel port doesn't seem to work,
anyone have any suggestions?

The error is:
-------------------------------------------------------------------
make: *** [bf] Error 1
bmentink@laptop:~/amForth/amforth-2.7/appl/avr-STK500$ make bf
avrdude -c stk500  -P /dev/parport0 -p atmega128 -e -U flash:w:bf.hex:i -U
eeprom:w:bf.eep.hex:i
avrdude: ser_open(): can't set attributes for device "/dev/parport0":
Inappropriate ioctl for device
make: *** [bf] Error 1

Thanks,
Bernard.

Re: [Amforth-devel] Deferred number?

[Tom H. <cel...@gm...>]

Matthias,

Thanks for the personal reply.

> emit/key and friends are deferred since they should easily
> re-targettable from the serial to say can or I2C. For number
> there was so such scenario. Why do you want to behave number
> differently? A solution could be to CATCH conversion errors,
> since number THROWs on them

One reason would be to allow the entering of different number bases
whatever BASE is set to, or to allow entering double precision
numbers.

Of course I am only playing, but when I used Forth as a sort of
super-monitor on an embedded system controlling a lathe I did this to
allow me to enter hex as $abcd whilst keeping BASE set to 10.

Regards

Tom Harris <cel...@gm...>

[Amforth-devel] XMEGA

[Bernard M. <bme...@gm...>]

Quick question: Does amforth support the XMEGA with just the addition of the
.asm in the /devices folder and the appropriate .inc file from ATMEL Studio?
Or is a bit harder than that?

Regards,
Bernard

Re: [Amforth-devel] AT90CAN128 Forth code

[Bernard M. <bme...@gm...>]

Thanks Mathias,

I may just do that. In the same vein, is there a mega128.asm around
somewhere?
If not I may create one using the at90can128.asm as a guide (I know there is
the associated m128def.inc file supplied by the atmel studio install).

Cheers,
Bernard.

On Mon, Apr 28, 2008 at 4:48 PM, Matthias Trute <mt...@we...> wrote:

> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Bernard Mentink wrote:
> > Hi All,
> >
> > Has anyone written and Forth code to support the CAN perpheral on this
> > device?
>
> The mcu core is supported, the perpheral to my knowledge not (yet).
>
> You are welcome to contribute it ;=)
>
> Matthias
> -----BEGIN PGP SIGNATURE-----
> Version: GnuPG v1.4.6 (GNU/Linux)
>
> iD8DBQFIFVcl9bEHdGEMFjMRAoWLAJoDw8ARTXxPA2QhqT3HOIFDXDh5rwCfRLcq
> 5nePEafyN1h6Q+VYtRKCb9w=
> =nQ3+
> -----END PGP SIGNATURE-----
>
> -------------------------------------------------------------------------
> This SF.net email is sponsored by the 2008 JavaOne(SM) Conference
> Don't miss this year's exciting event. There's still time to save $100.
> Use priority code J8TL2D2.
>
> http://ad.doubleclick.net/clk;198757673;13503038;p?http://java.sun.com/javaone
> _______________________________________________
> Amforth-devel mailing list
> Amf...@li...
> https://lists.sourceforge.net/lists/listinfo/amforth-devel
>

Re: [Amforth-devel] Deferred number?

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

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Tom Harris wrote:
> Greetings,
> 
> I am still trying to find my way around amforth, after a 20 year gap
> since I last used Forth. I have what I suspect is a stupid question.
> 
> On the Forth I last used I remember that the number word was vectored
> via a variable 'number, so that you could replace the number parser.
> Amforth does not use a deferred word for number, so it is not possible
> to replace number, is this correct? I can redefine number, but any
> previous usages of number, like the all-important quit word will still
> use the old definition. Is there a reason why number was not deferred
> like emit, key, pause, etc.?

emit/key and friends are deferred since they should easily
re-targettable from the serial to say can or I2C. For number
there was so such scenario. Why do you want to behave number
differently? A solution could be to CATCH conversion errors,
since number THROWs on them

Matthias
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQFIFVfh9bEHdGEMFjMRAjClAKCJqDErXIVFGpUqXJf8f2OLfYy1mgCgpSQL
SZ+Rb2gJc+9XJiXPjG1gPB8=
=Xizt
-----END PGP SIGNATURE-----

Re: [Amforth-devel] AT90CAN128 Forth code

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

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Bernard Mentink wrote:
> Hi All,
> 
> Has anyone written and Forth code to support the CAN perpheral on this
> device?

The mcu core is supported, the perpheral to my knowledge not (yet).

You are welcome to contribute it ;=)

Matthias
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.6 (GNU/Linux)

iD8DBQFIFVcl9bEHdGEMFjMRAoWLAJoDw8ARTXxPA2QhqT3HOIFDXDh5rwCfRLcq
5nePEafyN1h6Q+VYtRKCb9w=
=nQ3+
-----END PGP SIGNATURE-----

[Amforth-devel] Deferred number?

[Tom H. <cel...@gm...>]

Greetings,

I am still trying to find my way around amforth, after a 20 year gap
since I last used Forth. I have what I suspect is a stupid question.

On the Forth I last used I remember that the number word was vectored
via a variable 'number, so that you could replace the number parser.
Amforth does not use a deferred word for number, so it is not possible
to replace number, is this correct? I can redefine number, but any
previous usages of number, like the all-important quit word will still
use the old definition. Is there a reason why number was not deferred
like emit, key, pause, etc.?


On the bright side, what other language allows you to have enough
knowledge to poke around in the internals after a few days spent
learning it?

TomH

[Amforth-devel] AT90CAN128 Forth code

[Bernard M. <bme...@gm...>]

Hi All,

Has anyone written and Forth code to support the CAN perpheral on this
device?

Thanks,
Bernard Mentink

[Amforth-devel] New release: 2.7

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

Hi,

I just made a new release of amforth.

The most notable changes are a better ans94
conformance (stack order of double cell values)
and the assembler from Lubos Pekny (www.forth.cz).
The assembler may emit a warning like

Appnotes\m169def.inc(1199): warning: Register r31 (and r30) already 
defined by the .DEF directive

this is harmless and can be ignored.

The downloadable package contains now every
file to make it easier to use.

Have fun with it
Matthias

Re: [Amforth-devel] Atmega8 UART issues

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

Klaus Zerbe wrote:

> @Matt: If possible, try also another controller since Atmega8 ist a 
> little bit small for amforth anyway.

For the records: I just put an amforth from repository trunk on an
atmega8 with an external 12MHz quartz and everything worked as expected.

   amforth 2.7 ATmega8
 > unused cells decimal u.
  2298 ok
 >


The fuses are as follows

MCU=atmega168
LFUSE=0xee
HFUSE=0xdf
EFUSE=0xf9

MCU=atmega8
LFUSE=0xee
HFUSE=0xd9

Bye
Matthias

Re: [Amforth-devel] Block commands

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

Hi,

> are there yet any ideas on how the Block commands will be done? I guess
> the
> biggest problem is memory as the normal 1024 byte block size would be far
> to big.

I've implemented some words for the I2C EEPROM with the native block sizes
of the chips (around 32 or 64 bytes per block).

One approach to implement full block buffers could be a virtualization.
But this will be difficult since the normal memory access words have
to be used to access the block buffer. A word set like b@/b! to
access (virtual) block memory would certainly work, but does not help
porting code from other forth's .

In older forth version a word b/blk was defined to give the size of
a block buffer in bytes. The value can be different for each block
device (EEPROM, SD-Card etc).

A total different solution would be: full block buffer sizes but only on
systems that have enough memory (internal or external).

All solution should provide some kind of an abstraction layer, since
there are far too many different block devices available. That
layer can work with native block sizes.

Bye
Matthias

[Amforth-devel] Block commands

[Christian B. <cas...@ca...>]

Servus, (=bavarian for hello and goodbye)

are there yet any ideas on how the Block commands will be done? I guess the 
biggest problem is memory as the normal 1024 byte block size would be far to 
big.

Servus
  Casandro

Re: [Amforth-devel] Adding own code to amforth: reserving heap space

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

Christian Berger wrote:

> Ohh I was able to get it from therw. It seems like you did preety much what I 
> did. I think from what I have seen the only point where I could claim my code 
> to be better is in the vsync area, but that's largely irrelevant. Ohh and my 
> clock doesn't need an additional quartz.
> 
> Anyhow, are you interrested in getting some FM-Synthesis sound routines for 
> your application?

I'm collecting everything and publish all I'm allowed to do.
But the TV application is not mine.

Bye
Matthias

Re: [Amforth-devel] Adding own code to amforth: reserving heap space

[Christian B. <cas...@ca...>]

Am Donnerstag 13 März 2008 20:57:57 schrieb Matthias Trute:
> Erich Waelde wrote:
> > Hello,
> >
> > Christian Berger wrote:
> >> Am Mittwoch 12 März 2008 20:43:16 schrieb Erich Waelde:
> >>> not sure whether this helps: You did look at the application
> >>> in amforth/appl/tv, did you?
> >>
> >> No, but that sounds interresting. What is it and where do I find it? I
> >> don't have an appl directory.
> >
> > You need to look at the svn sources directly:
> > http://amforth.svn.sourceforge.net/viewvc/amforth/appl/tv/
>
> I'm sorry to tell, that the link does not work any more...

Ohh I was able to get it from therw. It seems like you did preety much what I 
did. I think from what I have seen the only point where I could claim my code 
to be better is in the vsync area, but that's largely irrelevant. Ohh and my 
clock doesn't need an additional quartz.

Anyhow, are you interrested in getting some FM-Synthesis sound routines for 
your application?

Servus
  Casandro

Re: [Amforth-devel] Adding own code to amforth: reserving heap space

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

Erich Waelde wrote:
> Hello,
> 
> Christian Berger wrote:
>> Am Mittwoch 12 März 2008 20:43:16 schrieb Erich Waelde:
>>> not sure whether this helps: You did look at the application
>>> in amforth/appl/tv, did you?
>> No, but that sounds interresting. What is it and where do I find it? I don't 
>> have an appl directory.
> 
> You need to look at the svn sources directly:
> http://amforth.svn.sourceforge.net/viewvc/amforth/appl/tv/

I'm sorry to tell, that the link does not work any more...

The reason is simple: a reorganization of the repository
structure. It is not yet fully finished but I think it won't change
a lot again, since I use the new structure internally for quite some
time. The only thing thats left is around applications that I cannot
work with, like the tv application. The actual link for them is

http://amforth.svn.sourceforge.net/viewvc/amforth/applications/

All new code is now in trunk, regardless of core or library or
(my) applications.

Matthias

Re: [Amforth-devel] Adding own code to amforth: reserving heap space

[Erich W. <ew....@on...>]

Hello,

Christian Berger wrote:
> Am Mittwoch 12 März 2008 20:43:16 schrieb Erich Waelde:
> > not sure whether this helps: You did look at the application
> > in amforth/appl/tv, did you?
> 
> No, but that sounds interresting. What is it and where do I find it? I don't 
> have an appl directory.

You need to look at the svn sources directly:
http://amforth.svn.sourceforge.net/viewvc/amforth/appl/tv/


Cheers,
Erich

[Amforth-devel] Forth-Computer first release

[Christian B. <cas...@ca...>]

Servus,

I just made the first release of the Forth-Computer, a modified version of 
amforth with video output and keyboard input.

Here's the release note:
http://www.mikrocontroller.net/topic/94193

Servus
  Casandro