amforth-devel

[Amforth] amforth-shell.py patch proposal

[Enoch <ix...@ho...>]

Hello Matthias & all:

As I promised to generalize crc8.frt (#1) a question was raised how would
the user configure it to its use, namely, select the generating
polynomial byte and the needed bit order (#2).

I answered the question through the following amforth-shell.py
improvement which introduces a local appl_owns.py substitutions file.

Without further ado here's the patch and its use for your perusal:

Patch: http://pastebin.com/uDK7fk9N

Example:

$ cat appl_owns.py
# amforth-shell.py application specific forth substitutions

APPL_OWNS = { 'DEBUG': '1' }

$ make shell

...snip...

(AT90CAN128)> DEBUG .
1  ok

Regards, Enoch.

#1 Assuming that Matthias places it in ${AMFORTH}/lib/crc8.frt :-)
#2 With no real constant-s to pay for in the user code.

Re: [Amforth] amforth-shell.py patch proposal

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

Hi Enoch,

>
> As I promised to generalize crc8.frt (#1) a question was raised how would
> the user configure it to its use, namely, select the generating
> polynomial byte and the needed bit order (#2).
> 
> I answered the question through the following amforth-shell.py
> improvement which introduces a local appl_owns.py substitutions file.
> 
> Without further ado here's the patch and its use for your perusal:
> 
> Patch: http://pastebin.com/uDK7fk9N

A few notes

+        sys.path.insert(0, ".")
         if os.environ.has_key("AMFORTH_LIB"):
           self._search_list = os.environ["AMFORTH_LIB"].split(":")
-        else:
-          self._search_list=["."]

During my tests I found that change less useful. I run the shell
not only from the appl directory but also from $HOME, so adding
"." to the search path leads to a huge amount of files, the
shell reads in upon start (on my system it took more than only
a few seconds. And I have 8GB RAM, but my $HOME is bigger). So
I reject this part of the patch. Sorry.

For the other stuff "appl_own.py" I've not opinion yet. What
exactly is the problem you want to solve? What I would like to
have is something that could the following (forth code!)

[undefined] foo
  #include foo
[endif]

for which the shell sends the content of foo unless foo is already
defined. The control words are not sent to the controller as well.

File oriented forth's have something like that, and I think the
shell is capable of doing something similiar.

Matthias

Re: [Amforth] amforth-shell.py patch proposal

[Enoch <ix...@ho...>]

Matthias Trute <mt...@we...> writes:

> Hi Enoch,
>
>>
>> As I promised to generalize crc8.frt (#1) a question was raised how would
>> the user configure it to its use, namely, select the generating
>> polynomial byte and the needed bit order (#2).
>> 
>> I answered the question through the following amforth-shell.py
>> improvement which introduces a local appl_owns.py substitutions file.
>> 
>> Without further ado here's the patch and its use for your perusal:
>> 
>> Patch: http://pastebin.com/uDK7fk9N
>
> A few notes
>
> +        sys.path.insert(0, ".")
>          if os.environ.has_key("AMFORTH_LIB"):
>            self._search_list = os.environ["AMFORTH_LIB"].split(":")
> -        else:
> -          self._search_list=["."]
>
> During my tests I found that change less useful. I run the shell
> not only from the appl directory but also from $HOME, so adding
> "." to the search path leads to a huge amount of files, the
> shell reads in upon start (on my system it took more than only
> a few seconds. And I have 8GB RAM, but my $HOME is bigger). So
> I reject this part of the patch. Sorry.

Point taken, I'll amend. My project resides outside the amforth
hierarchy, hence, I did not notice the hit.

> For the other stuff "appl_own.py" I've not opinion yet. What
> exactly is the problem you want to solve? What I would like to
> have is something that could the following (forth code!)
>
> [undefined] foo
>   #include foo
> [endif]

That all came in context of crc8.frt. It needs two configurable
constants as both the polynomial chosen and the reading order are
application specific (in fact MSB first is considered the reverse
order).

(
$D5 constant CRC8PLY 
-1 constant CRC8MSB
)

I said, how nice it would be if amforth-shell would do it for me just
like it does with Atmel's register/bit defs.

By the way, I surround the above constants with () which amforth-shell.py
respects but amforth itself does not. That's just perfect!

Regards, Enoch.


>
> for which the shell sends the content of foo unless foo is already
> defined. The control words are not sent to the controller as well.
>
> File oriented forth's have something like that, and I think the
> shell is capable of doing something similiar.
>
> Matthias
>
>
> ------------------------------------------------------------------------------
> Everyone hates slow websites. So do we.
> Make your web apps faster with AppDynamics
> Download AppDynamics Lite for free today:
> http://p.sf.net/sfu/appdyn_d2d_feb

Re: [Amforth] amforth-shell.py patch proposal

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

Enoch,

> I said, how nice it would be if amforth-shell would do it for me just
> like it does with Atmel's register/bit defs.

That makes sense. But appl_own.py as the filenname..... There should
be better ones. Or a more generic approach. I'm not a python expert,
but many python directories contain a file named __init__.py to do
something useful.

> 
> By the way, I surround the above constants with () which amforth-shell.py
> respects but amforth itself does not. That's just perfect!

Uhm. Other would have call it a bug ;)

Matthias

Re: [Amforth] amforth-shell.py patch proposal

[Enoch <ix...@ho...>]

Matthias Trute <mt...@we...> writes:

> Enoch,
>
>> I said, how nice it would be if amforth-shell would do it for me just
>> like it does with Atmel's register/bit defs.
>
> That makes sense. But appl_own.py as the filenname..... There should
> be better ones. Or a more generic approach. I'm not a python expert,
> but many python directories contain a file named __init__.py to do
> something useful.

Seems too "pythonic" for me :-)
In fact, I would have preferred it to be appl_owns.frt with
"$D5 constant CRC8PLY" lines, only that I am too lazy :-)

Regards, Enoch.

>> 
>> By the way, I surround the above constants with () which amforth-shell.py
>> respects but amforth itself does not. That's just perfect!
>
> Uhm. Other would have call it a bug ;)
>
> Matthias
>
>
> ------------------------------------------------------------------------------
> Everyone hates slow websites. So do we.
> Make your web apps faster with AppDynamics
> Download AppDynamics Lite for free today:
> http://p.sf.net/sfu/appdyn_d2d_feb

Re: [Amforth] amforth-shell.py patch proposal

[Enoch <ix...@ho...>]

Hello Matthias & all:

> Matthias Trute <mt...@we...> writes:
>> That makes sense. But appl_own.py as the filenname..... There should
>> be better ones. Or a more generic approach. I'm not a python expert,
>> but many python directories contain a file named __init__.py to do
>> something useful.
>
> Seems too "pythonic" for me :-)
> In fact, I would have preferred it to be appl_owns.frt with
> "$D5 constant CRC8PLY" lines, only that I am too lazy :-)
>

Thanks for the encouragement. I hope that my new attempt to patch
amforth-shell.py is better.

Patch: http://pastebin.com/ENbdSkj5

Usage:

Create a file in the application directory called appl_defs.frt:
Yes, a forth file :-)

\ example

0 constant false
-1 constant true
$d5 constant CRC8PLY
-1 constant CRC8MSB

Let's create the test.frt:

\ #error-on-output no

(
$d5 constant CRC8PLY
-1 constant CRC8MSB
)

CRC8PLY .
CRC8MSB .

$ make shell
AMFORTH_LIB=.:/home/enoch/avr/amforth/trunk/lib python /home/enoch/avr/amforth/trunk/tools/amforth-shell.py -p /dev/ttyUSB2
|I=appl_defs: example
|I=appl_defs: 4 loaded
|I=Entering amforth interactive interpreter
|I=getting MCU name..
|I=successfully loaded register definitions for at90can128
|I=getting filenames on the host
|I=  Reading .
|I=  Reading /home/enoch/avr/amforth/trunk/lib
(AT90CAN128)> #include test.frt
|D=#include test.frt
|I=getting filenames on the host
|I=  Reading .
|I=  Reading /home/enoch/avr/amforth/trunk/lib
|I=using test.frt from/home/enoch/private/project
**** /home/enoch/private/project
|I=getting MCU name..
|I=successfully loaded register definitions for at90can128
|F=/home/enoch/private/project/test.frt
|D|    1|\ #error-on-output no
|W|    2|
|C|    3|(
|C|    4|$d5 constant CRC8PLY
|C|    5|-1 constant CRC8MSB
|C|    6|)
|W|    7|
|S|    8|CRC8PLY .
|O|    8|213
|S|    9|CRC8MSB .
|O|    9|-1
|W|   10|
*

Regards, Enoch.