Re: [Flashforth-devel] For-next loops - the New Design

SourceForge Headquarters 1320 Columbia Street Suite 310 San Diego, CA 92101 +1 (858) 422-6466

Mike,

-fortest
marker -fortest

2variable count
: _for 0. count 2! for count 2@ 1. d+ count 2! r@ u. next cr count 2@ ud. ; 
: _for2 0. count 2! for2 count 2@ 1. d+ count 2! r@ u. next cr count 2@ ud. ; 

FOR:
----------------------------
25 _for 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
25  ok<#,ram>

1 _for 0
1  ok<#,ram>

0 _for
0  ok<#,ram>

65535 _for 65534 65533 ..
.. 13 12 11 10 9 8 7 6 5 4 3 2 1 0
65535  ok<#,ram>

FOR2:
___________________

25 _for2 25 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
26  ok<#,ram>

1 _for2 1 0
2  ok<#,ram>

0 _for2 0
1  ok<#,ram>

65535 _for2 65535 65534 ..
.. 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
65536  ok<#,ram>

My comments:
It does not behave as it should (this is my personal point of view):
_for works as before, zero skips, not good
_for2 works not good, it means you do N+1 _for or something like that..

I strongly believe it should work as follows ("My Design"):
standard FOR-NEXT:
-------------------------------------------------

65535 _for 65534 65533 ..
.. 13 12 11 10 9 8 7 6 5 4 3 2 1 0
65535  ok<#,ram>

25 _for 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
25  ok<#,ram>

1 _for 0
1  ok<#,ram>

0 _for 65535 65534 ..
.. 13 12 11 10 9 8 7 6 5 4 3 2 1 0
65536 ok<#,ram>

?FOR-NEXT with zero skip:
--------------------------------------------------
65535 _?for 65534 65533 ..
.. 13 12 11 10 9 8 7 6 5 4 3 2 1 0
65535  ok<#,ram>

25 _?for 24 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0
25  ok<#,ram>

1 _?for 0
1  ok<#,ram>

0 _?for
0  ok<#,ram>

IgorM

______________________________________________________________
> Od: Mikael Nordman <mik...@fl...>
> Komu: <fla...@li...>
> Datum: 26.05.2015 16:07
> Předmět: Re: [Flashforth-devel] For-next loop - full 16bit count?
>
>I added printing of the index to your pseudocode.
>
>So     2 for r@ u. next
>prints  2 1
>
>And     0 for r@ u. next
>prints  0 65535 .... 1
>
>So will also your pseudocode below, but
>
>I want  65535 for r@ u. next
>prints  65535 .... 0
>
>This is what I have implemented in the FOR2
>You can pull it from git.
>
>BR Mike
>
>
>uint16 index;
>uint16 count = 0;
>
>index = 2;
>
>for_loop:
>
>blink_a_led;
>
>count = count + 1;
>
>print index;              <----------------NOTE
>
>/ / decrement and skip if zero test
>index = index  - 1;   /// decrement
>if index==0 goto out_of_loop // skip if zero
>// end of decrement and skip if zero test
>
>goto for_loop:   // this will be skipped when zero
>
>out_of_loop:
>
>print count;
>
>
>On 26.05.2015 13:28, om1zz wrote:
>> A loop with "decrement and skip if zero" construct in asm means condition evaluation and action around "next"-  like this (an example code):
>>
>> uint16 joe;
>> uint16 count = 0;
>>
>> joe = 2;
>>
>> for_loop:
>>
>> blink_a_led;
>>
>> count = count + 1;
>>
>> / / decrement and skip if zero test
>> joe = joe  - 1;   /// decrement
>> if joe==0 goto out_of_loop // skip if zero
>> // end of decrement and skip if zero test
>>
>> goto for_loop:   // this will be skipped when zero
>>
>> out_of_loop:
>>
>> print count;
>>
>> end
>>
>> With joe=2 the count will be 2, with joe=0 the count will be 65536.
>>
>> Igor
>>
>> ______________________________________________________________
>>> Od: Mikael Nordman <mik...@fl...>
>>> Komu: <fla...@li...>
>>> Datum: 26.05.2015 11:51
>>> Předmět: Re: [Flashforth-devel] For-next loop - full 16bit count?
>>>
>>> If you do a decrement and skip if zero, it means that one index will
>>> not be counted (=0).
>>> So you would not get the full 16-bit range.
>>> 2 for ---> 2 1   (2 counts)
>>> 1 for ---> 1     ( 1 counts)
>>> 0 for ---> 65535 65534 .. 1 (65535 counts)
>>>
>>> So my NEXT uses decrement and skip if carry
>>>
>>> Therefore it has work like this
>>> 2 for ---> 2 1 0 (3 counts)
>>> 0 for ---> 0     ( 1 counts)
>>> 65535 for ---> 65535 65534 .. 0 (65536 counts)
>>>
>>> BR Mike
>>>
>>>
>>> On 26.05.2015 11:21, om1zz wrote:
>>>> Some reasoning to mu suggestion below:
>>>> 1. we talking here forth - a better macro assembler
>>>> 2. with your scenario we move towards a primitive function with "more
>>>> intelligence" than one would expected from a better assembler..
>>>> 3. imho for-next in forth is an equivalent of "decrement and skip if
>>>> zero" construct, where (at least me since '82) observed the behaviour
>>>> when putting zero in it loops for a lot of times around :)
>>>> Igor
>>>>
>>>>> Hmm, the old FOR shall behave, I think:
>>>>>
>>>>> 2 for ---> 1 0 (2 counts)
>>>>> 0 for ----> 65535 65534 .. 0 (65536 counts)
>>>>>
>>>>> new ?FOR
>>>>>
>>>>> 2 ?for ---> 1 0 (2 counts)
>>>>> 0 ?for ---> skip (none count)
>>>>>
>>>>> So no change with 99,99999% of old code.
>>>>> I.
>>>>>
>>>>> ______________________________________________________________
>>>>>> Od: Mikael Nordman <mik...@fl...>
>>>>>> Komu: <fla...@li...>
>>>>>> Datum: 26.05.2015 06:38
>>>>>> Předmět: Re: [Flashforth-devel] For-next loop - full 16bit count?
>>>>>>
>>>>>> I did some experiments with this and have now two for..next loops
>>>>>> : t for r@ u. next ;  ok<#,ram>
>>>>>> : tt for2 r@ u. next ;  ok<#,ram>
>>>>>> 9 t 8 7 6 5 4 3 2 1 0  ok<#,ram>
>>>>>> 9 tt 9 8 7 6 5 4 3 2 1 0  ok<#,ram>
>>>>>>
>>>>>> The only problem is the naming of the words.
>>>>>> The old FOR ideally should be called ?FOR
>>>>>> and the new one FOR.
>>>>>> But that introduces a problem with existing FF code...
>>>>>>
>>>>>> Mike
>>>>>>
>>>>>> On 24.05.2015 16:00, om1zz wrote:
>>>>>>> Yea, I do understand the zero loop skip intention, but I consider
>>>>>>> that a bug.
>>>>>>> There is no way to do full 65536 loops with 16bit without doing 0
>>>>>>> for..next.
>>>>>>>
>>>>>>> As the web knowledge says - there are only incompatible for-next
>>>>>>> loops and inefficient and lazy do-loops :)
>>>>>>>
>>>>>>> I would recommnd to change the behavior to full 16bit, entering 0
>>>>>>> should be the users responsibility (similar as divide by zero).
>>>>>>
>>>
>>>
>>> ------------------------------------------------------------------------------
>>> One dashboard for servers and applications across Physical-Virtual-Cloud
>>> Widest out-of-the-box monitoring support with 50+ applications
>>> Performance metrics, stats and reports that give you Actionable Insights
>>> Deep dive visibility with transaction tracing using APM Insight.
>>> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>>> _______________________________________________
>>> Flashforth-devel mailing list
>>> Fla...@li...
>>> https://lists.sourceforge.net/lists/listinfo/flashforth-devel
>>>
>>
>> ------------------------------------------------------------------------------
>> One dashboard for servers and applications across Physical-Virtual-Cloud
>> Widest out-of-the-box monitoring support with 50+ applications
>> Performance metrics, stats and reports that give you Actionable Insights
>> Deep dive visibility with transaction tracing using APM Insight.
>> http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>> _______________________________________________
>> Flashforth-devel mailing list
>> Fla...@li...
>> https://lists.sourceforge.net/lists/listinfo/flashforth-devel
>>
>
>------------------------------------------------------------------------------
>One dashboard for servers and applications across Physical-Virtual-Cloud 
>Widest out-of-the-box monitoring support with 50+ applications
>Performance metrics, stats and reports that give you Actionable Insights
>Deep dive visibility with transaction tracing using APM Insight.
>http://ad.doubleclick.net/ddm/clk/290420510;117567292;y
>_______________________________________________
>Flashforth-devel mailing list
>Fla...@li...
>https://lists.sourceforge.net/lists/listinfo/flashforth-devel
>