Shifting Right or Left, proper command? - WORKING!

Joms

I am trying to combine two seperate hex charactors into one 'two-place' hex, but can not seem to make the shift left/right work the way I want. Do I need to decode these to bitwise then shift left/right first? Or would I use a lookup table in this application?

  mystr[noparse][[/noparse]21] := 56    'debugs as 8
  mystr[noparse][[/noparse]22] := 65    'debugs as A

    Debug.tx(mystr[noparse][[/noparse]21]) 
    Debug.tx(mystr[noparse][[/noparse]22])
    Debug.tx(13)

I can get this to debug properlly and work, however I want to try and use them in calculations as a single variable. For example, I want a variable that would now contain '8A' in hex or '138' in dec.

Thanks in advance for the help...

Post Edited (Joms) : 4/21/2009 12:47:27 AM GMT

StefanL38

Hello,

so where is your code where you tried to bitshift it ?

best regards

Stefan

Timothy D. Swieter

One of the problems that I see here is you have an ascii representation of 8A - text, and then there is the numeric representation of 8A as 0x8A (or $8A). They are two different things. I gather you want the numeric 8A as a value for manipulation, right?

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Timothy D. Swieter, E.I.
www.brilldea.com - Prop Blade, LED Painter, RGB LEDs, 3.0" LCD Composite video display, eProto for SunSPOT
www.tdswieter.com

MagIO2

if mystr[noparse][[/noparse]21]>'@'
my_byte:=((mystr[noparse][[/noparse]21] & $47) - constant( 'a' - 10 ) ) << 4
else
my_byte:=(mystr[noparse][[/noparse]21] - '0') << 4

if mystr[noparse][[/noparse]22]>'@'
my_byte|=(mystr[noparse][[/noparse]22] & $47) - constant( 'a' - 10 )
else
my_byte|=mystr[noparse][[/noparse]22] - '0'

or

my_byte := lookdown( my_byte[noparse][[/noparse]21]&$4f: '0','1','2','3','4','5','6','7','8','9','a','b','c','d','e','f') << 4
my_byte |= lookdown( my_byte[noparse][[/noparse]22]&$4f: '0','1','2','3','4','5','6','7','8','9','a','b','c','d','e','f')

The & $4f is for converting uppercase letters to lowercase. So, this conversion will work for both.

Joms

Ok, here is what I have so far. It is a result of what I started with, which was lookup tables, and I added some of the code that MagIO2 posted.

    mystr[noparse][[/noparse]21] := 56    'debugs as 8
    mystr[noparse][[/noparse]22] := 65    'debugs as A

    src_total := lookup( (mystr[noparse][[/noparse]21]&$4f) - 47 : "0".."9",0..6,"a".."f") << 8    'also tried a '4' on this line.
    src_total |= lookup( (mystr[noparse][[/noparse]22]&$4f) - 47 : "0".."9",0..6,"a".."f")         'should this line have a := instead of the |=?
 
    Debug.tx(mystr[noparse][[/noparse]21]) 
    Debug.tx(mystr[noparse][[/noparse]22]) 
    Debug.tx(13)
    Debug.dec(src_total)

I have the mystr[noparse][[/noparse]21] & [noparse][[/noparse]22] debuging properly as I would suspect they would, however still having the problems with the lookup table section.

Timothy - Just want to make sure, ascii is the same thing as a charactor, right? so "0" as ascii should be equal to "48" as a decimal?

The result I am looking for in the variable 'src_total' should be $8A or 138 as a decimal.

jazzed

Letter "A" is 65 or $41 and "a" is 97 or $61 ... you are saying mystr[noparse][[/noparse]22] := "A" but are not looking for "A" in the lookup ... even if you got rid of the questionable "- 47" in each lookup line. Yah, I should lookup lookup but that would be too easy ;-)

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
--Steve


Propalyzer: Propeller PC Logic Analyzer
http://forums.parallax.com/showthread.php?p=788230

MagIO2

lookup is simply wrong. You feed lookup with one value (mystr[noparse][[/noparse]21]&$4f)-47 and it takes this as position the following list of elements and returns that element.
So, for the values 0-9 it returns "0" - "9" as ascii value.

You have to use lookdown, which returns the index of the place where it found the value.

<<8 shifts a number 8 bits to the left. Have a look at the binary representation of bytes:
$8A = %1000 1010
So, when you decode the 8 you have to shift it 4 bits to the left to get it to the right position.

Thanks, jazzed ... that's the problem in my code ... the lookdown should use the uppercase letters.

Ahh .. and another small bug .. lookdownz is our friend here, as the ascii value $30 = "0" is at the first place of the list and should return index 0.
Another example ... $35 = "5" is at the sixth place of the list and should return index 5.
$41 = "A" is at the eleventh place of the list and should return index 10.

Post Edited (MagIO2) : 4/17/2009 7:34:21 PM GMT

Joms

OK, I think I understand... The reason I added the -47 was to start counting at zero.

Anyways, when I run your lookdown commands I get an error. It is because of the ' around each letter, the program things that it is a comment. I am going to try replacing the all the ' with ". I will let you know what happens....




EDIT: Didn't fix it... Here is the code I am at now and just get zero's...

    mystr[noparse][[/noparse]21] := 56
    mystr[noparse][[/noparse]22] := 65

    src_total := lookdown( mystr[noparse][[/noparse]21]&$4f: "0","1","2","3","4","5","6","7","8","9","a","b","c","d","e","f") << 4
    src_total |= lookdown( mystr[noparse][[/noparse]22]&$4f: "0","1","2","3","4","5","6","7","8","9","a","b","c","d","e","f")
    
    Debug.tx(mystr[noparse][[/noparse]21]) 
    Debug.tx(mystr[noparse][[/noparse]22]) 
    Debug.tx(13)
    Debug.hex(src_total,2)

Post Edited (Joms) : 4/17/2009 7:39:02 PM GMT

MagIO2

Sorry for that ... I didn't check my code because the propeller was not at hand ... it's the conversion to uppercase which fails.

It has to be $7f. And the list has to be with uppercase letters.

src_total := lookdownz( mystr[noparse][[/noparse]21]&$7f: "0","1","2","3","4","5","6","7","8","9","A","B","C","D","E","F") << 4
src_total |= lookdownz( mystr[noparse][[/noparse]22]&$7f: "0","1","2","3","4","5","6","7","8","9","A","B","C","D","E","F")

Joms

Ok, Getting darn close now... But I can not seem to figure out what to put for 'a' thru 'f' on the lookdown table...

    mystr[noparse][[/noparse]21] := 56
    mystr[noparse][[/noparse]22] := 65

    src_total := lookdownz( mystr[noparse][[/noparse]21]&$4f: 0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15) << 4     'wont allow me to use "a,b,c,d,e,f"
    src_total |= lookdownz( mystr[noparse][[/noparse]22]&$4f: 0,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15)             'wont allow me to use "a,b,c,d,e,f"
    
    Debug.tx(mystr[noparse][[/noparse]21]) 
    Debug.tx(mystr[noparse][[/noparse]22]) 
    Debug.tx(13)
    Debug.hex(src_total,2)

This works fine except when I get above 9, then it just shows 0 where it should display a A. However, it does work correctly when I get to 16 it will display a 10, which is correct.

QUESTION - It will not allow me to put a 'a,b,c,d,e,f' inplace of the '10,11,12,13,14,15' on the lookdown table. How do I need to format that part correctly?

Joms

IT WORKS!!!! MagIO2, you got it! Thanks for all the help!

I am posting the correct code I used below so that other people can see it, and it is semi documented so if somebody runs accross this post they will know the solution...

    mystr[noparse][[/noparse]21] := 56
    mystr[noparse][[/noparse]22] := 65

    src_total := lookdownz( mystr[noparse][[/noparse]21]&$4f: 0,1,2,3,4,5,6,7,8,9,"A","B","C","D","E","F") << 4
    src_total |= lookdownz( mystr[noparse][[/noparse]22]&$4f: 0,1,2,3,4,5,6,7,8,9,"A","B","C","D","E","F")
    
    Debug.tx(mystr[noparse][[/noparse]21]) 
    Debug.tx(mystr[noparse][[/noparse]22]) 
    Debug.tx(13)
    Debug.dec(src_total)

Joms

Well, I thought I had it working but seem to have a weird problem going on... The following is the code I have working...

PUB GetSource(dst) : src 
  mystr[noparse][[/noparse]21] := 30
  mystr[noparse][[/noparse]22] := 30
  
  src := lookdownz( mystr[noparse][[/noparse]21]&$4f: 0,1,2,3,4,5,6,7,8,9,"A","B","C","D","E","F") << 4
  src |= lookdownz( mystr[noparse][[/noparse]22]&$4f: 0,1,2,3,4,5,6,7,8,9,"A","B","C","D","E","F")

  src += 1

  debug.dec(src)          '<-----------This will debug properly and display a 1
  debug.tx(13)

Here is the problem.... When I call on this section of code I would expect to receive the the same thing as when I debug it inside the pub, but I don't... This is what I am using to call on the Pub called GetSource...

  shop1 := Data.GetSource(31)
  
  debug.dec(shop1)      '<----- This does not display properly and shows a 113
  debug.tx(13)

The above will show up as a 113 in the PST software when I debug.... Am I missing something here? I would expect it to be the same?

jazzed

Works for me although using your code as is I get "11" for output. Not sure why it's not "1 <NEWLINE> 1 <NEWLINE>" ....

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
--Steve


Propalyzer: Propeller PC Logic Analyzer
http://forums.parallax.com/showthread.php?p=788230

Joms

Hmmmm... OK, I am am digging farther into this... Doesn't seem to make sense to me, but I must have something simple wrong...

Thanks for checking my work...

Timothy D. Swieter

I followed some of the discussion here, I think you guys are on the right track. I would like to point a reference though. I think it is the Simple Serial object or the PC Debug object that has functions for converting numbers into characters/ASCII to send to a terminal window. These functions might give some inspiration on the conversion process. I believe it can be done without the lookdown/lookup as the code can use the first character in the character/ASCII table as the base and do a calculation from there.

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Timothy D. Swieter, E.I.
www.brilldea.com - Prop Blade, LED Painter, RGB LEDs, 3.0" LCD Composite video display, eProto for SunSPOT
www.tdswieter.com

jazzed

There are many ways to do things. Reading other's code is an excellent way to learn. It is still not clear what Joms is trying to achieve. Maybe that's a secret. Got me [noparse]:)[/noparse]

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
--Steve


Propalyzer: Propeller PC Logic Analyzer
http://forums.parallax.com/showthread.php?p=788230

Joms

Sorry guys, I tried to be clear, but sometimes I get lost in it all.· I have a sensor that actually outputs 4 seperate hex charactors.· An example would be $30, $30, $30, $30 (or as humans would read it 0000).· The overall goal I am trying to achive is read that as one number and to have it be an actual number.

$30, $30, $30, $30 = 0000

$30, $30, $30, $31 = 0001

$30, $30, $30, $36 = 0006

$30, $30, $38, $41 = 008A

I think you get the idea now...· I completely understand the part of learning from others code.· I printed out the whole prop manual and referance guide and do use that alot, but I have problems actually building code from that.· I spend quite a bit of time searching the forums and reading through old posts where people were asking code questions.· That is one of the reasons I like to post the code after I figure out the solution to the problem, so someone else can learn from it also...

AS AN UPDATE - I am still having problems getting the number to transfer between the two spin files.· I still havn't analizyed the problem completely to ask any concrete questions...

StefanL38

Hello Joms,

therefore you can use the code from PUB RxHexof the Extended_FDSerial-object

this codesnippet assumes that datain[noparse][[/noparse] ] is a bytearray containing the characters as ASCII-coded hexadecimal digits
where the least digit is in dataIn[noparse][[/noparse] 0 ]

In you case variable "NoOfDigits" would be 4

      place := 1
      repeat x from (NoOfDigits - 1) to 0                             

        if (dataIn[noparse][[/noparse]x] => ("0")) and (datain[noparse][[/noparse]x] =< ("9"))
          value := value + ((DataIn[noparse][[/noparse]x]-"0") * place)         

        if (dataIn[noparse][[/noparse]x] => ("a")) and (datain[noparse][[/noparse]x] =< ("f"))
          value := value + ((DataIn[noparse][[/noparse]x]-"a"+10) * place) 

        if (dataIn[noparse][[/noparse]x] => ("A")) and (datain[noparse][[/noparse]x] =< ("F"))
          value := value + ((DataIn[noparse][[/noparse]x]-"A"+10) * place)         

        place := place * 16                                 

best regards

Stefan

MagIO2

Instead of multiplication I'd suggest shifting, because it is faster

replace place:=1 with place:=0
replace * place with << place
and replace place := place * 16 with place +=4

In the if I'd use the trick as shown above to convert to uppercase. Then you only need one if for the letters A-F.

If you can rely on the fact that the string only contains 0-9 and A-F or a-f, you can do the if statement easier (and faster)
if dataIn[noparse][[/noparse]x] =<"9"
value := value + (DataIn[noparse][[/noparse]x]-"0")<<place
else
value := value + (DataIn[noparse][[/noparse]x]&$4f-constant("a"+10))<<place
If transmission errors are possible you can use the code above or you can check errors during transmission (where you might have some time left to check) and skip conversion completely.

(Just to show some ways to do it faster and smaller ;o)

Ole Man Earl

I need the similar thing, but to convert a decimal string to a value.
Ex:
"12345" as a string convert to
12345 as a value.

Most languages have a Val statement. I need the equivalent in spin language.
Ex:
print Val("12345") or x = Val("12345")

I need this because my gps reports altitude as a string
Ex:
$GPGGA 1963.7
as meters and I want to convert that to feet. To do so I need the string for altitude needs to be a value I can work with.

Any help ?

Ole Man Earl

BTW, Tim's 3" composite video display looks good !
Thanks Tim !

StefanL38

@MagIO,

very good improvement !
after seeing your version I ask myself why the author did it not that way
(hm - maybe for easier u nderstanding of the code

@Earl,

sometimes the solution is in a SPIN-file you might be already using
the Extended_FDSerial-object provides two methods for this

PUB rxDec : Value | place, ptr, x
PUB rxDecTime(ms) :Value | place, ptr, x, temp

which receive a string and convert it to a number

you can use this as a template to use it with other bytearrays used for storing strings

best regards

Stefan

Joms

OK, Now I am lost.· I was working with the code that MagIO2 posted and must have been missing something because it didn't act as predicted.· So, I tried the code that Stefan found from the Extended FDS object.· I can not seem to make either set act the way I would have predicted.· I did go in to the Extended FDS object and try to analyzed that code, however, it seems like I am doing it right?

If someone could please take a look at the attached code and let me know where I am going wrong, it would help a ton...

As a note:· The 3 lines of inactive code actually works like I want.· It will debug properly as I need it to, however when I go to use it in the main program I am building, it will not return the right value.· It seems like it is using the Hex value of the charactors in the other parts of the program.· I am assuming this code that I am trying is to help fix that.

Sorry if any of this is hard to understand...

Joms

I am posting the following code because it is a lot easier to just look at then to try and open the file and figure out what I am trying to do with the sensor. Whenever I run this it will always debug as "1". I have also debuged 'X' and it appears to be working 22 through 19 only like I had hoped for...

I would have though I would have debugged "138" if I had the code working like it does in the extended FDS object

  mystr[noparse][[/noparse]19] := 48   'Debugs as 0
  mystr[noparse][[/noparse]20] := 48   'Debugs as 0
  mystr[noparse][[/noparse]21] := 56   'Debugs as 8
  mystr[noparse][[/noparse]22] := 65   'Debugs as A

  place := 1
  
    repeat x from 22 to 19  
      if (mystr[noparse][[/noparse]x] => ("0")) and (mystr[noparse][[/noparse]x] =< ("9"))
        value := value + ((mystr[noparse][[/noparse]19]-"0") * place)  
      if (mystr[noparse][[/noparse]x] => ("a")) and (mystr[noparse][[/noparse]x] =< ("f"))
        value := value + ((mystr[noparse][[/noparse]19]-"a"+10) * place)   
      if (mystr[noparse][[/noparse]x] => ("A")) and (mystr[noparse][[/noparse]x] =< ("F"))
        value := value + ((mystr[noparse][[/noparse]x]-"A"+10) * place)  
      place := place * 16 
      
  debug.dec(value)

MagIO2

Ok, Joms, let's give it another try.

In your original post you wanted to convert two hex digits. As you said in a reply the lookdownz worked. In your last attachment you try to convert 4 digits. Of course you'd have to extend the code for that. But with the lookdownz there are 2 problems: Code size (because of the list of elements) and runtime (because the PASM code behind has to loop over the list to find the index). To reduce code size you could do the conversion of a 4 digit number in a loop as well:

value := 0
repeat x from 19 to 22
  value<<=4
  value|=lookdownz( mystr[noparse][[/noparse]x]&$4f: 0,1,2,3,4,5,6,7,8,9,"A","B","C","D","E","F")

The problem in you current repeat loop is:

      if (mystr[noparse][[/noparse]x] => ("0")) and (mystr[noparse][[/noparse]x] =< ("9"))
        value := value + ((mystr[color=red][b][noparse][[/noparse]19][/b][/color]-"0") * place)         

      if (mystr[noparse][[/noparse]x] => ("a")) and (mystr[noparse][[/noparse]x] =< ("f"))
        value := value + ((mystr[b][color=red][noparse][[/noparse]19][/color][/b]-"a"+10) * place) 

This should be [noparse][[/noparse]x] instead.

As I already said in a previous post, this can be optimized in several steps. As a first step I would use shifting instead of multiplication because it's much faster:

  place := 0

    repeat x from 22 to 19 step -1                             

      if (mystr[noparse][[/noparse]x] => ("0")) and (mystr[noparse][[/noparse]x] =< ("9"))
        value |= ((mystr[noparse][[/noparse]x]-"0") << place)         

      if (mystr[noparse][[/noparse]x] => ("a")) and (mystr[noparse][[/noparse]x] =< ("f"))
        value |= ((mystr[noparse][[/noparse]x]-constant("a"+10)) << place) 

      if (mystr[noparse][[/noparse]x] => ("A")) and (mystr[noparse][[/noparse]x] =< ("F"))
        value |= ((mystr[noparse][[/noparse]x]-constant("A"+10)) << place)         

      place += 4

And depending on the reliability of the transmission of the digits, you can·do the checks a bit weaker but faster:

      ' in case the digit is lower/equal nine has to be in range of "0"-"9"       if (mystr[noparse][[/noparse]x] =< ("9"))
        value |= ((mystr[noparse][[/noparse]x]-"0") << place)
      ' in case the digit is higher than nine, it has to be "a"-"f" or "A"-"F"         
      else
        value |= (((mystr[noparse][[/noparse]x]&$4f)-constant("A"+10)) << place) 

·Hope it helps.



Post Edited (MagIO2) : 4/20/2009 9:46:08 PM GMT

Joms

OK, I am still missing something. When I changed the [noparse][[/noparse]19] back to X that made that part of code work. I screwed up and had 19 in all the spots and forgot to change two of them back. Anyways, I think I understand what you are trying to do with the code you posted, however I just can not seem to make it work... I did try both versions.

Now, something else interesting, in the 'slow' code I got kinda working, it will only do one position of numbers. What I mean is it will·work correctly from 0 to 15, however instead of going to '16' in dec or '10' in hex, it will start back at 0 and count up again. I do not see where this is happening, because by using the 'place += 4, that should shift everything over 1 byte, correct?

  place := 1
  value := 0
  
    repeat x from 22 to 19 step -1 
      if (mystr[noparse][[/noparse]x] => ("0")) and (mystr[noparse][[/noparse]x] =< ("9"))
        value := value + ((mystr[noparse][[/noparse]x]-"0") * place)  
      if (mystr[noparse][[/noparse]x] => ("a")) and (mystr[noparse][[/noparse]x] =< ("f"))
        value := value + ((mystr[noparse][[/noparse]x]-"a"+10) * place)   
      if (mystr[noparse][[/noparse]x] => ("A")) and (mystr[noparse][[/noparse]x] =< ("F"))
        value := value + ((mystr[noparse][[/noparse]x]-"A"+10) * place)  
      place += 4
      
  debug.dec(value)
  debug.tx(13)        'debug carriage return

Also, I added the 'value := 0", otherwise it would count up everytime the program ran...

Post Edited (Joms) : 4/20/2009 10:24:27 PM GMT

mpark

Try changing
place := 0
and
repeat x from 22 to 19 ' no step
and
value := value + (... << place) ' three occurrences

You want to left shift by 0, 4, 8, 12 because each hex digit is 4 bits.
The repeat loop is funny when counting down. It subtracts the step, so a negative step goes the wrong way.

Joms

Great! That fixxed it. I read up on the repeat command and it does work a little different then I did expect. Thanks EVERYONE for all your help...

MagIO2

Ups ... the repeat loop behaves strange in this case. Again I had to post without running the code on my prop - shame on me.

As your code still uses the range-check I suppose you want to be sure the string to be converted had no (obvious) transfer errors. As such that's ok, but if you find such an error (say by mistake you have a "k" in your mystr) what will happen? The code simply skips this digit, which means it looks like a 0 has been send. Instead you should stop conversion and drop that value. That can be achieved by having an else in each if which sets a faulty-flag and ends the loop.
Outside of the loop you can check the faulty-flag and skip whatever you want to do with "value" or trigger retransmission - whatever makes sense in such an error-case.

MagIO2

Just for curiosity:

I wonder if that conversion would work as well (assuming we don't have transmission problems). No if at all.
value:=0
place:=0
repeat x from 22 to 19
·value |= ((mystr[noparse][[/noparse]x]|%00100000)-"0"-(mystr[noparse][[/noparse]x]>"9")&("a"-"0"+10)) << place
·place+=4

mystr[noparse][[/noparse]x]|%00100000 converts the current digit to lowercase if it is a letter. If it is a number it does not change anything.
%00110010 = 2
%00100000 bitwise or of that will still be 2

%01000001 = A
%00100000 bitwise or is %00110001 = a

So, after that we can be sure that we have either "0"-"9" or "a"-"f". We subtract "0" by default.
(mystr[noparse][[/noparse]x]>"9") this is a boolean expression returning 0 for false or -1 for true. -1 is %11111111_11111111_11111111_11111111.
(0 & something) = 0
(-1 & something) = something
Usually in our if statement from digits having "a"-"f" we subtrace "a" and add 10 to get the value of "a" as number. As we already subtracted "0" we only need to subtract the difference. So, our something is
"a"-"0"+10

If you put that calculation in a constant(), then it's done during compile-time and not during runtime.

jazzed

Well you beat me to the char | $20 bit [noparse]:)[/noparse]

▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
--Steve


Propalyzer: Propeller PC Logic Analyzer
http://forums.parallax.com/showthread.php?p=788230