XOR question

Comments

• Don M Posts: 1,654

  2013-02-13 15:07 edited 2013-02-13 15:07

  Actually that works. Is there a better or more accepted way?
• Don M Posts: 1,654

  2013-02-13 15:22 edited 2013-02-13 15:22

  How do I get this chksum calculation to work correctly?
  

  PUB Demo  | d, i, s, chksum
  
    ser.Start(RX, TX, 0, 9_600)                        ' communicate with 4D display
    term.start(31, 30, %0000, 115_200)                    ' start terminal using PST
    outa[RST]~~
    dira[RST]~~                                           ' Set RST pin as output
     
    
    pause(500)
    term.str(@ver)                                        ' display version     
  
  
  ' Goto Form3
  
    repeat i from 0 to 4
      s := byte[@form3][i]
      ser.tx(s)
      chksum := (chksum ^ i)
    ser.tx(chksum)
    term.hex(chksum, 4)
    
  DAT
  form3   byte  $01, $0A, $03, $00, $00  
  

  
  It calculates wrong. It shows $04 instead of $08.
• Don M Posts: 1,654

  2013-02-13 15:24 edited 2013-02-13 15:24

  Sorry- i figured it out. My mistake.
  

  PUB Demo  | d, i, s, chksum
  
    ser.Start(RX, TX, 0, 9_600)                        ' communicate with 4D display
    term.start(31, 30, %0000, 115_200)                    ' start terminal using PST
    outa[RST]~~
    dira[RST]~~                                           ' Set RST pin as output
     
    
    pause(500)
    term.str(@ver)                                        ' display version     
  
  
  ' Goto Form3
  
    repeat i from 0 to 4
      s := byte[@form3][i]
      ser.tx(s)
      chksum := (chksum ^ s)
    ser.tx(chksum)
    term.hex(chksum, 4)
    
  
• Mike Green Posts: 23,101

  2013-02-13 15:27 edited 2013-02-13 15:27

  You have to initialize chksum to zero. Global variables (VAR) are initialized to zero and the method return variable (RESULT) is initialized to zero, but the local variables (d, i, s, chksum) are not. They take on whatever was left over in the stack area from before the method (Demo) was called. Sometimes that's zero, but it's not guaranteed.
• Don M Posts: 1,654

  2013-02-13 16:20 edited 2013-02-13 16:20

  Ok. Thanks Mike.
• Mark_T Posts: 1,981

  2013-02-13 17:35 edited 2013-02-13 17:35

  While on the subject of XOR I can't resist mentioning the hoary old trick for swapping two variables without using temporary storage:

    a ^= b
    b ^= a
    a ^= b
  

  (though this doesn't need XOR if your machine is 3-address or has a reverse-subtract instruction

    a -= b
    b += a
    a := b - a   ' reverse subtract
  

  (Of course PASM is 2-address and I don't think has a reverse-subtract but I bet there's a way to do this
  without XOR or an extra register - mini challenge)
• Don M Posts: 1,654

  2013-02-14 06:28 edited 2013-02-14 06:28

  This works too:
  

    chksum := 0 
    repeat i from 0 to 70
      s := byte[@text][i]
      ser.tx(s)
      chksum ^= s  
    ser.tx(chksum)
  
• Ariba Posts: 2,694

  2013-02-14 08:30 edited 2013-02-14 08:30

  And this should also work:

  chksum := 0 
    repeat i from 0 to 70
      s := text[i]
      ser.tx(s)
      chksum ^= s  
    ser.tx(chksum)
  

  
  Andy