Weird Quirk with ITOA

Comments

• Options

  Joe Hebert Posts: 2,159

  March 2009

  If you change:
  send_string 0,"'Test 1:',itoa(crc << 8)"
  
  To this equivalent:
  send_string 0,"'Test 1:',itoa(crc * 256)"
  
  Then you get the expected result of 65280 ($FF00)
  
  Weird indeed.
  

  alexanbo wrote: »

  Was doing some troubleshooting on a CRC calculation and noticed that when I used ITOA to show some data it gave me some odd results
  

  button_event[10128:1:0,1]
  {
     push:
     {
        stack_var integer crc
        crc = $ffff
        send_string 0,"'Test 1:',itoa(crc << 8)"
        send_string 0,"'Test 2:',itoa(crc + 1)"
     }
  }
  

  
  Returned:
  Test 1:16776960
  Test 2:0
  
  Which is odd because I expected Test 1 to be $FF00 not $FFFF00. In digging I found that ITOA's parameter is actually a Long which kinda explains the first result, but then it's not consistent with the second result?

  0
• Options

  mpullin Posts: 949

  March 2009

  It might be an odd behavior of the bitwise shove (<<) as opposed to itoa. What happens if you remove itoa and just print the numbers?

  0
• Options

  alexanbo Posts: 282

  March 2009

  I think it is in itoa, because the calcuation was actually being done correctly, just when I tried to print out intemediary values with itoa that it got funky, sort of a Schr?dinger's Cat sort of thing.
  
  For example if you added a line before the print statement
  crc = $ffff << 8
  
  and changed the print statement to:
  send_string 0,"'Test1:',itoa(crc)"
  
  it would print out the $ff00 decimal equivalent

  0
• Options

  jweather Posts: 320

  March 2009

  alexanbo wrote: »

  I think it is in itoa, because the calcuation was actually being done correctly, just when I tried to print out intemediary values with itoa that it got funky, sort of a Schr?dinger's Cat sort of thing.
  
  For example if you added a line before the print statement
  crc = $ffff << 8
  
  and changed the print statement to:
  send_string 0,"'Test1:',itoa(crc)"
  
  it would print out the $ff00 decimal equivalent

  
  Yes, because crc is an integer (16 bits).
  
  If you said
  

  STACK_VAR LONG_INTEGER test
  test = $ffff << 8 // or test = $ffff * 256
  

  
  then the result is $ffff00, because test is a 32-bit integer.
  
  It's nothing to do with itoa aside from the fact that its argument is a LONG_INTEGER. The $ffff+1 test may be running into some weird signed/unsigned arithmetic as well.

  0