BUG: Incorrect Result, No "Overflow" Error, in Compiled App (71239)
This article was previously published under Q71239
SYMPTOMS
The following code fails to give an "Overflow" error when compiled and
run as an application, and prints an incorrect value of -32768:
x% = 32767 + 1
print x%
while mouse(0)<> 1:wend
An "Overflow" error correctly occurs when the code is run in the
interpreter.
This problem occurs in Microsoft QuickBasic versions 1.00, 1.00a, and
1.00b for the Macintosh System. Microsoft is researching this
problem and will post new information here when it becomes available.
MORE INFORMATION
The happens because of "constant folding" (also known as "constant
arithmetic"). Constant expressions are evaluated at compile time and
replaced with the result. Thus, an expression such as x% = 5 + 5 is
folded by the compiler into a single assignment statement of the form
x% = 10.
Thus x% = 32767 + 1 is compiled as x% = -32768. In this case, the
maximum value for an integer is reached and the value gets rolled over
(the sign bit gets set and a negative number is received). The
expression x% = 32767 + 2 is translated to x% = -32767; the expression
x% = 32767 + 3 is translated to x% = -32766; and so on.
| Modification Type: |
Minor |
Last Reviewed: |
1/9/2003 |
| Keywords: |
kbbug KB71239 |
|