By: Stanley B. Emailed: 1728 times Printed: 2329 times
Using an uninitialized object is a common program error, and one that is often difficult to uncover. The compiler is not required to detect a use of an uninitialized variable, although many will warn about at least some uses of uninitialized variables. However, no compiler can detect all uses of uninitialized variables.
Sometimes, we're lucky and using an uninitialized variable results in an immediate crash at run time. Once we track down the location of the crash, it is usually pretty easy to see that the variable was not properly initialized.
Other times, the program completes but produces erroneous results. Even worse, the results can appear correct when we run our program on one machine but fail on another. Adding code to the program in an unrelated location can cause what we thought was a correct program to suddenly start to produce incorrect results.
The problem is that uninitialized variables actually do have a value. The compiler puts the variable somewhere in memory and treats whatever bit pattern was in that memory as the variable's initial state. When interpreted as an integral value, any bit pattern is a legitimate valuealthough the value is unlikely to be one that the programmer intended. Because the value is legal, using it is unlikely to lead to a crash. What it is likely to do is lead to incorrect execution and/or incorrect calculation.
Be the first one to add a comment