Data-Flow Analysis

The Data Flow Analysis requires a Control-Flow Analysis (precondition). The data-flow for a variable within a program path (class or method) is defined through the following actions:

d	Variable is defined and set
r	Variable is referenced
u	Variable is undefined

Valid combinations are:

• dr, rr
  no anomalies.
• dd, du
  decent anomalies.
• ur
  critical anomalies.

Advantages

• Is equal to intuitive developer tests
• Anomalies of error-prone program sequences can be easily identified
• Indicates, which source code location must be checked
• Evaluation in combination with compiler warnings shall be considered

Disadvantages

• Decent anomalies, identified by the Control Flow result in warnings and hints. They do not create external failures.

Example

Given the following error-prone pseudo code:

sort (int Min, int Max) {
    int Helper;
    if (Min > Max) {
        Max = Helper;
        Max = Min;
        Helper = Min;
    }
}

The example above provides the following characteristics regarding anomalies:

Min	dr, rr
Max	dr, dd
Helper	ur, du

The correct pseudocode must look liked illustrated below:

sort (int Min, int Max) {
    int Helper;
    if (Min > Max) {
        Helper = Max;
        Max = Min;
        Min = Helper;
    }
}