Research And Design On Abstract Memory Modeling Technology Supporting Type Conversion

Path-oriented automated unit test can generate test cases and execute these test cases automatically, so it can improve the efficiency of software testing and reduce the cost of the test greatly. The traditional path-oriented automated unit test can generate the correct test cases for most program.But when type conversion occurs in one path, it can't generate correct test cases because of analysis precision of symbolic is not good enough, so it can't generate the correct symbol, memory model, and constraints, thus unable to generate the correct test cases for the program.To counter the problem of basic data type conversion, symbolic expression reconstruction scheme and constraint fragmentation scheme is proposed in this paper. Symbolic expressions reconstruction scheme rebuild the recursive derivation of symbolic expressions, adds the derivation of types and the representations of symbols and constraints that support type conversion operations on the basis of the original recursive formulas, designs the interval arithmetic rules of new symbolic expression to support the type conversion test case generation. Constraint fragmentation scheme designs the symbol and constraint slices of variables involved in type conversion. The test case values of the slice symbols are obtained by these constraints, and the test values of the slice symbols are assembled into the original symbols according to certain rules of the test case values.In order to solve the problem of complex data type conversion, this paper expands the traditional abstract memory model. On the basis of the original model, we add the semantic and constraint access to the case of pointer to array and array to pointer. Data type conversion of the constraint extraction algorithm is designed, the final complex data types for the conversion of the code to generate the correct test cases.