This paper presents a novel approach for automated test data generation of imperative programs containing integer, boolean and/or float variables. It extends our previous work to programs with procedure calls and arrays. A test program (with procedure calls) is represented by an Interprocedural Control Flow Graph (ICFG). The classical testing criteria (statement, branch, and path coverage), widely used in unit testing, are extended to the ICFG. For path coverage, the specified path is transformed into a path constraint. Our previous consistency techniques, the core idea behind the solving of path constraints, have been extended to handle procedural calls and operations with arrays. For statement (and branch) coverage, paths reaching the specified node or branch are dynamically constructed. The search for suitable paths is guided by the interprocedural control dependences of the program. The search is also pruned by a new specialized consistency filter. Finally, test data are generated by the application of the proposed path coverage algorithm. A prototype has been implemented. Experiments show the feasibility of the approach.

The paper presents a novel approach for automated test data generation of imperative programs containing integer, boolean and/or float variables. Our approach is based on consistency techniques integrating integer and float variables. We handle statement, branch and path coverage criteria. Our purpose is thus to automatically generate test data that will cause the program to execute a statement, to traverse a branch or to traverse a specified path. For path coverage, the specified path is transformed into a path constraint which is solved by an interval-based constraint solving algorithm handling integer, boolean and real variables. A valid test input is then extracted from the interval solutions. For statement (and branch) coverage, a path reaching the specified statement or branch is dynamically constructed. Our algorithm for path coverage is then applied. The search for a suitable path and the solving of path constraints make extensive use of consistency techniques. We propose a simple consistency notion called eBox consistency, for generalizing box consistency to integer and float variables. The eBox consistency is sufficient for our purpose. A prototype has been developed and experimental results show the feasibility of our approach. This work is an extension of work by A. Gotlieb (2000) for float and boolean variables.