Presented by Rebekah Leslie.

The existing implementation of DDT uses a depth-first search algorithm to drive the exploration of new paths for testing. This algorithm provides full coverage of the program under test, but is limited by the fact that the number of paths increases exponentially with the size of the program. By employing the control-flow graph information of the program under test, we can direct the testing process towards program paths that contain unvisited points and therefore obtain full branch coverage in a smaller number of tests than would be required by the original depth-first search algorithm. We will present two uses of control-flow graph information in DDT. The first use is a refinement of depth-first search where control-flow graph information is used to prune the search space to eliminate unnecessary tests. The second use is in the context of a prioritized work-queue that forms the basis for a variety of sophisticated search algorithms that exploit different heuristics.