CPR-TOOL.GITHUB.IO

Given the buggy program, a repair budget (time limit, iteration count), the fault location(s), a user specification, the language components for the synthesis, a failing test-case and optionally, a set of initial functional(passing) test cases, CPR generate a refined set of patches that are less over-fitting. The user specification defines the desired behavior of the repaired program (in addition to satisfying the given test cases). If initial tests are available, we assume that at least one failing test is available, which our method seeks to repair, apart from making sure that the user provided specification holds for all paths traversed via concolic exploration. Finally, CPR produces a ranked set of patches based on the explored input space... We provide a replication package for our experiments with CPR on the selected benchmarks in a Docker container, which consists of setup scripts/configuration files to re-run the experiments. The Docker environment consists of all dependencies..