CPR - CardioPulmonary Resuscitation
CPR: A new automated program repair technique based on concolic execution which works on patch abstraction with the sub-optimal goal of refining the patch to less over-fit the initial test cases.
Automated program repair reduces the manual effort in fixing program errors. However, existing repair techniques modify a buggy program such that it passes given tests. Such repair techniques do not discriminate between correct patches and patches that overfit the available tests and break untested but desired functionality. We attempt to solve this problem with a novel solution that make use of the duality of space exploration in Input Space and Program Space. We implemented our technique in the form of a tool called CPR and evaluated its efficacy in reducing the patch space by discarding overfitting patches from a pool of plausible patches. Similar to Cardio-Pulmonary Resuscitation (CPR) does to a patient, our tool CPR resuscitate or recover programs via appropriate fixes.
In this work, we therefore propose and implement an integrated approach for detecting and discarding overfitting patches by exploiting the relationship between the patch space and input space. We leverage concolic path exploration to systematically traverse the input space (and generate inputs), while systematically ruling out significant parts of the patch space. Given a long enough time budget, this approach allows a significant reduction in the pool of patch candidates, as shown by our experiments.
Workflow
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.
Build and Dependencies
We provide a ready-made container which includes all necessary envrionment set-up to deploy and run our tool. Dependencies include:
- LLVM 3.4
- KLEE 1.4
- Python 3.7
- Z3 SMT Solver
- MathSAT Solver
Build and run a container:
docker build -t cpr .
docker run --rm -ti cpr /bin/bash
Example
We provide several examples you can run to test our tool, all test cases are included in the 'tests' directory.
Run examples:
pypy3 CPR.py --conf=tests/bug-types/div-zero/div-zero-1/repair.conf
pypy3 CPR.py --conf=tests/bug-types/div-zero/div-zero-2/repair.conf
Documentation
Developers
- Ridwan Shariffdeen
- Yannic Noller
Contributors
- Sergey Mechtaev
License
This project is licensed under the MIT License - see the LICENSE file for details