SimFix

I. Introduction of SimFix
II. Environment setup
III. Run SimFix Step-by-Step
IV. Evaluation Result
V. Structure of the project

I. Introduction

SimFix is an automatic program repair technique, which leverages exisiting patches from other projects and similar code snippets in the same project to generate patches. The following figure is the workflow of our approach.

$The workflow of this technique.\label{workflow}$

Mining Stage

mine repair patterns from existing open-source projects, after which we can obtain a set of frequent repair patterns. Those patterns can be reused for other repairing scenarios as well.

Repairing Stage

Fault Localization : obtain a ranking list of candidate faulty statements and extract corresponding code snippets.
Donor Snippet Identification : identify the similarity between faulty code snippet and each candidate similar code snippet by leveraging three similarity metrics, according to which we obtain a list of candidate similar snippets with decending order of similarity value.
Variable Mapping : establish the mapping relationship between variables in faulty and similar code snippets by leveraging similarity metrics and then obtain a mapping table, based on which the variables in the donor code snippet will be replaced with the corresponding variables.
Modification Extraction and Intersection : extract code modifications to faulty code snippet via AST (Abstract Syntax Tree) matching and differencing against the donor snippet, and then the frequent patterns from the mining stage will be used to take intersection with those modifications to further ruled out invalid ones.
Patch Generation & Validation : generate repair patches by applying extracted code modifications to the faulty code snippet with combining and ranking whose modifications, then using the test suite to validate the correctness of candidate patches until a correct patch found or timeout.

II. Environment

OS: Linux (Tested on Ubuntu 16.04.2 LTS)
Download and configure Defects4J running environment.
Configure the following path.
- DEFECTS4J_HOME="home_of_defects4j"

III. How to run

SimFix was traditionally developed as an Eclipse Java project, you can simply import this project to your workspace and run it as a common Java program. The main class is cofix.main.Main, and for the running option please refer to the Running Options.

Before running

unzip file sbfl/data.zip to sbfl/data : used for fault localization
using the command line provided by Defects4J to checkout a buggy version of benchmark for testing.
Example: defects4j checkout -p Chart -v 1b -w ${buggy_program_path}/chart/chart_1_buggy

NOTE : the path of the buggy version of benchmark have to be set as:

…/projectName/projectName_id_buggy => Example: /home/user/chart/chart_1_buggy

Step 1, Build The Project

Originally, SimFix was developed as an Eclipse Java Project, you can simply import the project into your workspace and the class cofix.main.Main is the entry of the whole program.

Step 2, Running Options

Our prototype of SimFix needs three input options for running.

--proj_home : the home of buggy program of benchmark. (${buggy_program_path} for the example)
--proj_name : the project name of buggy program of benchmark. (chart for the example)

--bug_id : the identifier of the buggy program. (1 for the example)

the option of --bugy_id supports multiple formats:
single_id : repair single bug, e.g., 1.
startId-endId : repair a series of bugs with consecutive identifiers, e.g., 1-3.
single_id,single_id,single_id : repair any bugs for the specific program, e.g., 1,5,9.
all : repair all buggy versions of a specific project, i.e., all.

 
xxxxxxxxxx
Usage: --proj_home=${proj_home} --proj_name=${proj_name} --bug_id=${bug_id}
Example: --proj_home=/home/user --proj_name=chart --bug_id=1
Another: --proj_home=/home/user --proj_name=chart --bug_id=1,4,8

OPTION 1 : run within eclipse.

From the Main class:
Run As→Run Configurations… →Arguments : set the above arguments as Program Arguments.

OPTION 2 : run using command line.

We also provide runnable jar file of SimFix in the home folder of the project i.e., simfix.jar.
set the home directory of the SimFix project as your correct path and then run as:
java -jar simfix.jar --proj_home=/home/user --proj_name=chart --bug_id=1

Step 3, Result Analysis

After finishing the repair, there will be two additional folders: log and patch.

log : debug output, including buggy statements already tried, patches and reference code snippet for correct patch generation.

patch : a single source file repaired by SimFix that can pass the test suite. In the source file, you can find the patch, which is formatted as (example of Chart_1):

 
xxxxxxxxxx
// start of generated patch
int index=this.plot.getIndexOf(this);
CategoryDataset dataset=this.plot.getDataset(index);
if(dataset==null){
return result;
}
// end of generated patch
/* start of original code
        int index = this.plot.getIndexOf(this);
        CategoryDataset dataset = this.plot.getDataset(index);
        if (dataset != null) {
            return result;
        }
 end of original code*/

IV. Evaluation Result

Totally, SimFix successfully repair 34 bugs among 357 bugs in Defects4j v1.0 benchmark with generating 22 plausible but incorrect patches. The following table and venn diagram are comparison with existing approaches. More details are presented in the sub-folder final.

The comparison with existing approaches.\label{compare}

Intersection among different approaches.\label{venn}

$The comparison with existing approaches.\label{compare}$	$Intersection among different approaches.\label{venn}$
The comparison with existing approaches.	Intersections.

V. Structure of the project

 
xxxxxxxxxx
  |--- README.md   :  user guidance
  |--- bin         :  binary code
  |--- d4j-info    :  defects4j information
  |--- doc         :  document
  |--- final       :  evaluation result
  |--- lib         :  dependent libraries
  |--- sbfl        :  fault localization tool
  |--- src         :  source code
  |--- test        :  test suite

ALL suggestions are welcomed.