trux-dtf / fl-vs-apr Goto Github PK

Hi,

According to this line Main.java#L19

System.out.println("Arguments: <Failed_Test_Cases_File_Path> <Suspicious_Code_Positions_File_Path> <Buggy_Project_Path> <defects4j_Path> <Project_Name> <FL_Metric>");

Main class takes as arguments:

• <Failed_Test_Cases_File_Path> which I think can be one of these files, right?
• <Suspicious_Code_Positions_File_Path>, what's this and where can I find it?
• <Buggy_Project_Path>, checkout directory of each bug, right?
• <defects4j_Path>, D4J home directory, right?
• <Project_Name>, e.g., Chart
• <FL_Metric>, which ones are supported?

--
Best,
Jose

Hi,

The README.md file mentions that kPAR requires Java-7, would it incorrect/wrong to run it under Java-8?

--
Best,
Jose

Hi.
I'm a PhD student and I want to run your code but the path you use in this project confuses me. Is the data you upload in github equals to your dataset?

Hi,

After I have read the whole project, which is really well-written and well organized, I have got a question in my mind: how to get FL results using GZoltar 1.6.0 jar.

It seems that the ways to use GZoltar 1.6.0 and 0.1.1 are quite different in this project. I mean:

• when using GZoltar 0.1.1, the project writes java code (see GZoltarFaultLoclaization.java) to pass parameters to GZolta 0.1.1 jar.
• but while using GZoltar 1.6.0, the project seems to use parseResults() function to read files from external folders (e.g., "GZoltar-1.6.0/crushMatrixOutput/").

So I would like to ask that are the fault localization results of GZoltar 1.6.0 version obtained from command line (as opposed to Java implementation) ?

Much appreciated if any guidance can be given.