Artifact for ICSE-22 submission "Striking a Balance: Pruning False-Positives from Static Call Graphs"

This artifact contains 3 compressed folders for the code, data and results that accompany the paper.

Code

The Readme in the code details the dependencies as well as the instruction on how to run the train and test-phases for the tool, and reproduce the main experimental results.

Data

This includes the following:

1) Pre-computed call-graphs

2) Pre-trained models for the cg-pruner

3) Train and test program lists

The benchmark set used is the publicly available NJR-1 dataset (https://zenodo.org/record/4839913#.YSluRS1h2Cg) 

Results

This contains detailed (i.e. per-benchmark) results for the paper, including the data points used to plot the graphs.