softdevteam/error_recovery_experiment

Name: error_recovery_experiment

Owner: Software Development Team

Description: null

Created: 2018-04-10 13:49:06.0

Updated: 2018-04-18 15:55:47.0

Pushed: 2018-05-24 16:38:15.0

Homepage: null

Size: 25

Language: Python

GitHub Committers

UserMost Recent Commit# Commits

Other Committers

UserEmailMost Recent Commit# Commits

README

Error recovery experiment

This experiment comes in two phases:

  1. Generate Blackbox data
  2. Run the experiments
Generate Blackbox data

You will need to ask the BlueJ developers for a Blackbox account. This comes with various conditions which you must adhere to.

Once you have done that, the scripts in the blackbox directory can be used to generate Java source files from user input. Run build_blackbox_data.sh on the Blackbox server and, many hours later, you will have a combos file (which is the list of pairs needed to regenerate the source files) as well as the Java source in a src_files directory. The latter must never be redistributed. However, the combos file is safe to redistribute.

Run the experiments

cd runner && ./build.sh will build the experiments. If files called runner/lrpar_Cargo.lock or runner/lrpar_rev_Cargo.lock are present, they will be used as Cargo.lock files for mf and mfref.

cd runner && ./run.sh will run the experiments. You need to have a directory src_files with all the relevant source files inside the runner/ directory. Output files will be called runner/*.csv.

Human friendly PDF and TeX files can be produced with cd runner && ./process.py.


This work is supported by the National Institutes of Health's National Center for Advancing Translational Sciences, Grant Number U24TR002306. This work is solely the responsibility of the creators and does not necessarily represent the official views of the National Institutes of Health.