ROBUST: ROS Bug Study

Installation

This project uses BugZoo for managing bugs, building Docker images and running tests in containers.

To setup the BugZoo environment, we suggest using a Python virtualenv created in a suitable location. In the following example instructions, the venv is placed in the home directory, but that is of course not required.

$ sudo apt install python3-dev
$ virtualenv --python=python3 $HOME/bugzoo_venv
$ source $HOME/bugzoo_venv/bin/activate
$ pip3 install bugzoo

Don't forget to source $HOME/bugzoo_venv/bin/activate in every terminal the bugzoo script needs to be available.

Usage

The ROBUST dataset can be added to BugZoo as either a remote source:

$ bugzoo source add robust https://github.com/robust-rosin/robust

or as a local source

$ cd path/to/robust-rosin
$ bugzoo source add robust .

After adding ROBUST as a BugZoo source, you can use bugzoo bug list to obtain a list of bugs in the ROBUST dataset (along with bugs from any other BugZoo datasets that you might happen to have installed on your machine).

$ bugzoo bug list
Bug                                             Program     Dataset    Source    Installed?
----------------------------------------------  ----------  ---------  --------  ------------
robust:b826eae                                  care-o-bot  robust     robust    No
robust:eed104d                                  kobuki      robust     robust    No
robust:ca23e58                                  ros_comm    robust     robust    No
robust:b4dc23c                                  tf2         robust     robust    Yes
...

Bug Installation

Now that you've added ROBUST as a source to your BugZoo installation, you can begin building containers for each of the bugs inside the ROBUST dataset by using the bugzoo bug build command, as shown below:

$ bugzoo bug build --force robust:ca23e58

where ca23e58 is replaced by the identifier (i.e., the SHA-8) of the bug that you wish to build. The --force option instructs BugZoo to attempt to rebuild the image even if it is already installed -- don't worry, thanks to Docker's caching mechanism, rebuilding takes a few seconds at most (if there have been no changes to the image).

To determine the current replication status of a particular bug, refer to the progress tracker.

Interacting with the bugs

To launch an interactive container for one of the bugs, execute the following:

$ bugzoo container launch robust:b4dc23c

where the robust:b4dc23c is replaced by the name of the bug.

The bugzoo container execute command can be used to perform headless interaction with the bugs. For instance, in the example below, the developer fix is applied to the source code, the package under test is rebuilt, and the test is executed.

$ bugzoo container execute robust:b4dc23c ./fix && ./build.sh && ./test.sh

Container Anatomy

Each container provided by this repository contains the following files, all of which are located at /ros_ws:

• build.sh: is used to (re-)compile the package under test (PUT). build.sh should exit with code 0 if the PUT was successfully built. Conversely if the PUT fails to build, build.sh may exit with any code other than 0.
• test.sh: provides a test script that tests for the presence or absence of the bug. For build-related bugs, test.sh simply calls build.sh. If the test passes and the bug is not detected, test.sh should produce an exit code 0. Failing the test may cause test.sh to return any exit code other than 0. The same behaviour should also apply to build-related issues; that is, test.sh should exit with 0 if the build was successful, and any other exit code if not.
• fix: switches the source code for the PUT to its fixed state.
• unfix: switches the source code for the PUT to its buggy state.
• fix.patch: provides the developer patch that was used to fix the bug.

Note that build.sh, fix.patch, fix, and unfix are automatically generated during the BugZoo build process. test.sh and any files related to testing are hosted by the directory for each scenario.

Acknowledgements

Supported by ROSIN - ROS-Industrial Quality-Assured Robot Software Components.
More information: rosin-project.eu

This project has received funding from the European Union’s Horizon 2020
research and innovation programme under grant agreement no. 732287.