Skip to content


Repository files navigation


A tool for automatically checking the software quality of a C/C++ program and giving it a score. It is developed as part of my Hiwi project at the Heidelberg Institute for Theoretical Studies (HITS). The background for doing this is the appearent lack of quality in evolutionary biology software as found in this paper:

SoftWipe runs various checks to produce a software quality score: it compiles your program using clang and checks for compiler warnings; it activates almost all available warnings here. It runs your program with clang sanitizers activated, which detect memory errors and undefined behavior. It checks for the usage of assertions. It runs the following static code analyzers: cppcheck, clang-tidy, KWStyle, and lizard. Lizard is also used to find your programs average cyclomatic complexity and code duplication.

Installation & Usage

We highly recommend to use the conda package (see below) for a cleaner installation of the dependencies.

SoftWipe was mainly developed on Ubuntu 20.04 LTS. Some of the used code analysis tools might be affected on other OS versions. Clone this repository to download SoftWipe:

git clone


There are some Python package dependencies, which need to be installed before running SoftWipe:

And there are also some tool dependencies, but fear not! For macOS and Debian-based systems, SoftWipe can automatically install the following dependencies for you. Just run SoftWipe (sudo ./ - sudo is most likely required for installing the dependencies); if a dependency cannot be found, it will ask you whether it should automatically install it. Note for Mac users: Homebrew must be installed for the automatic installation of dependencies.

The following tools must be installed:

For make-based projects, you also need:

Also, if the program you want to check uses make or CMake as build system, make and CMake must be installed respectively.


There is a conda package for SoftWipe, which can be installed using:

conda install softwipe -c angtft -c conda-forge

On the first execution of SoftWipe, it will install Infer locally (if Infer is enabled), as there is no conda package for this tool yet. We recommend installing SoftWipe in a separate conda environment, which can be done with:

conda create --name softwipe_env
conda install softwipe -c angtft -c conda-forge -n softwipe_env

Then, before running SoftWipe, you need to activate the environment with:

conda activate softwipe_env

Close the environment again with:

conda deactivate

Basic usage

SoftWipe can be run as follows: [-c | -C] [-M | -m | -l target [target ...]] [-e EXECUTEFILE] programdir


-c tells SoftWipe that your program is C, and -C tells SoftWipe that you program is C++.

-M tells SoftWipe to build your program using CMake, -m using make, and -l using clang. The -l option takes as arguments the files to compile.

-e specifies the path to an "execute file" which contains a command line that executes your program.

programdir specifies the root directory of your target program.

Example command line for a CMake-based C++ program: -CM path/to/program -e path/to/executefile

For more options and further information, run --help.

Docker usage

Docker enables an awesome way to use a out-of-the-box installation by calling

docker run -it --rm -u <USER_ID>:<GROUP_ID> -w /work -v $PWD:/work softwipe/softwipe <SOFTWIPE_ARGS>

where USER_ID/GROUP_ID should be the owner of the host directory to have write permissions within the docker container. The same image can also be used for continuous integration, shown here for the example of a Jenkins pipeline:

stage('Softwipe') {
  agent {
    docker {
      image 'softwipe/softwipe:0.1'
  steps {
    sh ' -CM -e . 2>&1 |tee softwipe_general.txt'


A. Zapletal, D. Höhler, C. Sinz, A. Stamatakis (2021) The SoftWipe tool and benchmark for assessing coding standards adherence of scientific software Sci Rep 11, 10015 (2021).