Skip to content


Switch branches/tags

Name already in use

A tag already exists with the provided branch name. Many Git commands accept both tag and branch names, so creating this branch may cause unexpected behavior. Are you sure you want to create this branch?

Latest commit


Git stats


Failed to load latest commit information.
Latest commit message
Commit time

SPLC18 Challenge Case Documentation

Instructions for retrieving the target systems for the SPLC 18 challenge case by Paul Gazzillo, Ugur Koc, ThanhVu Nguyen, and Shiyi Wei.

See the call for challenge cases here:

See our paper describing the challenge here:

Benchmark Programs and Datasets

Easy - Variability Bugs Database (Abal et. al ASE14)

This is a database of 98 software bugs occurring only under specific configurations, i.e., variability bugs. To create this database, Abal et. al (ASE14) studied software repositories of Linux, Busybox, Marlin, and Apache Httpd Server.

The online database is in here: In each entry, there is a table providing detailed information of the bug, such as the bug type and specific configurations that are involved. To find the configurations that lead to these bugs, there are two afternatives you could use in your experiments: Simplified bug and Single function bug

In this challenge, you are asked to find configrations for each the variability bug location in Linux, Busybox, and Marlin. Here, we provide a processed version of this database under the vbdb directory. For these programs, we modified the code snippets under the simple directory to resolve some compilation issues and to make the bug easier to observe (by adding assertions etc.). Moreover, for these programs, we also include options_list.txt files which include list of options that appear in all bug reports of that program. You should use these lists as the configuration space.

The all_locations.txt file contain all bug locations in question in this challenge.


Once you come up with a configuration for a variability bug with your approach, you can test it by preprocessing the program with the configration flags that appear in the configration you generated + the -E flag. I.e.,

gcc -E -Denabled_config_option1 -Denabled_config_option2 ... simple_code_snippet.c

This will generate pre-processeed c file. In this c file, you should see the bug location. However, having this location does not necessarily mean the bug is actually in there (in some cases). To test this, you can actually compile the program

gcc -Denabled_config_option1 -Denabled_config_option2 ... simple_code_snippet.c -o a.out

Then, if running a.out results in the error mentioned in the database, this means you found a configuration which leads to the variability bug (in some cases, your config might differ from what is reported in the database).

Medium - Axtls

The latest version of Axtls is available:

To see list of available configuration options you can check the config/.config file after running make allyesconfig. An example .config file can be found under examples directory

Hard - Busybox

Buysbox website:

You can work with the version 1_28_stable which can be found in here:

To see list of available configuration options you can check the .config file after running make defconfig.

Ultimate - Linux Kernel

linux-stable-4.14.13 which can be found in here:

To see list of available configuration options you can check the .config file after running make allyesconfig.

Configuring the Target Systems

For the variability bugs database source files, gcc -E -D CONFIG_VAR1 -D CONFIG_VAR2 ... can be used to configure the source code with the C preprocessor.

To configure axTLS, BusyBox, and Linux, first run allyesconfig (or defconfig). Then replace the .config file with the one from your solution program's output and run make oldconfig. The system is now ready to be built with make under the new configuration.


The solution programs will be evaluated on an 8-core, 2.90GHz processor with 32GB of RAM running a Linux distribution. The challenge evaluation period is about three weeks, so solution that cannot finish in a substantially. Please let the evaluation committee know of any special computing resource requirements (special machinery, runs requiring more than a day, etc).

Useful Program Analysis Tools


KMax can be found in here:


KBuildMiner can be found iin here:


Contact authors of iTree paper (ICSE2014).


iGen can be found in here:


GOLEM is a part of a bigger toolchain called undertaker which can be found at:


Paul Gazzillo, Ugur Koc, ThanhVu Nguyen, and Shiyi Wei. 2018. Localizing Configurations in Highly-Configurable Systems. 2018.


Instructions for the SPLC 18 challenge case






No releases published


No packages published