Problem Statement: Finding effective configurations for Csmith, a test generator for C compilers, to trigger more bugs in the compilers.
Insight: The bug history of the C compilers can provide insight about how to make the compiler fail.
Approach: Clustering features in code snippets appeared in the GCC bug reports, and using the centroids of the clusters to recommend
configurations for the test generator.
Md Rafiqul Islam Rabin, Mohammad Amin Alipour
University of Houston
K-CONFIG: Using Failing Test Cases to Generate
Test Cases in GCC Compilers
Most Related Literature
[1] Yang et al., "Finding and Understanding Bugs in C Compilers", PLDI 2011.
[2] Groce et al., "Swarm Testing", ISSTA 2012.
[3] Alipour et al., "Generating Focused Random Tests using Directed Swarm Testing", ISSTA 2016.
[4] Chen et al., "Learning to Prioritize Test Programs for Compiler Testing", ICSE 2017.
[5] …
Experimental Results
Csmith [1]
Software
Engineering
Research Group
http://www.uh.edu/serg/
K-CONFIG
Table 1. The number of statements, branches, and functions covered in test suites.
Table 2. Result for 6 hours test budget.
Fig. 2. Cross-entropy in the n-gram models for n ∈ {1 . . . 5}.
Authors
Md Rafiqul Islam Rabin
mrabin@central.uh.edu
Mohammad Amin Alipour
maalipou@central.uh.edu

Generate Test Cases for GCC Compilers (ASE-LBR 2019)

  • 1.
    Problem Statement: Findingeffective configurations for Csmith, a test generator for C compilers, to trigger more bugs in the compilers. Insight: The bug history of the C compilers can provide insight about how to make the compiler fail. Approach: Clustering features in code snippets appeared in the GCC bug reports, and using the centroids of the clusters to recommend configurations for the test generator. Md Rafiqul Islam Rabin, Mohammad Amin Alipour University of Houston K-CONFIG: Using Failing Test Cases to Generate Test Cases in GCC Compilers Most Related Literature [1] Yang et al., "Finding and Understanding Bugs in C Compilers", PLDI 2011. [2] Groce et al., "Swarm Testing", ISSTA 2012. [3] Alipour et al., "Generating Focused Random Tests using Directed Swarm Testing", ISSTA 2016. [4] Chen et al., "Learning to Prioritize Test Programs for Compiler Testing", ICSE 2017. [5] … Experimental Results Csmith [1] Software Engineering Research Group http://www.uh.edu/serg/ K-CONFIG Table 1. The number of statements, branches, and functions covered in test suites. Table 2. Result for 6 hours test budget. Fig. 2. Cross-entropy in the n-gram models for n ∈ {1 . . . 5}. Authors Md Rafiqul Islam Rabin mrabin@central.uh.edu Mohammad Amin Alipour maalipou@central.uh.edu