Installer klee avec Homebrew, Nix

Liens du projet

• Page d'accueil
• Dépôt

Commandes et alias

• kleaver
• klee
• klee-exec-tree
• klee-replay
• klee-stats
• klee-zesti
• ktest-gen
• ktest-randgen
• ktest-tool

Historique du projet

KLEE was initially developed at Stanford by Cristian Cadar, Daniel Dunbar, and Dawson Engler and presented at OSDI 2008. The project site describes it as a symbolic execution engine built on LLVM, and the repository describes the two central pieces: the core symbolic virtual machine and a POSIX/Linux emulation layer for running bitcode programs with symbolic operating-system inputs.

The OSDI evaluation is the historical anchor for KLEE: it ran on GNU Coreutils, BusyBox, MINIX utilities, and the HiStar kernel, using generated tests to expose crashes and correctness problems in heavily used systems code. That paper established the shape of KLEE's long-term identity: a research tool that is still packaged as a practical command-line engine.

Historique d'adoption

KLEE has been adopted most visibly in research, program-analysis teaching, and systems-testing workflows. A 2020 journal article describes KLEE as a popular dynamic symbolic execution engine that began at Stanford and was later primarily developed and maintained by the Software Reliability Group at Imperial College London.

Package-manager adoption followed from that academic and systems-tooling footprint. The Homebrew formula ships the `klee`, `kleaver`, `ktest-tool`, replay, stats, and test-generation utilities, reflecting KLEE's role as a suite rather than a single binary.

Modes d'utilisation

Users compile C or C++ programs to LLVM bitcode, mark inputs symbolic, and run KLEE to explore feasible paths and emit concrete test cases. The POSIX runtime lets KLEE model command-line arguments, files, environment variables, and other parts of a Unix process environment, making it especially attractive for testing command-line utilities.

Pourquoi les passionnés de paquets s'y intéressent

KLEE matters to package nerds because it turns the package archive itself into test material: Coreutils, BusyBox, and MINIX utilities were not toy examples but ordinary low-level programs. It is also a good example of why some research artifacts become durable packages: the build may be specialized, but the command-line behavior is useful enough for distributions to preserve.

Chronologie

• 2008: OSDI paper presented KLEE as a symbolic execution tool for high-coverage tests on complex systems programs.
• 2008: The OSDI evaluation reported KLEE runs over more than 452 programs and serious bugs in Coreutils, BusyBox, MINIX, and HiStar.
• 2020: A journal article described KLEE's post-Stanford maintenance and its academic and industry community.
• 2025-12-23: GitHub releases page listed KLEE 3.2.

Related projects

• KLEE is related to LLVM, STP and other SMT-backed solver tooling, S2E, angr, CBMC, AFL-style testing workflows, and the POSIX utility suites used in its original evaluation.

Sources

• https://github.com/klee/klee
• https://klee-se.org/
• https://link.springer.com/article/10.1007/s10009-020-00570-3
• https://www.usenix.org/event/osdi08/tech/full_papers/cadar/cadar.pdf