- What languages does ClusterFuzz support?
ClusterFuzz definitely supports C, C++ when compiled with clang. It has also been tested with Rust and may work with other languages that can be compiled with an LLVM-based toolchain (e.g. Swift).
An LLVM-based toolchain is needed for full ClusterFuzz support for two reasons:
Because ClusterFuzz relies on LLVM sanitizers for detecting and identifying bugs.
Because use of libFuzzer and AFL in ClusterFuzz requires coverage instrumentation only LLVM can provide.
We strongly recommend using an LLVM-based toolchain. That said, you may be able to use some parts of ClusterFuzz with GCC or extend ClusterFuzz to support your use of GCC. However, if you do this, you are in unexplored territory and we can not support this. GCC may work for blackbox fuzzing without modifying ClusterFuzz since GCC supports ASan. Using GCC with AFL or libFuzzer would probably require considerable changes to ClusterFuzz to get working. In the end, it is probably easier to make whatever changes are needed to build with clang.
In theory, ClusterFuzz is language agnostic and can be extended to support fuzzing code written in any language. In practice though, it is less easy to find interesting bugs in memory safe languages so ClusterFuzz doesn’t have support for fuzzing most of them.