mcpp 1.0.0

McCabe++ (mcpp): cyclomatic complexity and other vulnerability-related code metrics

McCabe++ (mcpp)

mcpp measures typical code complexity metrics like McCabe's cyclomatic complexity.

The goal of this project is to provide a re-usable script to analyze C/C++ source code and extract complexity metrics from it. The implemented metrics are taken from the paper

LEOPARD: Identifying Vulnerable Code for Vulnerability Assessment through Program Metrics

This tool is released as part of our research in vulnerability discovery and has been used in our paper

SoK: Where to Fuzz? Assessing Target Selection Methods in Directed Fuzzing"

See also the corresponding repo.

Complexity Metrics

Dimension	ID	Metric Description
CD1: Function	C1	cyclomatic complexity
CD2: Loop Structures	C2	number of loops
	C3	number of nested loops
	C4	maximum nesting level of loops

Vulnerability Metrics

Dimension	ID	Metric Description
VD1: Dependency	V1	number of parameter variables
	V2	number of variables as parameters for callee function
VD2: Pointers	V3	number of pointer arithmetic
	V4	number of variables involved in pointer arithmetic
	V5	maximum number of pointer arithmetic operations a variable is involved in
VD3: Control Structures	V6	number of nested control structures
	V7	maximum nesting level of control structures
	V8	maximum number of control-dependent control structures
	V9	maximum number of data-dependent control structures
	V10	number of if structures without else
	V11	number of variables involved in control predicates

Setup

Build a docker container which performs the setup automatically or run the installation on your local machine:

pip install .

Note: It is recommended to install packages in virtual environments.

Usage

From Python

Simply import mcpp and then use the extract function (or one of its variants).

from pathlib import Path
from mcpp import extract

input_dir = Path("some/dir")
in_files = list(input_dir.glob("**/*.c"))
result = extract(in_files)

# to extract only a subset of the metrics
result = extract(in_files, ["V1", "C3"])

# full list of metrics:
from mcpp import METRICS
print(list(METRICS.keys()))

CLI

Configuration parameters can be changed in config.yaml or directly on the CLI with e.g. mcpp paths.out_root=some/dir.

Using all defaults:

mcpp                # with default params like input directory, see config.yaml

Changing params from command line:

mcpp in_path=/some/dir/single_source out_path=single_source_metrics.json
mcpp metrics=\[C1,C2,V4\]

Or by passing a changed config.yaml:

• -cp (config_path) specifies the absolute path to the directory where the config file is located
• -cn (config_name) specifies the name of the config file

mcpp -cp /some/other/dir -cn myconfig.yaml

Try out the example:

mcpp in_path=examples/data/source paths.out_root=examples/data-out
cat examples/data-out/complexity.json