Tezcatl

Nahuatl for “mirror”. Tezcatl measures a C or C++ codebase and writes a code metrics baseline: lines
of code, cyclomatic complexity, Halstead measures, documentation coverage, imported test coverage,
and the include graph with its cycles and coupling. It reads the project’s own
compile_commands.json and parses every unit with libclang and the project’s real flags, so it
measures the code as compiled rather than as text.

One command writes the whole baseline:

tezcatl report -p build --root . --modules modules.txt --out metrics

It writes report.md for people, report.json for programs, every table as CSV, and the include
graph for Graphviz.

Where to go next

You want to Read
build Tezcatl and write a first report Getting started
know every option and output of report The report
import test coverage Coverage
know exactly what each figure means Metric definitions
see what a report looks like on a real system Sample reports
understand how Tezcatl is built and tested Architecture

How far it has been checked

Every figure is checked against an independent tool or a hand count before it is trusted.

Check Result
Earthworm, 929 C and C++ units 0 parse errors; 3.5 s at 16 threads, 23.7 s on one (Release build)
Complexity against lizard on Earthworm, in the same configuration 94.1% equal; 99.0% outside SQLite’s amalgamation, with the differences explained
Coverage against gcovr: libmseed built with gcc 14 lines, branches and functions equal in all 42 files
Coverage against gcovr: Tezcatl itself equal in all 85 files
The sabotage log 142 deliberate breaks of the code, each caught by a test

Tezcatl is MIT licensed. The source is on
GitHub.

Next Guides