Automating Quality Checks with CCCC-C and C++ Code Counter

How to Use CCCC-C and C++ Code Counter for Accurate Code Analysis

1) Install and run

• Install from your package manager (e.g., apt install cccc) or build from source (github.com/sarnold/cccc).
• Run on your project root:

  Code
  Copy CodeCopied
  cccc src/.c src/.cpp include/*.h –outdir=metrics_output 

• Default output: HTML site in .cccc (or the directory set via –outdir).

2) Key options to set

• –outdir=DIR — where reports go.
• –html_outfile=FILE — main HTML report filename.
• –lang=LANG — force language parsing (c, c++, java).
• –report_mask=HEX — control which metrics appear (useful for custom reports).
• –opt_infile/–opt_outfile — load/save option presets.

3) Important metrics CCCC produces

• Lines of Code (LOC) — module/file and total counts.
• Cyclomatic Complexity (McCabe) — function/module complexity; higher = harder to test/maintain.
• Fan-in / Fan-out — coupling indicators (higher fan-out often signals high responsibility).
• Object-oriented metrics (for C++): inheritance depth, class-level measures.
• Module counts and simple size metrics.

4) How to interpret results for accuracy

• Treat LOC as a size indicator only—normalize by module or feature.
• Use cyclomatic complexity thresholds (suggested practice):
  • 1–10: simple, low risk
  • 11–20: moderate complexity — review
  • 21+: high complexity — refactor or add tests
• Look for files with high LOC + high complexity + high fan-out — prioritize these for refactor/testing.
• Compare metrics trends over time (use saved DB file via –db_outfile) instead of single snapshots.

5) Integrate into workflow

• Add CCCC to CI (run on pull requests) and fail or warn on preset thresholds.
• Generate HTML output as an artifact for developer review.
• Store and diff cccc.db between runs to detect regressions.

6) Practical tips

• Use –lang when extensions are ambiguous.
• Exclude generated code or third-party libs from analysis to avoid noise.
• Combine CCCC output with unit-test coverage and code-review results for balanced decisions.
• For large projects, run in parallel over subtrees and merge results into a single outdir.

7) Troubleshooting

• If parsing fails on modern C++ idioms, try latest CCCC from GitHub or pre-filter problematic files.
• Use verbose/debug options (debug_mask) during config to see parser issues.

References: CCCC project on GitHub (sarnold/cccc) and its user guide/man page for option details.