Skip to main content

CPLX003-duplicate-code — Duplicate Code Block Detected

Dimension: Complexity Default severity: Medium Languages: All Last reviewed: 2026-05-06

What it detects​

Performs a project-wide sliding-window hash scan to identify blocks of identical non-blank, non-comment lines that appear more than once across any combination of files in the project.

Algorithm​

  1. For each parsed file, extract normalised non-blank, non-comment lines (stripping leading //, #, *, -- markers and whitespace).
  2. Slide a window of threshold lines over the normalised line list and compute a 64-bit hash per window.
  3. The first occurrence of each hash is recorded. Every subsequent occurrence at a different location emits one finding at that site, referencing the original location in the message.

Configuration​

[rules."CPLX003-duplicate-code"]
threshold = 6   # window size in lines; default 6

Why it matters​

Copy-pasted logic means bug fixes and refactors must be applied in multiple places. Missing even one copy creates a divergence that is hard to spot in code review and leads to subtle inconsistencies.

Example — flagged​

Two files (a.rs and b.rs) containing the same 6-line loop body will each produce a finding referencing the other file.

Fix guidance​

Extract the duplicated block into a shared function or module that both call sites can import.

Limitations​

  • Blank lines and comment-only lines are excluded from windows, so purely structural duplication (same structure, different comments) may be missed.
  • Hash collisions on 64-bit hashes are astronomically unlikely but theoretically possible.
  • Only cross-file detection is performed within a single project run; no historical comparison across scan sessions.

Implementation​