CLI Usage

中文

dup-code-check is a Rust CLI binary: it parses arguments, calls the Rust core, formats output, and decides exit codes.

Basic usage

dup-code-check [options] [root ...]

• root ...: directories to scan; defaults to current working directory when omitted
• supports -- to terminate option parsing (useful when a root starts with -)

Examples:

dup-code-check .                 # default: duplicate files
dup-code-check --code-spans .    # suspected duplicate code spans
dup-code-check --report .        # all detectors in one report
dup-code-check -- --repo         # root starts with '-' (use --)

Modes

1) Default: duplicate files

dup-code-check [root ...]

Outputs “duplicate file groups” (each group contains 2+ files).

2) --code-spans: suspected duplicate code spans

dup-code-check --code-spans [root ...]

Outputs “duplicate span groups” (each group contains 2+ occurrences with line ranges).

3) --report: report mode

dup-code-check --report [root ...]

Runs multiple detectors and outputs a consolidated report (useful for manual review or CI artifacts).

Output formats

• text (default): human-friendly
• JSON: --json for machine-readable output
• stats: --stats adds scanStats in JSON; prints to stderr in text mode

See Output & Report for a full field reference.

Flags reference

Flags apply across default / --code-spans / --report, but some only affect specific detectors (see Scan Options).

Behavior switches

• --localization <en|zh>: set help/text output language (default en; JSON output is unchanged)
• --report: run all detectors and output a report
• --code-spans: find suspected duplicate code spans (with line ranges)
• --json: JSON output
• --stats: scan stats (stderr in text; scanStats in JSON)
• --strict: non-zero exit code if scan was incomplete
• --cross-repo-only: only output groups spanning >=2 roots
• --no-gitignore: do not respect .gitignore (default: respect)
• --gitignore: explicitly enable .gitignore (mainly useful in scripts)
• --follow-symlinks: follow symlinks (default: off)

Thresholds & limits

• --min-match-len <n>: minimum normalized length for --code-spans (default 50)
• --min-token-len <n>: minimum token length for token/block/AST-ish detectors (default 50)
• --similarity-threshold <f>: similarity threshold 0..1 (default 0.85)
• --simhash-max-distance <n>: SimHash max Hamming distance 0..64 (default 3)
• --max-report-items <n>: max items per report section (default 200)

Scan budgets

• --max-files <n>: stop scanning after reading n files (scanStats.skippedBudgetMaxFiles > 0 indicates the budget was hit)
• --max-total-bytes <n>: skip files that would exceed total scanned bytes budget
• --max-file-size <n>: skip files larger than n bytes (default 10485760 = 10 MiB)
• --max-normalized-chars <n>: stop scanning once stored normalized code chars would exceed n (scanStats.skippedBudgetMaxNormalizedChars > 0)
• --max-tokens <n>: (report) stop scanning once stored tokens would exceed n (scanStats.skippedBudgetMaxTokens > 0)

Ignore rules

• --ignore-dir <name>: ignore directory name (repeatable)

Help

• -h, --help: show help
• -V, --version: show version

Exit codes

• 0: completed successfully (even if some non-fatal skips happened: NotFound/TooLarge/Binary)
• 1:
  • runtime error (e.g. root does not exist / is not a directory, scan failures)
  • with --strict: scan was incomplete due to PermissionDenied, outside_root, relativize_failed, traversal errors, bucket truncation, or budget limits (maxFiles/maxTotalBytes/maxNormalizedChars/maxTokens)
• 2: argument parsing error (unknown flags, non-integers for integer flags, etc.)