godedup

Find structurally duplicate functions in Go code.

godedup detects copy-pasted Go functions even when variable names, literals, and package calls were changed.

---

Example

godedup ./...

$ godedup --output=table --no-tests
 
GROUP  TYPE   SIM   FUNCTION                     LOCATION                     STMTS  LINES
------------------------------------------------------------------------------------------
1      EXACT  100%  store.(*UserRepo).findByID   internal/store/user.go:84    5      14
1      EXACT  100%  store.(*OrderRepo).findByID  internal/store/order.go:91   5      14
------------------------------------------------------------------------------------------
2      EXACT  100%  api.writeJSON                internal/api/user.go:31      4      9
2      EXACT  100%  api.writeJSON                internal/api/order.go:28     4      9
------------------------------------------------------------------------------------------
3      NEAR   88%   worker.(*EmailJob).validate  internal/worker/email.go:55  7      19
3      NEAR   88%   worker.(*SMSJob).validate    internal/worker/sms.go:48    8      21

---

How It Works

godedup hashes every function's AST structure, normalizing away:

• Identifier names - userID and orderID are treated the same
• Literal values - "users" and "orders" are treated the same
• Package qualifiers - fmt.Println and log.Println are treated the same

Two functions that do the same thing with different variable names, different string literals, and different package calls hash identically.

Exact clones - top-level hash collision: identical control flow and structure.

Near clones - edit distance on per-statement hash sequences: same structure with a few added/removed statements.

---

Install

go install github.com/hashmap-kz/godedup@latest

brew tap hashmap-kz/homebrew-tap
brew install godedup

---

Usage

godedup [flags] [path ...]

Flags:
  --min-similarity  float    minimum similarity threshold (default: 0.85)
  --min-stmts       int      minimum statements to analyze (default: 3)
  --exact                    report only exact structural clones
  --no-tests                 exclude test files
  --output          string   output format: text, table, json (default: text)
  --version                  print version

CI integration - exits with code 1 if any clones found:

- name: Check for duplicate code
  run: godedup --exact ./...

JSON output for custom tooling:

godedup --output=json ./... | jq '.[] | select(.exact) | .functions[].name'

---

Normalization Rules

What	How normalized
Variable names	all -> same token
String literals	all -> ""
Numeric literals	all -> 0
Package qualifiers	fmt.X and log.X -> same structure
nil / true / false	preserved (semantic meaning)
Operator types	preserved (+ != -)
Statement order	preserved (order matters)
Control flow structure	preserved (the whole point)

---

Limitations

godedup is intentionally simple. It does not try to prove semantic equivalence.

It finds structural duplication, not all possible duplication.

For example, these may not match:

• same behavior implemented with different control flow
• helper functions extracted in only one copy
• loops rewritten as recursion
• code generated by tools, unless you include it in the scanned paths

---

License

MIT License. See LICENSE for details.