# REVIEW — review phase (Adversary)

## Verdict: `review(all): PASS` — 2026-06-16T00:00:00Z

Comprehensive cold-verification of the entire calculator from a fresh pull.
No findings. No VETOs.

---

## D1 — Full cold re-verify (all prior-phase DoD items)

**Method:** `git pull --rebase` into clean clone at work-adv, then manual
inspection of all source + re-execution of all verification commands from
STATUS-eval.md.

### Lexer (lex phase DoD)
- Integer token (`42` → `Token('NUMBER', 42)`, value is `int`): PASS
- Float token (`3.14` → float): PASS
- Leading-dot float (`.5` → 0.5): PASS
- Trailing-dot float (`10.` → 10.0): PASS
- All operators tokenized correctly: PASS
- Whitespace skipped: PASS
- `LexError` on bad chars (`@`, `$`, letters): PASS — error message includes char and position

### Parser (parse phase DoD)
- Precedence: `* /` bind tighter than `+ -` (verified via AST shape): PASS
- Left-associativity for same-precedence operators: PASS
- Parentheses override precedence: PASS
- Unary minus (single, double, in expressions): PASS
- `ParseError` on malformed input (trailing op, unclosed paren, empty, juxtaposed numbers): PASS

### Evaluator + CLI (eval phase DoD)
- `2+3*4` → 14, `(2+3)*4` → 20, `8-3-2` → 3, `-2+5` → 3, `2*-3` → -6: PASS
- True division: `7/2` → 3.5: PASS
- `EvalError` on division by zero (not bare `ZeroDivisionError`): PASS
- Whole-float formatting: `4/2` → prints `2`, `-4/2` → prints `-2`: PASS
- Non-whole float: `7/2` → prints `3.5`: PASS
- CLI exit 0 on valid, exit 1 + stderr on error: PASS
- No traceback on error: PASS

---

## D2 — Full suite green

```
python -m unittest discover -v
Ran 53 tests in 0.336s
OK
```

**53 tests, 0 failures, 0 errors.** PASS

---

## D3 — Cross-feature break-it probes

All probes run directly via CLI.

| Probe | Expected | Actual | Pass? |
|-------|----------|--------|-------|
| `-(-(1+2))` | `3` | `3` (exit 0) | PASS |
| `2+3*4-5/5` | `13` | `13` (exit 0) | PASS |
| `1 @ 2` (lex error) | error stderr, exit 1 | `error: unexpected character '@' at position 2`, exit 1 | PASS |
| `1/0` (eval error) | error stderr, exit 1 | `error: division by zero`, exit 1 | PASS |
| `(1+` (parse error) | error stderr, exit 1 | `error: unexpected end of input`, exit 1 | PASS |
| `( 3.5 * ( 1.0 + 0.5 ) )` | `5.25` | `5.25` (exit 0) | PASS |
| `6.0/2.0` (whole-float) | `3` | `3` (exit 0) | PASS |
| `--5` (double unary) | `5` | `5` (exit 0) | PASS |
| `---5` (triple unary) | `-5` | `-5` (exit 0) | PASS |
| `3*-2*-1` (unary in chain) | `6` | `6` (exit 0) | PASS |
| `(2+3)*(4-1)/3` | `5` | `5` (exit 0) | PASS |
| `1-(2-3)` (paren forces right) | `2` | `2` (exit 0) | PASS |
| `-7/2` (negative float) | `-3.5` | `-3.5` (exit 0) | PASS |
| `.5+.5` (leading-dot floats) | `1` | `1` (exit 0) | PASS |
| `5.*2` (trailing-dot float) | `10` | `10` (exit 0) | PASS |
| `((((5))))` (deep nesting) | `5` | `5` (exit 0) | PASS |
| `999999999999999999999*999999999999999999999` (big int) | large int | correct result (exit 0) | PASS |
| ` ` (whitespace only) | error, exit 1 | `error: empty input`, exit 1 | PASS |
| `` (empty string) | error, exit 1 | `error: empty input`, exit 1 | PASS |
| `+5` (unary plus, unsupported) | error, exit 1 | `error: unexpected token 'PLUS'`, exit 1 | PASS |

**No defects found.** Error propagation across lex→parse→eval is clean. All
error paths produce user-readable messages with no tracebacks.

---

## D4 — Findings cleared

**No findings to clear.** No standing VETOs.

---

## Summary

Every DoD item from every prior phase verified from cold state. Full suite
green (53/53). All D3 cross-feature break-it probes pass. Calculator is
correct and complete. Builder may write ## DONE to STATUS-review.md.