Summary
StacklessEngine::execute_function_step (kiln-runtime/src/stackless/engine.rs:11524) is a dead pub fn with zero callers anywhere in the workspace (verified: grep execute_function_step outside its own definition matches nothing — not even tests). It is also a fabrication stub: it validates the function index, then returns zero-filled results without executing anything.
// Simulate step execution - in real implementation would execute instructions
// For now, return completed with default values
...
for result_type in &func_type.results {
let default_value = match result_type {
ValueType::I32 => Value::I32(0),
ValueType::I64 => Value::I64(0),
...
_ => Value::I32(0),
};
results.push(default_value)?;
}
This is the same fabrication family as SR-53 (#443, the arg-arity zero-fill, now fixed) and #412 — but for results, and unlike SR-53 it is currently not reachable (no live caller), so no user is receiving fake results from it today. That is the only reason this is low-severity rather than a correctness bug.
Why it should go
Two CLAUDE.md rules apply directly:
- NO DEAD CODE — "If code is unused and has no clear immediate purpose → DELETE IT." A
pub fn with zero callers, no test, and a // For now placeholder body qualifies.
- Stub methods that return Ok(0)/Ok(()) without doing anything are BANNED — this returns zero-filled results as if it executed.
Leaving it is a latent trap: a future caller wiring up "step execution" would silently get fabricated results (✓ completed with all-zero outputs), exactly the class of bug SR-53/#412 were about — but with no arity signal to catch it.
Suggested fix
Delete execute_function_step (and any now-unused helpers it alone pulls in). If step-wise execution is genuinely planned, replace the body with Error::not_implemented_error(...) and add a tracking note — do not leave a stub that returns fabricated success.
Found by a fable subagent while fixing SR-53 (#443); reachability (zero callers) confirmed by grep on main (6e249068).
Summary
StacklessEngine::execute_function_step(kiln-runtime/src/stackless/engine.rs:11524) is a deadpub fnwith zero callers anywhere in the workspace (verified:grep execute_function_stepoutside its own definition matches nothing — not even tests). It is also a fabrication stub: it validates the function index, then returns zero-filled results without executing anything.This is the same fabrication family as SR-53 (#443, the arg-arity zero-fill, now fixed) and #412 — but for results, and unlike SR-53 it is currently not reachable (no live caller), so no user is receiving fake results from it today. That is the only reason this is low-severity rather than a correctness bug.
Why it should go
Two CLAUDE.md rules apply directly:
pub fnwith zero callers, no test, and a// For nowplaceholder body qualifies.Leaving it is a latent trap: a future caller wiring up "step execution" would silently get fabricated results (
✓ completedwith all-zero outputs), exactly the class of bug SR-53/#412 were about — but with no arity signal to catch it.Suggested fix
Delete
execute_function_step(and any now-unused helpers it alone pulls in). If step-wise execution is genuinely planned, replace the body withError::not_implemented_error(...)and add a tracking note — do not leave a stub that returns fabricated success.Found by a fable subagent while fixing SR-53 (#443); reachability (zero callers) confirmed by grep on
main(6e249068).