"Separate subgraph" bug misfires on indirect Brillig call with return value

[michaeljklein]

Aim

Attempted to compile a program that calls an unconstrained function and asserts that the result is zero:

unconstrained fn foo<T>(_x: T) -> Field {
    0
}

fn bar<T>(x: T) -> Field {
    /// Safety: test
    unsafe { foo(x) }
}

fn main() {
    let x: Field = 0;
    assert(bar(x) == 0);
}

Expected Behavior

Expected the program to compile without errors or bug's.

Bug

nargo compile returns:

bug: Input to Brillig function is in a separate subgraph to output
  ┌─ src/main.nr:7:14
  │
7 │     unsafe { foo(x) }
  │              ------ There is no path from the output of this Brillig call to either return values or inputs of the circuit, which creates an independent subgraph. This is quite likely a soundness vulnerability
  │
  = Call stack:
    1. src/main.nr:12:12
    2. src/main.nr:7:14

However, the Brillig call is completely constrained:

• All arguments to foo are constrained (i.e. constant)
• foo's result is constrained to be zero

To Reproduce

Workaround

Yes

Workaround Description

Ignore bug warning

Additional Context

No response

Project Impact

Nice-to-have

Blocker Context

No response

Nargo Version

nargo version = 1.0.0-beta.1 noirc version = 1.0.0-beta.1+3a42eb5c68f9616f0ebe367c894f0376ba41e0ef (git version hash: 3a42eb5, is dirty: false)

NoirJS Version

No response

Proving Backend Tooling & Version

No response

Would you like to submit a PR for this Issue?

None

Support Needs

No response

[TomAFrench]

I think this is a case where the error message can be improved but we should still throw a bug/warning. While all of the values returned from the oracle call are fully constrained, it's in a separate subgraph from the rest of the circuit the unconstrained function isn't really doing anything.