Runtime error in read-only function but test passes

[MarvinJanssen]

I am writing a unit test for a public function. Inside that public function I call into a read-only function that encounters a runtime error (failed unwrap-panic), however the test runner still reports the test as OK.

Example:

(define-map example uint bool)

(define-read-only (this-fails)
	(unwrap-panic (map-get? example u0))
)

(define-public (test-something)
	(begin
		(this-fails)
		(ok true)
	)
)

clarinet test output:

Readonly Contract call runtime error: ST1....contract-name::test-something() -> Runtime Error: Runtime error while interpreting ST1....contract-call: Runtime(UnwrapFailure, Some([FunctionIdentifier { identifier: "ST1....contract-name:test-something" }, FunctionIdentifier { identifier: "ST1....contract-name:this-fails" }, FunctionIdentifier { identifier: "_native_:native_unwrap" }]))
.file.ts => test-something ... ok (19ms)

[smcclellan]

@hugocaillard to investigate.

[hugocaillard]

Hey @MarvinJanssen

Currently, a runtime error in clarity won't make the test fail.

    const block = chain.mineBlock([
      Tx.contractCall(
        "contract",
        "will-panic", // will run time error
        [],
        wallet_1.address
      ),
    ]); // won't throw an error

So the test has to actually check the result of the call or the block height or something like assertEquals(block.receipts.length, 1).

Does that make sense?

Also we are currently thinking about revamping clarinet test so we could definitely consider changing this behavior (so that a clarity runtime error throws an error on the JS side)

[hugocaillard]

I'll close for now, it'll be considered as part of #1052 (and following issues)