E2E test "panicked at 'error on call `sign_and_submit_then_watch_default`"

[HCastano]

If you try and instantiate or call a contract message from different E2E test using the
same signer you end up running into the following error:

❯ cargo +nightly test --manifest-path ./examples/flipper/Cargo.toml

running 4 tests
test flipper::tests::it_works ... ok
test flipper::tests::default_works ... ok
test flipper::e2e_tests::e2e_first_instantiation ... FAILED
test flipper::e2e_tests::e2e_second_instantiation ... ok

failures:

---- flipper::e2e_tests::e2e_first_instantiation stdout ----
thread 'flipper::e2e_tests::e2e_first_instantiation' panicked at 'error on call `sign_and_submit_then_watch_default`: Rpc(ClientError(Call(Custom(ErrorObject { code: ServerError(1014), message: "Priority is too low: (318 vs 318)", data: Some(RawValue("The transaction has too low priority to replace another transaction already in the pool.")) }))))', /Users/hcastano/Workspace/ParityTech/ink/crates/e2e/src/xts.rs:259:17
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace


failures:
    flipper::e2e_tests::e2e_first_instantiation

test result: FAILED. 3 passed; 1 failed; 0 ignored; 0 measured; 0 filtered out; finished in 0.17s

error: test failed, to rerun pass `--lib`

However, if you change the signer in one of the tests then everything passes.

I'm not sure if the error lies in the E2E tests, or in subxt, but either way I imagine
there needs to be some better nonce management here.

This can be reproduced with ink@7789caf, cargo-contract@3e24722, and
substrate-contracts-node@0.21.0 using the following tests with Flipper:

#[cfg(test)]
mod e2e_tests {
    type E2EResult<T> = std::result::Result<T, Box<dyn std::error::Error>>;

    #[ink_e2e::test]
    async fn e2e_first_instantiation(
        mut client: ink_e2e::Client<C, E>,
    ) -> E2EResult<()> {
        let constructor = flipper::constructors::default();
        client
            .instantiate(&mut ink_e2e::alice(), constructor, 0, None)
            .await
            .expect("Instantiate `flipper` failed");

        Ok(())
    }

    #[ink_e2e::test]
    async fn e2e_second_instantiation(
        mut client: ink_e2e::Client<C, E>,
    ) -> E2EResult<()> {
        // N.B, If we use `ink_e2e::bob()` this passes
        let constructor = flipper::constructors::default();
        client
            .instantiate(&mut ink_e2e::alice(), constructor, 0, None)
            .await
            .expect("Instantiate `flipper` failed");

        Ok(())
    }
}

[ascjones]

but either way I imagine there needs to be some better nonce management here

Indeed, possibly a race condition if the tests are running in parallel (i.e both tests use the same nonce). Possibly could prove that by setting the test threads to 1.

[HCastano]

Indeed, possibly a race condition if the tests are running in parallel (i.e both tests use the same nonce). Possibly could prove that by setting the test threads to 1.

Just doing that (-j=1) at the top level cargo test call doesn't help

[SkymanOne]

Tried to replicate this issue on ink! 5.0.0-rc

#[ink_e2e::test]
async fn e2e_first_instantiation(
    mut client: ink_e2e::Client<C, E>,
) -> E2EResult<()> {
    let mut constructor = FlipperRef::new_default();
    client
        .instantiate("flipper", &mut ink_e2e::alice(), &mut constructor)
        .submit()
        .await
        .expect("Instantiate `flipper` failed");
    
    Ok(())
}

#[ink_e2e::test]
async fn e2e_second_instantiation(
    mut client: ink_e2e::Client<C, E>,
) -> E2EResult<()> {
    let mut constructor = FlipperRef::new_default();
    client
        .instantiate("flipper", &mut ink_e2e::alice(), &mut constructor)
        .submit()
        .await
        .expect("Instantiate `flipper` failed");

    Ok(())
}

Can not see the problem to be persisted. Therefore, I am gracefully closing the issue