flaky/tests: full_client::frame::staking::tx_bond Transaction outdated

[lexnv]

Detected by: https://github.com/paritytech/subxt/actions/runs/6745755385/job/18338194383

--- STDERR:              integration-tests full_client::frame::staking::tx_bond ---
thread 'full_client::frame::staking::tx_bond' panicked at testing/integration-tests/src/full_client/frame/staking.rs:216:5:
assertion failed: `Err(Transaction(Invalid("Invalid transaction: Transaction is outdated")))` does not match `Err(Error::Runtime(DispatchError::Module(err)))`
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace

[niklasad1]

Is it possible that same nonce is used twice for instance by another test?
We are still using a single substrate process right to run all integration test on right?

[jsdw]

We are still using a single substrate process right to run all integration test on right?

Every test should spawn a new substrate instance at the mo!

Looking at the test, I can't see anything that should be suspect here; two tx's are submitted but we wait_for_finalized_success on each of them. Maybe there's some race in the UnstableBackend where the finalized block that the second TX uses isn't as uptodate as it should be by the time we return from the first one? def worth looking into :)