naps62
commented
8 months ago Allright. this seems to fix both issues for me, although it still has one less serious wrinkle.
nevermind. there's still a panic if I try to instantiate a ledger twice while the device is locked (1st call fails with Err(UnexpectedNullResponse). Second call panics at hid.rs:143
Other than that, there's still a less serious wrinkle. the code below, without the sleep, fails 80% of the time on the second call (Err, not panic). adding the sleep makes it work more consistently.
updated my reproduction repo to showcase this
#[tokio::main]
async fn main() {
// first call, works
let _ = dbg!(call_ledger().await);
// tokio::time::sleep(tokio::time::Duration::from_millis(10)).await;
// second call, if done without sleep first, fails with `hid_error is not implemented yet`
let _ = dbg!(call_ledger().await);
}
pub(crate) async fn call_ledger() -> Result<Ledger, LedgerError> {
let path: String = "m/44'/60'/0'/0/0".into();
Ledger::new(HDPath::Other(path), 1).await
}
including the sleep makes this work consistently. without the sleep, the 2nd call will fail ~80% of the time:
[src/main.rs:11] call_ledger().await = Ok(
LedgerEthereum {
transport: Mutex {
is_locked: false,
has_waiters: false,
},
derivation: Other(
"m/44'/60'/0'/0/0",
),
chain_id: 1,
address: ...........,
},
)
[src/main.rs:12] call_ledger().await = Err(
LedgerError(
NativeTransportError(
CantOpen(
HidApiError {
message: "hid_error is not implemented yet",
},
),
),
),
)
xJonathanLEI
prestwich
leaving this open so that @naps62 and @xJonathanLEI can check it before merge 👍