felipecrv opened a new pull request, #2788:
URL: https://github.com/apache/arrow-adbc/pull/2788
Since the removal of the global driver lock, there is no risk of deadlocks
between the locks on the driver and on the database objects.
IIUC, this was why member functions on the database struct took a mutable
`self` parameter:
```rust
/// Initialize the given database using the loaded driver.
///
/// This uses `&mut self` to prevent a deadlock.
fn database_init(
&mut self,
mut database: ffi::FFI_AdbcDatabase,
) -> Result<ffi::FFI_AdbcDatabase> {
```
Requiring `&mut self` is overkill and requires another layer of locking at
the application side if multiple threads share a database instance (as they
should for connection pooling).
This doesn't mean we are passing a forged mutable pointer to the driver via
the FFI, the locking on the inner database ensures exclusive access to the
`FFI_AdbcDatabase`.
```rust
/// Initialize the given connection using the loaded driver.
fn connection_init(
&self,
mut connection: ffi::FFI_AdbcConnection,
) -> Result<ffi::FFI_AdbcConnection> {
let driver = self.ffi_driver();
let mut database = self.inner.database.lock().unwrap();
// ConnectionInit
let mut error = ffi::FFI_AdbcError::with_driver(driver);
let method = driver_method!(driver, ConnectionInit);
let status = unsafe { method(&mut connection, &mut *database, &mut
error) };
check_status(status, error)?;
Ok(connection)
}
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
