Function execute_with_retry 

pub async fn execute_with_retry<E: RedisCommandExecutor>(
    executor: &mut E,
    cmd: &RedisCommand,
    exchange: &mut Exchange,
    is_idempotent: bool,
    policy: &NetworkRetryPolicy,
) -> Result<(), CamelError>

Executes a command with retry on transient errors, using NetworkRetryPolicy.

This is the core retry logic extracted for testability.

• policy: reconnection policy (max attempts, backoff, jitter, etc.)
• is_idempotent: whether the command is safe to retry

Implementation note

Uses a manual retry loop calling NetworkRetryPolicy::should_retry and NetworkRetryPolicy::delay_for rather than the shared [retry_async] helper. Manual loop needed because: (a) executor.reconnect() must run before each retry attempt (not the first attempt), so retry_async’s “always invoke op” model doesn’t fit; (b) &mut executor / &mut exchange borrows cannot be re-borrowed through FnMut() -> async move { ... } — the Future returned by the closure holds the borrow past the closure body, which the borrow checker rejects. retry_async_cancelable has the same FnMut constraint so it is also excluded.