[Bug] Acknowledge might fail with ResultAlreadyClosed during unload

### Search before asking

- [X] I searched in the [issues](https://github.com/apache/pulsar-client-cpp/issues) and found nothing similar.

### Version

main

### Minimal reproduce step

Enable `ackReceipt` for https://github.com/apache/pulsar-client-cpp/blob/main/tests/extensibleLM/ExtensibleLoadManagerTest.cc

```c++
ConsumerConfiguration consumerConf;
consumerConf.setAckReceiptEnabled(true);
Consumer consumer;
Result consumerResult = client.subscribe(topicName, "sub", consumerConf, consumer);
```

### What did you expect to see?

The test passed

### What did you see instead?

```
2024-02-27 21:51:30.938 WARN [0x16b36b000] AckGroupingTracker:93 | Connection is not ready, ACK failed for [(14,0,-1,-1)]
2024-02-27 21:51:30.938 INFO [0x16b96f000] ExtensibleLoadManagerTest:145 | acked i: 0 81 ms
/Users/xuyunze/github.com/bewaremypower/pulsar-client-cpp/tests/extensibleLM/ExtensibleLoadManagerTest.cc:147: Failure
Expected equality of these values:
ackResult
Which is: AlreadyClosed
ResultOk
Which is: Ok
```

Note: I changed the log level of `AckGroupingTracker:93` to warn locally.

### Anything else?

When `HandlerBase::getCnx()` returns a null `weak_ptr`, it's retryable if the state is not `Closing` or `Closed`. We should schedule the flush again.

### Are you willing to submit a PR?

- [X] I'm willing to submit a PR!

Start with tests/extensibleLM/ExtensibleLoadManagerTest.cc, especially the ackReceipt setup and the failing assertion around line 147. Trace AckGroupingTracker and HandlerBase::getCnx() to understand the unload path and retry condition. Done means the test passes with ackResult equal to ResultOk instead of AlreadyClosed when the client unloads.

Written by the indexing model from the issue text.