[Bug] moneybatcher sample race condition?

What are you really trying to do?

I am trying to learn how to do batching using temporal. I found the moneybatch sample and I wanted to make sure that it can handle requests in parallel.

To simulate the behavior, I changed the workflow to withdraw more times than a single batch size as well as added a delay during the deposit call.
What this lead to is that the workflow would continue to run and not complete, it won't even process a single batch.

Describe the bug

This is related to the condition in the workflow:
Workflow.await(() -> count == batchSize);

If count suddenly goes from 0 to a value greater than batchSize, the workflow will continue to run and not complete.

I am new to temporal, so I am not quite sure how to fix this. Is this a bug in the sample or am I doing something wrong?

Minimal Reproduction

Download this branch:
https://github.com/derwasp/samples-java/tree/derwasp/moneybatcher

Either run ./run-moneybatch-demo.sh and observe the logs.
Or run the test TransferWorkflowTest and see the failure.

Environment/Versions

• OS and processor: M4 Mac
• Temporal Version: (CLI version 1.4.1)
• Are you using Docker or Kubernetes or building Temporal from source? Docker

Additional context

Commencez par le moneybatch sample dans core/src/main/java/io/temporal/samples/moneybatch et exécutez TransferWorkflowTest depuis la branche de reproduction liée. Examinez la condition de Workflow.await et la configuration du test pour les requêtes dépassant la taille d’un batch. Le travail est terminé lorsque le workflow reproduit s’achève et traite les batches attendus au lieu de rester bloqué.