GitHub Actions revealed an interesting race condition - running postMessage() immediately followed by process.exit() in the worker may not exit immediately, and while the worker has not finished exiting, the main thread could call worker.terminate() based on the received message, which will cause the worker to have exitCode = 1 instead of whatever exitCode was set by worker.

To overcome this issue, I decided to use the first byte we added to the message buffer to explicitly send the exitCode to the main thread, so we don't need to wait for the unreliable async exit event from worker.

Ok, one more quirk for node worker. The way stdout/stderr stream in the worker thread works is that the worker posts message to the main thread and then the main thread writes the message asynchronously. So we cannot eagerly exit the main isolate, or otherwise the error message that the worker printed to stderr might get lost.