Synchronous Pull

Based in [some comments in this PR](https://github.com/GoogleCloudPlatform/pubsub/pull/243) with @dpcollins-google, I opened this Feature request. Pub/sub provides an option to retrieve records synchronously: https://cloud.google.com/pubsub/docs/reference/rpc/google.pubsub.v1#pullrequest

It isn't the recommended way but it works well in some scenarios.

Pub/Sub Lite doesn't have this feature so if you need to do polling, some buffer is needed in the middle to simulate that behavior. That already happen in the [PubSub Lite kafka library](https://github.com/googleapis/java-pubsublite-kafka) and the Apache Beam connector. It will happen again in the Source kafka connector.

To avoid to repeat the same workaround in each place, it would be nice to provide that API as part of the Java SDK. It would help also to clients moving from PubSub Lite to PubSub and viceversa.

Start by reading the Pub/Sub PullRequest reference linked in the issue and the linked Pub/Sub Lite Kafka and Apache Beam implementations to understand the polling workaround. Define the Java SDK API and its behavior for synchronous retrieval, then verify that clients can use it without duplicating an intermediate buffer.

Escrito por el modelo de indexación a partir del texto del issue.