DAL: Fix batching when depiling unknown messages

What

Add a function to handle application messages and call this function both when the message is first received or when the message is unstacked from the unknown message list.

Why

To share code, and mainly because with current implementation the batch_accumulator function was not called on the previously unknown messages. This is a bug, since currently if a DAL node is lagging, it might consider some shards as "in the future" and stack them as unknown validity. When inspecting them after the catch-up, the function GS.handle_receive_message will output To_treat_in_batch(_) but the message will never be added in any batch since batch_accumulator was not called.