block streamer should check max_head_age in addition to max_gap

[roadscape]

There was a case where a node was forked for several hours, stuck on the same block. When the endpoint recovered, hive-sync continued following blocks at a 3s interval but way behind head. Checking head block time would allow us to easily detect this condition. It should be treated the same way as the max-gap check; i.e. exit listen routine and return to fast-sync.