net: announce our height in the handshake, and say when we are behind

[Bitflash-sh]

Follows #55. That one made a node notice a peer it cannot hear. This one lets it notice that it is behind everyone.

The gap

The version message carries no height. Bitcoin 0.1.0 never had the field and this fork inherited the gap, so a node has no idea where the rest of the network is. The only evidence of trouble is a number on screen that stopped moving — which looks exactly like a network with nothing to report.

That is the structural reason a broken node here has always been indistinguishable from an idle one. It is the shape of #42, of #51, and of the seed sitting 8 blocks behind for 27 minutes this morning with all three relays reachable the whole time.

What this does

nBestHeight goes on the end of the version message and lands in CNode::nStartingHeight.

• Last field, deliberately. A node predating this stops reading after addr and treats the rest as trailing bytes, which costs it one log line.
• Never required. Reading is guarded by vRecv.empty(); peers that do not send it stay at -1 and are left out of the median. Every release so far sends the short form.
• Median, not maximum, so one peer claiming an absurd height cannot move it.

Then it gets used: after five minutes more than one block behind the peer median, the node says so. In the log, not behind a debug category — an operator who already enabled -debug is not the one who needs telling — and in the GUI status area.

WARNING: 433 blocks behind the network (height 3364, peers report 3797) for 2 minutes

Testing

Two nodes on the live network, connected over .btf:

• the one starting from an empty datadir read its peer at 3797, warned at 2636, then 1274, then 433 blocks behind as it caught up, and went quiet on its own once level
• the other logged peer height = -1 for the older nodes it met on the real network — the compatibility path doing its job, unprompted
• no exceptions, no disconnects caused by the extra field

Known limit

nStartingHeight is a snapshot from handshake time and goes stale on a long-lived connection. So this catches a node that is receiving something but not blocks; it does not catch one that has gone completely deaf, because then every peer's remembered height is old too.

The inactivity timeout from #55 is what covers that case, and it refreshes these numbers when it forces the reconnect. Carrying the height in ping would make it live — worth doing, but on its own evidence rather than bundled here.