Steven Roose

I might be interested in working on this. Especially with segwit, witness data should be prunable.

@huyoulong Could you perhaps provide some unit tests that test for this behavior?

What happens when a node has multiple mining addresses? Round Robin? Or random?

Idea: `ScanUnspentOutputSet`. Cfr the way Core implemented `scantxoutset`.

Yeah this is based on the gRPC implementation that I worked on in the past. I'm not doing much work on btcd anymore, but the gRPC impl is quite finished...

Thanks for the credits :) Cool that you managed to do a proper constant time, I struggled with it at that time! This change would be compatible with https://github.com/btcsuite/btcd/pull/621, right?

Also, you seem to have forgotten to run gofmt. It's always a good idea to run the `goclean.sh` command before committing, the Travis here runs it as well.

Wait until Schnorr arrives, we're gonna need you here! :)

I tried to sync testnet with this PR and got this: ``` 00:16:15 2017-01-22 [INF] BMGR: Rejected block 00000000036334b7b4272c4d3fc1a72e07159c7938859d1580aca99807232243 from 127.0.0.1:18333 (outbound): failed to validate input 95ec9f464983a63f393f3e7b20a974668213ad76b816206343718a13d9ce897a:0 which references output...

@davecgh I wanted to fix some of the remarks you made on the code in order to get this PR included. It seems like it could be a good performance...