Input/Ordering Block Implementation

[kushti]

Tests coverage todo:

• test that transactions in locally as well as remotely generated input blocks are cleared from the mempool
• test that transactions in locally as well as remotely generated input blocks are accounted for in the wallet (spent in input blocks boxes cant be spent again, boxes created in input blocks can be spent )
• test mempool and wallet processing on inputs chain fork (when support implemented)

[ross-weir]

Could be worth detailing weak transaction ids a little bit, why 6 bytes of hashes specifically?

[flysky9981]

when the sub-block time is 1 second, high probability to produce many orphan blocks, is it suitable to still use the longest chain rule? or consider using GHOST algo, greedy heaviest observed sub-tree, to choose the chain, but the security of this chain is also secured by ergo main-chain, so just a suggestion.