This is needed to restore proper CreateNewBlock functioning for unconfirmed transaction chains.
Added RPC tests and fixed an additional bug found with them. This should now be ready for merging to master and 0.12.
Weak concept NACK. I don't think we should be spending effort on maintaining the old priority system (which was completely arbirary anyway).
Merge #7149 bugfix_priority
Add GetStartingPriority method to reliably get the original priority …
…of a transaction in the mempool
Change GetPriority calculation to dynamically update
Track the value of inputs that get confirmed in the chain and keep a cached value of priority at a given height and return current priority by only assuming these in chain inputs are aging.
Add consistency check for on the fly priority calculations
QA: Add test for priority via getrawmempool
QA: Enable testing priority via getblocktemplate as well
QA: Extend txn_priority test to check across a reorganisation
Bugfix: Correct off-by-one errors in getrawmempool startingpriority a…
- Since the transaction arrives after block N, its starting priority is that for block N+1.
- The current priority is likewise as considered for mining the next block.
Bugfix: rpc-tests/txn_priority: When connecting disconnected nodes, t…
…heir mempools may not sync, so only sync blocks
CTxMemPool::check: Use height+1 for priority comparison, since height…
… is not guaranteed to work
Closing this due to inactivity.