A few code changes were needed to accompany the test:
* Adding setObjectName() calls for a few Qt controls to make them easily
accessible from the test.
* Calling contextMenu->popup() instead of contextMenu->exec() to open
the transaction list context menu without blocking the test thread.
* Opening the context menu at the contextualMenu event point rather than
the cursor position (this change was not strictly needed to make the test
work, but is more correct).
* Updating the bumped transaction row with showTransaction=true instead of
false. This is needed to prevent the bumped tx from being hidden, so the last
part of the test which attempts to bump the bumped tx can work. (Technically
this change is a more general bugfix not limited to the testing environment,
but the bug doesn't happen outside of the testing environment because in the
full Qt client, a queued NotifyTransactionChanged notification causes the row
to be updated twice, first with showTransaction=false, then immediately after
with showTransaction=true.)
```
$ git blame src/policy/fees.cpp | grep becuase
3810e976 (2017-03-07 11:33:44 -0500 789) * checks for 2*target becuase we are taking the max over all time
$ git blame src/policy/fees.h | grep successfullly
2d2e1705 (2017-04-12 12:29:03 -0400 54) * representing that a tx was successfullly confirmed in less than or equal to
$ git blame src/wallet/feebumper.cpp | grep "hasen't"
a3878374 (2017-05-11 09:34:39 +0200 258) // make sure the transaction still has no descendants and hasen't been mined in the meantime
```
Previously if we didn't have any local addresses, GetLocalAddress would return
0.0.0.0 and then we'd swap in a peer's notion of our address in AdvertiseLocal,
but then nServices would never get set.
Motivation for moving these is to make supporting IPC simpler (#10102), so
these lookups can be one-shot IPC requests, instead of back-and-forth
interactions over the IPC channel.
Also these functions are potentially useful outside of the bitcoin GUI (e.g.
for RPCs).
Start importwallet rescans at the first block with timestamp greater or equal
to the wallet birthday instead of the last block with timestamp less or equal.
This fixes an edge case bug where importwallet could fail to start the rescan
early enough if there are blocks with decreasing timestamps or multiple blocks
with the same timestamp.
An off-by-one-block bug in importmulti rescan logic could cause it to return
success in an edge case even when a rescan was not successful. The case where
this would happen is if there were multiple blocks in a row with the same
GetBlockTimeMax() value, and the last block was scanned successfully, but one
or more of the earlier blocks was not readable.
Verbose is changed to an int. This can have values from 0-2 for each level of verbosity.
Verbosity level 2 has transaction details displayed in the results.
For the per confirmation number tracking of data, introduce a scale factor so that in the longer horizones confirmations are bucketed together at a resolution of the scale. (instead of 1008 individual data points for each fee bucket, have 42 data points each covering 24 different confirmation values.. (1-24), (25-48), etc.. )
Store in fee estimate file the block span for which we were tracking estimates, so we know what targets we can successfully evaluate with the data in the file. When restarting use either this historical block span to set valid range of targets until our current span of tracking estimates is just as long.
Track the first time we seen txs in a block that we have been tracking in our mempool. Used to evaluate validity of fee estimates for different targets.
Change the logic of estimateSmartFee to check a 60% threshold at half the target, a 85% threshold at the target and a 95% threshold at double the target. Always check the shortest time horizon possible and ensure that estimates are monotonically decreasing. Add a conservative mode, which makes sure that the 95% threshold is also met at longer time horizons as well.
Pieter Wuille