More bestblock records in wallets #2682
Conversation
Write bestblock records in wallets: * Every 20160 blocks synced, no matter what (before: none during IBD) * Every 144 blocks after IBD (before: for every block, slow) * When creating a new wallet * At shutdown This should result in far fewer spurious rescans.
|
Mostly I think it's essential that we do a release with a forced rescan prior to doing a release that supports turning on pruning that will make rescans infeasible (or at least a slow network traffic generating event). |
|
Automatic sanity-testing: PASSED, see http://jenkins.bluematt.me/pull-tester/95c7db3dbf4012dd02e5f1f30e6d982341facaa6 for binaries and test log. |
|
ACK. Synced testnet chain while watching bestblock record in wallet.dat, it changed at the right times. |
|
ACK. In valgrind, tried three different wallets, including old ones with no bestblock record and a empty one. Tested starts and stops, and a reindex. Looks okay and does not rescan on a new wallet. |
More bestblock records in wallets
More bestblock records in wallets
4 participants
Write bestblock records in wallets:
This should result in far fewer spurious rescans.
To correct all previously missed rescans, we should probably also once change the bestblock record name (suggested by gmaxwell), to force a single rescan once for every wallet. As this is likely to be annoying for almost all, and most that have missed funds in the past have already done a full rescan already, this is perhaps controversial and not included in this pullreq.
Should close #2676 (but I haven't tested that).