I'm not sure if it was always the case, or whether it signifies an underlying problem, but it is weird:
2016-04-03 10:34:10 keypool reserve 1
2016-04-03 10:34:10 keypool return 1
I'm not sure if it was always the case, or whether it signifies an underlying problem, but it is weird:
2016-04-03 10:34:10 keypool reserve 1
2016-04-03 10:34:10 keypool return 1
It loads a reserve key in order to check its timestamp. We should really provide a wallet function that computes that without that, though...
https://github.com/bitcoin/bitcoin/blob/0.12/src/wallet/wallet.cpp#L2569 -> It reserves a keypool key and returns it immediately (after getting its time). This is kind of ineffective because there are db operations included.
Why not accessing setKeyPool.begin().nTime while holding cs_wallet directly?
Why not accessing setKeyPool.begin().nTime while holding cs_wallet directly?
Good idea, but let's wrap that in a CWallet method.
The same happens when you use getnewaddress.
Well with getnewaddress you'd expect a keypool line to be reserved (but not returned).