< jonatack>
any preference in naming for an assertion that an amount has the expected number of digits after the decimal? i'm thinking "assert_scale", "assert_fractional_part_size", "assert_mantissa_length", "assert_decimal_places", etc.
< jonatack>
going with "assert_scale" and an explanatory docstring for now
< ryanofsky>
phantomcircuit, to fix link errors you may need to move blockfilter.cpp from libbitcoin_server library to libbitcoin_util library in makefile
< ryanofsky>
wallet isn't linked against libbitcoin_server to prevent wallet code from trying to access node global variables
< ryanofsky>
but a lot of things in libbitcoin_server that are just utilities without server state don't really belong there
< bitcoin-git>
[bitcoin] MarcoFalke opened pull request #19185: [0.19] Fix GBT: Restore "!segwit" and "csv" to "rules" key (0.19...fix_gbt_buried) https://github.com/bitcoin/bitcoin/pull/19185
< bitcoin-git>
[bitcoin] MarcoFalke closed pull request #19185: [0.19] Fix GBT: Restore "!segwit" and "csv" to "rules" key (0.19...fix_gbt_buried) https://github.com/bitcoin/bitcoin/pull/19185
< jarthur>
Does the REST API use libevent goodness?
< fjahr>
jarthur: yes
< jarthur>
ty
< phantomcircuit>
eyyyy using a filter with fixed siphash parameters and 30 bits does a wallet rescan in 4.6 minutes
< sipa>
phantomcircuit: nice
< phantomcircuit>
sipa, i also realized last night i was still loading the blocks when i did the BASIC filter benchmarks so im redoing those correctly
< phantomcircuit>
i can tell from the heat of my laptop that it's much much slower though
< phantomcircuit>
sipa, there's some house keeping stuff to do in the ScanForWalletTransactions method that could be a separate pr and are trivial, should i do that?
< phantomcircuit>
im not sure what the thinking is on that right now
< phantomcircuit>
btw is master broken? make check "ERROR: ConnectBlock(): coinbase pays too much (actual=5010000000 vs limit=5000000000)"
< sipa>
phantomcircuit: when?
< phantomcircuit>
no it's just my branch that has nothing to do with the coinbase values
< phantomcircuit>
what
< phantomcircuit>
oh hmm findNextBlock has identical results for a reorg and a straight failure to load the block
< phantomcircuit>
so the wallet test is expecting FAILURE when im returning success because the reorg flag is set, but the reorg flag doesn't mean reorg it means reorg or the block couldn't be found
< phantomcircuit>
right there's no way to distinguish between the start block is invalid and the start block was reorg'd so this test is just bad