< bitcoin-git>
[bitcoin] hebasto opened pull request #15863: scripts and tools: Ensure repos are up-to-date in gitian-build.py (master...20190421-update-repos) https://github.com/bitcoin/bitcoin/pull/15863
< harding>
I'm playing around with getblockfilter and something is confusing me. IIUC, the number of elements in the set is prefixed to the filter using a normal compactSize uint. I see this e.g. for blocks 1 to 169 which all start with 0x01, presumably the sole output for each generation tx. Block 170 is the first block with a non-generation tx; for that I see 0x03. That confuses me; I think the
< harding>
filter is supposed to contain four elements: 1. the generation tx's vout0 scriptPubKey (sPK), 2. the regular tx's vin0 prevout sPK, 3. reg tx's vout0 sPK, reg tx's vout1 sPK. Does anyone know what I'm misunderstanding about BIP158? bitcoin-cli getblockfilter $( bitcoin-cli getblockhash 170 )
< sipa>
harding: there can be overlaps between them?
< harding>
sipa: indeed there can, but AFAICT there's not for block 170.
< harding>
sipa: I take that back.
< harding>
sipa: there is overlap, block170 tx1 vout1 sends its change back to the spending address.
< echeveria>
that's unusual. I don't think bitcoin.exe supported that.
< harding>
echeveria: me neither; that's why I discounted it.