sipa: about the CNetMessage serialization split: what do you think about using a CNetBuffer for keeping the buffer & hash state, and once complete, we convert it into a CNetMessage (pure command / payload)?
instagibbs: it really is true as far as I'm aware :(
instagibbs, #12134 adds a nice framework for writing upgrade tests, and i'd like it merged but others think it would complicate testing framework to support old rpc versions, and would prefer to check in binary wallet blobs
Hi, with functional tests why do I need to generate >100 blocks to have my wallet knowing about the funds ? `self.nodes.generate(1)` then `self.nodes.sendtoaddress(` fails while `self.nodes.generate(101)` then `self.nodes.sendtoaddress(` passes ? Is this a good usage or should I do something else ? Reading other functional tests some seems to use the 101 trick but others don't
darosior: maturity rule
coinbase outputs are unspendable until buried 100 blocks deep
so the wallet treats them as balance only when those blocks exist
Yes did not think to that, thank you for the quick answer !
oh nice scantxoutset is decently fast, gave it 100k p2pk descriptors and it finished in around 30 seconds.
jb55 yeah wallet batch pr from achow101 improved that
promag: I thought that slowness was due to saving keys to disk
promag: scantxoutset is completely unrelated to the wallet