--- Log opened Mon Feb 22 00:00:37 2021
03:05 -!- ghost43_ [~daer@gateway/tor-sasl/daer] has joined #utreexo
03:05 -!- ghost43 [~daer@gateway/tor-sasl/daer] has quit [Remote host closed the connection]
09:03 < dergoegge> adiabat: you said you will get rid of ProofPositions? i have started to look at hash to known for the c++ lib and i think that function is useful there. its also used for creating proofs.
09:03 < dergoegge> just curious if you remove it what you will use instead
13:04 -!- mdrollette [~mdrollett@cpe-70-123-125-237.tx.res.rr.com] has joined #utreexo
14:07 < adiabat> dergoegge: the new function I had taking the place of ProofPositions was proofNodes
14:08 < adiabat> the idea is instead of returning the positions (uint64) of everything in the proof up to the roots, it instead returns pointers to the nodes themselves
14:08 < adiabat> also it's a method on the pollard so it knows how far up to go, and doesn't keep going up to the roots if hashes are already there
14:09 < adiabat> so in the extreme case where everything is cached, it will just go through the leaves, returning leaf nodes
14:09 < adiabat> then it will check that they match the leaves being proven.  So the whole thing stays on the bottom row and never moves up
14:10 < adiabat> if you run it on a fully unpopulated pollard, then it should do the same as ProofPositions (except pointers instead of positions that then have tobe grabPos'd
14:11 < adiabat> .. then the hard / tricky part which I haven't gotten is what to do with extra proof hashes.  There are a bunch of options...
14:12 < adiabat> I think I will start with just server & client both using the same caching so they know each others state
14:12 < adiabat> which isn't very flexible but I can maybe start that way
14:17 -!- adiabat [~adiabat@63.209.32.102] has quit [Ping timeout: 264 seconds]
14:19 -!- adiabat [~adiabat@63.209.32.102] has joined #utreexo
--- Log closed Tue Feb 23 00:00:37 2021