--- Log opened Thu Feb 03 00:00:48 2022
00:36 < darosior> sanket1729, andytoshi: why do we set `ops_count_sat` for `multi` to the number of keys? Maybe i'm missing something but i think it should only be so if k > 16. Otherwise it should be set to `k`. https://github.com/rust-bitcoin/rust-miniscript/blob/b8d04accb1e5fcf02849113531783a860ef42bbc/src/miniscript/types/extra_props.rs#L225
00:39 < darosior> Arf, i conflated the sigop count and the op count, sorry for the noise.
00:40 < darosior> sipa: the same holds for my comment about CalcOps() yesterday: for a k > 16 the sigopcount will be set to 20, but the opcount (which we care about here) set to k.
05:16 <@sipa> @darosior The input isn't compatible anymore, we can't merge the seeds.
05:16 <@sipa> @darosior You're saying the code is correct, right?
05:16 < darosior> Yes
05:24 < darosior> sipa: did you have any failure?
06:04 <@sipa> No failures.
06:08 -!- enick_343 [~afilini-m@2001:bc8:1828:245::2] has joined ##miniscript
06:10 <@sipa> So, I'm currently a bit undecided about what fuzzing approach to use. There is a spectrum of possibilities. One extreme is e.g. having an encoding that uses a single specific byte for every fragment, which rejects any other byte, and basically defines kind of a binary encoding for miniscript. The other extreme is using an encoding where you carefully track which fragments are permitted in a particular context, and pick always one of those.
06:10 <@sipa> My code currently does something in the middle between those.
06:11 <@sipa> The advantage of going all-out on using a fixed encoding is that it's much easier to keep compatible with existing seeds.
06:25 < darosior> How would you approach composition/arguments in the binary encoding? Smth like a variable length "If this byte is `2`, which corresponds to pk_k, read the next byte as the index of the key in the test data array"?
06:25 < darosior> Is it substantially better than the current approach? I guess it formalizes it
08:50 -!- salvatoshi [~salvatosh@genymobile-2-6-86.fib.nerim.net] has quit [Ping timeout: 256 seconds]
14:33 -!- enick_343 [~afilini-m@2001:bc8:1828:245::2] has quit [Ping timeout: 250 seconds]
19:45 <@sipa> I implemented it, and reintroduced the ops counting / malleable signing bug... it has run for hours and not found it.
19:46 <@sipa> Will let it run overnight, but I think this is very significantly slower (it's expected that it's slower, but how much is hard to predict)
19:57 -!- midnight_ is now known as midnight
23:23 -!- salvatoshi [~salvatosh@genymobile-2-6-86.fib.nerim.net] has joined ##miniscript
--- Log closed Fri Feb 04 00:00:49 2022