--- Day changed Sun Nov 01 2020 02:25 -!- jonatack [~jon@134.19.179.203] has quit [Quit: jonatack] 02:32 -!- jonatack [~jon@88.124.242.136] has joined ##miniscript 02:41 -!- jonatack [~jon@88.124.242.136] has quit [Ping timeout: 260 seconds] 02:42 -!- jonatack [~jon@213.152.161.101] has joined ##miniscript 03:03 < sanket1729> shesek: this was a todo for a long time, but we never got along to do it 03:04 < sanket1729> Lightning Koala along raised another issue on github asking for support. I will support for it in the next week. 03:15 -!- shesek [~shesek@unaffiliated/shesek] has quit [Remote host closed the connection] 03:15 -!- shesek [~shesek@unaffiliated/shesek] has joined ##miniscript 03:16 < shesek> sanket1729, very cool, thanks! 03:56 < shesek> another thing that I noticed is performance. deriving 1500 child keys/addresses using ExtendedPubKey directly takes around 0.1-0.2 seconds, doing the same with an equivalent descriptor takes around 8 seconds (on my somewhat old machine). 03:56 < shesek> I ended up implementing a special-case optimization for simple ranged p2*pkh descriptors, using the underlying ExtendedPubKey directly without invoking the rust-miniscript descriptor mechanism. since these type of descriptors are very common, I wonder if it might be worth doing something similar in rust-miniscript itself? 03:56 < shesek> https://github.com/shesek/bwt/commit/6c741b21efda5a3bc11cbbd9f883ea8d4427312e 16:07 < sanket1729> shesek: sorry for the late response. That is indeed a performance bug, the API should have secp context as a parameter 16:16 < shesek> ah-ah, I did notice that to_public_key() initializes the context every time and wondered about the performance implications of that 16:17 < shesek> well, if that's the major contributor to the performance issue, then its great news. should be easily fixable 17:46 < shesek> okay, nice! with this optimization in place (as a quick & dirty hack via lazy_static), the time for generating 1500 addresses went down to 0.2-0.25 seconds, which is barely noticeable compared to direct ExtPubKey derivation 17:49 < shesek> well, there goes that special-case p2*pkh optimization I implemented, its not really relevant at all now. which is great, this will eliminate a bunch of extra code and complexity on my side