--- Log opened Thu Sep 29 00:00:34 2022
00:37 -!- tibo [~tibo@240d:1a:afa:4000:dc0d:ca3c:7de5:ce89] has joined #bitcoin-rust
07:17 -!- tibo_ [~tibo@240d:1a:afa:4000:6580:b811:9dd6:f4ba] has joined #bitcoin-rust
07:21 -!- tibo [~tibo@240d:1a:afa:4000:dc0d:ca3c:7de5:ce89] has quit [Ping timeout: 246 seconds]
07:56 -!- tibo [~tibo@fp5a950abc.tkyc110.ap.nuro.jp] has joined #bitcoin-rust
07:59 -!- tibo_ [~tibo@240d:1a:afa:4000:6580:b811:9dd6:f4ba] has quit [Ping timeout: 268 seconds]
09:00 -!- dunxen [~dunxen@2001:470:69fc:105::1:fec1] has quit [Quit: You have been kicked for being idle]
09:02 -!- afilini [~afilini@2001:bc8:1828:245::2] has joined #bitcoin-rust
09:02 < afilini> hey everyone, quick question: is there a reason why From<u32> is not implemented for Sequence and PackedLockTime?
09:30 < andytoshi> it may have been an oversight
09:31 < andytoshi> or it may have been that kixunil fights me whenever i try to use .into() instead of a method with a super-long name designed to re-educate the Bitcoin Core out of everyone
09:31 < andytoshi> ..yeah, there we go
09:32 < andytoshi> there's a `from_consensus` method which does exactly what you want
09:32 < andytoshi> please file an issue
09:32 < andytoshi> or you can pile onto https://github.com/rust-bitcoin/rust-bitcoin/issues/1298 where i made the same complained about sighash modes
09:35 < afilini> i would personally find useful to have From implemented because i'd like to use it in our descriptor macro
09:36 < andytoshi> can you file an issue and link to #1298 ?
09:36 < andytoshi> sorry to be snarky, it's not productive
09:36 < BlueMatt[m]> heh, I still think `Into` was a major mistake in the rust language.
09:36 < afilini> so whatever the user passes to the macro, i'll just call `Sequence::from` or `PackedLockTime::from` and i can accept a larger variety of types (it could be a sequence itself, or a u32, etc)
09:36 < andytoshi> BlueMatt[m]: interesting! because it exists at all, or because it's annoying to use?
09:36 < BlueMatt[m]> magic baddddddddd, code readability important :)
09:36 < andytoshi> i mean, it's explicit
09:36 < afilini> andytoshi: sure!
09:36 < andytoshi> i wouldn't mind if we required peopre to use `into::<u32>()` say
09:37 < BlueMatt[m]> it hides all relevant type information, and very often has many implementations.
09:37 < andytoshi> but at least the way to write this is obvious
09:37 < BlueMatt[m]> which means I find it highly unreadable
09:37 < andytoshi> and you don't have to go sifting through the docs reading the type signature of every function until you find the one that does the conversion for you
09:37 < BlueMatt[m]> yea, I guess if you _had_ to write out the type in order to use it I'd be okay with that.
09:37 < BlueMatt[m]> but at that point you might as well have a method thats explicit
09:37 < andytoshi> BlueMatt[m]: ok, that's fair that bare .into() is very hard to read
09:37 < andytoshi> BlueMatt[m]: no, because there's no standard name for explicit methods
09:38 < BlueMatt[m]> yea, fair point, though `to_*` or `into_*` are generally used
09:38 < BlueMatt[m]> if they were slightly more consistent or enforced at a rustc-warning level then it'd be pretty good imo
09:38 < andytoshi> ok, well then i have a separate fight which is that we keep sticking _consensus_ randomly into shit..
09:39 < BlueMatt[m]> heh, well that sounds like a rust-bitcoin-specific problem :p
09:39 < andytoshi> fair enough :P
09:39 < andytoshi> anyway afilini if you open an issue asking for From i will support you :P
09:39 < andytoshi> in fairness if we implement this, nobody is obligated to use it
09:41 < afilini> done! https://github.com/rust-bitcoin/rust-bitcoin/issues/1302
09:42 < afilini> yeah i think i agree with Matt about `into`, but `from` is very useful when you don't want to force users to do the conversion themselves
09:45 < afilini> i mean, now rust-bitcoin has like three different types for representing timelocks, and if you write a function you could just take a generic type that you can use to construct a LockTime for example
09:46 < afilini> i like having more types in general, it leads to less "accidents" with users confusing things and shooting themselves in the foot, but on the other end it needs to be easy to convert between these types, otherwise it just gets annoying imho
10:01 -!- pablomartin [~pablomart@82.180.147.233] has joined #bitcoin-rust
10:06 -!- pablomartin [~pablomart@82.180.147.233] has quit [Client Quit]
11:01 -!- andytoshi [~apoelstra@user/andytoshi] has quit [Quit: leaving]
11:10 -!- andytoshi [~apoelstra@user/andytoshi] has joined #bitcoin-rust
11:11 -!- andytoshi [~apoelstra@user/andytoshi] has quit [Client Quit]
11:17 -!- andytoshi [~apoelstra@user/andytoshi] has joined #bitcoin-rust
11:56 < andytoshi> btw i finally started a script which is running all the rust-bitcoin and miniscript fuzztests indefinitely on my server. if any stop i will receive a text message
11:56 < andytoshi> i set the niceness of all of them to 19 so hopefully i don't notice it messing with my daily work
12:02 < BlueMatt[m]> Nice, chaincode has it running as well, after some initial honggfuzz hiccups
12:10 < RiccardoCasatta[> Did someone of you found a failure with fuzzing after a very long run? I found bugs with fuzzing but usually quite soon then nothing (but I never let run more than a couple of days)
12:18 < andytoshi> yeah, i've seen it happen
12:18 < andytoshi> btw our miniscript compile_descriptor fails with  and(hash256(),pk())
12:18 < andytoshi> we compile this, lift it, then recompile, and it gets reordered
12:18 < BlueMatt[m]> For deserialization targets, no, for more complicated stuff, yea, I’ve seen stuff after years of cpu time
12:19 < andytoshi> heh i don't think i've ever seen years, but definitely hours 
12:19 < BlueMatt[m]> Yea I’ve seen it after a month or more on like 40-100 cores
12:22 < RiccardoCasatta[> Woah, what was the target?
12:24 < BlueMatt[m]> LDK has some super deep complex targets that basically fuzz the full library top-to-bottom (not very well, though)
12:32 < andytoshi> btw i don't think the niceness is helping, my server feels really super slow
12:32 < BlueMatt[m]> You want the priority stuff
12:32 < BlueMatt[m]> Not nice
12:32 < BlueMatt[m]> chrt
12:41 < andytoshi> oh nieat
12:42 < andytoshi> ok, tried that, we'll see if things improve
12:43 < andytoshi> yeah i think that's better
--- Log closed Fri Sep 30 00:00:35 2022