Oddbean new post about | logout
▲ ▼
laanwj | 10 days ago (raw) | root | parent | reply | flag +3 
 yes-mind that the work that was just merged has been underway in three PRs, since at least 2017:

https://github.com/bitcoin-core/secp256k1/pull/448
https://github.com/bitcoin-core/secp256k1/pull/636
https://github.com/bitcoin-core/secp256k1/pull/1579

it took a long time because there's an unfortunate lack of reviewers for any change to secp256k1
i wish people would get involved more instead of criticizing and screaming "bitcoin core devs suck" from the sidelines
▲ ▼
mleku | 10 days ago (raw) | root | parent | reply | flag +1 
 i would very much like to spend a whole month reviewing that library and porting it to pure Go and inserting some SIMD instructions where relevant
▲ ▼
ChipTuner | 10 days ago (raw) | root | parent | reply | flag +1 
 Reviewers are not contributors, that's a project issue. It doesn't look like there is a lack of contributions. I don't know what the processes now, but I don't remember it being worth many people's time to become part of the project because if their red tape.

I'm not suggesting Bitcoin-Core developers suck as a whole, most are probably far more intelligent than myself. I'm saying they made a pretty silly mistake that MANY if not MOST other developers make and that took way too long to make it to release (that should have been a patch release). 

I also think criticism shouldn't be stifled for the sake of "if you can do it better, just do it yourself" doesn't fix the problem that everyone else depends on X project and they have obvious problems they aren't fixing etc. 

Could I have worded my original note better, probably, do I think I shouldn't have said anything at all. No.
▲ ▼
ChipTuner | 10 days ago (raw) | root | parent | reply | flag +2 
 At this point I'm considering vendoring the entire project for noscrypt. I'll have to review open PRs to see what else is missing I could fast-track.