Dmitry Chestnykh
Dmitry Chestnykh
Thank you for your kind words! Since `Verify` returns a `bool`, distinguishing between errors will either require changing the API (so I'd do this in the major version with other...
Does it happen every time you try? Can't reproduce it on 10.12.4 (16E195), 2.6 GHz Intel Core i5, 8 GB 1600 MHz DDR3, even after 500 attempts. My guess would...
Thank you, this makes sense. I'll look into it.
Thanks, I'm working on supporting es modules, but without rollup as it's super slow for me. Question: how would browser resolve dependencies? For example, the chacha package has these imports:...
Also, we need to detect the system random number generator (Node or browser) at runtime, and we can't use asynchronous imports for Node's crypto in ES module.
I'm a strong believer in strict encoding/decoding. If the intention is to decode data with whitespaces, it should be explicitly handled by the user by removing them. Especially since this...
Agree, base32 would be great to have. Not sure when I'll have time to implement it, though.
@chrisveness I wouldn't do this, I think just accepting N is good enough and standardized between implementations, and my logN thing was a mistake, which I fixed in "[modern API](https://github.com/dchest/scrypt-async-js#modern-api)"...
>Colin Percival says “The parameter N must be a power of 2 greater than 1” (www.tarsnap.com/scrypt.html), so allowing any value of N would not seem a good idea for standardisation....
Yes, the reference implementation allows only powers of two.