8ec5dc0dec
keyfork-mnemonic-generate: test ensure entropy floats around what we expect from birthday values
keyfork mnemonic recover should accept words in a randomized pattern
Resolved. Vec::dedup only removes sequential duplicates.
keyfork mnemonic recover should accept words in a randomized pattern
use std::io::Read;
use rand::prelude::*;
fn get_u32_os(file: &mut std::fs::File) -> u32 {
let mut bytes = [0u8; 4];
file.read_exact(&mut bytes).unwrap();
u32::from_le_by…keyfork mnemonic recover should accept words in a randomized pattern
The math says this should be 12%. snail's Python code says this should be 12%. My Rust code - even with OS RNG - says this should be 12%. And yet, for some reason, this test stops at 1%.
https:/…
keyfork mnemonic recover should accept words in a randomized pattern
Figuring out the math for the possibility a 24 word mnemonic has duplicate words, using some rough birthday-problem math (https://en.wikipedia.org/wiki/Birthday_problem#Calculating_the_probability)…
keyfork mnemonic recover should accept words in a randomized pattern
Assumption is that each word is unique but this is only average case, not every case
keyfork mnemonic recover should accept words in a randomized pattern
36 looks nicer than 35, which is the actual smallest number whose permutation count (factorial) is > 2^128
keyfork mnemonic recover should accept words in a randomized pattern
additional: 24 words only provides about 90 bits of entropy, 12 words even less. have the user type in 36 words regardless, which provides 36! permutations which is more than 2^128 / 128 bits of…
keyfork mnemonic recover should accept words in a randomized pattern