openworldformat 0.5.0

//! SHA-256, hand-rolled — the one hash the format's entry identity
//! needs, kept in-crate so "serde-only, no engine required" stays
//! literally true on every target (spec/session.md's canonical-hash
//! rule names the algorithm; it doesn't get to name a dependency).
//!
//! Verified against the FIPS 180-4 test vectors in the tests below and
//! by the cross-language canonical-JSON golden in `tests/`. This is
//! content addressing, not cryptography: the threat is drift between
//! writers, not adversaries.

/// SHA-256 of `bytes`, as 32 raw bytes.
pub fn sha256(bytes: &[u8]) -> [u8; 32] {
    const K: [u32; 64] = [
        0x428a2f98, 0x71374491, 0xb5c0fbcf, 0xe9b5dba5, 0x3956c25b, 0x59f111f1, 0x923f82a4,
        0xab1c5ed5, 0xd807aa98, 0x12835b01, 0x243185be, 0x550c7dc3, 0x72be5d74, 0x80deb1fe,
        0x9bdc06a7, 0xc19bf174, 0xe49b69c1, 0xefbe4786, 0x0fc19dc6, 0x240ca1cc, 0x2de92c6f,
        0x4a7484aa, 0x5cb0a9dc, 0x76f988da, 0x983e5152, 0xa831c66d, 0xb00327c8, 0xbf597fc7,
        0xc6e00bf3, 0xd5a79147, 0x06ca6351, 0x14292967, 0x27b70a85, 0x2e1b2138, 0x4d2c6dfc,
        0x53380d13, 0x650a7354, 0x766a0abb, 0x81c2c92e, 0x92722c85, 0xa2bfe8a1, 0xa81a664b,
        0xc24b8b70, 0xc76c51a3, 0xd192e819, 0xd6990624, 0xf40e3585, 0x106aa070, 0x19a4c116,
        0x1e376c08, 0x2748774c, 0x34b0bcb5, 0x391c0cb3, 0x4ed8aa4a, 0x5b9cca4f, 0x682e6ff3,
        0x748f82ee, 0x78a5636f, 0x84c87814, 0x8cc70208, 0x90befffa, 0xa4506ceb, 0xbef9a3f7,
        0xc67178f2,
    ];

    let mut state: [u32; 8] = [
        0x6a09e667, 0xbb67ae85, 0x3c6ef372, 0xa54ff53a, 0x510e527f, 0x9b05688c, 0x1f83d9ab,
        0x5be0cd19,
    ];

    // Padding: a 1 bit, zeros, then the bit length as u64 big-endian.
    let bit_len = (bytes.len() as u64).wrapping_mul(8);
    let mut message = bytes.to_vec();
    message.push(0x80);
    while message.len() % 64 != 56 {
        message.push(0);
    }
    message.extend_from_slice(&bit_len.to_be_bytes());

    let mut w = [0u32; 64];
    for block in message.chunks(64) {
        debug_assert_eq!(block.len(), 64, "padding makes whole blocks");
        let block: &[u8; 64] = block.try_into().expect("whole blocks");

        for i in 0..16 {
            w[i] = u32::from_be_bytes([
                block[i * 4],
                block[i * 4 + 1],
                block[i * 4 + 2],
                block[i * 4 + 3],
            ]);
        }
        for i in 16..64 {
            let s0 = w[i - 15].rotate_right(7) ^ w[i - 15].rotate_right(18) ^ (w[i - 15] >> 3);
            let s1 = w[i - 2].rotate_right(17) ^ w[i - 2].rotate_right(19) ^ (w[i - 2] >> 10);
            w[i] = w[i - 16]
                .wrapping_add(s0)
                .wrapping_add(w[i - 7])
                .wrapping_add(s1);
        }

        let [mut a, mut b, mut c, mut d, mut e, mut f, mut g, mut h] = state;
        for i in 0..64 {
            let s1 = e.rotate_right(6) ^ e.rotate_right(11) ^ e.rotate_right(25);
            let choose = (e & f) ^ ((!e) & g);
            let temp1 = h
                .wrapping_add(s1)
                .wrapping_add(choose)
                .wrapping_add(K[i])
                .wrapping_add(w[i]);
            let s0 = a.rotate_right(2) ^ a.rotate_right(13) ^ a.rotate_right(22);
            let majority = (a & b) ^ (a & c) ^ (b & c);
            let temp2 = s0.wrapping_add(majority);

            h = g;
            g = f;
            f = e;
            e = d.wrapping_add(temp1);
            d = c;
            c = b;
            b = a;
            a = temp1.wrapping_add(temp2);
        }
        for (s, v) in state.iter_mut().zip([a, b, c, d, e, f, g, h]) {
            *s = s.wrapping_add(v);
        }
    }

    let mut out = [0u8; 32];
    for (chunk, word) in out.chunks_mut(4).zip(state) {
        chunk.copy_from_slice(&word.to_be_bytes());
    }
    out
}

/// SHA-256 of `bytes`, lowercase hex.
pub fn sha256_hex(bytes: &[u8]) -> String {
    let digest = sha256(bytes);
    let mut out = String::with_capacity(64);
    for byte in digest {
        out.push_str(&format!("{byte:02x}"));
    }
    out
}

#[cfg(test)]
mod tests {
    use super::*;

    #[test]
    fn fips_vectors() {
        // The three vectors every SHA-256 is judged by.
        assert_eq!(
            sha256_hex(b""),
            "e3b0c44298fc1c149afbf4c8996fb92427ae41e4649b934ca495991b7852b855"
        );
        assert_eq!(
            sha256_hex(b"abc"),
            "ba7816bf8f01cfea414140de5dae2223b00361a396177a9cb410ff61f20015ad"
        );
        assert_eq!(
            sha256_hex(b"abcdbcdecdefdefgefghfghighijhijkijkljklmklmnlmnomnopnopq"),
            "248d6a61d20638b8e5c026930c3e6039a33ce45964ff2167f6ecedd419db06c1"
        );
    }

    #[test]
    fn one_million_a_bytes() {
        // The long-message vector: 1,000,000 'a's, one multi-block hash.
        let expected = "cdc76e5c9914fb9281a1c7e284d73e67f1809a48a497200e046d39ccc7112cd0";
        let message = vec![b'a'; 1_000_000];
        assert_eq!(sha256_hex(&message), expected);
    }
}