CodeDweller/mishmash.cpp

// mishmash.cpp
//
// Copyright (C) 2019-2020 MicroNeil Research Corporation.
//
// This software is released under the MIT license. See LICENSE.TXT.
//
// Mishmash is a non-cryptographic hash optimized for short strings.

#include "mishmash.hpp"

namespace codedweller {

uint64_t masher::engine(
  const unsigned char* buffer, size_t length,
  uint64_t accumulator,
  const masher::PrimeSet& primes) noexcept {
    for(size_t index = 0; index < length; index++) {
        unsigned char byte = buffer[index];
        uint64_t accumulator1 = primes.select(accumulator) + byte;
        uint64_t accumulator2 = ~accumulator * primes.select(byte);
        uint64_t accumulator3 = accumulator >> (32 + ((byte & 0x1F) ^ (byte >> 5)));
        accumulator = accumulator1 + accumulator2 + accumulator3;
    }
    return accumulator;
}

uint32_t mishmash(const unsigned char* buffer, size_t length) noexcept {
    uint64_t accumulator = masher::engine(buffer, length);
    return static_cast<uint32_t>(accumulator & 0x00000000ffffffff);
}

uint32_t mishmash(const std::string& s) noexcept { 
    return mishmash((const unsigned char*) s.c_str(), s.length()); 
}

uint32_t mishmash(const std::vector<unsigned char>& v) noexcept { 
    return mishmash(v.data(), v.size()); 
}

} // end of namespace codedweller