GMP - 在 mpz_t/mpz_class 中存储 64 位整数并取回 64 位整数

Question

我想将 64 位整数中的值分配给 mpz_class/mpz_t 变量，然后再取回 64 位整数。但是 GMP 只为 32 位和更低的整数提供此功能。

那么我如何将 64 位整数转换为 mpz_class/mpz_t 变量，反之亦然。 （有符号和无符号整数都需要它）

Answer 1

这可以通过mpz_import() and mpz_export() 函数来实现。
代码示例（在 LP64 数据模型上测试）：

using Type = int64_t;

// We start with some 64bit integer
const Type builtIn64 = std::numeric_limits<Type>::min();
std::cout << builtIn64 << '\n';

// Import the integer into the mpz_class
mpz_t mpzNum;
mpz_init(mpzNum);
mpz_import(mpzNum, 1, 1, sizeof(Type), 0, 0, &builtIn64);
if (builtIn64 < 0) {
    mpz_neg(mpzNum, mpzNum);
}
std::cout << mpz_class(mpzNum) << '\n';

// Export the mpz_t value to a buffer allocated by the function and given
// the word size, get also the number of words required to hold the value
const size_t wordSize = sizeof(Type);
size_t wordCount = 0;
void* outRaw = mpz_export(nullptr, &wordCount, 1, wordSize, 0, 0, mpzNum);

// Make sure that our integer type can still hold the value
if (wordCount == 1) {
    const Type out = *static_cast<Type*>(outRaw);
    std::cout << out << '\n';
}

// Free the allocated memory by mpz_export
void (*freeFunction)(void*, size_t);
mp_get_memory_functions(nullptr, nullptr, &freeFunction);
freeFunction(outRaw, wordCount * wordSize);

// Don't forget to free the allocated memory
mpz_clear(mpzNum);

---

LIVE DEMO