vigna / cprng Goto Github PK

I'm trying to run mspec on the infamous RANDU number for fun and kicks, but it's just not cooperating. All of the commands in comp.sh throw the same error, so I will just post one:

g++ -std=c++17 -O3 -march=native spect.cpp -o spect -lntl

Here g++ is an alias to clang++. Anyways, it does the following:
out.txt

g++ --version gives:

Apple clang version 15.0.0 (clang-1500.1.0.2.5)
Target: x86_64-apple-darwin22.6.0
Thread model: posix
InstalledDir: /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin

Reading the error shows that there's some indecision between uint64_t and unsigned long. While on x64 mac both long and long long are 64 bits, the typedef goes to unsigned long long. Because NTL only gives conv in unsigned long, the compiler just won't take it.

I, uh, just threw unsigned long explicitly into the template arguments in common.cpp. Then it just worked:

template <> ZZ NTL_NAMESPACE::conv<>(const uint128_t& x) {
	return conv<ZZ, unsigned long>(uint64_t(x >> 64)) << 64 | conv<ZZ, unsigned long>((uint64_t&)x);
}

template <> uint128_t NTL_NAMESPACE::conv<>(const ZZ& x) {
	return uint128_t(conv<unsigned long>(x >> 64 & conv<ZZ>(0xFFFFFFFFFFFFFFFF))) << 64 | conv<unsigned long>(x & conv<ZZ>(0xFFFFFFFFFFFFFFFF));
}

template <> string hex<>(uint64_t a) {
	return hex(conv<ZZ, unsigned long>(a));
}

Now I'm happy. I don't know, should I complain to the guys at NTL?

Alright, there is one more issue in search.cpp, line 149. Again, specify the type as:

			a = ((((conv<ZZ>(0) + next()) << 192) + ((conv<ZZ>(0) + next()) << 128) + ((conv<ZZ>(0) + next()) << 64) + conv<ZZ, unsigned long>(next())) & multiplier_mask) | multiplier_surround_bits;

and it works.