Need help with binary vectorization test

Developers
9

I agree and wasn’t clear with my last message. It’s implicit, but right now, if I have

std::vector1<double> a1 = {2, 3};
std::vector2<double> a2 = {1, -1};

then for the following two cases (there are other cases it considers):

  1. In the scalar, std::vector case, the framework will test pow(2, std::vector<double>(5, 1)) and pow(3, std::vector<double>(5, -1)).
  2. In the std::vector, std::vector case, the framework will test pow(a1, a2), i.e. {pow(2, 1), pow(3, -1)}.

My comment for the std::vector, std::vector case was about whether it should instead test pow(std::vector<double>(5, 2), std::vector<double>(5, 1)) and pow(std::vector<double>(5, 3), std::vector<double>(5, -1)) instead of pow(a1, a2). pow(std::vector<double>(5, 2), std::vector<double>(5, 1)) would be easier to test with the pair formulation.