Draft: Use xtensor<char, N> instead of xtensor<int, N>
When XTensor has merged https://github.com/xtensor-stack/xtensor/pull/2708, we can replace the xt::xtensor<int, N> that were added in !1023 (merged) by
the smaller xt::xtensor<char, N>.
In the meantime !1484 (merged) updated PreFlagger so it uses uint8_t instead of int. We could consider doing the same in this MR.
Edited by Maik Nijhuis