to mimic the instructions from RNP::Eigensystem in file Eigensystems.cpp  because MKL lapack cannot handle NULL pointer for the matrix 	RNP::Eigensystem(n2, NULL, n2, q, NULL, 1, phi, n2, work_, NULL, lwork); //M.P

please edit for the library locations first

fixed bug of undeclared variable

Fix SpectrumSampler in Python extension

…ating systems

Added HPC compliant make file MakefileHPC

…assuming installation via Homebrew

…assuming installation via Homebrew (fix typo)

…ing parameters.

‘mp1’ is proportional to the ’Width’ option​. ’Jinc’ calculates J1(2 π z)/(π z) where z is the argument passed into the function. The Bessel function itself is 0 at z = 0, but because of the division by π z is 1 when the argument z is zero.​ ‘fr’ is calculated from the G-vectors describing the different diffraction orders in reciprocal space and is thus of the same order of magnitude as the G-vectors, which have an order of magnitude of 1/(length scale of unit cell), so if everything is specified in nm, a typical unit cell of e.g. 500 nm gives: 1/500 = order of 1 x 10-3. So in this case, the argument going into the Jinc function will always be << 1; the numerator (fr) is of the order 1 x 10-3, and is always divided by at least 1 because the denominator is ‘2*mp1 + 1’. This means the Bessel function always returns 1, or something extremely close to 1 (so the ‘power’ parameter, which just takes the calculated smoothing factor and raises it to the power specified, also does nothing, because 1 to any power is still 1). This meant that regardless of which options you choose to pass for LanczosSmoothing, nothing really happens, because all you are doing is pre-multiplying things by 1.​ In any case, the inclusion of the constant ‘1’ makes no sense to me because it clearly ruins the scale-insensitivity of the implementation. I “fixed” the problem by simply removing the ‘+1’. This seems like a minor change but it significantly affects the behaviour of the LanczosSmoothing parameters passed to S4.

remove lines causing implicit function declaration errors

Add Apple M1 makefile

checking which flags are necessary for compilation

…k to compiler

Fixed S4 build for macOS-arm64

* avoid variable name I in Interpolator.h and main_python.c