# netboost native code (unix / macOS).
#
# No build-host CPU detection and no SIMD flags are set here: the AVX kernel in
# boosting.cpp is compiled with a per-function target attribute and selected at
# runtime via __builtin_cpu_supports(), so the same binary runs correctly on
# heterogeneous clusters (AVX / non-AVX nodes) and on non-x86 architectures
# (e.g. ARM, which uses the portable scalar path).
#
# Rcpp and RcppParallel headers are provided through 'LinkingTo' in DESCRIPTION;
# no extra compile or link flags are required (RcppParallel uses its header-only
# TinyThread backend).
