Further modernization of the build backend to reduce `setup.py` usage

Creating an issue to capture the internal discussions.

Currently both cuda.bindings and cuda.core use setuptools as the build backend and have a mixture of pyproject.toml static declarations and setup.py dynamic logics. We have two choices here:

• Keep setuptools but further reduce the amount of code in setup.py:
  • This seems to become possible with recent setuptools, where Cython extensions can now be declared in pyproject.toml.
• Switch to use scikit-build-core:
  • This has the benefit of reusing CUDA discovery mechanism (FindCUDAToolkit) and other nice build system features provided by CMake (see the relevant discussion in #46).
  • The potential downside is to maintain a custom CMakeLists.txt which is arguably a burden for non-C++ (=C & Python) oriented developers. Our Cython extensions are very simple to build.
    • However, this might become necessary once we start integrating some features from cudax::, e.g. #528.

Empieza comparando las configuraciones actuales de pyproject.toml y setup.py para cuda.bindings y cuda.core; después, lee la discusión enlazada en #46. Evalúa si las versiones recientes de setuptools pueden declarar las extensiones de Cython o si se necesitan scikit-build-core y un CMakeLists.txt, incluidas las consideraciones futuras sobre cudax:: en #528. Se considerará terminado cuando el proyecto acuerde una dirección para el backend de compilación y su alcance.