SWIG VS Cython

How is this any different than a mature tool such as SWIG (https://www.swig.org/)?

I've used SWIG extensively with Python to call C code and import C headers for testing/tooling purposes.

For pure C, CGO. For C++ they are likely creating shims with Swig: https://www.swig.org/

I mostly write in C and C++ so for language bindings I use Swig a lot. Say Im creating a machine learning library in C++, its very easy to create a Python API that can call the C++ classes and methods using Swig. iirc, I am using the same swig interface file to create bindings for Python, OCaml, R and even Fortran. Feel free to DM me if you got any more questions or anything!

Swig can make c++ types and functions available to python.

The likes of SWIG is often used to link C library-like code to 11-ish other widely used languages.

But, you might find these links interesting: * https://github.com/swig/swig/issues/918 * https://github.com/kaby76/swigged.llvm

Or take a look at the tools CppSharp or SWIG, s.a. Example for SWIG to wrap C++ library in .Net 6.

This sounds a bit like SWIG. You give SWIG an "interface file" describing the API of a library, and it uses that to spit out a bunch of language-specific bindings for that library.

and again, everything that was released after using an older version of cython.

Packages using cython (note the linked issues) due to this change.

Just for reference,

* Nuitka[0] "is a Python compiler written in Python. It's fully compatible with Python 2.6, 2.7, 3.4, 3.5, 3.6, 3.7, 3.8, 3.9, 3.10, and 3.11."

* Pypy[1] "is a replacement for CPython" with builtin optimizations such as on the fly JIT compiles.

* Cython[2] "is an optimising static compiler for both the Python programming language and the extended Cython programming language... makes writing C extensions for Python as easy as Python itself."

* Numba[3] "is an open source JIT compiler that translates a subset of Python and NumPy code into fast machine code."

* Pyston[4] "is a performance-optimizing JIT for Python, and is drop-in compatible with ... CPython 3.8.12"

[0] https://github.com/Nuitka/Nuitka

[1] https://www.pypy.org/

[2] https://cython.org/

[3] https://numba.pydata.org/

[4] https://github.com/pyston/pyston

Profile and optimize the hotspots with cython (or whatever the cool kids are using these days... It's been a while.)

Cython - - embed demo

There are also just-in-time compilers available for some Python features, that compile those parts to machine code. That includes Numba (usable as a library within CPython) and Pypy (an alternative Python implementation that includes a JIT compiler to improve performance). There’s also Cython, which is a superset of Python that allows more directly interfacing with C and C++ functions, and compiling the resulting combined code.

Conceptually, Cython is mainly for accelerating Python code, and can _also_ access C code. Meanwhile CFFI is specifically for calling C code and nothing else. I recommend the video for the differences.

One concrete thing that pops to my mind is that Cython doesn't support Py_LIMITED_API which means that you need to ship a lot more binary wheels. At least the issue is still open (https://github.com/cython/cython/issues/2542) and Cython projects IME need new wheels for each minor Python release. Compare that to cffi projects that (musl & pypy aside) only have to ship wheels for one Python version / architecture: https://pypi.org/project/argon2-cffi-bindings/#files

As you can see, in this benchmark Starlite handily beats even blakchseep, a notoriously fast ASGI framework written in Cython