For computationally complex operations, often its impractical to write some things in pure python. But having python bindings to act as glue, or to integrate with a not-so-computationally-intensive code is helpf. Thats the usecase for building a rust library and generating a python module using PyO3.
Is it in the spirit of temper to eventually (long in the future Im sure) have an alternative build path for python that permits python-friendly bindings but with compiled/native code? I think the python+rust story is common, but you could generalize this to any kind of C library that get called from a higher level language's FFI.
Would this pattern ever been an interest to the project? If you happen to write something compitational complex in temper, I can see just moving it to native python might be impractical, preventing the idea of easily making bindings in every language.
For computationally complex operations, often its impractical to write some things in pure python. But having python bindings to act as glue, or to integrate with a not-so-computationally-intensive code is helpf. Thats the usecase for building a rust library and generating a python module using PyO3.
Is it in the spirit of temper to eventually (long in the future Im sure) have an alternative build path for python that permits python-friendly bindings but with compiled/native code? I think the python+rust story is common, but you could generalize this to any kind of C library that get called from a higher level language's FFI.
Would this pattern ever been an interest to the project? If you happen to write something compitational complex in temper, I can see just moving it to native python might be impractical, preventing the idea of easily making bindings in every language.