Open maichmueller opened 3 weeks ago
It's a good idea.
I am not sure how well this works with the automatic generation of the python stubs since the ordering of declarations matters. Please make sure that generating the stubs still works. I suggest starting by moving just a tiny bit and see what stubs you get.
Hi,
the current one-source binding solution is cumbersome to work with. Tiny changes in a single part of the bindings require lengthy recompilation, followed by even longer LTO times.
I would like to break up the binding code into individual compilation units
init_action
,init_state_space
etc. to prevent this. This procedure is also pybind11's recommended practice IIRC.Before doing this, would you agree with such a PR or are you generally against it?