导入模块时，如何指定解释器来编译模块？

Question

I wrote a module (mymodule32.py) that contains a library that only supports 32 bit Python. I want to import this module in a 64 bit Python file (main64.py). But after searching, I realized that if I import the module in 64 bit Python, it will be compiled with the 64 bit interpreter. I will import mymodule32.py in main64.py, but I want to interpret it with a 32 bit interpreter.

I installed Python 3.5 64 bit, and in addition I installed Python 3.5 32 bit under the name py35_32 in the Anaconda virtual environment. I wrote the pseudo code below I wanted(it does not work, of course).

import(interpreter = py35_32) mymodule32

Answer 1

You can't do what you want with import . The interpreter is a single process, and can't switch between 64-bit and 32-bit modes.

Your options are to

• Run all your code in a Python 32-bit binary. You can guard against using a 64-bit binary by testing for sys.maxsize == (2 ** 31 - 1) .

• Make your library work correctly on a 64-bit interpreter too (just make sure your 32-bit arithmetic is properly masked).

• Run another Python process as a subprocess; that child process would import your module and use some form of IPC (interprocess communication) to pass data and results between the parent 64-bit process and the child 32-bit interpreter. There are no ready-made solutions here, however.