Preparing packages for Pypi for different versions of python

I am preparing a package for Pypi that calculates stationary distributions for a Markov chain with discrete state spaces .

The main files with the code run without problems on both Python 2.7 and Python 3.5 (we tested this in a virtual environment). The setup.py file also works, at least for Python 2.7.

From now on, I don’t understand how to act. Should I run setup.py on all possible versions of Python to create separate egg files for each version of Python? This seems rather tedious given that the file works on both versions of Python, and you will have to redo it every time you update the package.

What would be good practice here?

+4
source share
1 answer

If you are looking for a good release checklist (answering the question “what should I do before I click on pypi?”), You won’t be able to bypass Audrey Roy PyPI release checklist . This is a living document and is updated as necessary, but the high points are:

  • Update Change Log
  • Update Version Number
  • Run tests
  • Build for pypi using both the original distributions (sdist) and the newer wheel format.
  • Click on pypi test instance (using twine )
  • Verify that the README is displayed correctly.
  • ,
  • Push to pypi production
  • / git

:

- ( ). - .gitignore, . , , , , sdists , .

" ", , , . tox , travis appveyor. , , -

+2

All Articles