When creating a package for PyPi, and naming an author, but not an author_email, you get a warning as follows:

warning: Check: missing meta-data: if 'author' supplied, 'author_email' must be supplied too

The specs ( https://packaging.python.org/specifications/core-metadata/#author ) do not enforce this behavior, so I'd like to change the wording from `must` to `should`.

This can be reproduced by creating a setup.py, providing `author`, but not `author_email`, and then calling `python setup.py check` or `python -m pep517.build .`.

This issue was discussed at:
https://discuss.python.org/t/which-fields-are-required-for-a-setup-py-especially-is-author-required/2705
and
https://github.com/pypa/pep517/issues/73

Background:
I ported a 16 year old package to Python 3, and tried to upload it to PyPi. I know the author name, but not his email address. Also, I think he does not like to get bothered with emails for a project he abandoned 16 years ago.

P.S.: I am working on a PR for this and update this issue accordingly.

For your project, if you define maintainer and maintainer-email (with your own info), it’s too bad that adding author results in a warning for missing author-email!

The goal of these checks as I understand them is a best effort to encourage projects to contain contact information (and recognition for the work).  It seems legitimate to know the original author name but not email, and too bad that the simplest way to avoid the warning is to remove the author info.

I wonder if it would be going too far to change the checks to avoid the warning if we have author, maintainer and maintainer-email (or even more combinations? given that email format allows embedding a name directly there)

Thank you for your feedback.

Paul Ganssle suggested a updated wording over at
https://github.com/pypa/pep517/issues/73
so this is why I created a pr accordingly.

The intent of this issue and the pr is to minimize the confusion for a beginner with Python packaging (e.g myself) by doing something with the probably wrong word "must".

Changing the checks is another way to fix this issue, but it is beyond my scope to say one is better or the other.

For the future, we generally tend to keep distutils pretty "frozen", only making minor changes or the changes needed to build Python itself. Instead we generally make changes in setuptools, which for the moment monkey-patches distutils (and into which distutils will eventually be merged). One of the big reasons is that setuptools is used across all versions of Python, so the changes are automatically backported, whereas changes to distutils will only be seen by people using the most recent Python versions.

In this case, it's not really a substantive change, so I think we can leave it in distutils, I just wanted to bring this up as an FYI.

Thank you, both of you!

I especially appreciate the further information about how setuptools and distutils play together.

New changeset 9f9dac0a4e58d5c72aa3b644701cb155c009cb2c by Miss Islington (bot) (Jürgen Gmach) in branch 'master':
bpo-38914 Do not require email field in setup.py. (GH-17388)
https://github.com/python/cpython/commit/9f9dac0a4e58d5c72aa3b644701cb155c009cb2c

This issue can be closed as the related PR was merged.