v1.8.4

What's Improved

• @hugovk cleaned up the double whitespaces in the OIDC flow logging in #140
• @woodruffw added a title and a docs link to the OIDC error output in #139

Full Diff: v1.8.3...v1.8.4

v1.8.3

What's New

This release improves the logging detalization of which authentication mode is selected when the action runs. It surfaces this detail to the workflow run summary page as annotations. The change was contributed by @woodruffw in #136.

Full Diff: v1.8.2...v1.8.3

v1.8.2

What's Changed

This release started printing out full OIDC error messages to console, instead of just one line -- by @woodruffw in #134.

Full Diff: v1.8.1...v1.8.2

v1.8.1

🐛 What's Fixed

💔 Unfortunately, a tiny mistake in v1.8.0 caused a far-reaching regression for the most used code path.
❗ But don't worry, it's fixed now thanks to @njzjz who promptly spotted it and @zhongjiajie who sent a bugfix.

🙌 New Contributors

• @zhongjiajie made their first contribution in #131

Full Diff: v1.8.0...v1.8.1

v1.8.0

The Coolest Release Ever!

In this release, @woodruffw implemented support for secretless OIDC-based publishing to PyPI-like package indexes. The OIDC flow is activated when neither username nor password action inputs are set.

The OIDC “token exchange”, is an authentication technique that PyPI (and TestPyPI, and hopefully some future others) supports as an alternative to long-lived username/password combinations or long-lived API tokens.

IMPORTANT: The PyPI-side configuration is only available to participants of the private beta test. Please, only try out the zero-config mode if you are a beta test participant having followed the PyPI configuration instructions.
It's gone GA during Python 2023 and is available to everyone now.

Setup prerequisites: https://github.com/marketplace/actions/pypi-publish#trusted-publishing
PyPI's documentation: https://pypi.org/help/#trusted-publishers
Beta test enrollment: pypi/warehouse#12965

New Contributors

• @woodruffw made their first contribution in #123

Full Diff: v1.7.1...v1.8.0

v1.7.1

Regression?

There was a small setback with v1.7.0 — the snake_case fallbacks didn't work because the check for the kebab-case env vars with default values set was always truthy. This bugfix release promptly fixes that.

Full Diff: v1.7.0...v1.7.1

v1.7.0

What should I care about?

TL;DR The action input names have been converted to use kebab-case and marked deprecated. But the old names still work.

This is made to align the public API with the de-facto conventions in the ecosystem. We've used snake_case names, which the maintainer considers a historical mistake. New kebab-case inputs will make the end-users' workflows look more consistent and and visually distinguishable from other identifiers one may encounter in YAML.

There is no timeline for removing the old names, but it will happen in v3 or later versions of the action. If the maintainer doesn't forget to do this, that is.

The patch is here: #125.

Full Diff: v1.6.5...v1.7.0

v1.6.5

What's Changed

• Added an explicit warning when the password passed into the action is empty — thanks @colindean

New Contributors

• @colindean made their first contribution in #122

Full Diff: v1.6.4...v1.6.5

v1.6.4

oh, boi! again?

This is the last one tonight, promise! It fixes this embarrassing bug that was actually caught by the CI but got overlooked due to the lack of sleep.
TL;DR GH passed $HOME from the external env into the container and that tricked the Python's site module to think that the home directory is elsewhere, adding non-existent paths to the env vars. See #115.

Full Diff: v1.6.3...v1.6.4

v1.6.3

Another Release!? Why?

In #112 (comment), it was discovered that passing a $PATH variable even breaks the shebang. So this version adds more safeguards to make sure it keeps working with a fully broken $PATH.

Full Diff: v1.6.2...v1.6.3