🔒 Python and Semantic-release versioning

[dusky halo]

Hey folks! I’m stuck with python-semantic-release on a “test” branch where it keeps trying to upload the same artifacts (version 0.1.0-rc.1), resulting in a 422 (Unprocessable Entity) error on GitHub. Here’s the situation:

• I have a “test” branch for dev/pre-releases.
• I originally ran python -m build before running semantic-release publish, so it always built artifacts labeled 0.1.0 (or 0.1.0-rc.1).
• When semantic-release checks for a new version, it never increments beyond 0.1.0-rc.1, reuses the same tag, and tries to re-upload the same .whl/.tar.gz to GitHub → 422 error because those files already exist under that release.

I’ve tried:

1. Removing the manual python -m build step (letting semantic-release do the build),
2. Creating commits with feat!: ... BREAKING CHANGE: lines to force major bumps,
3. Even environment overrides like SEMANTIC_RELEASE_LEVEL=minor.

No matter what, the logs say “found 1 previous tag (0.1.0-rc.1),” reuses that tag, and collides with existing assets. Any idea why it won’t bump to something like 0.1.0-rc.2 or 0.2.0-rc.1? Is it because semantic-release thinks there aren’t actually new recognized commits, or because the prior build is always the same version?

How should I ensure each run actually increments the version and builds unique artifacts so we don’t keep hitting the 422 error? Any tips or config suggestions would be greatly appreciated!

[tulip sequoia]

@dusky halo

Python help channel opened

Remember to:

• Ask your Python question, not if you can ask or if there's an expert who can help.
• Show a code sample as text (rather than a screenshot) and the error message, if you've got one.
• Explain what you expect to happen and what actually happens.

:warning: Do not pip install anything that isn't related to your question, especially if asked to over DMs.

Closes after a period of inactivity, or when you send !close.

[dusky halo]

name: Release to PyPI

on:
  push:
    branches: [ "test" ]

jobs:
  release:
    runs-on: ubuntu-latest

    steps:
      - name: Check out repo
        uses: actions/checkout@v3
        with:
          fetch-depth: 0

      - name: Set up Python
        uses: actions/setup-python@v4
        with:
          python-version: "3.11"

      - name: Install Dependencies
        run: |
          pip install --upgrade pip
          pip install build python-semantic-release

      - name: Build the package
        run: python -m build

      - name: Run Semantic Release
        env:
          GH_TOKEN: ${{ secrets.GH_TOKEN }}
          PYPI_USERNAME: "__token__"
          PYPI_PASSWORD: ${{ secrets.PYPI_TOKEN }}
        run: semantic-release -vv publish

[tool.semantic_release]
tag_format = "{version}"
build_command = "python -m build"

[tool.semantic_release.branches."test"]
match = "test"
prerelease = true
prerelease_token = "rc"
repository_url = "https://test.pypi.org/legacy/"
starting_version = "0.1.0"

[tool.semantic_release.branches."main"]
match = "main"
prerelease = false
repository_url = "https://upload.pypi.org/legacy/"

[silver ridge]

where does this thing get its version data from

[dusky halo]

I think the problem is that python -m build always sees version = "0.1.0" from the config, so it builds 0.1.0 artifacts.
but i don't really understand what should i do, remove the building maybe or what to change

[dusky halo]

tried to remove the starting_version = "0.1.0" it didn't resolve the problem,
also tried to remove the buliding from the workflow, as i assume the sematic release might handle the build and add the version etc but still didn't work,

[silver ridge]

I've been doing semantic releases for a long time w/o this tool

[dusky halo]

if you can guide me through it, I would really appreciate,

also one question just to make sure,
does semantic release handle the following scenarios?

• create git tag and releases
• build distributions and add it to either pypi or test pypi

[silver ridge]

this is my workflow: https://github.com/agronholm/anyio/blob/master/.github/workflows/publish.yml

I push a git tag and everything else happens automatically

sdist/wheel creation, uploading to pypi and publishing a github release w/ the changelog

[dusky halo]

I appreciate it!
will figure it out thanks!

The project that i am working on is also going to be open source and later on there might be some contributors as well,
I am thinking to create a main and test branch and publish on the test pypi first and then merge it to the main, will check your repo as well and see how you did it!

[silver ridge]

note that I am using trusted publishing which has to be configured separately on the pypi

[dusky halo]

do you manually add a new tag? like git tag -a v1.0 -m "Version 1.0 release"

[silver ridge]

I manually add a new tag, yes

[dusky halo]

do you also update the doc manually?

https://github.com/agronholm/anyio/commit/74022ec53cde6c2e5d7441de37d5a0dcf9d6daad

[silver ridge]

yes

[dusky halo]

ohh okkay, Thanks!
I will try to do the same!

the only wierd part for me is that your project.toml file deosn't have

tool.semantic_release

[silver ridge]

why is that weird? I told you I'm not using that tool

[dusky halo]

ohh my bad! I see

[tulip sequoia]

Python help channel closed for inactivity

This help channel has been closed. Feel free to create a new post in #1035199133436354600. To maximize your chances of getting a response, check out this guide on asking good questions.