Skip to content

Improve Documentation on PyPI Release Process #60

@medley56

Description

@medley56

The example of releasing a Poetry-built package to PyPI is using manual CLI commands, which should not be the recommended approach IMO. Releases should be managed through some kind of CI workflow (Jenkins, GitHub Actions, Travis, GitLab CI, or others). Using manual uploads from a developer laptop is error prone and should be discouraged as you can end up with "dirty" (uncommitted code) releases on PyPI, which are immutable once published. Automated pipelines can perform validation checks prior to publishing to ensure tags, versions, and package metadata is in sync. They can also streamline releasing to GitHub, PyPI, and Conda all at once, which is critical for consistent release management between ecosystems.

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions