Enhancements

• Allow channel_alias to interpolate environment variables.
• Support running conda with PyPy on Windows.
• Add ability to add, append and prepend to sequence values when using the conda config subcommand.
• Support Python 3.10 in version parser.
• Add XDG_CONFIG_HOME to the conda search path following the XDG Base Directory Specification (XDGBDS).

Bug fixes

• Fix the PowerShell activator to not show an error when unsetting environment variables.
• Remove superfluous eval statements in fish shell integration.
• Indent the conda fish integration file using fish_indent.
• Fix handling of environment variables containing equal signs (=).
• Handle permission errors when listing all known prefixes.
• Catch Unicode decoding errors when parsing conda-meta files.
• Fix handling write errors when trying to create package cache or env directories.

Docs

• Update path of conda repo in RHEL based systems to /etc/yum.repos.d/conda.repo.
• Fix the advanced pip example to stop using the now invalid file: prefix.
• Minor docs cleanup and adding Code of Conduct.
• Add auto-built architecture documentation for conda based on the C4 Model. See the conda documentation for more information.
• Expand the contributing documentation with a section about static code analysis and code linting.
• Add developer guide section to the documentation, including a conda architecture overview.
• Stop referring to updating anaconda when conda update fails with an error.

Other

• Build Docker images periodically on GitHub Actions for the continuous integration testing on Linux, storing them on GitHub Packages's registry for reduced latency and cost when using Docker Hub.

• Simplify the Linux GitHub actions workflows by combining used shell scripts.

• Add periodic GitHub Actions workflow to review old issues in the conda issue tracker and mark them as stale if no feedback is provided in a sensible amount of time, eventually closing them.

• Add periodic GitHub Actions workflow to lock the comment threads of old issues and pull requests in the conda GitHub repository to surface regressions with new issues instead.

• Refactor test suite to use more GitHub Actions runners in parallel, reducing total run time by 50%.

• Switched the issue tracker to use forms with additional questions for bug reporters to help in ticket triage.

• Add and automatically run pre-commit as part of the CI system to improve the code quality continuously and raise issues in contributed patches early on.

  The used code linters are: flake8, pylint and bandit.

  The Python code formatter black is used as well but is only enforced on changed code in a commit and not to the whole code base at once.

• Automatically build the conda package upon the successful merge into the master branch and upload it to the conda-canary channel on anaconda.org.

  To try conda out simply run:

  conda install -c conda-canary/label/dev conda
  

• Automate adding new issues to public GitHub project board to facilitate issue triage.

• Update GitHub issue and pull request labels to be more consistent.

• Start using rever for release management.

• (preview) Enable one-click gitpod and GitHub Codespaces setup for Linux development.

Contributors

• Benjamin Bertrand
• Chawye Hsu
• Cheng H. Lee
• Dan Meador
• Daniel Bast
• Daniel Holth
• Gregor Kržmanc
• Hsin-Hsiang Peng
• Ilan Cosman
• Isuru Fernando
• Jaime Rodríguez-Guerra
• Jan-Benedikt Jagusch
• Jannis Leidel
• John Flavin
• Jonas Haag
• Ken Odegard
• Kfir Zvi
• Mervin Fansler
• bfis
• mkincaid
• pre-commit CI