To learn more about the purpose of PEPs and how to go about writing one, please start reading at PEP 1. Also, make sure to check the README for information on how to render the PEPs in this repository. Thanks again for your contributions, and we look forward to reviewing them!
Prior to submitting a pull request here with your draft PEP, see PEP 1 for some important steps to consider, including proposing and discussing it first in an appropriate venue, drafting a PEP and gathering feedback, and developing at least a prototype reference implementation of your idea.
In general, most non-Draft/Active PEPs are considered to be historical documents rather than living specifications or documentation. Major changes to their core content usually require a new PEP, while smaller modifications may or may not be appropriate, depending on the PEP's status. See PEP Maintenance and Changing Existing PEPs in PEP 1 for more.
Copyediting and proofreading Draft and Active PEPs is welcome (subject to review by the PEP author), and can be done via pull request to this repo. Substantive content changes should first be proposed on PEP discussion threads. We do advise against PRs that simply mass-correct minor typos on older PEPs which don't significantly impair meaning and understanding.
If you're still unsure, we encourage you to reach out first before opening a PR here. For example, you could contact the PEP author(s), propose your idea in a discussion venue appropriate to the PEP (such as Typing Discourse for static typing, or Packaging Discourse for packaging), or open an issue.
The PEPs repository defines a set of pull request templates, which should be used when opening a PR.
If you use Git from the command line, you may be accustomed to creating PRs by following the URL that is provided after pushing a new branch. Do not use this link, as it does not provide the option to populate the PR template.
However, you can use the gh
command line tool. gh pr create
will allow
you to create a pull request, will prompt you for the template you wish to use,
and then give you the option of continuing editing in your broswer.
Alternatively, after pushing your branch, you can visit https://github.com/python/peps, and follow the link in the notification about recent changes to your branch to create a new PR. The in-browser interface will allow you to select a PR template for your new PR.
When adding or modifying a PEP, please include the PEP number in the commit
summary and pull request title. For example, PEP NNN: <summary of changes>
.
Likewise, prefix rendering infrastructure changes with Infra:
, linting
alterations with Lint:
and other non-PEP meta changes, such as updates to
the Readme/Contributing Guide, issue/PR template, etc., with Meta:
.
All contributors need to sign the PSF Contributor Agreement. to ensure we legally accept your work.
You don't need to do anything beforehand; go ahead and create your pull request, and our bot will ping you to sign the CLA if needed. See the CPython devguide for more information.
All interactions for this project are covered by the PSF Code of Conduct. Everyone is expected to be open, considerate, and respectful of others, no matter their position within the project.
You can run this repo's basic linting suite locally, either on-demand, or automatically against modified files whenever you commit your changes.
They are also run in CI, so you don't have to run them locally, though doing so will help you catch and potentially fix common mistakes before pushing your changes and opening a pull request.
This repository uses the pre-commit tool to install, configure and update a suite of hooks that check for common problems and issues, and fix many of them automatically.
If your system has make
installed, you can run the pre-commit checkers
on the full repo by running make lint
. This will
install pre-commit in the current virtual environment if it isn't already,
so make sure you've activated the environment you want it to use
before running this command.
Otherwise, you can install pre-commit with
python -m pip install pre-commit
(or your choice of installer), and then run the hooks on all the files in the repo with
pre-commit run --all-files
or only on any files that have been modified but not yet committed with
pre-commit run
If you would like pre-commit to run automatically against any modified files every time you commit, install the hooks with
pre-commit install
Then, whenever you git commit
, pre-commit will run and report any issues
it finds or changes it makes, and abort the commit to allow you to check,
and if necessary correct them before committing again.
To check for common spelling mistakes in your PEP and automatically suggest corrections, you can run the codespell tool through pre-commit as well.
Like the linters, on a system with make
available, it can be installed
(in the currently-activated environment) and run on all files in the
repository with a single command, make spellcheck
.
For finer control or on other systems, after installing pre-commit as in the previous section, you can run it against only the files you've modified and not yet committed with
pre-commit run --hook-stage manual codespell
or against all files with
pre-commit run --all-files --hook-stage manual codespell