These are things lacking a PEP which, if solved, would solve a demonstrated pain point in the ecosystem. # Installation ## Index priority - A lot of work for pip to implement # Metadata ## `pyproject.toml` ### SBOMs - Probably should be a list of globs - Record build back-end, dependencies, and Python interpreter used ### License for source code for the project - I.e. not about what's contained in the wheel ### How to run tests - [Discussion](https://discuss.python.org/t/pre-pep-standardizing-test-dependency-and-command-specification/71896/61) - Check if a project supports a specific Python version - Redistributors - Check a change in a project (including CPython) won't break anything ## Distributions ### Consistent metadata across all wheels in a release - What data is used to select a wheel? - Tags - Interpreter - ABI - Platform - `requires-python` - Probably could use `in` since lock files introduced the concept - Consider the set of what's compatible is complete much like how 'packaging' generates the full list - E.g. `"manylinux_2_28" in platform_tags` ## Indexes ### Project and release-level metadata - Allows for mutability ## Ability to require 2FA sign-off to publish a release - Requires staging releases - [API 2.0](https://peps.python.org/pep-0694/) will have that - Would need something for API 1.0 since it's going to be around for a while - NPM now does this