Skip to content

Latest commit

 

History

History
98 lines (78 loc) · 4.25 KB

CONTRIBUTING.md

File metadata and controls

98 lines (78 loc) · 4.25 KB

Contributing to winnow

Thanks for wanting to contribute! There are many ways to contribute and we appreciate any level you're willing to do.

Feature Requests

Need some new functionality to help? You can let us know by opening an issue. It's helpful to look through all issues in case its already being talked about.

Bug Reports

Please let us know about what problems you run into, whether in behavior or ergonomics of API. You can do this by opening an issue. It's helpful to look through all issues in case its already being talked about.

Pull Requests

Looking for an idea? Check our issues. If it's look more open ended, it is probably best to post on the issue how you are thinking of resolving the issue so you can get feedback early in the process. We want you to be successful and it can be discouraging to find out a lot of re-work is needed.

Already have an idea? It might be good to first create an issue to propose it so we can make sure we are aligned and lower the risk of having to re-work some of it and the discouragement that goes along with that.

Parsers

Design guidelines

  • Generally grammar-level Parsers are free-functions and output/error conversion are inherent functions on Parser. Parser::verify is an example of some nuance as the logic is coupled to the Parser its applied to.
  • Parsers that directly process tokens must support complete vs streaming parsing.
  • Parsers that work with slices have take in their name.
  • When taking slices or repeatedly calling a Parser, control the number of times with a range, rather than hard coding it with the range in the name.
  • Where possible, write Parsers in a straight-forward manner, reusing other Parsers, so they may serve as examples for the user.

Process

As a heads up, we'll be running your PR through the following gauntlet:

  • warnings turned to compile errors
  • cargo test
  • rustfmt
  • clippy
  • rustdoc
  • committed as we use Conventional commit style
  • typos to check spelling

Not everything can be checked automatically though.

We request that the commit history gets cleaned up. We ask that commits are atomic, meaning they are complete and have a single responsibility. PRs should tell a cohesive story, with test and refactor commits that keep the fix or feature commits simple and clear.

Specifically, we would encourage

  • File renames be isolated into their own commit
  • Add tests in a commit before their feature or fix, showing the current behavior. The diff for the feature/fix commit will then show how the behavior changed, making it clearer to reviewrs and the community and showing people that the test is verifying the expected state.

Note that we are talking about ideals. We understand having a clean history requires more advanced git skills; feel free to ask us for help! We might even suggest where it would work to be lax. We also understand that editing some early commits may cause a lot of churn with merge conflicts which can make it not worth editing all of the history.

For code organization, we recommend

  • Grouping impl blocks next to their type (or trait)
  • Grouping private items after the pub item that uses them.
    • The intent is to help people quickly find the "relevant" details, allowing them to "dig deeper" as needed. Or put another way, the pub items serve as a table-of-contents.
    • The exact order is fuzzy; do what makes sense

Releasing

Pre-requisites

  • Running cargo login
  • A member of winnow-rs:Maintainers
  • Push permission to the repo
  • cargo-release

When we're ready to release, a project owner should do the following

  1. Update the changelog (see cargo release changes for ideas)
  2. Determine what the next version is, according to semver
  3. Run cargo release -x <level>