Pico/CONTRIBUTING.md
2017-06-20 17:32:09 +02:00

12 KiB

Contributing to Pico

Pico aims to be a high quality Content Management System (CMS) but at the same time wants to give contributors freedom when submitting fixes or improvements.

As such we want to encourage but not obligate you, the contributor, to follow these guidelines. The only exception to this are the guidelines elucidated in the Prevent merge-hell section.

Having said that: we really appreciate it when you apply the guidelines in part or wholly as that will save us time which, in turn, we can spend on bugfixes and new features.

Issues

If you want to report an issue with Pico's core, please create a new Issue on GitHub. Concerning problems with plugins or themes, please refer to the website of the developer of this plugin or theme.

Before creating a new Issue on GitHub, please make sure the problem wasn't reported yet using GitHubs search engine.

Please describe your issue as clear as possible and always include the Pico version you're using. Provided that you're using plugins, include a list of them too. We need information about the actual and expected behavior, the steps to reproduce the problem, and what steps you have taken to resolve the problem by yourself (i.e. your own troubleshooting).

Contributing code

Once you decide you want to contribute to Pico's core (which we really appreciate!) you can fork the project from https://github.com/picocms/Pico. If you're interested in developing a plugin or theme for Pico, please refer to the development section of our website.

Prevent merge-hell

Please do not develop your contribution on the master branch of your fork, but create a separate feature branch, that is based off the master branch, for each feature that you want to contribute.

Not doing so means that if you decide to work on two separate features and place a pull request for one of them, that the changes of the other issue that you are working on is also submitted. Even if it is not completely finished.

To get more information about the usage of Git, please refer to the Pro Git book written by Scott Chacon and/or this help page of GitHub.

Pull Requests

Please keep in mind that pull requests should be small (i.e. one feature per request), stick to existing coding conventions and documentation should be updated if required. It's encouraged to make commits of logical units and check for unnecessary whitespace before committing (try git diff --check). Please reference issue numbers in your commit messages where appropriate.

Coding Standards

Pico uses the PSR-2 Coding Standard as defined by the PHP Framework Interoperability Group (PHP-FIG).

For historical reasons we don't use formal namespaces. Markdown files in the content-sample folder (the inline documentation) must follow a hard limit of 80 characters line length.

It is recommended to check your code using PHP_CodeSniffer using Pico's .phpcs.xml standard. Use the following command:

$ ./vendor/bin/phpcs --standard=.phpcs.xml [file]...

With this command you can specify a file or folder to limit which files it will check or omit that argument altogether, in which case the current working directory is checked.

Keep documentation in sync

Pico accepts the problems of having redundant documentation on different places (concretely Pico's inline user docs, the README.md and the website) for the sake of a better user experience. When updating the docs, please make sure to keep them in sync.

If you update the README.md or content-sample/index.md, please make sure to update the corresponding files in the _docs folder of the picocms.github.io repo (i.e. Pico's website) and vice versa. Unfortunately this involves three (!) different markdown parsers. If you're experiencing problems, use Pico's erusev/parsedown-extra as a reference. You can try to make the contents compatible to Kramdown (Pico's website) and Redcarpet (README.md) by yourself, otherwise please address the issues in your pull request message and we'll take care of it.

Versioning

Pico follows Semantic Versioning 2.0 and uses version numbers like MAJOR.MINOR.PATCH. We will increment the:

  • MAJOR version when we make incompatible API changes,
  • MINOR version when we add functionality in a backwards-compatible manner, and
  • PATCH version when we make backwards-compatible bug fixes.

For more information please refer to the http://semver.org website.

Branching

The master branch contains the current development version of Pico. It is likely unstable and not ready for production use.

However, the master branch always consists of a deployable (but not necessarily deployed) version of Pico. As soon as development of a new MAJOR or MINOR release starts, a separate branch (e.g. pico-1.1) is created and a pull request is opened to receive the desired feedback.

Pico's actual development happens in separate development branches. Development branches are prefixed by:

  • feature/ for bigger features,
  • enhancement/ for smaller improvements, and
  • bugfix/ for non-trivial bug fixes.

As soon as development reaches a point where feedback is appreciated, a pull request is opened. After some time (very soon for bug fixes, and other improvements should have a reasonable feedback phase) the pull request is merged and the development branch will be deleted. Trivial bug fixes which will be part of the next PATCH version will be merged directly into master.

Build & Release process

We're using Travis CI to automate the build & release process of Pico. It generates and deploys phpDoc class docs for new releases and on every commit to the master branch. Travis also prepares new releases by generating Pico's pre-built packages, a version badge and pushing both to our website (the picocms.github.io repo). Please refer to our .travis.yml and the _build directory for details.

As insinuated above, it is important that each commit to master is deployable. Once development of a new Pico release is finished, trigger Pico's build & release process by pushing a new Git tag. This tag should reference a (usually empty) commit on master, which message should adhere to the following template:

Version 1.0.0

* [Security] ...
* [New] ...
* [Changed] ...
* [Fixed] ...
* [Removed] ...

Before pushing a new Git tag, make sure to update the Pico::VERSION and Pico::VERSION_ID constants. If you've simultaneously developed a new version of Pico's official default theme and/or the PicoDeprecated plugin, make sure to create releases for them first, otherwise automatic deployment will fail.

Travis CI will draft the new release on GitHub automatically, but will require you to manually amend the descriptions formatting. The latest Pico version is always available at https://github.com/picocms/Pico/releases/latest, so please make sure to publish this URL rather than version-specific URLs. Packagist will be updated automatically.

Labeling of Issues & Pull Requests

Pico makes use of GitHub's label and milestone features, to aide developers in quickly identifying and prioritizing which issues need to be worked on. The starting point for labeling issues and pull requests is the type label, which is explained in greater detail below. The type label might get combined with a pri label, describing the issue's priority, and a status label, describing the current status of the issue.

Issues and pull requests labeled with info: Feedback Needed indicate that feedback from others is highly appreciated. We always appreciate feedback at any time and from anyone, but when this label is present, we explicitly ask you to give feedback. It would be great if you leave a comment!

  • The type: Bug label is assigned to issues or pull requests, which have been identified as bugs or security issues in Pico's core. It might get combined with the pri: High label, when the problem was identified as security issue, or as a so-called "show stopper" bug. In contrast, uncritical problems might get labeled with pri: Low. type: Bug issues and pull requests are usually labeled with one of the following status labels when being closed:

    • status: Resolved is used when the issue has been resolved.
    • status: Conflict indicates a conflict with another issue.
    • status: Won't Fix means, that the problem was identified as intended behavior.
    • status: Rejected is used when the issue was rejected for another reason (e.g. when the issue wasn't reproducible).
  • The type: Enhancement and type: Feature labels are used to tag pull requests, which introduce either a comparatively small enhancement, or a "big" new feature. As with the type: Bug label, they might get combined with the pri: High or pri: Low labels to indicate the pull request's priority. A pull request isn't necessarily mergeable when being opened, so it might get labeled with status: Pending Merge when development of the pull request is finished, but merging isn't intended yet. After merging or closing the pull request, it is labeled with one of the status labels as described above for the type: Bug label.

  • The type: Idea label is similar to the type: Enhancement and type: Feature labels, but is used for issues or incomplete and abandoned pull requests. It is otherwise used in the exact same way as type: Enhancement and type: Feature.

  • The type: Release label is used in the exact same way as type: Feature and indicates the primary pull request of a new Pico release (please refer to the Branching and Build & Release process sections above).

  • The type: Notice, type: Question and type: Discussion labels are used to indicate "fyi" issues, issues opened by users or developers asking questions, and issues with disucssions about arbitrary topics related to Pico. They are neither combined with pri labels, nor with status labels.

  • The type: Duplicate label is used when there is already another issue or pull request related to this problem or feature request. Issues labeled with type: Duplicate are immediately closed.

  • The type: Invalid label is used for everything else, e.g. issues or pull requests not related to Pico, or invalid bug reports.

The status: Deferred label might get added to any open issue or pull request to indicate that it still open and will be resolved later.

Issues and pull requests, which are rather related to upstream projects (i.e. projects Pico depends on, like Twig), are additionally labeled with info: Upstream.

When a issue or pull request isn't directly related to Pico's core, but the project as a whole, it is labeled with info: Meta. The same applies to the info: Website label, however, in this case it is usually expedient to move the issue to the picocms.github.io repo containing Pico's website.