Skip to content

Latest commit

 

History

History
163 lines (103 loc) · 11 KB

CONTRIBUTING.md

File metadata and controls

163 lines (103 loc) · 11 KB

Contributing

Hi there! We're thrilled that you'd like to contribute to this project. Your help is essential for keeping it great. This project adheres to the Contributor Covenant Code of Conduct. By participating, you are expected to uphold this code.

The majority of contributions won't need to touch any Ruby code at all.

Getting started

Before you can start contributing to Linguist, you'll need to set up your environment first. Clone the repo and run script/bootstrap to install its dependencies.

git clone https://github.com/github/linguist.git
cd linguist/
script/bootstrap

To run Linguist from the cloned repository, you will need to generate the code samples first:

bundle exec rake samples

Run this command each time a sample has been modified.

To run Linguist from the cloned repository:

bundle exec bin/linguist --breakdown

Dependencies

Linguist uses the charlock_holmes character encoding detection library which in turn uses ICU, and the libgit2 bindings for Ruby provided by rugged. These components have their own dependencies - icu4c, and cmake and pkg-config respectively - which you may need to install before you can install Linguist.

For example, on macOS with Homebrew: brew install cmake pkg-config icu4c and on Ubuntu: apt-get install cmake pkg-config libicu-dev.

Adding an extension to a language

We try only to add new extensions once they have some usage on GitHub. In most cases we prefer that extensions be in use in hundreds of repositories before supporting them in Linguist.

To add support for a new extension:

  1. Add your extension to the language entry in languages.yml, keeping the extensions in alphabetical and case-sensitive (uppercase before lowercase) order, with the exception of the primary extension; the primary extension should be first.
  2. Add at least one sample for your extension to the samples directory in the correct subdirectory. We'd prefer examples of real-world code showing common usage. The more representative of the structure of the language, the better.
  3. Open a pull request, linking to a GitHub search result showing in-the-wild usage.
    If you are adding a sample, please state clearly the license covering the code in the sample, and if possible, link to the original source of the sample.

Additionally, if this extension is already listed in languages.yml and associated with another language, then sometimes a few more steps will need to be taken:

  1. Make sure that example .yourextension files are present in the samples directory for each language that uses .yourextension.
  2. Test the performance of the Bayesian classifier with a relatively large number (1000s) of sample .yourextension files. (ping @lildude to help with this) to ensure we're not misclassifying files.
  3. If the Bayesian classifier does a bad job with the sample .yourextension files then a heuristic may need to be written to help.

Adding a language

We try only to add languages once they have some usage on GitHub. In most cases we prefer that each new file extension be in use in hundreds of repositories before supporting them in Linguist.

To add support for a new language:

  1. Add an entry for your language to languages.yml. Omit the language_id field for now.
  2. Add a syntax-highlighting grammar for your language using: script/add-grammar https://github.com/JaneSmith/MyGrammar
    This command will analyze the grammar and, if no problems are found, add it to the repository. If problems are found, please report them to the grammar maintainer as you will not be able to add the grammar if problems are found.
    Please only add grammars that have one of these licenses.
  3. Add samples for your language to the samples directory in the correct subdirectory.
  4. Add a language_id for your language using script/set-language-ids.
    You should only ever need to run script/set-language-ids --update. Anything other than this risks breaking GitHub search 😢
  5. Open a pull request, linking to a GitHub search results showing in-the-wild usage.
    Please state clearly the license covering the code in the samples. Link directly to the original source if possible.

In addition, if your new language defines an extension that's already listed in languages.yml (such as .foo) then sometimes a few more steps will need to be taken:

  1. Make sure that example .foo files are present in the samples directory for each language that uses .foo.
  2. Test the performance of the Bayesian classifier with a relatively large number (1000s) of sample .foo files. (ping @lildude to help with this) to ensure we're not misclassifying files.
  3. If the Bayesian classifier does a bad job with the sample .foo files then a heuristic may need to be written to help.

Remember, the goal here is to try and avoid false positives!

Fixing a misclassified language

Most languages are detected by their file extension defined in languages.yml. For disambiguating between files with common extensions, Linguist applies some heuristics and a statistical classifier. This process can help differentiate between, for example, .h files which could be either C, C++, or Obj-C.

Misclassifications can often be solved by either adding a new filename or extension for the language or adding more samples to make the classifier smarter.

Fixing syntax highlighting

Syntax highlighting in GitHub is performed using TextMate-compatible grammars. These are the same grammars that TextMate, Sublime Text and Atom use. Every language in languages.yml is mapped to its corresponding TextMate scopeName. This scope name will be used when picking up a grammar for highlighting.

Assuming your code is being detected as the right language, in most cases syntax highlighting problems are due to a bug in the language grammar rather than a bug in Linguist. vendor/README.md lists all the grammars we use for syntax highlighting on GitHub.com. Find the one corresponding to your code's programming language and submit a bug report upstream. If you can, try to reproduce the highlighting problem in the text editor that the grammar is designed for (TextMate, Sublime Text, or Atom) and include that information in your bug report.

You can also try to fix the bug yourself and submit a Pull Request. TextMate's documentation offers a good introduction on how to work with TextMate-compatible grammars. You can test grammars using Lightshow.

Once the bug has been fixed upstream, we'll pick it up for GitHub in the next release of Linguist.

Changing the source of a syntax highlighting grammar

We'd like to ensure Linguist and GitHub.com are using the latest and greatest grammars that are consistent with the current usage but understand that sometimes a grammar can lag behind the evolution of a language or even stop being developed. This often results in someone grasping the opportunity to create a newer and better and more actively maintained grammar, and we'd love to use it and pass on it's functionality to our users.

Switching the source of a grammar is really easy:

script/add-grammar --replace MyGrammar https://github.com/PeterPan/MyGrammar

This command will analyze the grammar and, if no problems are found, add it to the repository. If problems are found, please report these problems to the grammar maintainer as you will not be able to add the grammar if problems are found.

Please only add grammars that have one of these licenses.

Please then open a pull request for the updated grammar.

Testing

You can run the tests locally with:

bundle exec rake test

Sometimes getting the tests running can be too much work, especially if you don't have much Ruby experience. It's okay: be lazy and let our build bot Travis run the tests for you. Just open a pull request and the bot will start cranking away.

Here's our current build status: Build Status

Maintainers

Linguist is maintained with ❤️ by:

  • @Alhadis
  • @BenEddy (GitHub staff)
  • @Caged (GitHub staff)
  • @grantr (GitHub staff)
  • @kivikakk (GitHub staff)
  • @larsbrinkhoff
  • @lildude (GitHub staff)
  • @pchaigno
  • @rafer (GitHub staff)
  • @shreyasjoshis (GitHub staff)

As Linguist is a production dependency for GitHub we have a couple of workflow restrictions:

  • Anyone with commit rights can merge Pull Requests provided that there is a 👍 from a GitHub staff member.
  • Releases are performed by GitHub staff so we can ensure GitHub.com always stays up to date with the latest release of Linguist and there are no regressions in production.

Releasing

If you are the current maintainer of this gem:

  1. Create a branch for the release: git checkout -b cut-release-vxx.xx.xx
  2. Make sure your local dependencies are up to date: script/bootstrap
  3. If grammar submodules have not been updated recently, update them: git submodule update --remote && git commit -a
  4. Ensure that samples are updated: bundle exec rake samples
  5. Ensure that tests are green: bundle exec rake test
  6. Bump gem version in lib/linguist/version.rb, like this.
  7. Make a PR to github/linguist, like this.
  8. Build a local gem: bundle exec rake build_gem
  9. Test the gem:
  10. Bump the Gemfile and Gemfile.lock versions for an app which relies on this gem
  11. Install the new gem locally
  12. Test behavior locally, branch deploy, whatever needs to happen
  13. Merge github/linguist PR
  14. Tag and push: git tag vx.xx.xx; git push --tags
  15. Create a GitHub release with the pushed tag (https://github.com/github/linguist/releases/new)
  16. Build a grammars tarball (./script/build-grammars-tarball) and attach it to the GitHub release
  17. Push to rubygems.org -- gem push github-linguist-3.0.0.gem