What is CI?

Continuous integration (CI) is the practice, in software engineering, of merging all developer working copies to a shared mainline several times a day.

Wikipedia https://en.wikipedia.org/wiki/Continuous_integration

What is CI?

Set up Travis for an R package

1. Create .travis.yml file:

language: R
sudo: false
cache: packages

2. Add it to the repository

3. Omit it from the R package, create .Rbuildignore:

^\.travis\.yml$

4. Commit and push

Add a README.md file, and a badge:

[![Build Status](https://travis-ci.org/<user>/<repo>.svg?branch=master)]
(https://travis-ci.org/<user>/<repo>)

Exercises

1. Register on https://travis-ci.org

2. Turn on Travis for your repository

3. Add a .travis.yml file to your repository

4. Push the new version to GitHub

5. Add a badge to your README.md file

What is test coverage?

Code coverage is a measure of the amount of code being exercised by the tests. It is an indirect measure of test quality.

Jim Hester, covr package author

Set up test coverage runs on Travis

Extend your .travis.yml file:

...
r_github_packages:
  - jimhester/covr

warnings_are_errors: true

notifications:
  email:
    on_success: change
    on_failure: change

after_success:
  - Rscript -e 'covr::codecov()'

Exercises

1. Add (dummy?) tests to your repository

2. Add code coverage support on Travis

3. Push the new version to GitHub

4. Add a badge for code coverage to README.md

Value busy maintainers’ time

• No +1 comments or Thank you! comments.
• Do not submit large pull requests without asking first.
• Comply with the repo standards. Really, do what the owner says.
• Don’t be upset if
  • your issue is not answered
  • your PR is ignored
  • you are told to change your PR
  • your PR is rejected
• Friendly reminder comments are OK (private emails not!)

GitHub API

• Manipulating GitHub progratically
• From R: https://github.com/gaborcsardi/gh
• Get a GitHub token
• Set GITHUB_PAT environment variable

Not just code

• git & GitHub are great for R Markdown
• Data analysis, presentations, etc.
• Scientific articles (not in word)
• Anything that is text, really: see e.g. Awesome R

Pain points

• If something goes wrong, it will be hard to fix it. Just don’t tinker with the history. If you do, always make backups first.
• No code review tools with R support
• No multi-platform continuous integration support
• R package GitHub repo is also for distribution: auto-generated files are in the repository, diffs are bloated.
• Price

Additional material

Software carpentry:

http://swcarpentry.github.io/git-novice/

GitHub help pages

https://help.github.com/

Hadley Wickham’s R packages book

http://r-pkgs.had.co.nz/git.html