Join GitHub today
GitHub is home to over 28 million developers working together to host and review code, manage projects, and build software together.Sign up
Fetching latest commit…
Cannot retrieve the latest commit at this time.
|Failed to load latest commit information.|
IMPORTANT UPDATE: This is no longer supported/maintained, and will most surely stop working on Jan 27, 2012 when Pivotal deprecates the Tracker API's prior to v3. Also, the standard GitHub-Pivotal Tracker service hook now handles everything this service did, and in fact works better now (better handles branch commits and avoiding duplicate comments in tracker due to merging branches). This app is a small server to serve as a GitHub Post-Receive hook to add comments, and update state in Pivotal Tracker, similar to say the Lighthouse service integration. As of January 23, 2010, Pivotal's Tracker API itself will have some support for doing this. See: http://pivotallabs.com/users/dan/blog/articles/1135-pivotal-tracker-api-new-version-v3-to-be-released-on-jan-23 However, note that, as of this writing, they only support a single API token, which means that all commits will be attributed to the Tracker member whose token is being used in the post-commit hook. This project solves that by associating Tracker API tokens with GitHub users (so even if your GitHub and Tracker emails are different that's ok), and thus you have properly attributed comments in your Tracker stories when making a commit. Also, many thanks to Alan Pinstein (apinstein) for his many contributions. Configure your Tracker API key, and Project ID in a config.yml file placed in the same directory as this app. It should look something like: tracker_github_hook: github_url: 'http://github.com/chris/tracker_github_hook' tracker_api_token: a1230e72340e3babc96d5e2fab67c18d tracker_project_id: 123 ref: refs/heads/master user_api_tokens: chris: email: firstname.lastname@example.org tracker_api_token: a1b2c3d4e5f67890 alan: email: email@example.com tracker_api_token: 0987654321abcdef The label ('tracker_github_hook' in this case) is arbitrary and not used, it's just their to be a useful bit of info to humans/organize the nested settings, and is not used. This setup allows you to have one service that supports multiple Tracker/GitHub projects, just define one of the above blocks for each one, and then anytime GitHub sends a push, the service will tease out which GitHub repo it came from and correlate that to which Tracker project you've assigned to that. The "ref" field is also optional; it will tell the hook to ignore commits on any branches other than the one listed. This is useful to prevent duplication of comments being pushed into Tracker if you have multiple remote branches. The "user_api_tokens" is optional - the primary "tracker_api_token" will be used by default. But, if you do supply this block, then the hook will correlate the email address of the author of the GitHub commit, to that within this list, and if it finds a match, will use the specified Tracker API token. This makes it so that the comment in your Tracker story shows up as being made by the same person making the GitHub commit (instead of whoever owns the default API token). When you make commits to Git/GitHub, and want a comment and optionally a state update made to a story in Tracker, add the following text to your commit message: [Story#####] or [Story##### state:finished] where ##### is the story number (see the bottom of an expanded story in Tracker for its ID). A commit message can have more than one [Story####] block, but the entire commit message will be added to both stories. The duplication is unfortunate but better than the alternative of ignoring additional story references altogether. This project also requires the following rubygems: - sinatra - rest-client - json More information for, and thanks to: Pivotal Tracker API: http://www.pivotaltracker.com/help/api GitHub Post-Receive Hooks: http://github.com/guides/post-receive-hooks Sinatra: http://sinatra.rubyforge.org/ RestClient: http://rubyforge.org/projects/rest-client/ TODO: - cleanup, better testing - support other story changes, like assigned user, etc.