Grudge

Grudge is a stand-alone system, which means you run it on your own, that watches the comments of some git repository and provides a web interface to allow people to rate the comments.

Why would someone want to do this?

In my experience, code commit comments have been very funny; or at least very good reads. It's kind of an insight into some of the programmer's minds. At one point, I was scouring the commit comments and sending a daily email of the ones I thought were funny. The emails were well received, so I figured there must be something to this.

Thus, the goals are:

Make people laugh
Inform others - besides those in development - of the day-to-day

I also hope that something like this will make for better commit comments; since you know you are now being watched ... and judged.

Why is it called Grudge?

Well, basically I typed 'rate' into Apple's thesaurus app and noticed the word 'judge'. I wanted to add a 'G' to the front of the app - since this is a git commit watcher - and so hit upon 'grudge'.

Alas, Grudge = Git Judge

How does it work?

Once your system is bootstrapped, grudge will monitor for updates. For each commit, grudge will store the SHA hash of the commit-id (for referential reasons) and the date it was committed (for ordering reasons). Commit messages and other commit data (like author) stay in the git repository and are always retrieved from there; even at a possible speed reduction (though grudge seems pretty quick).

Each vote is recorded for a commit as either a +1 or a -1. Votes are themselves entities, which means you could get the voting history for any commit.

Commits have a net score, which is simply the sum of all of the votes. Popular commits are those with the highest net score. Unpopular are those with the lowest net score.

Using Grudge

Requirements

Sinatra
Ruby Git
HAML
Sqlite3-ruby
DataMapper - sudo gem install data_mapper
DataObjects - sudo gem install data_objects ... if not already installed
DataObject bindings for Sqlite - sudo gem install do_sqlite3
JSON Pure - sudo gem install json_pure - annoyingly needed by data_mapper

Setting up an app

Install the gem

$ sudo gem install thumblemonks-grudge

Setup an instance

$ mkdir foo && cd foo
$ grudge init

Tell Grudge about your repository

Edit the file in your Grudge instance called config/grudge.yml. In it edit the git repository url that you want Grudge to watch. For now, do this for both development and production entries. You don't need to edit anything else, unless you just can't help yourself.

Example:

...
repository:
  url: git://github.com/thumblemonks/grudge.git
...

Run the app

When just playing around, you can run:

$ grudge start

This will first clone the remote repository you defined, bootstrap the commits in a local Sqlite database, then run the Sinatra web-app. As long as you don't remove clone.git, the cloning will only happen on the first run. Grudge will, however, grab new updates from the repo each time you start it.

Point your browser at http://localhost:9292/ and you are done.

Oh yeah ... advertise it!

Phusion

It is suggested that you run grudge through Apache via Phusion Passenger. In order to do this, simply point a configured Phusion Virtual Host at the public directory wherever you ran grudge init.

Getting automatic updates

As of this writing the way to tell Grudge to automatically look for updates is to setup a remote repository via a post-receive hook. This means your instance of Grudge will need to be routable by your git server.

If this is so, just setup a post-receive hook to hit the URL with a POST request:

http(s)://your-grudge-instance/repository/pull

Manual updates

You can update manually by either restarting the app or via the Grudge console like so:

$ grudge pull

You could also setup a bash alias like the following:

wget -q --post-data='\'''\'' --no-cookies --delete-after http://your-grudge-instance/repository/pull

Not that I did that ;)

Acknowledgements

Someone for sure. I guess my wife for giving me the Saturday to write this.