This is an old revision of the document!


Gitorious

Display project repositories from database gitorious:

SELECT slug, name FROM repositories r INNER JOIN projects p ON r.project_id = p.id WHERE r.name NOT LIKE '%wiki' ORDER BY p.slug, r.name;

Runtime Init Dependencies

  • mysql
  • stompserver

Installation

  • git clone git://gitorious.org/gitorious/mainline.git gitorious
  • cp config/database.sample.yml to config/database.yml and edit
  • cp config/gitorious.sample.yml to config/gitorious.yml and edit
  • bundle exec rake db:create:all (creates MySQL databases)
  • bundle exec rake db:setup
  • script/poller start
  • script/gitorious -d -e production

Configuration

Gitorious is installed at /var/www/gitorious

The scripts that should be run are in the script directory, and are: polling and gitorious. Init scripts were created to start the services. Everything must be run under user 'git'.

The git repositories are stored at /var/git

If starting the gitorious script, the RAILS_ENV is “production”. See gitorious -h for more options.

Git Access

All git access is run through user 'git', and is accessible by using SSH public keys. The keys can be added through the gitorious web interface.

Apache

It is possible to run gitorious through Apache, but git ssh access would not work. Instead, we are running it as a separate daemon.

Search is made available by running ultrasphinx. There's a custom init script named gitorious-search which is the original searchd script shipped with the ebuild, modified to run as user git.

The configuration file in /etc/sphinx/sphinx.conf is copied from /var/www/gitorious/config/ultrasphinx/production.conf

Support