repo-man

(c) 2008 Jason Frame [jason@onehackoranother.com]

What is it?

One of the most tedious things about hosting Subversion repositories over HTTP is that staying organised requires frequent editing of your Apache configuration. At work we use a directory structure similar to that show below (repositories are denoted by [R]):

/
  repo/
    internal/
      project-1/ [R]
      project-2/ [R]
      project-/3 [R]
    client/
      client-1/
        project-4/ [R]
        project-5/ [R]
      client-2/
        project-6/ [R]
        project-7/ [R]

The repository roots in the example above are /repo/{internal,client/{client-1,client-2}}, and each requires its own <Location> block and SVNParentPath directive within your httpd.conf. Access control configuration may also be required. As the number of repository roots increases, maintenance becomes increasingly depressing.

Enter repo-man. It's a (very) simple Ruby tool capable of generating Apache/SVN config fragments for multiple repository roots.

Installation

sudo gem install jaz303-repo-man

Usage

repo-man path-to-your-repository-root

This will spit out Apache configuration code suitable for pasting or dynamic inclusion into a VirtualHost block. The path you supply is assumed to correspond to the target virtual host's DocumentRoot and the repository locations will be normalised accordingly.

If any of your repositories has specific access control requirements put them in a file called svn_access in the corresponding root, or any of its ancestor directories, and repo-man will import this using the AuthzSVNAccessFile directive. The format of this file is explained at http://svnbook.red-bean.com/en/1.1/ch06s04.html, about 2/3 of the way down.