The majority of the data is exposed via the environment and the actual body is
piped to the script. This change allows for a great deal of flexibility with
regards to templating---users may write "templates" in whatever language they
choose to generate whatever output they wish.
We should do as much processing beforehand as possible so that the template does
not need to, for example, start snooping through the repos (which will be more
complicated once we support multiple VCSs).
Certain things are thrown in place just to demonstrate the concept. Many things
are also hard-coded that should be configurable (such as the title, copyright
and RSS item count).
Licensed under the GPLv3 instead of the AGPLv3 because the license does not
apply to generated content unless that content uses a portion of the program.
Mike Gerwitz