Git and Web Content Tricks?

Folks,

I am interested in hearing about any tricks or scripts you have that help you manage your website content on a daily basis.

While I am aware of the post-update hooks that assure a properly checked-out content after a push, I'm also wondering if anyone has dealt with this issue in the face of either beta-test site pre-testing requirements or additional support files such as database files as well.

Specifically, if you had web content that needed to be published only after it has gone through a qual cycle on a beta test platform, how do you account for any minor changes in content before converting it to your production site? Do you run a separate branch for beta differences and merge them up or down to a production branch? Do you run a post- or pre- hook to alter or create config files during the {beta,production} install?

If you had content in your complete git repository that included stuff that wasn't to be deployed to your HTDOC root, how do you account for only a "partial" or subset deployment of your files? For example, if you kept a copy of your backend database scehmas as well as your actual web pages under two separate directories of the same repository, how do you deploy just the web pages and not the schema files? Do you checkout with path limiters? Do you run a Manifest install file and script? Do you rely on some Makefile target scheme?

Any tips or tricks here for the weary?

Thanks! jdl