"Test" Branch on GitHub Repo

Question

On my webserver I currently have a Virtual Host, with an almost exact copy of my main public website, set up for testing purposes so I can edit my website and test new features without editing my live website. It has a simple .htaccess and .htpasswd authentication using

AuthType Basic
AuthName "Restricted access"
AuthUserFile /home/USERNAME/.htpasswd
require valid-user

I am wondering if it is possible to have a branch from my main website GitHub repo that, when I merge the changes from the test branch to the master branch, it will not merge the .htaccess and .htpasswd files.

I'm also wondering if it is possible to have the test branch link to a different folder on my webserver. For example the master branch will be linked to /var/www/html and I would like the test branch to be linked to /var/www/test.

Please make answers idiot proof as I still only have a very basic understanding of Git and GitHub

Answer 1

This answer kind of side-steps the issue by avoiding the problem (and added as a more detailed follow-up to my comment). Let's say you have two branches (for simplicity), master where all dev & testing happens (note: normally I'd have separate branches for both dev & test, though), and a second branch for production, prod. I'll assume these are both already set up, and identical, and currently contains your production website. (I don't know which OS you use for development, but I'll assume linux?) Then, make some changes in master, deploy to your test staging area using a script something like deploy.sh:

#!/bin/bash
# usage: deploy [test | prod]
deploy_test() {
    git checkout master
    # copy site to remote
    scp -r ./path/to/site me@host:/var/www/test

    # or, you could generate htpasswd here
    read -p 'Input path to your htpasswd file: ' pw 
    scp -r $pw ./path/to/site me@host:~/
}

deploy_prod() {
    git checkout prod
    # copy site to remote
    scp -r ./path/to/site me@host:/var/www/html
}

if [ "$1" = "prod" ]; then 
    deploy_prod
elif [ "$1" = "test" ]; then 
    deploy_test
fi

Ok, that's a crazy simple example, but that same idea can be implemented in whatever you're using for deployment. Everything is checked-in to git, except anything related to authentication (passwords, users), which are either env vars or typed into the deploy script. Further, a tool like jenkins ( https://jenkins-ci.org/ ) could call these scripts, and be used for test / deployment from a QA server to your staging host ...ok, you can see how things can slowly move from "small script" to "deployment pipeline" -- the key is keep each step simple, documented, and build it out slowly. Keep doing the simplest thing possible, and build on that. (note: I just typed all this into S.O., so it's an example only, not production code :-) )

Answer 2

Whenever you need to keep different content based on branches, one way is to:

• version only a template file
• seek the value files outside the repo (if they are confidential like htpasswd should be) or in versioned files named after the branch (if you want to track them in the same repo, like .htaccess.master, .htaccess.test)
  That way, there is no merge issue.

For that, you would register (in a .gitattributes declaration) in your submodule repo a content filter driver.

^{(image from "Customizing Git - Git Attributes", from "Pro Git book")}

The smudge script, associate to the template file (.htpasswd.tpl, .htaccess.tpl), would generate (automatically on git checkout) the properties file by looking in outside source the actual values. The generated actual .htxxx properties files remain ignored (by the .gitignore).

See a complete example at "git smudge/clean filter between branches".

Your smudge script can determine the name of the checked out branch with:

branch=$(git rev-parse --symbolic --abbrev-ref HEAD)