Today I faced problems when tried to write a .gitignore file.

Motivation

We have a large project (300k files, 20Gb of source). My team uses a centralized version control system. Sometimes it’s not handy, so I want to use git to manage my local changes. The whole repository is too big. It takes about 30 seconds on a good machine to make a simple commit or to switch branches. I work only on a subset of files, so it’s reasonable to write .gitignore to keep track only of several subdirectories, right?

Goal

Write .gitignore, that ignores all files except a directory named source/good (with subdirectories and subfiles). source/good can contain arbitrary depth files in dirs hierarchy.

Check your solution

I have written two simple scripts (will work only on unix with installed git).

1. unzip task.zip in empty directory (scripts will remove .git/ directory).
2. sh create.sh.
3. create and write a proper .gitignore in the root directory.
4. sh test.sh. It will tell you the result.