A .gitignore file is an important part of any project that is being version controlled. Adding things like node_modules to your git ignored files is essential.
A .gitignore file is an important part of any project that is being version controlled via git.
Update: While talking with Thai Wood about this problem today, we looked around and found a generator that will create your .gitignore based on IDEs, Operating Systems and Programming Languages, which is awesome!
If you've ever worked on a project with someone who uses their finder to navigate their filesystem, odds are you've seen this ugly file rear it's head.
.DS_Store is a hidden file that gets generated whenever someone opens a directory in the finder. It holds metadata about the folder, such as "where the icons should be positioned in the finder" when in the directory. This is useless to anyone other than you, so it's a good file to add to your
As a developer, if you navigate your filesystem with Finder,
.DS_Store files will be generated and modified frequently, so ignoring these is incredibly important (I'd also suggest learning to use the Terminal...it's significantly more powerful and robust for developers).
If you use any language that keeps a project's dependencies inside of the project itself (such as
node_modules for NodeJS), it's best to add the dependency directory to your .gitignore. You don't want to version control someone else's code in your code. If you're worried about keeping the same versions, you can always use
If you happen to use something like bower for things like font-awesome or jQuery or if you just copy/paste third party code into your project, it's best to not version control this stuff either. I've seen bootstrap in version control in a few projects and if you're building your production application with a sensible build tool (and using some kind of dependency manager), you shouldn't ever need to keep any of these external dependencies in version control.
I've been successfully (and happily) using node modules for all of my server and client-side dependencies, so ignoring the
node_modules directory takes care of this for me as well.
Your logs are really only useful for you and only useful for a specific environment at a specific time, which basically means keeping them in version control is going to cause all kinds of weird merging issues and extra cruft in your project that's not useful.
If I'm building frontend code using something like Webpack or Gulp, I'll always add
dist which where I build my static assets, (some people use the
public directory) to my .gitignore. My rule of thumb is "Is this directory a source of truth or a representation of a source of truth somewhere else in my project?"
For example, if I use sass, it doesn't make sense to version control the css files that are the result of precompiling from the sass files.
You may have a file with API keys in it (for instance, AWS used to recommend keeping these files in your project and not in your home directory), you should NEVER version control anything that has keys or passwords because you're giving the entire world the ability to see those keys...you don't want pictures of sad kittens on your production eCommerce website, so keep these out of your project by adding them to your .gitignore.
Those are the specific files and general types of files I add to my .gitignore file. Ultimately, I think some good general rules of thumb for what to add to a .gitignore are:
I'd say those 4 questions cover 99% of the use cases you'd ever need to consider.
What things do you add to your .gitignore file that I've missed here?
Subscribe to my Newsletter
Every other week I publish the Curiously Crafted newsletter.
In it, I explore the intersection of curiosity and craft: the people who make stuff, what they make and the way they pursue the craft of making.
On The Web
See all of the places I have a presence on the web.