Git ignore generated extensions

I’d like to keep JavaScript (.js) files under git control.

However, if a TypeScript (.ts) file exists, I’d like to ignore the associated .js file.

  • Git continues to push a .c9 folder despite it being in .gitignore
  • Import a git repo into an empty branch
  • git format-patch < everything since last svn rebase >
  • git checkout throws an error on --assume-unchanged files
  • Remove remote branches in Git
  • What is the difference between `git diff` and `git difftool`?
  • Is there a way to achieve this with git?

    EDIT: Preferably without changing directories and refactoring all the existing code.

  • SVN history is lost in git
  • TeamCity - Git push from agent
  • LibGit2Sharp Find what files were updated/added/deleted after pull
  • How to manage repositories on GitHub
  • Remove contents of a commit from a git repo after its already been git -rm'ed and pushed
  • How to make git log cut long comments?
  • One Solution collect form web for “Git ignore generated extensions”

    A typical way to do this is to put any compiled output in some distribution directory. You might organized your directory structure like so:

    src/
      ... various JavaScript files, etc.
    
    dist/
      ... compiled output from any build tool you're using
    

    Then in git, you can just ignore the dist directory, and assume that anyone who is going to pull down that repo and work on it will either run the build tool manually, or it will built automatically as part of the installation process.

    In other words, it’s easier to organize it without depending on a specific set of git functionality. This is a pretty common pattern in the JavaScript world.

    If you don’t go for that method, another way to do it is to add some identifier to the files that you want to ignore. For example, you might prepend all your files you want to ignore with _ and then add this rule to your .gitignore (see this answer for why the following works):

    \_*
    

    There are some drawbacks to this second approach:

    1. Other libraries you use may use filenames that start with a similar identifier and there will be collisions
    2. It’s hard to maintain; you may want to change the identifier later if there is some collision or you change your mind about your structure, and you’ll have to go back through all the files and change the names (of course, this can be automated, but it might be a pain)

    Hope that helps.

    Git Baby is a git and github fan, let's start git clone.