How do Git ignore rules differ from files already tracked by Git?

Not necessarily. The key distinction is whether Git already tracks the file. Ignore rules are for files that are not already tracked; adding a rule does not affect a file that is already in Git’s index. [S1]

For example, *.log can prevent a newly created, untracked debug.log from being picked up by Git. But if debug.log was already added to the index, that same pattern will not stop Git tracking it. [S1]

So first ask: is this a new file Git has not tracked, or an existing tracked file you now want to leave out of future tracking? The remedy depends on that distinction.

If you know the file was already added to Git, an ignore rule alone will not untrack it. Ignore patterns are intended for files Git is not already tracking. [S1] For a new, untracked file, a matching pattern is the relevant approach. [S1]

For an already tracked file that you want to keep locally, add a matching rule to an ignore file, then remove the file from Git’s index without asking Git to remove your local copy. The documented command pattern is git rm --cached <path>. [S1]

For example, if the file is debug.log, the steps look like this:

# Add this pattern to .gitignore
*.log

# Remove this one file from Git's index
git rm --cached debug.log

Replace debug.log with the path of your file. The ignore pattern and index-removal command do different jobs; use both when you want the file kept locally but no longer tracked. [S1]

Choose the ignore location according to whether the rule should be shared. Repository .gitignore files are an option for patterns intended to be shared; Git also has other exclude sources for rules with a different scope. [S1]

For a file that is already tracked, choosing a local-only or shared ignore location does not change the central issue: an ignore pattern still does not untrack it. [S1] You would still need to remove it from the index with git rm --cached <path> if you want it to remain on your computer but stop being tracked. [S1]

Before changing anything, check that you have the right path and that you intend to stop tracking that file. In the example, *.log is broad: it matches log files, not only debug.log. If you want to ignore just one file, use a pattern appropriate to that file and its location.

Use this sequence:

1. Identify the exact file path and decide whether the rule should be shared.
2. Determine whether Git already tracks the file; an ignore pattern will not untrack it. [S1]
3. Choose a pattern that fits the intended file or files, rather than copying a broad example without checking its scope.
4. If the file is already tracked and should stay on your computer, add the pattern and run git rm --cached <path> for that file. [S1]
5. Review your intended path and pattern before applying the change.

Worked example: for a tracked debug.log you want to keep locally, add a matching *.log pattern if you intend to ignore log files, then run git rm --cached debug.log. A new untracked debug.log can be covered by that pattern without the index-removal step. [S1]

The key distinction is whether the file is tracked: the ignore pattern covers untracked files, while git rm --cached removes a tracked file from the index and retains the local copy. [S1]