A way to validate .gitignore file

Question

Is there a way to validate a .gitignore file so you quickly find doubles paths or paths that don\'t exist anymore? Normally when a project is small this isn\'t really necess

Answer 1

In order to also validate git's ** pattern in paths I had to write a one-liner inspired by the script above:

find . -type f | git check-ignore -v --stdin | perl -pe 's,.*\.gitignore:,,; s,\t," "x100,e' | sort | uniq -w 100 -c | perl -pe 's, {100}.*,,'

Not exactly pretty but it works.

Answer 2

Not exactly.
One command which could come close would be git check-ignore

Choose a file you know is supposed to be ignored, and check the output of:

git check-ignore -v -- /path/to/ignored/file

You will see the rules of your .gitignore which apply.

---

Update March 2016: Git 2.8 will add a new way to debug .gitignore files and their rules.

In the context of allowing a sub-folder to not be ignored (even if its parent folder is ignored: see example here), I have found this gem by Thái Ngọc Duy (pclouds):

dir.c: support tracing exclude

man git includes:

GIT_TRACE_EXCLUDE:

Enables trace messages that can help debugging .gitignore processing.
See 'GIT_TRACE' for available trace output options.

With GIT_TRACE_EXCLUDE set to 1, you will see (after a git status) stderr debug messages like:

exclude: from ...
exclude: xxx => n/a
exclude: xxx vs. yyy at line z: => www

Answer 3

You can do a script to check it. I have made one for you there:

#!/bin/bash
set -o noglob
for file in `cat .gitignore | grep -v \#`
do
    printf "$file"
    find . -name "$file" | wc -l
done

it will display the rules followed by the number of match in the current directory and recursively. Example:

*.log     31
*.gz      0
*~     42
*.swp      0
*.aux     33
*.pdf     51
*.out      7
*.toc      6
*.nav      1
*.snm      1
.DS_Store      0

You could restrict the output to the line containing 0 by piping into egrep "\b0\b" if you want.