Gitignore Tester

"Why doesn't my .gitignore negation work?" is one of the most common git questions there is, and the honest answer is almost always the same: a `!pattern` trying to re-include a file inside a directory that an earlier rule already excludes can never work, because git does not descend into ignored directories to re-check anything inside them. This tool reimplements git's actual .gitignore matching algorithm — not an approximation — specifically so it can catch and explain that exact situation, along with every other pattern rule git supports.

Paste your .gitignore rules into one box and a list of file paths (one per line, like `dist/index.js`, `node_modules/foo/bar.js`, or `.env`) into the other, and the tool evaluates each path exactly the way git would: comments and blank lines are skipped, trailing spaces are stripped unless escaped, a leading `!` negates a rule, a leading `/` anchors it to the root instead of letting it match at any depth, a trailing `/` restricts it to directories, and `*`, `?`, `[abc]` character classes and `**` (matching across directory boundaries) all work per the documented glob rules.

Critically, paths are checked directory by directory from the root down, the same way git's own ignore engine works — not by testing the full path against every rule in isolation. That is what makes it possible to correctly detect the dead-negation case: if `dist/` is ignored on line 3 and a later `!dist/keep.js` tries to rescue one file inside it, this tool reports `dist/keep.js` as still ignored, names the line-3 rule as the reason, and separately flags the negation as shadowed rather than silently pretending it worked.

The result table gives every tested path its own row with a clear yes/no ignored verdict, the exact rule text and line number that decided it, and — whenever relevant — a plain-language note explaining a shadowed negation or an inherited parent-directory rule. That is the detail that turns a confusing .gitignore into one you can actually reason about before committing.