`.gitignore` files outside of the git repository are taken in account.

[mefyl]

Description

I have a .gitignore file at the root of my home repository and tagref will erroneously apply its rules in any git repository in my account.

Instructions to reproduce the bug

$ mkdir repo
$ cd repo
$ git init
[...]
$ echo '(* some [tag:tag] *)' > source.ml
$ tagref
1 tag, 0 tag references, 0 file references, and 0 directory references validated in 1 file.
$ echo '*' > ../.gitignore
# tagref does not see the tag anymore.
$ tagref                  
0 tags, 0 tag references, 0 file references, and 0 directory references validated in 0 files.
# Yet this gitignore file does *not* apply here, see:
$ git status
On branch master

No commits yet

Untracked files:
  (use "git add <file>..." to include in what will be committed)
        source.ml

nothing added to commit but untracked files present (use "git add" to track)
$ tagref --version
Tagref 1.10.0
$

Environment information:

• Tagref version: [e.g. 1.10.0]

[stepchowfun]

Tagref uses the ignore crate for this functionality, which I think is what ripgrep uses. Can you check if ripgrep has the same issue? Maybe it needs to be fixed upstream? Or it could be related to this:

tagref/src/walk.rs

Line 28 in d06b705

.require_git(false)

[mefyl]

Ripgrep does not seem to be affected:

$ mkdir repo 
$ cd repo 
$ git init
[...]
$ echo '(* some [tag:tag] *)' > source.ml
# ripgrep OK
$ rg tag:tag
source.ml
1:(* some [tag:tag] *)
# tagref OK
$ tagref 
1 tag, 0 tag references, 0 file references, and 0 directory references validated in 1 file.
$ echo '*' > ../.gitignore
# ripgrep OK
$ rg tag:tag              
source.ml
1:(* some [tag:tag] *)
# tagref KO
$ tagref 
0 tags, 0 tag references, 0 file references, and 0 directory references validated in 0 files.
$ rg --version
ripgrep 14.1.1

features:+pcre2
simd(compile):+SSE2,+SSSE3,+AVX2
simd(runtime):+SSE2,+SSSE3,+AVX2

PCRE2 10.43 is available (JIT is available)
$ 

[stepchowfun]

Thanks for confirming that! I don't know what's causing the issue is at the moment, but I'd happily review a PR to fix it.