GNU patch 2.7.6 always needs a strip option

Fixes #52329
Without this, you get errors such as

>No file to patch.  Skipping patch.
>1 out of 1 hunk ignored
>can't find file to patch at input line 857
>Perhaps you should have used the -p or --strip option?

There may be additional bugs like this if the target is not a directory. The solution should be to default `strip=0`.

Author: OrangeDog

salt/states/file.py

@@ -7268,8 +7268,8 @@ def _patch(patch_file, options=None, dry_run=False):
             # Try to reverse-apply hunks from rejects file using a dry-run.
             # If this returns a retcode of 0, we know that the patch was
             # already applied. Rejects are written from the base of the
-            # directory, so the strip option doesn't apply here.
-            reverse_pass = _patch(patch_rejects, ["-R", "-f"], dry_run=True)
+            # directory, so the required strip level is 0.
+            reverse_pass = _patch(patch_rejects, ["-R", "-f", "-p0"], dry_run=True)
             already_applied = reverse_pass["retcode"] == 0
 
             # Check if the patch command threw an error upon execution