git: is_tracked: use ls_files instead of walking the index (iterative…

…#3677)

Current implementation doesn't work with directories, which results in
hard-to-debug bugs like iterative#3561. The bug that made us walk through index
manually is now fixed in gitpython, so we can simply use `ls_files` from
now on. Added some tests to prevent this from happening in the future.

Fixes iterative#3561

dvc/scm/git/__init__.py

@@ -271,11 +271,7 @@ def untracked_files(self):
         return [os.path.join(self.repo.working_dir, fname) for fname in files]
 
     def is_tracked(self, path):
-        # it is equivalent to `bool(self.repo.git.ls_files(path))` by
-        # functionality, but ls_files fails on unicode filenames
-        path = relpath(path, self.root_dir)
-        # There are 4 stages, see BaseIndexEntry.stage
-        return any((path, i) in self.repo.index.entries for i in (0, 1, 2, 3))
+        return bool(self.repo.git.ls_files(path))
 
     def is_dirty(self):
         return self.repo.is_dirty()

tests/unit/scm/test_git.py

@@ -37,3 +37,37 @@ def test_walk_with_submodules(tmp_dir, scm, git_dir):
     # currently we don't walk through submodules
     assert not dirs
     assert set(files) == {".gitmodules", "submodule"}
+
+
+def test_is_tracked(tmp_dir, scm):
+    tmp_dir.scm_gen(
+        {
+            "tracked": "tracked",
+            "dir": {"data": "data", "subdir": {"subdata": "subdata"}},
+        },
+        commit="add dirs and files",
+    )
+    tmp_dir.gen({"untracked": "untracked", "dir": {"untracked": "untracked"}})
+
+    # sanity check
+    assert (tmp_dir / "untracked").exists()
+    assert (tmp_dir / "tracked").exists()
+    assert (tmp_dir / "dir" / "untracked").exists()
+    assert (tmp_dir / "dir" / "data").exists()
+    assert (tmp_dir / "dir" / "subdir" / "subdata").exists()
+
+    assert not scm.is_tracked("untracked")
+    assert not scm.is_tracked(os.path.join("dir", "untracked"))
+
+    assert scm.is_tracked("tracked")
+    assert scm.is_tracked("dir")
+    assert scm.is_tracked(os.path.join("dir", "data"))
+    assert scm.is_tracked(os.path.join("dir", "subdir"))
+    assert scm.is_tracked(os.path.join("dir", "subdir", "subdata"))
+
+
+def test_is_tracked_unicode(tmp_dir, scm):
+    tmp_dir.scm_gen("ṭṝḁḉḵḗḋ", "tracked", commit="add unicode")
+    tmp_dir.gen("ṳṋṭṝḁḉḵḗḋ", "untracked")
+    assert scm.is_tracked("ṭṝḁḉḵḗḋ")
+    assert not scm.is_tracked("ṳṋṭṝḁḉḵḗḋ")