cmd: diff: don't load dir cache for diff (iterative#3468)

dvc/repo/diff.py

@@ -22,29 +22,25 @@ def _paths_checksums():
         A dictionary of checksums addressed by relpaths collected from
         the current tree outputs.
 
-        Unpack directories to include their entries
-
         To help distinguish between a directory and a file output,
         the former one will come with a trailing slash in the path:
 
             directory: "data/"
             file:      "data"
         """
-        result = {}
-
-        for stage in self.stages:
-            for output in stage.outs:
-                if not output.is_dir_checksum:
-                    result.update({str(output): output.checksum})
-                    continue
-
-                result.update({os.path.join(str(output), ""): output.checksum})
-
-                for entry in output.dir_cache:
-                    path = str(output.path_info / entry["relpath"])
-                    result.update({path: entry["md5"]})
 
-        return result
+        def _to_path(output):
+            return (
+                str(output)
+                if not output.is_dir_checksum
+                else os.path.join(str(output), "")
+            )
+
+        return {
+            _to_path(output): output.checksum
+            for stage in self.stages
+            for output in stage.outs
+        }
 
     working_tree = self.tree
     a_tree = self.scm.get_tree(a_rev)

tests/func/test_diff.py

@@ -3,6 +3,8 @@
 import pytest
 import shutil
 
+from funcy import first
+
 from dvc.compat import fspath
 from dvc.exceptions import DvcException
 
@@ -40,11 +42,7 @@ def test_no_cache_entry(tmp_dir, scm, dvc):
     dir_checksum = "5fb6b29836c388e093ca0715c872fe2a.dir"
 
     assert dvc.diff() == {
-        "added": [
-            {"path": os.path.join("dir", ""), "hash": dir_checksum},
-            {"path": os.path.join("dir", "1"), "hash": digest("1")},
-            {"path": os.path.join("dir", "2"), "hash": digest("2")},
-        ],
+        "added": [{"path": os.path.join("dir", ""), "hash": dir_checksum}],
         "deleted": [],
         "modified": [
             {
@@ -126,15 +124,13 @@ def test_directories(tmp_dir, scm, dvc):
                 "path": os.path.join("dir", ""),
                 "hash": "5fb6b29836c388e093ca0715c872fe2a.dir",
             },
-            {"path": os.path.join("dir", "1"), "hash": digest("1")},
-            {"path": os.path.join("dir", "2"), "hash": digest("2")},
         ],
         "deleted": [],
         "modified": [],
     }
 
     assert dvc.diff(":/directory", ":/modify") == {
-        "added": [{"path": os.path.join("dir", "3"), "hash": digest("3")}],
+        "added": [],
         "deleted": [],
         "modified": [
             {
@@ -144,16 +140,12 @@ def test_directories(tmp_dir, scm, dvc):
                     "new": "9b5faf37366b3370fd98e3e60ca439c1.dir",
                 },
             },
-            {
-                "path": os.path.join("dir", "2"),
-                "hash": {"old": digest("2"), "new": digest("two")},
-            },
         ],
     }
 
     assert dvc.diff(":/modify", ":/delete") == {
         "added": [],
-        "deleted": [{"path": os.path.join("dir", "2"), "hash": digest("two")}],
+        "deleted": [],
         "modified": [
             {
                 "path": os.path.join("dir", ""),
@@ -164,3 +156,24 @@ def test_directories(tmp_dir, scm, dvc):
             }
         ],
     }
+
+
+def test_diff_no_cache(tmp_dir, scm, dvc):
+    (stage,) = tmp_dir.dvc_gen(
+        {"dir": {"file": "file content"}}, commit="first"
+    )
+    scm.tag("v1")
+    tmp_dir.dvc_gen(
+        {"dir": {"file": "modified file content"}}, commit="second"
+    )
+
+    os.remove(first(stage.outs).cache_path)
+    shutil.rmtree("dir")
+
+    # invalidate_dir_info to force cache loading
+    dvc.cache.local._dir_info = {}
+
+    diff = dvc.diff("v1")
+    assert diff["added"] == []
+    assert diff["deleted"] == []
+    assert first(diff["modified"])["path"] == os.path.join("dir", "")