add support for the git option "index.skipHash"

dulwich/index.py

@@ -451,7 +451,7 @@ def read(self) -> None:
             self.update(read_index_dict(f))
             # FIXME: Additional data?
             f.read(os.path.getsize(self._filename) - f.tell() - 20)
-            f.check_sha()
+            f.check_sha(allow_empty=True)
         finally:
             f.close()
 

dulwich/pack.py

@@ -1605,9 +1605,13 @@ def read(self, num=None):
         self.sha1.update(data)
         return data
 
-    def check_sha(self) -> None:
+    def check_sha(self, allow_empty: bool = False) -> None:
         stored = self.f.read(20)
-        if stored != self.sha1.digest():
+        # If git option index.skipHash is set the index will be empty
+        if stored != self.sha1.digest() and (
+            not allow_empty
+            or sha_to_hex(stored) != b"0000000000000000000000000000000000000000"
+        ):
             raise ChecksumMismatch(self.sha1.hexdigest(), sha_to_hex(stored))
 
     def close(self):
@@ -2485,9 +2489,9 @@ def __iter__(self):
 
     def check_length_and_checksum(self) -> None:
         """Sanity check the length and checksum of the pack index and data."""
-        assert len(self.index) == len(self.data), (
-            f"Length mismatch: {len(self.index)} (index) != {len(self.data)} (data)"
-        )
+        assert len(self.index) == len(
+            self.data
+        ), f"Length mismatch: {len(self.index)} (index) != {len(self.data)} (data)"
         idx_stored_checksum = self.index.get_pack_checksum()
         data_stored_checksum = self.data.get_stored_checksum()
         if idx_stored_checksum != data_stored_checksum:

tests/test_index.py

@@ -84,6 +84,9 @@ def test_len(self) -> None:
     def test_iter(self) -> None:
         self.assertEqual([b"bla"], list(self.get_simple_index("index")))
 
+    def test_iter_skip_hash(self) -> None:
+        self.assertEqual([b"bla"], list(self.get_simple_index("index_skip_hash")))
+
     def test_iterobjects(self) -> None:
         self.assertEqual(
             [(b"bla", b"e69de29bb2d1d6434b8b29ae775ad8c2e48c5391", 33188)],

tests/test_pack.py

@@ -1001,9 +1001,9 @@ def _result(self, unpacked):
         )
 
     def _resolve_object(self, offset, pack_type_num, base_chunks):
-        assert offset not in self._unpacked_offsets, (
-            f"Attempted to re-inflate offset {offset}"
-        )
+        assert (
+            offset not in self._unpacked_offsets
+        ), f"Attempted to re-inflate offset {offset}"
         self._unpacked_offsets.add(offset)
         return super()._resolve_object(offset, pack_type_num, base_chunks)