index: write blobs via git hash-object, not gitdb's odb.store

Author: caroescm

git/index/base.py

@@ -11,15 +11,13 @@
 import contextlib
 import datetime
 import glob
-from io import BytesIO
 import os
 import os.path as osp
 from stat import S_ISLNK
 import subprocess
 import sys
 import tempfile
 
-from gitdb.base import IStream
 from gitdb.db import MemoryDB
 
 from git.compat import defenc, force_bytes
@@ -34,6 +32,7 @@
     LockedFD,
     join_path_native,
     file_contents_ro,
+    hex_to_bin,
     _is_path_rooted,
     _to_relative_path,
     to_native_path_linux,
@@ -58,7 +57,6 @@
 
 from typing import (
     Any,
-    BinaryIO,
     Callable,
     cast,
     Dict,
@@ -726,24 +724,35 @@ def _store_path(self, filepath: PathLike, fprogress: Callable) -> BaseIndexEntry
         """
         st = os.lstat(filepath)  # Handles non-symlinks as well.
 
+        fprogress(filepath, False, filepath)
         if S_ISLNK(st.st_mode):
             # In PY3, readlink is a string, but we need bytes.
             # In PY2, it was just OS encoded bytes, we assumed UTF-8.
-            def open_stream() -> BinaryIO:
-                return BytesIO(force_bytes(os.readlink(filepath), encoding=defenc))
+            #
+            # `git hash-object` on a symlink path follows the link and hashes the
+            # target file's content, whereas Git stores the link's target string as
+            # the blob content. So the target string is written to a regular
+            # temporary file and hashed from there instead of passing the symlink
+            # path directly.
+            target = force_bytes(os.readlink(filepath), encoding=defenc)
+            fd, tmp_path = tempfile.mkstemp()
+            try:
+                os.write(fd, target)
+                os.close(fd)
+                hexsha = self.repo.git.hash_object(tmp_path, w=True, no_filters=True)
+            finally:
+                os.remove(tmp_path)
         else:
-
-            def open_stream() -> BinaryIO:
-                return open(filepath, "rb")
-
-        with open_stream() as stream:
-            fprogress(filepath, False, filepath)
-            istream = self.repo.odb.store(IStream(Blob.type, st.st_size, stream))
-            fprogress(filepath, True, filepath)
+            # Let the `git` binary write the object. This matches `git add`'s
+            # behavior exactly, including applying any clean filters configured via
+            # `.gitattributes`, and avoids reimplementing object creation (and its
+            # object-file permission handling) in Python.
+            hexsha = self.repo.git.hash_object(filepath, w=True)
+        fprogress(filepath, True, filepath)
         return BaseIndexEntry(
             (
                 stat_mode_to_index_mode(st.st_mode),
-                istream.binsha,
+                hex_to_bin(hexsha),
                 0,
                 to_native_path_linux(filepath),
             )