fix(tag): split tag object headers at line feeds only

`TagObject._set_cache_` split the decoded tag object with `str.splitlines()` and then read the headers by position, taking `lines[3]` as the tagger. `str.splitlines()` also breaks at `\r`, `\v`, `\f`, `\x1c`-`\x1e`, `\x85`, U+2028 and U+2029, but git ends a header only at a line feed. A U+2028 inside the `tag` header of an untrusted tag object therefore shifts every later header by one line in GitPython only. With `tag v1<U+2028>tagger Linus Torvalds <torvalds@linux-foundation.org> 1700000000 +0000` followed by a real `tagger Mallory <m@evil.example> ...` line, `git cat-file tag`, `git show` and `git for-each-ref --format=%(taggername)` report Mallory, while GitPython reported `tag == "v1"` and Linus Torvalds as the tagger. The object passes `git fsck --strict` and a clone with `transfer.fsckObjects=true`, because git only warns `badTagName`.

The fix splits at `"\n"` and drops the trailing empty element, so tags that only use line feeds parse exactly as before. A `\r` or U+2028 inside the message is now kept, as git shows it. The new test in `test/test_refs.py` fails before the change and passes after it. `ruff`, `ruff format --check` and `mypy` are clean. On my machine the full suite has 3 failures (`test_empty_repo`, `test_rebasing`, `test_diff_with_staged_file`), and all 3 also fail without this change.

Author: Keerthana-64

git/objects/tag.py

@@ -106,7 +106,16 @@ def _set_cache_(self, attr: str) -> None:
         """Cache all our attributes at once."""
         if attr in TagObject.__slots__:
             ostream = self.repo.odb.stream(self.binsha)
-            lines: List[str] = ostream.read().decode(defenc, "replace").splitlines()
+            # Git ends a header only at a line feed, but str.splitlines() also
+            # breaks at "\r", "\v", "\f", U+2028 and similar. Splitting a tag
+            # object that way lets an untrusted "tag" header containing such a
+            # character shift every later header by a line. Split at "\n" only,
+            # and drop the trailing empty element so tags that use line feeds
+            # parse exactly as before.
+            data = ostream.read().decode(defenc, "replace")
+            lines: List[str] = data.split("\n")
+            if lines and lines[-1] == "":
+                lines.pop()
 
             _obj, hexsha = lines[0].split(" ")
             _type_token, type_name = lines[1].split(" ")

test/test_refs.py

@@ -7,10 +7,12 @@
 import os.path as osp
 import tempfile
 import unittest
+from io import BytesIO
 from itertools import chain
 from pathlib import Path
 from unittest.mock import patch
 
+from gitdb import IStream
 from gitdb.exc import BadName
 
 import git.refs as refs
@@ -139,6 +141,31 @@ def test_tags_author(self):
         tagger_name = tagobj.tagger.name
         assert tagger_name == "Michael Trier"
 
+    def test_tag_headers_split_at_line_feeds_only(self):
+        # A "tag" header value carrying a U+2028 (or any separator that
+        # str.splitlines() honors but git does not) must not shift the later
+        # headers. Git ends a header only at "\n", so GitPython must read the
+        # whole value as the tag name and the following line as the tagger,
+        # exactly as "git cat-file tag" does. See the TagObject parsing fix.
+        with tempfile.TemporaryDirectory() as tmp_dir:
+            with self._repo_with_initial_commit(Path(tmp_dir)) as repo:
+                commit_hexsha = repo.head.commit.hexsha
+                payload = (
+                    f"object {commit_hexsha}\n"
+                    "type commit\n"
+                    "tag v1
tagger Linus Torvalds <torvalds@linux-foundation.org> 1700000000 +0000\n"
+                    "tagger Mallory <m@evil.example> 1700000001 +0000\n"
+                    "\n"
+                    "release\n"
+                ).encode("utf-8")
+                istream = repo.odb.store(IStream("tag", len(payload), BytesIO(payload)))
+                tagobj = TagObject(repo, istream.binsha)
+
+                assert tagobj.tag == "v1
tagger Linus Torvalds <torvalds@linux-foundation.org> 1700000000 +0000"
+                assert tagobj.tagger.name == "Mallory"
+                assert tagobj.tagger.email == "m@evil.example"
+                assert tagobj.message == "release"
+
     def test_tags(self):
         # Tag refs can point to tag objects or to commits.
         s = set()