Merge pull request #456 from gitpython-developers/fix-for-invalid-data-in-commits

Byron · Byron · commit 79fdaf349fa8 · 2016-05-30T18:58:30.000+02:00
Add test case as example of Git commit with invalid data
diff --git a/git/objects/commit.py b/git/objects/commit.py
@@ -501,14 +501,14 @@ def _deserialize(self, stream):
 
         try:
             self.author, self.authored_date, self.author_tz_offset = \
-                parse_actor_and_date(author_line.decode(self.encoding))
+                parse_actor_and_date(author_line.decode(self.encoding, errors='replace'))
         except UnicodeDecodeError:
             log.error("Failed to decode author line '%s' using encoding %s", author_line, self.encoding,
                       exc_info=True)
 
         try:
             self.committer, self.committed_date, self.committer_tz_offset = \
-                parse_actor_and_date(committer_line.decode(self.encoding))
+                parse_actor_and_date(committer_line.decode(self.encoding, errors='replace'))
         except UnicodeDecodeError:
             log.error("Failed to decode committer line '%s' using encoding %s", committer_line, self.encoding,
                       exc_info=True)
@@ -518,7 +518,7 @@ def _deserialize(self, stream):
         # The end of our message stream is marked with a newline that we strip
         self.message = stream.read()
         try:
-            self.message = self.message.decode(self.encoding)
+            self.message = self.message.decode(self.encoding, errors='replace')
         except UnicodeDecodeError:
             log.error("Failed to decode message '%s' using encoding %s", self.message, self.encoding, exc_info=True)
         # END exception handling
diff --git a/git/test/fixtures/commit_invalid_data b/git/test/fixtures/commit_invalid_data
@@ -0,0 +1,6 @@
+tree 9f1a495d7d9692d24f5caedaa89f5c2c32d59368
+parent 492ace2ffce0e426ebeb55e364e987bcf024dd3b
+author E.Azer Ko�o�o�oculu <azer@kodfabrik.com> 1306710073 +0300
+committer E.Azer Ko�o�o�oculu <azer@kodfabrik.com> 1306710073 +0300
+
+add environjs
diff --git a/git/test/test_commit.py b/git/test/test_commit.py
@@ -306,6 +306,13 @@ def test_serialization_unicode_support(self):
         # it appears
         cmt.author.__repr__()
 
+    def test_invalid_commit(self):
+        cmt = self.rorepo.commit()
+        cmt._deserialize(open(fixture_path('commit_invalid_data'), 'rb'))
+
+        assert cmt.author.name == u'E.Azer Ko�o�o�oculu', cmt.author.name
+        assert cmt.author.email == 'azer@kodfabrik.com', cmt.author.email
+
     def test_gpgsig(self):
         cmt = self.rorepo.commit()
         cmt._deserialize(open(fixture_path('commit_with_gpgsig'), 'rb'))
