Make sure to decode manifests into utf-8 when necessary

This fixes a decoding error

image/docker/test/test_schema1.py

@@ -1,10 +1,13 @@
+# -*- coding: utf-8 -*-
+
 import os
-import hashlib
 import json
 
 import pytest
 
-from image.docker.schema1 import MalformedSchema1Manifest, DockerSchema1Manifest
+from app import docker_v2_signing_key
+from image.docker.schema1 import (MalformedSchema1Manifest, DockerSchema1Manifest,
+                                  DockerSchema1ManifestBuilder)
 
 @pytest.mark.parametrize('json_data', [
   '',
@@ -130,3 +133,32 @@ def test_validate_manifest_with_unicode_encoded():
   digest = manifest.digest
   assert digest == 'sha256:dde3714ce7e23edc6413aa85c0b42792e4f2f79e9ea36afc154d63ff3d04e86c'
   assert manifest.created_datetime
+
+
+def test_validate_manifest_with_unencoded_unicode():
+  test_dir = os.path.dirname(os.path.abspath(__file__))
+  with open(os.path.join(test_dir, 'manifest_unencoded_unicode.json'), 'r') as f:
+    manifest_bytes = f.read()
+
+  manifest = DockerSchema1Manifest(manifest_bytes)
+  digest = manifest.digest
+  assert digest == 'sha256:5d8a0f34744a39bf566ba430251adc0cc86587f86aed3ac2acfb897f349777bc'
+  assert manifest.created_datetime
+
+  layers = list(manifest.get_layers(None))
+  assert layers[-1].author == u'Sômé guy'
+
+
+@pytest.mark.parametrize('with_key', [
+  None,
+  docker_v2_signing_key,
+])
+def test_build_unencoded_unicode_manifest(with_key):
+  builder = DockerSchema1ManifestBuilder('somenamespace', 'somerepo', 'sometag')
+  builder.add_layer('sha256:abcde', json.dumps({
+    'id': 'someid',
+    'author': u'Sômé guy',
+  }, ensure_ascii=False))
+
+  built = builder.build(with_key, ensure_ascii=False)
+  built._validate()