Ensure we encode the config in manifest schema 2 via the canonical JSON format

image/docker/schemautil.py

@@ -25,8 +25,28 @@ class ContentRetrieverForTesting(ContentRetriever):
 
 
 def ensure_utf8(unicode_or_str):
-  """ Ensures the given string is utf-8 encoded and not unicode. """
+  """ Ensures the given string is a utf-8 encoded str and not a unicode type. """
   if isinstance(unicode_or_str, unicode):
     return unicode_or_str.encode('utf-8')
 
   return unicode_or_str
+
+
+class _CustomEncoder(json.JSONEncoder):
+  def encode(self, o):
+    encoded = super(_CustomEncoder, self).encode(o)
+    if isinstance(o, basestring):
+      encoded = encoded.replace('<', '\\u003c')
+      encoded = encoded.replace('>', '\\u003e')
+      encoded = encoded.replace('&', '\\u0026')
+    return encoded
+
+
+def to_canonical_json(value, ensure_ascii=True, indent=None):
+  """ Returns the canonical JSON string form of the given value,
+      as per the guidelines in https://github.com/docker/distribution/blob/master/docs/spec/json.md.
+
+      `indent` is allowed only for the purposes of indenting for debugging.
+  """
+  return json.dumps(value, ensure_ascii=ensure_ascii, sort_keys=True, separators=(',', ':'),
+                    cls=_CustomEncoder, indent=indent)