Split the storage and negotiation parts of Codecs

The codec factory should support two distinct interfaces - negotiating
for a serializer with a client, vs reading or writing data to a storage
form (etcd, disk, etc). Make the EncodeForVersion and DecodeToVersion
methods only take Encoder and Decoder, and slight refactoring elsewhere.

In the storage factory, use a content type to control what serializer to
pick, and use the universal deserializer. This ensures that storage can
read JSON (which might be from older objects) while only writing
protobuf. Add exceptions for those resources that may not be able to
write to protobuf (specifically third party resources, but potentially
others in the future).
This commit is contained in:
Clayton Coleman
2016-04-23 15:00:28 -04:00
parent 5622c8a471
commit e0ebcf4216
47 changed files with 543 additions and 240 deletions

View File

@@ -27,6 +27,7 @@ import (
// EnableCrossGroupDecoding modifies the given decoder in place, if it is a codec
// from this package. It allows objects from one group to be auto-decoded into
// another group. 'destGroup' must already exist in the codec.
// TODO: this is an encapsulation violation and should be refactored
func EnableCrossGroupDecoding(d runtime.Decoder, sourceGroup, destGroup string) error {
internal, ok := d.(*codec)
if !ok {
@@ -45,6 +46,7 @@ func EnableCrossGroupDecoding(d runtime.Decoder, sourceGroup, destGroup string)
// EnableCrossGroupEncoding modifies the given encoder in place, if it is a codec
// from this package. It allows objects from one group to be auto-decoded into
// another group. 'destGroup' must already exist in the codec.
// TODO: this is an encapsulation violation and should be refactored
func EnableCrossGroupEncoding(e runtime.Encoder, sourceGroup, destGroup string) error {
internal, ok := e.(*codec)
if !ok {