seq'ing a PersistentProtocolBufferMap yields only two of the four fields

[jmibanez]

Using the latest release of clojure-protobuf, and this particular .proto definition:

message Key {
  required bytes salt = 1;
  required bytes iv = 2;
  required bytes encryptedKey = 3;
  required bytes verifyingKey = 4;
}

When ever I try loading from a file via protobuf-load-stream or via protobuf-load, the returned protobuf object does not yield the encryptedKey and verifyingKey keys, and its str representation doesn't have those keys either, e.g.:

protobuf-raw-seq
-> {:salt #<ByteString com.google.protobuf.ByteString@ca222b04>, :iv #<ByteString com.google.protobuf.ByteString@ae9053d8>}

(:encryptedKey protobuf-raw-seq)
-> #<ByteString com.google.protobuf.ByteString@e2fd05ec>

(take 4 (seq protobuf-raw-seq))
-> ([:salt #<ByteString com.google.protobuf.ByteString@ca222b04>] [:iv #<ByteString com.google.protobuf.ByteString@ae9053d8>])

A function that I've written to unpack ByteStrings into byte arrays and translate the protobuf object into a Clojure map depends on being able to seq into the PersistentProtocolBufferMap without losing keys.

This is on Clojure 1.4.0

I can provide a minimal test case. I am not sure whether this is a bug in clojure-protobuf or protobuf itself, but doing (.. protobuf-raw-master-rec def type getFields) does yield the missing fields.

[eRaz0rHead]

I believe this is the same error I discovered -- If I am correct, this is because of the mixedCase in the missing keys.
If you were to check your schema you'd find that the keys are :encryptedkey and not :encryptedKey.

However, the actual value is stored with the camelCase key, and so the toString() method is calling .get( ) vs .get () , which returns null and so it's left out of the output

[amalloy]

Yes, this is a duplicate of #45: the default naming strategy doesn't work for mixedCase keys.