More encoding / decoding examples

[mattheworiordan]

See ably/ably-cocoa#95, I think client library developers will generally struggle to understand the spec and implement exactly as we expect.

As such, I recommend we introduce a new JSON file in https://github.com/ably/ably-common/tree/master/test-resources which includes all permutations of non-encrypted encodings & decodings to remove any ambiguity about what is expected.

@paddybyers @SimonWoolf WDYT? This should not change the spec, but I can add a note in the encoding & decoding that is helps to run the tests against the JSON file.

[paddybyers]

which includes all permutations of non-encrypted encodings & decodings to remove any ambiguity about what is expected.

Yes, but really we want examples that map a Message, or at least a data, to an encoded Message or ProtocolMessage, so it is obvious how the arguments presented in the API get turned into serialised entities. But .... we can't put binary data into a JSON file, so we need the file to include an extra level of encoding to be able to contain the data and the serialised entity. That could just add even more confusion unless we explain it carefully ...

[SimonWoolf]

👍 there's a few in ably-js here, but yeah, putting it all with expected decoded result in a common file is a good idea

[mattheworiordan]

Yes, but really we want examples that map a Message, or at least a data, to an encoded Message or ProtocolMessage

I don't see why a ProtocolMessage is relevant given it will contain an array of Message objects, and encoding & decoding is contained within the Message

But .... we can't put binary data into a JSON file, so we need the file to include an extra level of encoding to be able to contain the data and the serialised entity. That could just add even more confusion unless we explain it carefully ...

How is that any different to https://github.com/ably/ably-common/blob/master/test-resources/crypto-data-256.json#L9-L18?