Add msgpack to the crypto fixture data

[SimonWoolf]

Prereq for addressing https://github.com/ably/wiki/issues/22

@mattheworiordan @paddybyers any objections?

[paddybyers]

any objections?

No, but right now we don't know how you're going to use these in tests.

It's not valid, unfortunately, to assume key order in the msgpack serialisation, so you can't do an encode() and compare the text output.

Verification of the decode() can easily be by decoding the msgpack string and doing a deepEqual on the result.

Verification of the encode() strictly would need us to have an independent and valid decoder to decode the output of the client library, and then do a deepEqual on the results.

[SimonWoolf]

It's not valid, unfortunately, to assume key order in the msgpack serialisation,

That's annoying. A combination of decoding the fixture and comparing with deepEqual, plus comparing the results of msgpack-encoding the testMessage and the encryptedMessage, would hopefully be enough to catch most bugs (the latter is enough to catch the bug we just fixed), but still, would have been nice to compare the result of encode() directly. Oh well.

[SimonWoolf]

(Aaand I've just realised all the weird behaviour I've been debugging for the past hour is because I generated these without passing sparse: true, and the msgpack decoder can't handle keys with undefined values. headdesk. Will regenerate)