Merge upstream changes

[erikdubbelboer]

This branch is 31 commits behind ugorji:master. This pull will merge all upstream changes.

I have tested this with memberlist and all tests passed.

[armon]

This is a very large change and could potentially break things in subtle ways. We purposed forked to avoid these kinds of large changes, especially as there is very little testing upstream for breaking of compatibility. I think we will sit on this for now.

[erikdubbelboer]

If you're afraid of this why not just use encoding/json or encoding/gob? The structs being encoded are quite small and having one less external dependency plus having the stability of the encoding packages could actually be a good thing.

[ryanuber]

@erikdubbelboer We still prefer msgpack for a few reasons, mainly its efficiency over the wire. Our unit tests have fairly good coverage, but alone are not enough to guard us against this kind of upstream breakage, as we learned quickly with ugorji#37. The commit we forked at is known stable and performant, which is why we are sticking with it for now.

Longer term, it would be nice to move forward with this and not maintain our own msgpack repo, we just haven't prioritized it yet.

[hashicorp-cla]

Thank you for your submission! We require that all contributors sign our Contributor License Agreement ("CLA") before we can accept the contribution. Read and sign the agreement

Learn more about why HashiCorp requires a CLA and what the CLA includes

_{Have you signed the CLA already but the status is still pending? Recheck it.}