Push sig kind out of txn witness

[glyh]

A simple noop PR pushing sig kind out of txn witness.

It's previously suggested by @georgeee we should store txn kind in proof cache db. But it seems they're not actually attached to proofs stored in proof cache DB. hence I'm leaning toward pushing upward, and here is this PR.

[cjjdespres]

The code doesn't end up being too messy after all, if we keep pushing the signature kind upward? It would make sense to store the signature kind in the proof db if we needed to, because the signature kind is used as an input to the proofs through the hashes. So the proofs are all valid relative to a single signature kind.

But if we don't need to store it there then that's also fine.

[cjjdespres]

!ci-bypass-changelog

[cjjdespres]

!ci-build-me

[glyh]

because the signature kind is used as an input to the proofs through the hashes.

Indeed. The problem is that it seems it's still not strictly attached to a DB, nothing's preventing us to write proof of wrong sig kind into the DB. So I think it's probably better to keep it otherwise.

I'm fine if you feel like we should do it otherwise.

[glyh]

Okay, I'll prepare a PR against develop for a clean merge

[cjjdespres]

I think this might have been covered by some of George's recent PRs?