VK hash (sha256) should include the composer_type

[dbanks12]

I asked @suyash67 if composer_type be a part of a VK's hash as it is not included in the sha256. His answer was yes or at least probably yes.

[suyash67]

Yes, I think we should add composer_type.

Imagine if we have two circuits with the exact same configuration, therefore same commitments, circuit_size, num_public_inputs but both use different underlying proof systems (stdhonk and stdplonk for example). In such a case, we would like the vk hash to make the distinction that both circuits are different.

[suyash67]

Talked to Zac to confirm this. Zac suggested that we should even add some identifier of chain (for e.g. sha256("AZTEC_ETHEREUM")) so that the verification for same circuits across chains could also be unique. But we can do that later.

[codygunton]

This was done at some point. Cf #562 -- not sure this is a Barretenberg thing, and I'd like to get rid of various identification enums in favor of using the type system, functions like std::same_as, and derived concepts like HasPlookup and IsPlonkFlavor.