Per input/output transaction hash jets

[roconnor-blockstream]

The tx-hash and *-hash jets for Elements transactions generally hash parts of the transaction across all input and or outputs. This lets you do something like CTV and APO. However, for covenants you typically want to hash all the data that corresponds to a specific input or specific output.

• input-hash : TWO^32 ⊢ S(TWO^256) Similar to inputs-hash but only hashing the data of one input.
• output-hash : TWO^32 ⊢ S(TWO^256) Similar to outputs-hash but only hashing the data of one output.
• issuance-hash : TWO^32 ⊢ S(TWO^256) Similar to issuances-hash but only hashing the data of one input.
• input-utxo-hash : TWO^32 ⊢ S(TWO^256) Similar to inputs-utxos-hash but only hashing the data of one input.

This PR refactors some of the existing jet specifications changing their CMRs.

[apoelstra]

primitive/elements/jets.c: In function 'input_utxo_hash':
primitive/elements/jets.c:1110:17: error: declaration of 'utxo' shadows a global declaration [-Werror=shadow]
 1110 |     const utxo* utxo = &env->tx->input[i].txo;
      |                 ^~~~
In file included from primitive/elements/ops.h:7,
                 from primitive/elements/jets.c:3:
primitive/elements/primitive.h:138:3: note: shadowed declaration is here
  138 | } utxo;
      |   ^~~~

[roconnor-blockstream]

@apoelstra This is ready for your review.

[apoelstra]

LGTM.

Two observations:

• When hashing issuance data we also cover the rangeproof/surjectionproof of the issuance even though this is arguably witness data. Unsure if this is the right choice.
• When hashing inputs we don't hash the nonce of the spent output. I think this is correct, but in any case we have no choice, because Elements does not keep track of the nonce in the utxoset. (This is an old elements bug.)

[apoelstra]

Looks like primitiveEnumJet.inc is not updated

++++ diff primitiveEnumJet.inc /nix/store/571z6kp50kw0x4v2x5vsgg3qk4qqp4jl-source/C/primitive/elements/primitiveEnumJet.inc
174a175
> INPUT_HASH,
183a185
> INPUT_UTXO_HASH,
201a204
> ISSUANCE_HASH,
329a333
> OUTPUT_HASH,

[roconnor-blockstream]

Should be fixed now.