fix(wasm-utxo): commit spent scriptPubKey, not redeem script, in v6 transparent sighash

[veetragjain]

Compute_v6_transparent_sighash hashed the caller-supplied redeem/witness script into the ZIP-244 S.2g.iii per-input digest field. Cross-checking against the vendored zcash_primitives/zcash_transparent reference crates, that field is actually the spent output's scriptPubKey, not the redeem/witness script ("scriptCode").

The two are identical for P2PKH inputs, which is all the existing golden fixture exercised - so the bug was invisible until a real 2-of-3 P2SH multisig shielding transaction (BitGo's actual wallet shape) was signed and submitted to a live Zcash testnet node, where it was rejected with ScriptInvalid.

• v6.rs: drop the script_code param from compute_v6_transparent_sighash; hash input_script_pubkeys[index] for the per-input field instead.
• zcash_psbt.rs: update v6_transparent_sighash accordingly; keep the redeem/witness-script presence check as a spendability sanity check.
• Add a real-world regression test in both Rust and TS: a mempool-accepted testnet tx spending a 2-of-3 P2SH multisig input, with its two ECDSA signatures verified against the corrected digest independently (Rust: secp256k1 crate directly; TS: @bitgo/utxo-lib's ecc.verify) rather than via this codebase's own signing code.
• Add ZcashV6Transaction.transparentSighash (Rust wasm export + TS wrapper) so the TS test can compute a sighash from arbitrary raw v6 transaction bytes, independent of the PSBT build flow.

Ticket: CSHLD-1448

[linear-code]

CSHLD-1448

[Ranjna-G]

can you explain a bit on scriptPubKey and redeem script?

[veetragjain]

scriptPubKey is something that utxo is locked to, when spending it you need to reveal the redeem script that hashes to the scriptPubKey that the utxo is locked to.