[Ranjna-G]

Summary

Adds get_input_outpoints to the PSBT wasm surface: each transparent input's previous-output outpoint, txid in display order. This is the wasm counterpart of utxo-lib's raw txInputs + getOutputIdForInput introspection, so consumers that only need the spent outpoints — unspent-id extraction, e.g. wallet-platform's pending-approval creation — behave identically for v4 and v6 PSBTs.

Linear: CSHLD-2087

Changes

• src/wasm/psbt.rs — PsbtInputOutpoint {txid, vout} + input_outpoints(psbt) (reads unsigned_tx.input[i].previous_output; pure, unit-testable) + the get_input_outpoints wasm method on both macro-generated impls (BitGoPsbt + wrapped variants); a thin JsValue wrapper
• src/wasm/psbt_ops.rs / src/wasm/mod.rs / src/wasm/try_into_js_value.rs — trait method, crate export, JsValue serialization
• js/psbtBase.ts + js/index.ts — getInputOutpoints() wrapper and its type
• Tests: Rust (input_outpoints_are_version_symmetric — the same mock outpoint spent by a v4 and a v6 PSBT yields the identical display-order {txid, vout}) and JS (the same assertion through the TS wrapper)

Why

wallet-platform's PA creation extracts input ids from the built transaction. For v6 PSBTs it routes through the wasm layer; the only wasm API that exposed outpoints was the wallet-key-aware parseTransactionWithWalletKeys. This accessor reads the same outpoints directly off the embedded unsigned transaction — the same data, in the same display-order shape utxo-lib's raw parse provides for v4 — with no wallet keys involved, matching v4 behavior.

Sequencing

wallet-platform consumes this once it releases (its bms counterpart branch is written against this API and bumps the pin in lockstep).

Test Plan

• cargo test --lib — 765+ passing, including the new version-symmetry test
• npx tsc --noEmit — 0 errors against regenerated bindings
• npx mocha test/fixedScript/zcashIronwoodPsbt.ts — all passing, including the new wrapper-level parity test

[linear-code]

CSHLD-2087

[claude]

Claude Code Review

This repository is configured for manual code reviews. Comment @claude review for a one-time review, or @claude review always to subscribe this PR to a review on every future push.

_{Tip: disable this comment in your organization's Code Review settings.}