test: clarify M1-F-D-defer status + add LIVECBOR_DEBUG env

While checking whether last night's localUPLCEval: false fix unblocks
the 5 tx-submit-live tests skipped under TODO(M1-F-D-defer), found
that those tests still crash — but at a DIFFERENT spot than before.

Original UPLC issue (now fixed in production via 7ea7328 + fd24598):
- lucid's CML WASM UPLC interpreter falsely rejected our Plutus V3
  Submit/Claim/Accept txs. End-to-end chat + TTS lifecycle now works
  on Vector testnet by routing eval to Ogmios.

New / remaining test-side issue:
- Even with a proper evaluateTransaction mock response, lucid's
  applyUPLCEvalProvider → CML TransactionBuilder.set_exunits still
  panics with wasm "unreachable" inside the test fixture. The
  production path doesn't hit this (real Ogmios + real wallet UTxOs).
  Whatever the test-mock + CML interaction trips, it's separate from
  the original UPLC bug.

Refinements landed
- liveCbor.ts: opt-in LIVECBOR_DEBUG=1 env that prints the full inner
  err + WASM stack from rethrowAsTxError, so the next investigation
  doesn't need to re-add this scaffolding.
- tx-submit-live.test.ts:
  - setupSubmitHappyPath() now mocks "evaluateTransaction" with a
    proper EvalRedeemer-shaped response (useful baseline for any
    future un-skip attempt; not enough on its own to make the tests
    pass).
  - The first skipped test's TODO comment updated to record what's
    fixed in production vs. what remains as a test-infra bug.

Test counts unchanged (3 pass, 5 still skipped); production path
still verified end-to-end on testnet. Task M1-F-D-defer stays pending
but the original UPLC-interpreter blocker is gone.

packages/shared/src/tx/internal/liveCbor.ts

@@ -180,9 +180,25 @@ function chainUtxoToLucidScriptInput(
 }
 
 /** TxConstructionError wrapping arbitrary lucid failures, surfacing
- * collateral mismatches with a stable substring so tests can assert. */
+ * collateral mismatches with a stable substring so tests can assert.
+ *
+ * Set LIVECBOR_DEBUG=1 to see the full inner error + WASM stack — useful
+ * for diagnosing lucid 0.4.30 / CML edge cases (e.g. the set_exunits
+ * panic that affects test fixtures but not production). */
 function rethrowAsTxError(err: unknown, fallback: string): never {
   const msg = err instanceof Error ? err.message : String(err);
+  if (process.env.LIVECBOR_DEBUG) {
+    // eslint-disable-next-line no-console
+    console.error("[liveCbor] inner error:", err);
+    if (err instanceof Error && err.stack) {
+      // eslint-disable-next-line no-console
+      console.error("[liveCbor] stack:", err.stack);
+    }
+    if (err instanceof Error && "cause" in err) {
+      // eslint-disable-next-line no-console
+      console.error("[liveCbor] cause:", (err as Error & { cause?: unknown }).cause);
+    }
+  }
   const lower = msg.toLowerCase();
   if (
     lower.includes("collateral") ||

tests/unit/tx-submit-live.test.ts

@@ -85,6 +85,17 @@ function setupSubmitHappyPath(escrowDatumHex: string): void {
     if (method === "queryNetwork/tip") {
       return rpcOk({ slot: Math.floor((DELIVER_BY - 30_000) / 1000), id: "a".repeat(64) });
     }
+    // localUPLCEval: false delegates exec-unit calc to Ogmios's
+    // evaluateTransaction. The provider has a fallback when this is missing
+    // but its synthesized budget can exceed lucid's bundled preset maxes
+    // → CML set_exunits WASM panic. Return conservative budgets that fit
+    // comfortably within PROTOCOL_PARAMETERS_DEFAULT's
+    // maxExecutionUnitsPerTransaction.
+    if (method === "evaluateTransaction") {
+      return rpcOk([
+        { validator: { purpose: "spend", index: 0 }, budget: { memory: 100_000, cpu: 50_000_000 } },
+      ]);
+    }
     if (method === "submitTransaction") return rpcOk({ transaction: { id: "d".repeat(64) } });
     return rpcOk({});
   });
@@ -94,9 +105,17 @@ function setupSubmitHappyPath(escrowDatumHex: string): void {
 
 describe("buildSubmitTx() [live] — CBOR shape", () => {
   it.skip("returns real Cardano CBOR starting with 83 or 84", async () => {
-    // TODO(M1-F-D-defer): UPLC wasm crashes on this test's lucid cost-models setup.
+    // TODO(M1-F-D-defer): the ORIGINAL UPLC-WASM crash (CML interpreting
-    // Production path uses real Ogmios cost models and works on testnet.
+    // the Plutus V3 script ourselves) was fixed in production by passing
-    // Defer until we plumb usePresetProtocolParameters into the test-side LucidContext.
+    // localUPLCEval: false to txBuilder.complete() — lucid now delegates
+    // eval to Ogmios. End-to-end chat + TTS lifecycle works on Vector
+    // testnet (commits 7ea7328 + fd24598). However this test still crashes
+    // at a DIFFERENT spot: lucid's applyUPLCEvalProvider → CML
+    // TransactionBuilder.set_exunits → wasm "unreachable" panic, even with
+    // a proper evaluateTransaction mock response. The test-side mock +
+    // CML interaction has its own bug separate from the production-path
+    // fix. Defer until we either (a) bisect the CML version or (b) rewrite
+    // these tests to assert at a layer above lucid's set_exunits.
     const claimedUtxo = buildClaimedEscrowUtxo();
     setupSubmitHappyPath(claimedUtxo.datumHex!);
     const chain = buildLiveChain();