Add support for Plutus scripts on compatible Tx by carbolymer · Pull Request #1282 · IntersectMBO/cardano-api

carbolymer · 2026-08-07T23:14:35Z

Description

Adds plutus spending script support to the compatible transaction API, so pre-Conway script transactions (PlutusV1/V2, Alonzo/Babbage) can be built without the deprecated API.
Motivated by the tx-generator migration in IntersectMBO/cardano-node#6602.

Changes:

createCompatibleTx supports script-witnessed inputs: spending redeemer pointers, witness datums, script integrity hash, collateral inputs, transaction metadata and validity upper bound.
Breaking: all transaction content is now passed as a single CompatibleTxBodyContent record, and failures are reported via the new CompatibleTxError.
Breaking: mkSpendingScriptDatum is exported for building spending-purpose datums.
Fixes: certificate and proposal script witnesses were missing from the witness set; proposal reference inputs were overwritten.
Removes the redundant StakeCredential field from WitTxCert.

Note

The commit TODROP: Fix redeemer pointer indexing in the experimental witness machinery is part of #1288 and will be removed from this PR once #1288 is merged.
This branch is intentionally not rebased onto master yet; CI will start once the rebase lands after #1288.

Checklist

New tests: ledger-oracle redeemer pointer properties cover the compatible builder paths
Changelog fragment (.changes/20260807_cardano_api_compatible_plutus.yml)
Self-reviewed

…hinery Proposal redeemer pointers were computed against Ord-sorted order instead of the ledger's OSet insertion order, and certifying pointers did not count unwitnessed certificates' index slots. Both now follow the ledger's positional resolution, with regression tests pinning the behaviour and property tests checking every witnessable category's pointer against the ledger's own Indexable resolution. The same fix is applied to the deprecated legacy transaction builder (createTransactionBody); end-to-end and legacy-bridge regression properties now cover the deprecated path too. Also harden getVotes against voter-map gaps, remove the unused StakeCredential field from the Witnessable WitTxCert constructor, and correct the redeemer-index ordering documentation.

createCompatibleTx now takes a CompatibleTxBodyContent record with per-input AnyWitness values, and computes spending redeemer pointers, witness datums, the script integrity hash, collateral inputs, transaction metadata and the validity upper bound. Witness indexing is shared with the experimental transaction builder. Also export the mkSpendingScriptDatum smart constructor from Cardano.Api.Experimental.

Copilot

Pull request overview

Extends the Cardano.Api.Compatible.Tx builder to support Plutus spending scripts (and related era features) so callers can build pre-Conway Plutus script transactions without relying on deprecated APIs, while also fixing redeemer-pointer indexing behavior and adding regression/property coverage.

Changes:

Add Plutus spending-witness support and additional threading (collateral, metadata, validity upper bound, script integrity hash) to createCompatibleTx, with a new CompatibleTxBodyContent record and CompatibleTxError.
Fix witness/indexing behavior for proposals/certificates/votes by preserving insertion order (OSet/OMap) and ensuring unwitnessed items occupy redeemer index slots.
Add/extend property tests to validate redeemer pointer indexing against ledger resolution and cover new compatible-builder behavior.

Reviewed changes

Copilot reviewed 16 out of 16 changed files in this pull request and generated 2 comments.

Show a summary per file

File	Description
cardano-api/test/cardano-api-test/Test/Cardano/Api/Transaction/Body/Plutus/RedeemerIndex.hs	New property suite asserting API redeemer indices match ledger `Indexable`/`redeemerPointer` resolution.
cardano-api/test/cardano-api-test/Test/Cardano/Api/Experimental/Fee.hs	Updates compatible-tx tests and adds new properties for spending redeemer order, witness set inclusion, pparams requirement, and extra content threading.
cardano-api/test/cardano-api-test/Test/Cardano/Api/Experimental.hs	Adds properties verifying proposal insertion-order indexing and certificate indexing that counts unwitnessed certs.
cardano-api/test/cardano-api-test/cardano-api-test.hs	Wires the new redeemer-index test group into the test runner.
cardano-api/src/Cardano/Api/Tx/Internal/Body.hs	Keeps unwitnessed certificates in witnessable extraction to preserve redeemer index slots; updates proposal-indexing docs.
cardano-api/src/Cardano/Api/Tx.hs	Reorders/clarifies export comment for witnessable extractors.
cardano-api/src/Cardano/Api/Internal/Orphans/Serialisation.hs	Adds `ToJSON` for `DijkstraLedgerPredFailure`.
cardano-api/src/Cardano/Api/Experimental/Tx/Internal/BodyContent/New.hs	Includes unwitnessed cert placeholders; makes vote extraction total to avoid index shifts on misses.
cardano-api/src/Cardano/Api/Experimental/Plutus/Internal/ScriptWitness.hs	Exports `mkSpendingScriptDatum` helper for spending-purpose datum construction.
cardano-api/src/Cardano/Api/Experimental/Plutus/Internal/IndexedPlutusScriptWitness.hs	Preserves insertion order for certificates/proposals during stable sort by returning `EQ` intra-category.
cardano-api/src/Cardano/Api/Experimental/Plutus.hs	Re-exports `mkSpendingScriptDatum`.
cardano-api/src/Cardano/Api/Experimental.hs	Re-exports `mkSpendingScriptDatum`.
cardano-api/src/Cardano/Api/Compatible/Tx.hs	Major rewrite: `createCompatibleTx` now takes `CompatibleTxBodyContent`, supports Plutus spending witnesses, computes script integrity hash, threads metadata/ttl/collateral, and fixes witness inclusion.
cardano-api/cardano-api.cabal	Adds `cardano-data` to test-suite dependencies and includes the new test module.
.changes/20260811_cardano_api_redeemer_pointer_indexing.yml	Changelog fragment for redeemer pointer indexing fixes (includes breaking + tests).
.changes/20260807_cardano_api_compatible_plutus.yml	Changelog fragment for compatible builder Plutus support and breaking API changes.

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.

+data CompatibleTxError
+  = -- | The supplied protocol parameters could not be converted to the
+    -- ledger representation needed to compute the script integrity hash.
+    CompatibleTxProtocolParametersConversionError ProtocolParametersConversionError
+  | -- | Plutus script witnesses are present (non-empty redeemers, datums
+    -- or plutus languages), but 'compatibleTxProtocolParams' is
+    -- 'Nothing', so the ledger's required script integrity hash cannot
+    -- be computed.
+    CompatibleTxMissingScriptIntegrityPParams
+  deriving Show
+
+instance Error CompatibleTxError where
+  prettyError err =
+    case err of
+      CompatibleTxProtocolParametersConversionError e -> prettyError e
+      CompatibleTxMissingScriptIntegrityPParams ->
+        "Plutus script witnesses are present but no protocol parameters were supplied "
+          <> "to compute the script integrity hash."


+  txIn1@(Api.TxIn txId (Api.TxIx txIx)) <- H.forAll genTxIn
+  let txIn2 = Api.TxIn txId (Api.TxIx (txIx + 1))


palas

I think there are a couple of issues that should be resolved before merging, and I also included a couple of potential readability improvement opportunities that I saw.

I still haven't had time to look at the tests yet too.

But it is looking good 👍

palas · 2026-08-13T22:16:03Z

+              -- Order must stay OMap insertion order; the shared
+              -- Witnessable-based indexing (via 'extractWitnessableProposals') preserves it.
+              proposalWitnesses
+                :: [(Witnessable ProposalItem (ShelleyLedgerEra era), AnyWitness (ShelleyLedgerEra era))]
              proposalWitnesses =
-                [ (ix, witness)
-                | (_, (ix, witness)) <-
-                    indexWitnessedTxProposalProcedures conwayOnwards proposalProcedures
-                ]
+                Exp.obtainCommonConstraints (convert conwayOnwards) $
+                  Exp.extractWitnessableProposals $
+                    Just proposalProcedures


I think the function is getting too large (even ignoring the where bits).

And proposalWitnesses doesn't need to be in the case, it could be moved to the where (having its own small case), and that would allow moving allWitnesses and sData to the where too (txAuxData can moved to the where already), and I think that would make the body of the function much more manageable.

Actually, updateTxBody wouldn't need to be here either, because I think it doesn't make use of the Either monad, so it wouldn't need to be monadic, it could be moved to the where too, but that is out of scope of the PR

palas · 2026-08-13T22:35:14Z

+  -- certificates and witnessed spending inputs.
  setRefInputs :: Endo (L.TxBody L.TopTx (ShelleyLedgerEra era))
  setRefInputs = do
    let refInputs =
          [ toShelleyTxIn refInput
-          | (_, _, Just (_, wit)) <- indexedTxCerts
+          | wit <- witnessedCertWitnesses
          , refInput <- maybeToList $ getAnyWitnessReferenceInput wit
          ]
+            ++ [ toShelleyTxIn refInput
+               | (_, wit) <- ins
+               , refInput <- maybeToList $ getAnyWitnessReferenceInput wit
+               ]


Suggested change

-- certificates and witnessed spending inputs.

setRefInputs :: Endo (L.TxBody L.TopTx (ShelleyLedgerEra era))

setRefInputs = do

let refInputs =

[ toShelleyTxIn refInput

| (_, _, Just (_, wit)) <- indexedTxCerts

| wit <- witnessedCertWitnesses

, refInput <- maybeToList $ getAnyWitnessReferenceInput wit

]

++ [ toShelleyTxIn refInput

| (_, wit) <- ins

, refInput <- maybeToList $ getAnyWitnessReferenceInput wit

]

-- certificates and witnessed spending inputs.

setRefInputs :: Endo (L.TxBody L.TopTx (ShelleyLedgerEra era))

setRefInputs = do

let refInputs =

[ toShelleyTxIn refInput

| wit <- witnessedCertWitnesses <> map snd ins

, refInput <- maybeToList $ getAnyWitnessReferenceInput wit

]

This can be done with a single list comprehension

palas · 2026-08-13T23:41:50Z

+  setScriptWitnesses sData scriptWitnesses =
    appEndos
      [ monoidForEraInEon
          era
-          ( \aeo -> alonzoEraOnwardsConstraints aeo $ Endo $ do
-              let sData = convScriptData' sbe extraDatums scriptWitnesses
-              let (datums, redeemers) = case sData of
-                    TxBodyScriptData _ ds rs -> (ds, rs)
-                    TxBodyNoScriptData -> (mempty, L.Redeemers mempty)
-              (L.datsTxWitsL .~ datums) . (L.rdmrsTxWitsL %~ (<> redeemers))
+          ( \aeo ->
+              alonzoEraOnwardsConstraints aeo $
+                obtainAlonzoScriptPurposeConstraints aeo $
+                  Endo $ do
+                    let (datums, redeemers) =
+                          fromMaybe mempty sData
+                            :: (L.TxDats (ShelleyLedgerEra era), L.Redeemers (ShelleyLedgerEra era))
+                    -- 'getAnyWitnessScript' covers both plutus and simple
+                    -- scripts. The simple ones overlap harmlessly (same
+                    -- hash, same script) with the allegra-onwards branch
+                    -- below, which is still needed for pre-Alonzo eras
+                    -- that this branch does not run in.
+                    let plutusAndSimpleScripts =
+                          mapMaybe getAnyWitnessScript scriptWitnesses
+                    (L.datsTxWitsL .~ datums)
+                      . (L.rdmrsTxWitsL %~ (<> redeemers))
+                      . ( L.scriptTxWitsL
+                            %~ (<> Map.fromList [(L.hashScript sw, sw) | sw <- plutusAndSimpleScripts])
+                        )
          )
      , monoidForEraInEon
          era
          ( \aeo -> allegraEraOnwardsConstraints aeo $ Endo $ do
              let ledgerScripts = convSimpleScripts sbe scriptWitnesses
              L.scriptTxWitsL
-                .~ Map.fromList
-                  [ (L.hashScript sw, sw)
-                  | sw <- ledgerScripts
-                  ]
+                %~ ( <>
+                       Map.fromList
+                         [ (L.hashScript sw, sw)
+                         | sw <- ledgerScripts
+                         ]
+                   )
          )
      ]

 convSimpleScripts
  :: ShelleyLedgerEra era ~ ledgerera
  => ShelleyBasedEra era
-  -> [(ScriptWitnessIndex, Exp.AnyWitness (ShelleyLedgerEra era))]
+  -> [Exp.AnyWitness (ShelleyLedgerEra era)]


Suggested change

setScriptWitnesses sData scriptWitnesses =

appEndos

[ monoidForEraInEon

era

( \aeo -> alonzoEraOnwardsConstraints aeo $ Endo $ do

let sData = convScriptData' sbe extraDatums scriptWitnesses

let (datums, redeemers) = case sData of

TxBodyScriptData _ ds rs -> (ds, rs)

TxBodyNoScriptData -> (mempty, L.Redeemers mempty)

(L.datsTxWitsL .~ datums) . (L.rdmrsTxWitsL %~ (<> redeemers))

( \aeo ->

alonzoEraOnwardsConstraints aeo $

obtainAlonzoScriptPurposeConstraints aeo $

Endo $ do

let (datums, redeemers) =

fromMaybe mempty sData

:: (L.TxDats (ShelleyLedgerEra era), L.Redeemers (ShelleyLedgerEra era))

-- 'getAnyWitnessScript' covers both plutus and simple

-- scripts. The simple ones overlap harmlessly (same

-- hash, same script) with the allegra-onwards branch

-- below, which is still needed for pre-Alonzo eras

-- that this branch does not run in.

let plutusAndSimpleScripts =

mapMaybe getAnyWitnessScript scriptWitnesses

(L.datsTxWitsL .~ datums)

. (L.rdmrsTxWitsL %~ (<> redeemers))

. ( L.scriptTxWitsL

%~ (<> Map.fromList [(L.hashScript sw, sw) | sw <- plutusAndSimpleScripts])

)

)

, monoidForEraInEon

era

( \aeo -> allegraEraOnwardsConstraints aeo $ Endo $ do

let ledgerScripts = convSimpleScripts sbe scriptWitnesses

L.scriptTxWitsL

.~ Map.fromList

[ (L.hashScript sw, sw)

| sw <- ledgerScripts

]

%~ ( <>

Map.fromList

[ (L.hashScript sw, sw)

| sw <- ledgerScripts

]

)

)

]

convSimpleScripts

:: ShelleyLedgerEra era ~ ledgerera

=> ShelleyBasedEra era

-> [(ScriptWitnessIndex, Exp.AnyWitness (ShelleyLedgerEra era))]

-> [Exp.AnyWitness (ShelleyLedgerEra era)]

setScriptWitnesses sData scriptWitnesses =

plutusAdditions . simpleScriptAdditions

where

plutusAdditions :: L.TxWits (ShelleyLedgerEra era) -> L.TxWits (ShelleyLedgerEra era)

plutusAdditions =

forEraInEon era id $ \aeo ->

alonzoEraOnwardsConstraints aeo $

obtainAlonzoScriptPurposeConstraints aeo $

let (datums, redeemers) =

fromMaybe mempty sData

:: (L.TxDats (ShelleyLedgerEra era), L.Redeemers (ShelleyLedgerEra era))

-- 'getAnyWitnessScript' covers both plutus and simple

-- scripts. The simple ones overlap harmlessly (same

-- hash, same script) with 'simpleScriptAdditions',

-- which is still needed for pre-Alonzo eras where this

-- addition is 'id'.

plutusAndSimpleScripts =

mapMaybe getAnyWitnessScript scriptWitnesses

in (L.datsTxWitsL .~ datums)

. (L.rdmrsTxWitsL %~ (<> redeemers))

. ( L.scriptTxWitsL

%~ (<> Map.fromList [(L.hashScript sw, sw) | sw <- plutusAndSimpleScripts])

)

simpleScriptAdditions :: L.TxWits (ShelleyLedgerEra era) -> L.TxWits (ShelleyLedgerEra era)

simpleScriptAdditions =

forEraInEon era id $ \aeo ->

allegraEraOnwardsConstraints aeo $

L.scriptTxWitsL

%~ ( <>

Map.fromList

[ (L.hashScript sw, sw)

| sw <- convSimpleScripts sbe scriptWitnesses

]

)

convSimpleScripts

:: ShelleyLedgerEra era ~ ledgerera

=> ShelleyBasedEra era

-> [Exp.AnyWitness (ShelleyLedgerEra era)]

I think Endos in this function only add complexity. They were already here before the PR, but since we are changing most of the function...

palas · 2026-08-14T01:32:41Z

+      -- languages from. Only the redeemer pointer map (built inside
+      -- 'convScriptData'') needs per-category indexing.
+      allWitnesses :: [AnyWitness (ShelleyLedgerEra era)]
+      allWitnesses = witnessedCertWitnesses <> map snd proposalWitnesses <> map snd ins


I think we are missing TxVotingProcedures witnesses

palas · 2026-08-14T01:57:31Z

+  :: L.AlonzoEraScript ledgerera
+  => [(TxIn, AnyWitness ledgerera)]
+  -> [(Witnessable TxInItem ledgerera, AnyWitness ledgerera)]
+witnessableTxIns txIns' = L.nub [(WitTxIn txIn, wit) | (txIn, wit) <- txIns']


We should validate there are no repeated TxIns in the list. Or at least that there are no two TxIns with different witnesses, because that would mess up the redeemer pointers

carbolymer self-assigned this Aug 8, 2026

carbolymer force-pushed the mgalazyn/feature/compatible-plutus branch 7 times, most recently from 7139bf6 to d6488c0 Compare August 13, 2026 08:58

carbolymer force-pushed the mgalazyn/feature/compatible-plutus branch 3 times, most recently from 5e5e6af to 75f0875 Compare August 13, 2026 10:35

carbolymer force-pushed the mgalazyn/feature/compatible-plutus branch from 75f0875 to 26aaf4a Compare August 13, 2026 10:40

carbolymer marked this pull request as ready for review August 13, 2026 15:33

carbolymer requested review from Jimbo4350, erikd and palas as code owners August 13, 2026 15:33

Copilot AI lite review requested due to automatic review settings August 13, 2026 15:33

carbolymer requested review from CarlosLopezDeLara and disassembler as code owners August 13, 2026 15:33

Copilot started reviewing on behalf of carbolymer August 13, 2026 15:33 View session

carbolymer mentioned this pull request Aug 13, 2026

Update compatible transaction building for the cardano-api changes IntersectMBO/cardano-cli#1408

Open

4 tasks

Copilot AI reviewed Aug 13, 2026

View reviewed changes

carbolymer mentioned this pull request Aug 13, 2026

Update tx-generator to not use deprecated cardano-api API IntersectMBO/cardano-node#6602

Open

9 tasks

palas reviewed Aug 14, 2026

View reviewed changes

Navigation Menu

Uh oh!

Add support for Plutus scripts on compatible Tx - #1282

Add support for Plutus scripts on compatible Tx#1282
carbolymer wants to merge 2 commits into
masterfrom
mgalazyn/feature/compatible-plutus

carbolymer commented Aug 7, 2026 •

edited

Loading

Uh oh!

Copilot AI left a comment

Uh oh!

palas left a comment

Uh oh!

palas Aug 13, 2026

Uh oh!

palas Aug 13, 2026

Uh oh!

palas Aug 13, 2026

Uh oh!

palas Aug 14, 2026

Uh oh!

palas Aug 14, 2026

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

		txIn1@(Api.TxIn txId (Api.TxIx txIx)) <- H.forAll genTxIn
		let txIn2 = Api.TxIn txId (Api.TxIx (txIx + 1))

Uh oh!

Conversation

carbolymer commented Aug 7, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Description

Checklist

Uh oh!

Copilot AI left a comment

Choose a reason for hiding this comment

Pull request overview

Reviewed changes

Uh oh!

palas left a comment

Choose a reason for hiding this comment

Uh oh!

palas Aug 13, 2026

Choose a reason for hiding this comment

Uh oh!

palas Aug 13, 2026

Choose a reason for hiding this comment

Uh oh!

palas Aug 13, 2026

Choose a reason for hiding this comment

Uh oh!

palas Aug 14, 2026

Choose a reason for hiding this comment

Uh oh!

palas Aug 14, 2026

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

carbolymer commented Aug 7, 2026 •

edited

Loading