Return LedgerPeerSnapshot directly instead of Serialised LedgerPeerSnapshot, as there is no reason to keep the result serialized.
Add deserialiseFromTextEnvelopeJSON and deserialiseFromTextEnvelopeJSONAnyOf which accept a raw ByteString, enabling text envelope deserialization without file I/O. Refactor readFileTextEnvelope and readFileTextEnvelopeAnyOf to use the new functions.
Fix serialiseToCBOR to extract raw script bytes from PlutusRunnable instead of using L.serialize' which adds CBOR framing that deserialiseFromCBOR does not expect. Add CBOR and TextEnvelope roundtrip tests for PlutusScriptInEra and AnyPlutusScript. Add haddocks for the text envelope functions.
The hardcoded surplus range assumed a fixed F1 for all inputs, but the fee varies with generated address structure (credential types, staking references). Compute F1 per-test via calcMinFeeTx and set surplus to [F1+4, F1+10]. Also deduplicates genTinySurplusTx — it now lives only in Fee.hs — and adds a deterministic fee gap invariant test.
The hardcoded surplus range assumed a fixed F1 for all inputs, but the fee varies with generated address structure (credential types, staking references). Compute F1 per-test via calcMinFeeTx and set surplus to [F1+4, F1+10]. Also deduplicates genTinySurplusTx — it now lives only in Fee.hs — and adds a deterministic fee gap invariant test.
Fix serialiseToCBOR to extract raw script bytes from PlutusRunnable instead of using L.serialize' which adds CBOR framing that deserialiseFromCBOR does not expect. Add CBOR and TextEnvelope roundtrip tests for PlutusScriptInEra and AnyPlutusScript. Add haddocks for the text envelope functions.
Remove max-backjumps from cabal.project
Add Copilot instructions to ignore sha256 comment mismatches in cabal.project
Narrow the range from 237–250 to 240–246. The lower bound has headroom because F1 can be slightly lower than expected for some generated inputs. The upper bound stays well below F2 to avoid boundary cases where the surplus barely creates a valid change output.
Narrow the range from 237–250 to 240–246. The lower bound has headroom because F1 can be slightly lower than expected for some generated inputs. The upper bound stays well below F2 to avoid boundary cases where the surplus barely creates a valid change output.
Narrow the upper bound from 250 to 246 to stay well below the F2 fee threshold, avoiding boundary cases where the surplus is just barely enough to create a valid change output.
The max-backjumps setting is no longer needed for cabal to solve constraints. The comment itself said to remove it when not needed anymore.
Replace fromJust in votingScriptWitnessSingleton with proper error handling