GADT for TransactionScriptFailure and PlutusDebug

[aniketd]

NOTE:

• The first commit only renames all Plutus.* imports in a standard form of either PV1 (PlutusLedgerApi.V1), PV2 (PlutusLedgerApi.V2) or PCD (PlutusCore.Data)
• Only the second commit implements all the changes related to making PlutusDebug a GADT.

TODO:

• Get confirmation about CBOR codecs being correct
• Add CHANGELOG entry

[lehins]

Few things need adjusting, but overall this is great stuff!!!

After this PR will probably be able to adopt this approach in other places as well, specific type that comes to mind is CostParams.

[JaredCorduan]

This looks really great @aniketd , thanks for improving the type safety of our code!

[aniketd]

NOTE: My current standing on this PR is that I'm uncertain only about the CBOR instances that have changed, and would like for someone to take a closer look at them. Besides those, the rest of the changes seem to be okay, and the failing tests are likely unrelated. The unexpected failure caused in an earlier CI run is also reproducible otherwise, so investigating that could be considered separately from this PR.

[lehins]

So it does look good, but the changes to CBOR encoding/decoding got broken.

It would be nice if you could add some roundtrip tests for the PlutusDebug serialization. At least it would give us some confidence that it is ok. You'll need to add a few Arbitrary instances here: Test.Cardano.Ledger.Alonzo.Serialisation.Generators for it to work.

[aniketd]

I'm going to clean up the commit history in preparation for the merge.

[aniketd]

I have cleaned up the commit history and rebased on master. I think this is ready to be merged now. cc. @lehins 🙇 😅

[lehins]

This looks awesome!