Remove redundant ScriptHash in ReferenceScript datatypes

[Swordlash]

Changelog

- description: |
    Remove redundant ScriptHash in ReferenceScript datatypes
  type:
  - breaking       
  - refactoring

Context

Sometime ago I opened an issue #606 but I didn't get any response, so I've made a PR myself.
One of the members of my team asked why is the Maybe ScriptHash needed in the PReferenceScript and after some digging it seems it's completely redundant and not used anywhere (with the comment being obsolete). Therefore I removed it and everything "just compiled".

Fixes #606

Checklist

• Commit sequence broadly makes sense and commits have useful messages
• New tests are added if needed and existing tests are updated. See Running tests for more details
• Self-reviewed the diff

[smelc]

@Swordlash> can you fix the formatting of Haskell files? You need to run:

# order matters: fourmolu first, then stylish-haskell
fourmolu -i <modified files>
stylish-haskell -i <modified files>

[smelc]

LGTM

@carbolymer> WDYT?

[Swordlash]

Formatting should be good now. Redundant whitespace it seems.

[carbolymer]

Those are used in
https://github.com/IntersectMBO/cardano-cli/blob/a157c492a4664b0ff4216e1c54a87c04adb0067b/cardano-cli/src/Cardano/CLI/EraBased/Run/Transaction.hs#L1381 to gather minting witnesses here: https://github.com/IntersectMBO/cardano-cli/blob/a157c492a4664b0ff4216e1c54a87c04adb0067b/cardano-cli/src/Cardano/CLI/EraBased/Run/Transaction.hs#L1364

[Swordlash]

This is some very unfortunate coupling, isn't it? Having a library keep redundant information (and confuse users) for the sake of one of downstream tools. Maybe we could refactor it down to cardano-cli?

[Swordlash]

Of course we can argue that if the user wants to reference a script then he probably knows the hash. If the above is not possible I would at least not make it Maybe, for the sake of API cleanliness. I think a redundant field is better than Maybe redundant field as at least there is no ambiguity which option to use.

[Swordlash]

@carbolymer I did the refactor, see the following: Swordlash/cardano-cli@b5c76a0

User provides the Maybe PolicyId himself anyway in the CLI, so cardano-cli code just temporarily stores this info in the PReferenceScript to be unpacked later. We can change the signature of the appropriate function to return a pair, so that we avoid storing the information in the datatype.

[Swordlash]

@carbolymer did you have a chance to take a look at the above? Tl;dr cardano-cli uses this field as a temporary container for user provided data and it can be safely refactored out.

[carbolymer]

Dzięki. Looks sensible. Can you open a PR to cardano-cli with those changes?

[Swordlash]

Sure thing

[Swordlash]

@carbolymer here it is: IntersectMBO/cardano-cli#918

[carbolymer]

It's used in cardano-cli in handling of minting scripts.

[carbolymer]

The purpose of this field should be documented better.