tx-generator cleanups

[MarcFontaine]

• Merge createChangeInEra with createScriptChange.
• Make mkUTxOScript polymorphic on the era.
• Merge runBenchmarkInEra with runPlutusBenchmark.
  runBenchmark can now create transactions that contain any mix of normal and plutus inputs.
• Change ToUTxO era = [Lovelace] -> ([TxOut CtxTx era], TxId -> [Fund]) to
  ToUTxO era = Lovelace -> (TxOut CtxTx era, TxIx -> TxId -> Fund)
  (Deal with every output of a TX individually. Notice : The new new type is more restricted than the old version.
  This change DOES not add new capabilities (but simplifies some parts of the code )

[mgmeier]

Maybe a dummy question... if we don't actually want it, what's the motivation for ruling out era transitions in this context? Would be great if the comment could provide that info.

[MarcFontaine]

So there are two 'era's in play: The era when the fund is created (this is wrapped an InAnyCardanoEra) and the era from getFundWitness :: forall era. IsShelleyBasedEra era => Fund -> Witness WitCtxTxIn era (That is basically when trying to spend the fund.)
getFundWitness must be partial (for example it makes no sense to spend funds in an earlier era that that not support a specific feature). But the way it was done here simply enforced that both eras are the same.
That was just lazyness on my side because writing the most general possible function is clunky.
So it does not intentionally rule out era transitions but it is a feature is just not implemented at the moment.

[MarcFontaine]

bors r+

[iohk-bors]

Build succeeded:

• buildkite/cardano-node
• ci/hydra-build:required