create-testnet-data: Make --stake-delegators write credentials to disk. Keep current behavior with new --transient-stake-delegators.

[smelc]

Changelog

- description: |
    create-testnet-data: rename --stake-delegators to --transient-stake-delegators

    Introduce --stake-delegators, that generates delegators, but write credentials to disk.
# uncomment types applicable to the change:
  type:
  - feature        # introduces a new feature
  - breaking       # the API has changed in a breaking way
  # - compatible     # the API has changed but is non-breaking
  # - optimisation   # measurable performance improvements
  # - improvement    # QoL changes e.g. refactoring
  # - bugfix         # fixes a defect
  # - test           # fixes/modifies tests
  # - maintenance    # not directly related to the code
  # - release        # related to a new release preparation
  # - documentation  # change in code docs, haddocks...

Context

• Implements first task of Follow-ups on create-testnet-data v1 #506
• This is kind of reintroducing some of the behaviors that was possible before Reduce memory usage of create staked command cardano-node#4021

How to trust this PR

• Look at how delegations is computed in the new OnDisk case. Squint really hard.

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

[carbolymer]

What's the context for this comment here?

[Jimbo4350]

The benchmarking team asked us to stop writing credentials to disk (they were generating massive numbers of keys and I believe it was slowing down their tests). In the case where we want to write keys to disk we basically accept that if we are using large numbers of keys there will be a performance degradation.

[carlhammann]

I left a few nitpicky comments. Otherwise, LGTM 🎉!

[carlhammann]

In the best of all possible worlds, this should be a field of the GenesisCreateTestnetDataCmdArgs, no?

[smelc]

Yes, but I've decided to leave it as future work, until this is requested 👍

[carbolymer]

can be joined in a single ExceptT

[smelc]

I went for some code sharing instead, to make this more digestible:

-    payVK <- firstExceptT GenesisCmdFileInputDecodeError
-                  $ newExceptT $ Keys.readVerificationKeyOrFile AsPaymentKey (VerificationKeyFilePath payVKF)
-    stakeVK <- firstExceptT GenesisCmdFileInputDecodeError
-                  $ newExceptT $ Keys.readVerificationKeyOrFile AsStakeKey (VerificationKeyFilePath stakeVKF)
+    payVK   <- readVKeyFromDisk AsPaymentKey payVKF
+    stakeVK <- readVKeyFromDisk AsStakeKey   stakeVKF
     let paymentCredential = PaymentCredentialByKey $ verificationKeyHash payVK
         stakeAddrRef = StakeAddressByValue $ StakeCredentialByKey $ verificationKeyHash stakeVK
         dInitialUtxoAddr = makeShelleyAddressInEra ShelleyBasedEraShelley nw paymentCredential stakeAddrRef
@@ -514,6 +512,9 @@ computeDelegation nw delegDir dPoolParams = do
    where
      payVKF = File @(VerificationKey ()) $ delegDir </> "payment.vkey"
      stakeVKF = File @(VerificationKey ()) $ delegDir </> "staking.vkey"
+     readVKeyFromDisk role file =
+      firstExceptT GenesisCmdFileInputDecodeError $ newExceptT $
+        Keys.readVerificationKeyOrFile role (VerificationKeyFilePath file)

[Jimbo4350]

Looks good but a couple comments

[Jimbo4350]

The benchmarking team asked us to stop writing credentials to disk (they were generating massive numbers of keys and I believe it was slowing down their tests). In the case where we want to write keys to disk we basically accept that if we are using large numbers of keys there will be a performance degradation.

No need to block this PR