Tuple fuzzers

[ariady-putra]

Like,

test prop_tuple2((a, b) via tuple(int(), bytearray())) {
  expect _: Int = a
  expect _: ByteArray = b
}

..

test prop_tuple9((a, b, .., h, i) via tuple9(int(), bytearray(), .., fuzzer_h(), fuzzer_i())) {
  expect _: Int = a
  expect _: ByteArray = b
  ..
  expect _: H = h
  expect _: I = i
}

[ariady-putra]

Usecase example: https://github.com/Anastasia-Labs/aiken-linked-list/pull/6/files#diff-51e35fd3abb5d42a8369c9f66d32af9e2f278f41300d151945281397249a4835R227-R241

test mint_validator_insert(
  (blake2b_224s, o_refs, now, penalty_address, contract_address) via fuzzer.tuple5(
    blake2b_224_fuzzer(count: 3),
    fuzz.set_between(output_reference_fuzzer(), 2, 2),
    fuzz.int_at_least(1000),
    address_fuzzer(
      FromScript(Specified(mock_script_hash("P"))),
      WithFuzzedDelegation,
    ),
    address_fuzzer(
      FromScript(Specified(mock_script_hash("I"))),
      WithFuzzedDelegation,
    ),
  ),
) {
 ..
}

Can also be combined with #22

[MicroProofs]

Merged commits to main

[ariady-putra]

Hey @MicroProofs ,

• I have this unique-n fuzzers (https://aiken-extra.github.io/test_kit/test_kit/fuzzy.html#Unique%20Fuzzers)

  It's fuzz.set_between(.., n, n) but instead of returning a list, it returns a tuple (except for n outside 2-9)

• And fuzzers to generate address, credential, etc. with some control (https://aiken-extra.github.io/test_kit/test_kit/fuzzy/fuzzer.html)

  For example:

  // random address with known payment-script:
  address_fuzzer(FromScript(Specified("KNOWN_SCRIPT_HASH")), WithFuzzedDelegation)
  
  // random multipurpose validator:
  address_fuzzer(FromScript(Fuzzed), WithTheSameDelegationScriptOrKey)
  
  // etc.

---

1. Are these worthy additions to be included in aiken/fuzz? I'll happily move the code over and create a PR.
2. Will there be a new Set type in Aiken? If so, I'll wait for it instead of returning tuples for unique fuzzers.

[MicroProofs]

1. I think those belong more to stdlib

2. Have not heard anyone talk about Set type??

[ariady-putra]

1. I think those belong more to stdlib

2. Have not heard anyone talk about Set type??

1. But stdlib would depend on fuzz?

2. I saw this CDDL and I thought there will be a new type https://github.com/IntersectMBO/cardano-ledger/blob/master/eras/conway/impl/cddl-files/extra.cddl
   I must confess I'm not proficient reading CDDL though

[MicroProofs]

1. ahh I see now. I think it's ok to add as a PR, but I'm wondering about the use cases for unique
2. No new types on-chain that's exclusive to offchain

[ariady-putra]

1. ahh I see now. I think it's ok to add as a PR, but I'm wondering about the use cases for unique

It's for type-safety, for example:

test prop(creds via set_between(cred_fuzzer(), 3, 3))

this works, but inside the test block you'll need to have

expect [a, b, c] = creds

If the set_between count is changed, either accidentally or otherwise, it will cause a runtime error.

With unique you can have

test prop((a, b, c) via unique3(cred_fuzzer()))

preventing accidental test failure during arrangement instead of assertion.