Test payload serialization

[m-chrzan]

• Added a few more conditions in existing payload tests.
• Added payload generation with private key rather than prepared signature.
• Added tests with two data packages.
• (most importantly) Added tests that pass the serialized payload to RedStone's deserializer, verify it deserializes correctly.

Two slight concerns that I noticed while going through the deserialization code:

• The aggregateValues hook has to be view. This is the only place where we have access to the array of values in a report batch. Initially I had thought this would have been the function where we implement most of our functionality, but we can't write to storage/call other contracts from here. We might be able to work around this by encoding aggregated data (specifically, isWithinAllowedDeviation, isCertain) in the return value of this function (it returns uint256 while our rate feed values fit in uint64).
• The deserializer takes into account only uniqueSignersThreshold values. It won't bother passing more into aggregateValues if there's more in the batch. This could cause two problems:
  • Slightly reduces security, by forcing the Oracle to always operate at the minimum allowed number of operators taken into account.
  • Could bias the median, given we're expecting the list of values in a batch to be sorted (so values from one extreme will be cut off by this behavior).

[chapati23]

just a general question: are the tests supposed to be passing already or no? just saw failing CI

[m-chrzan]

just a general question: are the tests supposed to be passing already or no? just saw failing CI

The Oracles.sol ones don't need to - that's where I've set up some scaffolding tests prior to code implementation. The payload and OracleValue tests should be passing.