fix(sol-types): many ABI coder fixes

[DaniPopes]

Motivation

While writing tests for #433, I found that word calculation was off, so I found all these separate bugs relating to size calculation.

• rename SolCall::Arguments to Parameters
• fix fixed-length arrays and tuple decoding
  • when not dynamic, don't take a child and just decode normally
  • fixes [Bug] abi_decode panic when using sol types of fixed bytes #433
• fix String and Bytes encoded size calculation
  • before: 32 + self.len().next_multiple_of(32), this returns 0 -> 64, 1 -> 64; when it should be 0 -> 64; 1 -> 92, so it was a word off most of the time.
• fix array encoded size calculations
  • I really don't know what happened here. These should be identical to self.tokenize().total_words() * 32
• fix token word count calculations for arrays and tuples
  • dynamic should be 1 (offset) + elements.total_words(), while it was 1 + N
  • non-dynamic should be just elements.total_words(), while it was elements.head_words()
• fixed-size arrays and tuples should be treated identically
• most length tests were just wrong because of everything listed above, and were adapted to make them pass (by using either the wrong encoded array or wrapping in single-element tuple which doesn't really make sense)

cc @prestwich (:

Solution

PR Checklist

• Added Tests
• Added Documentation
• Breaking changes

[mattsse]

lgtm

[phil-ociraptor]

Woot! Was running into a confounding calldata ABI encoding issue regarding arrays in alloy 0.4.0 and I think 0.5.0 (and presumably this PR) fixes it! ty for the fix (whether here or elsewhere)