test for divergence for Fast USDC

[turadg]

What is the Problem Being Solved?

To qualify for Mainnet we must test for divergence.

Description of the Design

• private test (requires Ability to test validator divergence in multichain-testing #10896)
• on Devnet (but that's in use by u18a)
• an Instagoric if we add in Noble and Osmosis test chains

Security Considerations

Scaling Considerations

Test Plan

Cover the Golang code paths that support fast-usdc. Does not need load.

IBC coming in (Noble) and going out (e.g. Osmosis).
Cover address hooks.

Does not have to work in CI. Just a proof of it working that SRE can verify.

Upgrade Considerations

[dckc]

Devnet Deployment notes, for reference:

• 28 Jan: #81. USDC vbank registration (plain)
• 28 Jan: Fast USDC test deployment on devnet - Platform & BLDers - Agoric Community Forum
  • failed due to mistake in proposal files
• 7 Feb: #84 Fast USDC Test Deployment (updated)
  • failed due to lack of cosmosOrchestrationService
• 11 Feb: #86 init-orchestration to support Fast USDC testing
  • failed for lack of a getPortAllocator method

[turadg]

Given the test network hurdles, we instead did deeper static analysis of the Golang code. @michaelfig is confident there is no non-deterministic behavior in any Golang code related to Fast USDC, and other panic problems would have been discovered in multi-chain testing.