ICS20-V2 Path Forwarding E2E Tests

[chatton]

Summary

Tracking issue for ICS20-V2 Path Forwarding E2E tests

Tests:

• Given 3 chains with ics20-2 channels between A and B, and between B and C, perform a successful transfer of tokens from A to C through B. [Nikolas] (test) Test that tokens are successfully forwarded from A to C through B #6758
• Given 3 chains with ics20-2 channels between A and B, and between B and C, send first tokens from A to B, and then perform a successful unwind + forward transfer of the IBC tokens from B to C through A. Topology: A -> B -> C, and A -> C. [Nikolas] (test) - Three chains path forwarding with unwind #6778
• Given 3 chains with ics20-2 channels between A and B, and between B and C, perform a transfer of tokens from A to C through B, where the hop between B and C will fail. (tests): e2e test for failed forwarding #6876
• Given 3 chains with ics20-1 channels between A and B and B and C, perform a successful channel upgrade from v1 to v2 for both channels and send tokens from A to C through B. [Nikolas] (test) Add test for path forwarding with channel upgrade  #6793
• Integration test: a test where the memo triggers the execution of a callback only on the final destination. [Nikolas] (test) Add tests for callback execution when forwarding a packet #6805

---

For Admin Use

• Not duplicate issue
• Appropriate labels applied
• Appropriate contributors tagged/assigned

[bznein]

I'll start by taking the following one:

Given 3 chains with ics20-2 channels between A and B, and between B and C, perform a successful transfer of tokens from A to C through B.

[crodriguezvega]

Just a note to remember to update the unreleased compatibility tests when we have all the e2e tests ready.

[bznein]

Given 3 chains with ics20-2 channels between A and B, and between B and C, perform a transfer of tokens from A to C through B, where the hop between B and C will time out. [Nikolas]

I discussed with @chatton on this one and it doesn't seem like we have a way to cause a timeout on e2e tests. We could arrange something like creating a relayer that only watches A->B, wait for timeout, and then update it to watch B to C, but that would be ugly and not possible to do cleanly right now. Right now we do have an integration timeout test (TestOnTimeoutPacketForwarding) which might be enough. @crodriguezvega what do you think?

[crodriguezvega]

All tests have been completed 🥳 Shout-out to @bznein for taking care of all these tests!