Philippe Laferrière

In #164, there was some push-back against using `anyhow` for our errors. However, looking at projects that use ibc-rs, none of them actually make use of our `Error` enum: +...

I believe I found a good strategy for a better error systems (discussion with @romac heavily influenced the design). The new system recognizes 2 "error sources": + **Host errors:** the...

> Given that hosts will be free to introduce their desired errors through the associated Error type under the ValidationContext, why is it needed to include the Host(E) variant, and...

Reposting [this comment](https://github.com/cosmos/ibc-rs/pull/444#issuecomment-1440565785) here for better visibility: > I think the difference is that only this other crate would need to be upgraded quickly after tendermint-rs, and that would make...

Hmm, only now realizing I never actually described what I had in mind. The idea is to pull the light client (and our "host helpers") in a new crate (say...

> The recent refactoring of TokenTransferReader/Keeper to TokenTransferValidation/ExecutionContext revealed that there is a potential for host chain objects implementing core contexts to expand their capabilities by also implementing token transfer...

The big difference between `ClientState` and `ConsensusState` is how they're stored. + `ClientState`: `clients/{identifier}/clientState` + `ConsensusState`: `clients/{identifier}/consensusStates/{height}` Specifically, we remember the history of `ConsensusState`s (indexed by height), whereas we only...

Suggested grouping of methods: https://github.com/cosmos/ibc-rs/pull/386#issuecomment-1429959273

We should also keep in mind/revisit the host-based API that was discussed in [ADR 5](https://github.com/cosmos/ibc-rs/blob/main/docs/architecture/adr-005-handlers-redesign.md#host-based-api).

Additionally, note that we also check the lower bound in `check_client_consensus_height()`, as mandated by [ics-24](https://github.com/cosmos/ibc/blob/main/spec/core/ics-024-host-requirements/README.md#consensus-state-introspection). ibc-go doesn't, and neither does the spec. Maybe this check is redundant with the consensus...