#

https://github.com/cosmos/ibc-go/discussions/2191#discussioncomment-3709253

- TL;DR / action items
- Initial implementation will require a relayer
- Aditya to file specific IBC bugs for loopback
- Justin to schedule the work
- Implementation to proceed on the Strangelove fork
- Aditya will review
- how to manage issues/PRs?
- Aditya will post individual issues on IBC branch
- SL would upstream PRs for each issue

- open questions
- ? how do packets flow
- :- send and receive in one step
- \- has gas concerns when message is delivered
- :+ rely on a local relayer
- + simpler implementation
- RL needs to gets success/fail of the operations
- JS ICA library already provides that in the API
- v3 to v5 the chain has to have ICA auth
- custom application logic for auth
- ICA auth module will get the acknowledgement
- it will need to forward the ack back to contract
- Agoric JS ICA controller does ack processing
- instead of atomicity
- could trigger deliver in the endblock
- Dates
- v6 out
- target: end of Oct (ideally)
- has meta-data field
- also backported to other version
- v7
- target: TBD
- would include localhost

# Aditya notes
Aditya, [Oct 12, 2022 at 6:49:44 AM]:
Hi Dean, a bunch of high-priority stuff got dropped on my plate this morning

still worth having the meeting but would like to timebox my contributions so i can focus on the other issues

Here are my thoughts on the fork just so we don't need to go over that during the meeting:

Ideally you can wait for it to get into IBC main. We're wrapping up v6 right now which will take a couple weeks and then complete the release pipeline for it (now in alpha)

after that localhost will get into v7 along with the other client refactor changes

I can understand if Agoric can't wait that long. Definitely do not want to block you unnecessarily

If you need it sooner, you can for sure have Strangelove help out

I can review Strangelove prs but can't commit more resources on that front in the immediate future

I can also work with Jack on how we'd want that work eventually upstreamed to ibc-go

But until the code is in an ibc-go release, our team can't guarantee correctness of the solution. So it would be expected that Strangelove is willing to make that guarantee and will provide you guys support for any bugs that arise

Of course, we will work on getting localhost on our end as soon as possible with the same architecture outlined in the ibc-go discussion. And then agoric should be able to seamlessly switch back to using cosmos/ibc-go

Hope that all makes sense

If you're good with that direction, can we structure the agenda so that any technical questions needed from me for Strangelove to get started right away comes first?

sorry for having to streamline, hope i can unblock you guys today and then focus on putting out some fires