This PR seems to be combining a number of different changes, including rather huge, sweeping, and controversial ones like completely deleting cosmos-sdk-proto.

I think something like retiring cosmos-sdk-proto and naming ibc-proto as its official successor is a much, much bigger discussion, and especially shouldn't be mixed into other changes like trying to add no_std support to cosmrs.

i can spilt. Completely delete cosmos sdk proto because it does not support no std. I want to use ibc proto directly instead. I think it can also merge the proto that two different projects depend on to minimize multiple repositories.

That's a political decision as much as it is a technical one.

You need to get buy in from both maintainers to merge the two projects, and then there's a question of where the resulting code lives.

Also, right now this project benefits from having a "monorepo" where we can make cross-cutting changes between cosmos-sdk-proto and cosmrs, which is something we'd lose if this project switched to using ibc-proto. That's a pretty major disadvantage.

Retiring cosmos-sdk-proto would also force many downstream users to migrate, which would create a lot of churn. I am definitely not convinced it's a good idea.

I'm currently implementing ics06 solomachine. I want to use the pullickey type in cosmrs directly. I don't want to reimplement it in ibc-rs. I think there will be more problems with ibc-rs and cosmrs intersecting in the future