remove typed event usage to keep consistency

[tac0turtle]

Summary

Currently there are a few places that use typed events and the majority do not. To keep consistency we should remove typed events until the repo is ready for migration.

Once the team is ready to migrate the repo we can bring them back and roll them out all at once.

[aaronc]

Why not starting using typed events more widely?

[tac0turtle]

because there are issues with typed events like you mentioned in the issue like performance.

[aaronc]

What other issues? Why don't we just address them holistically as we're improving events?

x/authz, x/group, and x/nft are the only modules I know of using typed events. They're newer modules anyway and trying to reflect this newer design. Unless we want to abandon typed events altogether, I would say we leave the events in these modules alone and just fix any issues.

[tac0turtle]

authz, groups, and nft should not have used typed events as they were not implemented in the repo. If they were the whole repo would be using them.

why cant we be consistent in the sdk?

groups uses orm and a different module structure. we should aim to be consistent in the sdk.

[aaronc]

authz, groups, and nft should not have used typed events as they were not implemented in the repo. If they were the whole repo would be using them.

They were implemented in the repo and the plan was to migrate all events to them, there just wasn't time to do that. Are we backtracking on the plan for typed events altogether?

The performance issue relating to JSON marshaling was raised by one user, but I haven't seen actual numbers. There are alternate ways that could be more performant. We should just come up with a design to address this.

groups uses orm and a different module structure. we should aim to be consistent in the sdk.

The plan was always to adopt ORM and a new module structure more widely as well. Groups was an experiment in that direction. Instead of going backwards to the old approach just for consistency's sake, I'd rather see us figure out where we want to move towards and spend time on that.

[aaronc]

I propose #11278 as an alternative to this.

[tac0turtle]

I propose #11278 as an alternative to this.

this seems dependent on pulsar. that work is still unknown when it will land.

Ill resurrect the typed events work. I thought since it was never delivered it wasn't ready