Cosmos-sdk: General Amino audit / cleanup / IAVL changes

[aaronc]

Summary

In the process of migrating to protobuf, several pieces of the code outside of the core tx/query/state logic are also relying on amino in addition to

This issue is to track a task of auditing and migrating all the remaining bits:

• [x] [iavl](https://github.com/tendermint/iavl) completed in #265
• [x] store/ (https://github.com/cosmos/cosmos-sdk/pull/6984)
• [ ] audit the rest of the code for direct usage of amino methods and add them to this list
  
  
  
  • Legacy REST endpoints and queriers for clients backward compatibility
  
  
  • x/paramsstate encoding - ref: https://github.com/cosmos/cosmos-sdk/issues/6983
  
  
  • LegacyAmino still used with arrays of proto messages (implement MarshalJSONArray or wrap lists of proto.Message in some other message) ref: https://github.com/cosmos/cosmos-sdk/pull/7054#discussion_r472803644 related issue: https://github.com/cosmos/cosmos-sdk/issues/7559
  
  
  • x/gov/client/cli/query.go GetCmdQueryDeposits
  
  
  • x/gov/client/cli/query.go GetCmdQueryVotes
  
  
  • x/ibc/light-clients/07-tendermint/client/cli/tx.go NewCreateClientCmd
  
  

In general:

• all data types should be proto-marshalable
• all code that does marshaling should use codec.Marshaler instead of amino directly (IAVL may need to implement its own marshaler interface)

[marbar3778]

for iavl: https://github.com/tendermint/iavl/issues/242

[marbar3778]

is the plan to remove amino from IAVL before 0.39 release of the sdk?

cc @aaronc @clevinson

[aaronc]

That would be desirable

[alexanderbez]

From @erikgrinaker:

So about IAVL and Amino: it would actually be trivial to strip out Amino (the library)
but retain  the same encoding. The only thing it actually uses is varints
(which are included in the Go stdlib), and varint-prefixed byte slices.

I’m not convinced that moving to Protobuf schemas has any significant benefit over
this, especially since the Protobuf encoding is basically the same (varints everywhere).

I agree in that IAVL can be simplified and just use binary varint encoding. No need for anything special here.

[erikgrinaker]

Submitted an initial IAVL PR here: https://github.com/cosmos/iavl/pull/265

Need to figure out what to do with encoding of proofs.

[github-actions[bot]]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.

[aaronc]

This is done right? @blushi? Closing.