-
Notifications
You must be signed in to change notification settings - Fork 616
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
chore: add solomachine client storage #1144
chore: add solomachine client storage #1144
Conversation
// TODO: Should be telemetry and events be emitted from 02-client? | ||
// The clientID would need to be passed an additional arg here |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
I think we need to make a decision on whether telemetry and events should remain to be emitted from 02-client
or from within the ClientState
implementation. We may need to pass the client ID as an additional arg if we want to do this in the lightclient implementation, or otherwise we may need to return the consensus height. Any thoughts on this? cc. @colin-axner @seantking
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
We should not require light client implementations to emit events. This decision informed the proposed design. We should allow light client to optionally emit their own events and telemetry. This is technically possible by constructing a ClientMessage
which contains the clientID, but I'm not opposed to adding the clientID as an argument to the function since 02-client already has that information. It doesn't provide any usefulness outside of event emission/telemetry since the client store is prefixed by the clientID
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
To expand further on why we should not require light client implementations to emit events:
Relayers make use of the events we have in 02-client. By passing this to light client implementations, we lose standardization and risk light client implementators making mistakes. Since the functionality called by 02-client is very verbose now, 02-client can emit appropriate events when doing updates or updates on misbehaviour
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Okay, great, that makes a lot of sense! Thanks, I'll remove the events and telemetry from here.
// set default consensus height with header height | ||
consensusHeight := smHeader.GetHeight() | ||
if cs.ClientType() != exported.Localhost { | ||
clientStore.Set(host.ConsensusStateKey(consensusHeight), types.MustMarshalConsensusState(cdc, consensusState)) | ||
} else { | ||
consensusHeight = types.GetSelfHeight(ctx) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
where is this coming from? Solo machines don't store consensus states
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Slightly confused here, isn't this code run for solomachines too?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
It is, but it shouldn't be. This was another justification for why setting the consensus state within the light client implementation is useful. I thought I documented this clearly, but this was all I could find:
This would require light client developers to set the consensus states themselves (allowing header.GetHeight to be removed and stopping unnecessary state storage for solo machines).
Solo machines shouldn't be setting a consensus state ever. It has a single consensus state which is constantly updated and thus it is stored within the client state. The code linked above was an inefficiency, performing unnecessary storage
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Ahh, okay! Tbh I was confused why ConsensusState
existed as a field on the solomachine ClientState
and the extra storage was happening.
This should all be much easier to reason about when the refactor is done. Thank you for clarifying this and providing context, much appreciated!
Updated and removed this code :)
// TODO: Should be telemetry and events be emitted from 02-client? | ||
// The clientID would need to be passed an additional arg here |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
To expand further on why we should not require light client implementations to emit events:
Relayers make use of the events we have in 02-client. By passing this to light client implementations, we lose standardization and risk light client implementators making mistakes. Since the functionality called by 02-client is very verbose now, 02-client can emit appropriate events when doing updates or updates on misbehaviour
@@ -615,7 +623,7 @@ func (suite *SoloMachineTestSuite) TestUpdateState() { | |||
suite.Require().Equal(consensusState, cs.(*types.ClientState).ConsensusState) |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
can you add checks on the successful case that the correct client state is set in the store?
func (cs ClientState) UpdateStateOnMisbehaviour( | ||
_ sdk.Context, _ codec.BinaryCodec, _ sdk.KVStore, // prematurely include args for self storage of consensus state | ||
) (*ClientState, exported.ConsensusState, error) { | ||
func (cs ClientState) UpdateStateOnMisbehaviour(ctx sdk.Context, cdc codec.BinaryCodec, clientStore sdk.KVStore) (*ClientState, exported.ConsensusState, error) { |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Is there a test for this function? Can you check that the correct client state is set in the store?
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Yeah, there is a test function for this. I was initially just doing an assertion on the returned value but I can adapt it to query the store. Will do the same as you suggest for UpdateState
👍
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Sweet, I just figure better to do now so when we remove the return values we don't have to add testing code
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Definitely! Agreed 💯
…/ibc-go into damian/solomachine-client-storage
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Great work!
* WIP: solomachine UpdateState now handles storage of ClientState and ConsensusState * applying suggestions for review * removing unnecessary consensus state storage * updating tests to use store in favour of return values
## Overview This PR uses pkg `httptest` to avoid port conflict when running tests under `da/test` i.e. `TestLifecycle`, `TestRetrieve`. This allows the tests to pass even if an address is bound to the default port i.e. `127.0.0.1:26658` ## Checklist <!-- Please complete the checklist to ensure that the PR is ready to be reviewed. IMPORTANT: PRs should be left in Draft until the below checklist is completed. --> - [ ] New and updated code has appropriate documentation - [ ] New and updated code has new and/or updated testing - [ ] Required CI checks are passing - [ ] Visual proof for any user facing features like CLI or documentation updates - [ ] Linked issues closed with keywords
## Overview This PR improves upon cosmos#1144 by simplifying `mock.Server` to use `httptest` directly, avoiding a duplicate `http.Server`. `Server.Start` doesn't need a listener anymore, and it returns the URL at which the `httptest.Server` is running. ## Checklist <!-- Please complete the checklist to ensure that the PR is ready to be reviewed. IMPORTANT: PRs should be left in Draft until the below checklist is completed. --> - [ ] New and updated code has appropriate documentation - [ ] New and updated code has new and/or updated testing - [ ] Required CI checks are passing - [ ] Visual proof for any user facing features like CLI or documentation updates - [ ] Linked issues closed with keywords
Description
UpdateState
andUpdateStateOnMisbehaviour
closes: #XXXX
Before we can merge this PR, please make sure that all the following items have been
checked off. If any of the checklist items are not applicable, please leave them but
write a little note why.
docs/
) or specification (x/<module>/spec/
)godoc
comments.Unreleased
section inCHANGELOG.md
Files changed
in the Github PR explorerCodecov Report
in the comment section below once CI passes