-
Notifications
You must be signed in to change notification settings - Fork 155
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
New DALC API #341
New DALC API #341
Conversation
Codecov Report
@@ Coverage Diff @@
## main #341 +/- ##
==========================================
+ Coverage 58.61% 59.12% +0.50%
==========================================
Files 44 44
Lines 7092 7180 +88
==========================================
+ Hits 4157 4245 +88
- Misses 2363 2364 +1
+ Partials 572 571 -1
Continue to review full report at Codecov.
|
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.
This LGTM. I've run the simple ethermint RPC tests against the CI built evmos image using the ephemeral cluster and they pass.
I'm confused how the tests are passing since the ephemeral cluster setup using an older DALC image which shouldn't have a compatible API...
IIRC in ephemeral cluster we only have single optimint node, which is an aggregator. It shouldn't sync, so it's not using changed API. Only change in part related to block submission was adding single field in |
Pinging all approvors (@liamsi @jbowen93 @adlerjohn @evan-forbes) ;) PR is updated. All comments addressed. |
This mock DA implementation reflects all new requirements: - use new DA API (DA height oriented) - store multiple Optimint block in single DA block (at same height) - optimint blocks may be located in non-consecutive DA blocks New mock requires Optimint to implement block height mapping, improved sync logic. Many unit tests are going to fail now. Implementing mentioned features is required to fix the tests.
This time mock mimicks actual DA layer completely. It "creates blocks" at given intervals. Block height can be incremented by more than 1.
With new DALC API, syncing logic has to be changed. Block manager has to keep track of DA height (it's also saved in state).
2be9218
to
eb2e87c
Compare
m.logger.Error("failed to retrieve block from DALC", "daHeight", daHeight, "errors", err.Error()) | ||
break | ||
} | ||
atomic.AddUint64(&m.daHeight, 1) |
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.
Wait, are you sure this is the right use of atomics?
What if &m.daHeight
changes in between loading it and adding 1
to it here? Then you'd be adding 2
to it.
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.
This is the only place where m.daHeight
is changed. It's atomic change, because we're reading this variable in several other places (in other goroutines). TBH the LoadUint64
in the beginning of the loop is not necessary (normal read should be fine), but I decided to handle this variable "atomically" everywhere for consistency.
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.
My concern is if e.g. &m.daHeight
is N
, then it could be read as N
in two separate threads and incremented twice to N+2
, completely bypassing N+1
. Is that ever possible, if is so, is it a problem?
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's not possible right now, as there is only one thread incrementing this variable, and I don't see a reason to add more threads/goroutines.
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.
Not going to block on #341 (comment), but if it's an issue can you open an PR to further investigate.
This PR introduces new DALC API into optimint.
Most important changes:
Actually, there's no need to explicitly map DA layer height to Optimint height.
Resolves #281.