-
Notifications
You must be signed in to change notification settings - Fork 1.2k
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
primitives: use alloy Receipts
#12059
Conversation
@@ -17,12 +17,10 @@ where | |||
&mut self, | |||
first_tx_index: TxNumber, | |||
_: BlockNumber, | |||
receipts: Vec<Option<Receipt>>, |
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 would break receipt pruning during staged sync
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.
and the fact that all tests pass, probably means we should create a test to ensure this
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.
so you mean that for this specific function we should keep Vec<Option<Receipt>>
or does it affect also other methods? In summary you want to keep this behavior right? (Was thinking that this was equivalent to simply have no receipt)
If
receipt
isNone
, it has been pruned.
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.
@joshieDo This means also that in ExecutionOutcome
, we cannot use Receipts
directly here no?
pub receipts: Receipts, |
Because with Receipts
, we have a vec[vec[Receipt]]
but here you mention the need for vec[vec[Option<Receipt>]]
due to staged sync pruning flags no?
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 keep Vec<Option>
yes. im not sure on effects on other methods (cc @shekhirin )
But maybe we could just stick to the PR's goal of replacing the type and not touch anything else? Or is this somehow required?
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.
Yes the initial goal of this PR was indeed to replace the Receipts
type with its alloy equivalent but as discussed here alloy-rs/alloy#1247 (review), in alloy we no longer have an Option
around the Receipt
natively inside Receipts
.
To try to do things little by little I just created another PR #12162 to minimize the changes with:
pub type Receipts = alloy_consensus::Receipts<Option<Receipt>>;
What I propose:
- We merge the first small PR that eliminates the implementation of
Receipts
in reth primitives: use alloyReceipts
withOption
#12162. - We close this PR (where we are discussing right now) which will then become obsolete.
- We open other small issues to try to better prune and add unit tests according to consensus: add
Receipts
struct alloy-rs/alloy#1247 (review) - Gradually we try to remove the
Option
fromReceipts
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 misread this as replacing Receipt and not Receipts, my bad.
Maybe we should tacke first the pruning rather the type replacement?
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.
@joshieDo As you prefer, I think both ways are fine:
- We could maybe first merge primitives: use alloy
Receipts
withOption
#12162 first so that we can play easier with theOption
or not because the alloyReceipts<T>
has the generic type T so that maybe we can replace step by step. - Also it would maybe allow to directly focus on alloy and not cause additional regression on reth.
- But if you don't agree no worries, we can first fix the pruning and then merge primitives: use alloy
Receipts
#12059
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.
@joshieDo @mattsse After rethinking about this, maybe this is simpler to merge #12162 first and then we can play with the new type at the required places, progressively, by doing, at choice:
alloy::Receipts<Option<Receipt>>
alloy::Receipts<Receipt>
depending on the locations.
In this way we could progress slowly, what do you think?
Closed with #12162 |
Ref #7238