feat: Add TrieUpdatesSorted and HashedPostStateSorted in all ExEx notifications
#20333
Conversation
This changes the TrieUpdates field of the Chain type (which is used in ExEx notifications) from an Option into a BTreeMap mapping the block number to the updates pulled from the ExecutedBlock. With this in place all ExEx subscribes have full access to all trie updates produced by each block.
dhyaniarun1993
requested review from
Rjected,
mattsse,
mediocregopher,
rakita and
shekhirin
as code owners
dhyaniarun1993
changed the title
hashed_state [WIP]
dhyaniarun1993
changed the title
hashed_state [WIP] hashed_state
dhyaniarun1993
requested review from
DaniPopes,
fgimenez,
joshieDo,
klkvr and
yongkangc
as code owners
|
seems like needs rebase. probably easier if you simply cherry-pick the commits that are new to this pr (the ones that don't have a pr link in brackets at end), onto the base branch, and then force push @dhyaniarun1993 |
dhyaniarun1993
changed the base branch from
mediocregopher/ex-ex-trie-updates
to
main
dhyaniarun1993
changed the base branch from
main
to
mediocregopher/ex-ex-trie-updates
I think it's upto date with main. The base branch is not updated. |
emhane
changed the base branch from
mediocregopher/ex-ex-trie-updates
to
main
crates/exex/exex/src/wal/storage.rs
Outdated
| let block_number = block.header().number(); | ||
|
|
||
| // Create some non-empty trie updates and hashed state to ensure the 4-field format is used | ||
| let trie_updates = TrieUpdatesSorted::default(); |
There was a problem hiding this comment.
Can we include some nominal trie updates/state changes to make this fully end-to-end? If someone accidentally breaks either of those serialization formats we would want this new .wal file test to break.
There was a problem hiding this comment.
Sure. I add some cases here.
|
Just a minor nit about the test case, and then the feature propagation CI error, other than that this looks good, thanks! 👍 |
dhyaniarun1993
changed the title
hashed_stateTrieUpdatesSorted and HashedPostStateSorted in all ExEx notifications
| /// Bincode-compatible [`super::TrieUpdatesSorted`] serde implementation. | ||
| /// | ||
| /// Intended to use with the [`serde_with::serde_as`] macro in the following way: | ||
| /// ```rust | ||
| /// use reth_trie_common::{serde_bincode_compat, updates::TrieUpdatesSorted}; | ||
| /// use serde::{Deserialize, Serialize}; | ||
| /// use serde_with::serde_as; | ||
| /// | ||
| /// #[serde_as] | ||
| /// #[derive(Serialize, Deserialize)] | ||
| /// struct Data { | ||
| /// #[serde_as(as = "serde_bincode_compat::updates::TrieUpdatesSorted")] | ||
| /// trie_updates: TrieUpdatesSorted, | ||
| /// } | ||
| /// ``` | ||
| #[derive(Debug, Serialize, Deserialize)] | ||
| pub struct TrieUpdatesSorted<'a> { | ||
| account_nodes: Cow<'a, [(Nibbles, Option<BranchNodeCompact>)]>, | ||
| storage_tries: B256Map<StorageTrieUpdatesSorted<'a>>, | ||
| } |
There was a problem hiding this comment.
I think we want to create a new module (can be part of this file) called serde_bincode_compat, and feature gate this code behind the existing feature for this crate serde-bincode-compat. same for the hashed post state .
There was a problem hiding this comment.
Yep, we are already doing that https://github.com/op-rs/op-reth/blob/8605dc3b378f5bbce808e16a0dc7c0f0d7650433/crates/trie/common/src/lib.rs#L72-L78
emhane
removed
the
S-blocked
3 participants
Based onincludes changes from #19355