chore(crates): Alloy Types with Preserved Git History by refcell · Pull Request #750 · base/base

refcell · 2026-02-18T20:38:16Z

Summary

Pulls in alloy types with preserved git history.

smol helper

move to consensus, make them reusable

## Motivation  ## Solution  ## PR Checklist - [ ] Added Tests - [ ] Added Documentation - [ ] Breaking changes

## Motivation Bug arbitrary impl ## Solution Add missing `OpTxType::Eip7702` to `OpTxType::ALL` list ## PR Checklist - [ ] Added Tests - [ ] Added Documentation - [ ] Breaking changes

## Motivation Broken decoding for eip7702 ## Solution Adds conversion from u8 to `OpTxType::Eip7702` ## PR Checklist - [ ] Added Tests - [ ] Added Documentation - [ ] Breaking changes

## Motivation ref paradigmxyz/reth#12474 We need a separate `deposit_nonce` to account for deposit transaction responses which have it while it's not present in inner envelope. ## Solution Adds `nonce` field to `OptionalFields` helper. It would get deserialized only if envelope did not consume it and serialized only if present and inner envelope is a deposit ## PR Checklist - [ ] Added Tests - [ ] Added Documentation - [ ] Breaking changes

## Motivation ref https://t.me/paradigm_reth/36099 Makes envelope more consistend by making all variants hold a hash ## Solution  ## PR Checklist - [ ] Added Tests - [ ] Added Documentation - [ ] Breaking changes

adds encodeable+decodable+partialeq

### Description Adds an examples for transforming `Frame`s into `Batch`es and the reverse.

### Description Implements encoding for batch types. Steps in the direction to batch submission.

### Description Since #259 adds `Batch` encoding, book examples can decode and encode directly to and from the `Batch` types instead of the `SingleBatch` type.

### Description Removes public visibility from modules so crate docs don't show all the `pub use mod::*` types as re-exported. Also moves error types from `utils` into a new `errors` module.

### Description Adds batch compression to the `ChannelOut`.

## Motivation Protected bits are only present for legacy transactions and we should respect transaction type when getting them in `full_txs` ## Solution  ## PR Checklist - [ ] Added Tests - [ ] Added Documentation - [ ] Breaking changes

### Description Removes the `TryFrom` span batch for `RawSpanBatch`, inlining the coercion into a `to_raw_span_batch` method. Closes #266.

### Description Adds zlib compression to the `ChannelOut`. Closes #267

### Description Re-introduces the `BatchReader` into `op-alloy-protocol`. Batch compression already introduces the `brotli` and `miniz-oxzide` deps. The compression and decompression abstractions used here should be improved for both directions: compression + decompression.

### Description Now that `thiserror` supports `no_std`, we can use `thiserror` for error types over `derive_more::Display`. This PR updates the workspace to use `thiserror`.

### Description Removes error impls using `thiserror` instead.

### Description Removes re-exports from `op-alloy-genesis`.

### Description Adds Holocene timestamps for OP Sepolia and Base Sepolia hardcoded rollup configs.

### Description Small doc touchup pr for `op-alloy` crate.

### Description Cleans up `op-alloy-protocol` examples using the exported `decompress_brotli` method.

### Description Moves `OpTxType` from the `envelope` mod into it's own `tx_type` mod to make it more digestible. Also adds tests, including an `OpTxType::ALL` test that requires a newly-added variant is added to the list.

### Description Moves the `BatchTransaction` into the `batch` mod.

### Description Upstreams a small `OpTxType` conversion to `alloy_primitives::U8` from reth. --------- Co-authored-by: Matthias Seitz <matthias.seitz@outlook.de>

cb-heimdall · 2026-02-18T20:38:22Z

✅ Heimdall Review Status

Requirement Status More Info

Reviews

✅ 1/1

Denominator calculation

Show calculation

1 if user is bot 0

1 if user is external 0

2 if repo is sensitive 0

From .codeflow.yml 1

Additional review requirements

Show calculation

Max	0

0

From CODEOWNERS 0

Global minimum 0

Max 1

1

1 if commit is unverified 0

Sum 1

refcell · 2026-02-18T20:39:43Z

Placing in draft to fix ci and some minor nits

github-actions · 2026-02-18T20:41:42Z

crates/alloy/rpc-types/src/transaction.rs

+
+impl From<OpTransactionFields> for OtherFields {
+    fn from(value: OpTransactionFields) -> Self {
+        serde_json::to_value(value).unwrap().try_into().unwrap()


Double .unwrap() in a From impl — this will panic if serialization or conversion fails. Since From impls are expected to be infallible, either make this truly infallible (restructure to avoid the serde round-trip) or change to a TryFrom impl with proper error handling.

Suggested change

serde_json::to_value(value).unwrap().try_into().unwrap()

serde_json::to_value(value).expect("OpTransactionFields serialization is infallible").try_into().expect("OpTransactionFields value is a valid OtherFields map")

github-actions · 2026-02-18T20:41:44Z

crates/alloy/rpc-types/src/receipt.rs

+
+impl From<OpTransactionReceiptFields> for OtherFields {
+    fn from(value: OpTransactionReceiptFields) -> Self {
+        serde_json::to_value(value).unwrap().try_into().unwrap()


Same double-.unwrap() pattern as in transaction.rs:193. A From impl that panics on serialization failure is problematic. At a minimum, use .expect() with a message documenting why this is safe, or convert to TryFrom.

Suggested change

serde_json::to_value(value).unwrap().try_into().unwrap()

serde_json::to_value(value).expect("OpTransactionReceiptFields serialization is infallible").try_into().expect("OpTransactionReceiptFields value is a valid OtherFields map")

github-actions · 2026-02-18T20:41:45Z

crates/alloy/rpc-types-engine/src/envelope.rs

+        let mut sig = self.signature.as_bytes();
+        sig[64] = self.signature.v() as u8;
+        data.extend_from_slice(&sig[..]);
+        data.extend_from_slice(self.parent_beacon_block_root.as_ref().unwrap().as_slice());


parent_beacon_block_root is Option<B256> and .unwrap() here will panic if it's None. Since this method already returns Result, this should propagate an error instead:

Suggested change

data.extend_from_slice(self.parent_beacon_block_root.as_ref().unwrap().as_slice());

data.extend_from_slice(self.parent_beacon_block_root.as_ref().ok_or(PayloadEnvelopeEncodeError::WrongVersion)?.as_slice());

(Or add a dedicated error variant like MissingParentBeaconBlockRoot — WrongVersion is a rough fit but avoids adding a new variant.)

github-actions · 2026-02-18T20:41:47Z

crates/alloy/rpc-types-engine/src/envelope.rs

+        let mut sig = self.signature.as_bytes();
+        sig[64] = self.signature.v() as u8;
+        data.extend_from_slice(&sig[..]);
+        data.extend_from_slice(self.parent_beacon_block_root.as_ref().unwrap().as_slice());


Same .unwrap() on parent_beacon_block_root as encode_v3 — should return an error via ? instead of panicking.

github-actions · 2026-02-18T20:41:49Z

crates/alloy/rpc-types-engine/src/payload/mod.rs

+    }
+
+    /// Returns the transactions for the payload.
+    pub const fn transactions(&self) -> &Vec<Bytes> {


Return &[Bytes] instead of &Vec<Bytes> — exposing &Vec<T> in a public API is a Rust anti-pattern (clippy ptr_arg). Same applies to withdrawals() in envelope.rs:321 and versioned_hashes() in sidecar.rs:108.

Suggested change

pub const fn transactions(&self) -> &Vec<Bytes> {

pub const fn transactions(&self) -> &[Bytes] {

github-actions · 2026-02-18T20:41:50Z

crates/alloy/rpc-jsonrpsee/src/traits.rs

@@ -0,0 +1,38 @@
+#![allow(missing_docs)]


Per the project's CLAUDE.md: "Do not add #![allow(missing_docs)] or other allow-lints to suppress clippy warnings. Fix the underlying issue instead." Add doc comments to the public items in this file and remove this allow.

Note: The //! Various jsonrpsee docs module doc on line 3 doesn't compensate for missing docs on the individual trait methods.

github-actions · 2026-02-18T21:02:11Z

crates/alloy/rpc-types-engine/src/envelope.rs

+    fn ssz_bytes_len(&self) -> usize {
+        let mut len = 0;
+        len += B256::ssz_fixed_len(); // parent_beacon_block_root is always 32 bytes
+        len += self.execution_payload.ssz_bytes_len();
+        len
+    }


ssz_bytes_len is inconsistent with ssz_append: ssz_append conditionally skips parent_beacon_block_root for V1/V2 payloads, but ssz_bytes_len unconditionally includes B256::ssz_fixed_len(). This will cause callers that pre-allocate based on ssz_bytes_len to over-allocate for V1/V2 payloads, and the returned length won't match the actual encoded bytes.

Suggested change

fn ssz_bytes_len(&self) -> usize {

let mut len = 0;

len += B256::ssz_fixed_len(); // parent_beacon_block_root is always 32 bytes

len += self.execution_payload.ssz_bytes_len();

len

}

fn ssz_bytes_len(&self) -> usize {

let mut len = 0;

if !matches!(self.execution_payload, OpExecutionPayload::V1(_) | OpExecutionPayload::V2(_))

{

len += B256::ssz_fixed_len(); // parent_beacon_block_root is 32 bytes

}

len += self.execution_payload.ssz_bytes_len();

len

}

github-actions · 2026-02-18T21:02:12Z

crates/alloy/rpc-types-engine/src/flashblock/payload.rs

+    ) -> impl Iterator<
+        Item = alloy_rlp::Result<
+            alloy_eips::eip2718::WithEncoded<alloy_consensus::transaction::Recovered<T>>,
+            alloy_consensus::crypto::RecoveryError,
+        >,
+    > + '_


alloy_rlp::Result is a type alias Result<T, alloy_rlp::Error> (one type parameter). Using it with two type parameters here is likely a compile error when the k256 feature is enabled. This should be Result<..., RecoveryError> using core::result::Result:

Suggested change

) -> impl Iterator<

Item = alloy_rlp::Result<

alloy_eips::eip2718::WithEncoded<alloy_consensus::transaction::Recovered<T>>,

alloy_consensus::crypto::RecoveryError,

>,

> + '_

) -> impl Iterator<

Item = Result<

alloy_eips::eip2718::WithEncoded<alloy_consensus::transaction::Recovered<T>>,

alloy_consensus::crypto::RecoveryError,

>,

> + '_

github-actions · 2026-02-18T21:02:14Z

crates/alloy/rpc-types-engine/src/envelope.rs

+    fn from_ssz_bytes(bytes: &[u8]) -> Result<Self, ssz::DecodeError> {
+        if bytes.len() < B256::ssz_fixed_len() {
+            return Err(ssz::DecodeError::InvalidByteLength {
+                len: bytes.len(),
+                expected: B256::ssz_fixed_len(),
+            });
+        }
+
+        // Decode parent_beacon_block_root
+        let parent_beacon_block_root = {
+            let root_bytes = &bytes[..B256::ssz_fixed_len()];
+            if root_bytes.iter().all(|&b| b == 0) {
+                None
+            } else {
+                Some(B256::from_slice(root_bytes))
+            }
+        };
+
+        // Decode payload
+        let execution_payload =
+            OpExecutionPayload::from_ssz_bytes(&bytes[B256::ssz_fixed_len()..])?;
+
+        Ok(Self { parent_beacon_block_root, execution_payload })
+    }


SSZ decode for V1/V2 always tries to parse parent_beacon_block_root from the first 32 bytes, but ssz_append does not write it for V1/V2 payloads. This means from_ssz_bytes cannot correctly roundtrip V1/V2 payloads — it will consume 32 bytes of the payload data as the beacon root.

The decode logic needs to handle V1/V2 differently: either peek at the payload to determine the version first, or document that this SSZ codec only supports V3+ envelopes.

github-actions · 2026-02-18T21:16:51Z

crates/alloy/network/src/lib.rs

+        match ty {
+            OpTxType::Deposit => Err(vec!["not implemented for deposit tx"]),
+            _ => {
+                let ty = TxType::try_from(ty as u8).unwrap();


.unwrap() on TxType::try_from() will panic if OpTxType has a variant whose u8 value doesn't map to a valid TxType. Since complete_type returns Result<(), Vec<&'static str>>, propagate the error instead:

Suggested change

let ty = TxType::try_from(ty as u8).unwrap();

let ty = TxType::try_from(ty as u8).map_err(|_| vec!["unsupported transaction type"])?;

github-actions · 2026-02-18T21:16:52Z

crates/alloy/network/src/lib.rs

+
+    fn build_unsigned(self) -> BuildResult<OpTypedTransaction, Base> {
+        if let Err((tx_type, missing)) = self.as_ref().missing_keys() {
+            let tx_type = OpTxType::try_from(tx_type as u8).unwrap();


Same .unwrap() issue — build_unsigned returns BuildResult (a Result), so this conversion should propagate the error rather than panicking:

Suggested change

let tx_type = OpTxType::try_from(tx_type as u8).unwrap();

let tx_type = OpTxType::try_from(tx_type as u8).expect("TxType u8 value must be a valid OpTxType");

At minimum, use .expect() documenting the invariant. Ideally, convert the error via the BuildResult error path.

github-actions · 2026-02-18T21:16:54Z

crates/alloy/rpc-types-engine/src/envelope.rs

+    }
+
+    /// Return the withdrawals for the payload or attributes.
+    pub const fn withdrawals(&self) -> Option<&Vec<Withdrawal>> {


Return &[Withdrawal] instead of &Vec<Withdrawal> — same ptr_arg anti-pattern already flagged on transactions(). Applies here and to versioned_hashes() in sidecar.rs:108.

Suggested change

pub const fn withdrawals(&self) -> Option<&Vec<Withdrawal>> {

pub const fn withdrawals(&self) -> Option<&[Withdrawal]> {

github-actions · 2026-02-18T21:16:55Z

crates/alloy/rpc-types-engine/src/attributes.rs

+                encode_jovian_extra_data(
+                    params,
+                    default_base_fee_params,
+                    self.min_base_fee.unwrap(),


This .unwrap() is safe because line 138 checks self.min_base_fee.is_none() and returns early, but it's fragile — a future refactor could break the invariant. Use the guarded value directly via if let or map to make this structurally safe:

self.eip_1559_params .zip(self.min_base_fee) .map(|(params, min_fee)| encode_jovian_extra_data(params, default_base_fee_params, min_fee)) .ok_or(EIP1559ParamError::NoEIP1559Params)?

github-actions · 2026-02-18T21:17:34Z

Review Summary

This PR adds 6 new base-alloy-* crates under crates/alloy/ porting OP Stack alloy types with preserved git history. The code is generally well-structured and follows workspace conventions. Below are the findings organized by severity.

Correctness Issues

SSZ encode/decode mismatch (envelope.rs): ssz_bytes_len unconditionally includes parent_beacon_block_root but ssz_append conditionally skips it for V1/V2. Similarly, from_ssz_bytes always reads 32 bytes for beacon root, breaking V1/V2 roundtrips. (existing comment)
.unwrap() on parent_beacon_block_root in encode_v3/encode_v4 (envelope.rs): These methods already return Result, so panicking on None is unnecessary. (existing comment)
Wrong Result type in recover_transactions (flashblock/payload.rs:67): alloy_rlp::Result takes one type parameter but is used with two. (existing comment)
.unwrap() in complete_type and build_unsigned (network/src/lib.rs:146,186): TxType::try_from() can fail; the surrounding functions return Result, so errors should be propagated. (new comment)

API Design

&Vec<T> return types: transactions() in payload/mod.rs:448, withdrawals() in envelope.rs:321, and versioned_hashes() in sidecar.rs:108 all return &Vec<T> instead of &[T]. (existing + new comments)
#![allow(missing_docs)] in rpc-jsonrpsee/src/traits.rs: Violates project lint policy. (existing comment)
From impls with double .unwrap() (transaction.rs:192, receipt.rs): Infallible From impls should use .expect() or be changed to TryFrom. (existing comment)

Robustness

get_jovian_extra_data fragile .unwrap() (attributes.rs:146): Relies on an early return 8 lines above. Should use .zip() or destructuring to make the invariant structural. (new comment)

Minor Observations (not commented)

pub(crate) on rlp_encode_fields, rlp_encoded_fields_length, tx_type in transaction/deposit.rs — project guidelines prefer pub for all types/functions within modules
pub mod interop and pub mod predeploys in consensus/src/lib.rs expose modules publicly but not all items are re-exported at the crate root via pub use
Cargo.toml dependency ordering is not strictly waterfall-sorted by line length in several crates

Overall the types are well-ported. The SSZ encode/decode mismatch for V1/V2 payloads is the most impactful issue and should be addressed before merge.

* feat: u16a spec Proposed specification for U16a. OptimismPortal functions for super roots have been removed but the definitions remain to simplify the process of adding the code back in later. Added isFeatureEnabled logic to the SystemConfig as well as updates to the pause function for chains that do not have the ETHLockbox feature enabled. * fix: lint * fix: use split spec format * fix: typo

mattsse and others added 30 commits November 6, 2024 20:41

feat: add fn for decoded 1559 params (#236)

a56bae8

smol helper

chore: move eip1559 impls (#237)

28666b5

move to consensus, make them reusable

chore: add is dynamic fee (#245)

a7d1e44

chore: add deserde test (#248)

fce4d61

chore: bump alloy 064 (#249)

6135eb0

feat: add missing OpTxType trait impls (#258)

8b957f4

adds encodeable+decodable+partialeq

chore(book): Frames to Batches Example (#232)

cca2baa

### Description Adds an examples for transforming `Frame`s into `Batch`es and the reverse.

feat(protocol): Batch Encoding (#259)

96ca127

### Description Implements encoding for batch types. Steps in the direction to batch submission.

fix(book): Batch over SingleBatch (#260)

5b58074

### Description Since #259 adds `Batch` encoding, book examples can decode and encode directly to and from the `Batch` types instead of the `SingleBatch` type.

chore(protocol): Re-organizes Modules and Errors (#261)

5813596

### Description Removes public visibility from modules so crate docs don't show all the `pub use mod::*` types as re-exported. Also moves error types from `utils` into a new `errors` module.

feat(protocol): Brotli Compression behind std (#263)

83cf428

### Description Adds batch compression to the `ChannelOut`.

chore(protocol): Remove TryFrom (#268)

33a1d07

### Description Removes the `TryFrom` span batch for `RawSpanBatch`, inlining the coercion into a `to_raw_span_batch` method. Closes #266.

feat(protocol): ZLIB Compression (#264)

ac1ea4b

### Description Adds zlib compression to the `ChannelOut`. Closes #267

chore(workspace): Use thiserror for Error Types (#269)

94b316e

### Description Now that `thiserror` supports `no_std`, we can use `thiserror` for error types over `derive_more::Display`. This PR updates the workspace to use `thiserror`.

feat: add missing txtype tryfroms (#272)

7fe48d6

chore: Remove Error Impls (#273)

3464216

### Description Removes error impls using `thiserror` instead.

chore(genesis): Remove Re-exports (#276)

cd99b99

### Description Removes re-exports from `op-alloy-genesis`.

feat(genesis): Holocene Timestamps on Sepolia (#285)

cd91437

### Description Adds Holocene timestamps for OP Sepolia and Base Sepolia hardcoded rollup configs.

chore(op-alloy): Docs (#277)

5cd4295

### Description Small doc touchup pr for `op-alloy` crate.

chore(protocol): Cleanup Examples (#278)

71c9774

### Description Cleans up `op-alloy-protocol` examples using the exported `decompress_brotli` method.

chore(consensus): Move OpTxType and add tests (#282)

22a91c5

### Description Moves `OpTxType` from the `envelope` mod into it's own `tx_type` mod to make it more digestible. Also adds tests, including an `OpTxType::ALL` test that requires a newly-added variant is added to the list.

chore(protocol): Batch Transaction Mod (#284)

9bdce9a

### Description Moves the `BatchTransaction` into the `batch` mod.

chore(consensus): OpTxType Conversion (#283)

1a078a0

### Description Upstreams a small `OpTxType` conversion to `alloy_primitives::U8` from reth. --------- Co-authored-by: Matthias Seitz <matthias.seitz@outlook.de>

refcell requested review from cody-wang-cb and haardikk21 February 18, 2026 20:38

refcell self-assigned this Feb 18, 2026

refcell added the A-repo Area: general repository changes label Feb 18, 2026

refcell marked this pull request as draft February 18, 2026 20:39

github-actions bot reviewed Feb 18, 2026

View reviewed changes

cleanup and remove unused crates

39013d3

github-actions bot reviewed Feb 18, 2026

View reviewed changes

refcell marked this pull request as ready for review February 18, 2026 21:03

github-actions bot reviewed Feb 18, 2026

View reviewed changes

danyalprout approved these changes Feb 18, 2026

View reviewed changes

refcell added this pull request to the merge queue Feb 18, 2026

Merged via the queue into main with commit 72eada2 Feb 18, 2026
19 of 20 checks passed

refcell deleted the alloy branch February 18, 2026 21:39

refcell restored the alloy branch February 18, 2026 22:04

	serde_json::to_value(value).unwrap().try_into().unwrap()
	serde_json::to_value(value).expect("OpTransactionFields serialization is infallible").try_into().expect("OpTransactionFields value is a valid OtherFields map")

	data.extend_from_slice(self.parent_beacon_block_root.as_ref().unwrap().as_slice());
	data.extend_from_slice(self.parent_beacon_block_root.as_ref().ok_or(PayloadEnvelopeEncodeError::WrongVersion)?.as_slice());

	pub const fn transactions(&self) -> &Vec<Bytes> {
	pub const fn transactions(&self) -> &[Bytes] {

	let ty = TxType::try_from(ty as u8).unwrap();
	let ty = TxType::try_from(ty as u8).map_err(\|_\| vec!["unsupported transaction type"])?;

	let tx_type = OpTxType::try_from(tx_type as u8).unwrap();
	let tx_type = OpTxType::try_from(tx_type as u8).expect("TxType u8 value must be a valid OpTxType");

	pub const fn withdrawals(&self) -> Option<&Vec<Withdrawal>> {
	pub const fn withdrawals(&self) -> Option<&[Withdrawal]> {

Conversation

refcell commented Feb 18, 2026

Summary

Uh oh!

cb-heimdall commented Feb 18, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

✅ Heimdall Review Status

Uh oh!

refcell commented Feb 18, 2026

Uh oh!

github-actions bot Feb 18, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions bot Feb 18, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions bot Feb 18, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions bot Feb 18, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions bot Feb 18, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions bot Feb 18, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions bot Feb 18, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions bot Feb 18, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions bot Feb 18, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions bot Feb 18, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions bot Feb 18, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions bot Feb 18, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions bot Feb 18, 2026

Choose a reason for hiding this comment

Uh oh!

github-actions bot commented Feb 18, 2026

Review Summary

Correctness Issues

API Design

Robustness

Minor Observations (not commented)

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

20 participants

cb-heimdall commented Feb 18, 2026 •

edited

Loading