qkc/types: add transactions and signing by blockchaindevsh · Pull Request #35 · QuarkChain/goshard

blockchaindevsh · 2026-07-11T15:13:23Z

qzhodl · 2026-07-15T04:16:45Z

+	"github.com/ethereum/go-ethereum/qkc/account"
+)
+
+func bigIntToHex(data *big.Int) string {


The filename solidity_abi_utils.go seems a bit unrelated to the transaction-related changes. Maybe we should choose a more descriptive name.

Also, are all the functions in this file used by the transaction module?

Good point. I renamed this to transaction_typed_hash.go and renamed the test accordingly in a53fef1.

I also checked the helpers here: they are all used by the transaction signing path via EvmTransaction.typedHash() -> typedSignatureHash(evmTxToTypedData(tx)). The tests cover both the typed-data conversion and intermediate packed/hash outputs.

iteyelmp · 2026-07-23T11:12:48Z

+	TXList []*CrossShardTransactionDeposit `bytesizeofslicelen:"4"`
+}
+
+type CrossShardTransactionDepositList struct {


I cannot find CrossShardTransactionDepositList in pyquarkchain. The corresponding Python type seems to be CrossShardTransactionList.

Could you confirm this type is intended to match Python CrossShardTransactionList(version=1)? The field order looks different: Python serializes tx_list followed by version(uint32), while this implementation appears to encode version into the size prefix (bytes[0] = 1).

You’re right. The existing CrossShardTransactionDepositList was based on goquarkchain’s packed-prefix format and does not match pyquarkchain.

I replaced it with CrossShardTransactionList in 1e3f38e), matching Python v1 exactly: tx_list with its normal 4-byte count prefix, followed by version as a trailing uint32(1). I also removed the unused legacy-v0 list path and added Python-generated golden vectors for empty and one-entry v1 lists, plus an unsupported-version test.

syntrust · 2026-07-27T04:17:24Z

+
+// Hash returns the hash to be signed by the sender.
+// It does not uniquely identify the transaction.
+func (s EIP155Signer) Hash(tx *EvmTransaction) common.Hash {


The Hash function may also need to branch by tx version, as in Sender.

Fixed in 543f466.

syntrust · 2026-07-27T08:02:32Z

+func data(tx []map[string]string) []string {
+	t := make([]string, 0)
+	for _, v := range tx {
+		if v["types"] == "bytes" {


The key in the map is "type" instead of "types".

Fixed in 6245898 .

qzhodl · 2026-07-29T07:21:50Z

+}
+
+type CrossShardTransactionDeposit struct {
+	CrossShardTransactionDepositV0


We can remove the CrossShardTransactionDepositV0 type and merge it into CrossShardTransactionDeposit

Fixed in 0734b84.

qzhodl · 2026-07-29T07:28:53Z

+	updated       bool
+	hash          atomic.Value


Why are there three copies of hashes in Transaction, EvmTransaction, and txdata?

Removed the one in txdata in 77ed4e4, now it aligns with pyquarkchain:

EvmTransaction.hash is the RLP hash of the inner EVM transaction, while Transaction.hash is the hash of the outer typed QKC envelope.

qzhodl · 2026-07-29T07:34:36Z

+
+func (e *EvmTransaction) SetFromFullShardKey(data uint32) {
+	t := Uint32(data)
+	e.data.FromFullShardKey = &t


Why do some setters set updated to true while others don't? What does updated indicate?

Replaced the vague updated flag with hashDirty in 754a0de. It now tracks changes to fields encoded in the inner EVM RLP hash, and all relevant setters (SetGas, SetFromFullShardKey, SetNonce, and SetVRS) mark it dirty. Hash() recomputes once and then caches the new value.
SetSender does not invalidate it because it only updates the derived sender cache, which is not part of the RLP payload. Added regression coverage for each hash-affecting setter.

qzhodl · 2026-07-29T07:35:25Z

+
+func (e *EvmTransaction) SetFromFullShardKey(data uint32) {
+	t := Uint32(data)
+	e.data.FromFullShardKey = &t


Why do some setters set updated to true while others don't? What does updated indicate?

The same as above.

qzhodl · 2026-07-29T08:15:54Z

+func (tx *EvmTransaction) GasPrice() *big.Int { return new(big.Int).Set(tx.data.Price) }
+func (tx *EvmTransaction) Value() *big.Int    { return new(big.Int).Set(tx.data.Amount) }
+func (tx *EvmTransaction) Nonce() uint64      { return tx.data.AccountNonce }
+func (tx *EvmTransaction) CheckNonce() bool   { return true }


Do we really need this function?

Removed in a784c9d.

qzhodl · 2026-07-29T08:55:43Z

+//
+// Note, the input map is reowned so the caller should not interact any more with
+// if after providing it to the constructor.
+func NewTransactionsByPriceAndNonce(signer Signer, txs map[account.Recipient]Transactions) (*TransactionsByPriceAndNonce, error) {


New Geth includes this function in the miner module, not in this file

Removed in 3221b8b, it can be introduced in the future QKC mining or transaction-pool integration.

qzhodl · 2026-07-29T08:56:56Z

+// Message is a fully derived transaction and implements core.Message
+//
+// NOTE: In a future PR this will be removed.
+type Message struct {


New Geth includes this type in the state_transition.go, not in this file

Removed in 432ed4c, it can be introduced later with the state-transition integration.

qzhodl · 2026-07-29T10:58:56Z

+}
+
+// EIP155Transaction implements Signer using the EIP155 rules.
+type EIP155Signer struct {


EIP155Signer is misleading because this signer handles all QKC transaction versions, including the non-EIP-155 versions 0 and 1. The networkId field also has inconsistent semantics: it is the expected QKC network ID for v0/v1, but is ignored for v2, where tx.NetworkId() is interpreted as the Ethereum chain ID.

Could we rename this type to QKCSigner (or VersionedSigner) and explicitly carry both IDs?

type QKCSigner struct { qkcNetworkID uint32 ethChainID uint32 }

Sender could then validate qkcNetworkID for v0/v1 and ethChainID for v2. This would make signer construction unambiguous and ensure v2 chain-ID validation does not depend solely on a separate validation layer.

Fixed in f84ee04.

qzhodl · 2026-07-29T10:59:05Z

+	}
+	R = new(big.Int).SetBytes(sig[:32])
+	S = new(big.Int).SetBytes(sig[32:64])
+	V = new(big.Int).SetBytes([]byte{sig[64] + 27})


For version 2, this currently stores V as only 27/28. However, Sender() treats version 2 as EIP-155 and subtracts 2*chainID + 8, so a v2 transaction produced by SignTx cannot be recovered correctly. It also differs from pyquarkchain, which stores V = 35/36 + 2*chainID.

Please add the EIP-155 offset for v2, for example:

V = new(big.Int).SetUint64(uint64(sig[64]) + 27) if tx.Version() == 2 { V.Add(V, new(big.Int).SetUint64(8+2*uint64(tx.NetworkId()))) }

Please also add a regression test that signs a v2 transaction with SignTx, checks that V is 35/36 + 2*chainID, and verifies that Sender recovers the original address.

Fixed in 967072b.

qzhodl · 2026-07-29T11:00:18Z

+	return hexutil.Encode(new(big.Int).SetUint64(uint64(data)).Bytes())
+}
+
+func receiptToHex(data *account.Recipient) string {


should be recipientToHex

Fixed in 346858b.

blockchaindevsh requested review from iteyelmp, ping-ke, qizhou, qzhodl and syntrust July 11, 2026 15:13

blockchaindevsh mentioned this pull request Jul 13, 2026

EthStorage Devs Meeting #185 Agenda ethstorage/pm#267

Closed

qzhodl reviewed Jul 15, 2026

View reviewed changes

This was referenced Jul 15, 2026

qkc/types: add token balances #33

Merged

QKC/OP Devs Meeting #103 QuarkChain/pm#158

Closed

EthStorage Devs Meeting #186 Agenda ethstorage/pm#268

Closed

blockchaindevsh force-pushed the qkc-3-types-03-logs-receipts branch from 92c7c91 to 1d9f03d Compare July 22, 2026 15:55

blockchaindevsh force-pushed the qkc-3-types-04-transactions branch from a53fef1 to 45f102f Compare July 22, 2026 15:58

iteyelmp reviewed Jul 23, 2026

View reviewed changes

blockchaindevsh added 3 commits July 23, 2026 19:33

qkc/types: add transactions and signing

6a2137b

qkc/types: rename transaction typed hash helpers

34c979b

qkc/types: align cross-shard list with python

1e3f38e

blockchaindevsh changed the base branch from qkc-3-types-03-logs-receipts to goshard/base July 23, 2026 11:35

blockchaindevsh force-pushed the qkc-3-types-04-transactions branch from bd79a6f to 1e3f38e Compare July 23, 2026 11:35

blockchaindevsh mentioned this pull request Jul 23, 2026

QKC/OP Devs Meeting #104 QuarkChain/pm#159

Closed

syntrust reviewed Jul 27, 2026

View reviewed changes

qkc/types: hash transactions by version

543f466

syntrust reviewed Jul 27, 2026

View reviewed changes

qkc/types: fix typed transaction byte data

6245898

blockchaindevsh mentioned this pull request Jul 28, 2026

EthStorage Devs Meeting #187 Agenda ethstorage/pm#269

Closed

syntrust approved these changes Jul 28, 2026

View reviewed changes

qzhodl reviewed Jul 29, 2026

View reviewed changes

blockchaindevsh added 3 commits July 29, 2026 16:38

qkc/types: flatten cross-shard deposit fields

0734b84

qkc/types: remove unused transaction hash field

77ed4e4

qkc/types: invalidate transaction hash cache on updates

754a0de

qzhodl reviewed Jul 29, 2026

View reviewed changes

blockchaindevsh added 3 commits July 29, 2026 17:01

qkc/types: remove unused transaction nonce check

a784c9d

qkc/types: defer transaction execution message

432ed4c

qkc/types: defer transaction ordering helpers

3221b8b

qzhodl reviewed Jul 29, 2026

View reviewed changes

blockchaindevsh added 3 commits July 29, 2026 19:44

qkc/types: split signer network identifiers

f84ee04

qkc/types: encode version 2 signature V

967072b

qkc/types: rename typed transaction recipient helper

346858b

Uh oh!

Conversation

blockchaindevsh commented Jul 11, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

blockchaindevsh Jul 23, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

qzhodl Jul 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

4 participants

blockchaindevsh Jul 23, 2026 •

edited

Loading

qzhodl Jul 29, 2026 •

edited

Loading