implement logic to build genesis state for the sequencer #2371

rianhughes · 2025-01-13T12:24:28Z

This PR implements the genesis pkg that builds a state diff from declaring classes, deploying contracts and invoking functions. The sequencer can then commit this state diff to state, so that upon startup, a set of account contracts have already been deployed and funded (it is not possible to achieve this from an empty state by submitting transactions, since Starknet requires accounts to be funded to perform any actions on the state - originally this wasn't a requirement in Starknet). This is particularly useful as it allows users to start the sequencer with any number of accounts pre-deployed, and pre-funded, etc, and opens up the possibility of using Juno as a devnet

In essence:

we load classes, then declare them,
execute the constructor functions using the vm, which writes the changes to the pending state
execute any functions of interest (eg token transfers) using the vm (where similarly the vm writes the changes to the pending state)
extract the state-diff and new classes from the pending state.

weiihann · 2025-01-13T12:57:43Z

genesis/genesis.go

+	for k, v := range aux.Contracts {
+		key, err := new(felt.Felt).SetString(k)
+		if err != nil {
+			return err
+		}
+		g.Contracts[*key] = v
+	}


I don't think we need to do dereferencing? Just do:

for k, v := range aux.Contracts { var key felt.Felt key, err = key.SetString(k) if err != nil { return err } g.Contracts[key] = v }

Ah right, in your approach key is allocated to the stack and not the heap, whereas in the original case key is allocated to the heap?

Is the motivation for this change so that key is allocated to the stack rather than the heap?

The suggested change doesn't work, since SetString only operates on a pointer
func (z *Felt) SetString(number string) (*Felt, error) {
and initialising a nil pointer will result in a panic when calling SetString (ie using var key *felt.Felt)

wip - fix test fix stuff

codecov · 2025-01-13T14:48:27Z

Codecov Report

Attention: Patch coverage is 65.08475% with 103 lines in your changes missing coverage. Please review.

Project coverage is 74.51%. Comparing base (d99e28b) to head (aad7fc7).
Report is 2 commits behind head on main.

Files with missing lines	Patch %	Lines
genesis/genesis.go	64.94%	42 Missing and 19 partials ⚠️
adapters/vm2core/vm2core.go	56.09%	14 Missing and 4 partials ⚠️
sync/pending.go	68.00%	14 Missing and 2 partials ⚠️
core/state_update.go	73.91%	5 Missing and 1 partial ⚠️
node/throttled_vm.go	60.00%	1 Missing and 1 partial ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##             main    #2371      +/-   ##
==========================================
- Coverage   74.70%   74.51%   -0.19%     
==========================================
  Files         110      111       +1     
  Lines       12021    12313     +292     
==========================================
+ Hits         8980     9175     +195     
- Misses       2349     2419      +70     
- Partials      692      719      +27

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

genesis/genesis.go

core/state_update.go

AnkushinDaniil · 2025-01-15T08:36:40Z

adapters/vm2core/vm2core.go

@@ -39,3 +40,52 @@ func AdaptOrderedEvents(events []vm.OrderedEvent) []*core.Event {
 	})
 	return utils.Map(events, AdaptOrderedEvent)
 }
+
+func AdaptStateDiff(trace *vm.TransactionTrace) *core.StateDiff {


As I can see we only use trace.StateDiff in this function, why do we pass trace as a parameter?

AnkushinDaniil · 2025-01-15T10:58:35Z

sync/pending.go

+	newNonce := new(felt.Felt).SetUint64(1)
+	p.stateDiff.Nonces[*contractAddress] = newNonce.Add(currentNonce, newNonce)


We have felt.Zero, can we have felt.One?

Suggested change

newNonce := new(felt.Felt).SetUint64(1)

p.stateDiff.Nonces[*contractAddress] = newNonce.Add(currentNonce, newNonce)

p.stateDiff.Nonces[*contractAddress] = currentNonce.Add(currentNonce, &felt.One)

AnkushinDaniil · 2025-01-15T11:06:20Z

sync/pending.go

+}
+
+func (p *PendingStateWriter) SetCompiledClassHash(classHash, compiledClassHash *felt.Felt) error {
+	// assumption: SetContractClass was called for classHash and succeeded


Can we make these assumptions a part of the function’s doc comment?

AnkushinDaniil

I have finished reviewing everything except vm logic. Everything is good, just small changes and questions. I'll continue reviewing vm logic.

AnkushinDaniil · 2025-01-16T08:43:55Z

adapters/vm2core/vm2core.go

+		return nil
+	}
+	stateDiff := trace.StateDiff
+	newStorageDiffs := make(map[felt.Felt]map[felt.Felt]*felt.Felt)


I think we could reduce memory allocation this way

Suggested change

newStorageDiffs := make(map[felt.Felt]map[felt.Felt]*felt.Felt)

newStorageDiffs := make(map[felt.Felt]map[felt.Felt]*felt.Felt, len(stateDiff.StorageDiffs))

AnkushinDaniil · 2025-01-16T08:50:25Z

core/state_update.go

+	mergeStorageDiffs := func(oldMap, newMap map[felt.Felt]map[felt.Felt]*felt.Felt) {
+		for addr, newAddrStorage := range newMap {
+			if oldAddrStorage, exists := oldMap[addr]; exists {
+				maps.Copy(oldAddrStorage, newAddrStorage)
+			} else {
+				oldMap[addr] = newAddrStorage
+			}
+		}
+	}


Could we just inline this function?

AnkushinDaniil · 2025-01-16T09:01:27Z

genesis/genesis.go

+	type auxConfig struct {
+		Classes           []string                       `json:"classes"`            // []path-to-class.json
+		Contracts         map[string]GenesisContractData `json:"contracts"`          // address -> {classHash, constructorArgs}
+		FunctionCalls     []FunctionCall                 `json:"function_calls"`     // list of functionCalls to Call()
+		BootstrapAccounts []Account                      `json:"bootstrap_accounts"` // accounts to prefund with strk token
+		Txns              []*rpc.Transaction             `json:"transactions"`       // declare NOT supported
+	}
+	var aux auxConfig


As I can see, we use auxConfig once.

Suggested change

type auxConfig struct {

Classes []string `json:"classes"` // []path-to-class.json

Contracts map[string]GenesisContractData `json:"contracts"` // address -> {classHash, constructorArgs}

FunctionCalls []FunctionCall `json:"function_calls"` // list of functionCalls to Call()

BootstrapAccounts []Account `json:"bootstrap_accounts"` // accounts to prefund with strk token

Txns []*rpc.Transaction `json:"transactions"` // declare NOT supported

}

var aux auxConfig

var aux struct {

Classes []string `json:"classes"` // []path-to-class.json

Contracts map[string]GenesisContractData `json:"contracts"` // address -> {classHash, constructorArgs}

FunctionCalls []FunctionCall `json:"function_calls"` // list of functionCalls to Call()

BootstrapAccounts []Account `json:"bootstrap_accounts"` // accounts to prefund with strk token

Txns []*rpc.Transaction `json:"transactions"` // declare NOT supported

}

AnkushinDaniil · 2025-01-16T09:04:51Z

genesis/genesis.go

+	validate := validator.Validator()
+	return validate.Struct(g)


Can we inline validate?

Suggested change

validate := validator.Validator()

return validate.Struct(g)

return validator.Validator().Struct(g)

AnkushinDaniil · 2025-01-16T09:13:26Z

genesis/genesis.go

+	for classHash, class := range newClasses {
+		// Sets pending.newClasses, DeclaredV0Classes, (not DeclaredV1Classes)
+		if err = genesisState.SetContractClass(&classHash, class); err != nil {
+			return nil, nil, fmt.Errorf("declare v0 class: %v", err)
+		}
+
+		if cairo1Class, isCairo1 := class.(*core.Cairo1Class); isCairo1 {
+			if err = genesisState.SetCompiledClassHash(&classHash, cairo1Class.Compiled.Hash()); err != nil {
+				return nil, nil, fmt.Errorf("set compiled class hash: %v", err)
+			}
+		}
+	}


Could we use descriptive variable names instead of this comment?

// Sets pending.newClasses, DeclaredV0Classes, (not DeclaredV1Classes)

I don't quite understand why only DeclaredV0Classes in newClasses.

AnkushinDaniil · 2025-01-16T10:04:53Z

genesis/genesis.go

+			}
+
+		default:
+			panic("transaction type not supported")


Suggested change

panic("transaction type not supported")

return nil, nil, fmt.Errorf("unsupported transaction type: %v", txn.Type)

AnkushinDaniil · 2025-01-16T10:06:53Z

genesis/genesis.go

+		fmt.Println("==")
+		fmt.Println(genesisSD)
+		fmt.Println(traceSD)


Suggested change

fmt.Println("==")

fmt.Println(genesisSD)

fmt.Println(traceSD)

AnkushinDaniil · 2025-01-16T10:21:54Z

genesis/genesis.go

+		} else if compiledClass, cErr := compiler.Compile(response.V1); cErr != nil {
+			return nil, cErr
+		} else if coreClass, err = sn2core.AdaptCairo1Class(response.V1, compiledClass); err != nil {
+			return nil, err
+		}


I think this code is easier to understand

Suggested change

} else if compiledClass, cErr := compiler.Compile(response.V1); cErr != nil {

return nil, cErr

} else if coreClass, err = sn2core.AdaptCairo1Class(response.V1, compiledClass); err != nil {

return nil, err

}

} else {

var compiledClass *starknet.CompiledClass

if compiledClass, err = compiler.Compile(response.V1); err != nil {

return nil, err

}

if coreClass, err = sn2core.AdaptCairo1Class(response.V1, compiledClass); err != nil {

return nil, err

}

}

AnkushinDaniil · 2025-01-16T10:40:37Z

sync/pending.go

+// StateDiffAndClasses returns the pending state's internal data. The returned objects will continue to be
+// read and modified by the pending state.


If so, could we make these fields public?

rianhughes marked this pull request as draft January 13, 2025 12:29

weiihann reviewed Jan 13, 2025

View reviewed changes

implement logic to build genesis state for the sequencer

459db5d

wip - fix test fix stuff

rianhughes force-pushed the feature/sequencer-genesis branch from fb11e07 to 459db5d Compare January 13, 2025 14:37

rianhughes marked this pull request as ready for review January 13, 2025 14:38

weiihann reviewed Jan 14, 2025

View reviewed changes

genesis/genesis.go Outdated Show resolved Hide resolved

weiihann reviewed Jan 14, 2025

View reviewed changes

core/state_update.go Outdated Show resolved Hide resolved

rianhughes added 2 commits January 14, 2025 13:43

comments: merge method, const string

fbae35c

lint

aad7fc7

AnkushinDaniil reviewed Jan 15, 2025

View reviewed changes

AnkushinDaniil reviewed Jan 16, 2025

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

implement logic to build genesis state for the sequencer #2371

implement logic to build genesis state for the sequencer #2371

rianhughes commented Jan 13, 2025 •

edited

Loading

weiihann Jan 13, 2025

rianhughes Jan 14, 2025

rianhughes Jan 14, 2025

rianhughes Jan 14, 2025

codecov bot commented Jan 13, 2025 •

edited

Loading

AnkushinDaniil Jan 15, 2025

AnkushinDaniil Jan 15, 2025

AnkushinDaniil Jan 15, 2025

AnkushinDaniil left a comment

AnkushinDaniil Jan 16, 2025

AnkushinDaniil Jan 16, 2025

AnkushinDaniil Jan 16, 2025

AnkushinDaniil Jan 16, 2025 •

edited

Loading

AnkushinDaniil Jan 16, 2025

AnkushinDaniil Jan 16, 2025

AnkushinDaniil Jan 16, 2025

AnkushinDaniil Jan 16, 2025

AnkushinDaniil Jan 16, 2025

		newNonce := new(felt.Felt).SetUint64(1)
		p.stateDiff.Nonces[*contractAddress] = newNonce.Add(currentNonce, newNonce)

	newNonce := new(felt.Felt).SetUint64(1)
	p.stateDiff.Nonces[*contractAddress] = newNonce.Add(currentNonce, newNonce)
	p.stateDiff.Nonces[*contractAddress] = currentNonce.Add(currentNonce, &felt.One)

	newStorageDiffs := make(map[felt.Felt]map[felt.Felt]*felt.Felt)
	newStorageDiffs := make(map[felt.Felt]map[felt.Felt]*felt.Felt, len(stateDiff.StorageDiffs))

	validate := validator.Validator()
	return validate.Struct(g)
	return validator.Validator().Struct(g)

	panic("transaction type not supported")
	return nil, nil, fmt.Errorf("unsupported transaction type: %v", txn.Type)

		// StateDiffAndClasses returns the pending state's internal data. The returned objects will continue to be
		// read and modified by the pending state.

implement logic to build genesis state for the sequencer #2371

Are you sure you want to change the base?

implement logic to build genesis state for the sequencer #2371

Conversation

rianhughes commented Jan 13, 2025 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

codecov bot commented Jan 13, 2025 • edited Loading

Codecov Report

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

AnkushinDaniil left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

AnkushinDaniil Jan 16, 2025 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

rianhughes commented Jan 13, 2025 •

edited

Loading

codecov bot commented Jan 13, 2025 •

edited

Loading

AnkushinDaniil Jan 16, 2025 •

edited

Loading