Event subtype implementation by andreanistico · Pull Request #122 · HorizenOfficial/vela

andreanistico · 2026-03-18T11:41:49Z

https://horizenlabs.atlassian.net/browse/HZN-2797

…ficial/horizen-pes into HZN-2797/event-subtype

paolocappelletti · 2026-03-24T13:10:49Z

pkg/common/appdata/types.go

 func NewAppData(initialAppState []byte) *AppData {
 	return &AppData{
-		version:  Version_1,
+		version:  Version_2,


we can keep Version_1 until we will be on mainnet (we don't have to deal with backward compatibility until then)

paolocappelletti · 2026-03-24T13:13:27Z

pkg/common/appdata/types.go


 // Serialize converts the AppData struct into a byte slice.
-// The format is:
+// Version 2 format:


same as above

paolocappelletti · 2026-03-24T14:57:27Z

pkg/executor/executor.go


 	if req.RequestType == common.AssociateKey {
-		//request  of type associate key: the payload is not encrypted and contains the new key
+		//request  of type associate key: the payload is not encrypted and contains the new key + seed


I would keep the seed optional (so it's up to the app decide if the seed is used or not) => in the event generation the "seed unspecificed" usecase is already covered but here it is not.
The associateKeyPayloadSize could be 133 or 198 length => if 133 no seed is passed and we just store the key as before

fixed! Since the contract is changed, i ask @saratnt help for bindings regenerations

paolocappelletti · 2026-03-24T14:58:12Z

pkg/common/appdata/types.go

 	version  uint8
 	appNonce uint64
 	appKeys  KeyStore
+	appSeeds SeedStore


we could maybe rename: appEventSeeds for clarity

saratnt · 2026-03-24T15:25:37Z

pkg/executor/executor.go


 	if req.RequestType == common.AssociateKey {
-		//request  of type associate key: the payload is not encrypted and contains the new key
+		//request  of type associate key: the payload is not encrypted and contains the new key + seed


Shouldn't the seed be encrypted? Otherwise anybody could generate all the possible EventSubType related to a single user

correct! fixed

oh yes, we can encrypt it with the key just generated (the key part must must be in clear)

saratnt · 2026-03-24T15:43:41Z

pkg/executor/executor.go

+		if seed, hasSeed := seedStore[event.UserID]; hasSeed {
+			privSubtype, err := GenerateRandomSubtype(seed, DefaultSubtypeN)
+			if err != nil {
+				return nil, apperrors.New(apperrors.CodeRequestFuncFailed, "failed to generate random subtype")


This error shouldn't mark the request as failed, but it should be a transient one.
In addition, it is impossible that there is no seed for the user but there is the pub key, because the AssociateKey check that the payload is 198 bytes, this means the EventSubType field of the wasm event is never used and it can be removed

saratnt · 2026-03-24T16:05:51Z

pkg/executor/subtype.go

+// and returns GenerateSubtype(seed, index).
+func GenerateRandomSubtype(seed []byte, n int) (string, error) {
+	randVal, err := rand.Int(rand.Reader, big.NewInt(int64(n)))
+	if err != nil {


If I have understood correctly, there are at most n different event subtypes for each user and only these n can be used.
If this is correct, it is possible that all the subtypes can be discovered in the long run for a user who performs enough requests. In fact, every time a user sends a request that generates just 1 event, this is a possibility to retrieve 1 of these subtypes. Eg when a user requests just a Deposit, there is just 1 event that we know belongs to the user itself, so we can get the subtype.
What about using the requestId instead of the random number, eg hash(seed+requestId)? The seed is private, so only the user can calculate it, with the requestId it changes everytime and it is never reused

For the user (and web app) could be harder to retrieve all its used request IDs to get all the events, while it is easy to get all the numbers from 0 to N.

In the design we choose to use a random number between 0 and N as compromise between ease of use and privacy. In the long run only 1/N transaction could be bounded to the same address, so it is hard to build a story of the user operation just from 1/N for an high N (N=50 by default, but it could be increased).

But yours is also a good idea, maybe we can explore it for a next development
cc @paolocappelletti

You're right, I totally forgot that the problem is that we don't know when was the last event related to that user. The requestId doesn't make sense in this scenario.

yes, for now let'sstay simple.
another option could be changing the seed from time to time with a specific action by the user

# Conflicts: # pkg/common/appdata/types.go # pkg/common/appdata/types_test.go # pkg/executor/encrypt_events_test.go # pkg/executor/executor.go # pkg/executor/handle_process_request_test.go

# Conflicts: # contracts/contracts/ProcessorEndpoint.sol # pkg/blockchain/contracts/processorendpoint/ProcessorEndpoint.go

saratnt · 2026-03-25T11:00:48Z

pkg/common/appdata/types.go

+	}
+
+	// Write each seed key-value pair (address + 65-byte seed)
+	for addr, seed := range s.appEventSeeds {


It is not clear to me why for the appKeys map the size of the map keys (ethereum addresses) is checked but not the size of the map values (PublicKeyP521 keys), while for appEventSeeds is the opposite. If we want to be sure the data were not corrupted, both should be checked

should be fixed now

saratnt · 2026-03-25T11:03:28Z

pkg/common/appdata/types.go

 }

+func (s *AppData) AddSeed(user ethCommon.Address, seed []byte) {
+	seedCopy := make([]byte, len(seed))


Here the length of seed could be checked, to prevent a corrupted seed to be included

saratnt · 2026-03-25T11:04:59Z

pkg/common/appdata/types.go

 		version:         Version_1,
 		appNonce:        0,
 		wasmFingerprint: [WasmFingerprintSize]byte{},
 		appKeys:         make(map[ethCommon.Address]*cryptotypes.PublicKeyP521),


This should be make(KeyStore)

saratnt · 2026-03-25T11:09:27Z

pkg/common/appdata/types.go


+	// Read seed store
+	ss := make(SeedStore)
+	if version != Version_1 {


This should be moved at line 164

saratnt · 2026-03-25T11:14:27Z

pkg/common/appdata/types.go

+		return nil, fmt.Errorf("insufficient data for seed pairs: need %d, have %d", expectedSeedSize, reader.Len())
+	}
+	for i := uint32(0); i < numSeeds; i++ {
+		addrBuf := make([]byte, KeyStore_KeySize)


For avoiding confusion, it would be better or to duplicate KeyStore_KeySize and to have SeedStore_KeySize const, or to rename KeyStore_KeySize to something more generic like Store_KeySize

renamed to Store_KeySize

saratnt · 2026-03-25T11:18:05Z

pkg/executor/encrypt_events_test.go

 )

 // Ensure encryptEvents copies EventSubType and keeps UserID as recipient.
 func TestEncryptEventsCopiesSubType(t *testing.T) {


It would be nice to have another test for when the user has a seed, the EventSubType is changed

added TestEncryptEventsChangesSubTypeWhenSeedPresent

saratnt · 2026-03-25T11:21:05Z

pkg/executor/executor.go

@@ -635,6 +639,27 @@ func (e *StatelessExecutor) HandleProcessRequest(ctx context.Context, req *commo
 		totalFuel = totalFuel.Add(totalFuel, big.NewInt(10))


This line should be moved at the end, at line 663

wip

5a1705c

andreanistico changed the base branch from main to dev March 18, 2026 11:42

andreanistico and others added 5 commits March 18, 2026 15:04

implemented functionality

e6fb14f

Bindings

52e4bdb

fix test

c6bcdad

Merge branch 'HZN-2797/event-subtype' of https://github.com/HorizenOf…

9de68f0

…ficial/horizen-pes into HZN-2797/event-subtype

fix test

83a4ab6

paolocappelletti requested review from paolocappelletti and saratnt March 24, 2026 13:07

paolocappelletti requested changes Mar 24, 2026

View reviewed changes

saratnt requested changes Mar 24, 2026

View reviewed changes

andreanistico added 4 commits March 25, 2026 09:59

fix PR

1eb80ee

encrypt seed

49892d1

Merge remote-tracking branch 'origin/dev' into HZN-2797/event-subtype

387928f

# Conflicts: # pkg/common/appdata/types.go # pkg/common/appdata/types_test.go # pkg/executor/encrypt_events_test.go # pkg/executor/executor.go # pkg/executor/handle_process_request_test.go

bindings

3305ec5

andreanistico requested review from paolocappelletti and saratnt March 25, 2026 09:40

andreanistico added 4 commits March 25, 2026 10:53

fix tests

23bed64

Merge remote-tracking branch 'origin/dev' into HZN-2797/event-subtype

04773cd

# Conflicts: # contracts/contracts/ProcessorEndpoint.sol # pkg/blockchain/contracts/processorendpoint/ProcessorEndpoint.go

new bindings

d01ea93

fix test

7468f6e

saratnt requested changes Mar 25, 2026

View reviewed changes

fix PR

4ebcd9d

andreanistico requested a review from saratnt March 25, 2026 12:50

saratnt approved these changes Mar 25, 2026

View reviewed changes

paolocappelletti approved these changes Mar 25, 2026

View reviewed changes

paolocappelletti merged commit 509c933 into dev Mar 25, 2026
3 checks passed

		@@ -635,6 +639,27 @@ func (e StatelessExecutor) HandleProcessRequest(ctx context.Context, req commo
		totalFuel = totalFuel.Add(totalFuel, big.NewInt(10))

Conversation

andreanistico commented Mar 18, 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!

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!

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!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants