fix: tests for graphql #788

howjmay · 2026-01-19T17:59:22Z

No description provided.

howjmay · 2026-01-30T09:51:23Z

packages/testutil/peering_net_behaviour_test.go

 		avgDuration /= int64(len(durations))
 		require.Greater(t, avgDuration, int64(50))
-		require.Less(t, avgDuration, int64(100))
+		require.Less(t, avgDuration, int64(120))


it takes longer time now

howjmay · 2026-02-01T19:30:14Z

.github/workflows/build-test.yml

-        run: make test-short
+        run: |
+          TEST_SHORT_PKGS="$(go list ./... \
+            | grep -v '/packages/vm/core/testcore' \


All these triggered timeout

Let's not forget to reenable or fix them. These are important tests.

#790 is solving the timeout issues too, so we will enable soon in the future

lmoe · 2026-02-05T10:25:53Z

packages/evm/jsonrpc/jsonrpctest/jsonrpc_test.go

 	t.Run("callTracer", func(t *testing.T) {
 		_, err := env.traceTransactionWithCallTracer(tx.Hash())
-		require.ErrorContains(t, err, "expected exactly one top-level call")
+		require.ErrorContains(t, err, "transaction not found")


From what I remember, expected exactly one top-level call has been an annoying issue with the calltracer. Why is it suddenly "transaction not found"?

It seems like the transaction wasn't executed, therefore you get transaction not found.

+1

Also, the change in graphql should not affect the EVM tests? Unless I'm missing something.

fixed this. It was caused by the amount coin returned by faucet is different now. I have added a test to ensure the const we used match the real amount

dessaya · 2026-02-05T14:00:27Z

clients/iotagraphql/graphql_client.go

+	current *iotago.ObjectRef,
+	timeout time.Duration,
+) (*iotago.ObjectRef, error) {
+	ticker := time.NewTicker(200 * time.Millisecond)


what is the ticker for?

I use it as a retry here, because the time for the node to index the TXs is too slow now. Or I should remove it?

Ah ok... is polling like this the "official" way? I mean, is there something like a subscription API to be notified of the updated object ref? Just trying to avoid hardcoded delays and timeouts.

We are trying to implement the subscription of graphql. However, I think UpdateObjectRef() is more like a one-off call. If we don't do retry here, we add to add it to all the tests. Not sure what is the best way to do it

I don't know, we should try to avoid hardcoded delays if possible. I would think there must be a proper way to subscribe to updates, but I'm not familiarized with the graphql api.

Anyway, if the retry delay is unavoidable, then at least maybe it should be a parameter or a const?

I added the tickingTime in GraphQLClient

dessaya · 2026-02-05T14:01:12Z

.github/workflows/build-test.yml

-        run: make test-short
+        run: |
+          TEST_SHORT_PKGS="$(go list ./... \
+            | grep -v '/packages/vm/core/testcore' \


Let's not forget to reenable or fix them. These are important tests.

dessaya · 2026-02-05T14:02:38Z

clients/iotagraphql/graphql_client.go

+func (c *GraphQLClient) waitForNewerObjectRef(
+	ctx context.Context,
+	current *iotago.ObjectRef,
+	timeout time.Duration,


If we already receive a context, then maybe the timeout is redundant, because the caller could pass a ContextWithTimeout

dessaya · 2026-02-05T14:07:09Z

clients/iotagraphql/graphql_client.go

+	var txnResponse *IotaTransactionBlockResponse
+	var gasPayments []*iotago.ObjectRef
+
+	for i := 0; i < 5; i++ {


maybe the amount of retries and the delay should be parameters, or consts

dessaya · 2026-02-05T14:11:17Z

packages/evm/jsonrpc/jsonrpctest/jsonrpc_test.go

 	t.Run("callTracer", func(t *testing.T) {
 		_, err := env.traceTransactionWithCallTracer(tx.Hash())
-		require.ErrorContains(t, err, "expected exactly one top-level call")
+		require.ErrorContains(t, err, "transaction not found")


+1

Also, the change in graphql should not affect the EVM tests? Unless I'm missing something.

dessaya · 2026-02-05T14:13:38Z

packages/solo/solo.go

+	ctx, cancel := context.WithTimeout(env.ctx, 60*time.Second)
+	defer cancel()
+
+	ticker := time.NewTicker(250 * time.Millisecond)


Is this "permanent"? Time delays make tests unhappy.

I think we need it until we discard grahpql in solo

dessaya · 2026-02-06T12:54:22Z

clients/iotagraphql/graphql_client.go

+
+func getBalance(ctx context.Context, client *GraphQLClient, address *iotago.Address) *big.Int {
+	balance, err := client.GetBalance(ctx, GetBalanceRequest{Owner: address})
+	if err != nil || balance.TotalBalance == nil {


if err != nil then maybe this should panic or return error

dessaya · 2026-02-06T13:15:08Z

clients/iotagraphql/graphql_client.go

+	current *iotago.ObjectRef,
+	timeout time.Duration,
+) (*iotago.ObjectRef, error) {
+	ticker := time.NewTicker(200 * time.Millisecond)


I don't know, we should try to avoid hardcoded delays if possible. I would think there must be a proper way to subscribe to updates, but I'm not familiarized with the graphql api.

Anyway, if the retry delay is unavoidable, then at least maybe it should be a parameter or a const?

dessaya · 2026-02-06T13:23:43Z

clients/iotagraphql/graphql_client.go

+	var txnResponse *IotaTransactionBlockResponse
+	var gasPayments []*iotago.ObjectRef
+
+	for i := 0; i < maxRetries; i++ {


Having to retry operations is suspicious... is the graphql api unreliable? Is it expected to have to retry? Perhaps we are using the api wrong somehow?

Maybe this is fine as a temporary measure but IMO we should at least leave TODOs to remove all delays and retries at some point.

The problem was caused by the indexer performs really slow in the test scenario. Once we switch to Simulacrum, then the problem may be gone

howjmay force-pushed the ql-all-3 branch 13 times, most recently from b13165e to 29b6891 Compare January 26, 2026 19:43

howjmay force-pushed the ql-all-3 branch 2 times, most recently from e4b4a6c to 0630144 Compare January 30, 2026 09:30

howjmay commented Jan 30, 2026

View reviewed changes

howjmay changed the title ~~Ql all 3~~ fix: tests for graphql Jan 30, 2026

howjmay force-pushed the ql-all-3 branch 2 times, most recently from 7be2017 to 134ed18 Compare January 30, 2026 17:27

howjmay marked this pull request as ready for review February 1, 2026 14:01

howjmay force-pushed the ql-all-3 branch from 134ed18 to 5b6acdb Compare February 1, 2026 14:52

fix: tests for graphql

6a1b7f4

howjmay force-pushed the ql-all-3 branch from 7b598d7 to 6a1b7f4 Compare February 1, 2026 19:29

howjmay commented Feb 1, 2026

View reviewed changes

lmoe reviewed Feb 5, 2026

View reviewed changes

dessaya reviewed Feb 5, 2026

View reviewed changes

howjmay force-pushed the ql-all-3 branch 2 times, most recently from 0a3fa86 to c48f7c8 Compare February 5, 2026 22:08

howjmay requested review from dessaya and lmoe February 6, 2026 12:15

dessaya approved these changes Feb 6, 2026

View reviewed changes

howjmay force-pushed the ql-all-3 branch from c48f7c8 to 893bbfa Compare February 6, 2026 16:26

fix: Update SingleCoinFundsFromFaucetAmount value

3bcdb92

howjmay force-pushed the ql-all-3 branch from 893bbfa to 3bcdb92 Compare February 7, 2026 20:03

howjmay merged commit c225b9f into iotaledger:ql-dev Feb 8, 2026
3 checks passed

howjmay deleted the ql-all-3 branch February 8, 2026 12:00

fix: tests for graphql #788

fix: tests for graphql #788

Conversation

howjmay commented Jan 19, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

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!

howjmay Feb 5, 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!

howjmay Feb 5, 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!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

howjmay commented Jan 19, 2026 •

edited

Loading

howjmay Feb 5, 2026 •

edited

Loading

howjmay Feb 5, 2026 •

edited

Loading