ethclient/simulated: add timeout to fix flaky test

[holiman]

We have a flaky test:

--- FAIL: TestForkResendTx (0.02s)
    backend_test.go:234: sending transaction: nonce too low: next nonce 1, tx nonce 0

The reason the test fails is that after rolling back a block (reorging out a transaction), the head-event which should trigger the txpool to reset the internal state is not handled fast enough, and when we try to re-submit the tx, the pool uses the old data, saying the nonce should be 1.

This is a crappy fix, but I don't see how to properly make sure the head-event was processed by subpools, when all we have is a backend and an ethclient.

[lightclient]

Could you look at the pending nonce? Something like this

diff --git a/ethclient/simulated/backend_test.go b/ethclient/simulated/backend_test.go
index 0237c1a0d..c27fbf39e 100644
--- a/ethclient/simulated/backend_test.go
+++ b/ethclient/simulated/backend_test.go
@@ -230,9 +230,23 @@ func TestForkResendTx(t *testing.T) {
 
 	// 5.
 	sim.Commit()
+
 	// The new-head-event needs to trickle out to the tx pool(s), so they reset
 	// their internal state.
-	time.Sleep(1 * time.Second)
+	timeout := time.Now().Add(time.Second)
+	for {
+		if time.Now().After(timeout) {
+			t.Fatalf("timeout while waiting for subpools to reorg")
+		}
+		nonce, err := client.PendingNonceAt(ctx, testAddr)
+		if err != nil {
+			t.Fatalf("failed to check nonce: %v", err)
+		}
+		if nonce == tx.Nonce() {
+			break
+		}
+	}
+
 	if err := client.SendTransaction(ctx, tx); err != nil {
 		t.Fatalf("sending transaction: %v", err)
 	}

[karalabe]

Lemme try to fix this in a different way without the sleep

[holiman]

Lemme try to fix this in a different way without the sleep

This does seem fixed now, with #28837. Closing