support procedural test style for VM-defined workloads #1667

tsachiherman · 2024-10-16T21:11:18Z

What ?

Add the infrastructure that would allow custom-VM to implement their own integration and e2e testing, outside of the workload design pattern.

Why ?

Custom networks might try to achieve different goals that could not easily be tested using the workload patten being used. The workload patten is great when we want to ensure that the custom-VM perform the hyper-SDK functionalities correctly. However, when aiming to test custom network functionalities, this testing model comes short.

How would we solve that problem ?

Each VM would defined the following:

// TestsRegistry is global tests registry instance where all the tests are to be registered. Each VM would typically have only
// a single instance of such a registry.
var TestsRegistry = &registry.Registry{}

as well as one or more tests:

var _ = registry.Register(TestsRegistry, "Transfer Transaction", func(t ginkgo.FullGinkgoTInterface, tn tworkload.TestNetwork) error {
 ...

Then, when the e2e and integration tests would get executed, a new test node would be added [Custom VM Tests] which would invoke these custom test. The network would be the same network created during the e2e/integration tests.
In order to get access to that network, the workload.TestNetwork would be used :

type TestNetwork interface {
	ConfirmTxs(context.Context, []*chain.Transaction) error
	GenerateTx(context.Context, []chain.Action, chain.AuthFactory) (*chain.Transaction, error)
	URIs() []string
	Configuration() TestNetworkConfiguration
}

The implementation of the TestNetwork depends on whether we're running e2e or integration testing.
On integration testings, the ConfirmTxs would be executed against the underlying VM directly.
On e2e testings, the the ConfirmTxs would be implemented via the JSON RPC and Indexer.

examples/morpheusvm/tests/e2e/transfer_test.go

samliok

I'm unconvinced this testing pattern adds value to VM developers. Seems like this adds a lot of additional complexity, especially if the developer still needs to implement the workloadfactory

samliok · 2024-10-18T17:10:44Z

examples/morpheusvm/tests/e2e/transfer_test.go

+
+	require.NoError(err)
+
+	timeoutCtx, timeoutCtxFnc := context.WithDeadline(context.Background(), time.Now().Add(2*time.Second))


we have a const txCheckInterval. I think it makes sense to either pass the const in here or remove the const completely and pass in a value from the test.

samliok · 2024-10-18T17:10:48Z

examples/morpheusvm/tests/e2e/transfer_test.go

+	ginkgo "github.com/onsi/ginkgo/v2"
+)
+
+var _ = he2e.RegisterTest("Transfer Transaction", func(t ginkgo.FullGinkgoTInterface, tn workload.TestNetwork) error {


i'm concerned this test doesn't add any real value to a hypersdk VM developer. If anything, it adds less value than the current workload tests because the user doesn't have fine grained controls of what a "confirmed" transaction is. In the workload tests, at least devs can define their custom confirmation logic. In these e2e tests all a user gets back is whether or not the transaction was found in the block.

examples/morpheusvm/tests/integration/transfer_test.go

samliok · 2024-10-18T17:15:56Z

tests/e2e/network.go

+)
+
+const (
+	txCheckInterval = 100 * time.Millisecond


this is the const i mentioned earlier

samliok · 2024-10-18T17:24:13Z

examples/morpheusvm/tests/e2e/transfer_test.go

+	parser, err := lcli.Parser(context.Background())
+	require.NoError(err)
+
+	spendingKey, err := tn.WorkloadFactory().GetSpendingKey()


in these custom e2e tests, we only use the workloadfactory for this one method. This means if the user wants to only create custom e2e tests they need to also define the workloadfactory interface which is quite confusing 😅

maybe its worth adding some of the further work you mentioned to this PR. If the GenesisKeys are added to the test network then we wouldn't need to couple the WorkloadFactory here

tests/e2e/e2e.go

tests/registry/registry.go

aaronbuchwald · 2024-10-25T03:07:06Z

examples/morpheusvm/tests/transfer.go

+var _ = registry.RegisterTest("Transfer Transaction", func(t ginkgo.FullGinkgoTInterface, tn tworkload.TestNetwork) error {
+	require := require.New(t)


(future PR) - it would be nice to enable tests to declare their funding dependency ie. request that you pass in N keys with balance of at least 1K

examples/morpheusvm/tests/transfer.go

tests/integration/network.go

aaronbuchwald · 2024-10-25T03:26:35Z

tests/workload/network.go

+	SubmitTxs(context.Context, []*chain.Transaction) error
+	ConfirmTxs(context.Context, []*chain.Transaction) error
+	GenerateTx(context.Context, []chain.Action, chain.AuthFactory) (*chain.Transaction, error)


Can we remove TestNode and support a single ConfirmTxs call on TestNetwork instead of the node?

We should leave it to each environment (integration/e2e) to handle this since submitting a tx to a single node in integration will result in a different outcome than submitting a tx to a single node in the e2e environment.

Each environment should enforce that marking a tx as confirmed indicates all nodes have marked it as confirmed, so that we can check the state by making API calls directly to the URIs reported by the TestNetwork.

tests/workload/network.go

tests/e2e/e2e.go

aaronbuchwald · 2024-10-28T13:42:13Z

tests/integration/integration.go

+var _ = ginkgo.Describe("[Custom VM Tests]", func() {
+	require := require.New(ginkgo.GinkgoT())
+
+	for testRegistry := range registry.GetTestsRegistries() {


Could we pass the test registry into the integration/e2e tests rather than using a global registry this way?

Otherwise, I think this is just adding an extra step to the same pattern as #1667 (comment)

Unfortunately, that won't work with ginkgo. we need the tests to be available in the GetTestRegistries ( or another global variable ) by the time this method is being invoked.
What I can do is to set this to be a function placeholder that would later on invoke the various tests. However, that would not be a ginkgo node and any failure on these tests would not be reported correctly.

tests/workload/network.go

tests/integration/network.go

aaronbuchwald · 2024-10-28T13:49:58Z

tests/integration/network.go

+func (i *instance) confirmTx(ctx context.Context, txid ids.ID) error {
+	lastAcceptedHeight, err := i.vm.GetLastAcceptedHeight()
+	if err != nil {
+		return err
+	}
+	lastAcceptedBlockID, err := i.vm.GetBlockHeightID(lastAcceptedHeight)
+	if err != nil {
+		return err
+	}
+	blk, err := i.vm.GetBlock(ctx, lastAcceptedBlockID)
+	if err != nil {
+		return err
+	}
+	for {
+		stflBlk, ok := blk.(*chain.StatefulBlock)
+		if !ok {
+			return ErrTxNotFound
+		}
+		for _, tx := range stflBlk.StatelessBlock.Txs {
+			if tx.ID() == txid {
+				// found.
+				return nil
+			}
+		}
+		// keep iterating backward.
+		lastAcceptedBlockID = blk.Parent()
+		blk, err = i.vm.GetBlock(ctx, lastAcceptedBlockID)
+		if err != nil {
+			return ErrTxNotFound
+		}
+	}
+}


Can we switch to using the indexer API here instead so we don't need a manual search?

wasn't the intent of having the integration tests was to bypass the indexer and all other rpc endpoints ?
( in contrast to the e2e tests )

I don't think there's any need to explicitly avoid rpc endpoints. The integration tests should enable us to test the same logic much faster than starting a full e2e network by running all the VMs in a single process and skipping the overhead of running a full e2e network.

tests/integration/network.go

examples/morpheusvm/tests/workload/genesis.go

tests/e2e/network.go

Co-authored-by: aaronbuchwald <aaron.buchwald56@gmail.com> Signed-off-by: Tsachi Herman <24438559+tsachiherman@users.noreply.github.com>

…to tsachi/proc-testing

aaronbuchwald · 2024-10-30T13:41:29Z

examples/morpheusvm/tests/transfer.go

+// TestsRegistry is global tests registry instance where all the tests are to be registered. Each VM would typically have only
+// a single instance of such a registry.


Suggested change

// TestsRegistry is global tests registry instance where all the tests are to be registered. Each VM would typically have only

// a single instance of such a registry.

// Global test registry initialized during init to ensure tests are identical during ginkgo

// tree construction and test execution

// ref https://onsi.github.io/ginkgo/#mental-model-how-ginkgo-traverses-the-spec-hierarchy

aaronbuchwald · 2024-10-30T13:42:01Z

examples/morpheusvm/tests/transfer.go

+// a single instance of such a registry.
+var TestsRegistry = &registry.Registry{}
+
+var _ = registry.Register(TestsRegistry, "Transfer Transaction", func(t ginkgo.FullGinkgoTInterface, tn tworkload.TestNetwork) error {


Since we are using require internally can we remove the returned error?

aaronbuchwald · 2024-10-30T13:44:15Z

tests/e2e/e2e.go

+var _ = ginkgo.Describe("[Custom VM Tests]", ginkgo.Serial, func() {
+	tc := e2e.NewTestContext()
+	require := require.New(tc)
+
+	for testRegistry := range registry.GetTestsRegistries() {
+		for _, test := range testRegistry.List() {
+			ginkgo.It(test.Name, func() {
+				testNetwork := NewNetwork(tc)
+				require.NoError(test.Fnc(ginkgo.GinkgoT(), testNetwork), "Test %s failed with an error", test.Name)
+			})
+		}
+	}
+})


Hmm so I think the only requirement is that we populate it via a global/init function rather than in BeforeSuite or SynchronizedBeforeSuite as we do now.

Could we get rid of the nested for loop here in favor of a single registry of tests and remove the map[*Registry]bool type? This should satisfy registering tests as globals for ginkgo and remove the extra loop.

aaronbuchwald · 2024-10-30T13:58:46Z

tests/e2e/network.go

+)
+
+type Network struct {
+	nodes  []*Node


Can we remove the node type completely? It seems that it just uses the URI and the reference to the network's parser to generate a tx, which is not specific to the node. I think we could just get rid of this construct completely and only have the network type.

aaronbuchwald · 2024-10-30T14:06:42Z

tests/integration/network.go

+func (i *instance) confirmTx(ctx context.Context, txid ids.ID) error {
+	lastAcceptedHeight, err := i.vm.GetLastAcceptedHeight()
+	if err != nil {
+		return err
+	}
+	lastAcceptedBlockID, err := i.vm.GetBlockHeightID(lastAcceptedHeight)
+	if err != nil {
+		return err
+	}
+	blk, err := i.vm.GetBlock(ctx, lastAcceptedBlockID)
+	if err != nil {
+		return err
+	}
+	for {
+		stflBlk, ok := blk.(*chain.StatefulBlock)
+		if !ok {
+			return ErrTxNotFound
+		}
+		for _, tx := range stflBlk.StatelessBlock.Txs {
+			if tx.ID() == txid {
+				// found.
+				return nil
+			}
+		}
+		// keep iterating backward.
+		lastAcceptedBlockID = blk.Parent()
+		blk, err = i.vm.GetBlock(ctx, lastAcceptedBlockID)
+		if err != nil {
+			return ErrTxNotFound
+		}
+	}
+}


I don't think there's any need to explicitly avoid rpc endpoints. The integration tests should enable us to test the same logic much faster than starting a full e2e network by running all the VMs in a single process and skipping the overhead of running a full e2e network.

aaronbuchwald · 2024-10-30T18:28:34Z

tests/e2e/network.go

+		testNetwork.nodes = append(testNetwork.nodes, n)
+	}
+	testNetwork.parser = &parser{
+		Parser: networkConfig.Parser(),


Can we add a comment on the parser type explaining why we need to wrap the chain.Parser type that is already passed in? As I understand it, this is because we cannot currently know the blockchainID in advance since it's the hash of the transaction on the P-Chain that will create the chain.

yes. I'll add some comments around that.

aaronbuchwald · 2024-10-30T18:31:31Z

tests/e2e/network.go

+	return networkConfig
+}
+
+type Node struct {


Can we remove the Node type? This seems to only be used to keep a list of uris and then a pass through back to the network, so I think we can just keep the URIs on the network and move GenerateTx to be a function directly on the network.

aaronbuchwald · 2024-10-30T18:32:59Z

tests/e2e/network.go

+func (n *Network) ConfirmTxs(ctx context.Context, txs []*chain.Transaction) error {
+	err := n.nodes[0].confirmTxs(ctx, txs)
+	if err != nil {
+		return err
+	}
+	var targetHeight uint64
+	// check the accepted block height on all blocks.
+	for nodeIdx := 0; nodeIdx < len(n.nodes); nodeIdx++ {
+		_, nodeHeight, _, err := n.nodes[nodeIdx].accepted(ctx)
+		if err != nil {
+			return err
+		}
+		if nodeIdx == 0 {
+			// since we've already confirmed the tx on this node, just use the height as the target.
+			targetHeight = nodeHeight
+		} else if nodeHeight < targetHeight {
+			// take a short delay and try again.
+			select {
+			case <-ctx.Done():
+				return ctx.Err()
+			case <-time.After(txCheckInterval):
+			}
+			nodeIdx-- // try again the same node.
+			continue
+		}
+	}
+	return nil
+}


Suggested change

func (n *Network) ConfirmTxs(ctx context.Context, txs []*chain.Transaction) error {

err := n.nodes[0].confirmTxs(ctx, txs)

if err != nil {

return err

}

var targetHeight uint64

// check the accepted block height on all blocks.

for nodeIdx := 0; nodeIdx < len(n.nodes); nodeIdx++ {

_, nodeHeight, _, err := n.nodes[nodeIdx].accepted(ctx)

if err != nil {

return err

}

if nodeIdx == 0 {

// since we've already confirmed the tx on this node, just use the height as the target.

targetHeight = nodeHeight

} else if nodeHeight < targetHeight {

// take a short delay and try again.

select {

case <-ctx.Done():

return ctx.Err()

case <-time.After(txCheckInterval):

}

nodeIdx-- // try again the same node.

continue

}

}

return nil

}

func (n *Network) ConfirmTxs(ctx context.Context, txs []*chain.Transaction) error {

c := jsonrpc.NewJSONRPCClient(n.uris[0])

for _, tx := range txs {

_, err := c.SubmitTx(ctx, tx.Bytes())

if err != nil {

return err

}

}

indexerCli := indexer.NewClient(n.uris[0])

for _, tx := range txs {

success, _, err := indexerCli.WaitForTransaction(ctx, txCheckInterval, tx.ID())

if err != nil {

return err

}

if !success {

return ErrUnableToConfirmTx

}

}

_, targetHeight, _, err := c.Accepted(ctx)

if err != nil {

return err

}

for _, uri := range n.uris[1:] {

if err := jsonrpc.Wait(ctx, txCheckInterval, func(ctx context.Context) (bool, error) {

c := jsonrpc.NewJSONRPCClient(uri)

_, nodeHeight, _, err := c.Accepted(ctx)

if err != nil {

return false, err

}

return nodeHeight >= targetHeight, nil

}); err != nil {

return err

}

}

return nil

}

aaronbuchwald · 2024-10-30T18:38:40Z

tests/registry/registry.go

+// we need to pre-register all the test registries that are created externally in order to comply with the ginko execution order.
+// i.e. the global `var _ = ginkgo.Describe` used in the integration/e2e tests need to have this field populated before the iteration
+// over the top level nodes.


Can we avoid this by setting this in a global variable / setting via init function so that it's identical during test construction and execution?

I think that it is identical during test construction and execution.

Not sure how to implement that so that it would work ginkgo otherwise, as ginkgo is managing these in globals as well ( but keep them internal )

aaronbuchwald · 2024-10-30T18:54:31Z

tests/integration/network.go

+	for i := 0; i < len(instances); i++ {
+		if i == biggestHeightInstanceIndex {
+			continue
+		}
+		vm := instances[i].vm
+		for {
+			height, err := vm.GetLastAcceptedHeight()
+			if err != nil {
+				return err
+			}
+			if height == biggestHeight {
+				break
+			}
+			statefulBlock, err := instances[biggestHeightInstanceIndex].vm.GetDiskBlock(ctx, height+1)
+			if err != nil {
+				return err
+			}
+			err = vm.SetPreference(ctx, statefulBlock.ID())
+			if err != nil {
+				return err
+			}
+			blk, err := vm.ParseBlock(ctx, statefulBlock.Bytes())
+			if err != nil {
+				return err
+			}
+			err = blk.Verify(ctx)
+			if err != nil {
+				return err
+			}
+			err = blk.Accept(ctx)
+			if err != nil {
+				return err
+			}
+			if instances[i].onAccept != nil {
+				instances[i].onAccept(blk)
+			}
+		}
+	}
+	return nil


Suggested change

for i := 0; i < len(instances); i++ {

if i == biggestHeightInstanceIndex {

continue

}

vm := instances[i].vm

for {

height, err := vm.GetLastAcceptedHeight()

if err != nil {

return err

}

if height == biggestHeight {

break

}

statefulBlock, err := instances[biggestHeightInstanceIndex].vm.GetDiskBlock(ctx, height+1)

if err != nil {

return err

}

err = vm.SetPreference(ctx, statefulBlock.ID())

if err != nil {

return err

}

blk, err := vm.ParseBlock(ctx, statefulBlock.Bytes())

if err != nil {

return err

}

err = blk.Verify(ctx)

if err != nil {

return err

}

err = blk.Accept(ctx)

if err != nil {

return err

}

if instances[i].onAccept != nil {

instances[i].onAccept(blk)

}

}

}

return nil

for i := 0; i < len(instances); i++ {

if i == biggestHeightInstanceIndex {

continue

}

instance := instances[i]

for {

height, err := instance.vm.GetLastAcceptedHeight()

if err != nil {

return err

}

if height == biggestHeight {

break

}

statefulBlock, err := instances[biggestHeightInstanceIndex].vm.GetDiskBlock(ctx, height+1)

if err != nil {

return err

}

if err := instance.applyBlk(ctx, statefulBlock); err != nil {

return err

}

}

}

…to tsachi/proc-testing

commit b2ad4d3 Author: Tsachi Herman <24438559+tsachiherman@users.noreply.github.com> Date: Sun Nov 3 16:01:37 2024 -0500 support procedural test style for VM-defined workloads (#1667) Signed-off-by: Tsachi Herman <24438559+tsachiherman@users.noreply.github.com> commit dbcc6c9 Author: Richard Pringle <richard.pringle@avalabs.org> Date: Thu Oct 31 16:42:44 2024 -0400 Update wasmtime-go (#1705) commit 1015ae1 Author: Sam Liokumovich <65994425+samliok@users.noreply.github.com> Date: Thu Oct 31 12:47:45 2024 -0400 Remove HRP + Decimals dead code (#1703) commit 24bacf8 Author: Sam Liokumovich <65994425+samliok@users.noreply.github.com> Date: Wed Oct 30 10:31:38 2024 -0400 Action Test Reusing Store (#1700) commit 1ce6ad8 Author: Richard Pringle <richard.pringle@avalabs.org> Date: Tue Oct 29 13:23:56 2024 -0400 Don't benchmark serializing parameters (#1697) commit 1c43f7e Author: Sam Liokumovich <65994425+samliok@users.noreply.github.com> Date: Tue Oct 29 13:16:34 2024 -0400 Call Contract Benchmarks (#1696) commit 7bd1a49 Author: nathan haim <haim.nathan@icloud.com> Date: Sat Oct 26 20:41:26 2024 +0200 state: Make state.Keys JSON human readable (#1661) Signed-off-by: nathan haim <nathan.haim@free.fr> commit c592a1d Author: rodrigo <77309055+RodrigoVillar@users.noreply.github.com> Date: Fri Oct 25 14:02:08 2024 -0400 Add `balance` cmd to CLI (#1691) commit 7fac7fa Author: Sam Liokumovich <65994425+samliok@users.noreply.github.com> Date: Fri Oct 25 13:28:52 2024 -0400 Default Contract Manager Implementation (#1687) commit 0459bd4 Author: rodrigo <77309055+RodrigoVillar@users.noreply.github.com> Date: Fri Oct 25 12:23:04 2024 -0400 Add Getter to `BalanceHandler` (#1685) commit 960b790 Author: containerman17 <8990432+containerman17@users.noreply.github.com> Date: Sat Oct 26 00:14:29 2024 +0900 Universal CLI (#1662) Creates a single binary that can be used with any HyperSDK VM, as long as the VM uses standard functionality. Signed-off-by: containerman17 <8990432+containerman17@users.noreply.github.com> Co-authored-by: Copilot Autofix powered by AI <62310815+github-advanced-security[bot]@users.noreply.github.com> Co-authored-by: Aaron Buchwald <aaron.buchwald56@gmail.com> commit ef66271 Author: Richard Pringle <richard.pringle@avalabs.org> Date: Thu Oct 24 15:45:17 2024 -0400 Move vmwithcontracts out of examples (#1676) commit 3895ff5 Author: Richard Pringle <richard.pringle@avalabs.org> Date: Thu Oct 24 14:59:57 2024 -0400 Implement NFT example (#1684) commit 82dc621 Author: rodrigo <77309055+RodrigoVillar@users.noreply.github.com> Date: Thu Oct 24 09:54:31 2024 -0400 Add back `actionID` to `StateKeys()` (#1683) commit cf6d48a Author: Richard Pringle <richard.pringle@avalabs.org> Date: Wed Oct 23 15:11:25 2024 -0400 Force test-context to be initialized with an actor (#1682) commit 1a3c670 Author: Sam Liokumovich <65994425+samliok@users.noreply.github.com> Date: Wed Oct 23 14:26:09 2024 -0400 Refactor Workload (#1668) Signed-off-by: Sam Liokumovich <65994425+samliok@users.noreply.github.com> Co-authored-by: aaronbuchwald <aaron.buchwald56@gmail.com> commit 403bfb3 Author: rodrigo <77309055+RodrigoVillar@users.noreply.github.com> Date: Tue Oct 22 21:56:25 2024 -0400 Decouple State Manager (#1658) commit 61993f8 Author: Richard Pringle <richard.pringle@avalabs.org> Date: Mon Oct 21 14:25:25 2024 -0400 Revert "Make Morpheus build script take optional path (#1671)" (#1673) commit 87cf88a Author: aaronbuchwald <aaron.buchwald56@gmail.com> Date: Mon Oct 21 12:39:13 2024 -0400 Update min go version to go1.22.8 (#1672) commit 22241f5 Author: Richard Pringle <richard.pringle@avalabs.org> Date: Fri Oct 18 14:18:53 2024 -0400 Make Morpheus build script take optional path (#1671) commit 0658976 Author: Richard Pringle <richard.pringle@avalabs.org> Date: Wed Oct 16 18:37:50 2024 -0400 Move runtime tests into contracts-ci (#1666) commit e3d911a Author: Sam Liokumovich <65994425+samliok@users.noreply.github.com> Date: Sun Oct 13 17:45:32 2024 -0400 Remove global variables from spam script (#1659) commit bca7739 Author: Tsachi Herman <24438559+tsachiherman@users.noreply.github.com> Date: Fri Oct 11 11:41:27 2024 -0400 Create SignedTransaction out of the existing Transaction (#1640) commit 6f15b2f Author: Sam Liokumovich <65994425+samliok@users.noreply.github.com> Date: Thu Oct 10 20:47:49 2024 -0400 Include Spam Config Defaults (#1657) commit b3713a1 Author: Sam Liokumovich <65994425+samliok@users.noreply.github.com> Date: Thu Oct 10 20:16:23 2024 -0400 Separate Spam Broadcast Functionality (#1653) commit 182e023 Author: Sam Liokumovich <65994425+samliok@users.noreply.github.com> Date: Thu Oct 10 18:19:44 2024 -0400 Decouple Spam Functionality from CLI into Throughput Package (#1636) commit b8104ac Author: aaronbuchwald <aaron.buchwald56@gmail.com> Date: Thu Oct 10 14:44:08 2024 -0400 Fix incorrect naming of stateless block (#1652) commit 725a589 Author: aaronbuchwald <aaron.buchwald56@gmail.com> Date: Thu Oct 10 10:12:59 2024 -0400 Use CreateActionID to give simulated actions unique actionID (#1644) commit 0a2d991 Author: aaronbuchwald <aaron.buchwald56@gmail.com> Date: Thu Oct 10 09:57:25 2024 -0400 Update ActionRegistry to ActionCodec (#1650) commit 829b991 Author: containerman17 <8990432+containerman17@users.noreply.github.com> Date: Thu Oct 10 22:42:19 2024 +0900 Reflect Marshaler (#1592) Signed-off-by: containerman17 <8990432+containerman17@users.noreply.github.com> commit b3c991f Author: Tsachi Herman <24438559+tsachiherman@users.noreply.github.com> Date: Thu Oct 10 09:36:21 2024 -0400 add simulateTransaction endpoint (#1610) Signed-off-by: Tsachi Herman <24438559+tsachiherman@users.noreply.github.com> commit cbb6a33 Author: containerman17 <8990432+containerman17@users.noreply.github.com> Date: Thu Oct 10 22:24:22 2024 +0900 Add boolean support to the ABI spec (#1648) Signed-off-by: containerman17 <8990432+containerman17@users.noreply.github.com> commit 5d8c590 Author: Sam Liokumovich <65994425+samliok@users.noreply.github.com> Date: Thu Oct 10 09:21:46 2024 -0400 Remove marshaling from transfer action (#1647) commit b934dbd Author: Joshua Kim <20001595+joshua-kim@users.noreply.github.com> Date: Wed Oct 9 16:52:29 2024 -0400 Remove type alias for Action + Auth + Output Registry (#1547) Signed-off-by: Joshua Kim <20001595+joshua-kim@users.noreply.github.com> commit b5407c9 Author: containerman17 <8990432+containerman17@users.noreply.github.com> Date: Thu Oct 10 00:45:38 2024 +0900 Simulate a chain of actions (#1635) commit fee360f Author: Tsachi Herman <24438559+tsachiherman@users.noreply.github.com> Date: Fri Oct 4 11:23:58 2024 -0400 refactor marshalActions implementation (#1631) commit 2cb5530 Author: aaronbuchwald <aaron.buchwald56@gmail.com> Date: Thu Oct 3 18:49:12 2024 -0400 Add block indexing to Indexer API (#1606) commit f1bcc59 Author: aaronbuchwald <aaron.buchwald56@gmail.com> Date: Thu Oct 3 18:31:39 2024 -0400 Add JSON marshalling to Result (#1627) commit 2ea784a Author: Richard Pringle <richard.pringle@avalabs.org> Date: Thu Oct 3 12:54:10 2024 -0400 Implement multisig example (#1581) commit 6bbf236 Author: Richard Pringle <richard.pringle@avalabs.org> Date: Thu Oct 3 12:43:41 2024 -0400 Update codeowners (#1629) commit 47849d8 Author: aaronbuchwald <aaron.buchwald56@gmail.com> Date: Thu Oct 3 08:34:51 2024 -0400 Re-mark pubsub tests as flaky (#1625) commit d847132 Author: aaronbuchwald <aaron.buchwald56@gmail.com> Date: Wed Oct 2 14:23:24 2024 -0400 Reduce tests.unit.sh timeout value less than CI job (#1623) commit 404e74f Author: aaronbuchwald <aaron.buchwald56@gmail.com> Date: Wed Oct 2 14:07:04 2024 -0400 Add JSON marshalling for fee dimensions (#1622) commit 7e53003 Author: Joshua Kim <20001595+joshua-kim@users.noreply.github.com> Date: Wed Oct 2 12:38:34 2024 -0400 Replace usage of `internal/network` with `p2p` (#1613) Signed-off-by: Joshua Kim <20001595+joshua-kim@users.noreply.github.com> commit 79f363e Author: containerman17 <8990432+containerman17@users.noreply.github.com> Date: Wed Oct 2 11:53:59 2024 +0900 Standardize decimals (#1620) commit e191c88 Author: Joshua Kim <20001595+joshua-kim@users.noreply.github.com> Date: Tue Oct 1 20:02:30 2024 -0400 Update to avalanchego@v1.11.12-rc.2 (#1614) Signed-off-by: Joshua Kim <20001595+joshua-kim@users.noreply.github.com> commit 59c84d7 Author: aaronbuchwald <aaron.buchwald56@gmail.com> Date: Tue Oct 1 17:19:17 2024 -0400 Add ExecutedBlock type (#1601) commit 165a455 Author: Tsachi Herman <24438559+tsachiherman@users.noreply.github.com> Date: Tue Oct 1 09:26:32 2024 -0400 validate buffer length prior to calling Uint64 (#1588) commit c24f0d8 Author: rodrigo <77309055+RodrigoVillar@users.noreply.github.com> Date: Tue Oct 1 09:25:18 2024 -0400 Fix VM-With-Contracts Unit Tests (#1602) commit 8ab9d83 Author: aaronbuchwald <aaron.buchwald56@gmail.com> Date: Mon Sep 30 16:57:29 2024 -0400 Minor nits on pubsub implementation (#1593) commit 592cd7a Author: rodrigo <77309055+RodrigoVillar@users.noreply.github.com> Date: Mon Sep 30 12:46:35 2024 -0400 Remove `StateKeysMaxChunks()` (#1607) commit 3e358c1 Author: aaronbuchwald <aaron.buchwald56@gmail.com> Date: Mon Sep 30 12:46:20 2024 -0400 Update ActionBenchmark to use single ExpectedOutput (#1608) commit e78841e Author: Franfran <51274081+iFrostizz@users.noreply.github.com> Date: Mon Sep 30 16:42:00 2024 +0200 Write FIFO cache unit tests (#1201) * write FIFO cache unit tests * invert expected / actual * add limit parameter and cache fail on 0 size * simplify test table and create new test for empty cache * simpler tests by having a limit of 2 * fix test cases * typos * backout uselesss change * add a "not an LRU" test case * use switch in fifo unit tests commit f1150a1 Author: aaronbuchwald <aaron.buchwald56@gmail.com> Date: Fri Sep 27 17:46:28 2024 -0400 Update tx indexer to include tx action outputs (#1597) commit 6a3fd63 Author: Joshua Kim <20001595+joshua-kim@users.noreply.github.com> Date: Thu Sep 26 15:44:56 2024 -0400 Remove usage of mockgen (#1591) Signed-off-by: Joshua Kim <20001595+joshua-kim@users.noreply.github.com> commit 49376bd Author: rodrigo <77309055+RodrigoVillar@users.noreply.github.com> Date: Thu Sep 26 14:39:59 2024 -0400 Fix benchmarks (#1590) commit 77a73ba Author: Richard Pringle <richard.pringle@avalabs.org> Date: Tue Sep 24 12:38:27 2024 -0400 Ignore gas on mock-function-call (#1585) commit 1bbab7b Author: containerman17 <8990432+containerman17@users.noreply.github.com> Date: Tue Sep 24 23:19:18 2024 +0900 Remove special cases for Bytes, add arrays support (#1587) commit 3410599 Author: aaronbuchwald <aaron.buchwald56@gmail.com> Date: Mon Sep 23 12:34:55 2024 -0400 Add spam cmd to morpheus readme (#1577) commit dfefd27 Author: aaronbuchwald <aaron.buchwald56@gmail.com> Date: Mon Sep 23 12:34:32 2024 -0400 Improve chain comments (#1579) commit 6c15244 Author: Richard Pringle <richard.pringle@avalabs.org> Date: Fri Sep 20 16:10:21 2024 -0400 Series of changes that improve development (#1583) commit e0a3324 Author: Richard Pringle <richard.pringle@avalabs.org> Date: Fri Sep 20 13:12:58 2024 -0400 Use repr(packed) with state-keys (#1580) commit 75c6cfd Author: aaronbuchwald <aaron.buchwald56@gmail.com> Date: Thu Sep 19 07:58:39 2024 -0400 Separate proposer monitor from vm into internal package (#1574) Signed-off-by: Joshua Kim <20001595+joshua-kim@users.noreply.github.com>

initial revision.

6f2fc50

tsachiherman self-assigned this Oct 16, 2024

lint

1833c3f

richardpringle reviewed Oct 17, 2024

View reviewed changes

examples/morpheusvm/tests/e2e/transfer_test.go Outdated Show resolved Hide resolved

tsachiherman added 7 commits October 17, 2024 15:23

update

a4f51a1

merge

8e22542

update

fe528f5

update

baf90fd

lint

ee7974c

update

0b935bb

update

83afa8d

tsachiherman marked this pull request as ready for review October 18, 2024 14:36

tsachiherman requested review from aaronbuchwald and joshua-kim as code owners October 18, 2024 14:36

tsachiherman requested a review from samliok October 18, 2024 14:37

samliok requested changes Oct 18, 2024

View reviewed changes

joshua-kim linked an issue Oct 22, 2024 that may be closed by this pull request

switch from workload generator to procedural test style for VM-defined workloads #1359

Closed

Merge branch 'main' into tsachi/proc-testing

2523292

aaronbuchwald mentioned this pull request Oct 23, 2024

Refactor Workload #1668

Merged

tsachiherman added 3 commits October 24, 2024 11:54

Merge branch 'main' into tsachi/proc-testing

1716ecf

stage

b681a95

update

9fcc4dc

samliok mentioned this pull request Oct 24, 2024

Improve Testing For Action Developers #1686

Open

tsachiherman added 4 commits October 24, 2024 16:23

update

62048d7

lint

3f7f35c

Merge branch 'main' into tsachi/proc-testing

bb6ca7b

try to fix chainid

b27c3e7

aaronbuchwald requested changes Oct 25, 2024

View reviewed changes

tsachiherman added 2 commits October 25, 2024 11:06

move test registry instance to be created at the vm level.

c55e886

lint

53d05e2

tsachiherman added 5 commits October 25, 2024 17:04

Merge branch 'main' into tsachi/proc-testing

834bf53

lint

128d5fc

Merge branch 'main' into tsachi/proc-testing

93541c0

update per CR

c4621d6

sync e2e tests

bb12aa7

tsachiherman requested a review from aaronbuchwald October 27, 2024 16:10

aaronbuchwald reviewed Oct 28, 2024

View reviewed changes

tests/e2e/e2e.go Outdated Show resolved Hide resolved

aaronbuchwald requested changes Oct 28, 2024

View reviewed changes

tsachiherman and others added 7 commits October 28, 2024 10:28

Update tests/workload/network.go

473a08d

Co-authored-by: aaronbuchwald <aaron.buchwald56@gmail.com> Signed-off-by: Tsachi Herman <24438559+tsachiherman@users.noreply.github.com>

update per CR

66d5493

Merge branch 'tsachi/proc-testing' of github.com:ava-labs/hypersdk in…

4c55dc3

…to tsachi/proc-testing

update per CR

83231c7

use ginkgo.Serial instead of mutex'ing.

4bc7424

lint

f087929

Merge branch 'main' into tsachi/proc-testing

23f5b79

tsachiherman requested a review from aaronbuchwald October 29, 2024 20:55

Merge branch 'main' into tsachi/proc-testing

782f4b2

aaronbuchwald requested changes Oct 30, 2024

View reviewed changes

tsachiherman added 2 commits October 31, 2024 11:00

avoid re-calc of txid

920c45a

Merge branch 'tsachi/proc-testing' of github.com:ava-labs/hypersdk in…

e5bb823

…to tsachi/proc-testing

RodrigoVillar mentioned this pull request Oct 31, 2024

Update MorpheusVM Tutorial #1706

Merged

tsachiherman added 5 commits November 1, 2024 12:04

update per CR

5471cac

Merge branch 'main' into tsachi/proc-testing

2e7261f

fix lint

4021021

fix comment per lint

a346111

update per CR

3b5d436

aaronbuchwald approved these changes Nov 3, 2024

View reviewed changes

aaronbuchwald merged commit b2ad4d3 into main Nov 3, 2024
18 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

support procedural test style for VM-defined workloads #1667

support procedural test style for VM-defined workloads #1667

tsachiherman commented Oct 16, 2024 •

edited

Loading

samliok left a comment •

edited

Loading

samliok Oct 18, 2024

samliok Oct 18, 2024

samliok Oct 18, 2024

samliok Oct 18, 2024

samliok Oct 18, 2024

aaronbuchwald Oct 25, 2024

aaronbuchwald Oct 25, 2024

tsachiherman Oct 25, 2024

aaronbuchwald Oct 28, 2024

tsachiherman Oct 28, 2024

aaronbuchwald Oct 28, 2024

tsachiherman Oct 28, 2024

aaronbuchwald Oct 30, 2024

aaronbuchwald Oct 30, 2024

tsachiherman Oct 31, 2024

aaronbuchwald Oct 30, 2024

tsachiherman Oct 31, 2024

aaronbuchwald Oct 30, 2024

aaronbuchwald Oct 30, 2024

tsachiherman Nov 1, 2024

aaronbuchwald Oct 30, 2024

aaronbuchwald Oct 30, 2024

tsachiherman Oct 31, 2024

aaronbuchwald Oct 30, 2024

tsachiherman Nov 1, 2024

aaronbuchwald Oct 30, 2024

tsachiherman Nov 1, 2024

aaronbuchwald Oct 30, 2024

tsachiherman Nov 1, 2024

aaronbuchwald Oct 30, 2024

tsachiherman Nov 1, 2024


		require.NoError(err)

		timeoutCtx, timeoutCtxFnc := context.WithDeadline(context.Background(), time.Now().Add(2*time.Second))

		var _ = registry.RegisterTest("Transfer Transaction", func(t ginkgo.FullGinkgoTInterface, tn tworkload.TestNetwork) error {
		require := require.New(t)

		// TestsRegistry is global tests registry instance where all the tests are to be registered. Each VM would typically have only
		// a single instance of such a registry.

-// TestsRegistry is global tests registry instance where all the tests are to be registered. Each VM would typically have only
-// a single instance of such a registry.
+// Global test registry initialized during init to ensure tests are identical during ginkgo
+// tree construction and test execution
+// ref https://onsi.github.io/ginkgo/#mental-model-how-ginkgo-traverses-the-spec-hierarchy

support procedural test style for VM-defined workloads #1667

support procedural test style for VM-defined workloads #1667

Conversation

tsachiherman commented Oct 16, 2024 • edited Loading

What ?

Why ?

How would we solve that problem ?

samliok left a comment • 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

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

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

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

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

Choose a reason for hiding this comment

Choose a reason for hiding this comment

tsachiherman commented Oct 16, 2024 •

edited

Loading

samliok left a comment •

edited

Loading