Files
sanderling/internal/runner/runner_test.go
T
pj 69002b07b1 fix(runner,sample-app): surface silent errors and demo a failing property (#15)
* fix(runner): surface non-deadline WaitForIdle errors

Previously the WaitForIdle return value was discarded entirely, hiding
real driver failures (gRPC transport errors, sidecar crashes) behind
the expected deadline-exceeded case. Log non-deadline errors so they
are visible without changing control flow.

* chore(sample-app): drop unused uptime_millis extractor

Registered in SampleApplication but never consumed by spec.ts.

* fix(sample-app): drop trivial appIsRunning property

app_state was hardcoded to 'running' so the property was a tautology
that could never fail. Removing both the extractor and the property
is the simplest fix; demo-grade properties that can fail land next.

* feat(sample-app): add Reset button that zeroes clickCount

Pairs with the next commit's tap-reset action so the fuzzer can
violate clickCountNeverDecreases and demonstrate uatu actually
finding a property violation.

* feat(sample-app): add tap-reset action to exercise Reset button

Weighted at 10/122, fuzzer reaches it within a short run. Pairs with
the Reset button to demonstrate uatu detecting the
clickCountNeverDecreases violation.

* fix(runner): filter WaitForIdle errors via context state, not errors.Is

errors.Is(err, context.DeadlineExceeded) misses gRPC's wrapped
status.DeadlineExceeded, so every step under the maestro driver
logged a spurious warning. Check idleCtx.Err() instead — captures
both deadline-fired and parent-canceled cases regardless of how the
driver wraps them.

* chore(sample-app): tune action weights so demo violates in ~30s

Prior weights left tap-reset rare enough that short demo runs missed
the violation by chance. Bumped to 30/107, with typeUsername reduced
since username noise doesn't help exercise clickCount.

* refactor(runner): route warnings through slog

Adds Options.Logger (defaults to slog.Default()) and converts the
three warning sites that were using fmt.Printf. Progress line stays
on Printf since it's user-facing UI, not a log. Makes the warnings
testable via a capturing handler.

* test(runner): assert WaitForIdle driver errors are logged

Captures slog output via TextHandler into a buffer and asserts the
warning message + injected error text appear when the mock driver
returns a non-context error from WaitForIdle. Guards against a
regression of the silent-error swallow.
2026-04-18 18:48:25 +07:00

386 lines
11 KiB
Go

package runner
import (
"bytes"
"context"
"encoding/json"
"errors"
"log/slog"
"net"
"os"
"path/filepath"
"slices"
"strings"
"sync"
"testing"
"time"
"github.com/priyanshujain/uatu/internal/agent"
mockdriver "github.com/priyanshujain/uatu/internal/driver/mock"
"github.com/priyanshujain/uatu/internal/trace"
"github.com/priyanshujain/uatu/internal/verifier"
)
const fixtureSpec = `
const screen = __uatu__.extract(state => state.snapshots.screen ?? "");
const balance = __uatu__.extract(state => state.snapshots.balance ?? 0);
globalThis.properties = {
balanceNonNegative: __uatu__.always(() => balance.current >= 0),
};
globalThis.actions = __uatu__.actions(() => [__uatu__.tap({ on: "id:next" })]);
`
type harness struct {
server *agent.Server
listener net.Listener
clientWG sync.WaitGroup
conn *agent.Conn
mock *mockdriver.Driver
verifier *verifier.Verifier
writer *trace.Writer
snapshot []map[string]json.RawMessage
}
func newHarness(t *testing.T, snapshots []map[string]json.RawMessage) *harness {
t.Helper()
listener, err := net.Listen("tcp", "127.0.0.1:0")
if err != nil {
t.Fatal(err)
}
server := agent.NewServer(listener)
directory := t.TempDir()
writer, err := trace.NewWriter(directory)
if err != nil {
t.Fatal(err)
}
verifierInstance, err := verifier.New()
if err != nil {
t.Fatal(err)
}
if err := verifierInstance.Load(fixtureSpec); err != nil {
t.Fatal(err)
}
state := &harness{
server: server,
listener: listener,
mock: mockdriver.New(),
verifier: verifierInstance,
writer: writer,
snapshot: snapshots,
}
t.Cleanup(func() {
_ = listener.Close()
_ = writer.Close()
})
return state
}
func (h *harness) startSDK(t *testing.T) {
t.Helper()
h.clientWG.Go(func() {
conn, err := net.Dial("tcp", h.listener.Addr().String())
if err != nil {
t.Errorf("dial: %v", err)
return
}
defer conn.Close()
if err := agent.WriteMessage(conn, agent.Hello("0.0.1", "android", "com.fixture")); err != nil {
t.Errorf("hello: %v", err)
return
}
index := 0
for {
message, err := agent.ReadMessage(conn)
if err != nil {
return
}
if message.Type == agent.MessageTypePause {
snapshots := map[string]json.RawMessage{}
if index < len(h.snapshot) {
snapshots = h.snapshot[index]
}
if err := agent.WriteMessage(conn, agent.State(message.ID, snapshots)); err != nil {
return
}
index++
}
}
})
}
func (h *harness) acceptConnection(t *testing.T) {
t.Helper()
ctx, cancel := context.WithTimeout(context.Background(), 2*time.Second)
defer cancel()
connection, err := h.server.Accept(ctx)
if err != nil {
t.Fatalf("Accept: %v", err)
}
h.conn = connection
}
func TestRunner_HappyPathStepsAndTraces(t *testing.T) {
snapshots := []map[string]json.RawMessage{
{"screen": json.RawMessage(`"home"`), "balance": json.RawMessage(`100`)},
{"screen": json.RawMessage(`"home"`), "balance": json.RawMessage(`200`)},
{"screen": json.RawMessage(`"home"`), "balance": json.RawMessage(`300`)},
}
state := newHarness(t, snapshots)
state.startSDK(t)
state.acceptConnection(t)
ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
defer cancel()
summary, err := Run(ctx, Options{
Duration: 100 * time.Millisecond,
SnapshotTimeout: 2 * time.Second,
IdleTimeout: 50 * time.Millisecond,
Connection: state.conn,
Driver: state.mock,
Verifier: state.verifier,
TraceWriter: state.writer,
})
if err != nil {
t.Fatalf("Run: %v", err)
}
if summary.Steps == 0 {
t.Errorf("expected at least one step, got 0")
}
if len(summary.Violations) != 0 {
t.Errorf("no violations expected, got %v", summary.Violations)
}
actions := state.mock.Actions()
if !containsAction(actions, mockdriver.ActionTapSelector, "id:next") {
t.Errorf("expected TapSelector with id:next, got %v", actions)
}
}
func TestRunner_ViolationSurfacesInSummary(t *testing.T) {
snapshots := []map[string]json.RawMessage{
{"balance": json.RawMessage(`100`)},
{"balance": json.RawMessage(`-1`)},
{"balance": json.RawMessage(`50`)},
}
state := newHarness(t, snapshots)
state.startSDK(t)
state.acceptConnection(t)
ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
defer cancel()
summary, err := Run(ctx, Options{
Duration: 100 * time.Millisecond,
SnapshotTimeout: 2 * time.Second,
IdleTimeout: 50 * time.Millisecond,
Connection: state.conn,
Driver: state.mock,
Verifier: state.verifier,
TraceWriter: state.writer,
})
if err != nil {
t.Fatalf("Run: %v", err)
}
if len(summary.Violations) == 0 {
t.Errorf("expected at least one violation, got %v", summary.Violations)
}
if !containsProperty(summary.Violations, "balanceNonNegative") {
t.Errorf("expected balanceNonNegative in violations: %v", summary.Violations)
}
}
func TestRunner_RejectsMissingFields(t *testing.T) {
_, err := Run(context.Background(), Options{Duration: time.Second})
if err == nil || !strings.Contains(err.Error(), "Connection") {
t.Errorf("expected Connection-required error, got %v", err)
}
}
func TestRunner_RejectsZeroDuration(t *testing.T) {
_, err := Run(context.Background(), Options{
Connection: &agent.Conn{},
Driver: mockdriver.New(),
Verifier: mustNewVerifier(t),
TraceWriter: mustNewTraceWriter(t),
})
if err == nil || !strings.Contains(err.Error(), "Duration") {
t.Errorf("expected Duration-required error, got %v", err)
}
}
func TestRunner_RecordsScreenFieldFromSnapshot(t *testing.T) {
snapshots := []map[string]json.RawMessage{
{"screen": json.RawMessage(`"customer_ledger"`), "balance": json.RawMessage(`1`)},
}
state := newHarness(t, snapshots)
state.startSDK(t)
state.acceptConnection(t)
ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
defer cancel()
if _, err := Run(ctx, Options{
Duration: 100 * time.Millisecond,
SnapshotTimeout: 2 * time.Second,
IdleTimeout: 50 * time.Millisecond,
Connection: state.conn,
Driver: state.mock,
Verifier: state.verifier,
TraceWriter: state.writer,
}); err != nil {
t.Fatal(err)
}
body, err := os.ReadFile(filepath.Join(state.writer.Directory(), "trace.jsonl"))
if err != nil {
t.Fatal(err)
}
if !strings.Contains(string(body), `"screen":"customer_ledger"`) {
t.Errorf("screen field not in trace: %s", body)
}
}
func TestScreenFromSnapshot(t *testing.T) {
t.Run("string value returns screen", func(t *testing.T) {
snapshots := map[string]json.RawMessage{"screen": json.RawMessage(`"home"`)}
screen, err := screenFromSnapshot(snapshots)
if err != nil {
t.Fatalf("unexpected error: %v", err)
}
if screen != "home" {
t.Errorf("screen = %q, want %q", screen, "home")
}
})
t.Run("missing key returns empty with no error", func(t *testing.T) {
screen, err := screenFromSnapshot(map[string]json.RawMessage{})
if err != nil {
t.Fatalf("unexpected error: %v", err)
}
if screen != "" {
t.Errorf("screen = %q, want empty", screen)
}
})
t.Run("non-string value returns error", func(t *testing.T) {
snapshots := map[string]json.RawMessage{"screen": json.RawMessage(`{"nested":1}`)}
screen, err := screenFromSnapshot(snapshots)
if err == nil {
t.Fatalf("expected error for non-string screen, got nil")
}
if screen != "" {
t.Errorf("screen = %q, want empty on error", screen)
}
})
}
func TestRunner_LogsWaitForIdleDriverErrors(t *testing.T) {
snapshots := []map[string]json.RawMessage{
{"balance": json.RawMessage(`100`)},
}
state := newHarness(t, snapshots)
state.startSDK(t)
state.acceptConnection(t)
state.mock.Failures[mockdriver.ActionWaitForIdle] = errors.New("sidecar lost gRPC stream")
var logBuf bytes.Buffer
logger := slog.New(slog.NewTextHandler(&logBuf, &slog.HandlerOptions{Level: slog.LevelWarn}))
ctx, cancel := context.WithTimeout(context.Background(), 5*time.Second)
defer cancel()
if _, err := Run(ctx, Options{
Duration: 100 * time.Millisecond,
SnapshotTimeout: 2 * time.Second,
IdleTimeout: 50 * time.Millisecond,
Connection: state.conn,
Driver: state.mock,
Verifier: state.verifier,
TraceWriter: state.writer,
Logger: logger,
}); err != nil {
t.Fatalf("Run: %v", err)
}
output := logBuf.String()
if !strings.Contains(output, "wait_for_idle failed") {
t.Errorf("expected wait_for_idle warning, got: %q", output)
}
if !strings.Contains(output, "sidecar lost gRPC stream") {
t.Errorf("expected driver error message in warning, got: %q", output)
}
}
func TestApplyAction_InputTextSurfacesFocusTapError(t *testing.T) {
t.Run("selector focus tap fails", func(t *testing.T) {
driverMock := mockdriver.New()
driverMock.Failures[mockdriver.ActionTapSelector] = errors.New("adb unreachable")
action := verifier.Action{Kind: verifier.ActionKindInputText, On: "id:username", Text: "alice"}
err := applyAction(context.Background(), driverMock, action, nil)
if err == nil {
t.Fatalf("expected focus tap failure to surface, got nil")
}
if containsAction(driverMock.Actions(), mockdriver.ActionInputText, "") {
t.Errorf("InputText must not run after focus tap failed: %v", driverMock.Actions())
}
})
t.Run("coordinate focus tap fails", func(t *testing.T) {
driverMock := mockdriver.New()
driverMock.Failures[mockdriver.ActionTap] = errors.New("tap driver error")
action := verifier.Action{Kind: verifier.ActionKindInputText, X: 10, Y: 20, Text: "alice"}
err := applyAction(context.Background(), driverMock, action, nil)
if err == nil {
t.Fatalf("expected focus tap failure to surface, got nil")
}
if containsAction(driverMock.Actions(), mockdriver.ActionInputText, "") {
t.Errorf("InputText must not run after focus tap failed: %v", driverMock.Actions())
}
})
}
func mustNewVerifier(t *testing.T) *verifier.Verifier {
t.Helper()
verifierInstance, err := verifier.New()
if err != nil {
t.Fatal(err)
}
return verifierInstance
}
func mustNewTraceWriter(t *testing.T) *trace.Writer {
t.Helper()
writer, err := trace.NewWriter(t.TempDir())
if err != nil {
t.Fatal(err)
}
t.Cleanup(func() { _ = writer.Close() })
return writer
}
func containsAction(actions []mockdriver.Action, kind mockdriver.ActionKind, payload string) bool {
for _, action := range actions {
if action.Kind != kind {
continue
}
switch kind {
case mockdriver.ActionLaunch:
if action.BundleID == payload {
return true
}
case mockdriver.ActionTapSelector:
if action.Selector == payload {
return true
}
case mockdriver.ActionTerminate:
return true
default:
return true
}
}
return false
}
func containsProperty(records []ViolationRecord, property string) bool {
for _, record := range records {
if slices.Contains(record.Properties, property) {
return true
}
}
return false
}