feat(driver): native DoubleTap RPC for a tight inter-tap gap

Composing two Tap round trips from the Go client spread the taps by
hundreds of milliseconds on iOS, wide enough for the app to navigate
between them, so double-submission races could never reproduce. The
sidecar now lands both taps back-to-back next to the device transport.

internal/driver/sidecar/client.go

@@ -104,14 +104,18 @@ func (c *Client) TapSelector(ctx context.Context, selector string) error {
 	return err
 }
 
-// doubleTapGap is the inter-tap delay for DoubleTap: short enough to land both
+// doubleTapGap is the inter-tap delay for the selector fallback: short enough
-// events inside a sub-100 ms race window, long enough for the sidecar to
+// to land both events inside a sub-100 ms race window, long enough for the
-// serialize two MotionEvent streams. The sidecar exposes no native double-tap
+// sidecar to serialize two MotionEvent streams.
-// RPC, so the gesture is two Taps with this gap.
 const doubleTapGap = 50 * time.Millisecond
 
+// DoubleTap dispatches the native RPC so the backend can land both taps as
+// close together as the platform allows. Composing two Tap round trips from
+// here spreads them by hundreds of milliseconds on iOS, wide enough for
+// navigation to interleave between the taps.
 func (c *Client) DoubleTap(ctx context.Context, x, y int) error {
-	return doubleTap(ctx, func() error { return c.Tap(ctx, x, y) })
+	_, err := c.stub.DoubleTap(ctx, &driverpb.Point{X: int32(x), Y: int32(y)})
+	return err
 }
 
 func (c *Client) DoubleTapSelector(ctx context.Context, selector string) error {

proto/driverpb/driver.pb.go

@@ -1003,11 +1003,13 @@ const file_driverpb_driver_proto_rawDesc = "" +
 	"cpuPercent\x12\x1d\n" +
 	"\n" +
 	"heap_bytes\x18\x02 \x01(\x03R\theapBytes\x12,\n" +
-	"\x12total_memory_bytes\x18\x03 \x01(\x03R\x10totalMemoryBytes2\xd1\t\n" +
+	"\x12total_memory_bytes\x18\x03 \x01(\x03R\x10totalMemoryBytes2\x98\n" +
+	"\n" +
 	"\x06Driver\x12J\n" +
 	"\x06Launch\x12#.sanderling.driver.v1.LaunchRequest\x1a\x1b.sanderling.driver.v1.Empty\x12E\n" +
 	"\tTerminate\x12\x1b.sanderling.driver.v1.Empty\x1a\x1b.sanderling.driver.v1.Empty\x12?\n" +
-	"\x03Tap\x12\x1b.sanderling.driver.v1.Point\x1a\x1b.sanderling.driver.v1.Empty\x12J\n" +
+	"\x03Tap\x12\x1b.sanderling.driver.v1.Point\x1a\x1b.sanderling.driver.v1.Empty\x12E\n" +
+	"\tDoubleTap\x12\x1b.sanderling.driver.v1.Point\x1a\x1b.sanderling.driver.v1.Empty\x12J\n" +
 	"\vTapSelector\x12\x1e.sanderling.driver.v1.Selector\x1a\x1b.sanderling.driver.v1.Empty\x12D\n" +
 	"\tInputText\x12\x1a.sanderling.driver.v1.Text\x1a\x1b.sanderling.driver.v1.Empty\x12P\n" +
 	"\tEraseText\x12&.sanderling.driver.v1.EraseTextRequest\x1a\x1b.sanderling.driver.v1.Empty\x12H\n" +
@@ -1069,37 +1071,39 @@ var file_driverpb_driver_proto_depIdxs = []int32{
 	1,  // 6: sanderling.driver.v1.Driver.Launch:input_type -> sanderling.driver.v1.LaunchRequest
 	0,  // 7: sanderling.driver.v1.Driver.Terminate:input_type -> sanderling.driver.v1.Empty
 	2,  // 8: sanderling.driver.v1.Driver.Tap:input_type -> sanderling.driver.v1.Point
-	5,  // 9: sanderling.driver.v1.Driver.TapSelector:input_type -> sanderling.driver.v1.Selector
+	2,  // 9: sanderling.driver.v1.Driver.DoubleTap:input_type -> sanderling.driver.v1.Point
-	3,  // 10: sanderling.driver.v1.Driver.InputText:input_type -> sanderling.driver.v1.Text
+	5,  // 10: sanderling.driver.v1.Driver.TapSelector:input_type -> sanderling.driver.v1.Selector
-	4,  // 11: sanderling.driver.v1.Driver.EraseText:input_type -> sanderling.driver.v1.EraseTextRequest
+	3,  // 11: sanderling.driver.v1.Driver.InputText:input_type -> sanderling.driver.v1.Text
-	11, // 12: sanderling.driver.v1.Driver.Swipe:input_type -> sanderling.driver.v1.SwipeRequest
+	4,  // 12: sanderling.driver.v1.Driver.EraseText:input_type -> sanderling.driver.v1.EraseTextRequest
-	12, // 13: sanderling.driver.v1.Driver.PressKey:input_type -> sanderling.driver.v1.PressKeyRequest
+	11, // 13: sanderling.driver.v1.Driver.Swipe:input_type -> sanderling.driver.v1.SwipeRequest
-	2,  // 14: sanderling.driver.v1.Driver.LongPress:input_type -> sanderling.driver.v1.Point
+	12, // 14: sanderling.driver.v1.Driver.PressKey:input_type -> sanderling.driver.v1.PressKeyRequest
-	0,  // 15: sanderling.driver.v1.Driver.Screenshot:input_type -> sanderling.driver.v1.Empty
+	2,  // 15: sanderling.driver.v1.Driver.LongPress:input_type -> sanderling.driver.v1.Point
-	0,  // 16: sanderling.driver.v1.Driver.Hierarchy:input_type -> sanderling.driver.v1.Empty
+	0,  // 16: sanderling.driver.v1.Driver.Screenshot:input_type -> sanderling.driver.v1.Empty
-	0,  // 17: sanderling.driver.v1.Driver.Snapshot:input_type -> sanderling.driver.v1.Empty
+	0,  // 17: sanderling.driver.v1.Driver.Hierarchy:input_type -> sanderling.driver.v1.Empty
-	13, // 18: sanderling.driver.v1.Driver.RecentLogs:input_type -> sanderling.driver.v1.RecentLogsRequest
+	0,  // 18: sanderling.driver.v1.Driver.Snapshot:input_type -> sanderling.driver.v1.Empty
-	9,  // 19: sanderling.driver.v1.Driver.WaitForIdle:input_type -> sanderling.driver.v1.Duration
+	13, // 19: sanderling.driver.v1.Driver.RecentLogs:input_type -> sanderling.driver.v1.RecentLogsRequest
-	0,  // 20: sanderling.driver.v1.Driver.Health:input_type -> sanderling.driver.v1.Empty
+	9,  // 20: sanderling.driver.v1.Driver.WaitForIdle:input_type -> sanderling.driver.v1.Duration
-	16, // 21: sanderling.driver.v1.Driver.Metrics:input_type -> sanderling.driver.v1.MetricsRequest
+	0,  // 21: sanderling.driver.v1.Driver.Health:input_type -> sanderling.driver.v1.Empty
-	0,  // 22: sanderling.driver.v1.Driver.Launch:output_type -> sanderling.driver.v1.Empty
+	16, // 22: sanderling.driver.v1.Driver.Metrics:input_type -> sanderling.driver.v1.MetricsRequest
-	0,  // 23: sanderling.driver.v1.Driver.Terminate:output_type -> sanderling.driver.v1.Empty
+	0,  // 23: sanderling.driver.v1.Driver.Launch:output_type -> sanderling.driver.v1.Empty
-	0,  // 24: sanderling.driver.v1.Driver.Tap:output_type -> sanderling.driver.v1.Empty
+	0,  // 24: sanderling.driver.v1.Driver.Terminate:output_type -> sanderling.driver.v1.Empty
-	0,  // 25: sanderling.driver.v1.Driver.TapSelector:output_type -> sanderling.driver.v1.Empty
+	0,  // 25: sanderling.driver.v1.Driver.Tap:output_type -> sanderling.driver.v1.Empty
-	0,  // 26: sanderling.driver.v1.Driver.InputText:output_type -> sanderling.driver.v1.Empty
+	0,  // 26: sanderling.driver.v1.Driver.DoubleTap:output_type -> sanderling.driver.v1.Empty
-	0,  // 27: sanderling.driver.v1.Driver.EraseText:output_type -> sanderling.driver.v1.Empty
+	0,  // 27: sanderling.driver.v1.Driver.TapSelector:output_type -> sanderling.driver.v1.Empty
-	0,  // 28: sanderling.driver.v1.Driver.Swipe:output_type -> sanderling.driver.v1.Empty
+	0,  // 28: sanderling.driver.v1.Driver.InputText:output_type -> sanderling.driver.v1.Empty
-	0,  // 29: sanderling.driver.v1.Driver.PressKey:output_type -> sanderling.driver.v1.Empty
+	0,  // 29: sanderling.driver.v1.Driver.EraseText:output_type -> sanderling.driver.v1.Empty
-	0,  // 30: sanderling.driver.v1.Driver.LongPress:output_type -> sanderling.driver.v1.Empty
+	0,  // 30: sanderling.driver.v1.Driver.Swipe:output_type -> sanderling.driver.v1.Empty
-	6,  // 31: sanderling.driver.v1.Driver.Screenshot:output_type -> sanderling.driver.v1.Image
+	0,  // 31: sanderling.driver.v1.Driver.PressKey:output_type -> sanderling.driver.v1.Empty
-	7,  // 32: sanderling.driver.v1.Driver.Hierarchy:output_type -> sanderling.driver.v1.HierarchyJSON
+	0,  // 32: sanderling.driver.v1.Driver.LongPress:output_type -> sanderling.driver.v1.Empty
-	8,  // 33: sanderling.driver.v1.Driver.Snapshot:output_type -> sanderling.driver.v1.SnapshotResponse
+	6,  // 33: sanderling.driver.v1.Driver.Screenshot:output_type -> sanderling.driver.v1.Image
-	14, // 34: sanderling.driver.v1.Driver.RecentLogs:output_type -> sanderling.driver.v1.LogEntries
+	7,  // 34: sanderling.driver.v1.Driver.Hierarchy:output_type -> sanderling.driver.v1.HierarchyJSON
-	0,  // 35: sanderling.driver.v1.Driver.WaitForIdle:output_type -> sanderling.driver.v1.Empty
+	8,  // 35: sanderling.driver.v1.Driver.Snapshot:output_type -> sanderling.driver.v1.SnapshotResponse
-	10, // 36: sanderling.driver.v1.Driver.Health:output_type -> sanderling.driver.v1.HealthStatus
+	14, // 36: sanderling.driver.v1.Driver.RecentLogs:output_type -> sanderling.driver.v1.LogEntries
-	17, // 37: sanderling.driver.v1.Driver.Metrics:output_type -> sanderling.driver.v1.MetricsResponse
+	0,  // 37: sanderling.driver.v1.Driver.WaitForIdle:output_type -> sanderling.driver.v1.Empty
-	22, // [22:38] is the sub-list for method output_type
+	10, // 38: sanderling.driver.v1.Driver.Health:output_type -> sanderling.driver.v1.HealthStatus
-	6,  // [6:22] is the sub-list for method input_type
+	17, // 39: sanderling.driver.v1.Driver.Metrics:output_type -> sanderling.driver.v1.MetricsResponse
+	23, // [23:40] is the sub-list for method output_type
+	6,  // [6:23] is the sub-list for method input_type
 	6,  // [6:6] is the sub-list for extension type_name
 	6,  // [6:6] is the sub-list for extension extendee
 	0,  // [0:6] is the sub-list for field type_name

proto/driverpb/driver.proto

@@ -11,6 +11,11 @@ service Driver {
   rpc Terminate(Empty) returns (Empty);
 
   rpc Tap(Point) returns (Empty);
+  // DoubleTap lands two taps as close together as the platform allows.
+  // Composing two Tap calls from the client spreads them by hundreds of
+  // milliseconds on iOS, wide enough for navigation to interleave, which
+  // defeats double-submission testing.
+  rpc DoubleTap(Point) returns (Empty);
   rpc TapSelector(Selector) returns (Empty);
   rpc InputText(Text) returns (Empty);
   // EraseText deletes characters from the focused field so InputText can

proto/driverpb/driver_grpc.pb.go

@@ -22,6 +22,7 @@ const (
 	Driver_Launch_FullMethodName      = "/sanderling.driver.v1.Driver/Launch"
 	Driver_Terminate_FullMethodName   = "/sanderling.driver.v1.Driver/Terminate"
 	Driver_Tap_FullMethodName         = "/sanderling.driver.v1.Driver/Tap"
+	Driver_DoubleTap_FullMethodName   = "/sanderling.driver.v1.Driver/DoubleTap"
 	Driver_TapSelector_FullMethodName = "/sanderling.driver.v1.Driver/TapSelector"
 	Driver_InputText_FullMethodName   = "/sanderling.driver.v1.Driver/InputText"
 	Driver_EraseText_FullMethodName   = "/sanderling.driver.v1.Driver/EraseText"
@@ -44,6 +45,11 @@ type DriverClient interface {
 	Launch(ctx context.Context, in *LaunchRequest, opts ...grpc.CallOption) (*Empty, error)
 	Terminate(ctx context.Context, in *Empty, opts ...grpc.CallOption) (*Empty, error)
 	Tap(ctx context.Context, in *Point, opts ...grpc.CallOption) (*Empty, error)
+	// DoubleTap lands two taps as close together as the platform allows.
+	// Composing two Tap calls from the client spreads them by hundreds of
+	// milliseconds on iOS, wide enough for navigation to interleave, which
+	// defeats double-submission testing.
+	DoubleTap(ctx context.Context, in *Point, opts ...grpc.CallOption) (*Empty, error)
 	TapSelector(ctx context.Context, in *Selector, opts ...grpc.CallOption) (*Empty, error)
 	InputText(ctx context.Context, in *Text, opts ...grpc.CallOption) (*Empty, error)
 	// EraseText deletes characters from the focused field so InputText can
@@ -102,6 +108,16 @@ func (c *driverClient) Tap(ctx context.Context, in *Point, opts ...grpc.CallOpti
 	return out, nil
 }
 
+func (c *driverClient) DoubleTap(ctx context.Context, in *Point, opts ...grpc.CallOption) (*Empty, error) {
+	cOpts := append([]grpc.CallOption{grpc.StaticMethod()}, opts...)
+	out := new(Empty)
+	err := c.cc.Invoke(ctx, Driver_DoubleTap_FullMethodName, in, out, cOpts...)
+	if err != nil {
+		return nil, err
+	}
+	return out, nil
+}
+
 func (c *driverClient) TapSelector(ctx context.Context, in *Selector, opts ...grpc.CallOption) (*Empty, error) {
 	cOpts := append([]grpc.CallOption{grpc.StaticMethod()}, opts...)
 	out := new(Empty)
@@ -239,6 +255,11 @@ type DriverServer interface {
 	Launch(context.Context, *LaunchRequest) (*Empty, error)
 	Terminate(context.Context, *Empty) (*Empty, error)
 	Tap(context.Context, *Point) (*Empty, error)
+	// DoubleTap lands two taps as close together as the platform allows.
+	// Composing two Tap calls from the client spreads them by hundreds of
+	// milliseconds on iOS, wide enough for navigation to interleave, which
+	// defeats double-submission testing.
+	DoubleTap(context.Context, *Point) (*Empty, error)
 	TapSelector(context.Context, *Selector) (*Empty, error)
 	InputText(context.Context, *Text) (*Empty, error)
 	// EraseText deletes characters from the focused field so InputText can
@@ -276,6 +297,9 @@ func (UnimplementedDriverServer) Terminate(context.Context, *Empty) (*Empty, err
 func (UnimplementedDriverServer) Tap(context.Context, *Point) (*Empty, error) {
 	return nil, status.Error(codes.Unimplemented, "method Tap not implemented")
 }
+func (UnimplementedDriverServer) DoubleTap(context.Context, *Point) (*Empty, error) {
+	return nil, status.Error(codes.Unimplemented, "method DoubleTap not implemented")
+}
 func (UnimplementedDriverServer) TapSelector(context.Context, *Selector) (*Empty, error) {
 	return nil, status.Error(codes.Unimplemented, "method TapSelector not implemented")
 }
@@ -390,6 +414,24 @@ func _Driver_Tap_Handler(srv interface{}, ctx context.Context, dec func(interfac
 	return interceptor(ctx, in, info, handler)
 }
 
+func _Driver_DoubleTap_Handler(srv interface{}, ctx context.Context, dec func(interface{}) error, interceptor grpc.UnaryServerInterceptor) (interface{}, error) {
+	in := new(Point)
+	if err := dec(in); err != nil {
+		return nil, err
+	}
+	if interceptor == nil {
+		return srv.(DriverServer).DoubleTap(ctx, in)
+	}
+	info := &grpc.UnaryServerInfo{
+		Server:     srv,
+		FullMethod: Driver_DoubleTap_FullMethodName,
+	}
+	handler := func(ctx context.Context, req interface{}) (interface{}, error) {
+		return srv.(DriverServer).DoubleTap(ctx, req.(*Point))
+	}
+	return interceptor(ctx, in, info, handler)
+}
+
 func _Driver_TapSelector_Handler(srv interface{}, ctx context.Context, dec func(interface{}) error, interceptor grpc.UnaryServerInterceptor) (interface{}, error) {
 	in := new(Selector)
 	if err := dec(in); err != nil {
@@ -643,6 +685,10 @@ var Driver_ServiceDesc = grpc.ServiceDesc{
 			MethodName: "Tap",
 			Handler:    _Driver_Tap_Handler,
 		},
+		{
+			MethodName: "DoubleTap",
+			Handler:    _Driver_DoubleTap_Handler,
+		},
 		{
 			MethodName: "TapSelector",
 			Handler:    _Driver_TapSelector_Handler,

proto/driverpb/driver_test.go

@@ -16,6 +16,7 @@ func TestDriverServiceDescriptor(t *testing.T) {
 		"Launch":      true,
 		"Terminate":   true,
 		"Tap":         true,
+		"DoubleTap":   true,
 		"TapSelector": true,
 		"InputText":   true,
 		"EraseText":   true,

sidecar/src/main/kotlin/dev/sanderling/sidecar/DriverBackend.kt

@@ -4,6 +4,15 @@ interface DriverBackend {
     fun launch(bundleId: String, clearState: Boolean, env: Map<String, String> = emptyMap())
     fun terminate(bundleId: String)
     fun tap(x: Int, y: Int)
+
+    // doubleTap lands two taps as close together as the platform allows.
+    // The default composes two taps back-to-back; backends with higher
+    // per-tap latency override to tighten the gap.
+    fun doubleTap(x: Int, y: Int) {
+        tap(x, y)
+        tap(x, y)
+    }
+
     fun tapSelector(selector: String)
     fun inputText(text: String)
     fun eraseText(characterCount: Int)
@@ -735,6 +744,15 @@ class IosDriverBackend(private val udid: String) : DriverBackend {
 
     override fun tap(x: Int, y: Int) = withReconnect { driver.tap(maestro.Point(x, y)) }
 
+    // Both taps inside one reconnect scope, with nothing between them: the
+    // XCTest transport adds hundreds of milliseconds per round trip, so any
+    // client-side composition spreads the taps wide enough for the app to
+    // navigate between them.
+    override fun doubleTap(x: Int, y: Int) = withReconnect {
+        driver.tap(maestro.Point(x, y))
+        driver.tap(maestro.Point(x, y))
+    }
+
     override fun longPress(x: Int, y: Int) = withReconnect { driver.longPress(maestro.Point(x, y)) }
 
     override fun tapSelector(selector: String) = withReconnect {

sidecar/src/main/kotlin/dev/sanderling/sidecar/DriverService.kt

@@ -54,6 +54,13 @@ class DriverService(
         }
     }
 
+    override fun doubleTap(request: Point, responseObserver: StreamObserver<Empty>) {
+        runRpc(responseObserver) {
+            backend.doubleTap(request.x, request.y)
+            Empty.getDefaultInstance()
+        }
+    }
+
     override fun longPress(request: Point, responseObserver: StreamObserver<Empty>) {
         runRpc(responseObserver) {
             backend.longPress(request.x, request.y)

sidecar/src/test/kotlin/dev/sanderling/sidecar/DriverServiceTest.kt

@@ -83,6 +83,34 @@ class DriverServiceTest {
         assertEquals("hello world", backend.lastInputText)
     }
 
+    @Test fun doubleTapDefaultComposesTwoTaps() {
+        // Interface delegation would bind the default doubleTap to the
+        // delegate, bypassing the tap override, so implement the interface
+        // directly.
+        val taps = mutableListOf<Pair<Int, Int>>()
+        val backend = object : DriverBackend {
+            override fun launch(bundleId: String, clearState: Boolean, env: Map<String, String>) {}
+            override fun terminate(bundleId: String) {}
+            override fun tap(x: Int, y: Int) { taps.add(x to y) }
+            override fun tapSelector(selector: String) {}
+            override fun inputText(text: String) {}
+            override fun eraseText(characterCount: Int) {}
+            override fun swipe(fromX: Int, fromY: Int, toX: Int, toY: Int, durationMillis: Long) {}
+            override fun pressKey(key: String) {}
+            override fun longPress(x: Int, y: Int) {}
+            override fun screenshot(): Triple<ByteArray, Int, Int> = Triple(byteArrayOf(), 0, 0)
+            override fun hierarchy(): String = "{}"
+            override fun recentLogs(sinceUnixMillis: Long, minLevel: String): List<LogLine> = emptyList()
+            override fun waitForIdle(durationMillis: Long) {}
+            override fun healthy(): Boolean = true
+            override fun metrics(bundleId: String): MetricsSample = MetricsSample(0.0, 0L, 0L)
+        }
+        val client = newClient(backend)
+
+        client.doubleTap(Point.newBuilder().setX(120).setY(340).build())
+        assertEquals(listOf(120 to 340, 120 to 340), taps)
+    }
+
     @Test fun eraseTextForwardsCharacterCount() {
         val backend = StubDriverBackend("android")
         val client = newClient(backend)