revert(runner): drop selector-first tap; broke path/testTag selectors

internal/runner/runner.go

@@ -562,27 +562,23 @@ func settleForForeground(ctx context.Context, options Options) {
 func applyAction(ctx context.Context, drv driver.DeviceDriver, action verifier.Action, tree *hierarchy.Tree) error {
 	switch action.Kind {
 	case verifier.ActionKindTap:
-		// A selector resolves live on-device at tap time, so it lands correctly
+		x, y, ok := resolveCoordinates(action, tree)
-		// even when the layout shifted between observe and apply (e.g. the soft
+		if !ok {
-		// keyboard finished animating out, moving every control down). Tapping
+			if action.On == "" {
-		// the observe-time coordinates would miss the moved control. Coordinates
+				return nil
-		// are the fallback for selector-less actions (the web/V8 path only emits
+			}
-		// coordinates).
-		if action.On != "" {
 			return drv.TapSelector(ctx, action.On)
 		}
-		if x, y, ok := resolveCoordinates(action, tree); ok {
+		return drv.Tap(ctx, x, y)
-			return drv.Tap(ctx, x, y)
-		}
-		return nil
 	case verifier.ActionKindDoubleTap:
-		if action.On != "" {
+		x, y, ok := resolveCoordinates(action, tree)
+		if !ok {
+			if action.On == "" {
+				return nil
+			}
 			return drv.DoubleTapSelector(ctx, action.On)
 		}
-		if x, y, ok := resolveCoordinates(action, tree); ok {
+		return drv.DoubleTap(ctx, x, y)
-			return drv.DoubleTap(ctx, x, y)
-		}
-		return nil
 	case verifier.ActionKindLongPress:
 		x, y, ok := resolveCoordinates(action, tree)
 		if !ok {
@@ -600,18 +596,14 @@ func applyAction(ctx context.Context, drv driver.DeviceDriver, action verifier.A
 		}
 		return drv.Swipe(ctx, fromX, fromY, toX, toY, duration)
 	case verifier.ActionKindInputText:
-		// Focus the field by selector when one is present, for the same
-		// layout-shift reason as Tap: InputText is the action that raises the
-		// keyboard, so its follow-on focus tap is the most likely to land on a
-		// stale position. Coordinates focus the selector-less web/V8 path.
 		tapped := false
-		if action.On != "" {
+		if x, y, ok := resolveCoordinates(action, tree); ok {
-			if err := drv.TapSelector(ctx, action.On); err != nil {
+			if err := drv.Tap(ctx, x, y); err != nil {
 				return err
 			}
 			tapped = true
-		} else if x, y, ok := resolveCoordinates(action, tree); ok {
+		} else if action.On != "" {
-			if err := drv.Tap(ctx, x, y); err != nil {
+			if err := drv.TapSelector(ctx, action.On); err != nil {
 				return err
 			}
 			tapped = true

internal/runner/runner_test.go

@@ -673,14 +673,13 @@ func TestApplyAction_InputTextErasesExistingTextBeforeTyping(t *testing.T) {
 		t.Fatalf("applyAction: %v", err)
 	}
 	// The post-tap settle is now a brief internal sleep, not a WaitForIdle RPC,
-	// so the recorded driver actions are focus-tap, erase, input. The focus tap
+	// so the recorded driver actions are tap, erase, input.
-	// goes by selector (live on-device resolution) since the action carries one.
 	actions := driverMock.Actions()
 	if len(actions) != 3 {
 		t.Fatalf("want tap, erase, input; got %v", actions)
 	}
-	if actions[0].Kind != mockdriver.ActionTapSelector || actions[0].Selector != "id:username" {
+	if actions[0].Kind != mockdriver.ActionTap {
-		t.Errorf("first action = %+v, want tap_selector id:username", actions[0])
+		t.Errorf("first action = %v, want tap", actions[0].Kind)
 	}
 	if actions[1].Kind != mockdriver.ActionEraseText || actions[1].CharacterCount != len("stale-value") {
 		t.Errorf("second action = %+v, want erase_text of %d characters", actions[1], len("stale-value"))