Not directly related but I think 'pumpAndSettle control test' test in packages/flutter_test/test/widget_tester_test.dart should have been put in the pumping group.

Thanks @Hixie so much for the helpful discussion! I landed what's discussed together with some updates in the comments and redo the description of this PR for the motivation and explaining the PR. Let me know what you think about the new version of description as the document you suggested.

Another issue I discussed with @liyuqian shortly after our meeting is, probably I need to change the injected input class from PointerEvent to PointerData for enabling the integration test for #60558 . My previous trial failed on PointerData due to #60734 but I'll have a try if there's a way to get around tomorrow.

A conclusion I can draw now about preparing for #60558 is:

1. Pointer event resampler (#41118) #60558 's solution is to put an extra layer between window.onPointerDataPacket and GestureBinding for PointerData.
2. WidgetTester and all our test bindings are injecting PointerEvent to drive the test app, injecting PointerData with current binding will look like input from device, which is handled here:

   flutter/packages/flutter_test/lib/src/widget_tester.dart

   Lines 659 to 764 in 6eaaf16

   	@override
   	void dispatchEvent(PointerEvent event, HitTestResult result) {
   	if (event is PointerDownEvent) {
   	final RenderObject innerTarget = result.path
   	.map((HitTestEntry candidate) => candidate.target)
   	.whereType<RenderObject>()
   	.first;
   	final Element innerTargetElement = collectAllElementsFrom(
   	binding.renderViewElement,
   	skipOffstage: true,
   	).lastWhere(
   	(Element element) => element.renderObject == innerTarget,
   	orElse: () => null,
   	);
   	if (innerTargetElement == null) {
   	debugPrint('No widgets found at ${binding.globalToLocal(event.position)}.');
   	return;
   	}
   	final List<Element> candidates = <Element>[];
   	innerTargetElement.visitAncestorElements((Element element) {
   	candidates.add(element);
   	return true;
   	});
   	assert(candidates.isNotEmpty);
   	String descendantText;
   	int numberOfWithTexts = 0;
   	int numberOfTypes = 0;
   	int totalNumber = 0;
   	debugPrint('Some possible finders for the widgets at ${binding.globalToLocal(event.position)}:');
   	for (final Element element in candidates) {
   	if (totalNumber > 13) // an arbitrary number of finders that feels useful without being overwhelming
   	break;
   	totalNumber += 1; // optimistically assume we'll be able to describe it
   	final Widget widget = element.widget;
   	if (widget is Tooltip) {
   	final Iterable<Element> matches = find.byTooltip(widget.message).evaluate();
   	if (matches.length == 1) {
   	debugPrint(" find.byTooltip('${widget.message}')");
   	continue;
   	}
   	}
   	if (widget is Text) {
   	assert(descendantText == null);
   	final Iterable<Element> matches = find.text(widget.data).evaluate();
   	descendantText = widget.data;
   	if (matches.length == 1) {
   	debugPrint(" find.text('${widget.data}')");
   	continue;
   	}
   	}
   	final Key key = widget.key;
   	if (key is ValueKey<dynamic>) {
   	String keyLabel;
   	if (key is ValueKey<int> ||
   	key is ValueKey<double> ||
   	key is ValueKey<bool>) {
   	keyLabel = 'const ${key.runtimeType}(${key.value})';
   	} else if (key is ValueKey<String>) {
   	keyLabel = "const Key('${key.value}')";
   	}
   	if (keyLabel != null) {
   	final Iterable<Element> matches = find.byKey(key).evaluate();
   	if (matches.length == 1) {
   	debugPrint(' find.byKey($keyLabel)');
   	continue;
   	}
   	}
   	}
   	if (!_isPrivate(widget.runtimeType)) {
   	if (numberOfTypes < 5) {
   	final Iterable<Element> matches = find.byType(widget.runtimeType).evaluate();
   	if (matches.length == 1) {
   	debugPrint(' find.byType(${widget.runtimeType})');
   	numberOfTypes += 1;
   	continue;
   	}
   	}
   	if (descendantText != null && numberOfWithTexts < 5) {
   	final Iterable<Element> matches = find.widgetWithText(widget.runtimeType, descendantText).evaluate();
   	if (matches.length == 1) {
   	debugPrint(" find.widgetWithText(${widget.runtimeType}, '$descendantText')");
   	numberOfWithTexts += 1;
   	continue;
   	}
   	}
   	}
   	if (!_isPrivate(element.runtimeType)) {
   	final Iterable<Element> matches = find.byElementType(element.runtimeType).evaluate();
   	if (matches.length == 1) {
   	debugPrint(' find.byElementType(${element.runtimeType})');
   	continue;
   	}
   	}
   	totalNumber -= 1; // if we got here, we didn't actually find something to say about it
   	}
   	if (totalNumber == 0)
   	debugPrint(' <could not come up with any unique finders>');
   	}
   	}

   
   and will not go to widget listeners.
3. I don't think we can integrate test Pointer event resampler (#41118) #60558 without making significant change to the binding or to write a completely new binding. But new binding is not useful for the migration of flutter_driver to a e2e based host independent test because E2EWidgetsFlutterBinding is a subclass of LiveTestWidgetsFlutterBinding.
4. The other solution to test Pointer event resampler (#41118) #60558 (more like unit test in this case) is to have a PointerData input, push it through a stand-alone PointerDataResampler instance and convert the new PointerData to the PointerEvent, then inject it in the test app. This basically means we won't have some metric on the same benchmark test case, that can show improvement when introducing the resampling feature, like what we do for devicelab, because it's basically a different test app.
5. The above may be more meaningful when Pointer event resampler (#41118) #60558 settles to a public API. But for now I don't think I can plan for testing this.

With the above said, I'm reopening and ask for review for the current version.

@dnfield Instead of filing an issue I decided to implement it. Though I don't have a very good use case for now of using this method under LiveWidgetController, the major reason being LiveWidgetController is used in a very limited situation. But in terms of the feature it's working with a unit test. Would you mind reviewing it also, or if you still insist with leaving it unimplemented, I can revert the commit.

make it another PR.