Unstable test FlipBooleanPropertyGefTest.test_doFlip

[ptziegler]

Can be reproduced locally when executing the test ~5 times in a row.

org.opentest4j.AssertionFailedError: expected: <package test;
import java.awt.*;
import java.awt.event.*;
import javax.swing.*;
import javax.swing.border.*;
public class Test extends JPanel {
	public Test() {
		setLayout(new BorderLayout());
		{
			MyPanel panel = new MyPanel();
			panel.setMyExpanded(true);
			add(panel);
		}
	}
}
> but was: <package test;
import java.awt.*;
import java.awt.event.*;
import javax.swing.*;
import javax.swing.border.*;
public class Test extends JPanel {
	public Test() {
		setLayout(new BorderLayout());
		{
			MyPanel panel = new MyPanel();
			add(panel);
		}
	}
}
>
	at org.junit.jupiter.api.AssertionFailureBuilder.build(AssertionFailureBuilder.java:151)
	at org.junit.jupiter.api.AssertionFailureBuilder.buildAndThrow(AssertionFailureBuilder.java:132)
	at org.junit.jupiter.api.AssertEquals.failNotEqual(AssertEquals.java:197)
	at org.junit.jupiter.api.AssertEquals.assertEquals(AssertEquals.java:182)
	at org.junit.jupiter.api.AssertEquals.assertEquals(AssertEquals.java:177)
	at org.junit.jupiter.api.Assertions.assertEquals(Assertions.java:1145)
	at org.eclipse.wb.tests.designer.core.AbstractJavaTest.assertEditor(AbstractJavaTest.java:273)
	at org.eclipse.wb.tests.designer.swing.SwingGefTest.assertEditor(SwingGefTest.java:115)
	at org.eclipse.wb.tests.designer.core.model.generic.FlipBooleanPropertyGefTest.test_doFlip2(FlipBooleanPropertyGefTest.java:67)
	at org.eclipse.wb.tests.designer.core.model.generic.FlipBooleanPropertyGefTest.test_doFlip(FlipBooleanPropertyGefTest.java:46)
	at java.base/java.lang.reflect.Method.invoke(Method.java:580)
	at org.eclipse.wb.tests.designer.core.model.parser.AbstractJavaInfoTest$JavaInfoExtension.interceptTestMethod(AbstractJavaInfoTest.java:85)
	at java.base/java.util.ArrayList.forEach(ArrayList.java:1596)
	at java.base/java.util.ArrayList.forEach(ArrayList.java:1596)

[ptziegler]

The problem is the same as always: The widget isn't moved in time, so the bounds of the JPanel are wrong, which causes the double-click event to be sent to (x=4830 y=9795).

[ptziegler]

The tests are much more stable when I comment out the calls to setLocation(10000, 10000), but they still fail occasionally. So I'm definitely on the right track with my assumption, that the window manager doesn't process the event in time. Perhaps I need to listen to the "component moved" event, rather than just the belief that the window will be moved instantaneously?

windowbuilder/org.eclipse.wb.swing/src/org/eclipse/wb/internal/swing/utils/SwingImageUtils.java

Lines 281 to 300 in 9dc76f8

	public static void prepareForPrinting(Component component) throws Exception {
	component.setLocation(10000, 10000);
	// don't grab focus during printing
	if (component instanceof Window window) {
	m_fosucableStates.put(window, window.getFocusableWindowState());
	window.setFocusableWindowState(false);
	}
	// make visible
	setVisible(component, true);
	{
	// workaround to prevent window from flashing if the Window Manager
	// doesn't allow the window to appear off-screen.
	if (component instanceof Window window) {
	window.toBack();
	// do the location change once again, because sometimes setLocation()
	// for invisible windows could be ignored.
	component.setLocation(10000, 10000);
	}
	}
	}

[ptziegler]

Sadly the parent model is not guaranteed to be the parent component in the Swing screen. Otherwise I could've simply used getLocation(), which is always relative to its (direct) parent. However, it should be possible to use a similar translation logic to what exist is Draw2D. The main goal has to be to avoid the conversion to and from screen coordinates, which is what's causing the problem.