[Win32] Make Control bounds point/pixel conversion cope with rounding

[HeikoKlare]

The invertibility of point/pixel conversions is limited as point values are int-based and with lower resolution than pixel values. In consequence, values need to be rounded when converted between the two, which inevitably leads to rounded values that do not fit for every use case. This adds test cases that demonstrate such use cases, including simple parent/child scenarios, in which the child is supposed to fill the parent, and including layouting scenarios incorporating the client area of a composite, and how the current implementation is not capable of producing proper results for them.

This change also adapts the methods for setting bounds/size of controls to deal with the limited invertibility:

1. They transform the passed bounds into a "global" coordinate system (the one of the containing shell) such that all values are rounded in the same way. Otherwise, every control uses a coordinate system relative to its parent, leading to the same global point coordinates being transformed to different pixel coordinates, such that parent and child may not fit to each other even though their point coordinates perfectly match.
2. They shrink the calculated pixel values by at most the maximum error that can be made when transforming from point to pixel values, such that rounding errors due to layouts that calculated control bounds based on a composites client area are evened out. Without that, layouted controls may be up to one point too large to fit into the composite.

Examples

The different issues are demonstrated with the following snippet (conforms to the test cases):

public static void main(String[] args) {
	System.setProperty("swt.autoScale.updateOnRuntime", "true");
	Display display = new Display();
	Shell shell = new Shell(display);
	shell.setSize(200, 200);
	int width = 11;
	int height = 11;

	Composite c1 = new Composite(shell, SWT.NONE);
	c1.addPaintListener(e -> {
		e.gc.setBackground(display.getSystemColor(SWT.COLOR_BLUE));
		e.gc.fillRectangle(0, 0, c1.getSize().x + 5, c1.getSize().x + 5);
	});
	c1.setBounds(11, 9, width, height);
	Label l1 = new Label(c1, SWT.BORDER);
	l1.setBounds(0, 0, width, height);

	Composite c2 = new Composite(shell, SWT.NONE);
	c2.addPaintListener(e -> {
		e.gc.setBackground(display.getSystemColor(SWT.COLOR_BLUE));
		e.gc.fillRectangle(0, 0, c2.getSize().x + 5, c2.getSize().x + 5);
	});
	c2.setBounds(10, 30, width, height);
	Label l2 = new Label(c2, SWT.BORDER);
	l2.setBounds(0, 0, width, height);

	width += 2;
	height += 2;
	Composite c3 = new Composite(shell, SWT.NONE);
	c3.addPaintListener(e -> {
		e.gc.setBackground(display.getSystemColor(SWT.COLOR_BLUE));
		e.gc.fillRectangle(0, 0, c3.getSize().x + 5, c3.getSize().y + 5);
	});
	c3.setBounds(13, 53, width, height);
	Label l3 = new Label(c3, SWT.BORDER);
	l3.setBounds(0, 0, width, height);

	Composite c4 = new Composite(shell, SWT.NONE);
	c4.addPaintListener(e -> {
		e.gc.setBackground(display.getSystemColor(SWT.COLOR_BLUE));
		e.gc.fillRectangle(0, 0, c4.getSize().x + 5, c4.getSize().y + 5);
	});
	GridLayout layout = new GridLayout(2, true);
	layout.marginBottom = 0;
	layout.marginTop = 0;
	layout.marginLeft = 0;
	layout.marginRight = 0;
	layout.marginHeight= 0;
	layout.marginWidth= 0;
	layout.horizontalSpacing = 0;
	c4.setLayout(layout);
	c4.setBounds(13, 100, 101, 50);
	System.out.println(c4.getBounds());
	System.out.println(c4.getClientArea());
	Label l4a = new Label(c4, SWT.BORDER);
	l4a.setBackground(display.getSystemColor(SWT.COLOR_GREEN));
	l4a.setLayoutData(new GridData(GridData.FILL_BOTH));
	Label l4b = new Label(c4, SWT.BORDER);
	l4b.setBackground(display.getSystemColor(SWT.COLOR_GREEN));
	l4b.setLayoutData(new GridData(GridData.FILL_BOTH));
	c4.layout();

	shell.open();
	while (!shell.isDisposed()) {
		if (!display.readAndDispatch())
			display.sleep();
	}

	display.dispose();
}

Before at 125%:

After at 125%:

You can easily see the behavior in layouted controls (particularly using grid layouts), such as controls based on Forms UI:

Before at 125%:

After at 125%:

[github-actions]

Test Results

  115 files   -  3    115 suites   - 3   11m 11s ⏱️ - 7m 0s
4 618 tests  - 34  4 601 ✅  - 32  17 💤  - 2  0 ❌ ±0 
  323 runs   - 15    320 ✅  - 14   3 💤  - 1  0 ❌ ±0 

Results for commit 3924200. ± Comparison against base commit 0b21ff2.

This pull request removes 37 and adds 3 tests. Note that renamed tests count towards both.

AllGTKTests Test_GtkConverter ‑ test_HeuristicASCII_dollarSign
AllGTKTests Test_GtkConverter ‑ test_HeuristicASCII_emptyString
AllGTKTests Test_GtkConverter ‑ test_HeuristicASCII_letterA
AllGTKTests Test_GtkConverter ‑ test_HeuristicASCII_letters
AllGTKTests Test_GtkConverter ‑ test_HeuristicUTF16LE_null
AllGTKTests Test_GtkConverter ‑ test_HeuristicUTF16_AsciiLetters
AllGTKTests Test_GtkConverter ‑ test_HeuristicUTF16_Asciiletter
AllGTKTests Test_GtkConverter ‑ test_HeuristicUTF16_LotsOfLetters
AllGTKTests Test_GtkConverter ‑ test_HeuristicUTF16_letter
AllGTKTests Test_GtkConverter ‑ test_HeuristicUTF16_letters
…

org.eclipse.swt.widgets.ControlWin32Tests ‑ testChildFillsCompositeWithOffset
org.eclipse.swt.widgets.ControlWin32Tests ‑ testChildFillsScrollableWithBadlyRoundedClientArea
org.eclipse.swt.widgets.ControlWin32Tests ‑ testChildWithOffsetFillsComposite

This pull request skips 1 and un-skips 1 tests.

org.eclipse.swt.tests.junit.Test_org_eclipse_swt_widgets_Shell ‑ test_activateEventSend

org.eclipse.swt.tests.junit.Test_org_eclipse_swt_browser_Browser ‑ test_setUrl_remote_with_post

[amartya4256]

I tested this branch and I found out that it reintroduces the flickering CTabFolder border line issue. I tested at 150 and 250% and was able to reproduce the flickering behaviour on resizing the CTabFolder by dragging the sides in the Sashlayout.

[HeikoKlare]

Thank you, @amartya4256! I am still analyzing the issue you reported. All values (bounds in points and pixels of involved controls) look fine, but still something is not fitting.

Since this PR combines two improvements anyway, I have extracted the one part that does not seem to break anything (and the part that is more important anyway) to a separate PR, so that we can at least move on with that:

• [Win32] Fix control bounds exceeding parent size #2849

[github-actions]

Test Results (win32)

 26 files   -     8   26 suites   - 8   10s ⏱️ - 4m 36s
 87 tests  - 4 549   86 ✅  - 4 477  0 💤  - 73  1 ❌ +1 
113 runs   -    57  112 ✅  -    55  0 💤  -  3  1 ❌ +1 

For more details on these failures, see this check.

Results for commit 3c3a6fb. ± Comparison against base commit 558c4e6.

This pull request removes 4549 tests.

AllWin32Tests ImageWin32Tests ‑ testDisposeDrawnImageBeforeRequestingTargetForOtherZoom
AllWin32Tests ImageWin32Tests ‑ testDrawImageAtDifferentZooms(boolean)[1] true
AllWin32Tests ImageWin32Tests ‑ testDrawImageAtDifferentZooms(boolean)[2] false
AllWin32Tests ImageWin32Tests ‑ testImageDataForDifferentFractionalZoomsShouldBeDifferent
AllWin32Tests ImageWin32Tests ‑ testImageShouldHaveDimesionAsPerZoomLevel
AllWin32Tests ImageWin32Tests ‑ testRetrieveImageDataAtDifferentZooms(boolean)[1] true
AllWin32Tests ImageWin32Tests ‑ testRetrieveImageDataAtDifferentZooms(boolean)[2] false
AllWin32Tests ImageWin32Tests ‑ test_getImageData_fromCopiedImage
AllWin32Tests ImageWin32Tests ‑ test_getImageData_fromImageForImageDataFromImage
AllWin32Tests TestTreeColumn ‑ test_ColumnOrder
…

[HeikoKlare]

Closing this as it did not prove any benefits compared to the already applied improvements for handling rounding issues in control bounds.