class VclPixelProcessor2DTest : public test::BootstrapFixture
{ // if enabled - check the result images with: // "xdg-open ./workdir/CppunitTest/drawinglayer_processors.test.core/" static constexpr constbool mbExportBitmap = false;
// Test that drawing only a part of a gradient draws the proper part of it. void testTdf139000()
{
ScopedVclPtr<VirtualDevice> device
= VclPtr<VirtualDevice>::Create(DeviceFormat::WITHOUT_ALPHA);
device->SetOutputSizePixel(Size(100, 200));
device->SetBackground(Wallpaper(COL_RED));
device->Erase();
// I stumbled over this when hunting another problem, but I have to correct // this: This test does test something that is not supported. It seems to be // based on the *misunderstanding* that in the version of the constructor of // FillGradientPrimitive2D (and similar others) with two ranges the 2nd // B2DRange parameter 'OutputRange' is a 'clipping' parameter. This is *not* // the case --- it is in fact the *contrary*, it is there to *extend* the // usual definition/paintRange of a gradient: // It was originally needed to correctly display TextFrames (TF) in Writer: If you // have a TF in SW filled with a gradient and that TF has sub-frames, it inherits // the gradient fill. Since you can freely move those sub-TFs even outside the // parent TF there has to be a way to not only paint gradients in their definition // range (classical, all DrawObjects do that), but extended from that. This is // needed e.g. for linear gradients, but - dependent of e.g. the center settings - // also for all other ones, all can have geometry 'outside' the DefinitionRange. // This is now also used in various other locations which is proof that this is // useful and needed. It is possible to see that basic history/reason for this // parameter by following the git history and why and under which circumstances // that parameter was originally added. Other hints are: It is *not* named // 'ClipRange'. Using a B2DRange to define a ClipRange topology would be bad due // to not being transformable, a PolyPolygon would be used in that case. Using as // clipping mechanism would offer a 2nd principle to add clipping for primitives // besides MaskPrimitive2D - always bad style in a sub-system. A quick look // on it's usages gives hints, too. // This means that when defining an outputRange that resides completely *inside* // the definitionRange *no change* at all is done by definition since this does // not *extend* the target area of the gradient paint region at all. If an // implementation does clip and limit output to 'outputRange' that should do no // harm, but is not the expected/reliable way to paint primitives clipped. // That's why all DrawObjects with gradient fill (and other fills do the same) // embed the fill that is defined for a range (usually the BoundRange of a // PolyPolygon) in a MaskPrimitive2D defined by the outline PolyPolygon of the // shape. Nothing speaks against renderers detecting that combination and do // something optimized if they want to, especially SDPRs, but this is not // required. The standard embedded clipping of the implementations of the // MaskPrimitive2D do the right thing. // This test intends to paint the lower part of a gradient, so define the // gradient for the full target range and embed it to a MaskPrimitive2D // defining the lower part of that area to do that.
basegfx::B2DRange definitionRange(0, 0, 100, 200);
basegfx::B2DRange outputRange(0, 100, 100, 200); // Paint only lower half of the gradient.
exportDevice(u"test-tdf139000.png"_ustr, device);
Bitmap bitmap = device->GetBitmap(Point(), device->GetOutputSizePixel());
BitmapScopedReadAccess access(bitmap); // The upper half should keep its red background color.
CPPUNIT_ASSERT_EQUAL(BitmapColor(COL_RED), access->GetColor(Point(0, 99))); // First line of the gradient should not be the start color, but something halfway.
CPPUNIT_ASSERT_LESS(static_cast<sal_uInt16>(16),
access->GetColor(Point(0, 100)).GetColorError(COL_GRAY)); // Last line of the gradient should be the end color, or close.
CPPUNIT_ASSERT_LESS(static_cast<sal_uInt16>(16),
access->GetColor(Point(0, 199)).GetColorError(COL_BLACK));
}
Die Informationen auf dieser Webseite wurden
nach bestem Wissen sorgfältig zusammengestellt. Es wird jedoch weder Vollständigkeit, noch Richtigkeit,
noch Qualität der bereit gestellten Informationen zugesichert.
Bemerkung:
Die farbliche Syntaxdarstellung und die Messung sind noch experimentell.