Hi, using the current version 14.3.0 and SkiaSharp Version 3.116.1 I get an AccessViolationException when trying to load in a JPG file.
Is this a known issue?
Here is the stacktrace:
|>|SkiaSharp.dll!SkiaSharp.SkiaApi.sk_pixmap_scale_pixels(System.IntPtr cpixmap, System.IntPtr dst, SkiaSharp.SKSamplingOptions* sampling) Line 7461|C#|
| |SkiaSharp.dll!SkiaSharp.SKPixmap.ScalePixels(SkiaSharp.SKPixmap destination, SkiaSharp.SKSamplingOptions sampling) Line 216|C#|
| |SkiaSharp.dll!SkiaSharp.SKBitmap.ScalePixels(SkiaSharp.SKPixmap destination, SkiaSharp.SKSamplingOptions sampling) Line 714|C#|
| |SkiaSharp.dll!SkiaSharp.SKBitmap.ScalePixels(SkiaSharp.SKBitmap destination, SkiaSharp.SKSamplingOptions sampling) Line 709|C#|
| |ThinkGeo.Core.dll!8RQ=.HBU=.JBU=(object image, int width, int height)|Unknown|
| |ThinkGeo.Core.dll!ThinkGeo.Core.GeoImage.Clip(ThinkGeo.Core.DrawingRectangle clippingRectangle, int clippingWidth, int clippingHeight)|Unknown|
| |ThinkGeo.Core.dll!ThinkGeo.Core.NativeImageRasterSource.GetImageCore(ThinkGeo.Core.RectangleShape worldExtent, int canvasWidth, int canvasHeight)|Unknown|
| |ThinkGeo.Core.dll!ThinkGeo.Core.RasterSource.GetImage(ThinkGeo.Core.RectangleShape worldExtent, int canvasWidth, int canvasHeight)|Unknown|
| |ThinkGeo.Core.dll!ThinkGeo.Core.RasterLayer.DrawCore(ThinkGeo.Core.GeoCanvas canvas, System.Collections.ObjectModel.Collection<ThinkGeo.Core.SimpleCandidate> labelsInAllLayers)|Unknown|
| |ThinkGeo.Core.dll!ThinkGeo.Core.Layer.BCA=(ThinkGeo.Core.GeoCanvas canvas, System.Collections.ObjectModel.Collection<ThinkGeo.Core.SimpleCandidate> labelsInAllLayers)|Unknown|
| |ThinkGeo.Core.dll!ThinkGeo.Core.Layer.Draw(ThinkGeo.Core.GeoCanvas canvas, System.Collections.ObjectModel.Collection<ThinkGeo.Core.SimpleCandidate> labelsInAllLayers)|Unknown|
| |ThinkGeo.UI.Wpf.dll!ThinkGeo.UI.Wpf.LayerTileView.DrawLayerAsync.AnonymousMethod__0()|Unknown|
var layerOverlay = new LayerOverlay();
mapView.Overlays.Add(layerOverlay);
var geoTiffRasterLayer = new SkiaRasterLayer(@"d:\Downloads\Stadtkarte_farbig_75\Stadtkarte_farbig_75.jpg");
//var geoTiffRasterLayer = new WpfRasterLayer(@"d:\Downloads\Stadtkarte_farbig_75\Stadtkarte_farbig_75.jpg");
layerOverlay.Layers.Add(geoTiffRasterLayer);
geoTiffRasterLayer.Open();
mapView.CurrentExtent = geoTiffRasterLayer.GetBoundingBox();
_ = mapView.RefreshAsync();
Hi, thanks, are there any other special WPF layers that I should use? Or only this one?
It works with WpfRasterLayer. I also tried the test jpgs in the HowDoI Project and they work, for whatever reason only some jpgs don’t work. I sadly don’t have a code snippet as this is a WPF application from 2003 and extremely huge and convoluted. The general structure looks like this:
I have a special ProjMap : MapView which defines two overlays, ShapefileOverlay and ImageOverlay (image is the same as below)
public LayerOverlay ShapefileOverlay
{
get
{
if (!this.Overlays.Contains("ShapefileOverlay"))
{
this.Overlays.Add("ShapefileOverlay", new LayerOverlay());
((LayerOverlay)this.Overlays["ShapefileOverlay"]).DrawingQuality = DrawingQuality.HighSpeed;
}
return (LayerOverlay)this.Overlays["ShapefileOverlay"];
}
}
The layers I define as special Layers as I need meta information on, so we have special classes for this.
public ViewerImageLayer(IViewerLayer layer)
{
switch (layer.FileType)
{
case FileType.Tiff:
this.MapLayer = new GeoTiffRasterLayer(this.FilePath);
break;
case FileType.JPG:
case FileType.JPEG:
case FileType.PNG:
this.MapLayer = new WpfRasterLayer(this.FilePath);
break;
}
with
public interface IViewerLayer
{
ProjMap Map { set; }
Layer MapLayer { get; set; }
... other meta info
Again, this is an extremely old application from 2003 which used a legacy ThinkGeo version, so if there’s anything obvious that is necessary / unnecessary I’d appreciate it
E.g. is it good practice to separate Shapes and Images in two different Layers? Or is that unnecessary and better to keep them in one?
We have 3 Layers in total, one for Shapes, one for Images, and one for Temporary Highlights. We are increasing the amount of Layer Types that can be displayed (so adding WPF, WMS, File Geodatabases etc.) - how would the best way of structuring the Overlays be? Or just render everything in one & then the Temporary Highlights in the other?
In ThinkGeo.Core, the built‑in RasterLayer uses SkiaSharp and works cross‑platform.
WpfRasterLayer is the only “special WPF layer” we have at this time, it’s desktop‑only (WPF/WinForms) and have better performance.
About updating to the newer version
My suggestion is to rebuild your project and address any ThinkGeo warnings. Those deprecation hints will point you to the new APIs in the latest ThinkGeo release.
Overlay strategy
Q: Is it good practice to separate Shapes and Images in two different Layers? Or is that unnecessary and better to keep them in one?
A: I think you meant Overlays. They should be two different layers anyway, but whether we should put them into one or two overlays.
When to combine (one overlay):
Layers always drawn and refreshed together (e.g. shapes and images are both for your background so every time you pan/zoom around the map, they will be refreshed at the same time).
When to separate (multiple overlays):
Layers that update at different times (e.g. highlights, so you need to refresh the features in the highlight layers without redrawing the background layers).
Suggested overlay structure for your case
You can put highlight layer to a separate overlay and all the other layers to one overlay. Because all the other layers are always refreshed at the same time, I suppose.
You can consider putting WMS layer into its own overlay (or just use WmsOverlay instead) because WMS might be slow requesting the tiles, you don’t want it show down the other layers. Put it into a separate overlay so even the wms is slow, the other background layers will also be rendered.
