For a couple of days now I've tried to figure out why my nine-slice code does not work as expected. As far as I can see, there seems to be an issue with the Graphics.DrawImage method which handles my nine slice images incorrectly. So my problem is how to compensate for the incorrect scaling that is performed when running my code on the compact framework. I might add that this code of course works perfectly when running in the full framework environment. The problem only occurs when scaling the image to a larger image not the other way around. Here is the snippet:
public class NineSliceBitmapSnippet
{
private Bitmap m_OriginalBitmap;
public int CornerLength { get; set; }
/// <summary>
/// Initializes a new instance of the NineSliceBitmapSnippet class.
/// </summary>
public NineSliceBitmapSnippet(Bitmap bitmap)
{
CornerLength = 5;
m_OriginalBitmap = bitmap;
}
public Bitmap ScaleSingleBitmap(Size size)
{
Bitmap scaledBitmap = new Bitmap(size.Width, size.Height);
int[] horizontalTargetSlices = Slice(size.Width);
int[] verticalTargetSlices = Slice(size.Height);
int[] horizontalSourceSlices = Slice(m_OriginalBitmap.Width);
int[] verticalSourceSlices = Slice(m_OriginalBitmap.Height);
using (Graphics graphics = Graphics.FromImage(scaledBitmap))
{
using (Brush brush = new SolidBrush(Color.Fuchsia))
{
graphics.FillRectangle(brush, new Rectangle(0, 0, size.Width, size.Height));
}
int horizontalTargetOffset = 0;
int verticalTargetOffset = 0;
int horizontalSourceOffset = 0;
int verticalSourceOffset = 0;
for (int x = 0; x < horizontalTargetSlices.Length; x++)
{
verticalTargetOffset = 0;
verticalSourceOffset = 0;
for (int y = 0; y < verticalTargetSlices.Length; y++)
{
Rectangle destination = new Rectangle(horizontalTargetOffset, verticalTargetOffset, horizontalTargetSlices[x], verticalTargetSlices[y]);
Rectangle source = new Rectangle(horizontalSourceOffset, verticalSourceOffset, horizontalSourceSlices[x], verticalSourceSlices[y]);
graphics.DrawImage(m_OriginalBitmap, destination, source, GraphicsUnit.Pixel);
verticalTargetOffset += verticalTargetSlices[y];
verticalSourceOffset += verticalSourceSlices[y];
}
horizontalTargetOffset += horizontalTargetSlices[x];
horizontalSourceOffset += horizontalSourceSlices[x];
}
}
return scaledBitmap;
}
public int[] Slice(int length)
{
int cornerLength = CornerLength;
if (length <= (cornerLength * 2))
throw new Exception("Image to small for sliceing up");
int[] slices = new int[3];
slices[0] = cornerLength;
slices[1] = length - (2 * cornerLength);
slices[2] = cornerLength;
return slices;
}
}
So, my question is, does anybody now how I could compensate the incorrect scaling?
/Dan
After some more trial and error I've finally found a solution to my problem. The scaling problems has always been to the top-center, right-center, bottom-center and left-center slices since they're always stretched in only one direction according to the logic of nine slice scaling. If I apply a temporarely square stretch to those slices before applying the correct stretch the final bitmap will be correct. Once again the problem is only visible in the .Net Compact Framework of a Windows CE device (Smart Device). Here's a snippet with code adjusting for the bug in CF. My only concern now is that the slices that get square stretched will take much more memory due to the correction code. On the other hand this step is only a short period of time so I might get away with it. ;)