hello all

I find when call the function UDPointType, the GDI object will increase (because i will change the image with the map ExtentsChanged,so i call the UDPointType function frequently) . i find up the source code in MapWinGIS.ocx. the function define as follows: in the source code i find GetDC but not ReleaseDC.

	void CMapView::SetUDPointType(long LayerHandle, LPDISPATCH newValue)
	{
		Layer * l = get_ShapefileLayer(LayerHandle);
		if (l)
		{
			IImage * iimg = NULL;
			newValue->QueryInterface( IID_IImage, (void**)&iimg );

			if( iimg != NULL )
			{	ShapeLayerInfo * sli = (ShapeLayerInfo*)l->addInfo;
				if( sli->udPointType != NULL )
					sli->udPointType->Release();
				sli->udPointType = iimg;

				if( sli->udDC != NULL )
					sli->udDC->DeleteDC();
				sli->udDC=NULL;

				iimg->get_Width(&(sli->udPointTypeWidth));
				iimg->get_Height(&(sli->udPointTypeHeight));

				if( sli->udPointTypeWidth > 0 && sli->udPointTypeHeight > 0 )
				{
					CBitmap * bmp = new CBitmap();

					bmp->CreateDiscardableBitmap(GetDC(),sli->udPointTypeWidth,sli->udPointTypeHeight);
					sli->udDC = new CDC();
					sli->udDC->CreateCompatibleDC(GetDC());
					
					sli->udDC->SelectObject(bmp);

					long pad=sli->udPointTypeWidth*24;
					pad%=32;
					if(pad!=0)
					{	pad=32-pad;
						pad/=8;
					}

					if( sli->udPointTypeImage != NULL )
						delete [] sli->udPointTypeImage;
					sli->udPointTypeImage = new unsigned char[(sli->udPointTypeWidth*3+pad)*sli->udPointTypeHeight];

					long rowLength = sli->udPointTypeWidth*3 + pad;
					VARIANT_BOOL useImageTransparencyColor = FALSE;
					iimg->get_UseTransparencyColor(&useImageTransparencyColor);
					OLE_COLOR imageTransparencyColor = RGB(0,0,0);
					iimg->get_TransparencyColor(&imageTransparencyColor);
					sli->udTransColor=imageTransparencyColor;

					register int j, i;
					for( j = 0; j < sli->udPointTypeHeight; j++ )
					{	
						long * row = new long[sli->udPointTypeWidth+1];
						VARIANT_BOOL retval;
						iimg->GetRow( j, row, &retval );
						
						for( i = 0; i < sli->udPointTypeWidth; i++ )
						{	
							sli->udPointTypeImage[(sli->udPointTypeHeight-j-1)*rowLength+i*3] = GetBValue( row );
							sli->udPointTypeImage[(sli->udPointTypeHeight-j-1)*rowLength+i*3+1] = GetGValue( row );
							sli->udPointTypeImage[(sli->udPointTypeHeight-j-1)*rowLength+i*3+2] = GetRValue( row );
						}
						if( row != NULL )
							delete [] row;
					}

					BITMAPINFO bif;
					BITMAPINFOHEADER bih;
					bih.biBitCount=24;
					bih.biWidth=sli->udPointTypeWidth;
					bih.biHeight=sli->udPointTypeHeight;
					bih.biPlanes=1;
					bih.biSize=sizeof(BITMAPINFOHEADER);
					bih.biCompression=0;
					bih.biXPelsPerMeter=0; 
					bih.biYPelsPerMeter=0;
					bih.biClrUsed=0;
					bih.biClrImportant=0;
					bih.biSizeImage=(sli->udPointTypeWidth*3+pad)*sli->udPointTypeHeight;
					bif.bmiHeader = bih;

					SetDIBitsToDevice( sli->udDC->m_hDC,0,0,sli->udPointTypeWidth,sli->udPointTypeHeight,
									   0,0,0,sli->udPointTypeHeight,sli->udPointTypeImage, &bif, DIB_RGB_COLORS );
				}
				else
				{	if( sli->udPointTypeImage != NULL )
						delete [] sli->udPointTypeImage;
					sli->udPointTypeImage = NULL;
					sli->udPointTypeWidth = 0;
					sli->udPointTypeHeight = 0;
				}
			}
			else
				ErrorMessage(tkINTERFACE_NOT_SUPPORTED);
		}
	}

thanks

I'm also confused by the codes.

I think ReleaseDC is necessary for GetDC.
And maybe "delete sli->udDC;" is also needed for "sli->udDC = new CDC();".

Besides, I found some other similar strange codes:

virtual ~ShapeLayerInfo()
{
...
...
if( udDC )
{ CBitmap * newbmp = new CBitmap();
CBitmap * bmp = udDC->SelectObject(newbmp);
if( bmp )
bmp->DeleteObject();
udDC->DeleteDC();
}
}

The above codes look not so good.
"delete bmp;" and "delete udDC" are also needed here.
And then, where will newbmp go?

I hope they are not bugs.



Edited 1 time(s). Last edit at 10/08/2010 12:45AM by Robin Liu.