Allow creating DIBs storing pixels in non-premultiplied format in wxMSW.

wxDIB assumed that MSW always uses bitmaps in pre-multiplied pixel format but
this turns out to be wrong: some Windows functions such as ImageList_Draw()
or CreateIconIndirect() apply pre-multiplication internally and so must be
given data in non-pre-multiplied format on input.

This commit adds a possibility to create wxDIB in such format but doesn't use
it anywhere yet, this will be done in subsequent commits.

See #9050.

git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@65959 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775

include/wx/msw/dib.h

@@ -138,12 +138,28 @@ public:
     // ------------------
 
 #if wxUSE_IMAGE
-    // create a DIB from the given image, the DIB will be either 24 or 32 (if
-    // the image has alpha channel) bpp
-    wxDIB(const wxImage& image) { Init(); (void)Create(image); }
+    // Possible formats for DIBs created by the functions below.
+    enum PixelFormat
+    {
+        PixelFormat_PreMultiplied = 0,
+        PixelFormat_NotPreMultiplied = 1
+    };
+
+    // Create a DIB from the given image, the DIB will be either 24 or 32 (if
+    // the image has alpha channel) bpp.
+    //
+    // By default the DIB stores pixel data in pre-multiplied format so that it
+    // can be used with ::AlphaBlend() but it is also possible to disable
+    // pre-multiplication for the DIB to be usable with ImageList_Draw() which
+    // does pre-multiplication internally.
+    wxDIB(const wxImage& image, PixelFormat pf = PixelFormat_PreMultiplied)
+    {
+        Init();
+        (void)Create(image, pf);
+    }
 
     // same as the above ctor but with the return code
-    bool Create(const wxImage& image);
+    bool Create(const wxImage& image, PixelFormat pf = PixelFormat_PreMultiplied);
 
     // create wxImage having the same data as this DIB
     wxImage ConvertToImage() const;