This may not be a bug. It could be your programming. There is something I remember. When you create an Image with CreateImage() and you need mask/alpha channel, you have to fill it with the proper data first (quite a long time ago - I think I made the same mistake years ago until someone pointed me to it).
LoadImage does that automatic.
Means: Load a 24Bit Image and the bits 24-31 are filled with Ones ($FFrrggbb) - Creating the Image they are filled with Zeros ($00rrggbb)... check it out or wait until someone writes I am wrong ;-)
LoadImage does that automatic.
Means: Load a 24Bit Image and the bits 24-31 are filled with Ones ($FFrrggbb) - Creating the Image they are filled with Zeros ($00rrggbb)... check it out or wait until someone writes I am wrong ;-)