I spent all day today waffling over graphics options. Whether to use OpenFL rendering or AIR, the display list or direct blitting or tile drawing, how to handle filters, how to handle post-processing, and so on and so on and so on.
I made very little in the way of concrete progress.
I did eventually come to a decision, though.
Here’s what I’m going to do: I’m going to put the C++ native platform down for now, as intriguing as the possibilities it offers are, and return to my original approach of using Flash/AIR’s quick direct pixel rendering code to push a bunch of cached high-res images, but taking advantage of Haxe’s performance gains. I’m also going to start looking into the idea of constructing an external process to handle post-processing each frame far faster than Flash can handle: Whether I achieve this using the CPU or the GPU will be determined when I start developing it. The GPU can do these per-pixel processes more efficiently, but it takes time to load the texture in which can obviate the performance gain. I’ll have to experiment to see which works better.
At least, now, I know where I’m going with this. Today was kind of frustrating, but now I can move forward and start addressing the damn bugs that are still outstanding tomorrow.