Questions (and then sleep)

The whole idea of a shadow buffer seems fairly fragile.
It tends to break down easily, especially when you’re mapping it, and
using it to emulate palette changes and so on.

Since the SDL surface mapping is fairly robust and blit works well,
how much is the shadow surface really needed?

When you’re talking about speed, you need direct access to the framebuffer,
so you’ll always want to have code that can write to the native screen

When you’re talking about convenience, it is really nice to have a shadow
buffer that emulates the display you want. Should I move the shadow buffer
code out into the public API so people who want the convenience can create
it, but I don’t have to worry about automatically creating it and keeping
it transparent?

Essentially, it’s just a big surface that’s mapped to the real screen.

Your thoughts folks?

See ya!
-Sam Lantinga (slouken at–
Author of Simple DirectMedia Layer -