It goes something like: use SDL_SetRenderTarget() to divert drawing operations to an intermediate GPU-based SDL_Texture at your preferred size, rather than to the output window. When you've finished doing all your drawing of the current frame to this GPU texture, you call SDL_SetRenderTarget(NULL) to resume drawing to the window, and then use SDL_RenderCopy() to scale and draw your intermediate texture to fit the output window. I think.At some point I want to try to port the graphics to SDL 2, but as I don't know it at all yet that's going to be a tall order. However, it does run through the sdl-compat library which uses SDL2 under the lid, and through this it certainly does fullscreen mode much better than "real" SDL 1.2.
I can't remember any concrete details about thread safety, other than MacOS was the most twitchy of the desktop OSes w.r.t. multithreaded SDL2 graphics routines. I think I resorted to doing all graphics operations in a single CPU thread.
Statistics: Posted by Diminished — Thu Dec 19, 2024 7:41 pm