Özkan Sezer

Suggest rebasing and retesting especially after eae0b80c095fd6171e524bcca86c4627f547c6c2 which fixed incompatibility between dynapi and main body .

The `lstrcpyA` and `wsprintfA` dependencies are easily overcome by a local stpcpy() and itoa() implementation, and since those two are universal and standalone they work for everything: Patch below, OK...

> Yep, seems fine to me. Applied. (And it applies almost without any changes to sdl1.2-compat too: Should I?) Now `MessageBoxA`, and possibly `LoadLibraryA`, seem to be the only problematic...

Both MessageBoxW and LoadLibraryW are documented to be "destop apps only", so I can't see how that would help.

Maybe something like this? (Disclaimer: Not even compile tested, I don't do winrt or c++ and the cmake'ry is possibly bullsh?t: @icculus and @madebr to rescue.) [test.zip](https://github.com/libsdl-org/sdl2-compat/files/12301033/test.zip)

> > Maybe something like this? > > I think yes, but we probably need _someone_ to take the time to compile testsprite2 for UWP and see what the state...

OK, pushed to git main.

> With these 2 patches ([cmake](https://github.com/madebr/sdl2-compat/commit/5b0cba3cb1f695b94ca62aaf6933b33f04107095) + [ci](https://github.com/madebr/sdl2-compat/commit/b149921295721a3e45c28c3f092a36c98b8a54b0)), I can get SDL2-compat compiling, but not linking. Those changes should be good I guess and can go in > The linking...

> I know you aren't supposed to do `LoadLibraryA` in `DllMain`, but this works... Are you doing LoadLibraryA in your patch? Or am I looking at the wrong place? >...

> > > I know you aren't supposed to do LoadLibraryA in DllMain, but this works... > > Are you doing LoadLibraryA in your patch? Or am I looking at...