Ryan C. Gordon
Ryan C. Gordon
I don't think we're going to do this; we don't promise any sort of consistency in callback frequency (and even outside of vsync, we assume some platforms will refuse to...
Actually, let me reopen this and clarify.
So we have the hint (SDL_HINT_MAIN_CALLBACK_RATE) that controls this, but it is ignored if there's a window created...the original use was for things like loopwave.c that are just burning CPU...
Okay, this needs Apple Pencil still (it's the only "pen" you can use with iPads, and not on iPhones at the moment, so basically any pen in UIKit should be...
Oh! macOS! There's currently no such thing as a Mac with a touchscreen at all--and for the foreseeable future this will be the domain of iOS, right?!--so maybe we mark...
Of course, hahaha. We'll leave Cocoa as "unknown" for now, then.
This is resolved in SDL3_mixer.
That specific assert was to check that locking a mutex can _never_ fail, but as we just saw, one can destroy a mutex and use the free'd memory to trigger...
My thinking is this specific assert could just change to... ```c SDL_assert((rc == 0) && "pthread_mutex_lock failed! Did you accidentally destroy this SDL_Mutex?"); ``` ...if we're keeping it.
The info lives in a global linked list once an assert fires, for later recovery by the program, and lives as static memory at each use of SDL_assert, so it...