According to AddressSanitizer, when `sandbox_malloc` causes the
WebAssembly memory to grow in size, every single coroutine on the
sandbox stack gets corrupted. So if `sandbox_malloc` is going to cause
the memory to grow in size, we need to yield so that there are no
coroutines on the sandbox stack while the reallocation occurs.
The Emscripten build doesn't support stack protection. Besides, if we
really wanted to enable fortification and stack protection, we would
enable it for the entire application instead of enabling it for only
these two libraries.
Now it actually stops the audio instead of doing nothing.
Also, calling `Audio.bgm_stop` was causing a crash in Emscripten due to
an exception being thrown internally in OpenAL Soft and C++ exceptions
being disabled when targeting Emscripten. Shouldn't happen anymore.
This way Meson will automatically keep the Git hash updated as you make
new Git commits, rather than forcing you to reconfigure the build again
every time you make a new commit to keep the commit hash updated.
This allows more flexibility when loading games in libretro builds,
since we can now load games either from a directory or from a ZIP or 7Z
archive. Also, the path cache is now active for all filesystem calls
made from inside Ruby.
I don't think I'll be needing SDL. Pretty sure about 100% of the
functionality of SDL that's currently being used here is either trivial
to reimplement or needs to be replaced for the libretro core to work in
the first place. For example, I already know I'm going to need to
completely replace the SDL input handling with libretro's input API.
Also, I'll need to get rid of the use of SDL threads in the audio code
later because libretro cores need to render audio synchronously with the
video rendering.
I'm including the SDL headers, though, to reduce the number of
compilation errors.
