undo push-model SDL2 sound backend, use SDL1 approach
commitd096885b2b71000fd7292eb941eabeb19a44ab3e
authorrofl0r <rofl0r@users.noreply.github.com>
Sun, 29 Aug 2021 02:17:09 +0000 (29 02:17 +0000)
committerrofl0r <rofl0r@users.noreply.github.com>
Sun, 29 Aug 2021 08:50:13 +0000 (29 08:50 +0000)
tree4f3ff19fd5bb3f737a99d62c24bb711e46c751b5
parenta24b9df4f9853de378cca7ff466803e03f9a6234
undo push-model SDL2 sound backend, use SDL1 approach

while theoretically nicer, the push model causes to get sound
out of sync with the emulator; apparently the emulator needs to
wait until the buffer is actually consumed.
i'll look into fixing that, but until that's done use the SDL1
approach. it's also helpful to have this in commit history
so one can easily go to the SDL1 style sound and compare.
sys/sdl/sdl-audio.c
sys/sdl2/sdl-audio.c