Andre Weissflog

1881 isn't a multiple of any 2^N number, but it's a multiple of 3: 3*627 = 1881. But 627 also isn't a particularly "round number". This is very very odd.

Aha, interesting! When you try to initialize with sample rate 44100, does the ALSA initialization change this to 48000 (e.g. ```sample_rate before: 44100``` and ```sample_rate after: 48000```? (44100 / 48000)...

Hmm ok, I've been reading up a bit more on how ALSA setup-configuration works, and I did a bit more code cleanup as result, but I don't think this fixes...

...which is also strange because I would except that resampling is a good thing if a requested sample rate isn't supported. But maybe this is what enables picking a nearby...

Hmm this sounds like there's still some disagreement about sample rates and/or buffer sizes, maybe it's just the returned sample rate that's wrong. I really should try to get my...

...I wonder if the best short-term solution is to simply set the default sample rate in sokol_audio.h to 48kHz. E.g. if "native" 48kHz is more widely supported than 44.1kHz, this...

...another more robust solution is to rewrite the ring buffer code so that the audio streaming buffer size doesn't need to be a multiple of the ring-buffer packet size... I...

Hmm I wonder why I essentially undid that commit in the followup commit here: https://github.com/floooh/sokol/commit/5b5d9a7d2c2c7b2ec350e5ba16767527404ed2dc I really need to write more detailed commit messages for such things :/ > Disabling...

Hmm, that sounds like a weird use case tbh... wouldn't it be better to disable the aiming mechanism (and visualization) until the first 'proper' mouse event is received? From googling...

PS: please ignore the broken CI build, that's a problem with the sokol-d build.zig not being fixed for zig 0.14.x