I'm amazed at the sheer complexity of the 'workarounds' proposed here ('injecting DLL'...?????), as if melonDS was a closed-source black-box thing. reminds me of the NO$GBA/NO$ZOOMER situation, in a way.

also, it IS coming in 0.9. I coded the AR engine already. we just need to make the interface for it.

____________________
Kuribo64

probably something to do with that specific audio interface/driver and how often/when it's telling SDL to request more audio samples. basically, melonDS's audio sync is tied to the SDL audio stream callback, so if that's called at an inconsistent enough rate, it might affect how stable the emu runs.

____________________
Kuribo64

do I need a savefile to get to the crash point?

well, I find a crash point in the very beginning. from what I observe, it started breaking in melonDS 0.7.3.

gonna backtrack it to find the exact commit that broke it.

____________________
Kuribo64

related Github issue thread: https://github.com/Arisotura/melonDS/issues/359

____________________
Kuribo64

basically you made a WiiU gamepad

____________________
Kuribo64

this is pointless, we already have the whole forum for this, plus the issue tracker on Github.

also, we don't need a 'throw random ideas at the dev team' thread. things take time and effort to code, you know.

____________________
Kuribo64

well, good catch, I'll look into it

____________________
Kuribo64

it's weird because according to the code that shouldn't happen

____________________
Kuribo64

nah, that's fine; the point is to make sure the emu thread is paused before loading or saving a state, so we don't enter some race condition fucking things up

I'll look deeper into it, anyway

____________________
Kuribo64

this bug should be fixed now

____________________
Kuribo64

it can also help to disable the framerate limiter, vsync, audio sync, let both melonDS instances run as fast as possible

I'll grant that it's not too convenient as far as gameplay is concerned, tho

____________________
Kuribo64

the glitch is known and it's a timing issue. timings are something we need to work on, but it'll take a while.

____________________
Kuribo64

thanks!

well, we have an experimental branch that fixes some of these timing issues, we still need to adapt it to the JIT tho

____________________
Kuribo64

why do you need to constantly make these big overcomplicated posts with no linebreaks? it's kind of painful to read.

____________________
Kuribo64

well ok

____________________
Kuribo64