@r-type Would like to sync with you on Atari800

Libretro Cores
#1

Hi r-type, I’ve forked libretro-atari800, and did a whole series of changes that I’d like to sync with you on, to make things a lot better, namely:

  • new menu options to set system type, NTSC/PAL, Internal BASIC on/off, SIO acceleration on/off, Hi-Res artifacting on/off, etc.
  • addition of ATX as a supported format (to play copy protected disks)

and I’m currently trying to get netplay working (serialization works, but I need to talk with you on how best to handle mutex locking with the emulator thread, as currently rewind and netplay hammer the statesav_* functions with overlapping calls…

Sorry, sending it here, because I couldn’t figure out how to send a message with this forum.

My code is forked over @ http://github.com/tschak909/libretro-atari800

-Thom

#2

FYI, @r-type is pretty active on github, too, so if he doesn’t see this, try pinging him there.

#3

Hi , nice you are working on to add feature. I have not so much time for now to look at , but i added issues in the repo so we now can talk about this in. I hope also we can hook the @anon24419061 joypad RA mapper in .

#4

Sorry, which repo?

I am currently trying to track down state saving bugs.

Currently, I have state saving working by implementing the *OPEN macros as fmemopen() calls. I can load and save individual state very effectively, and on some games (Pogoman, Wall War, etc.) rewind works excellently. However, on other games (Track Attack, World Karate Championship, etc.) rewind always causes a CIM, and am wondering if this is because of some sort of potential threading problem, race condition perhaps… I tried to mitigate it by fashioning a mutex in the state saving code that the libretro-core.c spins on to ensure calls don’t overlap…but that doesn’t seem to help… still trying to find some variable that I can consistently reproduce.

-Thom