savestate_auto_save and savestate_auto_load do not work?

[S3riousSam]

That may be something obvious from the project perspective, but it does not look documented (known issue/limitation?) that this core does not handle libretro autosave feature (yet?).

I could not observe any evidence (although the save state file is created properly after quitting) that the emulator loads the state back when relaunching a game (experimenting on Raspberry Recalbox). Some (forum threads) mention that the core loads the state then resets right after.

Manual savestate/loadstate works well however.

[Darknior]

If it can ba repair, i love this option too

[markwkidd]

@S3riousSam and @Darknior what is a good game to test this with?

In other words, what is a game that supports saving and where it will be obvious that a savestate was loaded so that we can tell when the autosave/autoload feature is working?

[S3riousSam]

I think they all do? I mean, it could be Pac-Man, Mario Brothers, Donkey Kong...?

[markwkidd]

Unfortunately only some games support saving with mame2003.

It's not a feature I have ever used. But if pacman.zip works that's enough to know.

[S3riousSam]

...I think there is misunderstanding. libretro allows to autosave the emulator states when leaving/quitting a core. This feature "works" with mame2003-libretro core, but the autoload state (which is loading back the autosave state when launching the core) fails... Some claims that the core actually loads the autosave state, but then later on, reset it's state (and therefore breaks the "feature"). The savestate works if loaded manually afterward.

[markwkidd]

Thank you , this is all very helpful. These kinds of details will help start the debugging/implementation process in an informed way.

Please share anything else that you think we should know. Assume total ignorance on the autoload and the autosave functionality on my part.

[markwkidd]

I would like to report this as a RetroArch issue. @S3riousSam and @Darknior are you still available to report this to RetroArch at https://github.com/libretro/RetroArch/issues ?

If you are willing to get an issue going, tag me and I will join in. If we need to change something in mame2003 then that is fine, but with manual state loading I am willing to propose this as possibly a RetroArch/frontend issue.

[markwkidd]

@S3riousSam and @Darknior is this still an issue? It was looking like either we would need advice from RetroArch or that it was even a bug in RetroArch rather than this core. As far as I know it could have been an intermittent RetroArch regression bug that is solved, etc. If not the OP, can anyone else reproduce this?

[schmurtzm]

Hi @markwkidd , I think that this issue still happens with mame2003-plus, it is not crashing or reset but games doesn't resume correctly (but it works if you resume manually).

Here the issue that I have created, I give some steps and roms names to reproduce easily : libretro/mame2003-plus-libretro#1433