Netplay saves SRAM on both host and client

[GameDragon2k]

Description

Currently netplay will save SRAM to both the host and client end. This means the client joining will lose any personal progress or saves that may have. Decent multiplayer example of this is Super Mario All-Stars or Super Mario Kart.

Expected behavior

Battery save is loaded from host, it should be saved only to the host's PC.

Actual behavior

Battery save is loaded from host, but it saves on both the host and client.

Version/Commit

• RetroArch: Jan 25 2017 84cb05d

Environment information

• OS: Windows 10

---

Want to back this issue? Post a bounty on it! We accept bounties via Bountysource.

[andres-asm]

This is not really an issue but a matter of debate.
What if you're playing with a friend and decide to proceed on your own later?

[Alcaro]

Expected behavior 2: Netplay save files have another filename, on both host and client.

[andres-asm]

What if you want to continue your single player game with a friend?
Not saying it can be done but changing the savefile name in RA is a pita and can go wrong for many different reasons.

This is a real problem I guess.
Even achievements are posted when speccing netplay matches.

It's the kind of thing that must be discussed to find a good solution.

[Alcaro]

Expected behavior 3: RetroArch supports multiple save files for the same game. You can choose which to use when loading the game.

Expected behavior 4: What plausible games are enjoyable with both single and multiple players, and contain any data worth saving?

[andres-asm]

Secret of Mana

[GameDragon2k]

I can see how this could be useful, but if you're always going to be playing with your friend, then the host will likely be the same everytime. In that regard, the save should only be on the host in my opinion.

This way your friend can continue his own save file without problems.

I used the Super Mario All-Star example because it happened to me. Friend had clear data for each Mario game, but because I hosted with a fresh save, his save was overwritten and lost all of his work.

[Alcaro]

But what if the host wants to play alone?

[GameDragon2k]

There isn't any issue with the host playing alone. Host doesn't lose anything as it is currently.

[micochops]

I see how this can be a potentially serious issue, especially if the client is entirely unaware that his save file is overwritten and all his progress will be lost if he connects to a host whose progress in the game is different from him. What are the chances of connecting to a host with exactly the same progress in the game as you?

[hizzlekizzle]

We already have an option to block SRAM overwrites on state load, so having an option to do the same thing to netplay clients seems reasonable.

[claimer21]

How are you guys saving with netplay anyways? I can't get mine to save using autosave/autoload, and using the normal F2 and F4 hotkeys don't work.

[andres-asm]

@moggd it works just the same as in single player so I have no idea about that

Regarding the original issue. What I'll do is make clients use a subfolder in the saves folder.
That should be good enough.

I was gonna append a different extension but that won't scale for all cores so I think this solution is the best we can have atm

$SAVEDIR/netplay/

If you use savefile sorting it would go int $SAVEDIR/netplay/corename/ instead of $SAVEDIR/corename

We can't really block from saving since not all cores use the SRAM interface anyway so the best solution we can offer ATM is to save in a different location.

[andres-asm]

The only limitation there would be that it sets paths on core initialization so it could still lead to problems in manual connect if both ends were already in-game when they connect...

The other solution would be to add an env to libretro to make the cores netplay aware, then we can cover all bases and add a few extra features too.

[rrooij]

Just had this issue with the great game Seiken Densetsu 3. I wasn't that far in the game, so it didn't bother me much. But it is still worth to mention.

I think the folder option should be used as @fr500 stated. Making cores netplay aware would be much better, as cores can use netplay features their own way that fits the needs of the core. However, that's a lot of work and in the meantime I guess that the folder option should fit all the needs.

[andres-asm]

Actually I will go for @hizzlekizzle's solution, using block SRAM overwrite on all clients.
Maybe with a toggle