Add battery save import/export methods

[Konstantysz]

Summary

Adds Python bindings for battery save import/export to match the new DeSmuME interface API.

Dependencies

⚠️ Requires DeSmuME PR: https://github.com/TASEmulators/desmume/pull/930]

This PR temporarily points the desmume_src submodule to my fork until the upstream PR is merged.

Changes

• Added DeSmuME_Backup class with three methods:
  • backup.import_file(path) - auto-detects format (.sav, .dsv, .duc, .dss)
  • backup.import_raw(path, size) - manual size override
  • backup.export_file(path) - export to .dsv
• Added backup property to DeSmuME class
• Temporarly updated submodule to konstantysz/desmume fork

Usage Example

from desmume.emulator import DeSmuME

emu = DeSmuME()
emu.open('pokemon_black2.nds')

# Import battery save
if emu.backup.import_file('my_save.sav'):
    print("Save loaded!")

Use Case

I'm building a Pokemon shiny hunting tool that needs to load battery saves programmatically across multiple emulator instances. Currently there's no way to do this without manual GUI interaction.

Testing

Functional testing pending - will compile custom libdesmume.dll and test with Pokemon ROMs once DeSmuME PR is reviewed.
Once the upstream DeSmuME PR merges, I'll update the submodule to point back to SkyTemple/desmume.

---

[Konstantysz]

Update: DeSmuME PR has been merged! The backup import/export API is now part of the official DeSmuME repository.

Submodule updated I've updated the desmume_src submodule to point to upstream master instead of my fork. The changes are now based on the official codebase. Tested and working.

• Import: Successfully tested with .sav files
• Export: Successfully tested (exports to .sav format)

Python bindings work correctly with the merged C++ API. This PR is now fully ready for review. Let me know if you need any changes!

I think I can ping you @mike8699 as a maintainer.

[theCapypara]

Great stuff! Thanks a lot!
Unfortunately this project is currently without a maintainer, but as the original author it LGTM!

[Konstantysz]

Great stuff! Thanks a lot! Unfortunately this project is currently without a maintainer, but as the original author it LGTM!

So there is no possibility for merge?

[Konstantysz]

@mike8699 what do you think about this PR?

[mike8699]

@mike8699 what do you think about this PR?

Looks great. Unfortunately, I am not a maintainer but just a past contributor 🙃

[theCapypara]

@Frostbyte0x70

[Frostbyte0x70]

Hi, sorry for the delay, I was not receiving notifications until I was pinged.

Unfortunately, it seems like the build is failing for reasons unrelated to your changes (I opened #70 to fix the workflow, but that one fails with an error). There's currently no one available to fix the build, so I'm afraid that both #70 and by extension this PR will remain unmerged for now.
If someone happens to fix the build in the future, I'll let you know so you can rebase this on top of #70 after it's merged, then we can run the CI on this PR again.

Thanks for taking the time to contribute, and sorry for the disappointment.

[Konstantysz]

Hi, sorry for the delay, I was not receiving notifications until I was pinged.

Unfortunately, it seems like the build is failing for reasons unrelated to your changes (I opened #70 to fix the workflow, but that one fails with an error). There's currently no one available to fix the build, so I'm afraid that both #70 and by extension this PR will remain unmerged for now. If someone happens to fix the build in the future, I'll let you know so you can rebase this on top of #70 after it's merged, then we can run the CI on this PR again.

Thanks for taking the time to contribute, and sorry for the disappointment.

@Frostbyte0x70, I created fixes for the Github CI in #71. It passed all runs, beside macos-12, which was waiting indefinitely.

[Frostbyte0x70]

Doesn't this mean we'll lose the commits added in our fork of desmume? The commit that's currently checked out in the submodule has a few commits not present upstream, and the branch said commit is in has those plus a few more that were added later.

[Konstantysz]

So should I port my recent changes made to TASEmulators/desmume to your fork? It's actually just +43 lines of code.

[Frostbyte0x70]

Are those the only changes needed from the Desmume repo to get your feature to work? If so, I could just cherry-pick that commit into the branch used by the submodule.

[Konstantysz]

Yeah, I just need changes from this PR. But I don't get which branch should I pull from? I'll pull from the correct place, port changes and test, but I don't understand the starting point for me :(

[theCapypara]

IIRC I just backported some compilation fixes onto an older commit because I didn't want to fully update in case things break, there shouldn't be any functional difference and it's probably fine to update.
I'd recommend seeing what the base commit was that exists in the upstream repo and what commits I actually added on top of that, should be fairly easy with the git CLI, just kinda hard with GitHub's UI, though the network graph should also show it.

[Frostbyte0x70]

I'd recommend seeing what the base commit was that exists in the upstream repo and what commits I actually added on top of that

I believe the second link I posted above shows exactly that. For now, I'll just replicate the changes from Konstantysz's PR on the branch used by the submodule (which seems to be interface-build).

[Frostbyte0x70]

Alright, I've added the changes from the PR. @Konstantysz you can now checkout commit SkyTemple/desmume@42ee6ea on the Desmume submodule and push again.

[Frostbyte0x70]

It finally works! 🎉
Thanks!

[Konstantysz]

It finally works! 🎉
Thanks!

Thanks too!