JP NAND file can't be found

[facaikotei]

What version of melonDS DS are you using?

1.1.0

Does the problem happen in standalone melonDS?

No, it's a bug specific to melonDS DS.

What happened?

When I use JP NAND file, it say that none was found and ask me to place NAND file, but the file is already there. US, EU and CN NAND files can work normally.

What should've happened instead?

It should be found like US, EU or CN NAND file.

Which platforms do you see the problem on?

• Windows
• macOS
• Linux
• Android
• iOS
• tvOS
• Other (specify below)

Which versions of the OS?

Android 13

Which libretro frontend are you using?

retroarch 1.17.0

Terms

• I agree to follow the code of conduct.
• I understand that melonDS DS is not developed or maintained by the authors of melonDS.
• I understand that this form is for reporting bugs, not for requesting help.
• I understand that the author develops this core in his spare time, and that he is under no obligation to respond to this bug report.

[JesseTG]

Could you please post a log file with debug-logging enabled? You'll want to set the logging options like so:

This will provide information that may help me reproduce this issue. Thank you!

[facaikotei]

log file: retroarch.log

[facaikotei]

I solved the problem myself.
I opened the NAND file in hex editor then copied the values from address 0xff800-0xff83f to address 0xf000000-0xf00003f then the core found the NAND file.
So could you describe in core option more detail about NAND file? Because not everyone knows the address of footer and how to check.

[JesseTG]

I opened the NAND file in hex editor then copied the values from address 0xff800-0xff83f to address 0xf000000-0xf00003f then the core found the NAND file.

When you did this, did the NAND file size change? I don't currently validate the footer, I only check the file size. I suspect I may need to loosen this check.

[facaikotei]

Yes, from 240MB(251658240)(without footer) to 240MB(251658304)(footer merged).

[facaikotei]

Maybe checking the file name is better? Like NO$GBA and melonDS do.

[JesseTG]

Maybe checking the file name is better? Like NO$GBA and melonDS do.

I originally considered this, but decided against it. The convention is for DSi NAND images to end in .bin, right? Unfortunately a lot of other important system files (even for other cores) use .bin, so it wouldn't be enough to just use the name.

I'll give the GBATek docs a closer look to see if I can learn some more about NAND images.