Hacking G6/M3 FAT32 Lib v3 (13-10-06)

[jelbo]

Download here:

http://www.handheldsources.com/Backup/G6M3...%2013-10-06.rar

Readme says:

FAT32:
G6/M3 FAT32 source code.

moonshell13:
demo program base on "moonshell-013".

G6M3_write_demo.c:
write nand/sd/cf demo.

Thanks!

Files in the archive:

G6M3 FAT32Lib_V3
 G6M3_write_demo.c
 readme.txt
|_
 FAT32
  FAT32.h
 |_
  SOURCE
   disc_io.c
   disc_io.h
   gba_nds_fat.c
   gba_nds_fat.h
   io_g6.h
   io_g6.o
   io_m3cf.c
   io_m3cf.h
   io_m3sd.c
   io_m3sd.h
   m3sd.s
   mediatype.h
|_
 moonshell13
 Â


This agentq's, the porter of ScummVM to NDS, first reaction (as seen here):

QUOTE

After a quick look, it seems like the source for the G6 driver is missing. There's an object file, but no source.

Please, if you can find the source to G6_GetInterface() etc. let me know, because I can't.

I hope Chishm or others are able to implement this into FATlib so every homebrew is able to fully support the G6 and M3. Before jumping for joy and giving thanks to the G6/M3 team, I'll wait what the developers themselves can do with the files.

 

[Extreme Coder]

The M3/G6 team tricked the devs the first time they released their library, since that lib was only working with an official compiler($$$). Hope it's not the same case this time.

 

[retrohead]

The M3/G6 team tricked the devs the first time they released their library, since that lib was only working with an official compiler($$$). Hope it's not the same case this time.

They have again released the binary files and not the source code.

LAME

 

[Extreme Coder]

How hard is it to get their sources? It's not like they will lose anything:/

 

[jelbo]

Damn, I just knew it...

There's no way to reverse engineer them or something? I think chishm said he could do it, but it costs a lot of time and is rather useless since the needed files are right there in Hong Kong... Maybe we should organize some reward for the one who does it

If I were coder I'd probably do it

 

[JoeB]

I don't understand what the developers are complaining about. The G3/M3 team did release the object files (.o files). One should be able to simply link to them after they compile their code!

This is a common thing, unless they use a different, incompatible, compiler. At which point, how difficult would it be to switch your code base to use the same compiler as the M3/G6 team uses!?? Companies (like IBM, Microsoft, Oracle, etc. do this all the time!)

Rather than complain, why not publically state what compiler your want your object code to be generated in, and I'm sure the G6 team will be able to do so.

Then again, a lot of compilers do generated compatible .o objects!

Asking a company to just give the source is just crazy!

 

[Extreme Coder]

The .o objects are compiled using the official compiler(which costs a lot of $$$), not DevKitPro, which means it's going to be very hard to make it work with DevKitPro(which I guess all homebrew developers use). I
t's common sense we need the code in the compiler all homebrew developers use, they could've known that easily, after they released their first version of this library, and all devs started complaining and E-Mailing them.
Even NeoFlash released the complete source without anybody complaining at them...

 

[jelbo]

I think a lot of e-mails should be sent once again now the release is still fresh. I'm no sure if those Chinese programmers understand what exactly is needed (what Extreme Coder is saying).. Maybe a Chinese e-mail would help better?

 

[jpx86]

the problem is the object file provided is compiled with an old version of devkitarm. basically, its worthless.

you can disassemble the .o file and have the code in assembler, but not back to c code. chism has actually said even the assembler code would be fine if they'd release it, but he's not going to make an effort since they are being jackasses.

 

[rkenshin]

thought if they released their FATlibs, it would eventually lead to fakes?

Or is that the wrong idea I'm getting since this happened to the Supercard after they released their FATLib source?

 

[cory1492]

thought if they released their FATlibs, it would eventually lead to fakes?

Or is that the wrong idea I'm getting since this happened to the Supercard after they released their FATLib source?

Nope, the FAT lib itself would have little to nothing to do with reversing the hardware behind what makes the FAT lib workable on that hardware. Probably the biggest reasons the supercard got cloned (badly, I might add, since the clones dont work exactly like the origonal) is the fact the controller chip they used in it is getting dated (and thus is cheaper to reverse and cheaper to put into production than when the supercard was first released), it's simpler than M3's, and has been around for about a year longer.

I mean, a fully working homebrew driver has been around for M3 CF for quite a while, and there are no clones.

 

[js18]

does this mean I can make a custom media extend player?

 

[MajinGohan]

Still a no-go for homebrew on G6L and M3 I guess. Damn.