Homebrew [Release] FBI - Open source CIA installer

  • Thread starter Thread starter Deleted User
  • Start date Start date
  • Views Views 1,611,725
  • Replies Replies 4,207
  • Likes Likes 102
Yep, only Pokemons and a dozen more for europe.

So I could install those for free on sysnand I'm happy with that =D
gonna work a lot these hollydays on a rom loader cause I'm not satisfied of the haking scene actually so gonna try to at least patch signature check
 
  • Like
Reactions: Margen67
So I could install those for free on sysnand I'm happy with that =D
gonna work a lot these hollydays on a rom loader cause I'm not satisfied of the haking scene actually so gonna try to at least patch signature check

Please share it, should you manage to do it. LOL

Or you can just share it with me... It could be our little secret. ;)
 
  • Like
Reactions: Margen67
Steveice10 Did you check how much processing power consumed by onProgress? I suspect that if you remove floating point devision and make actual output more rare, not every callback, the actual installation will be significantly faster.

I seem to recall doing this at one point but changing it back so the cancel button check would be called more often. Of course, back then the read buffer was significantly bigger, so each loop would take longer. Anyway, I removed the use of floating point division from onProgress, added a counter so it'll only be called every few loops, and made a few more minor touch-ups.

Not sure how big of a speed boost it'll give as I usually only test with small CIA files, but it may improve install times a bit.
 
Steveice10 can I use this in Ninjhax to do region changing? By the way, it fails to load in rxEmuNAND 9.2.0-20U (Nothing installs, even legit games)...

I know next to nothing about region changing, but I'm pretty sure it involves more than simply installing the other region's titles. So it could be used for region changing, but there's more to it that FBI can't help you with.

Also, rxTools uses the 9.6 FIRM (firmware.bin), which has the ARM11 kernel exploit patched, so FBI cannot be used through Ninjhax in rxTools.
 
I know next to nothing about region changing, but I'm pretty sure it involves more than simply installing the other region's titles. So it could be used for region changing, but there's more to it that FBI can't help you with.

Also, rxTools uses the 9.6 FIRM (firmware.bin), which has the ARM11 kernel exploit patched, so FBI cannot be used through Ninjhax in rxTools.

if only there were a way to get the 9.2 FIRM... one day...
 
  • Like
Reactions: Margen67
By request, I've allowed disabling prompts when installing over the network. This makes batch installing over the network, which can be done with SocketPunch, much easier.
 
  • Like
Reactions: Margen67
I seem to recall doing this at one point but changing it back so the cancel button check would be called more often. Of course, back then the read buffer was significantly bigger, so each loop would take longer. Anyway, I removed the use of floating point division from onProgress, added a counter so it'll only be called every few loops, and made a few more minor touch-ups.

Not sure how big of a speed boost it'll give as I usually only test with small CIA files, but it may improve install times a bit.
One more proposal. You can keep input check but actual details output may be placed after
if ((totalsize>>17)&1)
That will work every 256MB of transferred data, with present speed it is about 8 times per second. We just don't need more precise progress display anyway and it really can speed up the whole process.
 
One more proposal. You can keep input check but actual details output may be placed after
if ((totalsize>>17)&1)
That will work every 256MB of transferred data, with present speed it is about 8 times per second. We just don't need more precise progress display anyway and it really can speed up the whole process.

I'd rather have a more up to date progress display. At that point, I/O speeds would be more of a bottleneck than progress updating.
 
I'd rather have a more up to date progress display. At that point, I/O speeds would be more of a bottleneck than progress updating.
Can't say for sure for the output routines you used, but I implemented text output based on gxtexturecopy and it drastically reduces performance. Just check how the speed is affected on a large CIAs. Anyway who needs an update over 10fps, while the current version shows 20+ fps.
 
Can't say for sure for the output routines you used, but I implemented text output based on gxtexturecopy and it drastically reduces performance. Just check how the speed is affected on a large CIAs. Anyway who needs an update over 10fps, while the current version shows 20+ fps.

I suppose it wouldn't hurt to reduce it to something like 4 times a second.

EDIT: Okay, updated it just now. I actually decided to just go ahead and make the progress display only redraw when the percentage changes.
 

Site & Scene News

Popular threads in this forum