Homebrew Homebrew app GarageNX - Title Management / Installer / Etc.

[RobZilla10001]

Before I begin, let me just say YES, I used AI in the development of this application. No, I will not go into detail (except for the fact that it was directly responsible for the Spanish language file); I don't feel the need to defend or debate that decision. Use it or not, entirely up to you.

GarageNX​

I'd like to introduce my first Switch homebrew project, GarageNX (so titled because it aims to "handle everything under the hood" of the Switch, so to speak). This is my attempt to replicate the functionality of DBI. Yes, there are other title management software applications out there, but many seem to appreciate DBI and as such, I figured it was the best template to work from.

What it does:

• File browser - ranger-style three-column navigation with split-pane copy/move, text and hex viewers (paged for large files), and archive peeking (NSP/XCI content listing).
• Network browser - navigate SMB/NFS shares using the same browser interface.
• Title management - enumerate installed titles; uninstall, move between SD and NAND, reset version requirements, edit metadata, dump, and repack.
• Maintenance tools - clean orphaned records, old updates, placeholders, unused tickets, firmware dumping, version manifests.
• System information - comprehensive firmware, CFW, hardware, battery, and activity reporting.
• Connectivity - USB-MTP, FTP server, and HTTP server with QR-code network sharing.
• Full localization - drop-in language files; no language is second-class.

What it doesn't do:

• Create forwarders - Sphaira handles this functionality well, no need to replicate.
• Launch Homebrew - Once again, Sphaira is the standard here.
• Browse HTTP/FTP/GitHub - saw no need for it. If there is demand, this may come in a future build.
• Look pretty - I felt no need to go overboard with a flashy UI. Plutonium has it's place, it's a wonderful framework and XorTroll is a peach for having created it. That being said, for what I'm trying to accomplish, a very basic, text based SDL2 implementation works just fine.

What I need:

• Testing - I can only think of so many different ways to test, and they're not necessarily real world scenarios. That's where you come in. Please, have at it and put it through the ringer. Submit any issues to the GitHub.
• Translation - I'd prefer not to rely on AI exclusively for translation; I know it's imperfect and I'd prefer more natural language results than what an AI can look up in a dictionary or online.

What will change:

• Splash - I previously had an AI generated splash that was...problematic to say the least. Would've put this project in the cross hairs of big N. I am working with an artist to come up with something more appropriate; in the meantime, there's the garish placeholder that is there now.
• Language support - as more translations are completed and verified, they will be added to the repo with haste. I don't want anyone left out in the cold on this; everyone deserves to have well-working software regardless of the language they speak.
• OPTIONAL: Themes - In the future, themes will be external files like the language files (they were at some point and, for reasons escaping me at this very moment, they were integrated into the build); see /source/ui/theme.cpp for the basics. Should be fairly simple to split it back off.

As I've said above, please feel free to put this through it's paces; it does no one any good if anything is broken. I've done as much testing and delayed this release as long as I dare; it's time to get some fresh eyes and have some constructive feedback. I'm sure I've forgotten something or overlooked some broken feature, please enlighten me so that I can correct it. Thanks for taking the time to check this out. The working build is under the branch "alpha-build"; the main branch is several commits behind. Please ignore it for now.

Click Here For The Repo.

 

[mathew77]

Another LLM slop software, for Gods' sake..
We definitely need a special tag for it, that's for sure.

 

[RobZilla10001]

Another LLM slop software, for Gods' sake..
We definitely need a special tag for it, that's for sure.

This you?

I only ask because you post "so we can avoid LLM at any cost." <-- Is that all LLM? Because if you look at your posts, many, many of your posts are on "AI slop" and globbing on about how great the projects are. So are we to avoid "at any cost" as you say? Or is it just projects you're not personally invested in? Many of the ports you post about are either fully AI or AI assisted in some way. Chansey doesn't hide it. I don't believe Nagaa or Sensei do either (I could be wrong here, apologies to the authors if I am). I'm not trying to change your mind, btw, just pointing out that you're inconsistent. If you truly hate AI/LLM, give up your ports, because they're all extensively AI assisted.

 

[albarapa]

This you?
View attachment 589838
I only ask because you post "so we can avoid LLM at any cost." <-- Is that all LLM? Because if you look at your posts, many, many of your posts are on "AI slop" and globbing on about how great the projects are. So are we to avoid "at any cost" as you say? Or is it just projects you're not personally invested in? Many of the ports you post about are either fully AI or AI assisted in some way. Chansey doesn't hide it. I don't believe Nagaa or Sensei do either (I could be wrong here, apologies to the authors if I am). I'm not trying to change your mind, btw, just pointing out that you're inconsistent. If you truly hate AI/LLM, give up your ports, because they're all extensively AI assisted.

Many people here believe they are entitled to demand that projects be created entirely without AI, without considering that these are nonprofit projects developed by people in their free time as a hobby. They often fail to understand how much time, effort, and technical expertise it actually takes to program and maintain such projects.

Edit: On the other hand, your disclaimer regarding the use of AI practically invites this kind of discussion. Personally, I'd keep it shorter so the post focuses more on your project.

 

[ChanseyIsTheBest]

I know it's a bit rich coming from me as a serial vibe-coder but I do worry about if you are writing stuff to NAND just make sure you test really well first and have a good understanding so you don't give people any issues.

I can get away with being sloppy vibe-coding I mean just look at how the Slay the Spire 2 port is at the moment because I'm not doing any system operations with my ports only SD card writes. I don't think I could bear putting out something that fried someone's filesystem. The stakes and responsibility are just higher with certain types of homebrew like this compared to ports.

 

[RobZilla10001]

I know it's a bit rich coming from me as a serial vibe-coder but I do worry about if you are writing stuff to NAND just make sure you test really well first and have a good understanding so you don't give people any issues.

I can get away with being sloppy vibe-coding I mean just look at how the Slay the Spire 2 port is at the moment because I'm not doing any system operations with my ports only SD card writes. I don't think I could bear putting out something that fried someone's filesystem.

Completely understandable. That's partially why the NAND writing features are disabled by default. You have to manually turn them on (not just because of that, but because it's a good policy in the first place when dealing with NAND). And I've been testing with extra switches I have laying around with NAND backups, in case of issue. Obviously, not 100% safeguard or a perfect replacement for more extensive testing, is why I'm hoping for more help in the testing department.