Awesome! I guess it's time to prepare a beta release of my Game Boy Color emulator :)

In b4 Texas Instruments' C&D orders. :P

too late for them. I wouldn't be surprised if the exploit alerady made it on russian/chineese servers

Didn't want to use the ARM GBZ80 core from Goomba Color, huh?

Ha, I thought about it. But I didn't want to figure out all the GBA-dependent stuff, plus I don't like using other people's code :P

I did reference your code occasionally if I was encountering emulation oddities though. It's kind of funny, we both used the same method for determining the half-carry flag O_o

This is great because of the NSPIRES capabilties.

woohoo. Nice work guys :) I'm excited to see all the stuff calc84maniac has been cooking up released to hardware.

Yay I'm glad it's out! I can't wait to try it (and eventually Calc84 Gameboy Color and TI-89 emulator)

Absolutely superb news, finally that calculator might be worth something. This Ndless thing looks to be java to me, is there a technical reason it is only for windows?

It uses TI's proprietary link components. :-( (That's actually one of the reasons it's in Java.)

Someone really needs to port it to the CalcForge libs (or the LPG libs, they're still almost the same thing). Note that those libs are GPL, so we'd need Ndless to get dual-licensed (as the MPL is not compatible with the GPL), but its authors seem amenable to that idea, so I think that shouldn't be an obstacle.

So far, the following functions already have an equivalent in the CalcForge libs:
pxConnectApp . connectedDevice().getDeviceInfo
pxConnectApp . setConnectAPI (internal function, the CalcForge libs have a different API there)
pxConnectApp . enumerateConnectedDevices
pxConnectApp . connect
pxConnectApp . installOS
pxConnectApp . getDeviceInfo
conApi.shutdown
conApi.sendFile
conApi.getFile
conApi.enumDirectory("/") ("/" being hardcoded, see get_dirlist in calc_nsp.c, it does a recursive enumeration, but I don't think that matters)

On the other hand, the following functions would have to be added to the CalcForge libs to make the Injector work:
conApi.getFileAttributes (low-level function already present: cmd_r_dir_attributes) (It would probably be possible to do without that and use a receive to test directory existence instead of that.)
conApi.mkDir (looks required to me)
conApi.copyFile (true local copy, required, copying through the PC won't work, we need to send the actual command for on-calc copying)
conApi.delete (could probably be done away with, but we'd like to avoid making a mess; del_var is already in the API, but not implemented for the Nspire yet)

Ok, I'm not a programer and I am trying to understand what this all means. So someone please answer a few questions.
1. Was this accomplishment with TI's co-operation and urging or in spite of their efforts to keep such things from happening.
2. Is there some way that TI can stop this progress?
3. How long will it take before some awesome games are probably going to be available?
4. Assuming that most people will have the latest version of the os installed, how do you install the older version needed.
5. What does this mean for the people that simply want better math capabilities, like 3d graphing of multiple surfaces.

Many thanks to anyone who will answer these questions for us non-programers.

This would better have been asked in a top-level comment, but let's answer it anyway:
1. The latter.
2. Probably. But let's not give them ideas!
3. Hard to tell, depends on the progress of the toolchain, on how incomplete a toolchain developers are willing to put up with and on how many developers are interested in the first place.
4. Ndless downgrades it for you.
5. Math programs are possible. On the other hand, you currently need to use an older OS with less Math features.

> Someone really needs to port it to libti* / libcalc*
True, but there's a significant number of tasks of higher priority than this one ;-)
I mean, working on the toolchain, the library, the documentation, the program loader, expanding the set of OS versions that can be opened, the emulator, etc.
All of those matter more for the enjoyment of calculator hobbyists out there, and will foster community growth more, than the task you're suggesting (which has clearly its use, but should be a mid-term goal - free software ideology should not trump efficiency) ;-)

The SVN repository indicates that work on Ndless, toolchain, etc. is going on.

Making the installer work on GNU/Linux is very important to reach the maximum possible userbase. Not everyone has Window$.

BTW, the emulator also still needs porting.

Having all the toolchain be Window$-only is a major issue.

Why don't you work on getting it ported? I'm sure they'd love to have your help with all of your expertise in the field.

The first step (adding the required functions to the CalcForge libs) is one of my top priorities at the moment (also as the primary maintainer of the CalcForge libs).

But I think somebody who actually has an Nspire would be more qualified to work on that stuff.

> Making the installer work on GNU/Linux is very important to reach the maximum possible userbase. Not everyone has Window$.
I know, re-read my message ;-)
Still, there are lots of tasks of higher priority than that one. Even a port of Ndless to MacOS X (using TI's software) would reach more users than a rewrite of Ndless to use libti*. I don't know whether it has been actually tested under MacOS X, but Ndless _does_ compile under MacOS X.

> BTW, the emulator also still needs porting.
IIRC, it works under Wine. That reduces the priority of porting it.

> Having all the toolchain be Window$-only
Wrong. For one thing, there are lots of decent ARM toolchains that work under other platforms.

The current version has registry access hardcoded to find Computer Link, so it won't work on anything other than Window$.

And the link tool and the emulator are essential parts of the toolchain.

Excellent work!

I am waiting for the awesome upcoming programs for classes, experiments and gaming!

That is so cool.
I had given up hope on my nspire, but this is great news.
Now about only 4 to 12 months for the good games to start rolling out.
Mario in 16 shade grayscale will be so cool.
This took forever.

It would be cool if someone makes Enhanced BASIC - a app that allows you to create BASIC programs with at least the capabilities of the 89. Better if it is a calculator app with advanced programming capabilities.

the funny thing is that some development for the Nspire, such as a Gameboy Color emu, started before the exploit is made public XD

I wouldn't be surprised if some games came out much before that

Though keep in mind I started the GBC emulator almost 4 months ago now

How exactly does that work? Do you just make a few assumptions about how things will possibly work in the future and go from there? Obviously we already know the thing is ARM, and we had the OS dumped from it a while ago. How much do you really need to know to get started with stuff?

At this point we had dumped the boot2 and found out a bit about the hardware, and Goplat started his emulator. So, I knew where the LCD RAM was and how to access the keypad, and that was pretty much all I needed to know until I added OS interfacing later.

ok ... now i see its already out (gbc emu)
wow_ that is cool
no wait
thanks to everybody who developed it

1) i need to read

Also, i dont wnt to do this if it will mess with my future abilities to upgrade firmware.
It wont, right?

Right. And I would recommend that until we see how TI responds code-wise, no one update to an OS beyond 1.7. Wait for word from a reliable source in the community that it's safe to upgrade before doing so. You wouldn't want to lock yourself out of something because TI fixed or restricted something.

The ability to downgrade may not be there forever.