This is just... perfect!
With this we can make our own "exec" and even transfer files without tns or compressed or below /documents or debug easier!

As I'm working on USB calc-to-calc connectivity with Ndless, I'm actually hooking into NavNet, which is the TI-Nspire specific protocol and TI driver for standard OS services (dir listing, screenshots, file transfers, ...). The new syscalls can be used to call services and expose new ones.

How did you find it out? I mean, the OS is big and the USB protocol and such things are quite complicated in software.
Searching for function addresses and finding out which parameter is what surely took long.

(Lionel, I'm at the moment loading navnet.dll, but switching to libticalcs when running on Linux shouldn't be too difficult I suppose as long as we can match the 2 APIs).

Oh, the API for TI-8* and nspires is the exact same, so I fear it's not possible without more functions exposed..

I'm not sure if NavNet requires the computer to initiate the service calls (all the standard services are computer to calc...).

TINC(L)S, and therefore libticalcs, does it, so I'd say "probably"

(Lionel, I'm at the moment loading navnet.dll, but switching to libticalcs when running on Linux shouldn't be too difficult I suppose as long as we can match the 2 APIs).

It's alright to use, in the beginning, an API that you already know because you already used it in the fairly distant past for Ndless.
At the time, libticables' special version of the libusb DLLs, required for working around bugginess in the Nspire's linking code, did not have either signature or filter driver support. Romain tried to mainline the changes a long time ago, he didn't receive a reply; I retried, and this time, I got a reply, made a patch, which got improved and integrated into the official version with filter driver support and signature.
Now that the main reason for not working with libti* has disappeared (although the belief that TI's software and libti*/TILP are mutually exclusive is so entrenched that more than two years later, some people still don't know about it), libti* support should be added quickly enough, so that community software such as pacspire doesn't depend (only) on TI's proprietary code

Oh, the API for TI-8* and nspires is the exact same, so I fear it's not possible without more functions exposed..

Wrong. Remember from above, all implemented Nspire functions are exposed by libticalcs nowadays
Functions which do not fit into the common API are actually, and currently, a thing of the TI-83+ family, far more than a thing of the Nspire family. There's a single Nspire operation which does not fit into the common API: Nspire keypresses, which do not fit in 16 bits, unlike TI-Z80 and TI-68k keypresses. However, the TI-83+ family has a dozen functions which do not exist on other models. A little while ago, I made a low-usefulness patch for adding a number of such functions; it's queued in a local branch, but I have not committed it to master yet because, precisely, the usefulness is low.
If support for quizzes, and various other functionality rarely used outside of special equipment in classrooms, baked into TI's higher-level libraries which use NavNet.dll, were to be added to libticalcs, then the number of functions which do not fit into the common API would rise.


EDIT: looking at the code:
* the group of initialization functions, from TI_NN_Init() to TI_NN_DestroyOperationHandle(), can be replaced by libticables / libticalcs equivalents, see the authoritative test_ticalcs_2 example in the libticalcs source.
* TI_NN_Connect() <-> nsp_session_open() defined in nsp_vpkt.h;
* in this situation, TI_NN_Write() is the exact equivalent (with a different SID) of nsp_cmd_s_echo() defined in nsp_cmd.h, which calls nsp_vtl_pkt_new_ex() + nsp_send_data() + nsp_vtl_pkt_del(), all from nsp_vpkt.h;
* likewise, TI_NN_Read is nsp_cmd_r_echo(), i.e. nsp_vtl_pkt_new() + nsp_recv_data() + nsp_vtl_pkt_del();
* TI_NN_Disconnect() <-> nsp_session_close() defined in nsp_vpkt.h;
The nsp_cmd_s/r_echo() pair is used by is_ready() in calc_nsp.c, used as a function pointer by the common ticalcs_calc_isready() API.

I guess I could add a pair of helper functions to nsp_cmd.c/.h:
* a near-copy of nsp_cmd_s_echo(), taking the last argument of nsp_vtl_pkt_new_ex() as parameter;
* a near-copy of nsp_cmd_r_echo(), taking the argument of ticalcs_info() as parameter.

This looks interesting. I'll see if I can add some features (once I install the latest Ndless SDK)

This project is going like a rocket it seems o.o You guys are doing an awesome job, just wanted to say that; keep it up

Indeed, pretty nice demo

I just checked the svn, and the TI_NN* syscalls were added in r878.

I have only commited the addresses, I'm currently working on the functions definitions and tests.