QuoteFirst, i tried the Tumbleweed and Leap OpenSUSE images and they don't work... they just won't boot.
OpenSUSE said the images were experimental so they probably do not work.

openSUSE Leap 42.2 images use the same code as SLES, so you probably did something wrong there.
The Tumbleweed image uses the upstream kernel without any patches, so in theory it should work,
but in practive too much stuff is missing or broken.

QuoteBut like @Streetwalrus said, i was going to be disappointed because it simply doesn't have VC4 or the proprietary drivers, only framebuffer support.
I have no idea how to compile VC4's driver for X11 so needless to say, i started to look at other options.

There seems to be some confusion here. VC4 is just the name of the chip. There are three options to use it:

* Framebuffer only, using the mailbox protocol (vc4 kernel module not loaded)
* Proprietary driver, replacing system libs ("rpi-userland"): Known to not work properly on aarch64 anyway
* FOSS VC4 driver in mesa: Works well enough for a few OpenGL apps, but unstable

X just needs modesetting, so if vc4 is loaded it just works (tm).
I'm running my RPi3 headless anyway so I just use the first option.

Quick update: v1.2 got released with an important bugfix: The snapshot path of kits is no longer ignored!

Great news for iOS users: This version supports the much quicker JIT also on 64bit phones, even if they are not jailbroken!

Download it here

Quote from: DJ Omnimaga on December 05, 2016, 07:39:10 AM
Wow, that's a lot of downloads! Do you plan to put the most recent stable version on ticalc.org? I forgot if it was there. In any case, I'M glad that people are using it.

No, I always like to keep binaries at a single place, in this case GitHub. It makes it easier to update it: All links that point to firebird automatically
show the latest release.

Quick update: v1.1.1 has been released with a small bugfix for touchpad cursor handling.
Previously the handling of dragging (like the graph view) was a bit bad, it should work fine now.

Also, v1.1 has already been downloaded a whopping 268 times (on GitHub, not counting the linux repo) and all releases together 45337 times!

New release: Firebird 1.1

Get it on GitHub!



New features:
* Add new configuration dialog
* Add easier way to manage multiple flash images with Kits
* External LCD window, useful for presentations or saving screen space

Core improvements:
* Emulation of different HW revisions (A, J and W) of CX
* Support boot1 4.0 found in HW-W+ calcs
* Support lcd_compat mode used by ndless
* General speedups for x86_64 and ARM JITs
* Supports CX installer images
* Add headless and emscripten platforms (no binaries, not really that useful)
* Emscripten demo on https://nspire-emus.github.io/firebird/firebird.html

Keypad improvements:
* Add labels to the keypad
* More intuitive keymap for mapping host keyboard to calc keypad
* Touchpad can now be used to move the OS cursor
* Make active area of buttons bigger to make typing with touchscreens easier
* Avoid buttons appearing stuck after hovering with mouse

Misc. / other fixes:
* The 32bit iOS binary detects JIT capability on runtime to avoid crashes
* Some minor bugs fixed, like potential deadlocks in the debugger
* Fix running functions from GDB
* Improve HiDPI scaling
* Mobile UI optimized for tablets in landscape orientation (right- and left-handed mode)
* Add an update-checking feature to the about dialog

Notes:
* The internal format of snapshots changed - firebird will refuse to load snapshots created with an earlier version. 
You can use the old version to save your work to the flash image first.
The macOS version isn't codesigned, so to launch it you may need to either right-click the .app and choose Open*, or simply disable GateKeeper entirely.
* The iOS version is 32-bit only (for now anyway) and is built with JIT (which will disable itself if your device is not jailbroken)
* The Windows version also has a mobile/tablet .exe that may suit those platforms better.

Repo for Arch, Fedora and openSUSE: https://software.opensuse.org/download.html?project=home%3AVogtinator%3Afirebird-emu&package=firebird-emu

It was actually much less work than I thought, this ndless_resources adds 128k of additional stack space before running programs on OS >= 4.2.
Please test and report back!

Quote from: bluman855 on November 01, 2016, 04:21:16 AM
Sorry for the extreme necro, but is this problem resolved? I am on a non CAS Nspire CX, and the calculator reboots when I try to run nquake.

The issue hasn't been resolved yet, as the underlying cause is fairly complex. However, I've got two weeks of vacation now so hopefully I'll be able to look into this in the next few days!
The approach of mapping additional pages is likely the best one, as it means not having to fiddle with sp at all. It just needs some modification to the nucleus task struct, so that the OS doesn't think that it overflowed.

Quote from: gameblabla on October 02, 2016, 10:19:27 PM

Quote from: Vogtinator on October 02, 2016, 08:38:46 AM
Hm, that means there was a division by close-to-zero in the perspective divide.
It should be gone if you decrease the rendering distance, although it shouldn't
occur that often with BETTER_PERSPECTIVE... Did you run make clean?

Yes, i ran make clean several times my sweety but it is still there.
Decreasing the rendering distance doesn't seem to do much...
Same with BETTER_PERSPECTIVE.

Quote from: Vogtinator on October 02, 2016, 08:38:46 AM
Why?

So It happens less often ?
I admit i'm not quite sure what causes this to happen but you can see it only bleeds in the blue sky background...
Any idea for a workaround ?

You could try to make the cubes smaller and change the camera so that it makes everything look right again,
that way the coords don't get out of bounds that easily. A proper fix would be to adjust the chunk's coordinates
to the player's position, so that the player is always at (0/0/0) and the chunks change coordinates instead of
the player. This will cost some vertex recalculation time, but will avoid large coordinates reliably.

Quote from: gameblabla on October 01, 2016, 10:05:42 PM

Quote from: Vogtinator on October 01, 2016, 09:20:24 AM
If it runs fast enough, you can tune some settings in glconfig.h, especially BETTER_PERSPECTIVE.

Sadly, i had already enabled this and it doesn't fix anything.
When i have enabled SAFE_MODE, it spams my terminal with this :

Code Select Expand


Warning : Texture coord out of bounds !

And here does it look like on my GCW0 (and PC) :

Hm, that means there was a division by close-to-zero in the perspective divide.
It should be gone if you decrease the rendering distance, although it shouldn't
occur that often with BETTER_PERSPECTIVE... Did you run make clean?

QuoteThe map file also becomes much bigger.
Maybe it is a good idea to add insivible walls ?

Why?

Quote

QuoteIt's great that you were able to port it, can you clean your port up a bit and submit a pull request?

Hmm... That will take some time because i changed my git repo a quite bit compared to yours.
But i will look into it.

It's much appreciated!

QuoteYour initial SDL port didn't have controls, which i had to implement.

Thanks! I never bothered to do so because it was mainly just for engine testing and performance tuning
(valgrind and perf on a raspberry pi)

QuoteFor example, at higher resolutions, the inventory menu doesn't control properly so i have fixed that.

Yeah, I never bothered with that either, at some places I hardcoded coordinates.

QuotePlaying it on my zero, it runs much faster than on my nspire but i eventually found a bug
where if you go far enough, the textures will severely bleed.
I'm not sure how it can be fixed though : maybe extending the world ?

If it runs fast enough, you can tune some settings in glconfig.h, especially BETTER_PERSPECTIVE.

It's great that you were able to port it, can you clean your port up a bit and submit a pull request?

Quote from: DJ Omnimaga on August 02, 2016, 06:02:56 PM
Now the next question, though, is did they patch the exploits used by nBoot and Ndless?

The exploit used by nBoot can't be patched, it's in the hardware (boot1). It is fixed in HW-W though.

QuoteI remember that using the function exit sometimes crashes the nspire for no good reason and unfortunely, Gpsp seems to rely on it...

Code that behaves correctly should not crash the nspire. Most likely there's a buffer overflow somewhere or the emulator doesn't reset the hardware to the correct state.

QuoteIt does seem to run a little faster than before though.

Yeah, newlib is much more optimized than the OS itself and GCC 6 has some more aggressive optimizations, although those may cause crashes non not quite standard compliant code.
See: "-fno-delete-null-pointer-checks", "-fno-lifetime-dse"

Quote from: gameblabla on June 28, 2016, 05:27:16 PM
I'm sure someone like calc84maniac could have done it well before me if he wanted to but it seems that he left the Nspire scene quickly after Gpsp...
(probably because he got too many death threats :p)
Speaking of Gpsp, well, he violated the gplv2 because he never released the source code to it...

He did: https://www.omnimaga.org/ti-nspire-projects/gpsp-nspire-(gba-emulator)/msg314055/#msg314055

Quote from: Duke "Tape" Eiyeron on June 18, 2016, 02:23:05 PM
Looks like the floating point bug is actually an issue with the combo "recent Ndless version" and Nspire OS 3.1. Vog just told me that 3.1 is just broken so I may have to switch for 3.4. :|

It's not 100% sure that it is actually the OS, it's just most likely.

Quote
@Vogtinator, if you're on any IM, I'll be very happy to know how I can join it to be able to talk at the pace my mind goes because it's been very frustrating to wait for hours if not days for a single clue. I've nothing against you, don't take it personnal, it's just my patience that just ran short and I *want* to move forwards from this piece of WTF.

If I'm on, I'm most likely on #firebird-emu on efnet.

Crafti v1.2

(Original thread with old versions and history on omnimaga: https://www.omnimaga.org/ti-nspire-projects/ngl-a-fast-(enough)-3d-engine-for-the-nspire)

No major changes, just some bugfixes and improvements.
I haven't tested classic support extensively, so if you're relying on that you should keep your old version around.
On a HW-W calc this should improve usability significantly, so please upgrade!

Download on GitHub

Improvements in crafti: Bugfix: Division by 0 when calculating collision while looking straight ahead Zehn compression for faster startup and less space requirements

Improvements in nGL: Several performance improvements Uses lcd_blit, so works natively on HW-W On PC, use SDL for graphics Add tools for converting wavefront obj Add VECTOR3 and nglDrawArray Add two new lessons to the tutorial