Omnimaga
Calculator Community => Other Calculators => Topic started by: critor on November 26, 2011, 05:34:21 pm
-
We've begun alpha-testing Ndless 3.1 for the ClickPad & TouchPad 3.1.0 OS (not yet compatible with CX, sorry).
All Ndless 2.0 programs tested up to now ran directly without any problem.
For those who might have missed them on TCAP this evening, first feedbacks and photos here:
http://tiplanet.org/forum/viewtopic.php?p=118236#p118236
-
Very nice!
I'm glad is already quite stable :)
-
O_O So excited!!! (Even though I have a CX)
-
Excited as well, even though I don't have any nspire whatsoever :P
Also I saw it on TCAP it was pretty interesting :)
-
I'm so excited !
(Has someone tried OSLauncher ?)
EDIT : And does it work with OS 3.0.1 ?
-
OMG! This is already working. Wonder what TI is going to do next? Release a new TI-nspire with OS 4.0? Cmon TI, we dare you!
-
not before we getz ndless 3 for CX. :P
-
Good news! I hope compatibility for those with the CX models isn't too far off. :)
-
(not yet compatible with CX, sorry).
Darn /me wants it for the cx
-
/me wants it for CX if he gets one
-
This is nice. I can't wait to try it, as I got a clickpad besides my CX. Also lol I didn't see this topic until after I cross-posted a news about Ndless on TCAP :P
-
(Has someone tried OSLauncher ?)
If I'm not mistaken, OSlauncher should still work on ClickPad/TouchPad models.
On CX models there will be a little problem. OSlauncher needs a decrypted OS image. We used to decrypt on the PC as all 1.X/2.X OSes were encrypted ith the same BlowFish key. But CX OSes are now using different BlowFish keys for each version, the key beeing coded in the new "boot2.dat" file. Which means the "old" decryption program the author has no intention to update won't work any more.
EDIT : And does it work with OS 3.0.1 ?
No. You'll have to install OS 3.1.0.
-
On the SVN it's still ndless2. So can I assume that the source for testing isn't yet released to the public?
Or did I miss something?
-
On the SVN it's still ndless2. So can I assume that the source for testing isn't yet released to the public?
Or did I miss something?
It is a private alpha =)
ExtendeD wanted a private help to keep some information about the exploit private (not public, that way TI won't correct it on the next OS).
-
Fine I'll wait till the public release then ;)
-
YES! :w00t:
Finally Pokemon in a color calculator!! :D
-
(not yet compatible with CX, sorry)
BalancedFury, not so. Part of the problem, as I recall, is a piece of the CX hardware meant to prevent things like Ndless.
-
D:
Darn :P
Can't people fix that? :D
*Fingers Crossed*
-
Like Adriweb said, not lose faith in ExtendeD ;)
-
The demon :P
(Profile Picture Unless He Changed It) :D
-
(not yet compatible with CX, sorry).
Darn /me wants it for the cx
why was I downrated?
-
This is amazing, I just wanted to help on something, if I could...
Are you going to wait for the next nspire OS to release ndless 3.1 publicly?
-
No, I dont think he's going to wait [this long] (well, we can think TI isn't going to release a big update before 2012...), and since Ndless 3.1 is almost ready (not for the CX ... but it will come as well), ndless should be released soon... I hope :)
-
me too
-
I would like to see a solid API for ndless 3.1 before it's released... :(
(I'm working on porting SDL, and possibly a GUI library to the Ndless platform for potential inclusion)
I'm fine with a release without one, especially with the potential on the CX :D
-
I want it for the cx!
-
(not yet compatible with CX, sorry)
BalancedFury, not so. Part of the problem, as I recall, is a piece of the CX hardware meant to prevent things like Ndless.
Is this about the watchdog timer? That's pretty easy to defeat:
*90060C00 = 0x1ACCE551
*90060008 = 0
*90060C00 = 0
will turn it off.
-
O.O hopefully that works
-
(not yet compatible with CX, sorry)
BalancedFury, not so. Part of the problem, as I recall, is a piece of the CX hardware meant to prevent things like Ndless.
Is this about the watchdog timer? That's pretty easy to defeat:
*90060C00 = 0x1ACCE551
*90060008 = 0
*90060C00 = 0
will turn it off.
So it's not different at all on the CX then?
-
It's the same hardware, it's just that non-CX OSes don't enable it.
-
Well, if the next version of Ndless is released before the next version of the OS, it's likely that TI will positively go out of their way to fix the vulnerability(ies) used by this new Ndless...
-
I think these guys also need to figure out how color and B/W works - that is, figure out the HW a bit, and then figure out how to let developers access the HW. They also need to figure out how to make old programs work on the CX (grayscale emulation).
It's a lot of work, but they're on it. ;) (Plus that better SDK that I've posted about above...)
-
I think these guys also need to figure out how color and B/W works - that is, figure out the HW a bit, and then figure out how to let developers access the HW. They also need to figure out how to make old programs work on the CX (grayscale emulation).
It's a lot of work, but they're on it. ;) (Plus that better SDK that I've posted about above...)
Grayscale emulation is easy -- the LCD controller still supports 4-bit palettes, which we can set up to be shades of gray. Some of the other hardware might be problematic, though. The keypad hardware reads inverted values compared to the non-CX calcs, and the timer hardware is also different (which would probably break gbc4nspire... again). And lastly, the interrupt controller is different, but I don't know of any currently released programs that mess with the interrupt controller directly.
-
I was going to post that some of the differences between the CX (probably CM as well ?) and Clickpad/Touchpad models could make it hard to run, on the CX, Ndless 2.0 programs for Clickpad/Touchpad running OS 1.7/2.0.1/2.1.0 on the CX. But calc84maniac posted first :)
FWIW, OSLauncher disables interrupts directly, through DC00000C.
-
I wonder how many colors the CX LCD can display at once? Was it 256 colors, 65536 like the PRIZM (switchable to 8 colors) or something higher?
-
16 million i suppose, or isn't it? as you can set real rgb colors
--edit--
it is certainly more than 256 :)
-
Well, setColorRGB(255,255,255) <- max value. Idk if it's really taken into account for each degree of precision.
-
Its a 16 bit screen I think, so 65536 (like the PRIZM).
-
Yes, the Nspire has a 16-bit screen. Among various things, the screen capture link command sends data in the same format as on the Clickpad & Touchpad, but the RLE is based on 32-bit chunks instead of 8-bit chunks (2 pixels in both cases).
Lua needs to perform some conversion from 24-bit to 16-bit, which probably does not help performance if the conversion is triggered in a tight loop... Maybe image loading in documents needs to perform such a conversion, too.
-
Ideally they'd only need to do the 24-bit to 16-bit conversion for the SetColorRGB() function. The overhead for Lua would probably overshadow that conversion anyway.
-
Certainly :)
The conversion can be done with 3 memory reads and few other (e.g. orr, lsl) operations, or even (but it's slightly less accurate than a good LUT) without any memory reads, by truncating the least 2 or 3 significant bits of R, G and B.
-
Also, the TI.Image format is only 15 bit, so I guess there will have to be conversion here too (maybe that is why its quite slow).
-
I would like to see a solid API for ndless 3.1 before it's released... :(
(I'm working on porting SDL, and possibly a GUI library to the Ndless platform for potential inclusion)
I'm fine with a release without one, especially with the potential on the CX :D
A port of SDL doesn't necessarily have to be integrated to Ndless, you could release it as an independent static library.
-
The keypad hardware reads inverted values compared to the non-CX calcs
I've tried to recompile mViewer, and I do confirm this problem.
isKeyPressed() should be update to return inverted results on CX hardware, if possible of course.