Omnimaga
Omnimaga => News => Topic started by: critor on April 17, 2011, 01:51:08 pm
-
The Nspire 3.0 OS doesn't open unencrypted TNS files anymore: you get the "this document format is not supported" dialog box.
Some unofficial softwares (TI-Nspire Text Editor, TNIC web) were generating unencrypted TNS files with text and images.
Those files aren't usable anymore on OS 3.0.
They probably did it on purpose: to import text or images on your Nspire, you now have no other way than using the official software, if you have a valid licence.
Starting from OS 1.2, TNS files were stored encrypted.
In OS 1.1, TNS files were stored unencrypted as XML files, exactly like with TI-Nspire Text Editor and TNIC web.
I've just tested: TNS files generated by the 1.1 OS aren't usable on the 3.0 OS.
So, in order to bother us, TI even took the right to break their own advertised ascending compatibility. :mad:
To get back at TI, we should really break the TNS encryption as soon as possible.
-
I agree with you about the fact that we must find a way to break the .tns encryption as soon as possible, but we're not experiencing any problem running the freshly created .tns files from lua code, and they happen do be unencrypted, so it means the Nspire can still read these file .... but I think that's because the nspire is only chacking if the Document.xml is encrypted.
If it is, it's all good, because it doesn't check if Problem1.xml is encrypted.
That's how we could make our LUA->TNS converters.
So, for now, the only solution to keep the old tns files compatible, woudl be to rewrite the Document.xml part with some script
-
I'm not going to be surprised if TI blocks the ability to use Lua tns files that are partially unencrypted... It seems like they're trying to do everything to stop us, even if it involves disadvantages to the customers (such as not being able to downgrade).
-
That's why I strongly recommend to NOT Update to an eventual soon update of the Nspire OS.
We must check if they did something to the lua compatibility first ....
-
Hehe, encrypted TNS? They must be joking. Simply taking a Java decompiler and running it through their program does wonders... :P (Aided with a C decompiler, if needed)
-
This majorly sucks, and I'd have a hard time believing that they needed to do that.
Some people among us do already know how to decompress data compressed with the unspecified 0D compression method. The ways of compressing data with the unspecified 0D compression method need to be extracted from where they are, and reimplemented :)
-
People as in people in the ndless community? Not sure about that... :P I'm pretty sure if the teacher software can do it, all we need to do is to rip apart the core and look for the key and compression stuff.
I'd rather not do that, but Google doesn't quite help with "0D" decompression... :(
-
Well, the periodic table written in Lua was translated into French (see the latest uploads to TI-Bank), so decompressing 0D-compressed data has been done :)
-
And we need to compress it back now...
-
That sucks. That text editor would have made a great Lua/C IDE with some modifications.
-
Question (maybe stupid):
Could that problem with TI-Nspire Text Editor and TNIC files be related to their Ndless fix?
(opening the installer file gives the same error message, instead of executing code...)
-
Indeed.
Note that an independent port of Lua would be much better than relying on TI's sub-par Lua. See http://ourl.ca/10458 :)
[that was a reply to willrandship, critor posted in-between]
-
but for now, at least all the progs will be easily ported.
If we're porting something anyways, why not port python? It's faster, and has more readable code, IMO.
-
The Python VM is not that fast, actually ;)
-
I didn't say it was amazingly fast, just that it's a little faster than Lua. Really, I want it for the awesome syntax :P not so much the speed.
Of course, you can't compare speeds of the two when you're talking about a whole different platform anyways.
-
I can understand that - but precisely, we need to watch out for the requirements of that complicated VM that yields mild performance, on a somewhat embedded platform like the Nspire :)
-
yeah, that's why I really want a C compiler for the calc :P much better than either.
-
I'm not going to be surprised if TI blocks the ability to use Lua tns files that are partially unencrypted... It seems like they're trying to do everything to stop us, even if it involves disadvantages to the customers (such as not being able to downgrade).
Note that if you send a document to OS 3.0.1, re-save it there, and send it back, it'll be fully encrypted and will work on any future OS ;)
-
Hmm.....
So, if we know the decryption process, why is it so hard to reverse it?
-
Some people among us do already know how to decompress data compressed with the unspecified 0D compression method.
err, the decryption is not "known". Just kind of hack in the built-in XML_parse function with the emulator
-
Moved to news. I think this is important, because this seems to be yet another attempt at breaking compatibility with third-party files.
I hope user-generated Lua files will not break in the CX version of OS 3.0 and future OSes... I wouldn't be surprised if TI discovered that we are using the Lua language and are trying to prevent us or discourage us from doing so.
-
Chances are they will eventually. But they have to learn from past experience, and from Apple and Sony (geohot) that we won't give up until we have what we want :D
The internet programmer/hacker community is a powerful force because we have a goal and something we really really really want.
-
It seems the CX is already available from some stores...
http://www.officedepot.com/a/products/840966/Texas-Instruments-TI-Nspire-CX-Graphing/ (http://www.officedepot.com/a/products/840966/Texas-Instruments-TI-Nspire-CX-Graphing/)
-
Now this is the final straw for me. I already knew about the instability and overall awfulness of TI's calcs but REALLY?!? Now they're trying to stop us from using third-party TNS?
/me facepalms
I'm going with Casio. 'Nuff said.
-
If we have a computer program that can encrypt it, then we should be able to reverse it to make a program for us to encrypt it.
I didn't say it was amazingly fast, just that it's a little faster than Lua. Really, I want it for the awesome syntax :P not so much the speed.
Of course, you can't compare speeds of the two when you're talking about a whole different platform anyways.
The little bit I've done with LUA on the PC it was awfully slow, I couldn't even do a simple memory copy 30 times a second.
-
Let's just hope :angel: TI is creating the SDK for Lua and will distribute it freely. It's not a matter of believing, but is what they should do... a custom IDE integrating the nspire player, saving lua files and
building packing into tns files, aall-in-one would be awesome to help programmers, but maybe plugins for gedit and notepad++ would be enough.
They wouldn't "give" us something to simply take out on the next version...
-
By the way, I've just newsed on TI-Bank, bashing at TI (in french):
http://ti.bank.free.fr/index.php?mod=news&ac=commentaires&id=1053
-
Now this is the final straw for me. I already knew about the instability and overall awfulness of TI's calcs but REALLY?!? Now they're trying to stop us from using third-party TNS?
/me facepalms
I'm going with Casio. 'Nuff said.
If the CX won't play third-party Lua files, I have planned to post a news article to start a community-wide boycott of the CX calculator, encouraging people to stick with the Touch/Clickpad, 84+ and Casio Prizm models and link to Cemetech's rant about the CX.If we have a computer program that can encrypt it, then we should be able to reverse it to make a program for us to encrypt it.
I didn't say it was amazingly fast, just that it's a little faster than Lua. Really, I want it for the awesome syntax :P not so much the speed.
Of course, you can't compare speeds of the two when you're talking about a whole different platform anyways.
The little bit I've done with LUA on the PC it was awfully slow, I couldn't even do a simple memory copy 30 times a second.
I wonder if it will be this slow on calc, though. So far, seeing the grayscale games that were done, I think it is at least decently fast. I haven,t tried them on-calc, though.By the way, I've just newsed on TI-Bank, bashing at TI (in french):
http://ti.bank.free.fr/index.php?mod=news&ac=commentaires&id=1053
Thanks. The community needs to know about TI actions.
-
This may or may not be on topic, but looking at the little Nspire cradles that TI has, it says that they automatically detect and update your OS. Oh, isn't that wonderful... So now we need a program to convince it we are in fact a 3.0 OS. or NdlessOS... I mean... Ndless 3.0.
-
I think this could possibly be a result of TI thinking nDoom's (relavitely) highly-publicized release caused schools to discourage the Ti-Nspire, and they want to reduce the risk of any further connections between their new calculator and gaming. As it is, schools and Doom have already had some bad history (Columbine?). I suppose they want a fresh start. After all, the Ti-83/84 has already been, from TI's perspective, irreparably "corrupted" by the sheer amount of third-party software for gaming/entertainment.
That's my speculation. Dunno how true it is... Most of it's probably after-the-fact coincidences.
-
If the CX won't play third-party Lua files, I have planned to post a news article to start a community-wide boycott of the CX calculator, encouraging people to stick with the Touch/Clickpad, 84+ and Casio Prizm models and link to Cemetech's rant about the CX.
The ti-nspire clickpad with OS 1.0
was the first step against the supposed "corruptness" of the ti-84 and other models (to a lesser extent) such as the ti-89.
The ti-nspire touchpad with OS 2.0
was the second step which restricted OS downgrading to prevent progress of the ndless progressing through the OSes
The ti-snpire CX blah with OS 3.0
...they've been planning this for a while, and as news of the calculator community finally hacking into the ti-nspire to the point where all you have to do is run the ndless document in OS 2.0/2.1 (which they originally put out there to stop ndless working on these OSes) to play games such as nDoom... the release of the CX was to once again make the calculator community play catch up again. the step up in version numbering shows this. Hopefully in time we will break this barrier again, and it remains to be seen whether TI just keeps releasing new OSes everytime the calculator community succeeds, or just give up and make development open.
So this boycott will at least show TI who has the power here. You release a new OS, we'll hack into it eventually.
Or more simply, just don't upgrade.
-
If we have a computer program that can encrypt it, then we should be able to reverse it to make a program for us to encrypt it.
inb4 Nkrypt! :D
.. or Ncrypt if your a grammar nazi :p
-
I waqsn't even thinking of getting a CX, but a boycott..... Idk.
-
Well a boycott would simply involves not buying it altogether nor supporting it. That way in long terms people would probably get the Prizm instead or other TI calcs, and TI couldn,t discontinue the older models because they sell well and not their new ones.
Also OS 3.0 seems to introduce bugs with some functions O.O
This is where I found the image and bug explanation (in French) http://tibank.forumactif.com/t6923-os-30-bug-equi-diff
-
You get the image much faster in the associated news:
http://ti.bank.free.fr/index.php?mod=news&ac=commentaires&id=1056
OS 3.0 == crap
-
Another new major bug with the 3.0 OS. The rechargeable battery status is not reported correctly (allways at 0%) which causes problems with USB transfers.
http://ti.bank.free.fr/index.php?mod=news&ac=commentaires&id=1057
Don't through away your rechargeable battery: it's still good!!!...
So they've removed the bugs that let Ndless run, but they've added many new bigger bugs for us :P
I'm beginning to think that this is the worst Nspire OS ever...
I wouldn't be surprised if it tries to explain me that 1+1=3 soon...
-
Darn this sucks. Why can't these companies test their crap more before releasing it?
At least it's not as bad as OS 2.71MP so far, though...
-
Well, it was designed to suck :P
I don't think it's quite fair to accuse too much, considering it was designed mostly for the CX, it's not as if they're going to get it perfect every time, and, IMO, you should compare it to 1.1, rather than 2.0 or 1.7.
Is the 84+ emu decent at all?
-
The 3.0 OS being just crap wouldn't be a problem if it wasn't installed allmost automatically and if you could uninstall it.
I have no proof, but I really think that there is some random bug in the boot2 update code since the 2.0.0 OS.
Check here -> http://tibank.forumactif.com/t6932-l-hecatombe-des-boot2#109963
3 different persons have reported deleted boot2 since last week (this hadn't been reported for months - we had several similar reports after the 2.0.0 OS release and then nothing else...).
When the boot2 is missing, the calculator stays stucked at the following screen:
(http://i23.servimg.com/u/f23/13/23/13/53/noboot10.gif)
Without the boot2, you cannot use the USB plug. You can only repair the calculator using my external RS232 interface.
Has someone any idea of what can trigger this major bug? Does the 3.0 OS really randomly destroy TI-Nspire calculators?
I'd need some proof before newsing...
-
We just got some kind of a proof!
For one of the erased boot2 reports, the user was just installing the 3.0 OS.
(the users didn't answer precisely or the other reports)
There is a random bug in the boot2 update code!!!
http://www.univers-ti-nspire.fr/forum/viewtopic.php?f=8&t=516&p=2596#p2596
-
Triple posting, but... it's really really bad!
I'm missing words to describe TI stupidity.
We got another answer on the topic:
http://www.univers-ti-nspire.fr/forum/viewtopic.php?f=8&t=516
The user could boot OS 3.0 one time but not 2 times, which is perfectly normal as the boot2 is only updated after the 1st OS successfull boot.
So something went wrong during the boot2 update and the boot2 was destroyed.
OS 3.0 is destroying calculators because of the new boot2 whose only goal is to prevent you from downgrading and installing Ndless.
Because of previous reports, it seems this boot2 update bug is present since at least OS 2.0.0.
You should not install OS 3.0...
And you MUST NOT install boot2 3.0! Use TNOC to save your calculator from dying!!!...
Does anybody have a clearer idea of what can go wrong during the boot2 update? Golat? Bsl?
-
Should we be going back to 2.0.1 right now?
I took boot2 3.0 out so am I safe?
-
I guess
-
This is real creepy ... TI should bug-test their software more (yes, we can dream)
Anyway, I think this is news worthy!
-
If TI bricks are calcs with their software can we get replacements for free or would we need to fix it on our own?
-
This is real creepy ... TI should bug-test their software more (yes, we can dream)
Anyway, I think this is news worthy!
So do I. What do the modos/admins say? Do we have enough proof?
I'm preparing the TI-Bank news.
-
Wow... This reminds me of Nintendo a bit. They once released an update for the boot2 on the Wii, and that update was supposed to replace all custom versions of boot2 that had been installed. Some Wiis were bricked. Everyone thought that this was because they had installed homebrew, but that was incorrect! The update procedure was simply so glitchy that it could brick Wiis. But from what I've seen, Nspire brickings now are much more common than what happened then... TI is getting worse than they already were.
-
If TI bricks are calcs with their software can we get replacements for free or would we need to fix it on our own?
If it's still under warranty, they'll repair/replace it for free. If not, I doubt they'll admit it's their fault.
That's why I've asked if someone who "understands" the decrypted OS was seeing any trace/proof of this major bug in the 3.0 OS boot2 update code.
-
I've just newsed on TI-Bank about the biggest bug in all graphic calculators history: the "Nspire 3.0 boot2 destroying bug".
http://ti.bank.free.fr/index.php?mod=news&ac=commentaires&id=1060
You want to kill a TI-Nspire? Just install OS 3.0! :P
-
Wow this is terrible. I'M glad I ran my OS through TNOC before upgrading.
-
Levak has just discovered that OS 3.0 is overclocking the hardware even further:
Core = 264MHz (240MHz with the 2.1 OS - 180MHz with previous OSes)
CPU = 132MHz (120MHz with the 2.1 OS - 90MHz with previous OSes)
AHB = 33MHz
It might be possible that the overclock does help the boot2 flashing failure to occur...
-
I suspect the problem is a race condition between the boot2 updater and the filesystem. There could potentially be filesystem access going on concurrently with a boot2 update and as far as I know there is no synchronization to prevent one from screwing up the other.
Edit: haven't been able to confirm this. The only two tasks accessing flash are the main task and a FlashFX background task. The boot2 updater is run from the main task, which has a higher priority than the FlashFX task. So even when inserting large delays into the boot2 updater, the FlashFX task doesn't do anything until long after the update is finished.
-
Wow. I bet the overclocking is because they wanted to make OS 3.0 look like it runs a bit faster, when in fact they didn't optimize everything.
Anyway I finally posted a news article: http://ourl.ca/10540
-
They try and stop the hacks the community have created up to os2.1 by enforcing more anti-downgrade protection in their NEW os but what they do is instead recreate a situation like os 2.0.0, where ti-nspire's were bricked. They better release the update soon, because the longer they leave it, the more chaos there will be.
So major upgrade (just to catch up with Casio, and force anti-downgrade) = Dead Ti-nspires
-
Indeed, and I hope the next OS update either contains an older Boot2 or that they fix it.