Good point, I was wondering if calcdude had a .Hex of CSX or something..... I think the OS is bad too

Bad as in corrupted, or bad as in poor quality? Keep in mind of course that none of these are full-fledged, completed OSes. They're all pet projects that didn't have enough time devoted to them to get them anywhere near a finished state.

Meh, it seems that I have to make the raw .HEX into an OS and then sign the formed OS. Other wise, error...

That sounds more or less correct; is that a problem? Surely you can compile the OS yourself, no?

Nope... I dont have the hex

You have the source code though, do you not?

nope, nvmd

I dont think I will be able to send it...

The reason you cannot use an existing, unsigned, .8xu intended for the 83+(SE) on and 84+(SE), even after signed, is that the OS file will still say that it's for the 83+(SE), even though you used the 0A key.
And qazz, I'll see if I still have it. If not, the CSX source is on ticalc.org (the reason I had to edit it is that brass doesn't like '.'s in label names)
Edit: Note that my above reason is only an educated guess, I don't really know

eh, I was thinking the same thing too


also, that OS, "BAOS" wont work on the TI-84+SE

thanks a lot man, if you could send me a .hex of CSX, I would be bery pleased

Actually, I can just give you the .8xu
linky: http://www.usaupload.net/d/5jehsuqq44a
Edit: Do note this will work on an 84+(SE) only.

err, thanks.. wasnt expecting that


/me now gives calcdude a cookie

I still like PongOS better...
CSX can receive files, but there aren't any programs for it anyway, not to mention it can't write flash.
BAOS showed promise, but it seems to have died, and the author used Makefile's instead of batch scripts, which means I won't build it for you
Edit: Actually, there seems to have been progress earlier this year, which is awesome http://bzc.sourceforge.net/baos/news.html

Bah, I already made BAOS work for my calc


it suxs

all it does is turn my calc on and off

Try it in WabbitEmu, perhaps utilizing rom8x. It's pretty amazing.

yeah, it works on wabbit


weird, CSX was being very button sensitive (as in one press of "o" wou lead to seeing a "ooooooooooooooooo" in the screen)

now that I did a few on+dels it works now

I wonder if there ever was a WFRNG OS for calcs...

@Qazz42 ouch, I hate when softwares/games does that. Seems like poor key detection design to me x.x.

dont worry DJ, the DEL+ON made it less sensitive, so now it is fine

Excuse me for the big up :/

I was trying to install CSX on a 83+, not a 84+, and I failed several times...

I read this topic, I didn't understand all...

You said the problem was that the OS is not signed, and that I could sign it myself in rabbitsign with this line, isn't it ? (0A ?)


Code:


Two questions:

1) I thought I read that I should have a "0A key" in the same directory.
What is that ? A key that I could download from brandonw.net ?

2) The calcs of the United TI's members in 2004 didn't verify the OSes before installing ? Devellopers never had tested thier OS before on a real calc ?

Thanks for your help, I have no experience in OS manipulation...

Time for a crash course on OS signatures and keys, I guess, with a bit of history.

The calculator requires that any OS or app loaded on it have a valid cryptographic signature. For apps we usually just use the 0104 key, which was publicly released by TI in conjunction with the 83+ SDK. In order to load an OS, we need an OS private key (04 on the 83+, 0A on the 84+), which TI would never release.

Back in 2009, Ben Moody (FloppusMaximus) made a cryptic post on UTI which turned out to be the private key for the 83+ line's OS signing key, which we call 04 (or 0004). That name comes from the key ID in the calculator's certificate.
Further efforts followed (see same thread) to factor the remaining private keys (notably 0A, which is the OS key for the 84+ line). TI (understandably so) didn't like that and demanded that the keys be taken down. The raw data remains available on wikileaks.

In response to this, TI began using 2048-bit keys on new calculators, which makes signature validation incredibly slow but is infeasible to factor on current computers.

In this thread, they signed the OS with the 0A key so it would validate on an 84+. Since you're trying to install it on an 83+, you need to use the 04 key.

Since you're using an 83+, you can probably skip the whole signing process though. On boot code 1.00 (and probably 1.01), you can simply remove a battery when the screen says "Validating" after sending an OS to force it to load. The calculator will restart and end up not validating the signature. This trick doesn't work on newer boot codes.

Thanks a lot for this explaination !
I didn't know.



I tried it, that's what the pdf says in the zip I downloaded from ti calc.

If I do that, the calc shutdowns, and I can't turn it back On.
The LCD never turns on, nothing happens when I press On, until I delete the OS by pressing [DEL] when I remove a battery.
Then comes the message "please install new software now", not CSX...

I will verify the version of my bootcode.

EDIT: Oh, sorry... It was just that I didn't know the user should turn On with [2nd][On]and not just [On], and the contrast was at the minimum.
But the CSX is corrupted, I think. I see the home screen a nanosecond, and then vertical lines appear and my screen turns to blue. Bad idea.