TI-83+/84+ OS glitches compilation thread (all OSes)

[Radical Pi]

I've gotten the Equ>String( glitch on my 84+SE with OS 2.30 o_o
I don't think it was caused the exact same way you did, but the results were definitely the same: huge equations show in memory, they turn out to be a gigantic mass of garble in the Y= editor, and then it just goes crazy

[DJ Omnimaga]

Really?

Did you have ASM programs on your calc tho? If not, then it could be that the glitch is present in this OS too. Could you try to recreate it again (with backing up first)?

[Radical Pi]

I didn't have any asm programs at the time, as far as I can remember. I don't even remember how I got the glitch, just that I did, at least twice. If I knew how to recreate it, I would try though.

[Builderboy]

I had it happen to me a couple of hours ago, i think it was caused by an Axe program though, since i was experimenting with Portals and things were going wrong.  The glitch was exactly how the screenshot looked though (the results anyway)

[Eeems]

Yeah I've had an error like that when playing with axe...it happened when I was writing to L5 or L6.

[DJ Omnimaga]

Edited the first post with a summary of the bug reported at http://ourl.ca/4593

[thepenguin77]

Good fail on TI's part. I have located the code that causes Err:Bad Address. TI made a beginner programming mistake. Here is their code:

Code: [Select]

;a is the page of the program
;b is the page of the last flash page available
;hl is the address of the program

cp b
jr z, somewhere
jr nc, error

ld bc, $4000
sbc hl, bc
jr c, error

Anyone see the error? The only way that execution proceeds is if carry is set. TI then doesn't clear the carry flag before subtracting. So what happens if HL = $4000?   $4000 - $4000 - 1 = carry = error.

The way to replicate the error is to fill a flash page up almost all the way and leave 11-18 bytes free. (11 for a 1 letter name, 18 for a 8 letter name) This way, when it it gets archived, the entire flash header gets left on one page and the program data is on the next.

I'll be working on a fix for 2.43 and 2.53.

[Happybobjr]

cool.

A couple of things with that ^

1.  make sure you can still use the patch (i am assuming you are making) if you have a patched os from brandonw
2. You should show ti this     see what they say.

[ztrumpet]

Awesome!  I can't wait to have it.

Can you please make this patch and the group one work on OS 1.19 too?  That'd make me incredibly happy.

[thepenguin77]

Ok, they're done. The fix works great, I managed to unarchive that program in the screenshot. When you run them, you'll get a little menu that lets you install/uninstall them.

Making them for 1.19 wouldn't be hard at all. I actually just made a template for OS mods, so it's as easy as plugging in a page, address, and new and old data. But as of right now, there are two things stopping me:

1. I don't have a 1.19 ROM
2. I don't know how to unlock flash on a 83+.

If I get both of those things, I should be able to do it. (Don't put ROM's or unlocking code in this forum. But PM's...  )

[Deep Toaster]

Yay! Another fix! Can't wait for 1.19...

Also, wow, TI.

[DJ Omnimaga]

wow nice fix. Also fail at TI

One thing, though, do not share ROMs via PM either. If you got links to any, please delete those PMs. E-mail, please. 1and1 just checked my site a few days ago to help me fix the 500 Internal Server Errors (in case it's on my side) so if they notice ROM links things could go bad...

[ztrumpet]

Awesome, thanks!  I hope you can figure out how to get them to work on the 83, though.

[DJ Omnimaga]

Btw can those be installed when Brandon's For() glitch patch and xLIB patch are installed?

[thepenguin77]

Hmmm... I would say yes. Most OS patches are just small rearranging of code. The xLib one would be to change about 20 different values. The For( glitch would either be a shift and insert of a small area, or a call to the end of the page where it does something then returns.

Either case, they shouldn't affect this fix. I only actually shift like 5 bytes.

The 1.19 fix is on the way. I've been having some troubles, (I killed about 20 wabbitemus.) But it's coming. It would have been done if it weren't for the fact that the flash paging ports of the 83+ don't AND the value of a flash page that doesn't exist. For instance, on the 84+ BE if you tell it to go to page $7D, it really goes to $7D & $3F = $3D. But on the 83+, if you tell it to go to $7D, it goes to like $02, which will not work.